Datasets:
keerthanapacific
/
sample

Dataset card Data Studio Files
xet
Community
1
pmcid
stringlengths
6
6
title
stringlengths
9
374
abstract
stringlengths
2
4.62k
⌀
fulltext
stringlengths
167
106k
file_path
stringlengths
64
64
516782
Encouraging physician appropriate prescribing of non-steroidal anti-inflammatory therapies: protocol of a randomized controlled trial [ISRCTN43532635]
Background Traditional non-steroidal anti-inflammatory drugs (NSAIDs) are a widely used class of therapy in the treatment of chronic pain and inflammation. The drugs are effective and can be relatively inexpensive thanks to available generic versions. Unfortunately the traditional NSAIDs are associated with gastrointestinal complications in a small proportion of patients, requiring costly co-therapy with gastro-protective agents. Recently, a new class of non-steroidal anti-inflammatory agents known as coxibs has become available, fashioned to be safer than the traditional NSAIDs but priced considerably higher than the traditional generics. To help physicians choose appropriately and cost-effectively from the expanded number of anti-inflammatory therapies, scientific bodies have issued clinical practice guidelines and third party payers have published restricted reimbursement policies. The objective of this study is to determine whether an educational intervention can prompt physicians to adjust their prescribing in accordance with these expert recommendations. Methods This is an ongoing, randomized controlled trial. All primary care physicians in Manitoba, Canada have been randomly assigned to a control group or an intervention study group. The educational intervention being evaluated consists of an audit and feedback mechanism combined with optional participation in a Continuing Medical Education interactive workshop. The primary outcome of the study is the change, from pre-to post-intervention, in physicians' appropriate prescribing of non-steroidal anti-inflammatory therapies for patients requiring chronic treatment. Three classes of non-steroidal anti-inflammatory therapies have been identified: coxib therapy, traditional NSAID monotherapy, and traditional NSAID therapy combined with gastro-protective agents. Appropriate prescribing is defined based on international clinical practice guidelines and the provincial drug reimbursement policy in Manitoba.
Background Traditional non-steroidal anti-inflammatory drugs (NSAIDs) are a widely prescribed class of therapy used to relieve pain and inflammation. The drugs have been shown to be effective for a variety of common disorders (hence their widespread use), most notably chronic osteoarthritis and rheumatoid arthritis. They are relatively inexpensive due to the available generic versions, but unfortunately have clinically important drawbacks related to their gastrointestinal (GI) toxicity [ 1 ]. Each year, about 1% to 1.5% of patients taking traditional NSAIDs experience serious GI side effects such as perforations, ulcers, and bleeding [ 2 - 5 ]. When multiplied by the total number of NSAID users this translates into significant patient morbidity and mortality [ 1 , 6 ] and is associated with considerable health care costs related to hospitalizations or to the prescribing of expensive gastro-protective agents (GPAs) [ 7 - 12 ]. The cause of this GI toxicity is the "non-selective nature" of traditional NSAIDs that block both cyclo-oxygenase-2 (Cox-2), an enzyme involved in the production of inflammation and pain, and Cox-1, a related molecule that functions in GI tract mucosal protection and platelet function [ 13 ]. In the last five years, a new class of non-steroidal anti-inflammatory agents has become available to physicians, fashioned specifically to be safer than the traditional NSAIDs but priced at least two to three times higher than the generic versions of traditional NSAIDs [ 14 ]. The new drugs preferentially inhibit Cox-2 enzymes as compared to COX-1 molecules, and therefore have been christened "Cox-2 selective inhibitors" or "coxibs" for short. Large clinical trials comparing the use of coxibs to traditional NSAIDs have lent support to the concept that the new agents offer an improved GI safety profile while maintaining comparable analgesic efficacy in patients with chronic arthritis [ 3 , 4 , 15 ]. The trials, however, have also hinted that the improved GI safety may be compromised by concomitant use of low-dose aspirin [ 4 ] and may come at the expense of some cardiovascular safety, although these data remain controversial [ 16 , 17 ]. Regardless, the introduction of these new anti-inflammatory agents has prompted the question: when is it appropriate and cost-effective to prescribe coxibs versus traditional NSAID monotherapy or traditional NSAIDs in combination with GPAs? To assist physicians in selecting from the different classes of anti-inflammatory agents, various scientific bodies have published clinical guidelines and third party payers have issued restricted reimbursement criteria [ 18 - 24 ]. These guidelines and reimbursement criteria in general propose that, for the treatment of chronic osteoarthritis and rheumatoid arthritis, coxibs should be used in lieu of traditional NSAID monotherapy when patients have an elevated risk for serious GI events. Traditional NSAID therapy in combination with certain GPAs (misoprostol or proton pump inhibitors) is also recommended as an alternative for most high-risk patients. High-risk patients are identified as individuals who have one or more of the following clinical characteristics: a history of peptic ulcer disease; advanced age (over 65 years); concomitant use of corticosteroids or anticoagulants; multiple comorbid conditions; or use of high doses or multiple NSAIDs [ 1 , 5 , 6 , 25 , 26 ]. These high-risk patients are considered to benefit from the improved GI safety of coxibs or from the GI protection afforded by GPAs. Furthermore, coxibs and GPA co-therapy have been shown to be cost-effective in such high-risk patients, as the increased cost of the drugs is partially offset by significant reductions in morbidity and mortality and related expenses [ 14 , 27 - 29 ]. With these guidelines and policies on prescribing in place, it is now important for physicians to adjust their prescribing practices accordingly. This is true both for rheumatology specialists and for physicians in the primary care setting where osteoarthritis patients are frequently managed. Unfortunately, experience and educational research show that simply making guidelines available does not elicit behavioural change from physicians [ 30 - 33 ]. In the case of anti-inflammatory drug prescribing, there is already evidence showing that despite the availability of guidelines and reimbursement policies, physicians' choices of drugs remain suboptimal both in terms of the inappropriate use of traditional NSAIDs and the non-cost-effective use of coxibs [ 34 - 38 ]. The current use of anti-inflammatory drugs suggests a need for strategies that will prompt physicians to change their prescribing practice in accordance with the expert recommendations. Preferably, strategies should be investigated that have proved successful at altering physician behaviour in other settings. In this paper, we describe a randomized controlled study, which we are presently conducting, to evaluate the impact of an educational intervention on primary care physicians' prescribing of non-steroidal anti-inflammatory therapies. The study, being conducted in the province of Manitoba, Canada, tests the hypothesis that this intervention will significantly improve physician appropriate prescribing of anti-inflammatory drugs in compliance with international clinical practice guidelines and Manitoba's restricted drug reimbursement policy [ 18 - 22 ]. The intervention being tested is modelled on proven approaches to changing physician behaviour; it consists of an audit and feedback mechanism with optional participation in a Continuing Medical Education (CME) interactive workshop. This manuscript summarizes Phases II to V of a larger initiative entitled the Manitoba Appropriate Anti-Inflammatory Utilization Initiative (MAAUI). The study began in November 2000 and is currently in the data analysis stage. In this paper, we relate in detail the protocol of this randomized controlled study. Methods Study population MAAUI is a province-wide population-based study. All primary care physicians (non-specialists) registered with the College of Physicians and Surgeons of Manitoba and practicing in Manitoba since July 1995 were eligible to enter MAAUI. The study excluded residents, new physician graduates, physicians registered as specialists, and physicians who started practicing in the province of Manitoba after July 1995. The eligible study population thus totalled approximately 884 physicians. All eligible physicians were automatically entered into the MAAUI protocol. Physicians who were randomized to intervention group of the research were given the option to withdraw from the study. Six family physicians who agreed to act as facilitators for the MAAUI CME workshop were not included in the study population. Study design The MAAUI study design is depicted in Figure 1 . Primary care physicians were allocated to the control and intervention arms of the study using a stratified randomization process. Specifically, study participants were divided into groups according to the geographical area of their practice within Manitoba. These groups were stratified by physicians' "urban" (within Winnipeg – the provincial capital, population size 676,700) or "rural" (outside of Winnipeg, population size 468,300) practice location. Randomization within each stratum was then carried out at the level of the group. The MAAUI research team originally identified 12 urban and 11 rural areas in Manitoba based on existing community boundaries within Winnipeg and the presence of distinct regional health districts outside of Winnipeg. Because one urban and rural area each contained very few primary care physicians (less than five), these areas were joined to neighbouring regions resulting in 11 urban and 10 rural groups. Urban and rural groups were randomized in such a way as to achieve a 1:2 ratio of control to intervention groups. Four of the urban groups and three of the rural groups were randomized to the control arm, and the remaining seven urban and seven rural groups were randomized to the intervention arm. This study design was chosen for several reasons. First, it offered the scientific rigour that is associated with randomization and the use of a control group. It also ensured that both rural and urban Manitoban physician groups had an equal opportunity to take part in and learn from the study intervention. The study design was also particularly amenable to testing an intervention that required physicians to congregate in common locations (i.e. for the CME workshop). The 1:2 ratio of study groups helped to ensure that there were a sufficient number of study participants who would consider attending the CME workshops. The educational intervention Physicians allocated to the intervention arm of MAAUI were mailed a package that included an introduction to the study, audit and feedback material, and an invitation to participate in a CME workshop. Physicians in the control arm received no package. Audit and feedback material The audit and feedback material consisted of a "Personalised Prescribing Profile" (see Figure 2 ). This profile illustrated for each physician his/her recent prescribing pattern of non-steroidal anti-inflammatory therapies and the appropriateness of this prescribing pattern in light of expert recommendations. Specifically, the profile contained a bar chart showing the number of patients to whom the physician had prescribed long-term treatment with coxibs, traditional NSAID monotherapy, or traditional NSAIDs in combination with GPAs between August 1999 and September 2000. The proportion of these patients with whom the prescription was appropriate was also included. Long-term treatment was defined as patients who received therapy for a minimum of 100 days during this period. The chart, in addition to showing the physician's own prescribing profile, also showed the average profile of primary care physicians in the same geographical area and the profile of Manitoban primary care physicians overall. The reverse side of the chart listed the criteria used to define appropriate prescribing. These criteria were based on international clinical practice guidelines and Manitoba's restricted drug reimbursement policy for coxibs, and were verified by a rheumatologist on the MAAUI research team [ 18 - 22 ]. The criteria identified coxibs or traditional NSAID/GPA combination therapy as the appropriate choice for patients at risk for serious GI complications. The criteria also identified coxibs as the only appropriate therapy for patients with bleeding disorders or patients taking concomitant anticoagulants (this latter recommendation was based in part on Manitoba's reimbursement policy for coxibs and in part on the Canadian Consensus guidelines, and reflected concerns over the anti-platelet effect of traditional NSAIDs [ 18 , 22 ]). The format of the Personalised Prescribing Profiles was developed by the MAAUI research team and was validated prior to the study intervention using two focus groups of family physicians. The Personalized Prescribing Profiles were generated by Manitoba Health and the Manitoba Centre for Health Policy (MCHP) at the University of Manitoba, using administrative data from the Population Health Research Data Repository. This repository contains anonymized encounter-based records of individuals' interactions with the provincial health care system and is derived from information received by the Department of Health, Province of Manitoba, as part of the routine provision of health care in the province. The repository includes Physician Registry files, Medical Claims and Hospital discharge files, as well as Drug Programs Information Network (DPIN) files. The DPIN contains records of all drugs dispensed by Manitoba pharmacies, regardless of who is responsible for payment. Because of the sensitive nature of the Personalised Prescribing Profiles, the individualized profiles were not seen by the MAAUI research team. Instead the packages containing the profiles were assembled and mailed out by Manitoba Health with the assistance of an independent researcher, hired by the MAAUI research team, who signed a confidentiality agreement with Manitoba Health. CME workshop The CME workshop for MAAUI was entitled "The Utilization and Prescribing of Anti-inflammatory Drugs in Osteoarthritis". The workshop focused specifically on the prescribing for osteoarthritis, as it was felt that people with this disease account for a large proportion of the chronic non-steroidal anti-inflammatory use by primary care practices [ 39 ]. Eleven CME workshops were held throughout Manitoba (four within Winnipeg and seven outside of Winnipeg) and physicians in the intervention group were invited to voluntarily attend one of the 11. The workshops were free of charge and were approved for 3.0 hours of MAINPRO-M1 credits (continuing education credits awarded by the College of Family Physicians of Canada for group learning activities). Each workshop was facilitated by a trained family physician. The workshops included three components: an introductory video, a decision tree, and a case study portion. The introductory video provided the history and rationale of MAAUI, explained aspects of the Personalised Prescribing Profile, and introduced the decision tree. The decision tree depicted a stepwise approach to the diagnosis and treatment of osteoarthritis. This aid was designed by a rheumatologist on the MAAUI research team specifically for use in the study and was based on international clinical practice guidelines and Manitoba's drug reimbursement policy [ 18 - 22 ]. A take-home copy of the decision tree was provided to each workshop participant (see Additional file 1 ). Finally, the case study portion of the workshop consisted of studies exploring both the diagnosis and appropriate treatment of osteoarthritis. These case studies were supplied from an existing CME curriculum entitled "Clinical Scenarios in Osteoarthritis" [ 40 ]. The overall MAAUI educational intervention, including both the audit and feedback material and the optional CME interactive workshop, was chosen by the MAAUI research team based on the success of similar interventions in the past. Audit and feedback mechanisms have shown moderate success in changing physician behaviour in past studies [ 41 ]. Equally, the addition of a continuing education workshop to an existing intervention has been shown to increase the impact on physician behaviour [ 42 ]. Workshops with interactive elements, such as case studies, have also been found to be more successful at influencing physician behaviour than purely didactic lectures [ 43 ]. Affordability, practicality, and reproducibility were also considered when designing the intervention. Outcomes Principal outcomes The primary goal of MAAUI is to evaluate the impact of the educational intervention on physician appropriate prescribing of non-steroidal anti-inflammatory therapies for patients requiring long-term treatment. To this end, four principal outcome measures have been developed: • the change in appropriate prescribing of all non-steroidal anti-inflammatory therapy (including coxibs, traditional NSAIDs, and traditional NSAID/GPA combination therapy) from pre-to post-intervention; • the change in appropriate prescribing of coxibs from pre-to post-intervention; • the change in appropriate prescribing of traditional NSAID/GPA combination therapy from pre-to post-intervention; and • the change in appropriate prescribing of traditional NSAID monotherapy from pre-to post-intervention. These primary outcomes will be evaluated at the level of the individual physician and will be calculated using administrative data similar to that used to create the Personalised Prescribing Profiles. The change in appropriate prescribing (AP) for a given physician will be calculated by the physician's rate of appropriate prescribing during a 6-month period post-intervention (October 1 st 2001 to March 31 st 2002) minus this rate during a 10-month period immediately pre-intervention (October 1 st , 2000 to July 31 st , 2001). The rate of appropriate prescribing will be defined as the number of patients prescribed long-term treatment in whom the treatment was appropriate, divided by the total number of patients prescribed long-term treatment, multiplied by 100%. Otherwise expressed: Physician's ΔAP = Post-intervention AP - Pre-intervention AP Appropriateness of prescribing will be calculated by a researcher who is blinded to participants' study-arm allocation, and will be based on the same criteria as listed in the Personalised Prescribing Profiles (see Figure 2a ). In the event that a physician has fewer than five patients on long-term treatment in either the pre- or post-intervention period, that physician's data will be withheld from analysis to ensure confidentiality of the patients. Secondary outcomes In addition to the four principal outcomes, the MAAUI study has a number of prospectively-defined secondary outcomes: • The study will re-evaluate the primary outcomes in the following two physician subgroups: physicians who received the audit and feedback material and attended the CME workshop; physicians who received the audit and feedback material and chose not to attend the CME workshop. This analysis will help to determine the impact that the different components of the educational intervention had on physician behaviour. Specifically in the subgroup of physicians who attended the CME workshop: • The study will measure the change in physician knowledge of osteoarthritis from pre-CME workshop to immediately post-CME workshop and the change in knowledge from pre-CME workshop to five months post-CME workshop. These two outcomes should provide insight into physicians' retention of information following a CME workshop and will allow us to examine any relationship between change in physician knowledge and change in physician prescribing behaviour. • The study will also survey physicians' perceived change in prescribing behaviour at five months post-intervention. This outcome, compared with the principal study outcomes, will provide insight into the accuracy with which physicians' discern their own prescribing practices. • The study will collect process-related measures in order to mark areas for improvement in the study interventions. Specifically, physicians' impressions of the Personalised Prescribing Profile and CME workshop will be surveyed using the questionnaire distributed immediately following the CME workshop and using focus-group discussions at the end of the study follow-up. All workshop participants will be invited to attend a follow-up focus group (one for urban participants and one for rural participants). A separate focus group will be held for the workshop facilitators. Baseline data collection The following data were collected at baseline (October 1 st , 2000 to July 31 st , 2001) as control measurements: i) physician demographics including sex, urban or rural practice location, and the number of years in practice; ii) physician volume of non-steroidal anti-inflammatory therapies prescribed annually for patients requiring chronic therapy; and iii) physician rate of appropriate prescribing of NSAIDs. These latter two measures were assessed for each of coxibs, traditional NSAID/GPA combination therapy, and traditional NSAID monotherapy. Sample size and statistics Sample size was calculated based on the need to detect at least a 10% improvement in physician practice patterns associated with the intervention, for clinical significance. Alpha error was set at 1% because of the multiple primary outcomes and power was set at 80%. It was also assumed in the sample size calculation that physician baseline demographic measures and baseline practice patterns would contribute an additional 10% variation to the outcomes. Based on these figures and using sample size techniques for multivariate regression analysis [ 44 ], we estimated the need for a minimum of 116 physicians in the control group and intervention group respectively. Multivariate regression analysis will be used to determine if the study intervention has a significant impact on the primary outcomes. Baseline data will be included as control measures and interaction effects between these baseline data and the intervention will also be evaluated. Ethics The MAAUI protocol has been approved by the Health Research Ethics Board, Faculty of Medicine, University of Manitoba, and by the Health Information Privacy Committee of Manitoba Health which is responsible for approving research projects that use personal health information held by a government department. Competing interests The MAAUI study is funded by Merck Frosst Canada Ltd., a pharmaceutical company that produces a range of non-steroidal anti-inflammatory therapies including a coxib and various traditional NSAIDs. Merck Frosst's involvement in the study is limited to representation (by L-JM) on the steering committee, which is the governing body responsible for approving all aspects of study development and implementation. Merck Frosst will not participate in data analysis and will not have access to confidential individual patient or physician data. AK, CM, and GT have received speaking fees from Merck Frosst Canada. KR is the scientific writer for this manuscript, employed by the PHCRU. All of these competing interests were not specifically disclosed to study participants. Merck Frosst's sponsorship of the study, however, was disclosed in the consent form signed by participants at the MAAUI CME workshops. Authors' contributions MD, AK, BK, and CM conceived of the study. MD, AK, L-JM, CM, and GT contributed to the development of the protocol. GT created the decision tree and list of criteria for appropriate prescribing. GT also chaired the MAAUI project steering committee. LM is project managing the study. KR performed the literature review for this manuscript and wrote the first draft and coordinated all subsequent revisions. All authors approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional file 1 Decision Tree: take-home copy The decision tree was a component of the MAAUI CME workshop. Physicians were introduced to the decision tree during the workshop and were provided with a two-sided take-home copy. The decision tree depicts a stepwise approach to the diagnosis and treatment of osteoarthritis. a/ Side 1. b/ Side 2. Of note, on side 2 of the decision tree the term "NSAID" denotes traditional NSAIDs. click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516782.xml
523856
Low autocrine interferon beta production as a gene therapy approach for AIDS: Infusion of interferon beta-engineered lymphocytes in macaques chronically infected with SIVmac251
Background The aim of this study was to evaluate gene therapy for AIDS based on the transduction of circulating lymphocytes with a retroviral vector giving low levels of constitutive macaque interferon β production in macaques chronically infected with a pathogenic isolate of SIVmac251. Results Two groups of three animals infected for more than one year with a pathogenic primary isolate of SIVmac251 were included in this study. The macaques received three infusions of their own lymphocytes transduced ex vivo with the construct encoding macaque IFN-β (MaIFN-β or with a vector carrying a version of the MaIFN-β gene with a deletion preventing translation of the mRNA. Cellular or plasma viremia increased transiently following injection in most cases, regardless of the retroviral construct used. Transduced cells were detected only transiently after each infusion, among the peripheral blood mononuclear cells of all the animals, with copy numbers of 10 to 1000 per 10 6 peripheral mononuclear cells. Conclusion Long-term follow-up indicated that the transitory presence of such a small number of cells producing such small amounts of MaIFN-β did not prevent animals from the progressive decrease in CD4 + cell count typical of infection with simian immunodeficiency virus. These results reveal potential pitfalls for future developments of gene therapy strategies of HIV infection.
Background Highly active antiretroviral therapy (HAART) effectively inhibits human immunodeficiency virus (HIV) replication, but it has been suggested that a combination of HAART and strategies for boosting the immune system would give more effective long-term control of HIV infection [ 1 , 2 ]. Interferon β (IFN-β) is an attractive candidate for such therapy: 1) it is a natural, potent antiviral protein that inhibits HIV at various stages of the viral cycle, from uptake to the release of virus particles [ 3 - 9 ]; 2) Type I IFNs display immunomodulatory properties that could improve the immune control of HIV replication [ 10 - 12 ]. During HIV infection, the induction of type I IFN production has been shown to be impaired in T cells and macrophages, which are considered to be the major targets of the virus [ 13 - 16 ]. However, the use of recombinant IFN in therapeutic strategies is limited by its poor bioavailability and the need for high doses to obtain an antiviral effect, resulting in deleterious side effects [ 17 ]. It has been suggested that the efficacy of type I IFNs for the treatment of HIV infection could be increased by developing a gene therapy strategy based on the modified production of IFN-β in genetically engineered lymphocytes [ 18 ]. For this purpose, a retroviral vector derived from Moloney murine leukemia virus, in which the human IFN-β coding sequence has been placed under the control of a fragment of the murine H2-K b gene promoter, has been used to ensure the continuous generation of low levels of IFN-β in transduced cells [ 10 , 19 ]. The transduction of peripheral blood lymphocytes (PBL) with this vector inhibits HIV replication in vitro and increases the survival of CD4 + cells in culture. Furthermore, IFN-β production in PBL from HIV-infected donors increases Th1-type cytokine production, improves cytotoxic responses against cells expressing HIV proteins, and the proliferative response to recall antigens [ 10 , 12 ]. These in vitro results have been confirmed in the SCID mouse model of HIV infection [ 20 ]. However, as the human-SCID mouse has a number of limitations as a model of AIDS, the efficacy and safety of this strategy should also be evaluated in a more appropriate model, such as macaques infected with simian immunodeficiency virus (SIV). SIV resembles HIV-1 and HIV-2 in its genomic organization and biological properties [ 21 ] and systematically causes a disease in macaques that is remarkably similar to AIDS in humans [ 22 ]. We have previously shown that PBL obtained from seronegative animals and transduced with a vector carrying the macaque IFN-β coding sequence placed under the control of a 0.6-kb fragment of the murine H2-K b gene promoter develop greater resistance to SIVmac251 in vitro [ 23 ]. In healthy seronegative macaques, infusion with autologous lymphocytes transduced ex vivo with the vector encoding IFN-β results in approximately 1 transduced cell per thousand peripheral blood mononuclear cells (PBMCs). The genetically modified cells were detected for at least 74 days after infusion, with no major side effects, in these experiments. Following infection with SIVmac251, macaques that had received the IFN-β construct infusion displayed lower peak plasma viral loads during primary infection than did control macaques. No adverse reaction was observed, and these macaques maintained high CD4+ T-lymphocyte counts for at least 478 days [ 24 ]. However, a gene therapy strategy for HIV infection would only be possible during the chronic phase of infection. At this stage, the immune system, and particularly CD4 + T cells – the major target of our gene therapy approach – may be strongly affected by the virus. We therefore investigated the safety and efficacy of this strategy in macaques chronically infected with a primary, pathogenic isolate of SIVmac251, but still in an asymptomatic state. The efficacy of our strategy has been examined according to two parameters. The eventual survival advantage of IFN-β transduced cells has been monitored by following the presence of such transduced cells in the blood stream as well as in the lymph nodes of infused macaques. This group of animals was compared to a controlgroup having received cells transduced with a retrovirus carrying a modified version of the MaIFN sequence with a deletion blocking mRNA translation. Animals were subjected to three infusions of autologous T lymphocytes transduced ex vivo with both constructs. The eventual clinical benefits of the presence of IFN-β-transduced cells have been monitored for two years, by examining, in both groups of animals, the absolute number of circulating CD4+ lymphocytes, cell associated viral load and plasma vial load. Results Status of animals before treatment The in vivo safety and anti-SIV efficacy of IFN-β-engineered lymphocytes in chronically SIV-infected macaques was assessed by following for two years animals that had received three infusions day 0, day 361 and day 613) of autologous T lymphocytes transduced with a construct encoding IFN-β (macaques IFN1, IFN2, IFN3) or, as a control, with a retrovirus carrying a modified version of the IFN-β that could not generate functional protein (macaques C1, C2, C3). These macaques had been infected with 4 AID 50 of a primary, pathogenic isolate of SIVmac251 more than one year before the start of the experiment. On day 0, the mean number of circulating CD4 + T lymphocytes was 767 ± 215 μl, and all animals had detectable SIV provirus in PBMCs (Table 1 ). Plasma SIV viremia was low or undetectable in most animals, the detection threshold being 1,500 copies of SIV RNA copies per milliliter of plasma. Table 1 Immunological and virological parameters of macaques at day 0 of the experiment. At the onset of the experiment, the six male cynomolgus macaques (Macaca fascicularis) were chronically infected by 4 AID 50 of a primary and pathogenic isolate of SIVmac251 for more than one year. They have been characterized for their mean number of circulating CD4+ T-lymphocytes, the time after SIV inoculation, and the cellular and plasma SIV viral loads. IFN group is represented by the three macaques that received their own cells transduced by the biologically active construct of IFN-β gene whereas the control group is represented by the three macaques that received their own cells transduced by the control construct. a: Immunophenotyping of Ficoll-purified PBMCs was performed by immunostaining with specific anti-CD4 and anti-CD8 antibodies, and analyzing by flow cytometry. The mean number of circulating CD4+ T-lymphocytes was determined at day 0 post first infusion with five points preceeding the onset of the experiment. b: Cellular viral load was estimated by a quantitative limit dilution nested PCR method allowing specific double amplification of a gagfragment of SIVmac251. Number of proviral copies was estimated by the last dilution that can display, in an agarose gel, a signal amplification. The number of SIVmac251 gag gene copies per 1 mg of DNA, for instance 131300 cells, was then brought back to a number of gene copies per 10 6 cells. c Plasma SIV viral load was determined by the branched-DNA method. Mean number of circulating CD4 + lymphocytes a Time after inoculation of SIVmac251 Mean number of SIV proviral DNA copies in PBMCs b (Copies per 10 6 cells) Plasma SIV load c (10 3 copies per ml) Mean +/- Standard Deviation Days Mean +/- Standard Deviation IFN IFN1 811 +/- 154 666 3.5 +/- 3.7 40 IFN2 767 +/- 215 686 116.6 +/- 170.0 <1.5 IFN3 977 +/- 254 1000 0.9 +/- 2.5 20 Control C1 1356 +/- 172 708 3.5 +/- 3.7 <1.5 C2 777 +/- 189 313 4.3 +/- 3.6 <1.5 C3 1055 +/- 478 1830 0.8 +/- 0.0 <1.5 Transduced PBLs After being transduced with the MFG-K b MaIFN-β and MFG-K b ΔMaIFN-β constructs, PBLs were readministered to the animal from which they were originally taken. Each macaque was infused with 10 8 to 4 × 10 8 lymphocytes. Semi-quantitative PCR analysis revealed that the mean transduction efficiencies for the transduction of PBL with the MFG-K b MaIFN-β and MFG-K b ΔMaIFN-β constructs were 10.33 % ± 7.42 % and 17.13 % ± 10.61 %, respectively. The IFN-β-transduced populations were characterized in culture by a low IFN-β production, ranging from 12 to 24 units per 5 × 10 5 cells per 3 days. We previously published that similar rates of Ma IFN-β-transduction results in a signficant reduction of SIVmac251 replication in vitro [ 23 ]. Such IFN-β-transduced cells remain detectable in the blood stream 485 days after reinfusion [ 24 ]. After the first inoculation (day 0), transduced cells was detected in peripheral blood, with about 10 transduced cells per 10 6 cells, for 14 days in macaque IFN1, and for 29 days in macaques IFN3 and C2 (Table 2 ). A transient peak of 1000 and 700 transduced cells per 10 6 circulating cells was observed in macaques C1 and C3, respectively. After completion of the series of infusions, with the last infusion occurring on day 613, transduced cells persisted at a low level (10 transduced cells per 10 6 cells) for only up to 60 days (Table 2 ). No transduced cells were detected at any time in the study for macaque IFN2. No significant difference was observed between the two groups of macaques in terms of transduced cell persistence (Table 2 ). The frequency of transduced cells was similar for CD4 + and CD8 + lymphocytes analyzed on day 673 (data not shown). No retroviral construct was detected in lymph nodes and splenic mononuclear cells. Table 2 In vivo follow up of transduced cells in blood. Absolute number of transduced cells per 10 6 PBMCs were evaluated by semiquantitative PCR amplification of IFN-β transgene in the two groups of animals. For in vivo f ollow up of transduced cells in blood from macaques, DNA samples of PBMCs were obtained at different dates following infusion of transduced PBL. This table indicates the minimum and maximum number of days following the first infusion of transduced cells in which the construct used was still detectable in PBMCs. Moreover, maximum transduction rate of PBMCs and detection treshold of the PCR method are indicated in the two groups of animals. IFN group is represented by the three macaques that received their own cells transduced by the biologically active construct of IFN-β gene whereas the control group is represented by the three macaques that received their own cells transduced by the control one. The relative intensity of the signals was compared to serial dilutions of lysate derived from plasmid-transfected cells that contained known numbers of IFN-β transgene copy per cell. a Day 0 is the first infusion day, other infusions occured at days 361 and 613. b Absolute number of transduced cells was below 10 per 10 6 PBMCs. 1st infusion a Days post-1st infusion 0 1 4 8 12 14 20 22 25 29 33 52 64 81 95 IFN IFN1 ND b 10 10 10 10 10 ND b ND b ND b ND b ND b ND b ND b ND b ND b IFN2 ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b IFN3 ND b 10 10 10 10 10 10 10 10 10 ND b ND b ND b ND b ND b Control C1 ND b 10 10 1000 100 10 10 10 10 10 ND b ND b ND b ND b ND b C2 ND b 10 10 10 10 10 10 10 10 10 ND b ND b ND b ND b ND b C3 ND b 10 700 10 10 10 10 10 10 10 ND b ND b ND b ND b ND b 2nd infusion a 3rd infusion a Days post-1st infusion 361 364 368 375 382 431 489 613 618 625 632 673 688 744 IFN IFN1 ND b 10 10 10 ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b IFN2 ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b ND b IFN3 ND b 10 10 10 ND b ND b ND b ND b 10 10 10 10 ND b ND b Control C1 ND b 10 10 10 ND b ND b ND b ND b 10 10 10 10 ND b ND b C2 ND b 20 20 20 ND b ND b ND b ND b 10 10 10 10 ND b ND b C3 ND b 10 10 10 ND b ND b ND b ND b 10 10 10 10 ND b ND b Clinical status and immunological follow-up Weight and rectal temperature remained fairly constant throughout the study (data not shown). No major variation in classical hematological parameters, including total lymphocyte and platelets counts, and hemoglobin concentration, was observed (data not shown). Immunological follow-up indicated that seven days after the first infusion (day 0), the number of circulating CD4 + lymphocytes significantly increased in all macaques studied (p < 0.05), except for C3. A similar significant increase (p < 0.05) was observed in the days following the second infusion (on day 311) for macaques IFN2, and C1, and following the third infusion (on day 613) for macaques IFN1, and C1 (Fig. 1A – 2A ). Figure 1 Evolution of immuno-virological parameters in SIVmac251 chronically infected macaques from the IFN group. Immunological and virological parameters were followed in macaques that received their own cells transduced by the retroviral construct allowing expression of the biologically active form of IFN-β. (A) Absolute number of circulating CD4+ lymphocytes was followed by immunophenotyping and flow cytometry. (B) Cell-associated viral load was estimated in PBMCs by a quantitative PCR method based on the specific amplification of the SIV gag gene. (C) Plasma viral load was estimated by a quantitative branched-DNA method based on the specific amplification of the SIV genome. Y axis split X axis at the first reinfusion date (D0) whereas black arrows indicate the second and third reinfusion dates. Figure 2 Evolution of immuno-virological parameters in SIVmac251 chronically infected macaques from the control group. Immunological and virological parameters were followed in macaques that received their own cells transduced by the deleted form of the retroviral construct. (A) Absolute number of circulating CD4+ lymphocytes was followed by immunophenotyping and flow cytometry. (B) Cell-associated viral load was estimated in PBMCs by a quantitative PCR method based on the specific amplification of the SIV gag gene. (C) Plasma viral load was estimated by a quantitative branched-DNA method based on the specific amplification of the SIV genome. Y axis split X axis at the first reinfusion date (D0) whereas black arrows indicate the second and third reinfusion dates. For all animals in both groups, absolute numbers of CD4+ T cells gradually decreased during the study (p < 0.05), and no significant difference in absolute numbers of CD4+ T cells was observed between the two groups of macaques (Fig. 1A – 2A ). The absolute numbers of circulating CD8 + T lymphocytes increased (p < 0.05) transiently during the days following each infusion of transduced cells, in both groups of macaques. However, with the exception of these peaks, absolute numbers of circulating CD8 + T lymphocytes did not change significantly during the study in any of the animals of either group (data not shown). Virological follow-up of animals We studied the course of SIV infection by determining the number of copies of SIV proviral DNA per cell, and the number of copies of SIV viral RNA per ml of blood. SIV provirus was detected in the PBMCs of all animals in both groups throughout the study. A transient and significant (p < 0.05) increase in cellular viral load was observed one to three weeks after each infusion in macaque IFN1 (Fig. 1B ) and in macaques C1, C2 and C3 (Fig. 2B ). A similar transient and significant (p < 0.05) increase in cellular viral load also occurred one to three weeks after the first and second infusions in macaque IFN2 and after the third infusion in macaque IFN3 (Fig. 1A ). Analysis of the number of SIV RNA copies in the plasma revealed that plasma viremia peaked (p < 0.05) one week after the first and the third infusions in macaque IFN1 (Fig. 1C ), after the first infusion in macaque IFN3 (Fig. 1C ) and after the second infusion in macaque C3, (Fig. 2C ). The other animals displayed no significant change in plasma viral load during the course of the experiment. Discussion In this study, we assessed the feasibility and efficacy of a gene therapy method based on the introduction into PBL of an IFN-β gene resulting in the constitutive production of low levels of IFN-β, in macaques chronically infected with SIVmac251. The present work was unable to bring new lighting on the efficacy of our gene therapy method since we encountered the problem of disappearence of transduced cells (control or IFN-β transduced cells) few days after each infusion. Throughout the study, significant, transient peaks of cell-associated and / or plasma viral loads were observed in most animals a few weeks after the infusion of transduced cells. These variations may reflect in vivo activation of viral replication, probably due to the infusion of activated cells. This phenomenon was also observed after the infusion into SCID mice of transduced human PBLs, resulting in up-regulation of CCR-5 HIV co-receptor expression in human CD4 + T cells [ 27 ]. Indeed, the SIVmac251 isolate used in our experiment is a CCR5-dependent virus, and its replication may have been activated by upregulation of the CCR-5 coreceptor after infusion. However, gene therapy strategies for the treatment of HIV infection could only be envisaged in combination with HAART. In this context, the activation of host virus replication observed after the infusion of transduced cells would be overcome by HAART treatment. The mean rates of transduction of PBL isolated from macaques chronically infected with SIVmac251 were 10.33 % ± 7.42 % and 17.13 % ± 10.61 % for the MFG-K b MaIFN-β and MFG-K b ΔMaIFN-β constructs, respectively, which is similar to the transduction efficiency previously reported for PBLs isolated from healthy non infected macaques [ 23 , 24 ]. The transduction efficacy for lymphocytes from healthy donors and HIV-seropositive patients has also been found to be similar [ 10 ], indicating that chronic infection does not affect the retroviral transduction of lymphocytes. After the first infusion, small numbers of engineered cells (control and IFN-β-transduced cells) were detected for only 29 days. Thus, the persistence of transduced cells in chronically infected macaques was lower than that previously reported in non infected macaques, in which IFN-β-engineered cells were detected for at least 70 days, and for more than a year after SIVmac251challenge [ 24 ]. This former study indicates also that immune response that may be induced by mouse cell components or FCS present in culture medium may not alter persistence of genetically modified immune cells. We carried out three infusions of engineered cells and, after each infusion, the engineered cells disappeared from the bloodstream within a few days. Poor persistence of circulating engineered cells has been reported in HIV-infected macaques and in SCID mice, and has been attributed to the delocalization of circulating transduced cells in the lymph nodes [ 28 ], and intestine [ 29 ]. In our study, we detected no engineered cells in the lymph nodes or spleen, indicating that the delocalization of transduced cells to these organs could not account for the absence of transduced cells in the blood. The short-term persistence of transduced cells has already been reported in other studies in which autologous engineered T cells were cleared rapidly from the bloodstream [ 30 ]. However another group reported the persistence of engineered cells for more than 25 weeks (0.1 to 10% of PBMC) in HIV-infected patients [ 29 , 31 , 32 ]. They hypothesized that the higher rate of T-cell survival was due to ex vivo stimulation through CD3 and CD28. Indeed, it has been demonstrated that the inhibition of HIV replication in CD3- CD28- stimulated CD4 + cells is due to the production of cytokines associated with Th-1 function [ 33 ] and to the downregulation of CCR-5 expression [ 34 ]. Thus, in our study, the disappearance of transduced cells may be due to ConA-stimulation, which may induce apoptosis in lymphocytes, as previously described [ 35 ]. IFN-β-producing cells and cells transduced with the control vector displayed similar levels of in vivo persistence. We previously reported higher levels of resistance to HIV in vitro following the transduction of human CD4 + T cells [ 19 ], human macrophages [ 36 ] and macaque PBL [ 23 ] with a construct encoding IFN-β. However, Vieillard et al. [ 10 ] reported inefficient protection of transduced lymphocytes against HIV replication in vitro for PBLs isolated from patients in an advanced state of HIV infection. This lack of protection probably resulted from the downregulation of interferon alpha/beta receptor expression in donors with AIDS, leading to hyporesponsiveness to type I IFN [ 37 ]. Thus, although we selected animals with CD4 + cell counts that were still high, the disease may have been so advanced that transducing PBLs with a construct encoding IFN-β had little effect, with the engineered lymphocytes subjected to the high rate of lymphocyte turnover observed during SIV infection [ 38 , 39 ]. Our previous work with the macaque model encouraged us to develop low-level autocrine IFN-β production as an approach to gene therapy for AIDS. The persistence of 1 transduced cell per 10 3 circulating cells before SIV challenge was correlated with low plasma virus load and the maintenance of CD4 + and CD8 + cell counts in macaques infused with the construct encoding IFN-β [ 24 ]. In this study, performed with animals infected for more than one year, cells transduced with the IFN-β construct rapidly disappeared from the bloodstream after infusion. This suggests that gene therapy by PBL transduction should be performed as soon as possible after primary infection. We are well aware that the number of transduced lymphocytes was too small for a major effect in this study and we believe that further exploration of IFN-β-based anti-HIV therapy will require the construction of high-titer vectors, with the aim of increasing the proportion of vector-transduced HIV target cells. An alternative method for IFN-β gene therapy involves the transduction of CD34 + hematopoietic stem cells. This method has been proposed for the treatment of HIV infection [ 40 , 41 ]. The transduction of these cells, which are able to generate all the main HIV target cells, will increase the proportion of transduced cells, extend IFN-β production to macrophages and dendritic cells, and should facilitate long-term expression of the therapeutic construct. We have already demonstrated that macrophages transduced with an IFN-β construct display enhanced HIV resistance, and that HIV transmission to CD4 + T cells is prevented in IFN-β-transduced dendritic cells [ 42 ]. We intend to investigate the possibility of transducing hematopoietic stem cells to inhibit viral replication in macaques chronically infected with SIVmac251, in the near future. Methods Animals Six male cynomolgus macaques ( Macaca fascicularis ), weighing between 3 and 7 kg, and negative for herpes B, filovirus, STLV-1, SRV-1, SRV-2, SIV, and hepatitis-B were used in this study. Before all experimental procedures, animals were anesthetized with chlorhydrate ketamine (Cenravet, France), and all procedures were conducted according to European guidelines for animal care (Official Journal of the European Communities L538, 18 December 1986). Macaques were housed in individual cages in biosafety level 3 facilities, as required by national regulations (Commission de Génie Génétique, Paris, France). Viral stock More than 300 days before infusion with the IFN construct, macaques were intravenously infected with 4 AID 50 of a primary, pathogenic SIVmac251 isolate. This virus stock was obtained by coculturing splenocytes obtained from an infected rhesus macaque with rhesus macaque PBMCs (Dr. R.C. Desrosiers, Harvard Medical School, MA, USA), and was amplified by a second passage on rhesus PBMCs (prepared and kindly provided by Dr. A.M. Aubertin, Université Louis Pasteur, Strasbourg France). Retroviral vectors The MFG-K b MaIFN-β retroviral vector used in this study has been described elsewhere [ 23 ]. It contains the macaque IFN-β coding sequence placed under the control of a 0.6 kb fragment of the murine H2-K b gene promoter, resulting in the continuous production of low levels of a biologically active macaque IFN-β. The MFG-K b ΔMaIFN-β retroviral vector used in this study as a control has been described elsewhere [ 23 ]. It contains a macaque IFN-β coding sequence with a 530 bp deletion, blocking IFN-β translation, under the control of the same promoter region. Vectors (MFG-K b MaIFN-β and MFG-K b ΔMaIFN-β were produced with two Ψ-CRIP packaging clones, each of which produced 2 × 10 5 infectious particles per ml, with no detectable replication-competent helper virus [ 23 ]. The Ψ-CRIP cells were maintained in Dulbecco's modified Eagle's medium (DMEM, InVitrogen, Grand Island, New York, USA) supplemented with 10 % heat-inactivated bovine serum (BS) (InVitrogen) and 0.2 μM antibiotics (penicillin / streptomycin / neomycin, PSN, InVitrogen). Isolation of macaque peripheral blood lymphocytes (PBL) Three macaques (IFN1, IFN2 and IFN3) received infusions of their own lymphocytes transduced with the biologically active MaIFN-β construct. Another three macaques (C1, C2, C3) were infused with their own lymphocytes transduced with the construct carrying the deleted form of the MaIFN-β, which cannot produce a translatable mRNA. We collected about 100 ml of blood from each macaque into heparin lithium tubes (Greiner, USA). Buffy coats were obtained by centrifugation (170 g / 15 min). Mononuclear cells were collected, and centrifuged (400 g / 30 min) on a Ficoll density gradient (Eurobio, Les Ulis, France). Plasma and erythrocytes, diluted 1 in 2 with 0.9% NaCl (InVitrogen), were washed and used immediately for infusion into the macaques. Transduction of macaque PBLs Isolated PBMCs (10 6 cells per ml) were activated by incubation for three days in RPMI-1640 medium, 10 % fetal calf serum (FCS), 2 mM L-glutamine (Bœhringer Mannheim, Mannheim, Germany), 0.2 μM antibiotics (penicillin / streptomycin / neomycin), 5 μg / ml concanavalin A (InVitrogen). Activated PBL were resuspended in transduction medium consisting ofn 45 % DMEM, 45 % IMDM (InVitrogen), 5 % FCS, 5 % BS, 4 μg / ml protamine sulfate (Sigma, Saint Louis, USA) and 20 IU / ml recombinant human (rHu) IL-2 (Bœhringer Mannheim). Cells were transduced by coculture for three days with subconfluent Ψ-CRIP packaging cells. At the end of the coculture period, the various cell populations were transferred twice to other culture plates to eliminate any residual adherent packaging cells. Transduced lymphocytes were washed, resuspended in 1× PBS at a concentration of 10 7 cells / ml, and injected intravenously into macaques. Transduction efficacy was estimated with transduced PBLs maintained in culture for 3 days. Evaluation of the transduction rate DNA was extracted from macaque PBMCs and the amount used for each sample was normalized based on data for amplification of the β-globin gene, using 5'-ACCATGGTGCTGTCTCCTGC-3' as sense primer, and 5'-CATGGCCACGAGGCTCCA-3' as an antisense primer. Both retroviral sequences were detected, using 5'-GTTCAGGCAAAGTCTTAGTC-3' as the sense primer, binding in the H2-K b gene promoter and 5'-TGAAGATCTCCTAGCCTGT-3 as the antisense primer, binding in the macaque IFN-β coding sequence. These primers amplified a 870-bp fragment from the MFG-K b MaIFN-β vector, and a 340-bp fragment from the MFG-K b ΔMaIFN- vector The PCR amplification products were identified by dot-blot hybridization with an IFN-β probe, and quantified with a PhosphorImager (Molecular Dynamics, Sevenoaks, England, UK), as previously described [ 19 ]. Relative signal intensity was compared with the signal intensity of serial dilutions of lysate derived from plasmid-transfected cells containing known numbers of transgene copies per cell. The detection threshold of the PCR assay used was estimated and found to be one copy of the IFN-β gene per 10 5 cells. Hematological and immunological follow-up of infused macaques All infused animals were followed during the months preceding the study, and for more than 700 days after the first autologous infusion. We carried out hematological analysis, and monitored weight, rectal temperature, and levels of lymphocytes transduced with the IFN-β construct. Blood formula and blood cell counts were determined with an automated hemocytometer (Coulter Corporation, Miami, USA). Axillary lymph nodes and spleens were removed from animals and ground in 1× PBS using a Potter homogenizer. Lymph nodes and splenic mononuclear cells (LNMC, SMC) were then collected and centrifuged (400 g / 30 min) on a Ficoll cushion (Eurobio, Les Ulis, France). DNA extraction and evaluation of in the rate of transduction of LNMC and SMC were performed as previously described. In vivo immunological follow-up of macaques receiving infusions We estimated the proportions of the various subtypes of circulating PBMCs by direct immunofluorescence assay (anti-CD3 clone FN18, Biosource International, CA, USA), anti-CD4 clone Leu 3a PE (Becton Dickinson, San Jose, Mountain View, CA, USA), anti-CD8 clone Leu 2a FITC (Becton Dickinson) antibodies and IgG isotypic controls (Immunotech, Marseille, France), and flow cytometry (Becton Dickinson). We used specific software (CellQuest, Becton Dickinson) as previously described [ 25 ] for the analysis. Sorting of CD4 + and CD8 + circulating lymphocytes Mononuclear cells isolated on Ficoll-Hypaque were positively separated using CD4-specific and CD8-specific immunomagnetic microbeads (MiniMACS, Miltenyi, Stadt, Germany) according to manufacturer's instructions. Subset purity was evaluated by flow cytometry, using secondary anti-CD4 clone OKT4-PE (Dako, Glostrup, Denmark) and anti-CD8 clone DK25-FITC (Dako) antibodies. The rates of transduction of the sorted CD4 + and CD8 + lymphocytes were evaluated, as described above. Plasma and cell-associated viral load Levels of SIV RNA in plasma were determined with the SIVmac-branched-DNA assay, using a detection threshold of 1,500 mEq per milliliter of plasma (Chiron Diagnostics, Amsterdam, The Netherlands). DNA was extracted from PBMCs with an extraction kit (Roche Diagnostics GmbH, Mannheim, Germany). Levels of SIV DNA in cells were determined using a two-step PCR method with two external gag -specific primers (1386-5': GAAACTATGCCAAAAACAAGT and 2129-5': TAATCTAGCCTTCTGTCCTGG) and two internal gag -specific primers (1731N 5': CCGTCAGGATCAGATATTGCAGGAA and 2042C 5': CACTAGCTTGCAATCTGGGTT), as previously described [ 26 ]. Statistical analysis Statistical significance was determined by paired or unpaired non parametric Wilcoxon and Mann-Whitney tests adapted for small samples. Competing interests The authors never received reimbursements, fees, funding, or salary from an organization that may in any way gain or lose financially from the publication of this paper in the past five years. The authors never any stocks or shares in an organization that may in any way gain or lose financially from the publication of this paper. The authors never have any other financial competing interests. The authors have no non-financial competing interests to declare in relation to this paper. Authors' contributions WG was the major contributor to this paper. EL participated in the design of the study and performed the cell cultures and transduction experiments. BB and JL participated in the animals manipulation. FM participated in the preliminary experiments. SP performed all PCR reaction for transduced cells in vivo follow-up. DD and EDM participated in the design and the coordination of the study. RLG performed the statistical analysis and participated in the design and the coordination of the study.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC523856.xml
546189
Saturated free fatty acids and apoptosis in microvascular mesangial cells: palmitate activates pro-apoptotic signaling involving caspase 9 and mitochondrial release of endonuclease G
Background In type 2 diabetes, free fatty acids (FFA) accumulate in microvascular cells, but the phenotypic consequences of FFA accumulation in the microvasculature are incompletely understood. Here we investigated whether saturated FFA induce apoptosis in human microvascular mesangial cells and analyzed the signaling pathways involved. Methods Saturated and unsaturated FFA-albumin complexes were added to cultured human mesangial cells, after which the number of apoptotic cells were quantified and the signal transduction pathways involved were delineated. Results The saturated FFA palmitate and stearate were apoptotic unlike equivalent concentrations of the unsaturated FFA oleate and linoleate. Palmitate-induced apoptosis was potentiated by etomoxir, an inhibitor of mitochondrial β-oxidation, but was prevented by an activator of AMP-kinase, which increases fatty acid β-oxidation. Palmitate stimulated an intrinsic pathway of pro-apoptotic signaling as evidenced by increased mitochondrial release of cytochrome-c and activation of caspase 9. A caspase 9-selective inhibitor blocked caspase 3 activation but incompletely blocked apoptosis in response to palmitate, suggesting an additional caspase 9-independent pathway. Palmitate stimulated mitochondrial release of endonuclease G by a caspase 9-independent mechanism, thereby implicating endonuclease G in caspase 9-indpendent regulation of apoptosis by saturated FFA. We also observed that the unsaturated FFA oleate and linoleate prevented palmitate-induced mitochondrial release of both cytochrome-c and endonuclease G, which resulted in complete protection from palmitate-induced apoptosis. Conclusions Taken together, these results demonstrate that palmitate stimulates apoptosis by evoking an intrinsic pathway of proapoptotic signaling and identify mitochondrial release of endonuclease G as a key step in proapoptotic signaling by saturated FFA and in the anti-apoptotic actions of unsaturated FFA.
Background Recent evidence suggests that intracellular accumulation of saturated free fatty acids (FFA) in vascular cells contributes to lipid-mediated cellular damage (see [ 1 - 4 ] for review). The cellular dysfunction associated with FFA overload, known as lipotoxicity, contributes to cell injury in settings of high FFA or triglycerides, such as obesity or type 2 diabetes [ 4 ]. Diverse mechanisms have been proposed to explain lipotoxicity including dysregulation of cell signaling, induction of a proinflammatory and prothrombotic state, or in some cases programmed cell death [ 1 , 4 ]. Indeed, saturated FFA have previously been shown to induce apoptotic cell death that is prevented in most cell types by unsaturated FFA [ 5 - 14 ]. In the microvasculature, the pro-apoptotic signaling pathways induced by saturated FFA and the anti-apoptotic pathways regulated by unsaturated FFA remain incompletely understood. Several mechanisms have been implicated in apoptotic cell death induced by saturated FFA. Some studies suggest that increased β-oxidation of FFA does not contribute to apoptotic cell death and suggest that unmetabolized FFA might be involved [ 1 , 13 ]. However, other studies contradict this observation and suggest a direct role for mitochondrial β-oxidation in the apoptotic response to palmitate.[ 10 ]. The finding that long-chain saturated but not unsaturated FFA cause apoptosis implicates a product made specifically from the saturated species. For instance, saturated but not unsaturated FFA are precursors for the pro-apoptotic lipid ceramide. Although palmitate does increase de novo ceramide synthesis in cultured cells, studies of the functional role of ceramide in palmitate-induced apoptosis have yielded conflicting results that might depend on the cell type in question [ 8 , 13 , 14 ]. Because saturated FFA are poor substrates for cardiolipin biosynthesis, decrements in cardiolipin and increased release of mitochondrial cytochrome-c have recently been implicated in apoptosis in breast cancer cells and cardiomyoctyes exposed to palmitate [ 12 , 13 ]. Another recent study demonstrated mitochondrial release of cytochrome-c in palmitate-treated pancreatic β-cells [ 14 ], which suggests that an intrinsic mitochondrial pathway of pro-apoptotic signaling might mediate the effects of saturated FFA on cell death. In the present study, we investigated the hypothesis that the saturated FFA palmitate induces apoptosis in microvascular mesangial cells and delineated the proapoptotic signals involved. We chose to study mesangial cells because lipids accumulate in mesangial cells in vivo in experimental models of type 2 diabetes, obesity, or hyperlipidemia [ 15 - 20 ], but the functional consequences of FFA accumulation are unclear. We report here that palmitate induces an intrinsic proapoptotic signaling pathway in mesangial cells that proceeds by a caspase 9-dependent pathway and by a caspase 9 -independent mechanism involving mitochondrial release of endonuclease G. In addition, we demonstrate that unsaturated FFA block both the caspase 9-dependent and -independent pathways of palmitate-stimulated apoptosis. Methods Reagents Antibodies used in these studies were as follows: human-specific active fragment of caspase 9 and human cytochrome-c (Cell Signaling, Beverley MA,), human active fragment caspase-8 and caspase-2 (BD Biosciences), endonuclease G (Chemicon, Temecula, CA), and β-Actin (Sigma, #A5316). Cell-permeable inhibitors of caspase 9 (Z-LEHD-FMK) and caspase-8 (Z-IETD-FMK) were from R&D systems (Minneapolis, MN). Etomoxir and 5-aminoimidazole-4-carboxamide-1-β-D-ribonucleoside (AICAR) were from Sigma and Toronto Research Chemicals (Ontario, Canada), respectively. Preparation of FFA-albumin complexes Fatty acid-albumin solutions were prepared by the protocol of Spector [ 21 ]. Briefly, sodium salts of FFA (Nu-Chek Prep, Elysian, MN) were added to PBS and gently warmed to facilitate solubility without damaging the fatty acid [ 21 ]. The warm, clear fatty acid salt solution was complexed to 5% fatty acid-free BSA in PBS at a 6:1 fatty acid to BSA molar ratio. The sterile filtered, complexed fatty acid solution was added to the serum-containing cell culture medium to obtain the indicated final FFA concentration. The final FFA concentration in the medium was confirmed with an enzymatic colorimetric assay (NEFA C, Wako). We also confirmed that addition of the complex to culture medium did not significantly alter the pH. Apoptotic cell death in cultured human mesangial cells Human mesangial cells (HMC), purchased from Cambrex Bioscience Inc. (Walkersville, MD), were maintained in Dulbecco's modified essential medium (Gibco-BRL) supplemented with 17% fetal bovine serum (FBS), 100 U/ml penicillin, 100 μg/ml streptomycin, 5 ng/ml selenite, and 5 μg/ml each of insulin and transferrin. Characterization was performed by phase contrast microscopy and by immunostaining for intermediate filaments and surface antigens as described previously [ 22 ]. Briefly, cells were positive for desmin, vimentin, and myosin, but did not stain for factor VIII, keratin, or common leukocyte antigen. To measure endogenous levels of cleaved caspase-3, cells were lysed in a buffer containing 20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 % Triton X-100, 2.5 mM Na pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na 3 VO 4 , and 1 μg/ml luepeptin. After adjusting for cell protein (DC Assay, BioRad, Hercules, CA), the amount of cleaved human caspase-3 (Asp 175) was measured by ELISA (Cell Signaling). For Western blotting of cleaved caspase proteins, the cells were washed with ice cold PBS and scraped in CHAPS extraction buffer(50 mM Pipes/HCI, pH 6.5, 2 mM EDTA, 0.1% Chaps, 20 μg/ml leupeptin, 10 μg/ml pepstatin A, 10 μg/ml aprotinin, 5 mM DTT, 2 mM Na pyrophosphate, 1 mM Na 3 VO 4 , and 1 mM NaF) and centrifuged at 2,000 × g for 10 min at 4°C. Protein content in the supernatant was assayed with the DC protein assay. An aliquot of the lysate (25 μg protein) was boiled in SDS sample buffer, resolved on a 4–12% SDS-PAGE gradient gel, and transferred to a 0.2 μm nitrocellulose membrane. After blocking in 5% non-fat dried milk in TBS-T (20 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.05% Tween 20) for 1 h, the membrane was washed 3 times with TBS-T for 5 min each and incubated overnight at 4°C with primary antibody in 3% BSA in TBS-T. After incubating with suitable HRP-labeled secondary Ab (1:2,000) and extensive washing, the proteins were detected by chemiluminescence with an average exposure ranging from 10–30 sec. As previously described [ 23 ], the Western blots were analyzed by densitometry in NIH Image by normalizing values for the relevant caspase fragment to the highest value within each experiments (maximum level = 1). To quantify the number of pyknotic nuclei, HMC on coverslips were washed once with PBS and fixed for 20 min with freshly-prepared 3.7% formaldehyde/20% sucrose in PBS. After washing twice with PBS the HMC were stained with 5 μg/ml Hoechst 33342 (Molecular Probes, Eugene OR) and mounted in Slow Fade Light (Molecular Probes). Using a Nikon Diaphot microscope, the number of pyknotic nuclei were counted and expressed as a percentage of the total number of nuclei counted (n = >300 nuclei per condition). DNA fragmentation was assessed by measuring release of nucleosomal fragments into the cytosol (Cell Death Detection ELISA Plus, Roche). Briefly, HMC in 24-well plates were centrifuged in situ for 10 min at 200 × g and the supernatant gently removed. The monolayer was incubated in lysis buffer for 30 min at room temperature and centrifuged again at 200 × g . The supernatant (i.e., cytosolic fraction) was assayed immediately for nucleosomal fragments. Enzymatic assay of caspase 9 in HMC HMC treated with FFA were lysed (20 mM Tris, pH 7.5, 150 mM NaCl, 1.0% Triton X-100) and frozen at -40°C. Equivalent amounts of total HMC protein were added to buffer containing the LEHD caspase 9 peptide substrate linked to a cleavable luciferase substrate, aminoluciferin (Promega). The amount of light produced in a coupled reaction with luciferase was measured once every hour for 3 hours in a Berthold Luminometer. Experiments with increasing amounts of cell protein confirmed that the assay was in the linear range under the conditions described. Measurements of cytochrome -c and endonuclease G redistribution To analyze cytochrome-c redistribution in HMC treated with FFA, cells were fractionated into cytosol and membrane fractions using 0.05% digitonin in an isotonic sucrose buffer exactly as described by Dong and coworkers [ 24 ]. Because cytochrome-c release occurs mostly from mitochondria, Western blot analysis of cytosol and membrane fractions is expected to reflect cytochrome-c translocation from mitochondria to the cytoplasm. In separate experiments, the same protocol was used to assess release into the cytoplasm of endonuclease G. Results Saturated but not unsaturated FFA cause apoptosis in cultured HMC Intracellular accumulation of long-chain saturated FFA (i.e., palmitate C16:0) has been shown to induce apoptosis in several cell types including cardiac myocytes and pancreatic β-cells [ 5 , 6 , 8 , 11 , 13 ]. To determine whether FFAs induce apoptosis in HMC, cells were incubated in medium supplemented with palmitate, stearate, oleate, or linoleate. The saturated FFA palmitate and stearate increased apoptosis in HMC as evidenced by cleavage of caspase-3 and DNA fragmentation (Fig. 1A,B ). In contrast, the unsaturated FFA oleate and linoleate did not increase caspase-3 cleavage or DNA fragmentation compared to cells incubated with albumin alone (Control, Fig. 1A,B ). The number of pyknotic nuclei, another prototypical feature of apoptotic cells, were significantly higher in palmitate-treated cells whereas the number of pyknotic nuclei in cell incubated with oleate were similar to control (Fig. 2A ). To determine whether mitochondrial β-oxidation was necessary for palmitate-induced apoptosis, the number of pyknotic nuclei were measured in cells treated with etomoxir, an inhibitor of carnitine palmitoyltransferase I. Etomoxir significantly amplified palmitate-induced formation of pyknotic nuclei at 48 h (Fig. 2B ). The number of pyknotic nuclei was unaffected by etomoxir alone compared to control. In contrast, stimulating fatty acid oxidation with AICAR, an activator of AMP-kinase, abolished the increase in pyknotic nuclei with palmitate (Fig. 2B ). The doses of etomoxir and AICAR used here have been previously demonstrated to inhibit and stimulate, respectively, fatty acid β-oxidation [ 13 ]. Thus, using three different criteria for identifying apoptotic cell death, these data demonstrate that the saturated but not unsaturated FFA induces apoptosis in HMC, similar to the proapoptotic effects of saturated FFA in other cell types [ 5 , 6 , 8 , 11 , 13 ]. In addition, enhanced β-oxidation of palmitate is not involved in this process; indeed, increased disposal of palmitate via oxidation apparently protects mesangial cells. Palmitate activates an intrinsic proapoptotic signaling pathway in HMC We next investigated the signaling pathways by which palmitate induces apoptosis in HMC. To begin answering this question, we examined activation of initiator caspases associated with different pathways of apoptotic signaling. Cleavage of procaspase zymogens is required to form active heterotetrameric caspase complexes, so we analyzed the time course of caspase cleavage by Western blotting in 3 independent experiments. The p35 cleaved fragment of caspase 9 was elevated in cells treated with palmitate (Fig. 3A,B ). After 48 h of palmitate, the amount of cleaved caspase 9 was similar to that in cells treated for 8 h with the robust apoptotic stimulus staurosporine (Fig. 3A , Pos lane). In contrast, palmitate did not increase cleavage of caspase-8 (Fig. 3A ), but treatment with a strong caspase-8 activator, etoposide, confirmed that HMC expressed caspase-8 and that the Western blot correctly measured caspase-8 cleavage (Fig. 3A , Pos lane). Caspase-2 has recently been implicated as an initiator caspase in signals involving stress in the endoplasmic reticulum or nucleus [ 25 ]. However, palmitate did not induce caspase-2 cleavage in HMC whereas the known caspase-2 stimulus camptothecin activated robust cleavage of the p15 fragment (Fig. 3A ). Reprobing of all blots for β-actin confirmed equal protein loading (Fig. 3A ). Collectively, these results suggest that palmitate activates caspase 9, an important initiator caspase in the intrinsic pathway of proapoptotic signal transduction. To confirm that palmitate activated caspase 9, we used a luminometric caspase 9 substrate to directly measure caspase 9 enzyme activity in cytosolic extracts of HMC. Palmitate at 0.4 mM increased caspase 9 activity 2.2- and 4.8-fold at 24 and 48 h, respectively (Fig. 4 ). A lower dose of palmitate (0.2 mM) that also stimulated apoptosis (Fig. 2 ) increased caspase 9 activity. Importantly, oleate (0.4 mM) did not stimulate caspase 9 activity above control levels at any time point tested (Fig. 4 ). These results confirm that the saturated FFA palmitate increases caspase 9 enzyme activity. Because activation of caspase 9 by palmitate points to an intrinsic pathway of proapoptotic signaling, we asked whether palmitate could stimulate release of cytochrome-c from mitochondria. Redistribution of cytochrome-c to the cytoplasm is an important step in apoptosome formation and greatly enhances the enzymatic activity of caspase 9 [ 25 , 26 ]. Western blotting of cytoplasmic and membrane-enriched fractions that contain mitochondria was used to assess cytochrome-c distribution. In control cells (5% FBS plus albumin alone), cytochome-c resided exclusively in the membrane fraction (Fig. 5A,B ). In cells treated with palmitate, a portion of total cytochrome-c was redistributed to the cytoplasmic fraction. Consistent with the inability of oleate to activate caspase 9, oleate did not appreciably redistribute cytochrome-c to the cytosol (Fig. 5 ). These results provide additional evidence that palmitate activates an intrinsic pathway of proapoptotic signaling in HMC. Inhibition of caspase 9 attenuates palmitate-induced apoptosis in HMC We next tested the functional role of caspase 9 activation in the palmitate-induced signal transduction cascade. Co-incubation with a cell-permeable selective inhibitor of caspase 9 blocked activation of caspase 9 enzyme activity by palmitate (Fig. 6A ). The caspase 9 inhibitor alone had a minor effect (<0.2-fold inhibition) on basal caspase 9 activity, and a caspase-8 inhibitor did not block caspase 9 activity stimulated by palmitate (Fig. 6A ). Under theses conditions, inhibition of caspase 9 abolished caspase-3 cleavage in cells treated with palmitate for 24 and 48 h (Fig. 6B ). Formation of pyknotic nuclei by 0.2 mM palmitate was blocked in cells treated with the caspase 9 inhibitor at both 24 and 48 h. However, the number of pyknotic nuclei was only partially reduced in cells treated with 0.4 mM palmitate for 48 h (Fig 6C ), even though under these conditions caspase 9 enzyme activity was completely inhibited (Fig. 6A ). Similarly, DNA fragmentation by palmitate was incompletely blocked by the caspase 9 inhibitor at 48 h (Fig. 6D ). These results show that inhibition of palmitate-stimulated caspase 9 blocks some but not all apoptotic death in HMC. Because the caspase 9 inhibitor completely blocked caspase-3 cleavage, these results also suggest that proapoptotic signaling by palmitate proceeds by both caspase 9/3-dependent and independent mechanisms. Mitochondrial release of endonuclease G in palmitate-treated HMC To elucidate the caspase 9/3-independent mechanisms of proapoptotic signaling by palmitate, we investigated the possibility that palmitate stimulates the release of other mitochondrial proapoptotic proteins that can contribute to nuclear changes independent of casapse-3. Endonuclease G is one such effector of apoptosis, a mitochondrial DNase released by a Bcl-2-dependent but caspase-independent mechanism [ 27 ]. To examine the extent to which palmitate can induce endonuclease G release, we treated HMC with palmitate and measured the presence of endonuclease G in the cytoplasm of fractionated cells. Palmitate increased the amount of endonuclease G released into the cytoplasm (Fig. 7A and 7B ). The release of endonuclease G in palmitate-treated cells was not altered by caspase 9 inhibition (Fig. 7A and 7B ). These results are consistent with the incomplete inhibition of DNA fragmentation observed when caspase 9 activation by palmitate was blocked in HMC (Fig. 6D ). Collectively, these results suggest that palmitate-induced apoptosis proceeds by caspase 9-dependent mechanisms and by a caspase 9 -independent mechanism involving endonuclease G. Caspase 9-dependent and -independent pathways of palmitate-induced apoptosis are abolished by unsaturated FFA Unsaturated FFA have been reported to block apoptosis by saturated FFA [ 8 , 11 , 13 ], so we next asked whether unsaturated FFA block palmitate-induced apoptosis in HMC and whether they act by inhibiting the caspase 9-dependent or -independent pathway. The unsaturated FFA oleate and linoleate blocked the increase in caspase-3 cleavage and DNA fragmentation in cells treated with palmitate (Fig. 8A,B ). Apoptosis induced by stearate was also effectively blocked in cells co-incubated with oleate or linoleate (Fig. 8A,B ). Oleate also prevented the increase in caspase 9 activity in cells treated with palmitate (Fig. 9 ). The specificity of this assay for caspase 9 was confirmed by showing that the cell-permeable caspase 9 inhibitor prevented palmitate-stimulated enzyme activity whereas a caspase 8-selective inhibitor had no effect (Fig. 9 ). Thus, similar to other cell types, unsaturated FFA inhibit the apoptotic response to saturated FFA in HMC. Moreover, oleate and linoleate completely prevent palmitate-induced apoptosis, which contrasts with the partial blockade observed with inhibition of caspase 9 (Fig. 6 ). Because the unsaturated FFA more effectively blocked palmitate-induced apoptosis, we investigated whether this was because they blocked the caspase-independent release of endonuclease G in cells treated with palmitate. As expected, palmitate increased mitochondrial release of cytochrome-c and endonuclease G in HMC (Fig. 10A,B ). Oleate alone had no effect on release of cytochrome-c or endonuclease G. Co-incubation of oleate with palmitate prevented the release of cytochrome-c and endonuclease G induced by palmitate (Fig. 10A,B ). These results are consistent with the notion that oleate blocks both the caspase 9-dependent and -independent mechanisms of proapoptotic signaling by palmitate. In particular, the ability of oleate to block release of endonuclease G in palmitate-treated cells is one explanation for the superior antiapoptotic effect of oleate compared to caspase 9 antagonism. Discussion Our results show that the saturated FFA palmitate induces an intrinsic pathway of proapoptotic signaling in HMC, a vascular target cell of lipid-mediated injury in the kidney. In contrast, the monounsaturated FFA oleate did not induce proapoptotic signaling and instead protected HMC from palmitate-induced apoptosis. Evidence that palmitate induced an intrinsic pathway of proapoptotic signaling is that it increased cytochrome-c release, caspase 9 cleavage, and caspase 9 enzyme activity. We also showed that the palmitate-stimulated intrinsic pathway proceeded by a caspase 9-dependent mechanism and by a caspase 9 -independent mechanism involving endonuclease G. Both the caspase 9-dependent and -independent pathways were effectively blocked by oleate. In this study we report that palmitate causes apoptosis in HMC, a microvascular cell that accumulates lipids in vivo in settings of high FFA such as obesity and type 2 diabetes [ 15 - 20 ]. We chose palmitate because it is the most abundant saturated FFA complexed to human serum albumin [ 28 ]. We used complexes where the molar ratio of FFA to albumin was 6:1. Although the normal physiologic ratio of FFA to albumin is approximately 2:1, serum FFA levels are greatly elevated in patients with obesity, type 2 diabetes, and proteinuric renal diseases, yielding ratios of 6:1 or higher [ 29 , 30 ]. In addition, normal circulating FFA levels are approximately 0.5 mM [ 3 ], so the concentrations of individual FFA species (i.e., 0.2 – 0.4 mM) used in this study seem reasonable. Therefore, our experiments were designed to evaluate mechanisms of FFA-induced apoptosis relevant to type 2 diabetes. Previous studies have demonstrated that saturated but not monounsaturated FFA cause apoptosis in other non-renal cell types [ 5 , 6 , 8 , 11 , 13 ], but the ability of FFA to induce apoptosis appears to vary with the specific cell type in question [ 1 ]. In addition, the signal transduction pathways by which saturated FFA induce apoptosis are incompletely defined. Similar to a previous report in breast cancer cells [ 13 ], palmitate-induced apoptosis in HMC was enhanced by inhibition of fat oxidation and reversed by increasing fat oxidation. Although we did not directly measure the effects of these compounds on fatty acid oxidation, these results suggest that palmitate must be metabolized to promote apoptosis and that mitochondrial β-oxidation of the saturated FFA does not participate in the proapoptotic response to palmitate. Several observations from our study suggest that palmitate induces an intrinsic pathway of apoptotic signaling in HMC. First, palmitate stimulated accumulation of cytochrome-c in the cytosol, which is an important step in the intrinsic pathway to promote apoptosome formation and activation of caspase 9. Palmitate-induced release of cytochrome-c has been previously reported in β-cells and breast cancer cells [ 13 , 14 ]. Palmitate has also been shown to stimulate release of uncharacterized proapoptotic proteins when added directly to isolated mitochondria [ 31 ]. Second, palmitate stimulated caspase 9 cleavage and activity in HMC. To our knowledge activation of caspase 9 by palmitate has not been previously shown. Caspase 9 is an initiator caspase in many but not all intrinsic pathways of proapoptotic signaling [ 25 , 26 ]. In a stimulus- and cell type-specific manner, caspase 2 can function upstream of caspase 9 [ 32 - 35 ], but in our experiments palmitate did not stimulate cleavage of caspase 2, an indicator of caspase 2 activation. Also, we did not observe cleavage of caspase 8, an initiator caspase in the death receptor pathway, in response to palmitate. Thus an important result of our study is that the intrinsic pathway in palmitate-induced apoptosis appears to involve caspase 9. A functional role for caspase 9 in palmitate-induced apoptosis was suggested by experiments in which a cell-permeable caspase 9 inhibitor, Z-LEHD-FMK, blocked apoptosis in response to palmitate. Z-LEHD-FMK completely inhibited the activation of caspase 9 at 24 and 48 h of 0.4 mM palmitate. Under these conditions, Z-LEHD-FMK reversed caspase-3 cleavage induced by palmitate, suggesting that caspase 9 is required for activation of caspase-3 by palmitate. While Z-LEHD-FMK effectively inhibited the nuclear changes characteristic of apoptosis at 24 h, the caspase 9 inhibitor did not completely reverse pyknotic nuclei or DNA fragmentation at 48 h. For example, partial inhibition of DNA fragmentation was observed in HMC treated with 0.2 or 0.4 mM palmitate for 48 h, even though the palmitate-induced increment in caspase 9 was blocked. A possible explanation of these results is that when HMC are exposed to palmitate for 48 h, the saturated FFA recruits additional caspase 9-independent mechanisms of apoptotic or non-apoptotic cell death that are not induced at 24 h. A possible caspase 9-independent mechanism for palmitate-induced apoptosis would be the mitochondrial release of endonuclease G, which we demonstrated in HMC. Endonuclease G is a DNase normally located in the intermembrane space of mitochondria. Some apoptotic stimuli cause caspase-independent release of endonuclease G after which it translocates to the nucleus and cleaves DNA [ 27 ]. Release of endonuclease G in palmitate-treated HMC was not blocked by Z-LEHD-FMK, which could explain the partial inhibition of DNA fragmentation by Z-LEHD-FMK under conditions where caspase 9 activation by palmitate was completely blocked. It is also possible that a small amount of caspase 3 activity remains even in the presence of the caspase 9 inhibitor, and it is possible that this low level of caspase-3 activity also contributes to endonuclease G release. In striking contrast to the partial inhibition of palmitate-induced apoptosis by the caspase 9 antagonist, we found that the unsaturated FFA oleate and linoleate completely prevented caspase 3 cleavage and DNA fragmentation in cells treated with either palmitate or stearate. Oleate prevented mitochondrial release of cytochrome-c and the increase in caspase 9 in cells treated with palmitate. In addition, oleate blocked mitochondrial release of endonuclease G in palmitate-treated cells. Taken together, these results support the notion that oleate completely prevents palmitate-induced apoptosis because, unlike inhibition of caspase 9 alone, oleate blocks both the caspase 9-dependent and -independent pathways. Conclusions These results show that palmitate stimulates apoptosis by evoking an intrinsic pathway of proapoptotic signaling. In addition, we have identified mitochondrial release of endonuclease G as a key step in proapoptotic signaling by saturated FFA and in the anti-apoptotic actions of unsaturated FFA. We believe that these results might be relevant to the pathogenesis of microvascular injury in type 2 diabetes because FFA accumulate in microvascular cells, including mesangial cells, in vivo in experimental models of type 2 diabetes, obesity, or hyperlipidemia [ 15 - 20 ]. Thus lipid-driven apoptosis might contribute to the microvascular remodeling that leads to numerous complications in type 2 diabetes. List of Abbreviations FFA, free fatty acid; HMC, human mesangial cells; Competing Interests The author(s) declare that they have no competing interests. Authors' contributions RM carried out most of the technical studies and helped draft the manuscript. MSS conceived of the study, participated in its design and execution, and helped to draft the manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546189.xml
550674
FADD adaptor in cancer
FADD (Fas Associated protein with Death Domain) is a key adaptor molecule transmitting the death signal mediated by death receptors. In addition, this multiple functional protein is implicated in survival/proliferation and cell cycle progression. FADD functions are regulated via cellular sublocalization, protein phosphorylation, and inhibitory molecules. In the present review, we focus on the role of the FADD adaptor in cancer. Increasing evidence shows that defects in FADD protein expression are associated with tumor progression both in mice and humans. Better knowledge of the mechanisms leading to regulation of FADD functions will improve understanding of tumor growth and the immune escape mechanisms, and could open a new field for therapeutic interventions.
The FADD molecule The FADD gene is located on chromosome 11q13.3 in humans and 7 in mice [ 1 ]. Mutations in the FADD gene containing locus are frequently observed in human malignancies [ 2 ]. For instance, the 11q13 region contains the fibroblast growth factor 3 and 4 genes which are coamplified in melanoma. It also includes the multiple endocrine neoplasia I gene whose mutation leads to tumor development of several endocrine glands including thyroid. Moreover, two genes implicated in leukemia are found in this locus: NUMA1 which is translocated in acute promyelocytic leukemia, and BCL1 which is located very close to the FADD gene and is mutated in B-cell leukemia/lymphoma. Although FADD has a central role in multiple receptor-induced cell death as discussed hereafter, no mutation of the FADD gene itself has been reported so far. Human and mouse FADD genes have the same quite simple organization consisting of two exons (286 bp and 341 bp in humans; 332 bp and 286 bp in mice) separated by a unique intron of approximately 2 kb. Interestingly, no cap site was reported on the human FADD mRNA [ 1 ], suggesting a particular regulation of FADD mRNA translation, although this topic has not been further investigated. Human and mouse FADD proteins are very similar (Figure 1 ). They consist of 208 and 205 amino acids (AA) respectively, and share 80% similarity and 68% identity [ 3 ]. FADD mRNA and protein are almost ubiquitously expressed in fetal and adult tissues, both in humans and mice [ 4 ]. Two domains are particularly well conserved between species: the death domain (DD) at the COOH-terminus of the protein, and the death effector domain (DED) at the NH2-terminus of the protein [ 5 , 6 ]. Both domains play a crucial role in transducing the apoptotic signal mediated by death receptors. Furthermore, a single serine (Ser) phosphorylation site essential for determining cell cycle progression is conserved in both species (human Ser 194 [ 7 ] and mouse Ser 191 [ 8 ]). Figure 1 Human and mouse FADD protein. Amino acids (AA) corresponding to the human FADD protein are marked in black, whereas AA corresponding to the mouse FADD protein are marked in grey. The death domain (DD) and death effector domain (DED) are essential for interaction with death receptors and transmission of the apoptotic signal. Human nuclear export sequence (NES in red) and nuclear localization sequence (NLS in blue) determine localization of the protein in the cytoplasm and the nucleus, which are associated with cell death and survival functions of the FADD protein, respectively. Human Ser 194 and mouse Ser 191 phosphorylation site (in purple) have a crucial role in survival/proliferation and cell cycle progression. Since the first role ascribed to FADD was to transmit apoptotic signals through its interaction with death receptors expressed at the cell membrane, it was assumed that FADD protein was exclusively localized in the cytoplasm of the cell. However, a nuclear localization sequence (NLS) and a nuclear export sequence (NES) were recently identified in the human FADD protein (Figure 1 ), and account for FADD protein expression in the nucleus and the cytoplasm of the cell, respectively [ 9 ]. The vast majority of the reports on FADD focused on the cytoplasmic FADD protein because of its pro-apoptotic function. In contrast, the role of the nuclear FADD is much more mysterious. It was recently reported that FADD expression in the nucleus protects cells from apoptosis, but the mechanism implicated in this survival function has not been investigated [ 9 ]. On the other hand, it has been shown that FADD could interact within the nucleus of adherent cells with the methyl-CpG binding domain protein 4 (MBD4) [ 10 ]. MBD4 is a GT mismatch repairing protein. Association between MBD4 and FADD within the nucleus could couple MBD4-mediated genome surveillance with FADD-mediated cell death. Thus, nuclear FADD could perhaps have a pro-apoptotic function, at least in response to DNA-damaging agents. Functions of the FADD protein An essential molecule for embryonic development The essential role of the FADD molecule was highlighted by generating FADD mutant null mice [ 11 , 12 ]. Indeed, FADD knockout mice were not viable. FADD null embryos died in utero at day 12.5 of development, due to underdevelopment, abdominal hemorrhage, and cardiac failure. Moreover, FADD loss of function did not result in a lymphoproliferative disorder as observed in viable Fas mutant mice [ 13 , 14 ]. These results indicated that in addition to its well known role in cell death, FADD was also implicated in survival/proliferation of some cell types. A main death transducer for DD-containing receptors FADD is the main signal transducing intermediate adaptor molecule of several death receptors including Fas, TNF-R1 (tumor necrosis factor receptor 1), DR3 (death receptor 3), TRAIL-R1 (TNF-related apoptosis-inducing ligand, DR4), and TRAIL-R2 (DR5) [ 4 , 11 , 12 , 15 ]. All these receptors possess, in their intra-cytoplasmic tail, a DD homologous to the DD of FADD allowing FADD recruitment to the activated receptor. FADD can be recruited either directly to Fas and TRAIL-Rs (Figure 2A ) or indirectly to TNF-R1 (Figure 2B ). In the latter case, FADD is recruited via another DD-containing adaptor molecule (TRADD, TNF receptor-associated protein with DD). Next, FADD recruits DED-containing initiator pro-caspase 8 or 10 through DED/DED interactions [ 16 - 18 ], thus forming the death-inducing signaling complex (DISC) [ 19 ]. Autoprocessing of initiator pro-caspase leads to activation of effective caspases which cleave intracellular substrates, causing the apoptotic death of the cell [ 20 , 21 ]. Figure 2 Apoptosis mediated by death receptors requires the FADD adaptor molecule. (A) Engagement of a Fas ligand trimer on a trimer of Fas leads to FADD adaptor molecule recruitment through homotypic DD interactions. FADD next binds initiator pro-caspase through DED interactions. This Fas/FADD/pro-caspase complex forms the Fas death-inducing signaling complex (DISC) since initiator pro-caspase activates a caspase cascade resulting in apoptotic death of the cell. Alternatively, c-FLIPs can promote cell survival by interacting with FADD through their respective DED, thus hindering recruitment and activation of initiator pro-caspase. (B) Signaling mediated by TNF-R1 implicates formation of two sequential complexes [61]. The complex I (in blue) contains the TNF-R1, the adaptor TRADD, the receptor interacting kinase (RIP), and the TNF-receptor associated factor 2 (TRAF2). It assemblies rapidly following TNF-α stimulation and activates the NF-kB pathway which in turn induces expression of survival genes, including c-FLIP. Later on, complex I dissociates from the TNF-R1 and is internalized. FADD can then bind the liberated DD of TRADD and recruits initiator pro-caspase, forming complex II (in red) which is cytoplasmic. Activation of initiator pro-caspase 8/10 in complex II results in apoptosis of the cell. Red box: DD; hatched red box: DED. Control of FADD recruitment to the DISC can occur following several mechanisms depending on the cell type and the death receptor [ 22 ]. The best characterized death receptor signaling inhibitors are DED-containing viral and cellular FLICE-inhibitory proteins (v-FLIPs and c-FLIPs, respectively) [ 23 , 24 ]. Inhibition of Fas-, TNF-R1-, and TRAIL-Rs-induced apoptosis by endogenous FLIPs results from binding of the c-FLIPs to the DED of FADD, thus hindering pro-caspase 8 activation (Figure 2A ). Similarly, v-FLIPs inhibit apoptosis mediated by death receptors either by binding to FADD and blocking pro-caspase 8 processing, or by binding to pro-caspase 8 and inhibiting FADD interaction. Therefore, equilibrium between FADD and the expression of its inhibitors determines the outcome of the death receptor-stimulated cell, i.e. apoptosis or survival. All the main death receptors described up to now require FADD adaptor for transmitting their apoptotic signal. Consequently, FADD is a central protein that controls multiple essential cellular processes including cellular homeostasis and elimination of pathological cells, particularly during the course of an immune response. Death receptor independent FADD induced apoptosis Formation of cytoplasmic death effector filaments (DEF) by oligomerization of DED-containing proteins, including FADD, is responsible for death receptor independent cellular apoptosis [ 25 ]. Indeed, FADD over-expression by itself is known to induce cell death through DEF formation that recruits and activates pro-caspase 8. However, the existence of DEF has not been established in vivo , and increasing evidences suggest that DEF could be artefactual structures resulting from protein over-expression. As a consequence, the ability of endogenous FADD to aggregate and form DEF in normal situation should be reconsidered. Functions in proliferation and cell cycle progression Beside being a main death adaptor molecule, FADD is also required for T cell proliferation. The first evidence of this property of FADD came from observations made in chimeric FADD knockout mice. Five-week-old chimeric FADD -/- mice presented a lack of thymocytes compared to wild type animals, with few or no CD4 + CD8 + double positive thymocytes remaining [ 12 ]. Moreover, several groups have demonstrated that FADD deficiency in peripheral T lymphocytes resulted in an inhibition of mitogen-induced T cell proliferation [ 12 , 26 - 30 ]. The mechanism leading to FADD-dependent T cell proliferation did not involve the early events associated with cell proliferation since expression level and functionality of the IL-2 receptor, level of IL-2 secretion, mobilization of intracellular calcium, and activation of NF-kB, p38-MAPK, and p44/42-MAPK appeared normal in FADD -/- T lymphocytes [ 12 , 29 , 30 ]. Recent data showed that FADD -/- T lymphocytes entered the cell cycle upon mitogenic stimulation, but died during progression through the cell cycle [ 30 ]. Therefore, lack of proliferation of FADD-deficient T cells results from defective survival associated with progression through the cell cycle rather than defective activation. Up to now, the molecular pathway implicated in FADD-mediated survival of lymphocytes has not been described. In addition to impairing survival during cell division, FADD deficiency also leads to a dysregulation of the cell cycle machinery [ 31 ]. The pattern of expression of molecules implicated in both G1/S and G2/M transitions was aberrant in FADD -/- lymphocytes, resulting in spontaneous entry and progression through the cell cycle of 10% of freshly isolated FADD -/- T cells (as compared to less than 2% of wild type T cells) [ 31 ]. The mechanisms responsible for FADD regulation of cell cycle progression are not fully understood. However, the phosphorylation of the 194 human and 191 mouse Ser of the protein (Figure 1 ) has recently drawn attention. Indeed, human FADD was phosphorylated at Ser 194, by a still unidentified 70 kDa protein kinase, in cells arrested in G2/M, whereas it was unphosphorylated in G1/S [ 7 ]. Generation of FADD Ser 191 mutant mice confirmed that FADD phosphorylation is involved in proliferation in vivo [ 8 ]. Replacement of Ser 191 by an aspartic acid resulting in constitutive phosphorylation of the FADD protein led to abnormal development of FADD mutant mice that shared the same phenotype as FADD deficient mice [ 8 ], including few CD4 + CD8 + thymocytes and defective progression through the cell cycle [ 12 , 31 ]. Tumor cells are constantly cycling cells. Although it does not seem to directly affect the cell cycle progression, FADD phosphorylation at Ser 194 sensitizes these cells to reagents that induce G2/M arrest such as the Taxol anticancer drug [ 32 ]. In human prostate cancer cell lines, treatment with Taxol resulted in Ser 194 FADD phosphorylation and G2/M arrest [ 33 ]. Moreover, etoposide or cisplatin chemotherapeutic drug-induced apoptosis of these cells was enhanced by pretreatment with Taxol, a process that was inhibited by cellular over-expression of an unphosphorylable FADD mutant [ 33 ]. Therefore, tumor cells that express a Ser 194 FADD mutant that cannot be phosphorylated or are unable to phosphorylate FADD at this Ser position are expected to resist apoptosis induced by anticancer drugs that induce G2/M arrest, and to be insensitive to the synergistic effect of chemotherapy. Obviously, a lack of FADD expression will have the same consequences. Role in tumor development FADD as a tumor suppressor The role of the FADD adaptor in cancer was initially demonstrated by generating RAG-1 deficient transgenic mice that target expression of a FADD dominant negative mutant in lymphocytes. With age these mice developed thymic lymphoblastic lymphoma, whereas FADD +/+ RAG-1 -/- mice did not [ 34 ]. Moreover, thymic lymphoblastic lymphomas were never observed in FADD -/- RAG-1 +/+ or FADD -/- RAG-1 +/- mice, demonstrating that absence of FADD expression was necessary but not sufficient to induce tumor development in this model [ 34 ]. These results were the first demonstration that FADD adaptor can act as a tumor suppressor in vivo . Using a mouse model of thyroid adenoma/adenocarcinoma, we showed spontaneous disappearance of FADD protein expression during the course of tumor development [ 35 ]. The so called gsp transgenic mice expressed an oncogene specifically in thyroid follicular cells (TFC), and developed thyroid hyperplasia that eventually transformed into hyperfunctioning adenomas or adenocarcinomas around the age of 8 months [ 36 ]. The fact that gsp mice developed hyperfunctioning adenomas or adenocarcinomas belatedly suggested that oncogene expression conferred a predisposition but that an additional event was necessary for thyroid tumor development. We found that FADD protein was highly expressed in all non-pathological and in almost all hyperplastic thyroid glands from gsp mice. In contrast, thyroid adenoma/adenocarcinoma expressed low or no FADD protein [ 35 ]. These results raised the possibility that loss of FADD protein expression could be an additional event contributing to tumorigenesis, and suggested that FADD plays a role as a tumor suppressor. We recently showed that absence of FADD protein expression in cancer cells is also a relevant phenomenon in human malignancies [ 37 ]. We looked for FADD protein expression in human acute myeloid leukemia (AML) cells. Leukemic cells of most AML patients are resistant to Fas-mediated cell death despite expressing the Fas receptor [ 38 ] and/or the FasL molecule [ 39 ]. Moreover, chemotherapeutic drugs used for AML treatment can kill target cells via several mechanisms, including death receptor-induced apoptosis [ 40 - 43 ]. We performed a retrospective study of 70 consecutive patients with de novo AML treated homogeneously, and found that leukemic cells of 2/3 of patients at diagnosis expressed low or no FADD protein [ 37 ]. Moreover, in this cohort of patients, we showed that absent/low FADD protein expression in AML cells at diagnosis was a new independent prognostic factor for poor response to chemotherapy (in terms of complete remission rate, event free and overall survivals) [ 37 ]. Importantly, absent/low FADD protein expression in AML cells at diagnosis was a prognostic factor even for patients classified as standard- or good-risk AML cases by cytogenetic and molecular criteria [ 37 ]. As a consequence, this new prognostic factor is of clinical importance since it will allow early identification of patients with chemoresistant AML who could benefit from more intensive post-remission therapy. Absence of FADD confers numerous advantages on cancer cells The fact that absence of FADD expression was found in different types of tumor cells both in mice and humans strongly suggested that absence of FADD contributed to tumor development. Indeed, lack of FADD protein can confer numerous advantages on pathological cells, which predominantly result in tumor survival/growth gain (Figure 3 ). Figure 3 Lack of FADD expression confers survival/growth advantages on tumor cells. Absence of FADD protein confers multiple death receptor-mediated apoptosis resistance and allows tumor cells to co-express death receptors and ligands without committing cell death. (A) Lack of FADD contributes to immune escape and resistance to chemotherapy. FADD deficient tumor cells resist death receptor-mediated apoptosis induced by TIL and chemotherapeutic drugs. Anthracyclines increase Fas and FasL expression on tumor cells. Etoposide induces Fas receptor trimerization, leading to Fas-mediated cell death independently of FasL expression. Both drugs enhance TRAIL-R2-mediated apoptosis. (B) Lack of FADD contributes to tumor counter-attack. Lack of FADD expression allows many types of tumor cells to express innocuously functional FasL that can kill TIL. Secretion of TNF-α by AML cells can have cytotoxic effects on TIL. (C) Lack of FADD contributes to tumor growth. In the absence of FADD, Fas signaling leads to a proliferative signal instead of an apoptotic one. Concomitant death receptor and ligand expression, in the absence of FADD, allows autocrine (in red) and paracrine (in orange) proliferation of tumor cells. Activated TIL can contribute to paracrine (in purple) proliferation of FADD-deficient tumor cells. Immune escape and resistance to chemotherapy Fas, TRAIL-Rs, TNF-R1, DR3, and potentially other receptors use FADD adaptor for transmitting the death signal. Thus, absence of FADD expression in tumor cells must confer multiple resistance of these cells to death receptor cytotoxicity. In agreement with this assumption, we observed that FADD -/- TFC as well as FADD -/- AML cells were resistant to Fas- and TNF-α-mediated apoptosis [ 35 ], and unpublished data). Since cytotoxic tumor infiltrating lymphocytes (TIL) use, among other mechanisms, death receptor-mediated cell death to kill pathological cells, tumor cells lacking FADD molecule expression may partially avoid immune attack (Figure 3A ). On the other hand, some anticancer drugs exert their cytotoxic effect by inducing death receptor and/or death ligand expression on tumor cells, thereby inducing suicidal/fratricidal apoptosis of the cells. For instance, anthracyclines and etoposide, two chemotherapeutic drugs used for AML treatment, enhance Fas- and TRAIL-R2-mediated cell death in vitro , a process requiring FADD molecule expression [ 40 - 42 , 44 , 45 ]. As a consequence, absence of FADD expression in AML cells of our patients may contribute to chemoresistance of leukemic cells (Figure 3A ). Tumor counter-attack As described above, absence of FADD allows co-expression of death receptors and ligands without inducing autocrine or paracrine apoptosis of tumor cells. For instance, although the Fas receptor was expressed at all stages of thyroid tumor development in gsp mice, FasL expression was gained with a high expression in adenomatous/adenocarcinomatous glands [ 35 ]. Using the same method, we found that leukemic cells of most AML patients expressed FasL [ 39 ] and secreted TNF-α despite expressing Fas and TNF-R1 [ 37 ], and unpublished data). Moreover, co-expression of Fas and FasL molecules that did not cause cell death was also observed in lymphoma [ 46 ], melanoma [ 47 ], astrocytoma [ 48 ], cancers of the colon [ 49 ], liver [ 50 ], lung [ 51 ], human thyroid [ 52 ]. The role of death ligand expression in tumor cells is still a controversial issue [ 53 - 55 ], but it is now well accepted that it allows at least some cancer cells to kill TIL that express death receptors, a process called the "tumor counter-attack" (Figure 3B ). Proliferative advantage Since the tumor counter-attack hypothesis cannot apply to all types of cancer cells, one could wonder whether other benefits of death ligand expression by tumor cells exists. Previous reports demonstrated that Fas signaling could lead to proliferation instead of apoptosis, depending on the cell type and the environmental conditions [ 56 - 58 ]. For instance, agonistic anti-Fas antibody-induced proliferation of hematological and non-hematological tumors has been described [ 59 ]. Moreover, we have shown that stimulation of FADD lacking thyrocytes by an agonistic anti-Fas antibody resulted in accelerated growth of TFC, via a particular Daxx adaptor-mediated pathway [ 35 ]. Therefore, Fas signaling, particularly in the absence of FADD, can confer proliferative advantage on tumor cells (Figure 3C ). Thus, FasL expression on Fas + FADD - adenomatous/adenocarcinomatous thyroid from gsp mice, as well as on human AML cells, may allow both autocrine and paracrine proliferation of these cells [ 35 , 37 ] (Figure 3C ). Furthermore, we can formulate the same hypothesis for TNF-α secretion by TNF-R1 expressing leukemic cells (Figure 3C ). If our hypotheses are correct, then activated TIL that express death ligands could contribute to FADD-deficient tumor proliferation (Figure 3C ). Restoring FADD protein expression- a new therapeutic issue Absence of FADD expression could confer multiple growth advantages on cancer cells (Figure 3 ), and is expected to contribute to disease progression. As a consequence, finding how to restore FADD protein expression in FADD-negative tumor cells represents a research field with potentially direct clinical applications. In fact, it is possible that some of the current cancer treatments act, at least partially, through this mechanism. For instance, carboplatin is a cytotoxic drug potentially effective at reestablishing functional FADD protein expression. Indeed, the human tongue carcinoma cell lines SCC-9 and SCC-25 express very low levels of FADD protein. Moreover, treatment with carboplatin enhances FADD protein expression, thus rendering cancer cells sensitive to Fas-mediated apoptosis [ 60 ]. The combination of carboplatin with chemotherapeutic drugs that induce death receptor-mediated cell death may result in improved treatment. Molecules implicated in FADD regulation of expression also represent new therapeutic targets. However, very little is known about the mechanisms leading to absent/low FADD protein expression in tumor cells. In SCC-9 and SCC-25 carcinoma cell lines, carboplatin upregulated FADD protein expression by increasing FADD mRNA [ 60 ]. However, FADD mRNA was normally expressed in mouse thyroid adenoma/adenocarcinoma and in human AML cells, and lack of FADD mRNA could not account for poor FADD protein expression in these cancer cells [ 35 , 37 ]. These results suggest that several mechanisms could be implicated in loss of FADD protein, depending on the type of cell and the environmental pressure. Besides, docking FADD protein away from the death receptor, in the nucleus for example, would have the same consequences. Understanding such mechanisms is the first necessary step towards development of new anticancer drugs targeting molecules that regulate FADD protein expression. Conclusion FADD is mainly known as a key adaptor molecule for numerous death receptors. However, increasing evidences have shown that FADD is a much more complex molecule implicated in apoptosis, survival, cell cycle progression, and proliferation of the cells. Therefore, FADD plays a central role in the frightening control of cell death and life. As a consequence, a defect in the FADD molecule can contribute to the development of diseases, and particularly cancer. Absence of FADD protein expression is a marker of tumor development in the mouse, and a prognostic factor for poor response to chemotherapy in humans. Since FADD deficiency could contribute to several malignancies, in view of its almost ubiquitous pattern of expression, study of the role of FADD in tumor development, growth, and resistance to treatment, and understanding how the expression of this puzzling molecule is regulated, are targets that merit further investigation. Authors' Contributions LT conceived the review and drafted the manuscript. GC participated in conceiving the review. AB and GC participated in the preparation of the manuscript. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC550674.xml
516796
What Is Life—and How Do We Search for It in Other Worlds?
If there is life on distant worlds, how would we go about finding it?
I need a “tricorder”—the convenient, hand-held device featured on Star Trek that can detect life forms even from orbit. Unfortunately, we don't have a clue how a tricorder might work, since life forms don't seem to have any observable property that distinguishes them from inanimate matter. Furthermore, we lack a definition of life that can guide a search for life outside Earth. How can we find what we can't define? An answer may lie in the observation that life uses a small, discrete set of organic molecules as basic building blocks. On the surface of Europa and in the subsurface of Mars, we can search for alien but analogous patterns in the organics. Life As We Know It The obvious diversity of life on Earth overlies a fundamental biochemical and genetic similarity. The three main polymers of biology—the nucleic acids, the proteins, and the polysaccarides—are built from 20 amino acids, five nucleotide bases, and a few sugars, respectively. Together with lipids and fatty acids, these are the main constituents of biomass: the hardware of life ( Lehninger 1975 , p 21). The DNA and RNA software of life is also common, indicating shared descent ( Woese 1987 ). But with only one example of life—life on Earth—it is not all that surprising that we do not have a fundamental understanding of what life is. We don't know which features of Earth life are essential and which are just accidents of history. Our lack of data is reflected in our attempts to define life. Koshland (2002) lists seven features of life: (1) program (DNA), (2) improvisation (response to environment), (3) compartmentalization, (4) energy, (5) regeneration, (6) adaptability, and (7) seclusion (chemical control and selectivity). A simpler definition is that life is a material system that undergoes reproduction, mutation, and natural selection ( McKay 1991 ). Cleland and Chyba (2002) have suggested that life might be like water, hard to define phenomenologically, but easy to define at the fundamental level. But life is like fire, not water—it is a process, not a pure substance. Such definitions are grist for philosophical discussion, but they neither inform biological research nor provide a basis for the search for life on other worlds. The simplest, but not the only, proof of life is to find something that is alive. There are only two properties that can determine if an object is alive: metabolism and motion. (Metabolism is used here to include an organism's life functions, biomass increase, and reproduction.) All living things require some level of metabolism to remain viable against entropy. Movement (either microscopic or macroscopic) in response to stimuli or in the presence of food can be a convincing indicator of a living thing. But both metabolism (fire) and motion (wind) occur in nature in the absence of biology. The practical approach to the search for life is to determine what life needs. The simplest list is probably: energy, carbon, liquid water, and a few other elements such as nitrogen, sulfur, and phosphorus ( McKay 1991 ). Life requires energy to maintain itself against entropy, as does any self-organizing open system. In the memorable words of Erwin Schrödinger (1945) , “It feeds on negative entropy.” On Earth, the vast majority of life forms ultimately derive their energy from sunlight. The only other source of primary productivity known is chemical energy, and there are only two ecosystems known, both methanogen-based ( Stevens and McKinley 1995 ; Chapelle et al. 2002 ), that rely exclusively on chemical energy (that is, they do not use sunlight or its product, oxygen). Photosynthetic organisms can use sunlight at levels below the level of sunlight at the orbit of Pluto ( Ravens et al. 2000 ); therefore, energy is not the limitation for life. Carbon, nitrogen, sulfur, and phosphorus are the elements of life, and they are abundant in the Solar System. Indeed, the Sun and the outer Solar System have more than 10,000 times the carbon content of the bulk of Earth ( McKay 1991 ). When we scan the other worlds of our Solar System, the missing ecological ingredient for life is liquid water. It makes sense, then, that the search for liquid water is currently the first step in the search for life on other worlds. The presence of liquid water is a powerful indication that the ecological prerequisites for life are satisfied. Orbital images, such as the canyon in Figure 1 , show clear evidence of the stable and repeated, if not persistent, flow of a low-viscosity fluid on Mars at certain times in its past history. The fluid was probably water, but the images could also suggest wind, ice, lava, even carbon dioxide or sulfur dioxide. Recently, results from the Mars Exploration Rover missions have shown that this liquid carried salts and precipitated hematite in concretions. The case for water, we could say, is tight. Figure 1 Water on Another World A Mars Global Surveyor image showing Nanedi Vallis in the Xanthe Terra region of Mars. The image covers an area 9.8 km ×18.5 km; the canyon is about 2.5 km wide. This image is the best evidence we have of liquid water anywhere outside the Earth. Photo credit: NASA/Malin Space Sciences. On Jupiter's moon Europa, the cracks and icebergs on the surface of the ice indicate water beneath the ice, but not necessarily at the present time. Present water on Europa is indicated by the magnetic disturbance Europa makes as it moves through Jupiter's magnetic field, not unlike the way coins in the pocket of a passenger will set off an airport metal detector. Europa has a large conductor, and this is most likely a global, salty layer of water. Viking on Mars: Been There, Tried That The Viking missions to Mars in the late 1970s were the first (and as yet, the only) search for life outside Earth. Each Viking conducted three incubation experiments to detect the presence of metabolism in the Martian soil. Each lander also carried a sophisticated Gas Chromatograph Mass Spectrometer for characterizing organic molecules. The results were unexpected ( Klein 1978 , 1999 ). There was a detectable reaction in two of the incubation experiments. In the “Gas Exchange” experiment, a burst of oxygen was released when the soil was exposed to water. The “Labeled Release” experiment showed that organic material was consumed, and that carbon dioxide was released concomitantly. In the Labeled Release experiment, this reaction ceased if the soil was first heated to sterilizing temperatures, but the reaction of the Gas Exchange Experiment persisted. If considered alone, the Labeled Release results would be a plausible indication for life on Mars. However, the Gas Chromatograph Mass Spectrometer did not detect the presence of any organic molecules in the soil at level of one part per billion ( Biemann 1979 ). It is difficult to imagine life without associated organic material, and this is the main argument against a biological interpretation of the Viking results ( Klein 1999 ; but cf. Levin and Straat 1981 ). It is also unlikely that the oxygen release in the Gas Exchange experiment had a biological explanation, because the reaction was so rapid and persisted after heating. It is generally thought that the reactivity observed by the Viking biology experiments was caused by one or more inorganic oxidants present in the soil, and was ultimately produced by ultraviolet light in the atmosphere. Consistent with the apparently negative results of the Viking biology experiments, the surface of Mars also appears to be too dry for life. Indeed, conditions are such that liquid water is rare and transient, if it occurs at all (e.g., Hecht 2002 ). It's Life, Jim, but Not As We Know It Table 1 shows a categorization of life as we have observed it. Using this diagram, we can speculate about how life might be different on Mars or Europa. At the bottom of the table, life is composed of matter—a reasonable assumption for now. Carbon and liquid water are the next level; this makes Mars and Europa likely candidates, because they have carbon and have, or have had, liquid water. Other worlds may have a different chemical baseline for life. The usual speculation in this area is that the presence of ammonia and silicon, rather than water and carbon, might be preconditions for life on other planets. Such speculation has yet to lead to any specific suggestions for experiments, or to new ways to search for such life, but this may just reflect a failure of human imagination rather than a fundamental limitation on the nature of life. Table 1 A Categorization of Structures That Comprise Terrestrial Life Life on Mars is also likely to be the same at the top of the table: at the ecological level. Primary production in a Martian ecosystem is likely to be phototrophic, using carbon dioxide and water. Heterotrophs are likely to be present to consume the phototrophs and in turn to be consumed by predators. Darwinian evolution would result in many of the same patterns we see in ecosystems on Earth. While it may be similar at the top (ecological) and bottom (chemical) levels, life on Mars could be quite alien in the middle, in the realm of biochemistry. Pace (2001) has argued that alien biochemistry will turn out to be the same as biochemistry on Earth, because there is one best way to do things and that natural selection will ensure that life everywhere discovers that way. Only observation will tell if there is one possible biochemistry, or many. Future missions to Mars might find microfossils in sedimentary rocks such as those at Meridiani Planum. Microbes don't readily form convincing fossils; the one exception may be the strings of magnetite formed by magnetotactic bacteria ( Friedmann et al. 2001 ). As interesting as fossils might be, we could not be sure that a fossil found on Mars was not merely another example of Earth life. We know that rocks have come to Earth from Mars, and it is possible that such rocks could have carried life between the planets ( Mileikowsky et al. 2000 ; Weiss et al. 2000 ). Finding fossil evidence for life on Mars does not demonstrate a second genesis in our Solar System. Finding a Way to Search for Alien Life If we were to find organic material in the subsurface of Mars or on the ice of Europa, how could we determine whether it was the product of a system of biology or merely abiotic, organic material from meteorites or photochemistry? If this life were in fact related to Earth life, this should be easy to determine. We now have very sensitive methods, such as PCR and fluorescent antibody markers, for detecting life like us. This case would be the simplest to determine, but it would also be the least interesting. If the life turned out to be truly alien, then the probes specific to our biology would be unlikely to work. What, then, could we do to determine a biological origin? The question is open and possibly urgent. On space missions already being planned, we may have the opportunity to analyze the remains of alien organics on the surface of Europa or frozen below ground on Mars. The instruments that will make this analysis must be designed in the next couple of years. One approach appears promising. I call it the “Lego Principle.” It is based on the patterns of the molecules of life. Biological processes, in contrast to abiotic mechanisms, do not make use of the range of possible organic molecules. Instead, biology is built from a selected set. Thus, organic molecules that are chemically very similar to each other may have widely different concentrations in a sample of biological organics. An example of this on Earth is the 20 amino acids used in proteins and the use of the left-handed version of these amino acids. The selectivity of biological processes is shown schematically in Figure 2 by the distribution of spikes in contrast to a smooth, nonbiological distribution. General arguments of thermodynamic efficiency and specificity of enzymatic reactions suggest that this selectivity is required for biological function and is a general result of natural selection. Different life forms are likely to have different patterns, and at the very least we might find the mirror symmetry of life on Earth, with d - instead of l -amino acids. Figure 2 Comparison of Biogenic with Nonbiogenic Distributions of Organic Material Nonbiological processes produce smooth distributions of organic material, illustrated here by the curve. Biology, in contrast selects and uses only a few distinct molecules, shown here as spikes (e.g., the 20 l-amino acids on Earth). Analysis of a sample of organic material from Mars or Europa may indicate a biological origin if it shows such selectivity. This approach has immediate practical benefit in the search for biochemistry in the Solar System. Samples of organic material collected from Mars and Europa can be easily tested for the prevalence of one chirality of amino acid over the other. More generally, a complete analysis of the relative concentration of different types of organic molecules might reveal a pattern that is biological even if that pattern does not involve any of the familiar biomolecules. Interestingly, if a sample of organics from Mars or Europa shows a preponderance of D-amino acids, this would be evidence of life, and at the same time would show that this life was distinct from Earth life. This same conclusion would apply to any clearly biological pattern that was distinct from that of Earth life. Organic material of biological origin will eventually lose its distinctive pattern when exposed to heat and other types of radiation, (examples of this include the thermal racemization of amino acids), but at the low temperatures in the Martian permafrost, calculations suggest that there has been no thermal alteration ( Kanavarioti and Mancinelli 1990 ). An interesting question, as yet unanswered, is how long organic material frozen into the surface ice of Europa would retain a biological signature in the strong radiation environment. On Europa, the organic material for our tests might be collected right from the dark regions on the surface. On Mars, there is ice-rich ground in the cratered southern polar regions ( Feldmann et al. 2002 ), which presumably overlies deeper, older ice. The surprise discovery of strong magnetic fields in the southern hemisphere of Mars ( Acuña et al. 1999 ; Connerney et al. 1999 ) indicates that the area may be the oldest undisturbed permafrost on that planet. Like the mammoths extracted from the ice in Siberia, any Martian microbes found in this ice would be dead, but their biochemistry would be preserved. From these biological remains, it would then be possible to determine the biochemical composition of, and the phylogenetic relationship between, Earth life and Martian life. We may then have, for the first time, a second example of life.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516796.xml
512296
Preliminary inventory and classification of indigenous afromontane forests on the Blyde River Canyon Nature Reserve, Mpumalanga, South Africa
Background Mixed evergreen forests form the smallest, most widely distributed and fragmented biome in southern Africa. Within South Africa, 44% of this vegetation type has been transformed. Afromontane forest only covers 0.56 % of South Africa, yet it contains 5.35% of South Africa's plant species. Prior to this investigation of the indigenous forests on the Blyde River Canyon Nature Reserve (BRCNR), very little was known about the size, floristic composition and conservation status of the forest biome conserved within the reserve. We report here an inventory of the forest size, fragmentation, species composition and the basic floristic communities along environmental gradients. Results A total of 2111 ha of forest occurs on Blyde River Canyon Nature Reserve. The forest is fragmented, with a total of 60 forest patches recorded, varying from 0.21 ha to 567 ha in size. On average, patch size was 23 ha. Two forest communities – high altitude moist afromontane forest and low altitude dry afromontane forest – are identified. Sub-communities are recognized based on canopy development and slope, respectively. An altitudinal gradient accounts for most of the variation within the forest communities. Conclusion BRCNR has a fragmented network of small forest patches that together make up 7.3% of the reserve's surface area. These forest patches host a variety of forest-dependent trees, including some species considered rare, insufficiently known, or listed under the Red Data List of South African Plants. The fragmented nature of the relatively small forest patches accentuates the need for careful fire management and stringent alien plant control.
Background Mixed evergreen forests form the smallest, most widely distributed and fragmented biome in southern Africa [ 1 ]. Originally classified as Undifferentiated Afromontane forest [ 2 ], 44% of this vegetation type within South Africa has been transformed [ 2 ]. The forest biome only covers 0.56 % of South Africa [ 3 ], yet it contains 5.35% of South Africa's plant species [ 4 ]. These forests have a relatively high species richness of 0.58 species km -2 , exceeding the grassland biome with 0.25 species km -2 and lagging only the fynbos with 1.36 species km -2 [ 5 ]. The forest biome occurring in Mpumalanga is now recognized as Mpumalanga Mistbelt Forest [ 6 ], and it covers only 0.51% of the province's surface area [ 7 ]. Cooper [ 8 ] conducted a study on the conservation status of indigenous forests within Transvaal, Natal and the Orange Free State, in which he estimated the Blyde River Canyon Nature Reserve's (BRCNR) forests at 352 ha. Recently, ownership of a 700 ha forest tract bordering on BRCNR at the base of the Drakensberg Escarpment was transferred to the Mpumalanga Parks Board (MPB). Therefore, it was thought that just over 1000 ha of forest occurred on BRCNR. To begin to understand the nature and distribution of forests in the BRCNR landscape, we need to have an accurate surface area inventory, a species composition list, and an identification and classification of the communities. Standard ecological fieldwork coupled with remotely sensed data and GIS allow for an efficient analysis. Prior to this investigation of the indigenous forests on BRCNR very little was known about the size and floristic composition of the forest biome conserved within the reserve. In comparison, the grassland biome has received much attention on BRCNR with a monitoring program set up to assess the status of the grasslands. We report here an inventory of forest size, fragmentation, species composition and the basic floristic communities along environmental gradients. Study area BRCNR is situated on the northern Drakensberg Escarpment, Mpumalanga Province, South Africa (24°40'S longitude, 30°51'E latitude; Figure 1 ). The Mpumalanga Parks Board administers this reserve, which is approximately 29 000 ha in size. The elevation on BRCNR ranges from 560 m to 1944 m above sea level. The stratigraphy of the northern Drakensberg Escarpment region is composed of sediment rock types (quartzite, shale, and dolomite) of the Transvaal Supergroup [ 9 ], dominated by Black Reef Quartzite and Wolkberg Group [ 10 ]. Rainfall varies from 541 mm to 2776 mm per annum. Variation in altitude and rainfall, associated with a landscape of geological and pedological extremes, has created a very diverse flora. The landscape is prone to lightning-induced burning [ 11 ] and is topographically complex, hosting a variety of habitats [ 12 ], including grassland plateaus, wetlands and sponge areas, grassland slopes, afromontane forest, riparian forest, moist woodlands, dry woodlands and shrublands. The vegetation type is classified as the Northeastern Drakensburg High-Mountain Sourveld ecoregion [ 13 ], and the reserve encompasses four veld types: Afromontane Forest, North-eastern Mountain Sour Grassland, Sour Lowveld Bushveld, and Mixed Lowveld Bushveld [ 14 ]. Figure 1 Reserve location. The Blyde River Canyon Nature Reserve is located in northeastern Mpumalanga Province, South Africa. Results Spatial distribution of forests We found a total area of 2111 ha of afromontane forest on BRCNR. This comprises 7.3% of the reserve's total surface area. Figure 2 shows the distribution of the forest on BRCNR. The forest is fragmented, with a total of 60 forest patches recorded, varying from 0.21 ha to 567 ha in size. The number of patches per size class illustrates this level of fragmentation (Figure 3 ). Unfortunately, 15 % of the larger forest patches have their patch size delimited by political boundaries. This means that a few forest patches continued over BRCNR's borders onto adjacent land and only the portions occurring on BRCNR were included in the calculations. If the average patch size is only calculated from patches with natural boundaries within the reserve, then the average size of forest patches is only 23 ha. This small patch size accentuates the fragmented nature of the afromontane forest on BRCNR. Figure 2 Forest distribution. Distribution of afromontane forest fragments in the Blyde River Canyon Nature Reserve. Figure 3 Forest patches per size class. Within the BRCNR, a total of 60 forest patches are recorded, varying from 0.21 ha to 567 ha in size. Flora and forest classification The forest flora recorded in 22 relevés is listed in Appendix A (see additional file 1 ). The total of 167 species includes 38 ferns and fern allies, 3 conifers and 140 flowering plants (11 monocots, 131 dicots). Previous botanical field work and logistical reconnaissance of the forest fragments within the park boundaries showed the forest flora to be qualitatively similar. However, some areas of the fragments were unreachable or technically difficult to visit and study due to the escarpment topography. For purposes of this vegetation analysis, the relevés were located (Figure 4 ) in two of the five largest forest patches, located in the central region of the BRCNR, where forest fragments were at least reasonably accessible. At this time, we consider these relevés to be representative. While the forest flora of BRCNR has not been exhaustively surveyed, this species list represents our best knowledge collected to date. Figure 4 Forest fragment study plots. Distribution of forest study plots (yellow dots) within the Blyde River Canyon Nature Reserve boundaries (red lines) shown on Landsat 7 satellite image. From our analysis of the forest vegetation in the relevés, two main plant communities – moist afromontane forest at high altitude and dry afromontane forest at lower altitude – are identified and described below. An eigenvalue of 0.33 was produced from the TWINSPAN algorithm at the first division into two communities. This value is strong enough to represent beta diversity and the division into two plant communities is accepted. Each community was then further broken down again into two communities, whose eigenvalues were lower and therefore taken as a measure of alpha diversity. The four proposed sub-communities are accepted as variants within the two larger communities. A complimentary analysis of forest communities using DECORANA was carried out and supports the results of the CANOCO analysis. Figure 5 displays the results from the CANOCO analysis. This joint plot data is analyzed for relationships with environmental data. These environmental variables are displayed as arrows radiating from the center of the diagram with the length of the arrows representing each environmental variable's contribution to explaining the variation in the sample scores on each axis. Altitude contributed most of the variation in samples along the bottom axis. Axis 1 is therefore based on an altitudinal gradient. Other environmental variables that contributed towards the floristic variation include drainage, slope, rockiness, shrub density, herb density, climber density and canopy density. The effects of aspect and topography were excluded, as their contribution was small. An altitudinal gradient accounts for most of the variation within the forest communities. Figure 5 Relationship between plant communities and environmental gradients. A CANOCO joint plot showing the relationship between the proposed plant communities and environmental gradients. The length and position of the arrows indicate the strength of the relationship and the direction of change. Moist high-altitude afromontane forest This was a tall forest community that was heavily logged at the turn of the 20 th century. The canopy is still very uneven as the sub-canopy trees have not yet matured or replaced those trees which were originally removed. This forest occurs in the mist-belt along the escarpment at altitudes of between 1450 m and 1700 m. The slopes are steep and generally scattered with large boulders. Preferential species identified for this community include Clivia caulescens , Ochna arborea , Podocarpus latifolius , Rapanea melanophloeos , Vernonia wollastonii and Blechnum attenuatum . Epiphytes are common in these forests. Within this community, two variants are recognized based on canopy development. Degraded moist high-altitude forest A very broken and open canopy with a large shrub component characterizes this forest type. A number of species within this community are shade-intolerant and have established themselves under an open canopy. This sub-community has been so over-utilized that the canopy has never been able to close and this community is sometimes dominated by the exotic invader species Acacia mearnsii . Other preferential species include Rubus spp., Myrsine africana and Helichrysum chrysargyrum . This community is only temporary as it will eventually be replaced by climax species during the next seral stage when the canopy closes. Developing moist high-altitude forest This forest type has a closed canopy with only a few pioneer, shade-intolerant species, such as Rhus tumulicola , Maesa lanceolata and Acacia mearnsii , found in the canopy. The shrub component is not nearly as dense as the degraded sub-community community. Preferential species include Behnia reticulata , Dovyalis lucida , Olea capensis subsp. macrocarpa , Asplenium rutifolium and Jasminum abyssinicum . Dry low-altitude afromontane forest Parts of this forest type were also logged at the turn of the 20 th century. This dry forest community occurs just below the escarpment mist-belt, from 1200 m to 1450 m above sea level. Slopes are steep to gentle, with scattered large boulders. Preferential species in this forest community include Brachylaena transvaalensis , Podocarpus falcatus , Adenopodia spicata , Lauridia tetragona and Pteris captoptera . Within this community, two variants are recognized based on degree of slope. Dry forest on gentle slopes or along drainage lines This forest type occurs in areas with a gentle slope or along drainage lines. Both the shrub and herb layer densities are high with the increase in soil moisture gained directly from drainage lines or as a result of poor drainage on gentle slopes. Differential species include Acacia ataxacantha , Combretum kraussii , Eugenia natalitia , Gymnosporia mossambicensis and Faurea galpinii . Dry forest along steep slopes These low altitude forests occur along steep slopes. Overall canopy density is high with a poorly developed herb layer. This sub-community eventually grades into the moist high-altitude afromontane forest at higher altitudes. Preferential species include Combretum edwardsii , Cryptocarya transvaalensis , Chionanthus peglerae , Xymalos monospora , Oxyanthus speciosus and Prosphytochloa prehensilis . Species responses Our analysis showed that altitude appears to be the primary environmental variable responsible for the distribution of forest plant species along environmental gradients. Specific species responses to environmental variables are depicted in Figure 6 . With 167 species recorded in the plots, only those species with a cumulative fit of more than 0.35 are displayed. Canopy density is used as a surrogate for measuring disturbance, or forest dynamics. It is expected that forest dynamics would have an influence on the distribution of forest species. Figure 6 Species responses to environmental variables. Upper graph is a scatter diagram produced from CANOCO using the CCA procedure, displaying species responses to environmental variables. Lower graph displays strength and position of environmental variables. Asparagus virgatus , Dicliptera clinopodia , Setaria megaphylla and Desmodium repandum are a few of the herbaceous plants that prefer the lower lying forests. Some of the trees include Euclea crispa , Rhamnus prinoides , Combretum kraussii , Myrsine africana , Lauridia tetragona , Protorhus longifolia and Scolopia mundii . Species that occur in the higher lying forests include Oxyanthus speciosus , Xymalos monospora , Rothmannia capensis and Cassipourea malosana . Some of the species that appear to prefer rocky areas include Clivia caulescens and Blechnum attenuatum . On the other hand, Prosphytochloa prehensilis, Rawsonia lucida and Ptaeroxylon obliquum seem to favor areas almost devoid of rocks or boulders. Species that favor low canopy densities, or higher levels of disturbance, are Rubus sp., Blotiella natalensis , Acacia mearnsii and Schefflera umbellifera . Discussion Geldenhuys [ 1 ] states that species diversity within forest patches is determined by patch size and proximity to other forests, which together explain 82% of the species richness in South Africa's forests. The floristic variation within BRCNR's forests can be attributed predominantly to variation in species composition along an altitudinal gradient. Extremes in environmental variables and gradients, associated with a history of over-utilization, have resulted in a floristically and dynamically diverse forest type. Fire exclusion, clearing of old plantation sites, and a history of intensive and selective logging have all contributed to the variation in forest dynamics currently occurring on BRCNR. The afromontane forests occurring on BRCNR are extremely fragmented, and yet over 2100 ha of forest patches are conserved within its boundaries. A large proportion of forest-dependent tree species are offered protection within BRCNR's borders. Of special interest was the discovery of Jasminum abyssinicum , Combretum edwardsii and Olinia radiata , all of which occurred in relatively large numbers in the forest. Jasminum abyssinicum had a status of Insufficiently Known under the old Transvaal threatened plants programme [ 15 ], and Hilton-Taylor [ 16 ] currently lists this species under the Red Data List of South African Plants. Jasminum abyssinicum occurred in 64% of the sample plots and in both forest communities identified. Combretum edwardsii is similarly listed as Insufficiently Known and also occurs under the Red Data List of South African Plants. Combretum edwardsii occurred in 73% of the sample plots. Pooley [ 17 ] lists O. radiata as very rare in KwaZulu Natal and Transkei and fails to mention its occurrence north of Natal. Palmer & Pitman [ 18 ] describe O. radiata as a rare species restricted to Natal forests from Pondoland to Zululand. This species was surprisingly common in the canopy of BRCNR's forests and occurred in 82% of the sample plots. Some severely over-utilized forest patches are currently in a state of recovery; however, current and future damage from invasive trees is a threat to this recovery. Some of the forest patches on BRCNR have a forest margin largely comprised of the invasive black wattle ( Acacia mearnsii ). The flammable nature of this species, compared to natural forest margin species, allows grassland fires to penetrate forests, resulting in a reduction in forest patch size. With a large edge effect resulting from many small forest patches, there is a need for careful fire management and stringent alien plant control. According to Geldenhuys [ 5 ], conservation status implies the extent to which populations, species or communities have been modified by the influences of man and the degree to which they might be expected to maintain their genetic diversity and ecological processes in the medium term (10 to 100 years). We see two different aspects to the conservation of the afromontane forest biome in the BRCNR. Firstly, it is the maintenance of the components and critical processes within a forest ecosystem. The disturbed and unstable state of the forest margins are identified as an area requiring further investigation. The effects of the alien tree species (e.g., black wattle) and the destructive burning of forest margins should be of concern to management authorities, as the forest patches are being reduced in size and the impact of edge effects is being amplified on the forest interior. Secondly, it is the maintenance of gene flow between the fragmented forest patches through management of the land surrounding the forest and forest corridors. As the BRCNR forest vegetation is situated along an altitudinal gradient, it therefore seems possible to identify certain forest patches (which may have been harvested in the past, or possibly will in the future) as critical adjuncts for conserved forest patches at the same altitude. From the evidence we gathered, no "climax" forests exist on BRCNR. Although it is known that BRCNR's forests were utilized, the impact, extent and degree of the utilization are still not quantified. An investigation into the successional and dynamic state of the five largest forest patches is currently underway. Very little of the neighbouring forest on Mariepskop was harvested for its timber, and this forest seems to be the obvious control site for further comparative research. Conclusions This study shows that BRCNR has a fragmented network of small forest patches that together make up 7.3% of the reserve's surface area, almost twice the area that was generally known. Two afromontane forest communities are recognized and associated along an altitudinal gradient, one within the moist mist belt and one within drier micro-climates outside the mist belt; further, within each community, variants were recognized based on either available soil moisture or degree of past utilization. These forest patches host a variety of forest-dependent trees, including some species considered rare, insufficiently known, or listed under the Red Data List of South African Plants. The fragmented nature of the relatively small forest patches accentuates the need for careful fire management and stringent alien plant control. Methods Forest size and fragmentation All forest patches greater than 0.25 ha in size were mapped on a Geographical Information System (GIS). Forest boundaries were marked on 1:10 000 orthophotos, and the data were digitized onto the Mpumalanga Parks Board's GIS program, SPANS [ 19 ]. Total forest size and patch numbers are calculated from the digitized data. Floristic description Twenty-two plots (0.04 ha each) were subjectively distributed throughout two of the five largest forest patches occurring on BRCNR (Figure 4 ), including the Op-de-Berg and Hebronberg forests. Plots covered an altitudinal range from 1240 to 1660 m above sea level and were sampled along environmental gradients, including the factors of slope, surface rockiness, drainage, topography, and disturbance (from logging). Sample relevés included a list of all the species present in a sample plot as well as cover-abundance values for each species, according to the Braun Blanquet cover-abundance scale [ 20 ]. A forest flora (see additional file 1 – Appendix A) is compiled from the species composition lists recorded in the 22 relevés. Data processing and analyses Eigenvalues produced from the TWINSPAN and DECORANA programs are a measure of the degree of separation in the data. According to Jongman, ter Braak & van Tongeren [ 21 ], low eigenvalues would represent a poor separation of samples and can be regarded as a measure of alpha diversity (species turnover within a plant community). High eigenvalues would represent a strong separation of samples, which can therefore be a measure of beta diversity (species turnover between plant communities). The 22 relevés were analyzed for a circumscription of possible plant communities. Firstly, a complementary analysis was run using the hierarchical classification program TWINSPAN [ 22 , 23 ] and the indirect ordination program DECORANA [ 24 , 25 ]. Complementary analyses of the two different multivariate analysis results provide for an accurate interpretation and description of plant communities [ 26 ]. Qualitative cover-abundance values were used for data input instead of quantitative values. Appendix A lists the 167 species recorded in the forest and used in the classification of plant communities. Secondly, the relevés were analyzed for relationships between plant communities and environmental gradients. Canonical correspondence analysis is a direct ordination technique that analyzes and presents such relationships between many species and numerous environmental variables. For this study, we used the program CANOCO [ 27 ]. No formal syntaxonomical classification was done in this study. An informal classification was performed with preferential species indicating the names of different vegetation associations. Specific species responses to environmental variables are a useful way of displaying the impact a certain environmental variable may have on a species. Direct ordinations relate species presence to environmental variables on the basis of species and environmental data from the same set of sample plots [ 28 ]. A scatter diagram produced from the CANOCO program depicts the relationship between species and environmental variables. Only those species with a cumulative fit of more than 0.35 are displayed. This ensures that only those species that most positively contributed to the ordination scores in the scatter plot are displayed. Canopy density is used as a surrogate for measuring disturbance, or forest dynamics. Authors' contributions ML conducted the field work and the map digitization. Both authors performed the statistical analyses. Both authors read and approved the final manuscript. Supplementary Material Additional File 1 Appendix 1. Forest flora in 22 relevés on Blyde River Canyon Nature Reserve Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC512296.xml
340947
The Effect of Learning on the Function of Monkey Extrastriate Visual Cortex
One of the most remarkable capabilities of the adult brain is its ability to learn and continuously adapt to an ever-changing environment. While many studies have documented how learning improves the perception and identification of visual stimuli, relatively little is known about how it modifies the underlying neural mechanisms. We trained monkeys to identify natural images that were degraded by interpolation with visual noise. We found that learning led to an improvement in monkeys' ability to identify these indeterminate visual stimuli. We link this behavioral improvement to a learning-dependent increase in the amount of information communicated by V4 neurons. This increase was mediated by a specific enhancement in neural activity. Our results reveal a mechanism by which learning increases the amount of information that V4 neurons are able to extract from the visual environment. This suggests that V4 plays a key role in resolving indeterminate visual inputs by coordinated interaction between bottom-up and top-down processing streams.
Introduction It is well established that learning can have a strong impact on neural responses to visual stimuli in high-level association cortices such as inferior temporal (IT) or prefrontal (PF) cortex, where the activity of single neurons reflects learning in pair association, object identification, or categorization tasks ( Sakai and Miyashita 1991 ; Logothetis et al. 1995 ; Booth and Rolls 1998 ; Kobatake et al. 1998 ; Erickson and Desimone 1999 ; Rainer and Miller 2000 ; Freedman et al. 2002 ; Sigala and Logothetis 2002 ). In these studies, learning is thought to modify neural activity to represent task-relevant attributes, such as trained views of three dimensional objects ( Logothetis et al. 1995 ) or associations between paired visual stimuli ( Sakai and Miyashita 1991 ; Erickson and Desimone 1999 ). The learned representations often exhibit invariance for stimulus features such as size ( Logothetis et al. 1995 ), rotation ( Booth and Rolls 1998 ), or stimulus degradation ( Rainer and Miller 2000 ). Similar neural activity to within-category stimuli during categorization ( Freedman et al. 2002 ) can also be thought of as a learning-dependent form of invariance. Several lines of evidence suggest that these learning effects involve synaptic plasticity and thus represent long-lasting modifications to visual association cortices. Recent evidence suggests that neurons in early visual sensory areas can also modify their response properties with learning. In particular, several studies have revealed learning-related changes in primary visual cortex (V1) ( Crist et al. 2001 ; Schoups et al. 2001 ; Ghose et al. 2002 ), although the extent and functional significance of these learning effects remains somewhat controversial ( Schoups et al. 2001 ; Ghose et al. 2002 ). Available evidence suggests that classical V1 response properties such as receptive field size or orientation tuning parameters are affected relatively little by learning, while learning does appear to cause general reduction in activity for trained stimuli as well as a task-dependent increase in the influence of nonclassical surround stimulation on the neuron's response. Learning thus appears to affect both low and high level areas of the ventral visual stream. The results obtained by studies in these two areas are, however, difficult to compare directly, owing to substantial differences in experimental design. In studies of IT or PF cortex, studies typically employ ‘complex' visual stimuli such as Fourier descriptors ( Sakai and Miyashita 1991 ), computer-rendered animals ( Freedman et al. 2002 ), or colored photographs and artwork ( Erickson and Desimone 1999 ). These stimuli are generally presented at the center of gaze and can be from 1° up to 10° of visual angle in size. Many studies also include a selection process that determines which of the neurons encountered in a given penetration are chosen for further quantitative study. By contrast, available learning studies in early visual areas follow well-established rules for investigation of primary and extrastriate visual areas. These studies employ ‘simple' visual stimuli such as oriented bars ( Crist et al. 2001 ) or gratings ( Schoups et al. 2001 ; Ghose et al. 2002 ). These stimuli are generally presented at eccentric locations, with stimulation parameters adjusted to the receptive field and orientation selectivity of the single neuron currently under investigation. Thus, both stimulus type and experimental procedure generally differ substantially, depending on whether a study investigates low-level sensory or high-level associative visual cortex. For a comprehensive account of how learning affects visual processing, the same stimuli and experimental procedure must be used to study different levels of the visual processing hierarchy. What kind of stimuli might be suitable to study visual areas as different as early sensory visual and PF cortex? We decided to use natural images for several reasons: The primate visual system evolved in the natural environment under conditions of ‘natural' stimulation; much is known about their statistical properties and they can therefore be well-controlled; they contain structure at all spatial scales and thus can be expected to activate a large fraction of visually responsive neurons. We avoid subjectively biasing our sample of recorded neurons by always recording from the first neurons whose waveforms we are able to reliably isolate. This ensures that our population of recorded neurons represents an unbiased sample in each brain region under study, and this in turn allows us to compare data obtained from different brain regions. We obtain a sensitive measure of behavioral performance and associated neural activity by employing a stimulus degradation procedure that makes stimuli harder to discriminate by adding various amounts of noise (see Figure 1 A). With degradation, stimuli become increasingly indeterminate because all stimuli in a given session are combined with the same noise pattern. Noise is newly generated for every session so that monkeys cannot rely on the specific individual characteristics of a particular noise pattern. Instead, they need to extract task-relevant information from degraded displays, whose particular details vary from day to day. Similarly, outside the laboratory we are rarely presented with familiar stimuli in canonical views and conditions of standard lighting, but instead need to extract this information from complex scenes in which it is embedded. Previously these kind of stimuli were used to study neural activity in the PF cortex ( Rainer and Miller 2000 ), where learning made neural activity more robust to stimulus degradation. After learning, PF neurons tended to fire in a similar manner to undegraded and moderately degraded versions of the same stimulus. Learning thus resulted in a form of neural response invariance, because degradation no longer had an impact on PF neural activity. Figure 1 Stimuli and Behavioral Task (A) An example natural image is shown at three coherence levels, corresponding to 100% (undegraded), 45% (degraded), and 0% (pure visual noise). (B) The sequence of trial events for the DMS task used in this study. After a fixation period, a sample stimulus (S) is briefly presented, followed by a delay period and the presentation of a probe stimulus (P). While sample stimuli were presented at different coherence levels, probe stimuli were always presented in undegraded form (100% coherence). The monkeys were required to release a lever if the probe matched the sample. Here our aim is to use similar stimuli and behavioral procedures to characterize how learning modifies neural activity in extrastriate visual cortical area V4. Area V4 was chosen because it is considered to be a sensory visual area at an intermediary processing stage in the ventral stream and because it is directly connected to parts of the PF cortex ( Petrides and Pandya 1999 ). Our task was a modified version of delayed-matching-to-sample (DMS) (see Figure 1 B). After grasping a metal lever and subsequently attaining central fixation, monkeys viewed a sample stimulus presented at one of six coherence levels ranging from undegraded (100% coherence) to fully degraded (0% coherence). After a brief delay, monkeys were presented with a probe stimulus (always at 100% coherence) and had to release a lever if the probe matched the sample (i.e., if the sample was identical to or was a degraded version of the probe stimulus). During each session, we employed four highly familiar stimuli and four ‘novel' stimuli that monkeys had not seen previously. Great care was taken to ensure that novel and familiar images differed only in terms of their familiarity to the animal (see Materials and Methods ). Using novel and familiar stimuli allowed us to ask whether learning had any effect on monkeys' ability to identify degraded and undegraded versions of natural images. Intermixing novel and familiar images in the same session had the additional advantage of allowing us to estimate for each single neuron in our population, whether there were any learning-related changes in the amount of stimulus-specific information these neurons communicated. Results We found that learning resulted in significant and robust improvements in monkeys' ability to identify degraded stimuli. Behavioral performance varied systematically with coherence ( Figure 2 A). Monkeys performed at chance level (50% correct) when stimuli were presented at 0% coherence and thus contained no task-relevant information. For degraded stimuli (35%–65% correct), monkeys performed significantly better with familiar than with novel stimuli ( t -test, p < 0.01). For undegraded stimuli at 100% coherence, the monkeys' performance was near ceiling for both novel and familiar stimuli (92% and 95% respectively; t-test, p = 0.12). Learning-dependent performance improvements for degraded stimuli were highly consistent across stimuli and monkeys. There were in fact no significant differences in the monkeys' performance to each of the familiar stimuli across sessions at all coherence levels (one-way ANOVAs, p > 0.1), and this was also true for novel stimuli. In addition, performance for novel and familiar stimuli did not differ significantly between the two monkeys at any coherence level (t-tests, p > 0.1). Note that the monkeys' excellent perfor-mance with undegraded novel objects reflects the fact that they have acquired the rule of the DMS task and are thus able to perform it near ceiling with novel stimuli. The timecourse of this learning-dependent difference in performance is shown in Figure 2 B. Session 1 represents a session in which a set of four initially novel stimuli is arbitrarily chosen and kept constant in subsequent sessions, thus becoming more and more familiar. Comparing performance for these stimuli with performance of novel stimuli that are randomly chosen in each session reveals that it takes several sessions for the learning effect to appear. Performance averaged across the first five session was similar for novel and familiar stimuli ( t -test, p = 0.43). Furthermore, the learning-dependent difference in performance appeared to asymptote after around ten sessions. In summary, learning led to robust improvements in the monkeys' ability to identify degraded natural images while the monkeys performed near ceiling for novel and familiar undegraded images. Figure 2 Learning Improved Monkeys' Ability to Identify Degraded Stimuli (A) Behavioral performance for the sessions during which neural data was collected ( n = 11) is shown as a function of the coherence of the sample stimulus for novel and familiar stimuli. Asterisks denote significant differences in performance for novel and familiar stimuli. (B) The performance at 45% coherence (%Correct 45 ) is shown for a set of novel stimuli that is introduced in the first session and then used during all subsequent sessions and thus becomes more and more familiar during subsequent sessions (circles). For comparison, performance with novel stimuli that are new and unique to each session is shown (diamonds). Sessions 1–20 represent purely behavioral training sessions (TRAIN), and sessions 21–26 represent combined behavioral and single unit recording sessions (REC). We now examine whether there were any learning-depen-dent changes in the activity of V4 neurons. Results described in this report are based on a population of 83 V4 neurons. We first asked whether there was any general difference in mean activity elicited by novel and familiar undegraded stimuli. We found that the response of V4 neurons to novel (〈FR nov 〉 = 36.7 ± 2.8 Hz) and familiar stimuli (〈FR fam 〉 = 34.2 ± 2.7 Hz) was similar ( t -test, p = 0.14; see also Table 1 ). Out of the 14 neurons that individually showed a significant difference in activity between novel and familiar stimuli ( t -test, p < 0.05), similar fractions preferred familiar or novel stimuli (6/14 or 43% and 8/14 or 57% respectively; χ 2 test, p = 0.45). We thus found that learning did not lead to a change in the average activity of V4 neurons for undegraded stimuli. Next, we examined whether learning resulted in any change in the amount of stimulus-specific information that V4 neurons communicated. To do this, we computed the mutual information between the set of four familiar or novel stimuli and the associated neural responses ( see Materials and Methods ). We found that V4 neurons on average communicated similar amounts of information about novel and familiar undegraded stimuli ( Figure 3 A). The average information communicated by each neuron in the entire population of 83 V4 neurons was similar for novel stimuli 〈I nov 〉 = 0.48 bits and for familiar stimuli 〈I fam 〉 = 0.45 bits ( t -test, p = 0.16). We selected 25% of the population (21 out of 83 neurons), which communicated most information about novel or familiar objects ( see Materials and Methods ). For this population of most informative neurons (white circles in Figure 3 A), we also found no difference between novel and familiar stimuli (〈I nov 〉 = 0.67 bits, 〈I fam 〉 = 0.65 bits; t -test, p = 0.48). Thus, for natural images (undegraded stimuli) we saw no significant learning-dependent difference in performance and also no learning-dependent changes in the average activity or in the amount of stimulus-specific information communicated by V4 neurons. Figure 3 Learning Led to an Increase in V4 Neural Information about Degraded but Not Undegraded Stimuli Here we summarize how much information V4 neurons communicated about novel (I nov ) and familiar (I fam ) stimuli for undegraded (A) and degraded (B) stimuli. Each symbol in the scatter plot represents a single neuron and shows how much information this neuron communicated about familiar (x-axis) and novel (y-axis) stimuli. In each scatter plot, white-shaded symbols represent the 25% most informative neurons, i.e., the one-quarter of the population communicating most information about either familiar or novel stimuli. The remaining three-quarters of the population are shown in gray shading. The single neuron example in Figure 5 is represented by the star. The black ‘x' represents the population mean for the 25% most informative neurons. Table 1 Mean Sample-Evoked Activity and Information Values This table summarizes the mean sample-evoked activity and information values for the entire population and for the set of informative V4 neurons (open circles in Figure 3 B) as a function of degradation and learning. Significance of paired t -tests comparing estimated values for familiar versus novel stimuli are reported to the right of each pair of values NS, not significant; * p < 0.05; ** p < 0.001 At intermediate coherence levels, the monkeys' ability to correctly identify degraded stimuli was improved by learning, and we asked whether this behavioral improvement was associated with any changes in the activity of V4 neurons. We found that V4 neurons indeed communicated significantly more information about degraded familiar than about degraded novel stimuli ( Figure 3 B). Considering the entire population, learning led to a significant increase in information about degraded stimuli from 〈I nov 〉 = 0.34 bits to 〈I fam 〉 = 0.40 bits ( t -test, p < 0.05). For the 25% most informative neurons (white circles in Figure 3 B), we observed an even larger change from 〈I nov 〉 = 0.47 bits to 〈I fam 〉 = 0.67 bits ( t -test, p < 0.001), corresponding to a 40% increase in information with learning. We further characterized this effect by examining how degradation affected the amount of information separately for novel and familiar stimuli. For both novel ( Figure 4 A) and familiar ( Figure 4 B) stimuli, V4 neurons communicated on average more information about undegraded (I 100 ) than degraded (I degrad ) stimuli (paired t -tests, p < 0.001), reflecting the fact that behavioral perfor-mance was better for undegraded than degraded stimuli (see Figure 2 A). The ΔI distributions (I 100 − I degrad ) for familiar and novel stimuli shown in the insets ( Figure 4 A and 4 B), however, differed significantly (paired t -test, p < 0.001), and learning was associated with a rightward shift in this distribution (〈ΔI fam 〉 = 0.06, 〈ΔI nov 〉 = 0.13). Interestingly, the kurtosis or skewness of the ΔI distribution changed by an order of magnitude from 0.13 for novel stimuli to 5.5 for familiar stimuli, similar to experience-dependent effects that have been observed on hippocampal place cell activity ( Mehta et al. 2000 ; Mehta 2001 ). As a consequence of these learning-dependent changes, many V4 neurons actually communicated more information about degraded than undegraded familiar stimuli (25/83 or 30%), whereas only a small minority did so for novel stimuli (6/83 or 7%). This difference in proportions was significant (χ 2 test, p < 0.001). Taken together, learning accordingly resulted in an increase in the amount of information communicated by V4 neurons about degraded stimuli and many neurons actually communicated more information about degraded than undegraded familiar stimuli. Figure 4 Many Neurons Communicated More Information about Degraded than Undegraded Familiar Stimuli Here we replot the data from Figure 3 to illustrate how much information V4 neurons communicated about degraded (I degrad ) and undegraded (I 100 ) stimuli separately for novel (A) and familiar (B) stimuli. Each symbol in the scatter plot represents a single neuron. The insets show how degradation affected the information communicated by V4 neurons, by plotting histograms of the ΔI distributions (I 100 − I degrad ) for novel and familiar stimuli. While 25 neurons (30% of the population) communicated more information about degraded than undegraded familiar stimuli, only six neurons (7% of the population) did so for novel stimuli. How did single V4 neurons mediate this learning-depen-dent increase in information about degraded stimuli? The activity of an example neuron is shown in Figure 5 in histogram and raster format for its preferred and nonpreferred familiar stimulus. This neuron showed little or no response to pure visual noise (0% coherence) or to its nonpreferred stimulus at any coherence level ( Figure 5 B). It was activated to a peak firing rate of about 20Hz by its preferred stimulus (red curve in Figure 5 A). Degradation of the preferred stimulus resulted in brisk activity of this neuron, and activity was greater to the preferred stimulus at all intermediate coherence levels (35%–65%) than to the undegraded preferred stimulus (paired t -tests, p < 0.01). For this neuron (see star in Figure 3 ), degradation resulted in a large increase in information about familiar stimuli from I 100 = 0.18 bits to I degrad = 0.74 bits. This example neuron thus displayed a nonmonotonic, inverted U-shaped response as a function of degradation. The responses of this neuron for the preferred and nonpreferred familiar stimuli and also for the corresponding novel stimuli are summarized in Figure 5 C. While the preferred novel undegraded stimulus also activated the neuron, degradation of this stimulus was not associated with significant response enhancement. To examine whether the inverted U-shaped response was in fact characteristic of the V4 neurons that communicated most information about degraded stimuli, we plotted the activity of the neurons which were highly selective for degraded stimuli (see white circles in Figure 3 B), as a function of coherence, using the preferred stimulus for each neuron ( Figure 6 ). We found that across this population, neural activity was indeed significantly enhanced for familiar stimuli at intermediate coherence levels of 55% and 65% relative to activity to undegraded familiar stimuli (paired t -tests: p < 0.05). By contrast, activity to novel stimuli systematically decreased with degradation and was significantly below activity to undegraded stimuli at coherence levels of 35% and 45% (paired t -tests, p < 0.05). As expected, V4 neurons generally showed greater activity to novel and familiar stimuli than to pure noise at 0% coherence (paired t -tests, p < 0.05). As detailed in Table 1 , mean activity was similar for undegraded familiar and novel stimuli, but significantly greater for degraded familiar than degraded novel stimuli (paired t -test, p < 0.05). Taken together, learning resulted in an increase in information communicated by V4 neurons about degraded or indeterminate stimuli. This increase in information was mediated by neurons that showed an enhancement in neural activity to degraded compared to undegraded familiar stimuli. Figure 5 Learning-Dependent Enhancement for Degraded Stimuli—Single Neuron Example (A and B) The activity for an example neuron for its preferred (A) and nonpreferred (B) familiar stimulus is shown in peri-stimulus-time-histogram (PSTH) and raster format. (C) The average firing rate during stimulus presentation as a function of coherence is summarized for this neuron for its preferred (+) and nonpreferred (−) familiar (fam) and novel (nov) stimuli. Figure 6 Learning-Dependent Enhancement for Degraded Stimuli—Population Activity These panels show the activity of neurons that communicated most information about degraded stimuli (i.e., white-shaded symbols in Figure 3 B) as a function of degradation for familiar (A and B) and novel (C and D) stimuli. The preferred stimulus was used for each neuron. The left column shows activity in PSTH format and the right column shows the mean stimulus-evoked activity at each coherence level; asterisks denote significant differences between activity at each coherence level and activity to undegraded stimuli at 100% coherence (paired t -tests, p < 0.05). We performed additional behavioral experiments to assess whether learning led to any changes in fixational eye movements, because such changes might shed light on what mediates monkeys' behavioral advantages for familiar degraded stimuli. In these studies, we allowed the monkeys to freely view sample stimuli during task performance and then estimated a fixation probability map (FPM) for each familiar and novel stimulus presented at 45% and 100% coherence ( see Materials and Methods . We applied a threshold to this map to identify regions where monkeys tended to fixate with high probability. The thresholded FPMs for 45% and 100% coherence versions of an example familiar and novel stimulus, along with the overlap between these regions, are shown in Figure 7 . As can be seen, there was substantially more overlap between the regions of focused eye position at 45% and at 100% after learning. This effect was significant across sessions and stimuli: On average, the overlap region increased by a factor of 2.8 from 0.54 ± 0.14 dva 2 (degrees of visual angle squared) for novel stimuli to 1.47 ± 0.16 dva 2 for familiar stimuli (unpaired t -test, p < 0.0001). There were also significant learning-dependent increases in the high-probability FPM areas at 45% and 100% coherence (at 45% from 1.04 ± 0.25 dva 2 to 1.88 ± 0.19 dva 2 , unpaired t -test, p < 0.01; at 100% from 0.84 ± 0.21 dva 2 to 1.74 ± 0.21 dva 2 , unpaired t -test, p < 0.01). This learning-dependent increase in the high probability FPM regions and their overlap was highly consistent across sessions and monkeys, and we observed it during all six sessions in both monkeys. Note that the lower FPM values for novel stimuli indicate that eye position was less focused and therefore more distributed before learning, whereas for familiar stimuli robust regions of focused eye position developed. Figure 7 Eye Movement Analysis during Free Viewing Regions of high fixation probability during free viewing of an example familiar and novel stimulus are shown. Monkeys viewed stimuli at 100% coherence (red-shaded regions) and at 45% coherence (yellow-shaded regions). The green-shaded regions represent regions with high fixation probability at both 45% and 100% coherence. Discussion V4 neurons are generally conceptualized as detectors of visual features of intermediate complexity, such as non-Cartesian gratings ( Gallant et al. 1996 ) or contour features ( Pasupathy and Connor 1999 ). We have found that learning does not affect how V4 neurons respond to undegraded natural images, both in terms of mean firing rate and information communicated about these stimuli. This absence of learning-dependent differences suggests that this V4 selectivity for features of intermediate complexity is not modified by learning, at least during the several weeks of training in the adult monkey during our task. Basic response properties of V4 neurons thus appear not be altered by learning, similar to findings in V1 that have found that parameters such as receptive field size or orientation tuning width remain unchanged even after extensive training ( Crist et al. 2001 ). Learning does however lead to robust changes in how V4 neurons respond in the presence of degradation. For novel stimuli, V4 neurons tend to act as simple passive feature detectors for which the addition of increasing amounts of noise to the display results in successive reduction in neural activity. Consistent with this finding, we observed a systematic decrease of blood-oxygen level-dependent (BOLD) levels with decreasing stimulus coherence in area V4 of anesthetized monkeys using novel stimuli ( Rainer et al. 2001 ) . After learning, many V4 neurons showed increased activity with the degradation of familiar stimuli, suggesting that they were specifically recruited for difficult discriminations involving the processing of these indeterminate visual inputs. The extraction and amplification of task-relevant elements from visual scenes is a key problem of intermediate-level vision. Our results suggest that V4 neurons play a crucial part in resolving indeterminate visual stimuli and signaling the presence of salient stimulus features. Consistent with this interpretation, studies have found that deactivation or ablation of V4 in monkeys has little impact on basic visual functions, but severely affects shape discrimination ( Girard et al. 2002 ), the identification of images that are occluded or have incomplete contour information ( Schiller 1995 ) or the visual selection in the presence of salient distracters ( De Weerd et al. 1999 ). A recent study found severe deficits after V4 ablation in tasks that required making judgments about oriented line segments embedded in distracter arrays ( Merigan 2000 ), a task that has many similarities to the extraction of task-relevant features from degraded displays in our study. We suggest that lesion-induced deficits are a result of disrupting the operation of V4 neurons which are engaged in selective amplification of task-relevant elements of the visual scene. This idea is consistent with our analysis of eye movements, because monkeys focused more reliably on particular stimulus features for familiar than for novel stimuli. This raises the possibility that allocation of focused attention during task performance under central fixation might have contributed to our results, since attention can greatly enhance the response of V4 neurons to visual stimulation ( Moran and Desimone 1985 ; Connor et al. 1997 ). Indeed, we suggest that the enhancement in activity and information about degraded familiar stimuli can be conceptualized as a learning-dependent form of attention. Our findings in V4 are in stark contrast to data obtained in the PF cortex using similar task and stimuli ( Rainer and Miller 2000 ). In the PF cortex, learning resulted in qualitatively different changes in neural activity. Learning resulted in a robust reduction in average neural activity to undegraded stimuli in PF cortex, whereas we found no general differences in activity in V4. This implies that while PF cortex may play a particularly important role in processing novel stimuli ( Ranganath and Rainer 2003 ), extrastriate visual areas communicate feature-specific information largely in the absence of learning-related changes for easy-to-discriminate stimuli. Learning led to neural response invariance across degradation in the PF cortex: neurons that responded differentially to two stimuli maintained this response difference for degraded stimuli after learning, whereas the difference in neural response collapsed with degradation for novel stimuli. Response invariance across degradation implies that the PF cortex does not differentiate between degraded and undegraded versions of a stimulus. Learning thus builds response invariance in the PF cortex. In V4, we found that learning led to a selective enhancement of activity for degraded stimuli over and above the response for undegraded stimuli. While PF neurons showed invariant activity, V4 neurons showed inverted U-shaped noise tuning and were thus most active during difficult discriminations, showing responses consistent with selective amplification of feature-specific activity. Our results suggest that the enhancement observed in V4 may be instrumental in establishing invariance in PF cortex and that interaction between these areas may be required to maintain it. Further experiments using simultaneous recordings from both regions are needed to directly test such a hypothesis. Several studies have identified learning-dependent increases in BOLD signals in extrastriate and temporal visual areas ( Dolan et al. 1997 ; Grill-Spector et al. 2000 ). Because BOLD measures aggregate activation across many neurons, these studies cannot dissociate whether learning-dependent increases are due to building of invariance or selective enhancement of a subpopulation of neurons. This kind of question is certainly important for characterizing functional properties of brain regions and can be answered definitively only by detailed comparison of neural population activity with simultaneously acquired BOLD signal ( Logothetis et al. 1999 , 2001 ). The task dependence of learning effects in V1 ( Gilbert 1998 ; Gilbert et al. 2001 ) has been taken as evidence that top-down modulation plays an important role in the learning-dependent modifications seen in V1 neurons and that, accordingly, these changes are reflections of plasticity in higher areas of the visual system. Our findings are certainly consistent with this view and suggest that vision is an active process involving recurrent interaction of different brain regions rather than a purely feed-forward process ( Thorpe et al. 1996 ), although our data are consistent with largely feed-forward processing for familiar undegraded stimuli. A possible biophysical mechanism for this interaction was identified by a recent study, which demonstrated that subthreshold activation of the distal apical dendrite of layer V pyramidal neurons can greatly enhance their response to more proximal inputs ( Larkum et al. 1999 ). Because feedback projections from higher cortical areas tend to arrive in upper cortical layers, this represents a mechanism by which feedback could exert control over activity in sensory cortices ( Siegel et al. 2000 ) and thus contribute to the inverted U-shaped responses observed in the present study. Several computational models have investigated how brain regions might interact during stimulus identification. A key feature of such models is the interaction between bottom-up and top-down processing ( Carpenter and Grossberg 1987 ; Ullman 1995 ). Consider a neuron in an intermediate visual area such as V4, receiving bottom-up feature-tuned visual input from visual areas lower in the hierarchy and top-down feedback from higher areas representing possible interpretations of the stimulus. It has been hypothesized that a match between top-down and bottom-up inputs could result in elevated activity or nonlinear response enhancement. We have observed such enhancement for familiar but not for novel stimuli, indicating that learning plays a critical role in facilitating interaction between top-down and bottom-up processing streams. Another type of model has suggested that top-down feedback may represent a predictive code, where top-down signals effectively cancel predictable responses in the bottom-up signal ( Mumford 1992 ; Rao and Ballard 1999 ). In this scheme, activity would be reduced for undegraded stimuli because it can be accurately ‘predicted away' by higher level areas. Degraded stimuli containing noise might not be accurately predicted, leaving more remaining activity compared to undegraded images. However, based on this model, one would predict lower activity for familiar than for novel degraded stimuli, because more of the familiar stimuli can be predicted away—exactly the opposite of what we have found. Thus, our results are more consistent with theories that conceptualize top-down feedback as high-level stimulus interpretations rather than as an error signals. Materials and Methods Behavioral and electrophysiological methods. Two adult male rhesus monkeys ( Macaca mulatta ) participated in these experiments. All studies were approved by local authorities and were in full compliance with applicable guidelines (EUVD 86/609/EEC) for the care and use of laboratory animals. Stimuli were 10° × 10° in size, 24-bit color depth, and presented at the center of gaze on a γ-corrected 21-inch monitor with linear display characteristics placed at a distance of 97 cm from the monkeys. Stimuli were generated using Fourier techniques that have been described in detail elsewhere ( Rainer et al. 2001 ). In brief, a large set of natural images was first normalized to have identical Fourier amplitude spectra. Degraded versions of natural images were generated mixing the Fourier phase spectra of natural images with a random phase spectrum corresponding to visual noise, independently for each of the RGB color channels. A different random phase spectrum was used during each session, and it was mixed with all images used during that session. Each trial began when the monkey grasped a lever and then acquired fixation on a central fixation point. After 1000 ms, a sample stimulus was presented for 320 ms, which could be any one of eight different images at six coherence levels (0%, 35%, 45%, 55%, 65%, and 100%). After a delay of 1000 ms, a probe stimulus was presented for 600 ms, which could be any one of the eight undegraded images (100% coherence). The monkeys were required to release the lever if the probe matched the sample (i.e., if the sample had been identical to or a degraded version of the probe). In case of a nonmatch, a second brief delay (200 ms) followed the probe, and this delay was always terminated by the presentation of the correct matching stimulus, ensuring that monkeys had to make a behavioral response on every trial. The monkeys were rewarded with apple juice for making correct responses and were rewarded randomly at 0% coherence where the sample contained no task-relevant information. During each session, the monkeys performed the task with a set of four familiar stimuli, with which they had many weeks of practice, as well as with a set of four novel stimuli that they had never seen before. Matches occurred on 50% of trials; the other 50% were non-matches selected randomly from the remaining stimuli. Owing to the normalization procedure, familiar and novel stimuli did not differ in terms of low-level characteristics of spatial frequency content and image intensity. Familiar stimuli from four categories were used (faces, flowers, birds, and landscapes), and one of the four novel stimuli also came from each of these categories. Fixation was monitored with a scleral search coil and sampled at 200Hz (CNC Engineering, Enfield, Connecticut, United States), and the monkeys were required to maintain fixation within a ±1.25° window at all times during the trial. The monkeys completed at least ten trials per condition during each session. Recordings were made from V4 using standard electrophysiological techniques. We employed a grid system (CRIST, Damascus, Maryland, United States) with eight tungsten microelectrodes (FHC Inc., Bowdoinham, Maine, United States). Preoperative magnetic resonance imaging (MRI) was used to identify the stereotaxic coordinates of V4, which was then covered by a recording chamber. To ensure an unbiased estimate of neural activity, we made no attempt to select neurons based on task selectivity. Instead, we advanced each electrode until the activity of one or more neurons was well isolated and then began collecting data. Comparison of the monkeys' performance during the last six training sessions to performance during recording sessions revealed that performance was unchanged for novel objects ( t -test, p = 0.87), but significantly lower during recording sessions for familiar stimuli ( t -test, p < 0.01), likely owing to nonspecific factors such as additional wait periods during these sessions. Eye movement analysis. To determine whether there were any learning-related changes in the monkeys' fixational eye movements, we performed separate behavioral experiments in which we allowed the monkeys to freely view the sample stimulus for a period of about 2 s. As before, we presented four familiar and four novel stimuli during each session, but we only used two coherence levels of 100% and 45% to allow us to assess whether learning led to changes both for degraded and undegraded stimuli. Monkeys performed around 20 trials for each stimulus at each degradation levels during each session, and we report here the results from a total of six sessions. We identified periods of fixation during free-viewing as periods as periods of at least 100 ms duration during which eye position did not change by more than 0.3°. We then marked off a region of 0.3° × 0.3° around this position and superimposed these regions for all fixations during all relevant trials. By normalizing the volume under this function to a value of 1, we created an FPM for each stimulus. We then applied a single threshold to the FPM for all stimuli and degradation levels. The threshold η was chosen to be an order of magnitude greater than the FPM value corresponding to randomly distributed eye position, i.e., to a value of η = 10 × 1/256 2 , and these areas were converted to degrees squared of visual angle (dva 2 ). The thresholded FPMs shown in Figure 7 depict the regions of the FPM that passed threshold for each of the two stimuli during an example session and thus represent the foci of eye position or regions of high fixation probability for that stimulus. Because FPMs are all normalized, a small or absent thresholded FPM region indicates that eye position was distributed on the stimulus without a clear focus. Note that for familiar stimuli, thresholded FPMs were highly consistent across sessions confirming the robustness of this measure. Data analysis. Neural activity was assessed during a fixed period of 310 ms duration, beginning 50 ms after the onset of the visual stimulus to take response latency into account. Such a period roughly corresponds to the time between saccades during natural viewing conditions. Out of a total population of 116 neurons, 83 task-related neurons were identified as showing significant differences in activity between any of the eight stimuli at any coherence level using a Bonferroni-corrected t -test evaluated at p < 0.05. Mean firing rates, reported in Table 1 , were computed using the preferred stimulus for each neuron. To assess whether learning had any systematic effect on the amount of stimulus-specific information communicated by V4 neurons, we quantified how much information was contained in the pattern of neural firing rates about novel and familiar stimuli separately. This quantity is given by the mutual information between the set of four familiar or novel stimuli and the set of associated firing rates ( Shannon 1948 ). We thus computed the mutual information (I) among the set of stimuli ( s ) and the neural responses ( r ): where P ( s ) is the probability of showing stimulus s , P ( r | s ) is the probability of observing a response r when stimulus s is presented, and P ( r ) is the probability of observing response r . Because calculation of information requires many trials, we computed information for two conditions: degraded and undegraded stimuli. For degraded stimuli, we pooled the coherence levels from 35% to 65%. For undegraded stimuli, we estimated the mutual information for 100% coherence stimuli during the sample period as well as during the probe period on nonmatch trials (to exclude possible movement-related activity). We report here estimates during the probe period because they are based more trials, but results were similar for the sample period. This ensured that information measures for degraded and undegraded stimuli were based in a similar number of trials. For each neuron we estimated four different information values, describing how much stimulus-specific information was contained in its firing rate distributions about undegraded, as well as degraded, familiar (I fam,100 , I fam,degrad ) and novel (I nov,100 , I nov,degrad ) stimuli. Note that although across all sessions we employed many more novel than familiar stimuli, each individual neuron from which we recorded during a given session ‘saw' exactly the same number of four familiar and four novel stimuli. We identified highly selective neurons in each population by selecting the 25% neurons that communicated most stimulus information about either novel or familiar stimuli ( n = 21 out of 83 neurons total); i.e., we chose the top 25% of the distribution max(I fam ,I nov ). We did this because, owing to our unbiased procedure, our sample contains neurons that did not communicate large amounts of information, and we thus wanted to establish that our conclusions also applied to the neurons that communicated most information. These neurons are shown as white filled circles in Figure 3 A and 3 C, whereas the remaining 75% of neurons ( n = 62) are shown as gray filled circles. There was significant overlap (13/21, 62%) between the populations of informative neurons for degraded and undegraded stimuli (χ 2 test, p < 0.05), indicating that the majority of neurons that were informative for undegraded stimuli were also informative for degraded stimuli. There were no significant differences between informative neurons and the entire population in terms of mean firing rate. Unless otherwise noted, we used paired t -tests to compare information measures obtained for novel and familiar stimuli.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC340947.xml
512282
The Retinome – Defining a reference transcriptome of the adult mammalian retina/retinal pigment epithelium
Background The mammalian retina is a valuable model system to study neuronal biology in health and disease. To obtain insight into intrinsic processes of the retina, great efforts are directed towards the identification and characterization of transcripts with functional relevance to this tissue. Results With the goal to assemble a first genome-wide reference transcriptome of the adult mammalian retina, referred to as the retinome, we have extracted 13,037 non-redundant annotated genes from nearly 500,000 published datasets on redundant retina/retinal pigment epithelium (RPE) transcripts. The data were generated from 27 independent studies employing a wide range of molecular and biocomputational approaches. Comparison to known retina-/RPE-specific pathways and established retinal gene networks suggest that the reference retinome may represent up to 90% of the retinal transcripts. We show that the distribution of retinal genes along the chromosomes is not random but exhibits a higher order organization closely following the previously observed clustering of genes with increased expression. Conclusion The genome wide retinome map offers a rational basis for selecting suggestive candidate genes for hereditary as well as complex retinal diseases facilitating elaborate studies into normal and pathological pathways. To make this unique resource freely available we have built a database providing a query interface to the reference retinome [ 1 ].
Background The mammalian retina is a highly structured tissue developmentally originating from neuroectodermal evagination of the diencephalon and subsequent invagination processes resulting in the formation of two cellular layers which ultimately give rise to the inner neural retina and the outer retinal pigment epithelium (RPE) monolayer [ 2 ]. In the adult, the neural retina consists of approximately 55 distinct cell types histologically structured into three layers of cells (photoreceptors, intermediate neurons and ganglion cells) and two layers of neuronal interconnections (outer and inner plexiform layers) [ 3 ]. The RPE is differentiated into polarized cells with an apical and a basal orientation separating the neural retina from the underlying choroidal blood supply. With its apical microvilli-like processes, the RPE establishes an intimate contact with the photoreceptor outer segments to sustain their metabolic support and maintain photoreceptor integrity [ 4 ]. Together, the neural retina and the RPE provide the structural and functional basis for light perception by ensuring the capture of photons, the conversion of light stimuli into complex patterns of neuronal impulses and the transmission of the initially processed signals to the higher visual centers of the brain. Recent progress in retinal research has greatly enhanced our current understanding of basic functional processes in the adult retina (e.g. [ 4 , 5 ]). A great deal of effort has focused on the molecular dissection of the phototransduction pathway and the retinoid cycle (e.g. ref. [ 6 ]). Besides elucidating physiological mechanisms in normal tissue, the identification of genes involved in hereditary retinal disease has provided another valuable source of insight into functional pathways of the retina and the RPE (reviewed in [ 7 , 8 ]). Despite these advances, a remaining challenge is to obtain a reference genome-wide expression map of the retina/RPE transcriptome, further facilitating the identification of retinal susceptibility genes, but most importantly, offering an invaluable resource for functional genomics studies. Initial analyses of human [ 9 , 10 ] and mouse [ 11 ] whole genome sequences and the use of more recent comparative gene prediction algorithms [ 12 , 13 ] suggest an overall number of mammalian gene loci in the range of 35,000 to 45,000. These estimates have largely been validated by experimental data on gene transcription [ 14 , 15 ] although alternative promoter usage, differential exon splicing during mRNA maturation, alternative usage of polyadenylation sites and other post-transcriptional modifications may further increase the genetic diversity required to encode the full complement of cellular transcripts [ 16 , 17 ]. In addition, there may be a considerable number of non-coding genes unaccounted for by current annotations [ 18 ]. In recent years, a number of approaches and technologies were adopted to identify genes expressed in the retina/RPE of human, cow, dog and mouse including data-mining and assembly of publically available expressed sequence tag (EST) information [ 19 - 23 ], sequencing of cDNA libraries generated via conventional methods [ 24 - 29 ] or via normalization techniques [ 30 , 31 ], hybridization to gene arrays of various formats [ 32 , 33 ] and serial analysis of gene expression (SAGE) [ 34 , 35 ]. Suppression subtractive hybridization (SSH) has been shown to be an efficient technique with which differentially expressed genes can be normalized and enriched over 1000-fold in a single round of hybridization [ 36 ]. Subsequently, applications of SSH to identify retina and RPE-enriched genes have been reported [ 37 - 39 ]. Based on a comprehensive survey of data available from 27 independent studies applying a wide spectrum of gene identification approaches we have now assembled a first genome-wide reference transcriptome of the adult mammalian retina/RPE. This reference transcriptome comprises 13,037 non-redundant transcripts and likely reflects up to 90% of the mammalian retinome. Results A total of 481,137 primary datasets on gene transcripts from the adult mammalian retina/RPE tissues have been generated in 27 independent studies (Table 1 ). Of these, 52,630 datasets (31,814 from retina, 11,632 from RPE and 9,184 from retina/RPE) were available and attributable to unique LocusLink identifiers (IDs). Correcting for gene redundancy within and between studies yielded a catalogue of 15,645 retinal/RPE genes. A survey of incidence and origin of each of these genes in the various studies analyzed demonstrated that 2,608 transcripts were found only once (see additional File 1 ) while the remaining 13,037 genes (see additional File 2 ) were confirmed in at least two and up to 16 independent gene identification approaches (Table 2 ). Thus, the latter compilation of genes may represent a more conservative description of the retinome minimizing a potential bias in data ascertainment. Of the 13K retinome, 1,411 genes were solely identified in retinal studies (see additional File 3 ) while 246 genes were exclusively found in the RPE datasets (see additional File 4 ). Table 1 Studies identifying adult mammalian retina / RPE transcripts and details on gene data retrieval Reference Source / method a) Species b) Original dataset No. of genes retrieved c) (non-redundant) No. of genes identified in ≥ 2 studies In-silico projects [20] Retina, TIGR (ID: version 3.3) Hs 1,047 30 30 [19] Retina, TIGR (ID: version 3.3) Hs 1,315 11 11 [21] Retina, UniGene (ID: build 118) Hs 4,974 1,485 1,480 [22] Retina, UniGene (ID: build 113) Hs 6,190 46 45 [23] Retina, EST (dbEST) Hs 40,000 117 110 UniGene (Build 162) d) Retina Hs 3,560 1,612 1,211 UniGene (Build 162) d) RPE Hs 1,760 1,506 1,116 UniGene (Build 162) d) Retina & RPE Hs 11,976 9,178 9,178 cDNA library sequencing [24] Fovea, conventional Hs 209 40 38 [25] Retina, conventional Hs 607 475 465 [37] Retina & RPE, SSH Hs 401 6 6 [30] Retina, subtracted Hs 137 49 48 [31] Retina, PC and subtracted Cf 173 66 65 [26] RPE, primary and amplified, conventional Hs 2,101 336 330 [38] RPE, SSH Bt 1,000 35 34 [27] and online e) Retina, PC Hs 2,701 2,096 2,059 [28] and online f) RPE, PC Hs 6,182 3,657 3,608 [29] Retina, PC (ID: MRA) Mm 1,793 421 412 [39] Retina, SSH Hs 1,080 310 301 [39] RPE, SSH Bt 2,350 343 329 Microarray Analysis [32] Retina, Affymetrix (ID: Mu11K subB) Mm 6,540 67 67 [33] Retina, custom array Hs 10,034 530 508 Serial Analysis of Gene Expression [34] Retina, adult Mm 54,009 4,233 3,974 [35] Retina (ID: HMAC2) Hs 102,359 7,269 6,919 [35] Retina (ID: HPR1) Hs 59,661 5,689 5,512 [35] Retina (ID: HPR2) Hs 105,312 7,268 6,955 [35] RPE (ID: HRPE1) Hs 53,666 5,755 5,211 Total (with inter-study redundancy) 52,630 50,022 Total (without redundancy) 15,645 13,037 a) TIGR: The Institute for Genomic Research; EST: expressed sequence tag; RPE: retinal pigment epithelium; SSH: suppression subtractive hybridization; PC: primary, conventional b) Hs: Homo sapiens; Mm: Mus musculus; Cf: Canis familiaris; Bt: Bos taurus c) Genes with LocusLink ID identified based on publically available information d) [67], For query terms see (see additional File 15 e) [68], March 2003 f) [69], March 2003 Table 2 Frequency of unique genes in studies No. of studies No. of unique genes 1 2,608 a) 2 4,162 3 2,368 4 2,002 5 1,736 6 1,282 7 768 8 385 9 178 10 84 11 37 12 23 13 6 14 2 16 4 Total 15,645 a) not included in assembly of 13K retinome (see additional File 1) To assess the degree of completeness of the adult mammalian retinome, we compared the LocusLink IDs of the 13,037 transcripts to partial lists of genes known i) to be specifically expressed in the retina/RPE (category I, n = 43) (see additional File 5 ), ii) to play a role in the phototransduction pathway/vitamin A cycle (category II, n = 57) (see additional File 6 ), iii) to encode retinal/RPE proteins verified by immunohistochemistry (category III, n = 260) (see additional File 7 ), and iv) to be associated with syndromic and non-syndromic retinal disease (category IV, n = 102) (see additional File 8 ). The data show that the compiled retinome covers all retina/RPE-specific transcripts (43/43) and 53/57 (93%) of the phototransduction pathway/vitamin A cycle genes. Known retinal/RPE proteins are represented by 204/260 (79%) transcripts while 87/102 (85%) genes known to be involved in retinal diseases are found in the 13K retinome collection (Table 3 ). To further evaluate the significance of these findings, partial transcriptomes of heart (n = 3,660; see additional File 9 ), liver (n = 5,780; see additional File 10 ) and prostate (n = 7,018; see additional File 11 ) were assembled and compared to the four selected categories. In category 1, none of the retina/RPE-specific genes were present in the heart, liver, or prostate partial transcriptomes while category 2 revealed four of 16 expected (25%, heart), 7 of 25 (28%, liver) and 9 of 31 (29%, prostate) genes. In category 3, 59/73 (81%, heart), 82/115 (71%, liver) and 92/140 (66%, prostate) genes were present. Retinal disease genes (category 4) were found at a rate of 18 of 29 (62%, heart), 26 of 45 (58%, liver) and 28 of 55 (51%, prostate) (Table 3 ). The expected values for the partial transcriptomes were calculated by adjusting the respective transcriptome sizes relative to the total number of transcripts of the retinome. Table 3 Representation of retinome and partial assemblies of heart, liver, and prostate transcriptomes in defined retina/RPE gene groups Transcriptome No. of non-redundant genes identified in ≥ 2 studies a) Retina / RPE-specific genes b) (n = 43) Vitamin A / phototrans-duction pathway c) (n = 57) Retina / RPE genes (verified by immunohisto-chemistry d) ) (n = 260) Retinal disease genes e) (n = 102) Source / Reference Retinome 13,037 43 53 204 87 Present publication Heart (partial) 3,660 0 4 59 18 UniGene Build 166, SAGE library GSM1499 Liver (partial) 5,780 0 7 82 26 UniGene Build 166, SAGE library GSM785 Prostate (partial) 7,018 0 9 92 28 UniGene Build 166, SAGE libraries GSM685, GSM739, GSM764 a) See additional Files 2, 9, 10 and 11 b) See additional File 5 c) See additional File 6 d) See additional File 7 e) See additional File 8 A comparison of the 13K retinome with partial transcriptomes of heart, liver, and prostate suggests a high degree of overlapping expression between retina/RPE and heart (3,496/3,660), liver (5,343/5,780) and prostate (6,471/7,018). A total of 2,330 genes are expressed in all tissues and represent putative "housekeeping" genes (see additional File 12 ). It should be noted that the low number of ubiquitously expressed genes is largely due to the fragmentary nature of the heart, liver, and prostate transcriptomes. With increasing transcriptome complexities this number is likely to increase. Analysis of the least complete transcriptome, the heart, reveals that 2,330/3,660 (64%) transcripts can be classified as ubiquitously expressed (see additional Files 9 and 12 ) while a maximum of 1,330/3,660 (36%) genes may display tissue-restricted or tissue-specific expression. A comparison of more complete transcriptomes may significantly reduce the latter estimate. So far 5,051 genes are only found in the retinome representing a collection of "retinome-enriched" transcripts, while 7,986 are also present in at least one of the partial transcriptomes of the heart, liver or prostate. Thirty-two genes were found to be expressed in heart, liver and prostate but not in the retinome (see additional File 13 ). The distribution of the mammalian retinome across the human genome was assessed by a paired comparison of the number of reference retinome genes (13,037) versus the number of human non-redundant syntenic gene predictions (SGPs) (43,109) [ 40 ] along 618 five-Mb windows (Fig. 1a ). Correction for the total number of SGPs relative to the number of retinome genes positioned on average 21.1 (median 19.9) SGPs per window compared to 21.3 (median 15.0) retinal genes per window. Based on the Wilcoxon two-sample paired signed rank test, the null hypothesis assuming similar distribution of the SGPs and the retinome genes across the five-Mb windows can be rejected at p < 0.01. While their overall distribution greatly parallels that of the SGPs, retinome genes tend to cluster in several chromosomal regions most prominently on chromosomes 6p22.1-p21.31, 11q12.2-q13.1, 16p13.3, 19p13.3 and 19q13.32-q13.33 (Fig. 1a,1b ). Figure 1 Chromosomal distribution of transcripts defining the reference retinome. ( A ) The distribution of 13,037 retinome genes over the human genome is shown as the difference between the number of observed and expected transcripts in window sizes of 5 Mb along the chromosomes (abscissa). The number of expected genes was based on 43,109 SGP-predicted transcripts. To correct for gene density per bin, the SGPs were adjusted by a factor of 0.30 (13,037/43,109). Positive/negative ordinate values indicate regions of enrichment/depletion of retinome-encoded transcripts. ( B ) Close-up of chromosomes 6 and 19 calculated for a window size of 1 Mb along the two chromosomes. To provide positional candidates for syndromic and non-syndromic hereditary retinopathies, the 13K reference retinome as well as the "retinome-enriched" transcripts (5,051 transcripts) were superimposed onto the disease intervals of 42 thus far uncloned retinal disorders (Table 4 ). In many instances, this results in a significant reduction of genes in the respective intervals offering a manageable number of candidates for retinal diseases (e.g. the RP29 locus contains 28 SGPs of which 5 are present in the reference retinome including GPM6A , WDR17 , FLJ22649 , VEGFC , AGA ). The number of possible candidates is further reduced in the "retinome-enriched" transcript category to GPM6A and VEGFC . To make the information on the reference retinome available, we have created the interactive RetinaCentral database, a research portal which collects and stores information on genes and proteins functionally relevant to the retinal tissues [ 1 ]. We have implemented an interactive data retrieval system that presently contains linked information on the 13,037 genes of the 13K reference retinome. Database scripts were programmed to synchronize the data with LocusLink index files [ 41 ] which are updated daily [ 42 ]. Table 4 Number of genes mapped to retinal disease loci Disease Chr. Flanking DNA marker Size of locus (Mb) No. of SGPs (n = 43,109) No. of retinome transcripts a) (n = 13,307) No. of "retinome-enriched" transcripts (n = 5,051) AA 11 D11S1323 D11S902 11.21 213 65 13 AIED X DXS7 DXS72 37.33 479 129 49 AXPC1 1 D1S2692 D1S414 3.00 43 7 3 BBS3 3 D3S1603 D3S1271 2.28 28 10 1 BBS5 2 D2S142 D2S326 16.81 353 50 24 BCD 4 D4S3051 D4S1652 2.06 17 2 1 CACD 17 D17S1810 CHLC.GAT7B03 3.17 118 76 36 COD2 X DXS292 DXS1113 6.1 37 5 3 COD4 X DXS10042 DXS8060 32.44 436 117 45 CORD1 18 marker position not available - - - - CORD4 17 marker position not available - - - - CORD8 1 D1S442 D1S2681 19.86 485 222 74 CORD9 8 D8S1820 D8S532 12.79 179 43 16 CRV 3 D3S3564 D3S1578 11.2 271 157 72 CYMD 7 D7S435 D7S526 1.56 25 13 7 EVR3 11 GATA34E08 D11S4102 14.00 192 35 13 LCA3 14 D14S606 D14S610 5.95 178 36 15 LCA5 6 D6S1551 D6S1694 36.08 541 263 93 LCA9 1 D1S1612 D1S228 5.52 130 51 22 MCDR1 6 D6S249 D6S1671 2.44 45 2 0 MCDR2 4 D4S3023 D4S3022 21.00 253 57 20 MCDR3 5 D5S1981 D5S2031 19.94 184 36 14 MCDR4 14 D14S261 - 10.00 b) 58 11 4 MRST 15 D15S211 D15S152 4.70 106 24 11 OPA2 X DXS993 DXS991 14.31 245 69 24 OPA4 18 D18S34 D18S548 2.55 23 3 1 PRD X MAOB DXS426 3.92 67 18 6 RCD1 6 marker position not available - - - - RNANC 10 D10S1225 D10S1418 5.76 58 9 2 ROA1 8 D8S1702 D8S1794 11.87 150 30 13 RP6 X DXS28 DXS164 4.92 60 4 2 RP17 17 D17S1607 D17S948 7.18 130 44 17 RP22 16 D16S287 D16S420 6.33 92 40 14 RP23 X DXS1223 DXS7161 10.51 140 30 15 RP24 X DXS8094 DXS8043 7.75 74 6 3 RP28 2 D1S1337 D2S286 45.78 674 194 68 RP29 4 D4S3035 D4S2417 3.66 28 5 2 STGD4 4 D4S1582 D4S2397 16.56 187 33 10 USH1A 14 D14S99E D14S292 6.98 134 44 18 USH1E 21 D21S1905 D21S1913 11.48 77 19 10 USH2B 3 D3S1578 D3S3658 12.80 287 161 74 USH2C 5 D5S428 D5S433 18.54 269 34 13 VRNI 11 INT2 D11S873 22.41 311 89 37 WFS2 4 D4S2366 D4S3023 2.18 26 10 5 WGN1 5 D5S626 D5S2103 7.11 84 9 4 a) See additional File 14; b) Linkage to a single DNA marker reported. An arbitrary locus size of 10.0 Mb was assigned. Discussion Compiling the transcriptome of a cell or tissue is arguably more demanding than establishing the number of gene loci encoded by a given genome sequence [ 43 ]. This may mainly be explained by the dynamic nature of mRNA itself which frequently produces alternative transcripts from a single gene locus by usage of tissue-specific promoters, cryptic splice sites or variable polyadenylation signals [ 44 , 45 ]. In addition, variation in gene expression is known to occur within and between populations [ 46 , 47 ] and allele-specific expression, even from non-imprinted genes, appears to be common [ 48 ]. Further complicating transcriptome definition are effects of gender and age on RNA expression [ 49 ] as well as agonal and postmortem factors which greatly affect RNA integrity and thus frequently influence subsequent analyses [ 50 ]. Finally, differences in experimental technologies and data post-processing add an additional level of variability. Taken together, the complexities in mRNA metabolism and experimental data handling strongly suggest that there is not a single transcriptome for a given cell or tissue but implies an arbitrary number of individual transcriptomes which need to be defined by a series of parameters such as age, gender, ethnicity, cause and time of death of the tissue donor besides many others. It is therefore advisable to initially aim for a reference transcriptome providing a blueprint of an expression profile within a broadly defined time-frame. Following this line of reasoning, we here present a framework of a first reference transcriptome of the retina/RPE consisting of 13,037 unique transcripts which broadly characterize the mature state of expression in this tissue. The present meta-analysis has integrated information from 27 studies employing diverse technologies to identify retinal/RPE transcrips. Among these, SAGE represents a sensitive tool to detect low level transcription [ 51 ] while the PCR-based SSH method is well suited to enrich for differentially expressed genes [ 36 ]. The combined use of these approaches together with conventional cDNA library sequencing and microarray-based techniques provides a more solid assessment of gene expression than would each method alone. For example, SAGE is based on sequencing of hundreds of thousands of short (10, 14, or 21 bp) tags, ideally derived from a unique location of a single transcript. Rare tags could originate from infrequently expressed transcripts but could also reflect minor genomic contamination or minor sequencing errors. For the assembly of the reference retinome we have addressed these concerns by including only those transcripts that have independently been confirmed in a second unrelated study. This has led to a conservative assembly of the 13K retinome. It should be kept in mind however that this proceeding likely excludes a number of authentic transcripts. This is illustrated by the finding that the 15K retinome which comprises 15,645 transcripts including those which were solely found in a single study (Table 2 ), contains an additional five of the 102 known retinal disease genes ( RHOK , MTATP6 , CHM , LRAT , RIMS1 ) not included in the 13K retinome. Similarly, an additional three genes ( RHOK , LRAT , GPRK7 ) involved in the vitamin A/phototransduction pathway are part of the 15K but not the 13K retinome. With additional transcription data on the retina/RPE becoming available, a second generation retinome map will need to address this issue. The estimation of transcriptome size represents one of the fundamental questions in molecular biology. Early studies using reassociation kinetics have calculated the number of distinct mRNA transcripts present in various mouse tissues to be between 11,500 and 12,500 [ 52 ]. Initial SAGE analyses have led to the conclusion that the number of different transcripts observed in normal and tumorous tissue may lie between 14,247 and 20,471 [ 53 ]. Recent data from comprehensive EST sequencing of a number of tissues including brain, breast, colon, head/neck, kidney lung, ovary, prostate, and uterus suggest expression of between 7,500 and 13,500 distinct genes for each tissue [ 54 ]. Although the size of the reference retinome is consistent with these estimates, the question of adequate transcript representation by the current compilation remains open. We have addressed this by defining a number of gene groups with known expression in retina/RPE and comparing these to the reference retinome. Genes exclusively expressed in retina/RPE are highly represented in the retinome (100%), as are mainly tissue-specific genes known to play a role in the vitamin A/phototransduction pathway (93%) (Table 3 ). A partial list of 260 genes whose encoded proteins were shown by immunohistochemistry to be expressed in the retina/RPE (but may also be present in other tissues), were represented in the reference retinome at a rate of approximately 79%. Similar numbers were obtained for the retinome coverage of retinal disease genes (85%). From these data we conclude that the 13K reference retinome is highly representative of retina/RPE-expressed genes and may describe as much as 90% of the transcript complement in the adult state. Another point of interest concerns the proportion of retinome transcripts which is uniquely expressed in this tissue. Brentani et al. [ 54 ] estimate that any two tissues may share between 73% and 84% of their transcriptomes. Comparing transcription in three tissues (breast, colon, head/neck) the authors found overlapping expression in 47% of transcripts. To investigate this in more detail, we have compiled three partial transcriptomes from heart (n = 3,660), liver (n = 5,780) and prostate (n = 7,018) by applying the same stringent criteria as defined for the retinome. Limited by the size of the partial heart transcriptome, we determined 2,330 transcripts (termed "housekeeping" genes) to be expressed in all four tissues (i.e. 64% of the heart transcriptome). Comparing the retinome to any of the partial transcriptomes revealed overlapping gene profiles between 92 % and 95 %. This would suggest that only a minor proportion of retinome transcripts is indeed unique to the retina/RPE. Thus far, we have identified a group of so called "retinome-enriched" genes comprising 5,051 transcripts which are not present in the partial transcriptomes of heart, liver and prostate. This group most likely contains additional "housekeeping" or tissue-restricted transcripts and needs further adjustment by more refined in-silico normalization to comprehensive reference transcriptomes of other tissues. Highly expressed genes including those with a ubiquitous or a tissue-specific transcription profile, have been shown to cluster in chromosomal regions of increased gene expression (termed RIDGEs) [ 55 , 56 ]. Functionally, this higher order structure has been related to transcriptional regulation [ 56 , 57 ]. To search for a possible correlation, we have determined the chromosomal distribution of the reference retinome independent of gene density. Our data show good agreement with the previously established regional expression map defining approximately 30 RIDGEs within the human genome. Overlaps are most evident for chromosomes 6, 9, 11, 17, and 19. From this we conclude that the majority of transcripts assembled in the reference retinome share characteristics of the RIDGEs including moderate to high level expression. This finding may be ascribed to the stringent selection criteria we have applied to assemble the reference retinome by excluding all transcripts (n = 2,608) that were reported in only a single study. Conversely, the RIDGE-like pattern of the reference retinome could be an indication that missing transcripts may have features compatible with chromosomal domains defined as anti-RIDGEs [ 56 ]. As opposed to RIDGEs, clustering of genes in anti-RIDGEs seems associated with significant decreased expression [ 56 ]. In contrast to their fractional occurrence in transcriptomes, the identification of such low abundant transcripts are likely to require significant resources in order to compile more complete transcriptomes. To provide positional candidates for retinal disease genes, we have mapped the transcripts representing the reference retinome to the minimal regions defined for 42 retinal disease loci with as yet undefined gene mutations. To further limit the number of candidate genes, in particular for loosely defined disease loci such as RP28 or VRNI , we have similarly integrated the "retinome-enriched" transcripts. This also accommodates for the fact that approximately 50% of retinal disease genes are retina/RPE-specific [ 58 ]. For 41 of 42 unknown disease genes we have now identified strong candidates although for some disease loci including AIED , COD4 , CORD8 , CRV , LCA5 , RP28 , and USH2B , the number of candidates may still exceed capacities of most laboratories for direct analysis. For other disease loci (e.g. BCD , BBS3 , COD2 , CYMD , MCDR4 , OPA4 , PRD , RNANC , RP24 , RP29 , RP6 , WFS2 and WGN1 ), a restricted number of candidates are now available (see additional File 14 ). Conclusions We here present a first near-complete transcriptome of a defined tissue, the retinome, which may serve as a reference for further efforts to establish spatial, i.e. cell-specific, and developmental transcriptomes of the retina/RPE. A fundamental aspect of the current study was to integrate the available information on gene identification generated by a wide range of techniques. This ensures robustness and reliability of transcript data providing a stringent framework for further expression studies in systems biology. A similar approach for other tissues/cells would be advisable as this may greatly facilitate in-silico identification of tissue-specific genes to elucidate functional pathways vital for a defined cell population. In addition, the reference retinome may prove valuable for providing strong candidates for hereditary as well as genetically complex diseases and thus may help to further our understanding of retinal biology in health and disease. Methods Data retrieval and analysis To assemble a list of genes expressed in the adult mammalian retina and RPE we reviewed 27 studies reporting raw or processed transcript data derived from several mammalian species including H. sapiens , B. taurus , C. familiaris and M. musculus (Table 1 ). The data were generated by cDNA library sequencing, microarray studies, and SAGE. Publically available data analysing transcripts from adult mammalian retina/RPE tissues published until December 2003 were included. Excluded were studies investigating transcription in retina/RPE by using RNA sources such as fetal tissues, cell lines or non-mammalian species. Gene identifiers such as GenBank accession number, gene nomenclature symbol, gene description, UniGene cluster ID, cDNA sequences or tags were available from sources as detailed in additional File 15 and were used to retrieve the unique human LocusLink ID for each gene (as of December, 2003). Only genes with established LocusLink ID were included in the present study. For SAGE data, tag-to-gene assignment was done by querying the SAGEmap_tag_ug-rel dataset [ 59 , 60 ]. Tags assigned to multiple genes were excluded from further analysis. Human orthologous genes were established via the NCBI-curated homology database[ 61 ] or by BLAST sequence comparison [ 62 ]. To assemble partial transcriptomes of heart, liver and prostate, for each tissue data were mined from at least one SAGE library, in addition to expressed sequence tag (EST) sources (see additional File 16 ). Similar to the criteria for the assembly of the retinome, genes identified in only one study were disregarded. EST retrieval was facilitated by use of the Gene Library Summarizer [ 63 ] which retrieves the known genes represented by at least one EST and generated from a tissue sample with normal histology. Partial lists of genes known to play a role in the retina and/or the RPE were assembled from the literature (see additional Files 5 , 6 , 7 and 8 ). Additional File 5 summarizes genes known to be exclusively expressed in retina and/or RPE, while additional File 6 includes genes involved in the phototransduction cascade and the vitamin A cycle. Additional File 7 is a partial compilation of genes/proteins verified by immunohistochemistry to be present in adult mammalian retina and/or RPE. A list of 102 genes involved in retinal diseases was retrieved from the RetNet database, January 2004 [ 58 ] (see additional File 8 ). Assignment of genes and disease loci to the human genome A total of 43,109 human non-redundant syntenic gene predictions (SGP) were retrieved (as of December 2003) and chromosomally mapped to the reference sequence of the human genome (July 2003) utilizing the USCS Genome Table Browser [ 64 ]. Based on the position of their putative transcription start sites, the SGPs were assigned to 5 Mb bins along the human chromosomes. In addition, one-megabase bins were defined for refined analysis of chromosome 6 and 19 (Fig. 1b ). Similarly, the chromosomal map positions of the retinome transcripts were determined by querying the USCS Genome Table Browser with the respective LocusLink, UniGene or RefSeq IDs. Mapped loci of retinal dystrophies with unknown genetic basis (n = 45) were taken from RetNet, January 2004 [ 58 ] and placed on the human genome sequence by querying the USCS Genome Table Browser with DNA marker sequences shown to flank the minimal candidate region. Three disease loci ( CORD1 , CORD4 and RCD1 ) are insufficiently mapped on the respective human chromosomes and were therefore not included in the analysis. Statistical analysis of gene distribution To determine if either of the two datasets, the 43,109 human non-redundant SGPs and the 13K retinome transcripts, is distributed in a non-parametric and distribution free manner over the genome, the Kolmogorov-Smirnov Goodness-of-Fit Test was used [ 65 ]. Statistical significance of the median difference in paired chromosomal distribution of retinome transcripts versus the SGPs was then evaluated by the non-parametric Wilcoxon two-sample paired signed rank test [ 66 ]. To carry out the test we calculated the difference between all genes versus retinal genes per 5-Mb bin. To correct for the total number of genes within the two groups, the SGPs per bin were adjusted by a factor of 13,037/43,109 = 0.30. Mean and median values per bin were 21.05 and 10.93 for all genes and 21.26 and 19.93 for retinal genes, respectively. Authors' contributions HLS participated in the design of the study, collected and processed the information from the published reports related to the retina and RPE genes. TG conceptually developed the RetinaCentral database and is involved in the curation of the site. JK carried out the statistical analyses and helped with the computational handling of data. BHFW was involved in all aspects of data assembly and prepared the manuscript. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Retina/RPE genes reported only in a single study List of genes expressed in the retina or RPE reported only in one study Click here for file Additional File 2 Gene list of the reference retina / RPE transcriptome List of 13,037 genes expressed in the retina/RPE Click here for file Additional File 3 List of genes identified exclusively in retina studies Click here for file Additional File 4 List of genes identified exclusively in RPE studies Click here for file Additional File 5 Partial list of genes known to be expressed specifically in the retina and/or RPE Click here for file Additional File 6 Partial list of genes known to be involved in the vitamin A cycle and phototranduction pathway Click here for file Additional File 7 Partial list of genes known to encode retina / RPE proteins Click here for file Additional File 8 List of known genes involved in retinal diseases Click here for file Additional File 9 Partial gene list of the heart transcriptome Click here for file Additional File 10 Partial gene list of the liver transcriptome Click here for file Additional File 11 Partial gene list of the prostate transcriptome Click here for file Additional File 12 List of genes present in the reference retinome and partial transcriptomes of heart, liver, prostate Click here for file Additional File 13 List of genes present in partial transcriptomes of heart, liver, and prostate but not in reference retinome Click here for file Additional File 14 List of retinome transcripts mapping to retinal disease intervals. Click here for file Additional File 15 Data sources used to compile the retina / RPE transcriptome Click here for file Additional File 16 Data sources used to compile partial transcriptomes of heart, liver, and prostate Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC512282.xml
518963
Development of real-time PCR for detection of Mycoplasma hominis
Background Mycoplasma hominis is associated with pelvic inflammatory disease, bacterial vaginosis, post partum fever, sepsis and infections of the central nervous system often leading to serious conditions. Association with development of female infertility has also been suggested, but different publications present different results. We developed a sensitive and fast diagnostic real-time PCR to test clinical samples from women undergoing laparoscopic examination before fertility treatment. To develop a test for the detection and quantification of M. hominis we selected a housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase ( gap ), as a target. Results Real-time PCR was optimized to detect 10 copies of M. hominis PG21 genomic DNA. A fluorescence signal was measured for all 20 other M. hominis isolates, and melting curves analysis showed variations in the melting temperature in agreement with sequence variation in the region of the probes. There was no amplification of other mycoplasmal DNA and human DNA. Eighty-three patient cervical swab samples from infertile women were cultured for M. hominis in the BEa medium. Two of the samples (2.4%) were positive after 48 hours of incubation. The real-time PCR detected the same two samples positive, and the DNA concentrations in the clinical specimens were calculated to 37.000 copies/ml and 88.500 copies/ml, respectively. Conclusion The results demonstrate that real-time PCR may prove to be a rapid alternative to the traditional cultivation method. Information on bacterial load in genital swabs can be obtained. The assay allowed detection of M. hominis in a closed system reducing the risk of contamination by amplicon carry-over.
Background Mycoplasmas are the smallest living prokaryotes known, capable of self-replication. They belong to the class Mollicutes and are distinguished phenotypically from other bacteria by their minute size and lack of a cell wall. Genetically they differ by having a small genome size and low G+C content {1} [ 1 ]. Mycoplasmas have adapted to a wide variety of hosts and can colonize man, other animals and plants. The colonising organisms are host specific. In humans, mycoplasmas colonize mainly the upper respiratory tract and the genitourinary tract. The first human Mycoplasma isolated was Mycoplasma hominis [ 2 ]. It is a heterogeneous genital mycoplasma [ 3 ] found in at least two-thirds of women with bacterial vaginosis (BV), compared to 10% of healthy women [ 4 , 5 ]. M. hominis has also been isolated from the endometrium and fallopian tubes of 10% women with salpingitis. However, its role as a primary pathogen is doubtful since it co-exists with many other bacteria in BV [ 6 ]. Studies made on women undergoing in vitro fertilization showed the presence of M. hominis only in 2.1% of the women [ 7 ]. Isolation from other sites than the genitourinary tract has been reported. M. hominis has been found to cause wound, joint and central nervous system infections [ 8 ], it has been isolated from cavernous angioma, but was not the cause of the disease [ 9 ] and cases of brain and scalp abscesses and meningoencephalitis were also reported [ 10 - 15 ]. Those cases demonstrated the pathogenic potentials of M. hominis and indicated a need for rapid recognition. So far culture is most commonly used for detection of genital Mycoplasma, but it requires special handling, complex media, and cultivation positive samples need further testing to determinate the species cultivated [ 16 ]. A case report of brain abscess in a 22 year-old female patient with postpartum infection [ 10 ] showed that culture took 4 – 5 days during which the patient's symptoms continued to worsen before the antibiotic treatment was changed. Comparison between culture method and PCR has been performed and showed that a PCR assay was as sensitive as culture for detection of M. hominis from clinical samples. In addition it was very specific [ 17 ]. An advantage of using PCR is that the system can detect the presence of both live and dead microorganisms in the sample. When comparing the original PCR protocols with the newly developed real-time protocol, it offers interesting advantages such as rapidity, closed system, which eliminates the risk of carry-overs, real-time monitoring of PCR activity, quantification of amplification product and, if required, mutation analysis. A recent study designed for the detection of M. hominis by real-time PCR in HIV-positive patient swab samples, suggested the use of SYBR Green with primers targeting another housekeeping gene, the 16S rRNA gene [ 18 ]. The 16S rRNA gene is the most conserved microbial gene, though, in M. hominis minor sequence variation was observed [ 19 ]. Because sequences from M. hominis isolates were available (see table 1 for accession nos.) and possible to compare with other mycoplasmas sequence [ 28 - 30 ], we selected the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase ( gap ) as target for development of a quantitative real-time PCR. Comparison of the DNA sequences from different M. hominis isolates showed, however, some small variations in the amplified DNA sequence [ 31 ]. We determined sensitivity and specificity of the M. hominis LightCycler real-time PCR and tested it on clinical swab samples using specific hybridization probes for detection. Results Design of primers and probes The principles of the system we used are based on two specific hybridization probes located internally to the amplification primers, each of them labelled with a different fluorescent dye [ 32 ]. The DNA sequence of primers and probes is shown (table 2 ). We selected M. hominis PG21 type strain as a template for primers and probes. The primers and probes were designed with respect to conservation of the DNA sequence within the M. hominis gap gene [ 31 ] and difference from the related genital Mycoplasma genitalium gap gene (Accession no. U39710) (fig. 1 ). The amplified DNA fragment was of 144 bp in size. Swissprot Protein Database was used to determine the amino acids sequence of GAPDH enzyme (E.C. no. 1.2.1.12). By use of the MOTIFS program the active site was predicted to consist of the amino acids: ASCTTNCL, located at nucleotides 451 to 474 (fig. 1 ) [ 31 ]. When comparing DNA sequences of the active site between M. hominis and M. genitalium, only 3 out of 24 nucleotides are mismatching (fig. 1 ), and therefore the probes and primers were placed before this region. As seen from fig. 1 , the G + C content of the M. hominis gap gene is very low. It was difficult to find a suitable location for the reverse primer, and therefore two reverse primers were designed (table 2 ). The sequence of the reverse primer II was partly overlapping the encoding region of the active site of GAPDH enzyme. When compared it was found that use of reverse primer I gave the highest sensitivity of the LightCycler PCR assay. Concentration of 5 copies/μl of PG21 DNA was not always detected in PCR runs when reverse primer II was used, whereas with use of reverse primer I, 5 copies/μl of PG21 were present in every run, and therefore this primer was chosen for the following experiments. In order to obtain optimal detection in the annealing phase of the PCR we designed the probes to anneal to the same strand as the forward primer and placed them as far as possible from that primer (fig. 1 ). One hybridization probe was labelled with fluorescein (FL) in the 3' end, the other with LightCycler Red705 (LCred705) in the 5' end. When the probes are hybridized less than 5 nucleotides apart, Fluorescence Resonance Energy Transfer (FRET) will be induced. The distance between the two probes was one nucleotide after annealing to template DNA allowing FRET light to be measured. Sensitivity and specificity Dilution series of the standard DNA from M. hominis PG21 were used to examine the sensitivity of the LightCycler PCR assay. The fluorescence curves are shown (fig. 2a ). Detection of PCR product was possible for the lowest DNA concentration: 1 copy/μl, equal to 2 copies in the reaction mixture (fig. 2a ). To reduce the noise, the cut-off value was set to 0.1. The standard curve had an average slope equal to -3.5, which means that the efficiency of the PCR reaction was 1.93 (oscillating between 1.9 and 2.0) (fig. 2b ). Calculations of the efficiency derived from the function for the amount of PCR product that was formed, represented by equation: N = N 0 × 2 n , where N is the number of amplified product, N 0 is the initial number of molecules and n is the number of the PCR cycles. Ideally the efficiency equals 2. Additionally, the sensitivity of the assay was determined for two other M. hominis isolates by making dilution series of DNA from M. hominis 132 and 4712. As for PG21, detection of a PCR product was possible for the lowest DNA concentration of 1 copy/μl indicating similar sensitivity of the two other isolates. The standard curves had slopes equal to -3.552 for isolate 132 and -3.576 for isolate 4712, the efficiency of the PCR reaction was 1.91 and 1.9, respectively. To determine the reproducibility of the assay, ten dilution series of PG21 DNA were analyzed in different PCR runs, and the values of crossing-points (also known as threshold cycles – C t ) were compared by calculation of the coefficients of variation ( CV ). The values of CV were between 3.8% and 6.7% (table 3 ). In all runs 10 out of 10 samples were positive except for 1 copy/μl where a PCR product was seen in 6 out of 10 PCR runs. The reproducibility of the assay for the two other M. hominis isolates 132 and 4712 was analyzed in three different PCR runs. In both isolates dilution of 5 copy/μl was present in three out of three runs, similarly to PG21 the dilution of 1 copy/μl was present in two out of three LightCycler runs. Our detection limit was therefore 5 copy/μl, equals to 10 copies in the reaction mixture. The reproducibility of the assay was acceptable. As a next step twenty M. hominis isolates (table 1 ) were tested with the designed primers and probes in the LightCycler real-time PCR. All isolates gave a positive fluorescence signal and the concentration of DNA was similar to 10 4 copies/μl of PG21, measured by the LightCycler instrument (fig. 3 ). The specificity of the LightCycler assay targeting the gap gene was evaluated by testing human DNA and DNA from different Mycoplasma species. With the specific probes there was no cross-reactivity to other Mycoplasma species or human DNA (fig. 4a ). For human DNA we used a concentration of 10 4 copies/μl calculated by OD measurements of the purified DNA and the genome size. Since DNA from different mycoplasmas was extracted by proteinase K treatment of PBS washed pellets, we tested such extracted DNA for presence of inhibitors. Two μl of DNA from five different mycoplasmas ( M. arginini, M. bovis, M. hyorhinis, M. pulmonis, M. salivarium ) were spiked with 2 μl of PG21 DNA of 100 copies/μl and added to the reaction mix. As illustrated (fig. 4b ) there was no inhibition in the proteinase K treated samples. Melting curve analysis Melting curve analysis can be used to determine the presence of non-specific amplification products. The melting temperature ( Tm ) is defined as the temperature at which half of a duplex-DNA becomes single-stranded [ 33 ]. As it was impossible to place probes in completely conserved regions (fig. 5 ), we analyzed the melting curves of the real-time PCR products of the different M. hominis isolates. Tm of PG21 DNA was 66°C and equal for high and low concentrations of DNA. It was shown that the melting temperature of the PCR products of M. hominis DNA clustered in 3 major groups (fig. 6 ). The isolates V2785 and P71 had the same temperature of 66°C as PG21, the second group 93, 7357, 132, P2, P7, SC4, DC63, 7808, 183, 1893, 10, W2 had a melting temperature of 64°C, while the last group 5941, 4712, 3105, M1449, 6188, had the melting temperature of 62°C. These different melting points were in agreement with variation in the DNA sequence of the probe regions (fig. 5 ). Coloured media and possible PCR inhibition Many Mycoplasma species were cultured in SP-4 or BEa media for specificity of LightCycler analysis. Additionally, the clinical samples were transported in SP-4 medium, which can be used for the recovery of M. hominis [ 34 ], and BEa medium was used for cultivation of M. hominis . Therefore, to see the effect of the coloured media on the LightCycler assay, we constructed artificial samples consisting of SP-4 or BEa medium spiked with DNA of known concentration (10 5 , 10 4 , 10 3 ). Two μl of the samples were analyzed by the LightCycler PCR. As shown (fig. 7 ), BEa medium completely inhibited the reaction, whereas SP-4 inhibited only partially, but markedly reduced the PCR efficiency. Analysis of clinical samples Culture Eighty-three endocervical samples from women attending fertility clinics in Denmark were cultured for the presence of M. hominis . Two samples were found positive. Three passages were examined for colour change, at each passage the colour of the medium turned pink, and samples were thus considered as positive cultures. Proteinase K treated culture materials were analyzed by Mycoplasma -genus-specific PCR [ 35 ], which gives PCR products from 16S rRNA gene of 265 bp in size. The PCR products from the two positive clinical samples were sequenced and the resulting DNA sequence confirmed them to be M. hominis . Quantification by culture of these two positive samples was performed by limited dilution in growth medium. There was a colour change in 9 wells in both samples, which corresponds to 25.600 CCU/ml in the swab sample, calculated from titration. LightCycler PCR on DNeasy treated samples The cervical swab samples were DNeasy treated and tested in duplicates by LightCycler PCR using the M. hominis gap -assay. Two samples (nos. 56 and 83) were positive when examined by LightCycler PCR (2.4%). The copy numbers were measured to be 220 (for 83) and 530 (for 56) copies/μl respectively. The amount of DNA copies per ml in the original sample was calculated to be 37.000 for sample 83 and 88.500 for sample 56. The two quantification methods for estimating bacterial load in positive samples thus showed that the number of live Mycoplasma cells was 69% (patient 83) and 29% (patient 56) of what was found by PCR. The positive clinical samples showed melting temperature of 64°C corresponding to the second group, in which the majority of the M. hominis isolates were found. Reproducibility of the quantification of DNA in the clinical specimens was analyzed in 10 negative patient samples. Four μl of samples were spiked with 2 μl of PG21 DNA of known concentration (100 copies/μl). The six μl were added to the LightCycler PCR reaction. No inhibition was observed (fig. 8 ). In some DNeasy treated clinical samples a slight fluorescence response was seen at the very late PCR cycle. To analyze whether this slight fluorescence response was unspecific, different DNeasy samples were analyzed. We used different concentrations of human DNA from " Mycoplasma free" Hep2 cells, which were purified with Blood & Cell Culture DNA Mini Kit, and these samples did not give a positive fluorescence signal when run in LightCycler PCR (fig. 9a ). In addition the results showed that there was no cross-reaction to human DNA. However, when DNA free water and standard dilutions of the human DNA extracted by the Blood & Cell Culture DNA Mini Kit were treated with DNeasy Tissue Kit, we experienced a slight fluorescence signal at the very late cycle numbers. Even DNA free water gave a positive signal (fig. 9a ). The calculated copy numbers were between 1–8 copies/μl. The melting curve analysis showed an atypical flat and broad melting curve with the melting peak below the range of Mycoplasma hominis isolates (fig. 9b ). Clinical samples showing the low concentration (between 1–8 copies/μl) also had the flat melting curve with the melting peak below 61°C (fig. 9b ) and were therefore considered as negative. One additional sample (no. 9) had an average concentration of 10 copies/μl, but when comparing the melting curve data, it had a melting curve identical to the DNeasy treated water and was therefore considered negative. There was thus 100% agreement between cultivation and detection by real-time PCR. LightCycler PCR on proteinase K treated clinical samples To additionally confirm that samples which gave a slightly positive fluorescence signal in LightCycler PCR after DNeasy treatment were not containing M. hominis DNA, we analysed proteinase K treated samples from those patients by the LightCycler PCR. None of the samples gave a positive fluorescence signal. After confirming that all samples were negative, we spiked 2 μl of each sample with 2 μl of PG21 DNA of 100 copies/μl, as the control for inhibition. None of the proteinase K treated samples showed inhibition. This clearly indicates that those samples were negative. Discussion A rapid quantitative real-time PCR for detection of M. hominis from cervical swab samples was developed. To our knowledge it is the first LightCycler PCR protocol where quantification of M. hominis is combined with melting curve assay. The LightCycler PCR reproducibility was able to detect down to ten copies/reaction of the genomic PG21 DNA. All other 20 M. hominis isolates were positive in the assay. Additionally, isolates 132 and 4712, which based on melting temperature belong to the two other groups than PG21, were used to document detection limits of M. hominis isolates. Similarly to PG21, the detection limits were 5 copies/μl, equals to 10 copies per reaction for both isolates. The average efficiency was high, but small differences were seen. This can be caused by different factors such as presence of inhibitors in the sample and treatment of the sample. The artificially constructed samples consisting of PG21 DNA and coloured BEa and SP-4 media showed that these media inhibited the PCR reaction indicating that it was crucial to wash the pellets of other Mycoplasma species with PBS. Additionally, a spiking assay where DNA from other mycoplasmas was used together with DNA from PG21 showed that there was no inhibition in the proteinase K treated samples. Since swab samples may contain PCR inhibitors we introduced the DNeasy procedure to additionally purify DNA from proteinase K treated clinical samples. Analysis of the ten negative clinical samples spiked with DNA from PG21 (100 copies/μl) showed no inhibition after DNeasy treatment. We experienced, however, some slight fluorescence response in the very late cycle number, which corresponds to the low concentration seen in DNeasy treated water or Hep2 cells (fig. 9a ) and flat melting curves with lower melting temperature than those of M. hominis isolates (fig. 9b ). DNeasy treated patient samples that gave such a weak fluorescence signal were considered as negative. Additionally, using the original proteinase K treated patient samples, for those patients that gave a weak fluorescence signal after DNeasy treatment, we did not see any reaction with LightCycler PCR. This strongly indicates that such slight fluorescence response is generated only in the DNeasy treated samples. A previous study showed that women with bacterial vaginosis showed presence of M. hominis in 38.5% compared to 8.3% of women with normal microbial flora [ 36 ]. Since none of our patients had bacterial vaginosis we had expected approximately 8% of our samples to be positive. We found two positive samples that were confirmed to be positive by the culture and had a high concentration of DNA by LightCycler PCR (220 and 530 copies/μl). The low number of positive patients (2.4%) was surprising but comparable with a previous study made on patients undergoing the in-vitro fertilization [ 7 ]. Comparison of the number of DNA copies and CCU showed that the CCU was lower than the number of DNA copies. This can be explained by the presence of dead bacteria in the samples. Even though the assay was designed for quantification, the variability in the melting peaks gave additional information of M. hominis . Differences in the melting temperatures between M. hominis isolates prove heterogeneity of the housekeeping gene sequence. The Tm value determines how well the sequence of probes matches the sequence of template DNA, and it will decrease if mismatched DNA is amplified. Single mismatch can decrease the Tm from 1°C up to 30°C [ 37 , 38 ] depending on many factors, such as pH, duplex length and G + C content. This kind of analysis is used in detection of subtypes of Herpes simplex virus, since the Tm discriminates between two different subtypes [ 33 , 39 , 40 ]. In the present study, Tm of the clinical samples can suggest to which group of isolates they belong and how different they are from the PG21 template DNA. In this study, patient samples nos. 56 and 83 had melting peaks similar to isolates from the second and largest group. The real-time technology where measurement of the fluorescence emitted during amplicon production is performed during each PCR cycle is considered as a breakthrough in PCR. Conventional PCR is an open, contamination-susceptible system where it is necessary to transfer the amplified product to other detection systems to confirm a positive result. Real-time PCR benefits by a closed system in which formation of a product is measured immediately without transfer [ 32 ]. Interpretation of LightCycler PCR results, presented as graphs and calculation of crossing points, introduces many advantages, but such a parameter of real-time PCR should be evaluated. Conclusions LightCycler PCR appears very promising for detection of organisms that are difficult to culture or whose growth is slow. We have developed a quantitative, specific LightCycler protocol for detection of M. hominis, which offers rapid diagnosis of one hour after DNA extraction. The DNA extraction method used was not the best choice as unspecific fluorescence did occur in the late number of cycles. Results from cultivation and LightCycler PCR were identical. The method is both sensitive and specific. All tested isolates gave a positive fluorescence response, and final amplification and quantification was performed in closed tubes, which reduces the risk of contamination. The described target gap gene sequence should be preferred to more varying parts. A small variation in this part of the gap gene among different M. hominis isolates was observed by the melting curve analysis. Methods Microorganisms and human DNA used for the study Organisms used in this study are listed in table 1 . Subjects A total of 83 consecutive women attending fertility clinics in Denmark (Brædstrup/Horsens and Holstebro) were studied. All patients were undergoing hysteroscopy and transvaginal hydrolaparoscopy (culdoscopy) [ 41 ] or laparoscopy due to infertility. The endocervical specimens for detection of presence of M. hominis were collected before the scopic examination. Material collection Endocervical specimens were obtained using a sterile chlamydial swab and the contents transferred immediately into a tube containing 2 ml of transport medium, SP-4 [ 42 ], containing thallium acetate (0.01%), which inhibits growth of other microorganisms. Such prepared samples were then sent to the laboratory of Department of Medical Microbiology and Immunology, Aarhus University, where they were stored at -70°C. Cultivation and harvesting of microorganisms and human Hep2 cells used in the study All isolates of M. hominis, M. buccale, M. salivarium, M. orale, M. arthritidis, M. arginini, M. lipophilum, M. primatum were cultivated in 1.7 ml of broth BEa medium [ 22 ]. U. urealyticum and U. parvum were grown in 1.7 ml of SU medium [ 43 ]. M. pneumoniae and M. genitalium were grown in 100 ml of SP-4 medium [ 42 ] as described in detail elsewhere [ 44 ]. M. fermentans, M. pulmonis, M. hyorhinis were cultivated in 1.7 ml of BEg medium [ 43 ]. Finally, M. bovis was cultured in BE medium [ 43 ]. All cultures were incubated at 37°C. The BEa medium changed colour from orange to pink in 48 hours due to reduction of phenol red by arginine hydrolysis. SP-4 and BEg both changed colour from orange to yellow, whereas SU changed from yellow to orange. BE medium changed from yellow to light pink in 72 hours. Except for M. hominis, M. genitalium and M. pneumoniae , which were cultivated up to bigger volumes (100 ml), 500 μl of the colour changed cultures were then placed in 7 ml of the new medium, and harvested after the medium changed for the second time. One ml of logarithmic-phase culture was centrifuged in Eppendorf tubes at 20.000 × g for 30 min. Each pellet was washed twice by phosphate-buffered saline (PBS) and the pellets were stored at -70° prior to use. Human Hep2 cells were cultured as described elsewhere [ 45 ]. DNA extraction and purification from microorganisms and human Hep2 cells used in the study DNA from M. arginini, M. bovis, M. hyorhinis, M. pulmonis, M. primatum , M. lipophilum, M. buccale, M. salivarium, M. orale, M. arthritidis, U. urealyticum, U. parvum, M. pneumoniae and M. fermentans was extracted by suspending pellets in 160 μl TE buffer, and adding 40 μl 10 mg/ml proteinase K. The proteins were digested by incubation at 55°C for one hour, followed by inactivation of the enzyme by boiling for 10 minutes at 100°C. The DNA from human Hep2 cells as well as M. genitalium DNA was extracted and purified by Blood & Cell Culture DNA Mini Kit (QIAGEN GmbH, Hilden, Germany). Genomic DNA from all M. hominis isolates was extracted as described [ 46 ] and followed by ultracentrifugation in CsCl-ethidium bromide density gradient [ 47 ]. Concentrations of DNA from isolates PG21, 132 and 4712, and human Hep2 DNA were calculated after measuring OD (optical density) at 260 and 280 nm with Unicam 8625 UV/VIS spectrometer (ATI Unicam, Cambridge, United Kingdom). Artificial "Mycoplasma free" samples treated with DNeasy Purification Kit Fourfold dilution series of human Hep2 DNA purified by Blood & Cell Culture DNA Mini Kit (QIAGEN GmbH, Hilden, Germany) of initial concentration 8.000 copies/μl were prepared. Twenty-five μl of each dilution and DNA free double distilled water were then treated with DNeasy™ Tissue Kit (QIAGEN GmbH, Hilden, Germany) procedure. Samples were diluted twice, because of elution step with 50 μl of elute buffer, and therefore 4 μl instead of two of each sample were used in the LightCycler PCR run. Cultivation of patient cervical samples Broth BEa medium was used for culture of M. hominis . To avoid overgrowth by other bacteria, present in the urogenital tract, a special mixture of antibiotics (Niels Friis; containing: 0.15 mg/ml cycloserine, 0.2 mg/ml vancomycin, 0.2 mg/ml bacitracin and 0.2 mg/ml mecillinam) was used. Twenty μl of the swab sample was placed in 1 ml of medium with Niels Friis antibiotics and incubated at 37°C. The BEa medium changed colour from orange to strawberry pink in 48 hours due to reduction of phenol red by arginine hydrolysis. Finally, samples were described as positive when it was possible to pass them further 2 times. Mycoplasma -genus-specific PCR was performed on the two positive cultures and the PCR products were sequenced. Sequencing reactions were carried out bidirectionally using the ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer, Narwalk, USA) on the purified PCR products according to the instructions supplied by the manufacturer. Sequencing was performed on an ABI PRISM 377 DNA Sequencer (Perkin Elmer, Narwalk, USA). DNA extraction from cervical samples All 83 patient samples were treated identically. Briefly, three hundred μl of the original swab sample was subjected to microcentrifugation at 20.000 × g , and to remove transportation medium the pellets were washed twice with PBS. The pellets were suspended in 49 μl TE buffer, and 1 μl 10 mg/ml Proteinase K was added, proteins were digested while incubating at 55°C for one hour, which was followed by inactivation of the enzyme by boiling for 10 minutes at 100°C. Finally, in order to remove possible inhibitors present in swab samples, DNA extraction with DNeasy™ Tissue Kit (QIAGEN GmbH, Hilden, Germany) was performed on the resulting solution without repeating the proteinase K treatment. Twenty-five μl of the proteinase K treated sample was diluted twice because of elution with 50 μl of elute buffer, 4 μl of each sample was used for the PCR. The filter pipette tips were used for all DNA preparation steps to reduce possibility of sample contamination. Quantification of the number of viable microorganisms by culture Estimation of the number of M. hominis in patient swab samples was performed by titration in BEa growth medium. Two-fold dilutions were made in ELISA trays by adding 0.01 ml of the clinical sample to 0.19 ml of BEa medium. The plates were incubated at 37°C and reading was performed on the third day. Plates were left in incubator and observed for 3 additional days but no further change appeared. The last well with visible colour change was considered to contain one colour changing unit (CCU) allowing us to calculate the number of viable microorganisms in the original clinical sample. Primer and probe design Primers and probes were designed from the gap gene of Mycoplasma hominis type strain PG21 (Accession No. AJ243692). This gene belongs to the housekeeping genes and is therefore very conservative in all organisms. Primer and probe sequences and their locations are present in table 2 . Both primers and probes were placed in front of the conserved region of the gap gene that is almost identical in all organisms. Primers were obtained from DNA Technology, Aarhus, Denmark, and probes from TIB-MOLBIOL, Berlin, Germany. Probe and primer sequences were analysed by BIOBASE (The Danish Biotechnological Database, the University of Aarhus, Denmark) and BLAST (National Centre of Biotechnology Information, National Institutes of Health, Bethesda, MD, USA). Real-time PCR assay with hybridization probes The PCR product was 144 bp in size, which according to the manual (Roche Molecular Biochemicals Technical Note No. LC 11/2000) is preferable to perform an efficient quantification of DNA. Real-time PCR was performed in glass capillary tubes. The reaction mixture was composed of 0.5 μM of each primer, 0.2 μM of each probe, 5 mM of MgCl 2 (PCR buffer), 2 μl of ready-to-use Fast Start DNA Master Hybridization Probes (Roche Diagnostics, Mannheim, Germany) (contains a hotstart Taq DNA polymerase and reaction mixture), 1.5 μl Uracil-DNA Glycosylase (heat-labile) and 2 μl of the DNA template. Water was added up to 20 μl, which was the final volume of all reaction-mix. For the DNeasy treated patient samples 4 μl of DNA was used. When proteinase K treated patient samples were examined (without DNeasy treatment), we used 2 μl of the undiluted DNA. To avoid contamination, mixing of the reagents (except of the DNA template) was performed in a separate room, away from rooms where culturing and DNA purification were done. The DNA template from M. hominis PG21 was added by use of filter pipette tips. Uracil-DNA Glycosylase (Roche Diagnostics, Mannheim, Germany) was used to prevent the samples from possible PCR "carry-over" contaminations from previous DNA synthesis reactions. Reaction mixes contained dUTPs instead of dTTPs, and therefore it was possible to avoid contamination of samples by adding an enzyme that hydrolyzes uracil-glycosidic bonds at U-DNA in single and double-strained DNA [ 48 ]. As negative control a sample containing all reagents except DNA was used in every PCR run. Quantification of DNA concentrations performed by LC-PCR Software was based on standard dilution series with known concentrations of genomic M. hominis PG21 DNA. The concentrations of standard dilution series were: 10 5 , 10 4 , 10 3 , 10 2 , 10 1 , 5 and 1 copy/μl. Siliconized tubes were used to prevent DNA from sticking to the wall of the plastic tubes. For carrier, yeast RNA was used in concentration 10 μg/ml. The LightCycler PCR program was composed by: Hotstart Taq DNA polymerase activation done in 95°C for 10 minutes, followed by cycling: 95°C (20°C/s) for 15 s, 58°C (20°C/s) for 8 s and 72°C (20°C/s) for 8 s, repeated 45 times. Melting assay ended the analysis: samples were heated to 95°C (20°C/s) without hold, cooled to 55°C (20°C/s) hold for 15 s and then heated slowly at 0.1°C/s up to 95°C, finally cooled to 40°C (20°C/s). Fluorescence emitted at 705 nm was measured at each annealing step since the fluorescence signal is emitted when both probes are hybridized. After annealing the temperature is raised and the hybridization probes are displaced by the Taq polymerase during the elongation step. Probe fluorescence was detected in canal F3 (measures at 705 nm) and F1 (measures at 530 nm). F3/F1 was used to correct differences in volume of the samples made during pipetting. Quantification of LightCycler products The results were interpreted with LightCycler software Vers. 3.5 (Roche Diagnostics). Quantification software performs all additional steps for generation of a standard curve. First step involves Baseline Adjustment with the use of a "fit points" method, second step allows background reduction using Noise Band correction, and the last step is Analysis where the standard curve is generated from the threshold cycles ( C t ) of the standard dilution series (fig. 1b ). Samples with high DNA load had low C t values, and low DNA load had high C t values. The concentration of DNA in clinical samples was set as "unknown". Each sample was run in duplicate. Calculation of the DNA concentration in the unknown sample was based on the standard curve slope. The average of the two concentration measurements was used for further analysis. Competing interests None declared. Authors' contributions Author A.B. carried out the real-time PCR experiments, the analyses of data, and drafted the manuscript. Author H.F.S. participated in designing the LightCycler PCR method, analysis of data and coordination of the manuscript. Author J.F. participated in coordination of the study and provided clinical samples. Author S.B. participated in design and coordination of the study. Author G.C. participated in design, data analyses, coordination of the manuscript and study.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC518963.xml
546002
Gene expression patterns associated with blood-feeding in the malaria mosquito Anopheles gambiae
Background Blood feeding, or hematophagy, is a behavior exhibited by female mosquitoes required both for reproduction and for transmission of pathogens. We determined the expression patterns of 3,068 ESTs, representing ~2,000 unique gene transcripts using cDNA microarrays in adult female Anopheles gambiae at selected times during the first two days following blood ingestion, at 5 and 30 min during a 40 minute blood meal and at 0, 1, 3, 5, 12, 16, 24 and 48 hours after completion of the blood meal and compared their expression to transcript levels in mosquitoes with access only to a sugar solution. Results In blood-fed mosquitoes, 413 unique transcripts, approximately 25% of the total, were expressed at least two-fold above or below their levels in the sugar-fed mosquitoes, at one or more time points. These differentially expressed gene products were clustered using k -means clustering into Early Genes, Middle Genes, and Late Genes, containing 144, 130, and 139 unique transcripts, respectively. Several genes from each group were analyzed by quantitative real-time PCR in order to validate the microarray results. Conclusion The expression patterns and annotation of the genes in these three groups (Early, Middle, and Late genes) are discussed in the context of female mosquitoes' physiological responses to blood feeding, including blood digestion, peritrophic matrix formation, egg development, and immunity.
Background Hematophagy, blood-feeding, is a behavior exhibited by most arthropod vectors of human pathogens. In anautogenous mosquitoes, the female generally feeds to repletion on a single blood meal and then proceeds to use this nutrition as the basis for the development of a batch of eggs. The cycle of host seeking, blood feeding, egg development, and oviposition is generally called the gonotrophic cycle, a term coined by Beklemishev in 1940 [ 1 ]. For most mosquitoes living in optimal field or laboratory conditions, this cycle requires about forty-eight hours and involves a complex series of biological events, including peritrophic matrix formation, blood digestion, oocyte development, vitellogenesis, and excretion. Digestion of the proteinaceous blood meal is required for oocyte development and vitellogenesis, and consequently these are coordinated processes. Multiple hormones interact to alter tissue states and to activate genes involved in these processes. The two hormones juvenile hormone (JH) and 20-hydroxyecdysone (20-E) are most fundamental to ovarian development. Within several days after emergence of female mosquitoes from the puparium, juvenile hormone (JH) stimulates the separation of ovarian follicles from germaria and limited growth of the ovarian follicle to its pre-vitellogenic resting state [ 2 ]. JH also confers competence to fat body cells and ovarian follicles for uptake of ecdysteroidogenic hormone (OEH). Then, in response to a blood meal, gonadotrophins are released from cerebral neurosecretory cells and cause the ovaries to become OEH-responsive [ 3 ]. OEH stimulates the ovaries to secrete ecdysone, the precursor to 20-E, as well as 20-E during vitellogenesis [ 4 - 6 ]. Fat body cells take up ecdysone, convert it to 20-E and use it to activate transcription of vitellogenin genes [ 7 ], the genes encoding the major egg-yolk proteins, as well as a large number of other genes, many of whose products will be incorporated into eggs [see [ 8 , 9 ] for reviews]. Prior to the blood meal, female mosquitoes access sugars for nutritional sustenance. During the first several hours following a blood meal, the mosquito undergoes physiological changes in addition to hormonal ones. Acquisition of a blood meal stimulates midgut proteolytic activity such that approximately 80% of the protein content is digested within one day [ 10 - 14 ]. Serine proteases including trypsins and chymotrypsins are responsible for the majority of endoproteolytic activity [ 11 , 12 , 15 ]. The role of trypsins in blood digestion has been well documented in Aedes aegypti , and more recently it has been investigated in An. gambiae . Despite the digestive proteolysis peak at 24 hours post blood meal, digestive enzymes exhibit two phases of transcription [ 16 , 17 ]. In Ae. aegypti there are three trypsins, early trypsin, which is constitutively expressed prior to blood feeding and two late trypsins which are blood induced. These two types of trypsins are also found in An. gambiae . The An. gambiae trypsin family includes seven genes clustered within 11 kb on chromosome 3R, in division 30A, that encode five functional proteins [ 18 ]. Trypsins 1 and 2 are both induced by a blood meal and exhibit similar expression profiles. In contrast to Trypsins 1 and 2, Trypsins 3, 4, and 7 are constitutively expressed in unfed females [ 18 ]. Trypsins 3 and 7 are down-regulated following a blood meal and not expressed again at levels detectable by RT-PCR until 28 hours post blood meal [ 18 ]. In addition to the trypsins, three chymotrypsin genes have been isolated and characterized in An. gambiae , two of which are located in tandem on chromosome 2L, in division 25D [ 19 , 20 ]. Both of these genes, AnChym 1 and 2, are expressed in the midgut by 12 hours post blood meal and their transcripts are abundant until 48 hours, as determined by PCR, unlike the levels of Trypsins 1 and 2 that have decreased dramatically by this time [ 19 ]. In contrast, the other characterized chymotrypsin, AgChyL, exhibits transcript level changes more similar to those of Trypsins 3–7 [ 20 ]. Two types of exopeptidases, carboxypeptidases and aminopeptidases, have been characterized in Anopheline mosquitoes. Edwards et al. [ 21 ] cloned a carboxypeptidase that was rapidly induced in An. gambiae midguts following blood meal ingestion. Multiple aminopeptidases have been isolated from hematophagous insects, and it has been suggested that they may play different roles in digestion [ 22 - 25 ]. Additional enzymes including glycosidases and lipases are also required for the digestion of non-proteinaceous blood constituents [ 26 , 27 ]. In addition to dramatic changes in physiology, blood feeding also induces changes in mosquito morphology. Following gut distension by blood ingestion, midgut epithelial cells secrete a Type I peritrophic matrix (PM) that is continuous along the length of the midgut [ 14 , 28 , 29 ]. Prior to the blood meal, the midgut epithelial cells contain high concentrations of apically located, morphologically granular, secretory vesicles. Presumably these apical granules contain precursors of the peritrophic matrix: as early as an hour after the adult female has taken a blood meal, they are no longer detectable [Staubli et al., 1966, as cited in [ 30 ]]. In An. gambiae , the PM can be visualized by electron microscopy as early as 12 hours PBM and it is fully formed by 48 hours PBM [ 28 , 31 ]. The PM is a biochemically complex structure containing not only chitin and other proteoglycans, but as many as 20–40 different proteins [ 31 - 33 ]. However, only one gene encoding a peritrophic matrix protein has been cloned in An. gambiae [ 34 ]. The exact functions of the PM remain unknown, but it has been suggested that this semi-permeable porous structure may function as a restrictive layer protecting the midgut epithelium from proteolytic digestive enzymes, from haematin crystals that form following hemoglobin breakdown and as a barrier to blood-borne pathogens including bacteria and malaria parasites [reviewed in [ 35 ]]. Once the adult mosquito acquires a blood meal, she spends approximately 48 hours converting about 20% of it into egg constituents [ 36 ], using another fraction of it to support the intense biosynthetic activities of this period and defecating the rest. Oogenesis in the mosquito ovary actually begins post-eclosion but oocyte growth attenuates at a resting stage until blood meal ingestion. Once reinitiated, egg development continues until oviposition. Successful egg production not only requires ovarian events for development and maturation of oocytes, but also synthesis of yolk constituents, both protein and lipid, in the fat body, followed by their uptake by oocytes and storage for later use during embryogenesis. Collectively, the events of yolk synthesis, uptake and storage constitute the process of vitellogenesis [ 30 ]. Vitellogenesis and oogenesis require the coordination of molecular events in at least these two different abdominal tissues, the fat body and ovary. Based on morphological and physiological criteria, the ovarian cycle can be divided into four phases: 1) Pre-vitellogenic, 2) Initiation, 3) Trophic, and 4) Post-trophic Phase [ 30 ]. The meroistic ovary of An. gambiae contains approximately 50 functional egg-production structures, the ovarioles. Each ovariole is comprised of two parts, a distal germarium and a vitellarium proximal to a common oviduct through which eggs will pass as they are laid. In the germarium, mitosis of the primordial germ cells creates a syncitium with an oocyte and seven nurse cells interconnected by intracellular bridges, or ring canals as a result of incomplete cytokinesis. Both the germ cell and the nurse cells are surrounded by a somatically derived follicular epithelium [ 37 - 39 ]. The first pre-vitellogenic phase is completed within three days of eclosion and ends with the separation of these follicles from the germaria and entry into the vitellaria. At the end of this phase, oocytes may have undergone some growth but then arrest until events initiated by acquisition of a blood meal cause them to become competent for ovarian vitellogenic events. Ingestion of a blood meal reinitiates ovarian development and follicle growth resumes. In Ae. aegypti and Anopheles albimanus , this period appears variable, lasting 3–10 and 8–16 hours, respectively, and ends with the initiation of vitellogenin synthesis [ 30 ]. In the next two days, during the trophic phase, the mosquito generates large amounts of vitellogenin, the secreted precursor to the major yolk protein vitellin. In addition to vitellogenin, the developing oocytes also accumulate other proteins, and lipids from the hemolymph, as well as ribosomes and mRNAs synthesized in the syncitial nurse cells. These latter constituents are transported to the germ cell through the ring canals connecting the oocytes and nurse cells by a process of cytoplasmic streaming [ 40 ]. Following delivery, several maternal mRNAs become localized within the oocyte. These maternal transcripts are fundamental for dorsal/ventral and anterior/posterior patterning of the embryo that will develop from the oocyte. This pattern of deposition and the patterning of the eggshell also depend on a complex signaling process involving both the somatic cells of the follicular epithelium and the oocyte. Once oocyte growth has ceased, vitellogenin synthesis terminates. This signals the onset of the post-trophic phase. During this time, the oocytes mature and eggshell structures begin to develop. The chorion, part of the eggshell, is secreted by the follicular epithelium and contains two layers, the first secreted, inner endochorion and the later secreted, outer exochorion [ 30 ]. It is the endochorionic layer that will harden and melanize after oviposition. Specialization of eggshell structures necessitates communication between cells. The RAS 1 signaling cascade is an important means of communication during the processes of oocyte and eggshell patterning, as it is during eye development and differentiation of structures late in embryogenesis [ 41 , 42 ]. During patterning in Drosophila , developing oocytes produce the TGFα protein Gurken that binds to the epidermal growth factor receptor (EGFR), a receptor tyrosine kinase (RTK) localized to the posterior follicle cells, to initiate RAS 1 signaling. Downstream from the activation of this RTK, the GTP-binding protein RAS 1 initiates a series of enzymatic events propagated successively by three protein kinases, RAF, MAPK, and MAPK kinase (MEK), resulting in the translocation of nuclear factors and possibly the concomitant reorganization of the cytoskeleton [reviewed in [ 43 ]]. Thus, a cascade of events leads to the establishment of the posterior follicle cell fate. The posterior follicle cells then signal back to the germ cells. This results in the reorganization of the oocyte cytoskeleton, and regulates the localization of anterior/posterior determinants. Similar to the eggshell, the oocyte also undergoes dorsal-ventral patterning. Following patterning of the follicle cells, maternal gene products are regulated by the Toll signaling pathway to generate a transcription factor gradient that will spatially regulate activity of specific zygotic genes within the fertilized oocytes [ 44 ]. Vitellogenic events in the fat body have also been divided into phases: 1) Pre-vitellogenic, 2) Vitellogenic, and 3) Termination. The pre-vitellogenic phase in the fat body coincides with the pre-vitellogenic phase of the ovarian cycle. During this phase, RNA synthesis increases in the fat body and the rough endoplasmic reticulum and the Golgi complex proliferate to prepare for the production of vitellogenin. At the start of the vitellogenic phase, the release of mosquito hormones initiated by digestion signal the onset of vitellogenesis [ 45 ]. Synthesis of large amounts of vitellogenins is facilitated by the large quantities of biosynthetic machinery generated during pre-vitellogenic stages, but also depends on the presence of multiple vitellogenin genes (Romans, unpublished). Following synthesis, vitellogenin is released into the hemolymph and eventually diffuses through channels between the cells of the follicular epithelium, whereupon it is accumulated by the oocytes by a process of receptor-mediated endocytosis in clathrin-coated pits [ 8 ]. When vitellogenesis has ceased, during the termination phase, the biosynthetic machinery in the fat body is degraded via a lysosomal pathway, at least in Ae. aegypti [ 46 ]. Thus, blood feeding initiates a complex series of physiological events in at least three tissues that are integrated by the actions of JH, 20-E and peptide hormones. These events may be required for parasite development; they certainly can be modulated by the presence of parasites [ 47 , 48 ] and may provide points of intervention for mosquito control. Microarray analysis provides a tool to study global expression patterns of thousands of genes simultaneously. By comparing the level of transcription of a gene over time between two states, e.g . blood-fed vs . sugar-fed, an expression signature for each gene can be defined in response to blood feeding. Consequently, these expression patterns may indicate how these genes are regulated and interact, and also the biological processes in which the act. In this study we performed microarray analysis of genes in female mosquito abdomens during the first 48 hours after a blood meal. We have implicated many of these genes in different processes stimulated de novo by blood feeding. The elucidation of the expression profiles of abdominal genes will provide a broadened basis for understanding vector-parasite interactions. Our study certainly provides insights into the physiology of the malaria vector Anopheles gambiae . Results Array composition Microarray analysis was conducted on 3057 cDNA clones generated from three different adult female An. gambiae mosquito abdomen-derived cDNA libraries to elucidate major patterns of gene expression through 48 hours post ingestion of a blood meal. Arrays were constructed from triplicate spotted negative controls (purified water, 3 × SSC with no DNA, and empty wells), positive controls for blood-fed samples consisting of 3 clones whose ESTs corresponded to rat ( R. norvegicus ) α and β hemoglobin chains, and PCR-amplified fragments obtained from 1132, 721, and 1204 clones randomly picked from the sugar-fed (harvested after 30 hours at 19°C), rat blood-fed (harvested 30 hours PBM at 19°C), and P. berghei infected rat blood-fed (harvested 30 hours PBM at 19°C) abdomen libraries, respectively (Table 1 ). Approximately 84% of PCR-amplified fragments were visualized on ethidium bromide stained 1% agarose, 1 × TBE gels prior to spotting (data not shown). Of these PCR-amplified fragments, 2219 clones (87% of electrophoresed PCR products) were represented by a single defined band (Table 2 ). Table 1 Microarray composition Controls Clones Negative Controls 108 Positive Controls 3 Libraries Sugar-fed Adult Female (incubated 30 hours at 19°C) Abdomen library 1132 Blood-fed Adult Female (incubated 30 hours PBM at 19°C) Abdomen library 721 Plasmodium berghei Blood-fed Adult Female (incubated 30 hours PBM at 19°C) Abdomen library 1204 Total 3168 Table 2 Appearance of PCR product following gel electrophoresis Total Clones Singlet Doublet Smear No Product Visible Amplification of 2558 Clones 2558 2098 (82%) 117 (5%) 6 (<1%) 337 (13%) Re-amplification of 183 Clones previously amplified with "No Product Visible" 183 121 (66%) 14 (8%) 0 48 (26%) Cumulative 2558 2219 (87%) 131 (5%) 6 (<1%) 202 (8%) Note: PCR-amplified fragments were visualized on ethidium bromide stained 1% agarose, 1 × TBE gels. Approximately 13% of PCR products could not be visualized following the first amplification due to product yield below the threshold of ethidium bromide detection; 54% of these (183) were re-amplified. ESTs corresponding to these spotted cDNAs were screened for mitochondrial contamination, filtered based on sequence trace file quality, and assembled (EST clustered) using the DNAstar Seqman II software (DNAstar, CA) (Table 3 ). The high quality ESTs clustered into 491 contigs (consensus sequence generated from ≥2 overlapping ESTs) and 1415 singletons (ESTs with no sequence similarity to any other EST in the assembly) for a total of 1906 unique transcripts (Table 4 ). Table 3 EST composition of array Number Percentage High Quality Sequence Data 2707 88.56% Poor Quality Sequence Data 131 4.28% Mitochondrial DNA 222 7.25% Total 3060 100% Note: EST analysis includes the ESTs of the positive controls ( Rattus norvegicus hemoglobin chains). Table 4 Putative transcripts represented on the array following EST assembly Number ESTs Represented Contigs 491 1292 Singletons 1415 1415 Total 1906 2707 Note: EST assembly includes the ESTs of the positive controls ( Rattus norvegicus hemoglobin chains). Microarray and bioinformatic analyses Global patterns of greater than two-fold up-regulation or down-regulation for these cDNAs were established by comparing transcript levels in blood-fed An. gambiae adult females at ten time points during and post ingestion of a blood meal to the levels in sugar-fed females. First strand cDNA was generated from total RNA collected at 5 min and 30 min after initiation of blood feeding and at 0, 1, 3, 5, 12, 24, and 48 hr post-blood meal. All cDNA populations were labeled and hybridized to arrays. For each PCR-amplified insert, Cy3 and Cy5 fluorescent dye levels were measured from 3 replicate spots on each of 50 arrays to generate average signal intensities, and an expression ratio depicting transcript fold change between sugar-fed and blood-fed mosquitoes calculated. Following quality control filtering and normalization, 456 cDNAs and the rat β-hemoglobin gene, the positive control, were expressed more than twofold above or below control, sugar-fed levels at one or more of the 10 blood-feeding time points. Following EST analysis, the 456 cDNAs were found to represent 413 unique mosquito transcripts, 10 of which were present in more than one set. This anomaly is due to EST clustering of alternatively spliced transcripts with different expression patterns. More unique transcripts are up-regulated than down-regulated in response to blood feeding, while 10% of them are both up-regulated and down-regulated over the time course of this study: 192 are up-regulated at least twofold, 173 are down-regulated at least twofold, and 48 are down-regulated and up-regulated. Bioinformatic analyses of these 413 unique transcripts showed that all sequences shared sequence identity with the An. gambiae genome (Table 5 ), 90% of which shared sequence identity with an entry in Nr of dbEST (Table 5 ). In this analysis Blast hits with an E value ≤1 × 10 -4 were considered significant. Table 5 Sequence similarity of consensus sequences Blastn (WGS An. gambiae ) Blastx (Nr) Blastn (Nr) Blastn (dbEST) Unannotated Total 3 Positive Controls No Significant Hit 3 3 403 Unique Transcripts 403 240 8 112 43 403 Early Genes 144 77 2 45 20 144 Middle Genes 130 78 6 36 10 130 Late Genes 139 93 2 31 13 139 Total 413 248 10 112 43 413 Microarray gene clustering and principal components analysis The behaviors of the gene products identified as at least twofold up/down-regulated were grouped into three sets using k -means clustering (Figure 1 ) and named according to the time of their induction during the 48-hour time course following the initiation of blood feeding. Set 1, hereafter referred to as the "Early Genes", contains 144 unique transcripts derived from 152 cDNAs, which are expressed mainly during the early time points (Table 6 , Figure 1A ). The majority of these genes are appear induced at least twofold more abundantly than in sugar-fed mosquitoes during the first five minutes of blood feeding. Many of these transcripts remain induced until 1-hour PBM, although some remain induced until 5 hours PBM. After 5 hours post blood meal, the majority of Early Genes is down-regulated and they remain down-regulated even 48 hours after blood meal ingestion. A small subset of the Early Genes shows a variant pattern of gene expression in which the transcripts are up-regulated from the first 5 minutes of blood uptake through 1 hour PBM followed by a repression in expression from 3 to 24 hours PBM and then a greater than twofold induction at 48 hours PBM. The 130 unique transcripts represented by 147 cDNAs in Set 2, the "Middle Genes", follow a more dynamic pattern of gene expression than the Early Genes (Table 6 and Figure 1B ). Most Middle Genes are down-regulated in blood-fed versus sugar-fed mosquitoes until 3 hours PBM followed by an increase in expression commencing at 5 hours PBM and peaking between 12 and 24 hours PBM. Subsequently, Middle Genes are down-regulated to initial transcript abundances by 48 hours PBM. Also, in a behavior largely exhibited by the Middle Genes, approximately 40% of genes are down-regulated when the mosquitoes completed feeding and left the rat (0 hours PBM). Set 3, the "Late Genes" contains 139 unique transcripts, 157 cDNAs, which are either down-regulated or constitutively expressed until 12 to 16 hours PBM after which they are up-regulated and, in contrast to the Middle Genes, continue to be highly expressed even at 48 hours PBM (Table 6 , Figure 1C ). Figure 1 Gene trees displaying the microarray generated expression profiles of abdomen-derived cDNAs in blood-fed compared to sugar-fed adult female mosquitoes during the following times: 5 and 30 minutes during blood meal (DBM), 0, 1, 3, 5, 12, 16, 24, and 48 hours post blood meal (PBM). k -means clustering of all genes up-regulated and down-regulated at least two-fold during at least one of the ten time points generated three sets of genes. These k -means-derived groups of genes were hierarchically clustered for visualization and include Set 1 designated the Early Genes (A), Set 2 designated the Middle Genes (B), and Set 3 designated the Late Genes (C). Each gene is represented by a single row of colored boxes; each time point is represented as a single column of colored boxes. The expression scale is represented as a gradation of color ranging from 5 fold induced genes indicated by saturated red to 2.5 fold repressed genes indicated by saturated green. Table 6 Bioinformatic analysis of two-fold expressed genes Early Genes Middle Genes Late Genes Total Twofold Expressed cDNAs (represented by consensus sequences >100 bp in length) 152 147 157 456 Replicate Consensus Sequences within k -means Sets 8 17 18 43 Unique Transcripts 144 130 139 403* Unique Transcripts included within PCA Analysis 82 69 98 249 *The complete data set (Early, Middle and Late genes combined) contains 413 unique transcripts. However, ten transcripts are present in two different gene sets and were counted twice as a result. Principal components analysis (PCA) of these 413 unique differentially expressed transcripts was also conducted to support the k -means defined sets. Following PCA, each transcript was plotted in a scatter plot comparing the PCA 1 and PCA 2 values. The three k -means-defined sets of genes did not overlap on these scatter plots. Thus the PCA results support classification of these transcripts into three groups. For each of the k -means defined groups, genes that had a PCA 1 or PCA 2 value greater than 0.5 or less than -0.5 were plotted on a parallel coordinates display (Figures 2 and 3 ). The resulting data sets contained 82, 69, and 98 unique transcripts representing Early Genes, Middle Genes, and Late Genes, respectively (Table 6 ). Figure 2 Scatter plot of the first two components for each gene that is either up-regulated or down-regulated at least two-fold during one time point in the course of this experiment. The Early Genes, Middle Genes, and Late Genes are colored red, blue and green, respectively. The genes colored in grey include Early, Middle and Late genes that did not have a value greater than 0.5 or less than -0.5 of both PCA 1 and PCA 2 components. Figure 3 Parallel coordinates display of expression profiles of differentially expressed PCA-filtered genes from the Early Genes (A), Middle Genes (B), and Late Genes (C). X-axes correspond to a successive time point. Y-axes denote the ratio of fluorescent intensities of blood-fed to sugar-fed samples at each time point for each gene in Panels (A), (B), and (C). Plotted genes had a PCA 1 value greater than 0.5 or less than -0.5, or a PCA 2 value greater than 0.5 or less than -0.5. In Panel (D), the activity levels of juvenile hormone (JH) and ecdysone (20-E) are plotted on a similar parallel coordinate graph (modified from Dhadialla and Raikhel 1994). Gene annotation and gene ontology assignments To identify An. gambiae genes whose products are involved in related processes, the EST consensus sequences of the transcripts differentially expressed in these 3 patterns were annotated using sequence similarity and categorized using the molecular functions listed by the Gene Ontology Consortium (GOC) and the biological processes defined by Holt et al. [ 49 ]. Gene annotations for all 413 at least twofold differentially expressed gene products are given in the Supplementary Table S1. They were then categorized into 9 major categories with 31 subdivisions (Table 7 ). 48% of the genes could not be annotated and therefore were categorized as "Unknown". The three most numerous categories containing annotated gene products were "Metabolism", "Protein Synthesis", and "Egg Production". Table 7 Functional annotation of AS represented by microarray expression group Early Middle Late Metabolism Simple/Complex Carbohydrate Metabolism and Transport 3 0 2 Oxidative Phosphorylation 8 5 1 Lysosomal Enzymatic Digestion 0 0 2 Protein Digestion 4 4 3 Protein Modification, Metabolism, Transport and Localization 5 12 8 Amino Acid and Derivative Metabolism and Transport 1 8 2 Nucleobase/Nucleoside/Nucleotide/Nucleic acid Metabolism and Transport 4 1 3 Fatty Acid/Lipid Metabolism and Transport 1 0 2 Vitamin/Vitamin Derivative/Cofactor Metabolism and Transport 2 0 0 Xenobiotic Metabolism and Transport 1 0 2 Total 29 30 25 Transport Ion Transport 3 3 0 Receptor-mediated Endocytosis 1 0 2 Total 4 3 2 Protein Synthesis Transcription and mRNA Processing 2 5 6 Translation 11 10 4 Protein Folding 3 4 4 Total 16 19 14 Cellular Processes Cell Cycle 0 3 3 Cellular Proliferation 0 0 2 Chromatin Assembly/Disassembly 0 0 5 Apoptosis 1 0 0 Senescence 1 0 0 Total 2 3 10 Egg Production Vitellogenesis/Oogenesis/Embryogenesis 2 10 7 Melanization 0 1 0 Total 2 11 7 Cellular Communication Signal Transduction 1 1 2 Cell-cell Signaling 3 0 1 Total 4 1 3 Intra-/Extra-cellular Architecture Maintenance Structural 3 4 2 Muscle-related 1 0 0 Cell Adhesion 1 1 0 Cytoskeleton Organization and Biogenesis 1 0 2 Total 6 5 4 Response to Stress/External Stimulus Response to Oxidative Stress 2 1 3 Immune/Defense Response 2 2 3 Total 4 3 6 Unknown Total 77 55 68 During the 48 hours PBM, the majority of gene products involved in metabolism were up-regulated Early and Middle Genes. Largely different metabolic biological processes were up-regulated in Early vs. Middle and Late Genes. More than half of the Early metabolic gene products, 20/29 unique transcripts, appear to be involved in carbohydrate metabolism, oxidative phosphorylation, and protein digestion. In contrast, 80% of the metabolic genes, 24/30 unique transcripts, represented in the Middle Genes contribute to various processes in protein digestion and metabolism, and metabolism of amino acids and their derivatives. One third of the Late genes involved in metabolism, 8/25 unique transcripts, are involved in protein metabolism. Five of these annotated sequences, ASs 368, 807, 1179, 279, and 922, encode products involved in post-translational modification. Reflecting the necessity of biosynthetic machinery in cell maintenance and growth, and probably also the highly conserved nature of proteins involved in housekeeping functions, the protein synthesis category contained the second largest number of genes functioning in a known process. 69%, 11/16 unique transcripts and 53%, 10/19 unique transcripts, respectively of the Early and Middle Genes, in this category are involved in translation. Approximately 25%, 5/19 unique transcripts, and 40%, 6/14 unique transcripts, of the protein synthesis genes represented among Middle and Late genes, are required for transcription and mRNA processing. This result seems almost paradoxical because transcription and mRNA processing necessarily precede translation. A number of biological processes were related by their involvement in nuclear events or the overall activity of the cell. There were three times as many genes involved in cellular processes among the Late genes than in the other two sets combined. Although all the genes involved in the cell cycle are Middle and Late Genes, the most strikingly up-regulated cellular process genes were those involved in chromatin assembly/disassembly. 5 of the 10 cellular process Late genes (ASs 1136, 59, 592, 1011, and 1792) are involved in maintenance of chromatin structure, a biological process represented only in the Late genes. These Late genes include both histones and high mobility group proteins. The majority of significantly up and down regulated genes appear to function in egg production, either in the development and maturation of oocytes or in the fat body synthesis of products that will be deposited in oocytes. Almost 90% of them, 17/19 unique transcripts, are Middle and Late genes. These genes are described in detail in the Discussion section. In contrast, half of the genes categorized as cellular communication genes, 4/8 unique transcripts, are Early genes. The majority of the cellular communication gene products in the combined sets of Middle and Late genes, are involved in different signal transduction pathways. Additionally, almost half of the intra-/extracellular architecture maintenance genes are Early genes. This category includes a wide variety of gene products such as peritrophin, both muscle-related and cytoskeletal actins, α-catenin and β-integrin. The Middle and Late genes in this category were mainly structural and included two peritrophins (ASs 13 and 642). The biological process categorized as transport included not only the movement of ions such as zinc, sodium and potassium, but also transport of molecules via receptor-mediated endocytosis. All three transport gene products, ASs 1336, 1605, and 432, in the Middle genes are responsible for the movement of ions. In contrast, in the Early and Late genes, several gene products (ASs 1974, 1071, and 2086) appear to be involved in receptor-mediated transport via clathrin-coated vesicles. A number of genes responding to oxidative stress (6 genes in total) were found in all three sets of genes indicating that they are transcribed throughout the 48 hours PBM. Seven additional gene products most probably involved in immunity, a response to external stress, were found among these three gene sets. qRT-PCR analysis Expression profiles of eight selected genes and the RP S7 control gene were confirmed using a quantitative real-time PCR strategy (Figure 4 ). Transcript levels for each of the eight genes were quantified using SYBR Green technology and differences in their expression between sugar-fed and blood-fed mosquitoes at 0, 5, 12, 24 and 48 hours PBM determined. Although the magnitudes of the changes in transcript abundances of all the genes whose expression levels were quantified by both techniques differed between the techniques, the changes in direction of expression, whether positive or negative, remained consistent for the majority of them. In addition, the overall patterns of expression exhibited by the three sets of genes were also apparent in the expression profiles created by qRT-PCR analysis. For the two Early genes, microarray analysis overestimated transcript levels between 2- and 30-fold more than qRT-PCR analysis. In contrast, for the majority of the Middle and Late gene expression measurements, microarray analysis underestimated transcript abundances relative to qRT-PCR analysis. Figure 4 Comparison of microarray and qRT-PCR gene expression profiles for selected genes. X-Y plots were generated from the ratio of transcript levels in the blood-fed adult female mosquitoes to the transcript levels in the sugar-fed adult female mosquitoes for eight selected genes at five time points including 0, 5, 12, 24 and 48 hours post blood meal (PBM). Genes were randomly selected from the three sets and included ASs 648 and 1786 from the Early Genes (A), ASs 996, 1158, 1949, and 679 from the Middle Genes (B), and ASs 12 and 1357 from the Late Genes (C). Triangles indicate fold expression data generated by qRT-PCR analysis; circles indicate fold expression data generated by microarray analysis. A horizontal line connecting either the diamonds or circles illustrates each gene expression profile. For each gene, the expression profiles created by qRT-PCR and microarray analysis are indicated using the same color. Discussion We have determined the gene expression patterns of 3,068 abdomen-derived cDNAs from adult female An. gambiae mosquitoes representing 1906 unique transcripts were determined in the first two days following ingestion of a blood meal by microarray analysis. 413 unique transcripts were shown to be up-or down-regulated at least twofold in blood fed mosquitoes relative to sugar-fed mosquitoes at one or more of the following times: 5 min and 30 min following initiation of blood feeding and 0, 1, 3, 5, 12, 24, and 48 hr post-blood meal. These transcripts were clustered into three sets with different temporal patterns of expression that may reflect the major hormonal changes occurring within the mosquito during a gonotrophic cycle. These differentially expressed gene products were annotated putatively using sequence similarity searches and categorized by biological process to identify the major events occurring post blood meal ingestion in the female mosquito. Multiple hormones interact to alter tissue states and to activate genes involved in the female mosquito's digestion of a blood meal, in oocyte development and in vitellogenesis. The three sets of differentially transcribed genes discerned in this study, the Early, Middle and Late genes may reflect differential hormonal responsiveness. After acquisition of a blood meal, the transcript levels of the Early genes which were abundant during blood feeding showed general declines until 24 hours PBM, after which a subset of transcript levels began to rise again. Expression of Early genes may be linked to the relatively high titers of JH present at the beginning of the first gonotrophic cycle and may then be repressed as a result of declining JH titers or of increasing 20-E. The expression of Middle gene transcripts followed an expression pattern that reflects the titers of 20-E: levels sharply increased by 12 hours PBM, remained stable or increased only slightly until 24 hours PBM, and then declined rapidly. OEH secreted by median neurosecretory cells stimulates the ovaries to secrete 20-E during vitellogenesis and the activity of this hormone begins to rise by 3–5 hours, peaks between 12 and 24 hours, and then declines to baseline levels by 48 hours post blood ingestion [ 4 - 6 ]. In contrast to the Middle genes, most Late gene transcripts exhibited baseline, steady state levels until 12 hours PBM after which they were induced at least twofold and continued to exhibit increased transcript levels at 48 hours PBM. This increase in transcript levels mirrored the increase in JH titer observed by 48 hours post blood meal ingestion [ 50 ]. These results suggest that Middle genes products may be ecdysone-responsive whereas Late genes products may be JH-responsive. Based on their finding that the Drosophila minidiscs gene product showed a primary response to JH, Dubrovsky et al. [ 51 ] have suggested that JH may transcriptionally regulate genes encoding maternally inherited products. Our Late gene, AS 806, shares sequence similarity with minidiscs . Additionally, the majority of mosquito gene products showing sequence similarity to maternally active Drosophila genes are categorized as Late genes. Whether transcription of Early genes is directly influenced by levels of JH or 20-E, cannot be determined easily because neither hormone is present at high levels during the first few hours following blood meal ingestion. Regardless of expression pattern, the gene products in each set reflect a diverse array of processes occurring in the female mosquito within 48 hours following initiation of blood feeding. The major processes initiated in response to blood feeding including digestion, peritrophic matrix formation, oogenesis and vitellogenesis, are discussed below with emphasis on the likely roles of particular gene products. Digestion Digestion of the two different food sources, nectar sugars and blood, requires changes in the types of enzymes present within the digestive tract of the mosquito. The numbers of genes associated with sugar and protein metabolism within each set of genes may reflect the switch from sugar to protein metabolism. For instance, there are increases in transcript abundance of genes involved in carbohydrate metabolism and oxidative phosphorylation in the Early genes. However, both Early and Middle gene categories are enriched in genes involved in protein digestion. This result stems from the involvement of certain gene products in blood meal digestion that initiate a signaling cascade resulting in the up-regulation of other related proteolytic enzymes. The Middle genes contain the majority of gene products involved in amino acid metabolism, a process that follows protein digestion, whereas by 48 hours PBM, the time at which the majority of Late genes are induced, there is a generalized decrease in digestive enzyme transcripts. Considering that blood contains large quantities of protein, the mosquito requires a variety of proteolytic enzymes to digest the recently acquired meal. In the present study, 11 genes were identified whose products are most likely required for protein digestion. These include 5 previously characterized digestive enzyme genes, two trypsins, a chymotrypsin, a serine protease and a carboxypeptidase. The majority of these digestive enzyme genes were transcribed at levels greater than twofold induction after 6 hours PBM. The An. gambiae Trypsins 1 and 2 are both induced by a blood meal and exhibit similar expression profiles although Trypsin 1 is expressed at higher levels. Muller and coworkers, using an RT-PCR strategy, showed that Trypsin 2 mRNA is present at 8, 12, 24 and 28 hours post blood meal [ 18 , 52 ]. In our study, the Trypsin 2 gene (AS 569) also exhibited increased transcript abundance at 12 hours PBM with maximal expression occurring at 24 hours PBM, but these levels decreased by 48 hours. In contrast to Trypsins 1 and 2, Trypsins 3, 4, and 7 are constitutively expressed in non-bloodfed females. By 4 hours following a blood meal, levels of Trypsin 4 become undetectable by Northern and RT-PCR analysis and do not reappear until 20 hours PBM [ 18 ] and unlike Trypsins 1 and 2, Trypsin 4 reaches maximal expression by 48 hours, near the end of the gonotrophic cycle. Our study identified two clones corresponding to Trypsin 4 (AS 568) but the two cDNAs exhibited different expression patterns. One (Accession no. CD747033) reached maximal transcript abundance at 48 hours PBM, the expected expression pattern. However, the other cDNA (Accession no. CD747029) was expressed at high levels prior to 6 hours PBM and also reached at least twofold increased levels by 48 hours PBM. These cDNAs may not have clustered together for technical reasons intrinsic to Seqman II, or they may correspond to alternatively spliced variants of the same gene. The sequence alignment showed 97% nucleotide sequence identity between the two ESTs over the region common to both. However, the CD747033 EST is only 253 bp in length and it is possible that the full-length cDNA represents an alternatively spliced transcript. In addition to Trypsins 2 and 4, we identified a trypsin-like serine protease (AS 648) which shared the greatest amino acid similarity with Trypsin 4. However, this serine protease exhibited highest nucleotide identity to a different region of chromosome 3R than that containing all previously identified digestive trypsins. This gene product was induced greater than twofold within 6 hours PBM and was repressed during the height of digestion. Unlike Trypsins 3–7, it was not expressed at higher levels at 48 hours PBM. It is possible that this trypsin is not involved in digestion but in another proteolytic process that is down-regulated following a blood meal. Barillas-Mury et al . [ 53 ] demonstrated that the early trypsin activity is essential to the transcription and subsequent expression of the late trypsins in Ae. aegypti . An. gambiae Trypsins 3–7 may indirectly activate transcription and increase the expression of Trypsins 1 and 2, the major endoproteolytic enzymes required for blood meal digestion [ 18 ]. In both Ae. aegypti and An. gambiae , these early expressed digestive enzymes are presumed to act as signal transducers causing transcriptional up-regulation of the late expressed ones [ 14 , 53 ]. Thus the digestive process is regulated by an elaborate biphasic expression pattern of serine proteases. Additional evidence suggests that not only tryptic peptides but cleaved amino acids may serve as systemic signals regulating subsequent processes. In contrast to the trypsin by-products, cleaved amino acids may also function as negative regulators of food intake. AS 1158 shared weak sequence similarity with the Drosophila pumpless protein, a larval fat body-expressed enzyme involved in glycine catabolism. In the fruit fly, larvae expressing the pumpless mutation are unable to pump food from the pharynx to the esophagus [ 54 ]. These mutant animals do not feed, neither do they upregulate genes normally involved in responses to starvation. Because feeding amino acids to wild type larvae phenocopied effects of the pumpless mutation, Zinke et al . [ 54 ] proposed that amino acids released from the fat body normally act as signals for cessation of feeding. In addition to the trypsins, three chymotrypsin genes have been characterized in An. gambiae [ 19 , 20 ]. The expression of two of these digestive enzymes, AnChym 1 and 2, has been localized to the mosquito midgut by analysis of Northern blots [ 19 ]. RT-PCR showed that both chymotrypsin genes are expressed at 12 hours PBM and are abundant until 48 hours PBM, unlike transcripts of Trypsins 1 and 2 which have decreased dramatically by this time [ 19 ]. The cDNA representing chymotrypsin 2, AS 99, was categorized as a Late gene since maximal transcript levels were achieved after 24 hours PBM. In contrast, the other characterized chymotrypsin, AgChyL, exhibits changes in transcript abundance which is more similar to those of Trypsins 3–7. mRNA is present in non-blood fed females, detectable until 8 hours post blood meal after which mRNA can no longer be measured until 48 hours PBM [ 20 ]. The cDNA corresponding to this chymotrypsin-like serine protease, AS 994, was induced more than twofold prior to the peak of digestion and clustered with the Early Genes. In addition to the aforementioned chymotrypsins, our study identified a previously uncharacterized chymotrypsin (AS 2243) also located on chromosome 2L but in a different region from both AnChym1 and AnChym2. Exhibiting an expression pattern different from both AnChym2 and AgChyL, this gene product was characterized as a Middle Gene with maximal transcript levels achieved between 12 and 24 hours PBM and a return to baseline by 48 hours PBM, similar to the expression patterns of Trypsins 1 and 2. Edwards et al . [ 21 ] investigated expression levels of An. gambiae carboxypeptidase A following blood meal ingestion. Northern blot analysis indicated that levels of Carboxypeptidase A mRNA rose rapidly to a ten-fold increase within 3–4 hours following a blood meal, then dropped to the pre-induction state by 24 hours PBM. We identified a carboxypeptidase gene located on chromosome 2L (AS 1742) that is transcribed in a manner similar to that of carboxypeptidase A. However another cDNA representing a carboxypeptidase (AS 44) exhibited a radically different expression pattern. Transcripts were present at low levels 1–5 hours PBM but increased more than twofold between 12 and 24 hours PBM, a pattern that resembled the enzymatic activity in An. stephensi observed by Jahan et al . [ 13 ], namely a rapid increase until 12 hours PBM, with a peak at 24 hours, followed by a steady decline over the next day. In contrast to the An. gambiae carboxypeptidase A, the levels of aminopeptidase peaked around 30 hours in An. stephensi [ 12 ]. Additionally, Lemos et al . [ 14 ] recorded peak aminopeptidase activity at 24 hours PBM in An. gambiae . We identified two aminopeptidases with at least two-fold increased expression during the 48 hours PBM. The first aminopeptidase, AS 340, reached peak transcript levels at 24 hours PBM, showing a similar expression pattern to the enzyme activity levels reported by Billingsley and Hecker [ 12 ] and Lemos et al . [ 14 ]. In contrast, the other aminopeptidase, AS 430, showed amino acid similarity to the Ae. aegypti aminopeptidase N. This aminopeptidase was classified as a Late gene, due to increased transcript levels at 24 hours PBM but maximal levels were not achieved until 48 hours PBM. Jahan et al . [ 13 ] documented two different kinetic profiles of aminopeptidase enzymes in An. stephensi depending on whether the enzyme was soluble or membrane-associated. The soluble aminopeptidase exhibited a kinetic profile similar to AS 340 and to that presented by Billingsley and Hecker [ 12 ] with peak activity at 24 hours PBM. Geering [ 55 ] had suggested that phospholipase activity plays a role in blood digestion in Ae. aegypti although no conclusive evidence was demonstrated. However, Geering and Freyvogel [ 56 ] demonstrated that lipolytic activity increased 15 hours after blood feeding. Of the three gene products that are characterized as being involved in Fatty Acid/Lipid Metabolism and Transport, two, ASs 1177 and 997, encoding an acetate-CoA ligase and a fatty acid binding molecule, respectively, are expressed at more than twofold greater abundance at 24 hours PBM. These gene products may be involved in fatty acid degradation of blood meal components and the transport of these lipids to the oocytes. The erythrocyte membrane contains a number of glycoproteins. It is therefore possible that enzymes normally associated with carbohydrate metabolism of nectar meals could also be involved in blood digestion. Almost all of the Simple/Complex Carbohydrate Metabolism and Transport genes identified in our study as being at least twofold upregulated were categorized as Early genes. Within the first 6 hours PBM, their transcripts are present at higher levels than in sugar-fed females and thereafter, they decrease steadily until 12 and 24 hours PBM. This expression profile does not exclude these genes from having a role in RBC glycoprotein metabolism. Several glycosidases are present within the midgut of An. stephensi , either associated with the lumen or with epithelial lysosomes [ 12 ]. The enzymatic activity of α-glucosidase, the major midgut glycosidase in An. stephensi , increased from 6 hours PBM to maximal levels by 24 hours and decreased to basal levels by 36 hours PBM in the anterior midgut. The transcript abundance of the α-glucosidase, AS1786, characterized in this study followed a different pattern than the enzymatic activity of An. stephensi α-glucosidase. It showed greatest amino acid sequence similarity to Drosophila melanogaster gene product CG8690 α-glucosidase and was categorized as an Early Gene with minimal transcript levels occurring at and after 24 hours PBM. Peritrophic matrix formation Shen and Jacobs-Lorena [ 34 ] characterized the An. gambiae peritrophic matrix protein Peritrophin 1 (Ag-Aper1) by analysis of Northern blots, and demonstrated that transcripts were present 6 hours PBM, increased by 12 hours, and remained elevated between 24–48 hours PBM. The present study identified several genes encoding proteins with a chitin-binding domain (InterPro ID IPR002557: Chitin binding Peritrophin-A) similar to the one found in Peritrophin 1. The Early gene AS 928 contains the Peritrophin-A chitin binding domain and maps to chromosome 3L in silico , corresponding to agCP10685. Another Early gene, AS 13, shows high identity to Peritrophin 1. A Middle gene, AS 516, is expressed by 6 hours PBM but does not reach maximal levels until 24 hours PBM. This gene product maps in silico to chromosome 2L, exhibits a similar transcript profile, and also shares 98% amino acid identity with Peritrophin 1. It is not clear why two sets of ESTs, both identified as Peritrophin 1, should exhibit different transcription patterns unless they are derived from differentially regulated genes. Another Middle gene, AS 1164, contains the Peritrophin-A chitin binding domain in addition to a prenyl-group binding site (InterPro ID IPR001230: CAAX box). This Middle gene may not be involved in peritrophic matrix formation but in some other process coinciding with protein digestion. Two other cDNAs, Accessions CD746211 and CD746202, both in AS 13, could also be localized to chromosome 2L. However, their ESTs exhibited greatest nucleotide identity to the gene predicted as ENSANGG00000020776 located 4 kb 3' to Ag-Aper1. These two cDNAs clustered as Early and Late genes respectively, and may represent alternatively spliced gene products. As early as an hour after the adult female has taken a blood meal, secretory vesicles previously present in the apical brush border of midgut epithelial cells are no longer detectable [Staubli et al., 1966 as cited in [ 30 ]]. These apical granules presumably contain precursors of the peritrophic matrix. The Middle and Late peritrophin gene products may be packaged into vesicles in preparation for a subsequent blood meal. In contrast, the Early peritrophin genes may be transcribed in response to blood meal acquisition and their products used immediately in the formation of the peritrophic matrix. Ovarian cycle and oogenesis The extensive literature on genes involved in Drosophila ovarian development and early embryogenesis opens windows into interpreting our An. gambiae microarray results and understanding mosquito egg development. The majority of An. gambiae genes upregulated at least twofold following a blood meal appear to function in egg production. The only gene possibly involved in oogenesis during the early phases of the ovarian cycle is the Early gene AS 670. It shares sequence similarity with peter pan , a Drosophila gene product required during oogenesis. Oocytes in peter pan mutants often have an incorrect number of associated nurse cells, suggesting that the peter pan protein influences the separation of cells within the germaria [ 57 ]. The identification of other genes involved in the early stages of mosquito oogenesis may be facilitated by the construction of cDNA libraries from the abdomens or ovaries of recently blood fed female mosquitoes. Several differentially expressed gene products found in the present study may be involved in the formation of ring canals, structures necessary for the delivery of maternal factors to oocytes. In particular, bulk transfer of cytoplasmic content from nurse cells to oocytes depends on actin structures [ 40 ]. A Middle gene product, AS 679, shares sequence similarity to the Drosophila gene CG13388 encoding the protein kinase anchor protein 200, Akap200 . Akap200 protein localizes to ring canals during oogenesis, regulates protein kinase C activity, and controls their morphology [ 58 ]. Late gene product AS 1317 shows sequence similarity to the Drosophila pendulin gene product, encoded by CG4799. This gene product is also required for assembly of fully functional ring canals. pendulin encodes an importin-α2, a protein necessary for the localization of the kelch gene product, CG7210. kelch encodes an actin organizer without which the ring canals become occluded and nurse cell-oocyte cytoplasmic transport is inhibited [ 59 , 60 ]. Though we found an apparent pendulin gene, we did not find kelch . A Late gene product, AS 578, is the homolog of Cdc42 , which encodes a small monomeric RHO GTPase involved in signal transduction. Rohatgi et al . [ 61 ] suggested that Cdc42 protein most likely links signal transduction to the actin cytoskeleton in Xenopus . In Drosophila ovaries, mutations in Cdc42 caused nurse cells to deflate and coalesce, and inhibited transfer of nurse cell cytoplasm to oocytes in late stage egg chambers [ 62 ]. Drosophila nurse cells transcribe the bicoid anterior determinant gene and the resulting mRNA is transported to the anterior region of developing oocytes via polarized microtubules [ 63 ]. bicoid does not appear to be an anterior determinant in other insects, but other genes important for its localization are conserved. The Early gene AS 2047 shares similarity with the Drosophila cornichon gene, CG5855. In Drosophila , cornichon is required for formation of a functional microtubular cytoskeletal scaffold used to transport bicoid mRNA and the posterior group oskar gene product to their proper location within the embryo [ 64 ]. The Late gene product, AS 1044 exhibits greatest sequence similarity to D. virilis exuperantia , a gene whose product is also required for proper bicoid mRNA localization D. melanogaster [ 65 , 66 ]. The Late gene AS 2222 is putatively identified as the An. gambiae homolog of Drosophila Notch . In Drosophila , Notch signaling regulates a large number of ovarian events beginning with cyst development in the germarium and extending through oogenesis [ 67 ]. The mechanisms by which Notch signaling activates transcription of its target genes are reviewed by Barolo and Posakony [ 68 ]. Since we identified Notch as a Late gene, its activities may be more restricted in An. gambiae , and/or reflect fundamental differences in ovarian biology. AS 1391, a Late gene product, shares sequence similarity with Drosophila Rab-protein 11 . This Drosophila small monomeric RAB GTPase is also involved in the polarization of the microtubules for the organization of the posterior pole and for oskar localization there [ 69 ]. To regulate the progress of oogenesis and embryogenesis, stored maternal mRNAs are translationally repressed during early oocyte development. The Middle gene product AS 2031 shares sequence similarity with the Drosophila gene product Bicaudal C , a RNA binding protein that may play a role in translational silencing of maternal mRNAs in addition to its role in eggshell patterning [ 70 ]. Mutations in Bicaudal C result in premature translation of oskar mRNA before it has reached the posterior region of the oocyte [ 71 ]. The Middle gene product AS 1490 is the putative homolog for the Drosophila gene product vasa (CG3506), an ATP dependent helicase involved in pole plasm assembly that may also be involved in translational modification of maternal mRNAs [ 72 ]. The Late gene AS 453 shares sequence similarity with the Drosophila cup protein (CG11181). cup protein interacts with nanos , the posterior determinant, and a translational regulator of the gap gene hunchback mRNA during oogenesis, although the exact function of the cup protein still remains unknown [ 73 ]. A DEAD box protein encoded by vasa also influences oocyte differentiation and the development of the Drosophila embryo body plan via translation of oskar , nanos , and g urken during oogenesis [ 74 - 77 ]. In amphibians, several mRNA binding proteins have been identified that are only present in oocytes and not post cleavage embryos [ 78 , 79 ]. One An. gambiae Middle gene product, AS 2449, shared sequence similarity with the Xenopus laevis poly(A)-specific ribonuclease and also has mRNA binding motifs, thus it may also repress translation of mRNAs in embryos. Maternal nurse cells not only provide the biosynthetic machinery and mRNA needed for oocyte axis determination, but also many transcripts and proteins required for zygotic development through the cellular blastoderm stage. The Middle gene product AS 1032 shares sequence similarity with nop5 , encoding a maternally derived product of the Drosophila CG10206 gene, a component of the small nucleolar ribonucleoprotein (snoRNP) complex involved in rRNA processing [ 80 ]. The Late gene product AS 806, referred to above in the context of its possible regulation by JH, shares sequence similarity with the Drosophila minidiscs gene product, an amino acid transporter. In Drosophila ovarian nurse cells, JH induces the expression of minidisks and its transcripts are most likely transferred to the oocyte during nurse cell cytoplasmic streaming [ 51 ]. Similar to the oocyte, the eggshell undergoes dorsal-ventral patterning. Crucial to this process is the correct placement of the oocyte relative to the maternal somatic follicle cells. In Drosophila , the localization of the oocyte depends on cadherin-associated adhesion [ 81 ]. The Late gene product AS 1890 is the homolog of α-catenin, the CG17947 gene product. A cytoskeletal anchor protein, α-catenin is required for positioning of the oocyte relative to the posterior follicle cells during germ cell rearrangement in Drosophila [ 81 ]. The RAS 1 signaling cascade is an important means of cell communication during embryo and eggshell patterning [ 41 , 42 ]. The Middle gene product AS 657 is weakly similar to the Drosophila Star protein, a RAS 1 enhancer involved in the EGF receptor signaling pathway, either upstream or in parallel to EGFR, during formation of the embryonic ventral midline. Star encodes a single pass transmembrane protein that may be involved in the processing of gurken protein. The DNA damage checkpoint 14-3-3epsilon protein also participates in RAS 1 signaling, normally functioning downstream or in parallel to RAF, but upstream of transcription factors. The Middle gene AS 106 shares sequence similarity with Drosophila 14-3-3epsilon . The 14-3-3epsilon protein is also capable of binding to a large number of other proteins in a phosphorylation-dependent manner. One of its functions may be to alter the cell cycle by binding Cyclin B and appears to have homologs in most if not all eukaryotes [ 82 , 83 ]. The An. gambiae Cyclin B homolog, AS 1357, grouped as a Late gene. Mutation screens in Drosophila have led to the identification of a number of other gene products that may be involved in RAS 1 signaling. TppII (tripeptidyl-peptidase II) , and smt3 (SUMO) were discovered in a search for lethal mutations that could enhance a weak RAS 1 eggshell phenotype [ 84 ]. The Middle gene product AS 1268 shares sequence similarity with Drosophila tripeptidyl-peptidase II , the CG3991 gene, encoding a serine protease that degrades neuropeptide signals [ 85 ]. The Late gene product AS 922 is the homolog of Drosophila smt3 , CG4494, whose product is ubiquitin-like protein that may tag proteins for nuclear localization or retention in the cytoplasm [ 86 , 87 ]. smt3 protein may modulate activity of transcription factors in the follicle cells downstream of EGFR activation. However, we feel that smt3 is likely to be a minor player in RAS 1 signaling in the events following blood ingestion in the mosquito, because its mRNA reaches maximal expression after 12–24 hours PBM, unlike the other gene products we identified as potentially influencing RAS 1 signaling. smt3 protein may also play a role in Toll signaling. This signal transduction pathway is known to be necessary for dorsal/ventral patterning of the Drosophila embryo. smt3 protein binds the NFκB homolog dorsal protein and targets this Rel transcription factor for migration to the nucleus [ 88 , 89 ]. Bhaskar et al . [ 88 ] demonstrated that smt3 conjugation to the dorsal protein enhanced its transcriptional activity. smt3 protein may play other roles in the cell by altering the interactions of septins, cytoskeletal proteins involved in cytokinesis [ 90 , 91 ]. In Drosophila , septins have been found in the cytoplasm of nurse cells and at the baso-lateral surfaces of follicle cells [ 92 ]. These results suggest even more pleiotrophc roles for the smt3 gene product in oogenesis. We also found that the Late gene AS 2034, a homolog of Drosophila Aos1 , the CG12276 gene, the smt3 (SUMO) activating enzyme, was also expressed at least twofold more abundantly during the height of smt3 expression. This result reinforces the importance of smt3 in the events occurring between 24–48 hours PBM. In addition to genes regulating the polarity of the embryo and eggshell, genes involved in cellular growth and differentiation were differentially expressed during the 48 hours PBM ingestion. AS 337 and 495 shared sequence similarity with the Ae. aegypti ornithine decarboxylase antizyme, a protein that modulates polyamine synthesis. The homologous Drosophila ornithine decarboxylase antizyme gene, formerly known as gut feeling , has been shown to be important in developing oocytes. It is one target of Sex lethal which encodes an RNA binding protein that regulates mRNA splicing and the mitotic events in early germ cells via regulating Cyclin B [ 93 ]. The Late gene product AS 2073 shares sequence similarity with the Drosophila polo CG12306 gene product, a protein kinase required for cytokinesis and another regulator of Cyclin B [ 94 ]. The Early gene AS 1972 shows identity with the Drosophila black pearl CG5268 gene product. This protein contains DnaJ domains implying that it is necessary for cellular growth [ 95 ]. Northern blot analysis of black pearl RNA from various developmental stages showed two transcripts with greatest expression in Drosophila embryos 0–6 hours old [ 95 ], the stages in which DNA replication recurs most rapidly. The Late gene AS 2268 shares sequence similarity with the Drosophila Imaginal disc growth factor4 ( Idgf4 ), a mitogen with a non-functional chitinase domain. Transcripts of Idgf4 are detected in Drosophila nurse cells, oocytes, and in the yolk cytoplasm of early embryos [ 96 ]. The Middle gene product AS 2273 shares sequence similarity with Drosophila β Integrin . An. gambiae β Integrin may interact with the Middle gene AS 985 product to promote somatic cell adhesion and cell migration during oogenesis and embryogenesis. This is due to the similarity of the AS 985 gene to Drosophila Receptor of activated protein kinase C , RACK1. RACK1 can bind a number of different signaling and cell adhesion molecules including the activated form of protein kinase C (PKC), Src family kinases, and β Integrins [ 97 - 99 ]. Cox et al . [ 100 ] demonstrated that, in a mammalian system, RACK1 organizes focal adhesions and directional cell migration via its Src-binding site. Mahairaki et al . [ 101 ] found that the An. gambiae β Integrin gene was expressed at highest levels 48 hours PBM, whereas we found that the β integrin homolog reached at least twofold increased expression by 24 hours PBM. A number of genes have been implicated in the development of the egg shell structures. Our screen does not appear to have identified any homologs of the several endochorionic structural proteins characterized in Ae. aegypti [ 102 , 103 ]. This was unexpected because Northern blot analysis had indicated that transcription of the vitelline membrane proteins 15a-1, 15a-2, 15a-3 was induced rapidly between 10 and 24 hours PBM, reached maximal levels between 30 and 40 hours PBM, and decreased to baseline levels between 50 and 60 hours PBM [ 102 , 103 ]. Our study also did not identify a Dopa decarboxylase (Ddc) gene, Ddc is an enzyme involved in the tyrosine metabolic pathway necessary for eventual chorion melanization in Ae. aegypti , and other melanization events. The gene is up-regulated in response to blood meal with transcripts initially detectable by 12 hours PBM, and maximal levels achieved between 24 and 48 hours PBM [ 104 ]. However, we did identify a gene encoding another enzyme involved in tyrosine metabolism. AS 1340, a Middle gene product, shared sequence similarity with the Ae . aegypti Dopachrome conversion enzyme [ 105 ]. This enzyme is required for processing of dopachrome to melanin. It is interesting that its mRNA is constitutively expressed in Ae. aegypti females, but becomes upregulated when they are infected with Dirofilaria [ 105 ]. Since insect melanins can be produced via any of three intermediates, Dopa, Dopamine, or Dopachrome, it may be that An. gambiae differs from Ae. aegypti in the substrate metabolized to produce chorionic melanin. Several Middle and Late genes encoding antioxidants were upregulated at least twofold 12–48 hours PBM. The Middle gene, AS 2033, a glutathione S-transferase D3, and the three Late genes, ASs 1684, 35, and 2156, encoding glutathione S-transferase 1–6 class theta, and homologs of Drosophila thioredoxin and Ae. aegypti 2-Cys thioredoxin peroxidase, may have roles in regulating reactive oxygen species that can be produced from the highly reactive quinones which are normally cross-linked into melanin immediately after they are formed. Ovarian cycle and vitellogenesis Paramount to the development of the embryo is the massive accumulation of vitellogenin by the oocyte. In An. gambiae there is a small, polymorphic tandem array of vitellogenin genes and a single dispersed vitellogenin gene, all located on Chromosome 2R in division 18B (P. Romans and M. Sharakhova, unpublished observations). Vitellogenin mRNA becomes detectable by Northern blot analysis by 8 hrs PBM, though it is detectable earlier by RT-PCR, increases dramatically by 12 hours, reaches maximal levels by 24 hours, and declines to undetectable levels by 48 hours PBM [ 47 ]. Our microarray study identified three cDNAs, all Middle gene products and greater than twofold induced, as vitellogenin gene homologs. Two of the ESTs were not conjoined during EST assembly because they represented non-overlapping 5' and 3' ends of the Vg1 gene. The third EST included the more closely resembled the sequence of the dispersed vitellogenin gene (P. Romans and A. Dana, unpublished). As expected, all three vitellogenin clones exhibited expression profiles similar to the overall pattern previously described [ 47 ]. Following synthesis in the fat body, vitellogenins are released into the hemolymph. Eventually, they diffuse through channels between the cells of the follicular epithelium and are accumulated by the oocyte by receptor-mediated endocytosis in clathrin-coated pits [ 8 ]. The increased number of gene products involved in receptor-mediated endocytosis before and after the height of vitellogenin gene transcription, 12–24 hours PBM in this study, may reflect a preparation for the increase in receptor-mediated endocytosis when the oocytes are accumulating vitellogenins and other yolk constituents during the trophic phase of the ovarian cycle. These genes included an Early gene, AS 1974, similar to the Drosophila Adaptin subunit, AP-1σ , CG5864, and the Late gene, AS 2086, homolog of another Drosophila clathrin-associated protein, AP-50 , CG7057. When vitellogenesis has ceased, the biosynthetic machinery in the fat body is degraded in lysosomes [ 46 ]. In Ae. aegypti , the lysosomal cathepsin D-like aspartic protease (AeLAP) exhibited a similar transcription profile to vitellogenin [ 106 ]. Cho et al . [ 107 ] also identified a Cathepsin B-like thiol protease, vitellogenic Cathepsin B or VCB, which is secreted from the fat body with a peak at 24 hours PBM and incorporated into oocytes. It appears to be involved in the degradation of vitellin in embryos. The Middle gene AS 996, a Cathepsin B, shares identity with this Ae. aegypti protein, exhibits the same expression profile, and may be its homolog. At approximately 30 hours PBM, 6 hours after peak production of vitellogenin, the activity of four other lysosomal enzymes, arylsulfatase A, acid phosphatase-1, β-galactosidase, and Cathepsin D, has dramatically increased to reach maximal levels by 36–42 hours PBM [ 108 , 109 ]. The two Late gene products, ASs 1254 and 2231, were identified putatively as the lysosomal enzymes, acid phosphatase-1 and Cathepsin F, respectively. These genes also may be involved in the termination phase of vitellogenesis. Cathepsin F is necessary for oocyte growth in a teleost fish and has been suggested to be associated with yolk protein processing [ 110 ]. It will be a very interesting example of gene co-evolution, should processing of vitellogenins, proteins conserved between egg-laying vertebrates and non-Brachyceran insects, actually be accomplished by similarly conserved cathepsins. Conclusions Holt et al . [ 49 ] performed the first genomic-scale study of hematophagy in An. gambiae by identifying 168 ESTs that differed in statistical abundance between cDNA libraries made from adult female mosquitoes fed on sugar and 24 hours PBM. Ribeiro [ 111 ] extended this study by describing an additional 267 such genes. We have expanded on these studies by identifying additional 359 ESTs and by examining virtually a complete first gonotrophic cycle experimentally. In addition, we found 18 ESTs present in our microarray and Ribeiro's [ 111 ] studies. All but one (AS 205) showed the same expression patterns at 24 hours PBM (Table 8 ). These highly synchronous expression profiles of those ESTs further validate that experimental microarray and in silico data can complement each other. However, our study is unique in that we have determined the temporal patterns of expression of the genes we identified. The observed similarities between the gene expression patterns and production of the two principal insect hormones suggest that gene transcription may be influenced by changes in JH titers as well as by 20-E levels, a phenomenon that has been well-studied in the context of Drosophila metamorphosis and in Ae. aegypti vitellogenesis. Future analysis may reveal genes co-regulated via the same promoters. Indeed, this now appears possible for organisms whose genomes have been sequenced [ 112 - 114 ]. As new regulatory sequences are identified, the arsenal of transcriptional regulators to drive their tissue- and stage-specific gene expression will be increased. We expect that this increased promoter availability will supplement current vector-control strategies. Table 8 List of genes differentially expressed* in female A. gambiae at 24 hours post-blood meal in both microarray and in silico (Ribeiro, 2003) gene expression studies. AS ID Ensembl ID Microarray in silico Molecular Functions Biological Processes 99 agCP3123 Up Up enzyme Protein Digestion 996 agCP14019 Up Up enzyme Egg Development 2222 agCP8969 Up Up unknown Unknown 1949 agCP12846 Up Up unknown Unknown 1317 agCP8818 Up Up transporter Transport 516 agCP3409 Up Up binding Structural 1044 agCP3927 Up Up unknown Egg Development 995 agCP5701 Up Up enzyme Protein Digestion 230 agCP2518 Up Up nutrient reservoir Egg Development 180 agCP1111 Up Up unknown Unknown 2207 agCP15442 Up Up transporter Ion Transport 2256 agCP2731 Up Up unknown Unknown 2243 agCP3610 Up Up enzyme Protein Digestion 205 agCP5849** Down Up unknown Unknown 86 agCP6049 Down Down unknown Unknown 553 agCP11425 Down Down transporter Oxidative phosphorylation 2123 agCP11416 Down Down transporter Transport 642 agCP8191 Down Down structural molecule Cuticle biosynthesis *; These genes displayed at least 2-fold up- or down-regulation relative to the control. **: AS 205 shows discrepancy between microarray and in silico expression data. Up: up-regulation; Down: down-regulation. Great progress has been made in the annotation of the An. gambiae genome, culminating in the public announcement of the genome sequence in 2002 and its subsequent updates. Yet, although we have identified 413 differentially expressed gene products, we could not assign almost half of them to a biological process. Of the 200 "Unknowns," 43 unique transcripts shared no significant identity with sequences in the Nr and dbEST databases. The genes corresponding to these transcripts may be identified following the second gene build of the An. gambiae genome. Functional studies using microarray analysis verified by qRT-PCR must confirm in silico predicted annotations and provide biological information about gene products. Many of the gene products identified in this study share sequence similarity with Drosophila proteins. Much of the information generated by studies of fruit fly cell biology and development may also apply to mosquitoes, although it will be more difficult to test in An. gambiae , since it is not easily manipulated genetically. This study underscores the importance of ongoing functional studies including tissue-specific expression profiling using microarray analysis and qRT-PCR. Understanding how the events following blood feeding are related to each other on a molecular level will provide a more comprehensive picture of this unique behavior and may also delineate new vector-control strategies. Methods Microarray chip fabrication Three cDNA libraries were constructed from abdomens of adult female An. gambiae which had been sugar-fed (harvested 30 hours post-eclosion), rat blood-fed (harvested 30 hours PBM), and P. berghei -infected rat blood-fed (harvested 30 hours PBM), all at 19°C (Dana, unpublished PhD thesis). Clones from all three libraries were subjected to PCR-based insert amplification using λTriplEx2 vector specific primers (3' LD Amplimer Primer 5'-ATACGACTCACTATAGGGCGAATTGGC-3'; 5' LD Amplimer Primer: 5'-CTCGGGAAGCGCGCCATTGTGTTGG-3'). Amplification reactions contained 1.0 μL eluted phage, 0.03 pmol of each primer, 1 × Taq Polymerase Buffer (Invitrogen), 3 mM MgCl 2 , 1 mM of each dNTP, and 0.2 U Taq Polymerase (Invitrogen), in a total volume of 100 μL. Reactions were conducted in 96-well plates on a Perkin-Elmer 9700 Thermocycler using the following cycling conditions: initial denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing/elongation at 70°C for 2 min, and a final elongation step at 68°C for 3 min. Samples of all PCR products were electrophoresed on 1% agarose, 1 × TBE gels and visualized by ethidium bromide staining. PCR products were purified on a Beckman Biomek FX using Montage PCR 96 Cleanup kits (Millipore), eluted in 100 μL of water, evaporated overnight and the pellets resuspended in 30 μL of 3 × SSC microarray spotting buffer. A total of 3060 resuspended cDNA inserts and 108 controls were spotted in triplicate on CMT-Gaps II slides (Corning, NY) using the Affymetrix Arrayer 417 at 19 – 20°C and relative humidity between 50 – 60%. Slides were post-processed by baking at 80°C for three hours, incubation in 1% SDS for 2 min, in 95°C purified water for a further 2 min, and then plunged 20 times into 100% ethanol kept at -20°C and air-dried via centrifugation at 500 RPM for 5 min. Microarray target preparation and hybridization Total RNA was extracted from blood-fed and sugar-fed whole adult female mosquitoes of the malaria susceptible 4Arr strain, 5–7 days post eclosion, using Trizol (Molecular Research Center, Inc) according to the manufacturer's directions. Mosquitoes were blood-fed on anesthetized white rats and maintained under conditions similar to those for sugar-fed mosquitoes, 25°C with 80% humidity and a 12-h light/dark cycle with available 20% sucrose solution, until collection. Fully-engorged mosquitoes were identified visually and harvested at the following 10 time-points: 1) 5 min after initiation of blood feeding during the acquisition of the blood meal (DBM), 2) 30 min DBM, 3) 0 hr post-blood meal (PBM), immediately after they ceased feeding on the rat, 4) 1 hr PBM, 5) 3 hr PBM, 6) 5 hr PBM, 7) 12 hr PBM, 8) 16 hr PBM, 9) 24 hr PBM, and l0) 48 hr PBM. For each blood-fed sample, total RNA was extracted from batches of approximately 10–15 females. Total RNA was also extracted from batches of 100 sugar-fed females for reference samples. RNA samples were then treated with 1.0 μL DNase I (Life Science Technology) according to manufacturer's instructions. Following DNase I treatment, total RNA was re-extracted with Trizol. First strand cDNA synthesis and labeling with Cyanine 3 (Cy3) or Cyanine 5 (Cy5), were performed on 15 μg of total RNA from each sample using the Genisphere 3DNA Array 50 kit according to the manufacturer's protocol. Hybridizations were conducted following the two step protocol recommended by the manufacturer: 1) cDNA hybridization to the amplified cDNA probes spotted on the slides, 2) hybridization of 3-DNA fluorescent dendrimers (Genisphere) to cDNAs via the capture sequences incorporated into them during first strand synthesis. All cDNA and fluorescent dye hybridizations were performed in a volume of 50 μL using the formamide-based hybridization buffer provided by the manufacturer. The cDNA hybridizations were performed at 45°C overnight. The slides were then washed according to the 3DNA Array 50 kit protocol and air dried by centrifugation for 3 min at 800 RPM. The 3-DNA hybridizations were performed at 53°C for 2 hours as described above, except that 0.5 mM DTT was added to the first two wash solutions to protect the fluorochromes from oxidation. Five replicate slides were generated for each of the ten time points for a total of 50 hybridized and labeled slides. These included two dye-swap experiments performed to eliminate dye fluorescence bias. Pilot experiments conducted with total RNA from the same sample labeled with both Cy3 and Cy5, self-self hybridization, indicated that there was no dye labeling bias following data normalization (data not shown). Microarray data acquisition and statistical analysis Following hybridization and washing, microarray slides were scanned successively at 532 and 635 nm using the Affymetrix 428 Array Scanner. Raw signal intensities were acquired using the adaptive circle algorithm and spot intensities quantified using the Jaguar 2.0 segmentation and data analysis software (Affymetrix, CA). Average signal intensities were normalized using the Loess curve for intensity dependent normalization followed by a per gene median normalization using the Genespring 5.1 software (Silicon Genetics, CA). Signal intensities were filtered such that only gene products exhibiting a raw signal intensity value greater than 300 pixels in one channel and greater than two-fold expression difference between the sugar-fed and blood-fed samples from at least one time point hybridized to the same array were utilized in further analysis. Gene expression level measurements falling outside one standard deviation from the mean signal intensity of each gene product calculated from the five replicates were excluded from further analysis. As an additional quality control, only genes whose PCR amplified products migrated as a single band in agarose gel electrophoresis and that generated high quality sequences for use in EST assembly were analyzed. Gene products that were induced or repressed at least twofold during blood feeding were initially clustered hierarchically using the Genespring software to determine the user-defined number of centroids (clusters) to be used in k -means clustering (data not shown). From this preliminary analysis it was determined that three major clades existed and the genes were clustered using Genespring software using a k -means clustering algorithm with a centroid number of 3 and the Pearson Correlation distance metric. Finally, an independent analysis using principal components analysis (PCA) was conducted on the genes induced or repressed at least twofold, using the Genespring software. qRT-PCR Transcript levels of several selected genes were measured using SYBR dye technology (Applied Biosystems, CA) and quantitative real-time PCR (qRT-PCR) analysis in order to validate microarray data,. The Primer Express Software v. 1.5 (Applied Biosystems, CA) was used to design the following primers to nine genes: the two Early Genes agCP4871 (AS 648; Forward 5'-TGATTCGTGCCAGGGTGAT-3'; Reverse 5'-CACCACACCAACAAGGACATC-3') and CG8690 (AS 1786; Forward 5'-GCTGACTTTGAGCGGTTGG-3'; Reverse 5'-CACAAAGTCCATGATCACCTTCA-3'), the four Middle Genes agCP8064 (AS 679; Forward 5'-TGGCGAGGTCGATCAGCTA-3'; Reverse 5'-CATTATCGCCATCGTTGTGTTG-3'), agCP12846 (AS 1949; Forward 5'-TTTGTGGTTCGGTATCGATCTG-3'; Reverse 5'-CGAGCACTTTGGCGAACTTC-3'), CG7758 (AS 1158; Forward 5'-CACGGTTGGCATTTCGAAC-3'; Reverse 5'-GCAGCTGTGCGAACACCA-3'), and agCP14019 (AS 996; Forward 5'-GTCGGGCGATTCCAATGA-3'; Reverse 5'-TGTAACCGGGCTGGCAAA-3'), and the two Late Genes agCP14623 (AS 12; Forward 5'-CGGCAAATCGGTTCAGCT-3'; Reverse 5'-TGAATCGGTGCCTTGCG-3') and agCP2112 (AS 1357; Forward 5'-CCTGCATGAAGGTGGAATGA-3'; Reverse 5'-TTGCCAAGCTCTCCCAACAC-3'), and the ribosomal protein S7 (RP S7) gene control (Forward 5'-CATTCTGCCCAAACCGATG-3'; Reverse 5'-AACGCGGTCTCTTCTGCTTG-3'). RP S7 was used as an internal control since its expression is constitutive during blood-feeding [ 115 - 118 ]. All amplifications and fluorescence quantification were performed using an ABI 7700 Sequence Detection System and associated Sequence Detector Software v. 1.7 (Applied Biosystems, CA). Standard curves were generated using 10-fold serial dilutions of genomic DNA (ranging from 0.0116 to 116 ng per reaction). These qPCR reactions were performed in duplicate in a total volume of 25 μL containing 12.5 μL of SYBR green PCR Master Mix, 300 nmol of each primer, and nuclease free water (Gibco, UltraPURE) using the following conditions; 50°C for 2 min, then denaturation at 95°C for 10 min followed by 45 cycles of denaturation at 95°C for 15 s, annealing and extension at 60°C for 1 min. qRT-PCR reactions for quantification of transcript levels were conducted using 50 ng of first strand cDNA prepared from RNA samples isolated for the microarray analysis. The abundance of each transcript in an RNA sample was estimated from the corresponding gene's standard curve and normalized against RP S7 transcript abundance in the same RNA sample. Authors' contributions AND carried out the cDNA library construction, microarray fabrication, data analysis, and drafted the manuscript. YSH performed the microarray experiment, data analysis, and helped draft the manuscript. MKK performed the qRT-PCR experiments and MEH annotated the ESTs. BWH constructed the microarray genechips and carried out microarray data acquisition. NFL sequenced the cDNA library and JRH maintained and provided mosquito samples throughout the project. PR assisted AND and YSH to draft the manuscript and reviewed it. FHC (P.I.) initiated and supervised the project. All authors read and agreed on the final version of this manuscript. Supplementary Material Additional File 1 Supplementary Table S1 in a Microsoft Excel format where gene annotations for all 413 at least twofold differentially expressed gene products are given. Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546002.xml
544946
Autoantibodies and autoantigens in autoimmune hepatitis: important tools in clinical practice and to study pathogenesis of the disease
Autoimmune hepatitis (AIH) is a chronic necroinflammatory disease of the liver characterized by hypergammaglobulinemia, characteristic autoantibodies, association with HLA DR3 or DR4 and a favorable response to immunosuppressive treatment. The etiology is unknown. The detection of non-organ and liver-related autoantibodies remains the hallmark for the diagnosis of the disease in the absence of viral, metabolic, genetic, and toxic etiology of chronic hepatitis or hepatic injury. The current classification of AIH and the several autoantibodies/target-autoantigens found in this disease are reported. Current aspects on the significance of these markers in the differential diagnosis and the study of pathogenesis of AIH are also stated. AIH is subdivided into two major types; AIH type 1 (AIH-1) and type 2 (AIH-2). AIH-1 is characterized by the detection of smooth muscle autoantibodies (SMA) and/or antinuclear antibodies (ANA). Determination of antineutrophil cytoplasmic autoantibodies (ANCA), antibodies against the asialoglycoprotein receptor (anti-ASGP-R) and antibodies against to soluble liver antigens or liver-pancreas (anti-SLA/LP) may be useful for the identification of patients who are seronegative for ANA/SMA. AIH-2 is characterized by the presence of specific autoantibodies against liver and kidney microsomal antigens (anti-LKM type 1 or infrequently anti-LKM type 3) and/or autoantibodies against liver cytosol 1 antigen (anti-LC1). Anti-LKM-1 and anti-LKM-3 autoantibodies are also detected in some patients with chronic hepatitis C (HCV) and chronic hepatitis D (HDV). Cytochrome P450 2D6 (CYP2D6) has been documented as the major target-autoantigen of anti-LKM-1 autoantibodies in both AIH-2 and HCV infection. Recent convincing data demonstrated the expression of CYP2D6 on the surface of hepatocytes suggesting a pathogenetic role of anti-LKM-1 autoantibodies for the liver damage. Family 1 of UDP-glycuronosyltransferases has been identified as the target-autoantigen of anti-LKM-3. For these reasons the distinction between AIH and chronic viral hepatitis (especially of HCV) is of particular importance. Recently, the molecular target of anti-SLA/LP and anti-LC1 autoantibodies were identified as a 50 kDa UGA-suppressor tRNA-associated protein and a liver specific enzyme, the formiminotransferase cyclodeaminase, respectively. Anti-ASGP-R and anti-LC1 autoantibodies appear to correlate closely with disease severity and response to treatment suggesting a pathogenetic role of these autoantibodies for the hepatocellular injury. In general however, autoantibodies should not be used to monitor treatment, predict AIH activity or outcome. Finally, the current aspects on a specific form of AIH that may develop in some patients with a rare genetic syndrome, the autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED) are also given. Autoantibodies against liver microsomes (anti-LM) are the specific autoantibodies detected in AIH as a disease component of APECED but also in cases of dihydralazine-induced hepatitis. Cytochrome P450 1A2 has been identified as the target-autoantigen of anti-LM autoantibodies in both APECED-related AIH and dihydralazine-induced hepatitis. The latter may indicate that similar autoimmune pathogenetic mechanisms can lead to liver injury in susceptible individuals irrespective of the primary defect. Characterization of the autoantigen-autoantibody repertoire continues to be an attractive and important tool to get access to the correct diagnosis and to gain insight into the as yet unresolved mystery of how hepatic tolerance is given up and AIH ensues.
1. Introduction Autoimmune hepatitis (AIH) is a rare chronic liver disease of unknown etiology. The estimated prevalence of AIH in Northern European countries is approximately 160–170 patients/10 6 inhabitants [ 1 , 2 ]. The disease predominates among women and is characterized by hypergammaglobulinemia even in the absence of cirrhosis, characteristic autoantibodies, association with human leukocyte antigens (HLA) DR3 or DR4 and a favorable response to immunosuppressive treatment [ 3 - 5 ]. The onset of AIH disease is usually insidious, with unspecific symptoms, such as, fatigue, malaise, arthralgias, and fluctuating jaundice, right upper quadrant pain or lethargy [ 5 - 8 ]. However, a substantial proportion of patients may have no obvious signs or symptoms of liver disease, while occasionally the presentation may be severe and almost identical to an acute or fulminant episode of viral hepatitis [ 5 - 8 ]. Although AIH brings in mind the archetypal patient being a young female with endocrine abnormalities, there is nowadays increasing evidence that the disease can also affect males and can present at almost any age (the large majority of patients being between 50 and 70 years of age) [ 6 - 12 ]. Liver histology is not pathognomonic for AIH and there is no single serologic test of sufficient specificity for the diagnosis of AIH as for the diagnosis of viral hepatitis A to E. Although the presence of autoantibodies is one of the distinguishing features of AIH, there is no single autoantibody with the diagnostic significance and specificity that antimitochondrial autoantibodies (AMA) demonstrate for the diagnosis of primary biliary cirrhosis (PBC). For this reason, autoantibodies can not be employed as a single marker for the diagnosis of AIH. It is rather a diagnosis reached by the exclusion of other factors leading to chronic hepatitis that include viral, toxic, genetic and metabolic causes [ 6 ]. Under this context, it is clear that sometimes AIH may be difficult to diagnose [ 7 , 8 ]. In 1992, the International Autoimmune Hepatitis Group reported a descriptive set of criteria that could be applied in the routine clinical practice for the diagnosis and classification of patients as having either 'definite' or 'probable' AIH [ 13 ]. In addition, a diagnostic scoring system was devised to provide an objective method for selection of relatively homogeneous groups of patients for research purposes [ 13 ]. The same group has remarkably simplified the descriptive set of criteria and the diagnostic scoring system in late 1998 (Tables 1 and 2 ) [ 6 ]. The diagnostic score demonstrates that the presence of defined autoantibodies is an integral part of the diagnosis of AIH but not its single diagnostic tool [ 6 , 14 ]. Table 1 Revised Scoring System for the Diagnosis of Autoimmune Hepatitis 6 . Parameter/Features Score Gender - Female/Male +2/0 Degree of elevation above upper normal limit of ALP vs. aminotransferases - <1.5 +2 - 1.5 – 3.0 0 - >3.0 -2 Total serum globulins, γ-globulins, or IgG above normal - >2.0 +3 - 1.5 – 2.0 +2 - 1.0 – 1.5 +1 - <1.0 0 ANA, SMA or LKM-1 (titers by immunofluorescence on rodent tissues or HEp2-cells) - >1 : 80 +3 - 1 : 80 +2 - 1 : 40 +1 - <1 : 40 0 - AMA positive -4 Hepatitis viral markers (IgM anti-HAV, HBsAg, IgM anti-HBc, anti-HCV and HCV-RNA) - Positive/Negative -3/+3 Recent or current use of known or suspected hepatotoxic drugs - Yes/No -4/+1 Average alcohol intake - <25 g/day +2 - >60 g/day -2 Other autoimmune disease(s) in patient or first degree relatives - Yes/No +2/0 Optional additional parameters (should be allocated only if ANA, SMA or LKM-1 are negative) - HLA DR3, DR4, or other HLA with published association with AIH) +1 - Seropositivity for any of ANCA, anti-LC1, anti-SLA/LP, anti-ASGPR and anti-sulfatide +2 Liver histology - Interface hepatitis +3 - Predominant lymphoplasmacytic infiltrate +1 - Rosetting of liver cells +1 - None of the above -5 - Biliary changes -3 - Other changes -3 Response to therapy (as defined in Table 2) - Complete/Relapse +2/+3 Definite AIH if greater than 15 before treatment or greater than 17 post-treatment; probable AIH if varies between 10–15 before treatment or 12–17 post-treatment. ALP: alkaline phospatase, IgM anti-HAV: hepatitis A virus IgM antibody, anti-HCV: hepatitis C virus antibody, HBsAg: surface antigen of hepatitis B virus, IgM anti-HBc: IgM antibody against the nuclear antigen of hepatitis B virus. Other abbreviations are the same as shown in the text. Table 2 Definitions of Response to Therapy. Response Definition Complete Either or both of the following: marked improvement of symptoms and return of serum AST or ALT, bilirubin and immunoglobulin values completely to normal within 1 year and sustained for at least a further 6 months on maintenance therapy, or liver biopsy specimen at some time during this period showing at most minimal activity. or Either or both of the following: marked improvement of symptoms together with at least 50% improvement of all liver tests during the first month of treatment, with AST or ALT levels continuing to fall to less than twice the upper normal limit within 6 months during any reductions toward maintenance therapy, or a liver biopsy within 1 year showing only minimal activity. Relapse Either or both of the following: an increased in serum AST or ALT levels of greater than twice the upper normal limit or a liver biopsy showing active disease, with or without reappearance of symptoms, after a "complete" response as defined above. or Reappearance of symptoms of sufficient severity to require increased (or reintroduction of) immunosuppression, accompanied by any increase in serum AST or ALT levels, after a "complete" response as defined above. AST: aspartate aminotransferase, ALT: alanine aminotransferase. In recent years however, significant progress has been made in the characterization of liver-related target-autoantigens. This has led to the notion that some of the major target-autoantigens in AIH are active enzymes of the human hepatic and non-hepatic microsomal xenobiotic metabolism [ 14 - 16 ]. The latter serve as a means to investigate this still enigmatic liver disease. This article will focus on the data that have evolved in the course of the characterization of autoantibody-autoantigen "system" in AIH by giving the current aspects on the role and significance of this "system" in the differential diagnosis and study of pathogenesis of AIH. 2. Classification of AIH According to the pattern of autoantibodies detected in AIH patients, a subclassification of the disease into three types was proposed in 1994 [ 17 ]. AIH type 1 (AIH-1) is characterized by the presence of antinuclear antibodies (ANA) and/or smooth muscle autoantibodies (SMA) which may associate with perinuclear anti-neutrophil cytoplasmic antibodies (p-ANCA) [ 3 , 5 , 6 , 14 , 15 ]. AIH type 2 (AIH-2) is characterized by the detection of specific autoantibodies against liver and kidney microsomal antigens (anti-LKM type 1 or infrequently type 3) [ 14 - 16 , 18 ] and/or antibodies against liver cytosol type 1 antigen (anti-LC1) [ 14 , 15 , 19 ]. AIH type 3 (AIH-3) is characterized by autoantibodies against soluble liver antigens (anti-SLA) [ 20 ] or to liver-pancreas antigen (anti-LP) [ 21 , 22 ]. The serological diversity of autoantibodies found in AIH supports the aforementioned subclassification and provides a framework for the scientific analysis of this heterogeneous disease group [ 5 , 15 ]. It also demonstrates that AIH may not be a single disease with a single underlying mechanism but most likely is a group of diseases with a similar clinical presentation [ 14 , 15 ]. This is further substantiated by the finding of an unusual form of AIH in 10–18% of patients with a rare autosomal recessive disorder, the autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED) [ 23 - 25 ]. This syndrome is characterized by chronic mucocuteneous candidiasis, ectodermal dystrophy and autoimmune tissue destruction particularly of the endocrine glands (hypoparathyroidism, adrenocortical failure and gonadal failure in females) [ 26 - 29 ]. However, due to recent clinical, serologic and genetic findings, it has been suggested that anti-SLA seropositive patients do not define a subgroup of AIH, but rather belong to the AIH-1 group [ 30 - 32 ]. For this reason, subdivision into AIH-1 (ANA, SMA, p-ANCA and/or anti-SLA/LP positive) and AIH-2 (anti-LKM-1, anti-LKM-3 and/or anti-LC1 positive) is in common usage (Table 3 ). Apart from serological differences, AIH-2 seems to be clinically and genetically distinguishable from AIH-1 [ 5 , 8 , 14 ]. Indeed, patients with AIH-2 are younger at presentation, usually have higher levels of bilirubin and transaminases, and are characterized by more severe disease than patients with AIH-1 [ 5 , 6 , 14 , 18 , 33 ]. In addition, contrary to what has been recorded in patients with AIH-1, no sustained remission has been observed after the discontinuation of immunosuppressive therapy in patients with AIH-2 [ 4 , 33 ]. Taking into consideration the genetic markers, it has been found that the association between HLA DR3 and AIH-2 is rather weaker than that reported in AIH-1, while an association between HLA DQ2 and AIH-2 has been reported [ 14 , 18 , 33 , 34 ]. However, AIH-2 patients represent only a small proportion of the total cases of AIH [ 3 - 6 ]. In addition, the long-term outcome of the affected patients appears to be similar both in AIH-1 and AIH-2 [ 6 , 33 ]. Therefore, the classification of AIH in these two major subgroups is still uncertain and controversial [ 6 , 8 , 13 ]. Table 3 Classification of Autoimmune Hepatitis (AIH) According to Autoantibodies Detection Type of AIH Characteristic autoantibodies AIH-1 ANA, SMA, p-ANCA anti-ASGP-R, anti-SLA/LP AIH-2 anti-LKM-1, anti-LKM-3, anti-LC1, anti-ASGP-R Abbreviations are the same as shown in the text. 3. Detectable Autoantibodies in AIH-1 3.1. Anti-nuclear antibodies (ANA) and smooth muscle autontibodies (SMA) ANA and/or SMA are almost exclusively requisites for the diagnosis of AIH-1 [ 3 , 5 , 6 , 14 , 15 , 30 ]. In typical cases of AIH-1, these autoantibodies are detected in significant titers (≥1;80 in adults and ≥1:40 in children) in almost half of Caucasians patients with AIH-1, while ANA alone are detected in 15% and SMA alone in 35% [ 5 , 30 , 35 ]. The most frequent and conventional method for ANA detection is the indirect immunofluorescence (IIF) assay on cryostat sections of rodent tissues and HEp-2 cells slides (Figures 1 and 2 ) [ 14 , 15 , 35 , 36 ]. Different patterns of fluorescence are found by this assay due to the large variability of target-autoantigens in the nuclei of HEp-2 cells that have been recognized [ 35 - 40 ]. Actually, ANA are found to be directed against single or double stranded DNA, tRNA, SSA-Ro, snRNPs, laminins A and C, cyclin A or histones [ 35 - 40 ]. Most commonly, a homogenous (34–58%) or speckled (21–34%) immunofluoerescence pattern is demonstrable [ 14 , 15 , 35 - 37 ]. So far however, neither a liver-specific nuclear antigen nor a liver-disease-specific ANA has been identified. For this reason, subtyping of ANA in cases of AIH-1 seems to have limited clinical implication and diagnostic relevance in routine clinical practice [ 6 , 14 , 15 , 35 - 40 ]. Figure 1 High titer antinuclear antibodies (ANA) of the homogeneous pattern by indirect immunofluoerescence on immobilized HEp-2 cells in a female patient with autoimmune hepatitis type 1 (AIH-1). Homogeneous ANA are frequently found in AIH-1 (original magnification 40×). Figure 2 Typical staining of antinuclear antibodies in the serum of a patient with autoimmune hepatitis type 1 visualized by indirect immunofluoerescence on cryostat sections of rat liver (original magnification 40×). SMA are detected by IIF on rodent liver and kidney, due to staining of vessel walls (Fig. 3 ), and stomach due to staining of the muscle layer (Fig. 4 ). SMA are directed against structures of the cytoskeleton such as actin, troponin, vimentin and tropomyosin. In AIH-1, SMA are predominantly directed against F-actin [ 41 ]. The latter seems to be diagnostically more relevant in pediatric patients where SMA may be the only marker of AIH-1 even in titers as low as of 1:40 [ 14 , 15 ]. Czaja et al [ 41 ] have suggested that antibodies to actin are associated with a younger age of disease onset, the presence of HLA-A1-B8-DR3 haplotype and a greater frequency of treatment failure, death from liver disease and earlier requirement for transplantation than actin-antibody negative AIH-1 patients. Figure 3 Smooth muscle antibodies by indirect immunofluoerescence on rat kidney (from a female patient with autoimmune hepatitis type 1). The immunofluorescence involves smooth muscle fibers within blood vessels (original magnification 40×). Figure 4 Smooth muscle autoantibodies by indirect immunofluoerescence on rat stomach (serum from a female patient with autoimmune hepatitis type 1). The immunofluorescence involves smooth muscle fibers within muscularis mucosa (original magnification 40×). ANA and/or SMA – usually in low titers – may occur in patients with chronic viral hepatitis B or C, but in most of these cases SMA lack F-actin specificity [ 14 , 15 , 42 , 43 ]. From the clinical point of view, interferon-alpha administration is generally safe in most cases of viral hepatitis with ANA and/or SMA, although occasionally may provoke mild self-limited autoimmune disorders compared to viral hepatitis patients without ANA or SMA autoantibodies [ 44 - 46 ]. During immunosuppressive treatment, disappearance of ANA and/or SMA is observed in the majority of patients with AIH-1 [ 47 ]. However, autoantibody status is unable to predict immediate outcome after cessation of corticosteroid administration. Additionally, neither autoantibody titers at first diagnosis nor autoantibody behaviour in the time course of the disease are prognostic markers for AIH-1 [ 14 , 15 , 47 ]. These findings indicate that ANA and SMA are not involved in the pathogenesis of AIH-1 and furthermore, their determination is more of diagnostic than prognostic value [ 5 , 14 , 15 , 47 ]. 3.2. Anti-neutrophil cytoplasmic autoantibodies (ANCA) These autoantibodies are directed against cytoplasmic constituents of neutrophil granulocytes and monocytes. Classically, they are detected by IIF using ethanol-fixed granulocytes as substrate [ 48 ]. Using the above method, two major subtypes can be distinguished. ANCA showing a diffuse or granular cytoplasmic staining (c-ANCA) and ANCA characterized by a perinuclear-staining (p-ANCA). Both c-ANCA and p-ANCA are valuable diagnostic and prognostic markers in systemic vasculitides in particular Wegener's granulomatosis and microscopic polyangiitis, respectively [ 48 , 49 ]. Proteinase-3 has been identified as the major target-autoantigen of c-ANCA in cases with Wegener's granulomatosis, while myeloperoxidase is the documented autoantigen of p-ANCA in most patients with microscopic polyangiitis [ 48 , 49 ]. Since then, ANCA (in most cases of p-ANCA type) were detected in a high prevalence in other inflammatory disorders of unknown aetiology such as, inflammatory bowel disease (more frequently in ulcerative colitis than in Crohn's disease) [ 50 , 51 ] and primary sclerosing cholangitis (PSC), a liver disease that is frequently associated with ulcerative colitis [ 51 , 52 ]. Several recent studies however, have also documented the presence of high titers of p-ANCA in the sera of patients with AIH-1 (Fig. 5 ; prevalence range 40–96%) [ 53 - 58 ] and to a much lesser extent in PBC patients (prevalence range 0–39%) [ 54 , 58 , 59 ]. Occasionally, high titers of c-ANCA can be detected in AIH-1 (Dalekos GN, 2002, unpublished observations). In contrast, p-ANCA have not been detected in serum samples from patients with AIH-2 [ 57 ]. Low ANCA titers are detected infrequently in patients with alcoholic or chronic viral liver diseases [ 54 , 57 , 60 ]. However, a recent large study in 516 patients with hepatitis C virus (HCV) infection revealed the presence of ANCA in as high as 55.6% of patients [ 61 ]. Interestingly these investigators have shown that all HCV positive sera with ANCA had c-ANCA pattern on IIF and contained proteinase-3 specificity [ 61 ]. The clinical relevance of this finding remains to be determined. In patients with AIH-1, PBC or PSC the detection of ANCA appears to be associated with a more severe disease course or the presence of cirrhosis [ 54 , 62 ]. The latter suggestion however, was not confirmed by more recent studies [ 53 , 54 ]. Figure 5 Perinuclear staining of anti-neutrophil cytoplasmic autoantibodies (p-ANCA) by indirect immunofluoerescence on ethanol fixed human granulocytes (serum from an ANA negative patient with autoimmune hepatitis type 1). Original magnification 40×). To determine the antigenic specificities of ANCA, antigen-specific enzyme linked immunosorbent assays (ELISAs) and Western blotting followed by immunodetection can be performed [ 14 , 63 ]. Using these techniques it became obvious that in AIH-1 the target-autoantigens recognized are multiple including cathepsin G, catalase, alpha-enolase, lactoferrin, actin and high mobility group (HMG) non-histone chromosomal proteins HMG1 and HMG2 [ 14 , 54 , 56 , 58 , 62 - 65 ]. However, the determination of antigenic specificities of ANCA seems to have limited clinical relevance in patients with AIH-1 [ 14 , 15 , 54 , 56 ]. In conclusion, the detection of ANCA may be a useful additional marker in searching for AIH-1, in particular in ANA/SMA/anti-LKM-1 negative cases of AIH. With the exception of a recent paper by Wu et al [ 61 ] the detection of ANCA is rather rare in chronic viral hepatitis [ 14 , 54 , 57 , 60 ]. The latter may prove useful in the differential diagnosis between patients with AIH and those with viral hepatitis who tested positive for ANA or SMA. Furthermore, since ANCA appear to be relatively rare in PBC [ 54 , 59 ], these autoantibodies may prove useful for distinguishing between genuine cases of AIH and cases of PBC with features overlapping with those of AIH [ 6 ]. However, due to the lack of specificity for the diagnosis of AIH and to its obscure role – if any – in AIH, their routine determination is not recommended [ 14 , 15 ]. 3.3. Autoantibodies against the asialoglycoprotein receptor (anti-ASGP-R) The asialoglycoprotein receptor (ASGP-R) is a liver-specific glycoprotein of the cell membrane. Its main function is the internalization of asialoglycoproteins by binding a terminal galactose residue to coated pits. Anti-ASGP-R autoantibodies are detected in 88% of patients with AIH (both types) [ 66 , 67 ]. However, these autoantibodies are also found in some patients with PBC, chronic viral hepatitis B and C and alcoholic liver disease although at lower frequency and lower titers [ 14 , 15 , 66 , 67 ]. The ASGP-R is preferentially expressed on the surface of periportal liver cells where piecemeal necrosis is found as a marker of severe inflammatory activity in patients with AIH [ 68 ]. This finding may suggest a possible immunopathogenetic involvement of anti-ASGP-R autoantibodies in AIH [ 69 ]. The general presumption is that target of potentially tissue-damaging autoreactions in AIH must be liver-specific and available to the immune system in vivo (e.g. expression on the surface of hepatocytes). So far, ASGP-R is the only target-autoantigen that has been positively identified and fulfils these criteria [ 68 , 69 ]. Additional support to this emerged from the determinations of anti-ASGPR autoantibodies in consecutive AIH patients. The levels of anti-ASGP-R autoantibodies vary according to the inflammatory activity of the disease. In addition, anti-ASGP-R antibody titers decreased significantly in response to immunosuppression, while they reappear when the disease has relapsed [ 66 , 70 ]. These autoantibodies may be diagnostically helpful when other autoantibodies are not detected and AIH is suspected. However, due to the belief that anti-ASGPR antibody represents a general marker of liver autoimmunity and the limitations in its detection (requires chemically purified ASGP-R, which is not yet widely available), its routine use is not generally recommended. 3.4. Antibodies against soluble liver antigens (anti-SLA) or to liver-pancreas antigen (anti-LP) The anti-SLA autoantibodies were described for the first time in 1987 [ 20 ]. They cannot be detected by IIF on common substrate. A competitive ELISA or a radioimmunoassay usually detects these autoantibodies [ 20 , 32 , 71 ]. SLA is found in 100000 g supernatant of liver homogenate and represent a cytosolic protein which is neither organ nor species specific [ 72 ]. However, the highest concentrations are found in liver and kidney tissues. The anti-SLA autoantibodies are detected in patients with AIH alone or in combination with SMA and/or ANA [ 30 - 32 , 73 ]. As noted above, similarities in the clinical profile between patients with AIH-1 (ANA and/or SMA positive) and AIH patients with anti-SLA alone in addition with an approximately 30% seropositivity overlap between anti-SLA and SMA and/or ANA suggest that anti-SLA is rather an additional important marker for the diagnosis of AIH-1, than a marker of a third type of AIH [ 6 , 14 , 30 - 32 ]. A scientific group from Tuebingen, Germany described for the first time the anti-LP autoantibodies in 1981 [ 21 ]. The LP antigen was predominantly detected in the S100 supernatant of liver and pancreas homogenates, indicating that this antigen was a soluble protein. Until recently, anti-LP and anti-SLA autoantibodies were thought to be different [ 20 - 22 ]. However, Wies et al report [ 74 ] provides convincing evidence and confirms previous suggestions that anti-SLA and anti-LP are one and the same autoantibody (anti-SLA/LP). In addition, the same study demonstrated that the identified target-autoantigen of anti-SLA/LP autoantibodies (a 35–50 kDa protein) was neither cytokeratins 8 or 18 [ 71 ] nor glutathione-S-transferase isoenzyme [ 75 ]. The results from two independent groups [ 76 , 77 ] were similar with those found by Wies et al [ 74 ]. After screening of cDNA expression libraries they identified a previously unknown amino acid sequence, which presumably encodes a UGA-suppressor tRNA-associated protein, as the targen-autoantigen of anti-SLA/LP autoantibodies [ 76 , 77 ]. The UGA-suppressor serine tRNA-protein complex is likely to be involved in cotranslational selenocysteine incorporation in human cells [ 78 ]. It was then obvious that the identification of SLA/LP autoantigen would allow the establishment of a reliable, universally available diagnostic test for AIH but also it would provoke the investigation in the area of autoimmune liver diseases. Regarding disease specificity, anti-SLA/LP autoantibodies have not been detected in patients with AIH-2, PBC, PSC, chronic viral hepatitis, alcoholic liver disease and non-hepatic autoimmune diseases by standardized ELISAs using reference autoantibody or recombinant antigen [ 20 , 32 , 73 , 79 ]. Ballot et al [ 32 ] also showed that these autoantibodies are different from anti-LC1. For these reasons, anti-SLA/LP has been considered as a valuable and specific diagnostic marker of AIH [ 31 , 32 , 73 , 74 , 76 , 77 , 79 ]. However, a recent study from the United Kingdom [ 80 ] has shown that anti-SLA/LP autoantibodies can also be detected in AIH-2 and in children with PSC. These investigators used eukaryotically expressed tRNP ((Ser) Sec)/SLA as target in a radioligand assay (RLA) which is well known as a more sensitive test than ELISAs and immunoblot due to its ability to identify antibodies directed to conformational epitopes [ 81 - 83 ]. Their novel findings need confirmation from other research groups and particularly to address whether anti-SLA/LP reactivity is also present in adult PSC. Recent data confirmed the previous finding that patients with anti-SLA/LP display a more severe course of AIH [ 79 , 80 , 84 ]. The latter suggest that anti-SLA/LP may be linked to the pathogenesis of the autoimmune process although the exact function and its role in autoimmunity are so far unclear [ 14 , 15 ]. From the clinical point of view however, this autoantibody may be helpful in an attempt to reduce the group of cryptogenic hepatitis and/or cirrhosis. 4. Detectable Autoantibodies in AIH-2 4.1. Autoantibodies against liver-kidney microsomes (anti-LKM) Three types of anti-LKM autoantibodies have been identified [ 3 , 5 , 14 , 15 , 18 , 30 , 63 , 72 , 85 ]. The LKM type 1 autoantibody (anti-LKM-1) is the characteristic serologic marker for the diagnosis of AIH-2 [ 5 , 18 , 63 , 72 ]. These autoantibodies were first described by Rizzetto et al [ 86 ], using the IIF method on rodent liver and kidney sections. The characteristic features of anti-LKM-1 autoantibodies are the diffuse staining of cytoplasm of the entire liver lobule and the exclusive staining of the P3 portion of the proximal renal tubules (Fig. 6 ) [ 18 ]. Due to this staining pattern of kidney sections anti-LKM-1 can be easily distinguished from AMA, which stain proximal and distal renal tubules (Fig. 7 ). Western blots with hepatic and renal microsomes revealed a protein band at 50 kDa [ 5 , 14 , 15 , 63 , 72 , 87 ]. Figure 6 Antibodies against liver-kidney microsomes type 1 (anti-LKM-1) react to the proximal tubules of the rat kidney. The absence of reactivity against thedistal tubules of the rat kidney (see also Fig. 6B) and parietal cells of the rat stomach distinguishes anti-LKM-1 autoantibodies from antimitochondrial antibodies (original magnification 40×). Figure 7 Antimitochondrial antibodies react to the proximal and distal tubules of the rat kidney (original magnification 40 ×). In these cases there is also reactivity to the parietal cells of the rat stomach. The major target-autoantigen of anti-LKM-1 autoantibodies in AIH-2 has been identified as the cytochrome P450 2D6 (CYP2D6) [ 87 - 89 ]. It has been shown that anti-LKM-1 autoantibodies inhibit the enzymatic activity of CYP2D6 in vitro, but not in vivo [ 90 ]. Epitope mapping experiments of CYP2D6 autoantigen have defined at least four different linear epitopes [ 91 , 92 ]. The most immunodominant epitopes of CYP2D6 were amino acids 257–269 and 321–351, which are recognized in about 70% and 50% of AIH-2 cases, respectively [ 91 , 92 ]. Two infrequent epitopes consisting of amino acids 373–389 and 410–429 are also recognized by anti-LKM-1 in some cases [ 92 ]. Recently, Klein et al [ 93 ] and Kerkar et al [ 94 ] reported another immunodominant epitope of CYP2D6 (amino acids 193–212) recognized in about 70% and 93 % of AIH-2 patients, respectively. However, due to failure of inhibition of CYP2D6 enzymatic activity using epitope specific antibodies and since the absorption of the above linear epitopes was unable to absorb inhibitory anti-CYP2D6 autoantibodies, the existence of additional conformational epitopes on CYP2D6 autoantigen has been postulated [ 95 ]. It is noteworthy to state here that depending on the geographic origin, 0–7% of patients with chronic hepatitis C – irrespective of the HCV genotype – develop anti-LKM-1 autoantibodies [ 6 , 14 , 43 , 63 , 96 , 97 ]. Recently, two studies have shown a higher prevalence of anti-LKM autoantibodies (up to 10%) in a small number of children or adult patients with HCV infection [ 98 , 99 ]. As stated for AIH-2, CYP2D6 is the major target autoantigen recognized by anti-LKM-1 autoantibodies in HCV patients [ 14 , 15 , 81 - 83 , 88 , 92 - 96 ]. However, we and others have failed to document CYP2D6 as the major target autoantigen of anti-LKM antibodies in HCV/anti-LKM positive sera [ 98 , 99 ]. In addition, recently Miyakawa et al [ 100 ] identified CYP2E1 and CYP3A4 as target autoantigens of anti-LKM autoantibodies in two patients with anti-LKM-positive chronic hepatitis C. Taking together, these findings may further indicate the heterogeneous autoimmune reactions that might take place in anti-LKM positive patients with chronic hepatitis C. The antigenic sites on CYP2D6 autoantigen recognized by anti-LKM-1 autoantibodies are different in AIH-2 and HCV/anti-LKM-1 positive cases [ 92 - 95 , 101 - 104 ]. For example, the major linear epitope of 257–269 amino acids, as well as the newly reported peptide of 193–212 amino acids are recognized in 70–93% of AIH-2 patients but only in 18–50% of HCV/anti-LKM-1 positive patients [ 83 , 93 , 94 , 101 ]. Additional support to the presence of conformation-dependent anti-LKM-1 autoantibodies in HCV/anti-LKM-1 positive serum samples has emerged from previous studies [ 99 , 102 , 105 ]. In the latter studies only about 30% of HCV/anti-LKM-1 positive sera reacted with 50 kDa component using Western blot assays, while additional bands at 59 kDa, 70 kDa and 80 kDa were detected [ 99 , 102 , 105 ]. However, even taking into account the above additional bands, no more than 45% of all sera tested reactive by Western blot. In contrast, a significant proportion of the previous negative sera tested positive for anti-LKM-1 using a specific competitive ELISA, while denaturation of the antigens prior to perform the ELISA resulted in complete loss of the signal [ 105 ]. Recently, the development of a more sensitive and specific assay for the detection of anti-LKM-1 autoantibodies was achieved [ 14 , 15 ]. This novel assay is a quantitative RLA based on immunoprecipitation using 35 S-methionine-labelled CYP2D6 antigen obtained by in vitro transcription and translation synthesis [ 81 - 83 , 99 , 104 ]. Using this assay it was shown that the anti-LKM-1 titers do not differ significantly between AIH-2 and HCV/anti-LKM-1 positive patients [ 81 - 83 ]. The presence of anti-LKM-1 in some patients with HCV infection led to the proposal for a further division of AIH-2 into AIH-2a (younger, predominantly female patients without evidence of HCV infection) and AIH-2b (older, predominantly male patients with HCV infection) [ 13 , 106 ]. Nowadays however, after the marked improvements in the reliability and availability of tests for HCV detection such a subdivision of AIH-2 appears unreasonable and tends to be deleted. Actually, HCV/anti-LKM-1 positive patients represent cases of "true" HCV infection with autoimmune features [ 6 , 107 ]. From the clinical point of view, screening for anti-LKM autoantibodies is recommended before the initiation of interferon-alpha therapy in HCV patients and if found positive a careful monitoring appears reasonable because occasionally interferon-alpha may unmask, or provoke autoimmune hepatic reactions and even "true"AIH [ 6 , 43 , 104 , 108 - 110 ]. Dalekos et al [ 104 ] studied antibody titers and performed epitope mapping of LKM-1-positive sera from patients with chronic hepatitis C. Interestingly, a patient with a high LKM-1 titer and autoantibodies directed against an epitope of amino acids 257–269, which are preferentially recognized by patients with AIH-2, showed exacerbation of the disease under interferon-alpha treatment. In contrast to other patients with HCV infection, this patient further recognized a rarely detected epitope on the C-terminal third of the protein. These results suggest that determination and monitoring of CYP2D6 autoantibody titers by both IIF and the RLA in combination with epitope mapping of CYP2D6 in HCV/anti-LKM positive patients before the initiation of interferon-alpha treatment, might be helpful in an attempt to identify those patients at risk of developing undesired autoimmune reactions [ 104 ]. The mechanism(s) of the development and the pathogenic role of anti-LKM-1 autoantibodies in hepatocellular injury are still unclear. It has been suggested that viral infections by herpes simplex virus (HSV) and related viruses may trigger the autoantibody formation through molecular mimicry in at least some individuals with AIH-2 [ 91 ]. Manns et al [ 91 ] tested 26 LKM-positive sera using Western blot with partial sequences of recombinant CYP2D6. Eleven sera recognized a short minimal epitope of eight amino acids with the sequence DPAQPPRD. Twelve other clones recognized a larger epitope containing this eight-amino acid core sequence. The search of electronic databases revealed an interesting match of the minimal epitope with the primary structure of the immediate early protein IE 175 of HSV-1 now known as infected cell protein 4 (ICP4) of HSV-1 (Fig. 8 ). Sequence identity was present for the PAQPPR sequence. This hypothesis was further supported by a case study in a pair of identical twins [ 111 ]. In this study, one sister suffered from AIH-2 but the other one was healthy. Interestingly, only the sister suffering from AIH-2 was HSV positive, and her serum recognized the viral protein ICP4 in lysates of HSV-infected cells [ 111 ]. So far however, overall evidence for mimicry as a driving force of AIH is not convincing. Figure 8 Linear B-cell epitopes on cytochrome P450 2D6 in autoimmune hepatitis type 2. The immunodominant epitope 257–269 aa shares sequence homology with the immediate early protein IE 175, a transcription factor of herpes simplex virus type 1 (now known as infected cell protein 4 or ICP4). Although this is an attractive model for the hypothesis of molecular mimicry, overall evidence for mimicry as a driving force of autoimmune hepatitis is not convincing. Besides molecular mimicry, chemical modification of self-proteins and/or immunological cross-reactivity to homologous autoantigens may also provide potential triggers for autoimmune responses. The latter has been suggested by Choudhuri et al [ 112 ] who have shown that in AIH-2 patients the linear epitope 321–351 of CYP2D6 cross reacts with amino acids 33–51 of carboxypeptidase-H (the target autoantigen of islet cell autoantibodies in insulin dependent diabetes mellitus), as well as, with amino acids 307–325 of 21-hydroxylase (major target autoantigen in Addison's disease). These findings possibly indicate the presence of a common motif of CYP2D6, carboxypeptidase-H and 21-hydroxylase, which may contribute through a cross reactive immune response to the development of multiple endocrinopathies in the course of AIH-2. Additional support to this hypothesis emerged from two recent studies by Kerkar et al [ 94 ] and Bogdanos et al [ 113 ]. In the first study the authors were able to show the similarity and cross-reactivity between the immunodominant epitope 193–212 of CYP2D6 and homologues of two unrelated viruses (HCV 2977–2996 and CMV 121–140) [ 94 ]. In the second study the researchers investigated whether the immunodominant epitope 252–271 of CYP2D6 in anti-LKM-1 positive AIH-2 and homologues from the NS5B and E1 proteins of the HCV polyprotein and the ICP4 of HSV-1 are targets of humoral immune response in anti-LKM-1 positive and anti-LKM-1 negative HCV infected patients and furthermore whether this response is cross-reactive [ 113 ]. The hypothesis of molecular mimicry and cross-reactivity in LKM-1 production has not been addressed experimentally. The authors for the first time gave experimental support to the notion that molecular similarity between CYP2D6, HCV and HSV can result in LKM-1 production via a cross-reactive response in genetically susceptible individuals (interestingly only the HCV positive/LKM-1 positive patients with viral/self cross-reactivity possessed the HLA B51 allotype) [ 113 ]. Taking together, the above studies suggest that multiple exposure to viruses mimicking self may represent an important pathway to the development of autoimmunity [ 94 , 113 ]. Two possible mechanisms have been proposed for the involvement of anti-LKM-1 autoantibodies in the pathogenesis of liver injury. The first appears to be a direct binding of these autoantibodies to hepatocytes, leading to lysis of liver cells, while the second is associated with an anti-LKM-1 induction of activating liver-infiltrating T lymphocytes, which indicates the combination of B and T cell activity in the autoimmune process involved [ 114 - 117 ]. A prerequisite for both anti-LKM-1 production and the activation of pathogenetic mechanisms involved in liver injury, is the expression of CYP2D6 on the surface of the patients' hepatocytes. Under this context, Ma et al [ 118 ] showed that key residues of a major CYP2D6 epitope (316–327) are exposed on the surface of the molecule and may represent key targets for anti-CYP2D6 production. In addition, recent data provides convincing evidence that anti-LKM-1 autoantibodies recognize CYP2D6 exposed on the plasma membrane of hepatocytes from either AIH-2 or HCV/anti-LKM-1 positive patients [ 114 , 115 ] suggesting a pathogenetic role for these autoantibodies in hepatic tissue damage either in AIH-2 or in some cases of HCV/anti-LKM-1 positive patients [ 104 , 109 , 110 , 115 ]. So far, anti-LKM type 2 autoantibodies (anti-LKM-2) have been detected only in some cases of drug-induced hepatitis caused by tienilic acid [ 14 , 63 ]. The target autoantigen of anti-LKM-2 has been documented as the CYP2C9 [ 85 ]. A proposed mechanism for the induction of anti-LKM-2 could be the binding of an active metabolite of the drug to the CYP2C9 protein, which then becomes antigenic [ 14 , 63 , 72 , 85 ]. Anti-LKM type 3 autoantibodies (anti-LKM-3) alone or in combination with anti-LKM-1 are also detected in about 5–10% of patients with AIH-2 [ 16 , 119 ]. In contrast to anti-LKM-1 and anti-LKM-2 autoantibodies, which on immunofluorescence stain liver and kidney tissues only, with anti-LKM-3 additional fluorescence signals may be present with tissue from the pancreas, adrenal gland, thyroid, and stomach. Family 1 of UDP-glycuronosyltransferases (UGT1) is the main target autoantigen of anti-LKM-3 autoantibodies (molecular weight of 55 kDa) [ 119 , 120 ]. These autoantibodies were first described in about 13% of patients with chronic hepatitis D, but not in patients with chronic hepatitis B or C [ 121 ]. However, three recent reports have shown the presence of anti-LKM-3 autoantibodies in some patients with HCV infection [ 99 , 122 , 123 ]. These findings may further support the heterogeneous phenomenon of the HCV-induced autoimmunity. 4.2. Autoantibodies against liver cytosolic protein type 1 (anti-LC1) In 1988 a second autoantibody marker of AIH-2 was recognized [ 19 ]. This autoantibody was found to react to a liver cytosolic protein. The autoantibody is organ specific but not species specific and was therefore called anti-LC1 [ 19 ]. The anti-LC1 autoantibodies are characterized by a cytoplasmic staining of the periportal hepatocytes when the IIF assay is used for their detection. The hepatocellular layer around the central veins is not stained [ 19 , 124 ]. These findings indicate that the target autoantigen of anti-LC1 autoantibodies is not uniformly distributed in rodent liver tissues. They are detected in about 30% of patients with AIH-2 [ 19 , 124 ] and in approximately 50% of all anti-LKM-1 positive cases [ 125 ]. It is noteworthy that the anti-LC1 autoantibodies proved to be the only serological marker in 10% of patients with AIH [ 19 ]. The detection of anti-LC1 autoantibodies by IIF is obscured due to the anti-LKM-1 pattern that frequently found in most of the anti-LC1 positive sera. For these reasons other techniques such as, the ouchterlony double diffusion, immunoblot or counter-immunoelectrophoresis are required for their detection [ 19 , 124 - 126 ]. By immunoblotting, anti-LC1 positive serum samples recognize a liver specific cytosolic protein of 58–62 kDa [ 124 - 126 ]. Recently the molecular target of anti-LC1 was identified as the formiminotransferase cyclodeaminase (FTCD) [ 127 ], which is a polymeric bifunctional enzyme involved in folate metabolism. However, another group demonstrated the arginninosuccinate lyase (ASL) as the target autoantigen of a weak precipitin line detected by the ouchterlony double diffusion assay in patients with autoimmune or viral hepatitis [ 128 ]. Anti-LC1 autoantibodies have been proposed as a more specific marker of AIH-2 than anti-LKM-1 autoantibodies, since in the original reports their presence was never associated with HCV infection [ 19 , 124 ]. However, a recent study by Lenzi et al [ 125 ] confirmed the above aspect only in the pediatric subset of their patients, while a substantial proportion of the adults with anti-LC1 autoantibodies had also markers of HCV infection. The significance of the association between anti-LC1 autoantibodies and HCV infection remains uncertain and has to be established [ 106 , 129 ]. In contrast to what has been found for anti-LKM-1, the titers of anti-LC1 autoantibodies appear to parallel with disease activity [ 130 ]. The latter may indicate a possible involvement of anti-LC1 in the pathogenesis of AIH-2. However, the clinical significance of anti-LC1 is not yet completely defined. 5. Detectable Autoantibodies in AIH in APECED Chronic hepatitis as a disease component of APECED may develop in 10–18% of patients [ 14 , 15 , 23 - 25 , 28 , 29 , 63 ]. APECED appears to be caused by mutations in a recently identified gene, the autoimmune regulator gene (AIRE), and represents the only known autoimmune disease with a monogenetic mutation today [ 26 , 27 , 131 ]. It is interesting that patients with AIH in the absence of APECED do not display mutations of the AIRE gene and are therefore genetically distinct from patients with AIH as a component of APECED [ 132 ]. Similar to AIH-2, hepatitis in APECED is associated with autoantibodies directed against cytochrome P450 proteins. In a large study with APECED patients, a typical LKM staining pattern and a predominant staining of the perivenous hepatocytes in the absence of staining of the kidney were observed [ 23 ]. The latter pattern is due to autoantibodies called liver microsomal autoantibodies (anti-LM). In this study each of anti-LKM and anti-LM antibodies were found in 8% of the patients [ 23 ]. These findings indicate that two or more different microsomal antigens are hepatic target-autoantigens in APECED. Indeed, screening of APECED sera with recombinant antigens using Western blots has shown reactivity against four different hepatic cytochromes P450: CYP1A1, CYP1A2, CYP2A6 and CYP2B6 [ 23 - 25 , 133 ]. CYP1A1, CYP2A6 and CYP2B6 are expressed both in liver and in kidney resulting to an LKM staining pattern, while CYP1A2 is not expressed in the kidney leading to the LM staining. Among the four autoantibodies anti-CYP2A6 were detected with the highest prevalence in a Finnish APECED patients group (15.6%), while anti-CYP1A2 were found in 6.3% [ 23 ]. These results were confirmed by quantitative immunoprecipitation assays with recombinant 35 S-labeled CYP1A2 and CYP2A6. Contrary to a previous work in Sardinian patients with APECED [ 25 ], the detection of anti-CYP2A6 autoantibodies in a larger group of Finnish patients was not associated with the presence or absence of hepatitis, while anti-CYP1A2 autoantibodies were detected only in APECED patients with hepatitis [ 23 ]. These findings indicate that anti-CYP1A2 is a specific marker for AIH as a component of APECED, albeit with a low sensitivity [ 23 , 24 ]. Anti-CYP2A6 autoantibodies may be used as an indicator for APECED, if they are present in a patient with AIH. In parallel with the above conclusions is the anti-LKM/LM detection by IIF in about 50% of patients with AIH as part of the APECED and in only 11% of APECED patients without hepatitis [ 23 ]. The same study showed that the prevalence of ANA detection in APECED patients was high (22%) but irrespective of the presence or absence of hepatitis. Therefore ANA are not useful laboratory markers for AIH in APECED [ 23 ]. To the contrary, none of the patients' sera tested positive for anti-SLA, anti-CYP2D6 or anti-FTCD autoantibodies, which are specific markers of AIH-1 and AIH-2 [ 23 ]. On the other hand, CYP1A2 and CYP2A6 could not be identified as hepatic autoantigens in the disease control groups consisting of patients with idiopathic AIH or patients with autoimmune rheumatic diseases [ 23 ]. These findings indicate that idiopathic AIH and AIH in APECED are characterized by different molecular targets of autoimmunity, which do not overlap. Therefore, AIH and hepatitis as part of the APECED may be distinguished on the basis of differences in autoantibody profile (Tables 4 and 5 ). Table 4 Detectable autoantibodies in AIH-1, AIH-2 and AIH as part of APECED AIH-1 or AIH-2 AIH in APECED ANA, SMA, ANCA, anti-ASGP-R anti-SLA/LP (molecular target: UGA suppressor tRNA-associated protein), anti-LKM-1 (molecular target: CYP2D6), anti-LKM-3 (molecular target: UGT1), anti-LC1 (molecular target: FTCD) ANA, anti-LC (molecular target: unknown), anti-LKM (molecular targets: CYP2A6, CYP1A1 and CYP2B6), anti-LM (specific autoantibody; molecular target: CYP1A2) Abbreviations are the same as shown in the text. Table 5 Differential diagnosis of chronic liver diseases according to the presence or absence of autoantibodies against molecularly defined autoantigens of cytochrome P450 complex using the radioligand assay. Anti-CYP2D6 Anti-CYP2A6 Anti-CYP1A2 Chronic liver disease + - - AIH-2 (94–100%), HCV (0–10%) - + - HCV, APECED with or without hepatitis - - + AIH in APECED, drug induced hepatitis + + - HCV (0–7%) - + + AIH in APECED Abbreviations are the same as shown in the text. Dalekos et al [ 134 ] using the sensitive quantitative RLA reported for the first time the presence of anti-CYP2A6 autoantibodies in about 2% of HCV-positive sera in general and in 7.5% of LKM-1-positive/HCV-positive sera. The latter further supports the low specificity of this autoantibody as a marker for AIH in APECED. Interestingly, anti-CYP2A6 autoantibodies were not detected in patients with AIH-2 who exhibit high titers of anti-LKM-1 autoantibodies [ 134 ]. The clinical relevance of this finding in HCV infection remains to be determined. Anti-LM autoantibodies are first described in dihydralazine-induced hepatitis [ 135 ]. The major target autoantigen of anti-LM in both conditions (hepatitis as part of the APECED and drug-induced hepatitis) has been documented as the CYP1A2 [ 23 - 25 , 133 , 135 ]. In cases of dihydralazine-induced hepatitis the production of anti-LM autoantibodies has been attributed to adduct formation of CYP1A2 with an activated metabolite of the drug [ 136 ]. By contrast, in APECED patients no relationship between CYP1A2 and drug usage is known. In addition, it is not known whether in APECED patients a close monitoring of anti-LM may lead to early, or even prophylactic, treatment of hepatitis as a new disease component. Evidence that autoantibodies may be found before the clinical and/or laboratory manifestation of a new disease component in APECED comes from adrenal and ovarian insufficiencies, where the respective autoantibodies are detected 2–3 years before the clinical presentation of the autoimmune components [ 137 ]. Another hepatic autoantigen in APECED, the aromatic-L-amino acid decarboxylase (AADC) has also been identified recently [ 133 , 138 ]. This enzyme is expressed in the liver cytosol and was originally described as a β-cell autoantigen [ 133 ]. The prevalence of anti-AADC autoantibodies is significantly increased in APECED patients with vitiligo (88%) and hepatitis (92%) [ 5 , 14 , 29 ]. So far, anti-AADC autoantibodies have only been reported in APECED and their role in AIH and vitiligo as disease components of APECED deserve further investigation. 6. Concluding Remarks In clinical practice the recognition of AIH is of great importance since most of the patients respond favorably to antiinflammatory and immunosuppressive treatment. In addition, recent novel findings dealing with the bone marrow hemopoietic progenitor cells and bone marrow stromal cells of patients with AIH suggests alternative therapeutic options even in refractory cases [ 139 ]. Diagnostic criteria for this disease have been codified recently [ 6 ]. These include descriptive criteria and scoring system based on clinical, serologic and histologic features of AIH (Table 1 ), which contribute substantially to the differential diagnosis of the disease from other forms of chronic hepatitis associated with autoimmune phenomena (Table 6 ). The discrimination between AIH and HCV infection is of particular importance, since the immunosuppression used in the former can deteriorate liver disease in HCV patients, while interferon-alpha treatment used in HCV infection may lead to exacerbation of AIH [ 104 , 108 - 110 ]. Table 6 Differential Diagnosis of Autoimmune Hepatitis. Other autoimmune liver diseases - Overlap syndromes - Primary biliary cirrhosis - Primary sclerosing cholangitis Chronic viral hepatitis - Chronic hepatitis B with or without hepatitis delta - Chronic hepatitis C - Chronic hepatitis non A to G Cholangiopathy due to human immunodeficiency virus infection Alcoholic liver disease Drug-induced hepatitis Non-alcoholic steatohepatitis Granulomatous hepatitis Hemochromatosis α 1 -antithrypsin deficiency Wilson's disease Systemic lupus erythematosus The detection of non-organ specific autoantibodies remains the hallmark of AIH. A step by step diagnostic application of autoantibody tests is mandatory for the evaluation of acute or chronic hepatitis of unknown cause. ANA, SMA and anti-LKM-1 autoantibodies should be first tested in patients with acute or chronic elevation of aminotransferases when virologic tests are negative and there is no current or past history for drug or alcohol abuse. Determination of ANCA, which occur in up to 90% of patients with AIH-1, may be useful in the identification of individuals who are seronegative for the above conventional autoantibody markers but should be kept in mind that this autoantibody lacks specificity. Many target autoantigens of the non-organ specific autoantibodies have been identified, but the latter has not led to the characterization of specific subpopulations of patients or changes in the treatment strategies. In addition, most of the non-organ specific autoantibodies do not seem to be involved in the pathogenesis of liver injury in AIH. Anti-LKM-1 autoantibodies could be an exception to the above aspect since recent data have demonstrated the expression of CYP2D6 on the surface of hepatocytes, while AIH-2 has not been observed in individuals who are deficient for CYP2D6. These findings provide arguments for an antigen-driven autoimmune process. It is possible that mutations in the autoantigen itself can lead to alterations in the three dimensional structure of the antigen, which induces autoimmunity. Antibodies directed against liver-related antigens have had similar limitations. Anti-ASGP-R and anti-LC1 autoantibodies appear to correlate with disease severity and response to treatment suggesting a pathogenetic role to the hepatocellular damage. In general however, autoantibodies should not be used as a tool for monitoring of treatment or to predict AIH activity and outcome. Anti-SLA/LP autoantibodies have been considered as valuable and specific markers for the diagnosis of AIH-1. However, a recent study has shown that anti-SLA/LP autoantibodies can also be detected in AIH-2 and in children with PSC. Irrespective of the disease specificity of this marker, it is obvious that testing for anti-SLA/LP will help to reduce the group of cryptogenic liver disease, by recognizing previously misdiagnosed patients with AIH who were seronegative for ANA, SMA or anti-LKM-1. In APECED, autoantibodies are directed against specific cytochrome P450 enzymes (e.g. CYP1A2, CYP2A6, CYP21, CYP17, and CYP11A1), that are expressed in organs affected by the disease process. These observations argue against the idea that antibodies against cytochrome P450 complex are simply epiphenomena secondary to tissue damage and that they have no relation to the etiology and pathogenesis of APECED. It is not known what triggers autoimmunity in AIH. The hypothesis that different causes may lead to loss of tolerance against the same molecular target autoantigen seems attractive. For instance, CYP1A2 is the molecular target in dihydralazine-induced hepatitis and AIH as a component of APECED, CYP2D6 in AIH-2 and in some patients with HCV infection, CYP2A6 in APECED and in a proportion of patients with HCV infection and UGT1 in some cases of AIH-2 and chronic hepatitis D or C. Research protocols in order to define AIH pathogenesis, disease susceptibility, determinants of disease severity, and to understand the epidemiology of AIH are future challenges in the investigational and clinical arena of this disease [ 139 - 141 ]. Competing interest The author(s) declare that they have no competing interests.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544946.xml
449906
Whose Copy? Whose Rights?
Primary scientific and medical literature is written for academic communities and the public good -- so should be governed by copyright licenses that permit the full range of uses that could benefit stakeholders in research.
This is the third installment in a series of editorials on the implications of open-access publishing for established publishing practices and stakeholders in scientific and medical research. The questions, tensions, and social concerns surrounding copyright and the Internet are very different for scientific and medical literature than for other kinds of easily reproducible digital works. Peer-reviewed publications are often the sole tangible products of the tremendously time-consuming and expensive process of conducting primary research in biology and medicine. Who should own the primary research articles that are the culmination of years of work by scientists and staggering financial investments by governments, universities, and tax-exempt foundations? What uses should the documents' owners permit? In practice, academic authors typically assign full copyrights to their articles to the publishers of the journals in which the works appear. Scientists do not benefit financially from the transaction; indeed, they often subsidize the cost of their articles' publication in the form of page charges, color charges, and other fees levied by publishers. The prestige associated with publishing in a selective journal is sufficiently valuable that scientists are generally willing to abdicate the legal rights to their own work without remuneration. In recent years, however, several technological and legal innovations have led a growing number of scientists to begin to question the sagacity of this arrangement. The advent of electronic publishing and the Internet itself have made technically possible a slew of novel uses of primary research papers. Simultaneously, the traditional “all rights reserved” copyright license has been supplemented by a variety of alternative licenses—of equal legal validity and available at no charge to anyone who wants them—that allow copyright holders to prevent some uses of a work without permission, but to authorize others. Different licenses created by the nonprofit organization Creative Commons ( www.creativecommons.org ), for example, allow copyright holders to mark their work with freedoms—to permit a work's reproduction for any noncommercial purpose (the Noncommercial License) or for any purpose at all provided that the original authorship is properly attributed (the Attribution License). The upshot of these developments is that copyright holders can now permit a spectrum of uses of a paper by prospective researchers, anthologizers, archivists, teachers, patients, policy makers, journalists, and other interested parties. Precisely which uses are permitted and which are not is far from a trivial matter. The particular copyright license under which an article is published largely determines how the document can be stored, searched, and built upon by other scientists. Authors' Rights and Users' Rights One implication of the variety of copyright licenses now widely available is that the right to use an article in one way or another is largely independent of its accessibility online. A paper that is touted as “freely available” or “free access” is very different from one that is “open access.” (See http://www.plos.org/openaccess for the formal definition of an open-access article, drafted at the 11 April 2003 Bethesda Meeting on Open Access Publishing.) When a document is “freely available,” someone who comes across it may be permitted to do nothing more than read it online on a publisher's Web site; the right to use the article in any other way is typically granted only at the publisher's discretion. When a document is open access, however, a wide range of additional uses are perpetually and irrevocably allowed— from the reproduction and distribution of the paper by a professor for his or her students to the archiving of the paper in a searchable online repository available to anyone in the world with an Internet connection, and more. The Creative Commons Attribution License (CCAL), which governs this editorial and all other content in PLoS Biology , permits a number of uses of articles that are typically restricted and for which there is an immediate demand. All articles published by PLoS can be included in coursepacks—a use-right that most authors would want to allow without exception, but that most traditional publishers grant only for a substantial fee (which they rarely share with authors). The CCAL also ensures that institutions are permitted to archive not only articles written by their own faculty, but all other works published under the same legal terms as well, thereby facilitating their permanent accessibility and preservation. For example, the LOCKSS (“Lots of Copies Keep Stuff Safe”) program ( http://lockss.stanford.edu/projectdescbrief.htm ), an ongoing project to support libraries' efforts to “create, preserve, and archive local electronic collections,” is viable only insofar as institutions are permitted to store information themselves, rather than access it exclusively via publishers' Web sites. Many collaborative projects between libraries and publishers have been complicated by legal constraints, including the stipulation that archives remain “dark,” or inaccessible to users, until any commercial incentive for restricting access to articles has been exhausted—clearly a suboptimal arrangement for researchers, and one that is unnecessary for collections of works governed by the CCAL. One of the truly revolutionary implications of open-access articles, however, is that we simply do not know the full range of their potential applications. They are available for any use that any entrepreneur can envision, so long as the authors of the papers are properly credited. The only certainty, then, is that the utility of open-access research articles will be limited solely by the imagination of those that are inspired by the possibilities—rather than by legal constraints. Authors' Protections Authors retain the copyright to all articles in PLoS Biology and license their works under nonexclusive terms that reserve only some—rather than all—rights. There are several common objections, generally leveled by publishers, against this practice. For example, it is sometimes argued that the traditional copyright arrangement in scientific publishing protects against uses of articles that authors would object to—while the CCAL permits such uses and renders authors helpless to prevent them. To the extent that the uses in question are for academic or archival purposes, such as those discussed above, it is certainly true that the CCAL permits practices that “all rights reserved” licenses do not. Indeed, the expanded range of legitimate academic uses of articles is among the primary selling points of the CCAL in the context of scientific publishing. To the extent that the ostensibly objectionable uses are commercial, the problem is easily remedied with the Creative Commons Noncommercial License, which prohibits commercial reuse of a work without the copyright holder's consent. PLoS has chosen, for reasons both philosophical and pragmatic, to permit the commercial use of works we publish. As a matter of principle, all of our policies reflect the view that scientific publishers are service providers and should not themselves restrict the potential applications of the largely publicly funded work in their journals. More concretely, if a commercial enterprise is interested in repackaging the articles that PLoS has published, we are loath to prevent an author's work from wider distribution. Any risk that a company will use an article for a purpose its author would be uncomfortable with is, in our view, substantially outweighed by the benefits of allowing—not on a case-by-case basis, but across the board—the reproduction of the article for inclusion in online encyclopedias, or for distribution in countries in which Internet access is unreliable, or, indeed, for creative uses we hope to inspire by making primary research articles legally available to commercial interests. Another recurring objection to the copyright arrangement that PLoS employs is that authors are inappropriate copyright holders because they are ill-equipped to protect their own works against plagiarism, misattribution, and other misuse. Most scientists, however, have enough familiarity with cases of plagiarism in their own field to know that their strongest protection against mis- or nonattribution is derived not from the threat of prosecution for copyright infringement, but from community standards of conduct. Furthermore, among the benefits of open-access articles is the fact that their full texts, rather than just their abstracts, are searchable—which, as any teacher knows, makes plagiarism much easier to detect. Beyond plagiarism and misattribution, it is not clear what uses of primary research articles authors would actually want to prevent (other than, perhaps, the commercial uses that their work is already susceptible to, in many cases, when publishers hold copyrights). Scientists do not receive royalties for their published work. The more widely their articles are read and cited, the more their professional reputations are bolstered. Certainly, research articles have a wide range of uses that publishers typically object to— and indeed often file suit over—such as their compilation in coursepacks by copy shops. Those applications, however, tend not to constitute “misuse” in many authors' eyes. Authors' Voice There is no question that the licensing arrangement PLoS employs is relatively novel—and therefore untested over the long term—in biomedical publishing. However, it hardly takes a radical understanding of the interests of authors and users of primary research articles to conclude that the open-access terms of copyright promise substantial benefits for both groups. What, then, can scientists do to encourage other publishers to follow suit and strike similar legal arrangements with authors as a matter of course? One answer is to “vote with your submissions;” that is, authors should submit their work preferentially to journals with copyright and licensing practices that genuinely serve their interests. Another equally important action for scientists is to raise the issue with their professional societies. Scholarly associations exist, among other reasons, to serve the needs of their members—and society members should actively urge their society journals to employ the CCAL or a similar license for their research articles. Scientific and medical literature is different from fiction or movies or music. The United States government invests more than $28 billion per year in the National Institutes of Health alone to fund research in biology and medicine. The scientists who conduct that research and the research paid for by other public-minded institutions in the United States and abroad have an affirmative moral obligation to share the knowledge they create—not just with students and faculty at elite Western universities, but with everyone who could use it and build upon it. When authors publish their work in journals with restrictive copyright practices, it becomes illegal (often for even the authors themselves) to store primary research articles in many archives or include them in coursepacks or use them for other responsible purposes. Those obstacles to sharing knowledge can be avoided without legislative intervention, however, if scientists and publishers alike embrace a legal paradigm for disseminating new discoveries that maximizes their utility.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC449906.xml
549523
Chromatography paper strip sampling of enteric adenoviruses type 40 and 41 positive stool specimens
Background The enteric subgroup F adenoviruses type 40 (Ad40) and 41 (Ad41) are the second most important cause of acute infantile gastroenteritis after rotaviruses. Repeated community outbreaks have been associated with antigenic changes among the Ad40 and Ad41 strains due to host immune pressure. Therefore large field epidemiological surveys and studies on the genetic variations in different isolates of Ad40 and Ad41 are important for disease control programs, the design of efficient diagnostic kits and vaccines against subgroup F adenoviruses. A novel method using sodium dodecyl sulphate SDS/EDTA-pretreated chromatography paper strips was evaluated for the collection, storage and shipping of Ad40/41 contaminated stool samples. Results This study shows that adenoviral DNA can be successfully detected in the filter strips by PCR after four months storage at -20°C, 4°C, room temperature (20–25°C) and 37°C. Furthermore no adenoviral infectivity was observed upon contact with the SDS/EDTA-pretreated strips. Conclusions Collecting, storing and transporting adenovirus type 40 and 41 positive stool samples on SDS/EDTA-pretreated chromatography filter strips is a convenient, biosafe and cost effective method for studying new genome variants and monitoring spread of enteric adenovirus strains during outbreaks.
Background Enteric adenoviruses (EAds) are considered to be the second most important causative agent of acute infantile gastroenteritis after rotaviruses. The fastidious subgroup F adenoviruses type 40 (Ad40) and 41 (Ad41) account for the majority of cases of severe acute diarrhea in children less than 2 years of age [ 1 , 2 ]. These viruses usually cause sporadic infantile gastroenteritis, but they have also been implicated in outbreaks and nosocomially acquired diarrhea [ 3 - 5 ]. The course of the disease is mild and self-limiting in most cases, but in immunocompromised patients these infections are associated with an increased morbidity and prolonged hospitalization [ 6 , 7 ]. Repeated community outbreaks and shift in the prevailing subgroup F adenovirus type have been associated with antigenic changes among the Ad40 and Ad41 strains due to host immune pressure [ 8 - 12 ]. Therefore, large field epidemiological surveys and studies on the genetic variations in different isolates of Ad40 and Ad41 are important for disease control programs, the design of efficient diagnostic kits and vaccines against subgroup F adenoviruses. For these purposes stool samples need to be collected, stored and transported to reference laboratories for genetic analysis. In many developing countries and remote areas, collection and storage of samples for laboratory diagnosis is difficult due to a restricted infrastructure. Moreover field conditions may limit the handling, transportation and refrigeration of the specimens. Previous studies have demonstrated the application of different filter papers for the collection and storage of blood [ 13 ], saliva [ 14 ] and stool [ 15 ] samples for further analysis. Filter paper sampling has been successfully used for screening studies of rotaviruses [ 16 ], noroviruses [ 17 ], human herpesviruses 6 and 7 [ 14 ], human immunodeficiency virus [ 13 , 18 ], hepatitis C virus [ 19 ], measles virus [ 20 ] and others viruses. This study describes the use of SDS/EDTA-pretreated filter paper strips in collection, transportation and storage of adenovirus type 40/41 positive stool samples for subsequent genetic analysis. Results and Discussion In the current study we describe the use of chromatography paper strips for the collection, transportation, and storage of EAds type 40/41 positive stool samples. In order to inactivate the adenoviruses and other microorganisms upon contact with the strips, the latter were pretreated with SDS, a surfactant with protein denaturising ability. This can allow safe transportation of the strips without extensive biohazard precautions. To protect the viral DNA from degradation by deoxyribonucleases (DNases) the chromatography strips were also preincubated with EDTA, and Tris-HCl. EDTA chelates magnesium ions, a necessary co-factor for most nucleases and the weak organic base Tris-HCl ensures the proper action of the chelating agent in binding the divalent cations. A diarrheal stool sample containing 2.6 × 10 6 adenoviral particles per ml was serially diluted 1:8 (dilution a ), 1:80 (dilution b ), 1:800 (dilution c ), 1:8000 (dilution d ) and 1:80,000 (dilution e ). The SDS/EDTA-pretreated filter paper strips were infected with each stool dilution and stored at -20°C, 4°C, room temperature (20 to 25°C), and 37°C. The presence of adenoviral DNA on the chromatography filter strips was detected by PCR amplification of a 301 bp fragment of the adenoviral hexone gene after storage for 7 days, 14 days, 56 days and 120 days (Figure 1 ). Figure 1 Polyacrylamide gel electrophoresis of the PCR products amplified from the DNA of the subgroup F adenovirus positive stool sample, extracted from the SDS/EDTA pretreated chromatography paper strips that have been stored at four different temperature conditions. Five tenfold dilutions of the original stool sample were tested ( a = 1:8, b = 1:80, c = 1:800, d = 1:8000, e = 1:80,000). Our results show that adenoviral DNA can remain stable even at higher than room temperature conditions for at least 4 months, indicating that the collection and storage of the infected filter strips is possible where freezers are not available. To be sure that the EAd40/41 contaminated filter strips are not infectious, we carried out a biosafety test to find out if any adenovirus could survive onto the SDS/EDTA-pretreated strips. Previously we showed that pathogenic bacteria such as Vibrio cholerae , enterotoxigenic E. coli , enteropathogenic E. coli , Salmonelle typhimurium and Shigella dysenteriae , were not able to survive on the SDS/EDTA strips [ 16 ]. The biosafety test performed in this study demonstrated that adenoviruses also lost infectivity upon contact with the SDS/EDTA strips (Figure 2 ). In HeLa cell line, no cytopathic effect was observed after incubation with the dialyzed eluate of the SDS/EDTA strips loaded with adenovirus type 1 (10 6 TCID 50/ml) after three passages of the infected cell line. The eluate of the untreated chromatography strips loaded with adenovirus type 1 caused cytopathic effect in the HeLa cell line. It can be concluded that the SDS/EDTA-pretreated strips can be used for the collection and shipping of adenovirus positive stool samples from remote areas to reference laboratories in a biosafe way. Figure 2 (A) Normal HeLa confluent monolayer. (B) CPE in the HeLa cells at 3 days after infection with adenovirus type 1. (C) CPE in the HeLa cells at 3 days after infection with 500 μl eluate of the infected adenovirus type 1 untreated filter paper strips. (D) CPE in the HeLa cells after 3 days infected with the dialysed adenovirus type 1 positive cell cultured sample. (E) No CPE was observed when the HeLa cells were infected with 500 μl of the dialysed eluate from the adenovirus type 1 infected SDS/EDTA-pretreated paper strip. Conclusions We conclude that the use SDS/EDTA-pretreated filter strips for retrieval and subsequent analysis of adenoviral DNA from EAds type 40/41 positive stool specimens is a feasible method for sample collection. The described filter paper strips facilitate collection, transport and storage of adenoviral positive stools because they are biosafe, cost effective and require minimal storage space. This study shows that adenoviral DNA can remain stable for at least 4 months at 37°C temperature conditions making this method especially attractive for field research or population screening in tropical countries where freezers are not available. Materials and Methods Chromatography paper strips Highly absorbent (870 g of water/m 2 ) Whatman grade 17 chr pure cellulose chromatography paper with thickness of 0.92 mm and a flow rate of 190 mm/30 min (Whatman, Kent, United Kingdom) was used. Strips of 80 mm × 4 mm were cut from the chromatography paper and soaked for two minutes in a solution of 2% (w/v) sodium dodecyl sulphate (SDS), 10 mM EDTA and 60 mM TrisHCl. The chromatography paper strips were left to dry overnight at room temperature. Disposable gloves were used during the preparation of the filter paper strips. Adenovirus sample loading on the chromatography paper strips A diarrhea stool sample that was positive for adenovirus type 40/41 hexon antigen by the Premier Adenoclone ® -Type 40/41 solid-phase sandwich enzyme immunoassay (Meridian Bioscience, Cincinnati, Ohio) was used for this study. The undiluted feces sample contained approximately 2.6 × 10 6 particles per ml of stool, as calculated from a standard curve supplied with the antigen enzyme immunoassay kit. The stool sample was diluted in 1 ml (dilution 1:8) DNase/RNase free water (Sigma) and the following dilutions were used: 1:8 (dilution a ), 1:80 (dilution b ), 1:800 (dilution c ), and 1:8000 (dilution d ), and 1:80000 (dilution e ). The pretreated chromatography strips were infected with 100 μl of the different dilutions of stool sample and were left to air dry overnight at room temperature. After complete drying, the infected strips were stored under four different temperature conditions: -20°C, 4°C, room temperature (20 to 25°C) and 37°C. PCR detection Half of the strip (160 mm 2 ) was used for the DNA extraction performed at the following storage time intervals: 7, 14, 56 and 120 days. The filter paper was inserted into an Eppendorf tube with 500 μl of Dnase/Rnase free water (Sigma) and thoroughly squeezed out. An aliquot of 200 μl of the squeezed eluate was used for DNA extraction using the QIAamp DNA Blood Mini Kit (Qiagen/Westburg, Leusden, The Netherlands) according to the manufacturer's instructions. A set of degenerate consensus primers (forward primer 5'-GCCSCARTGGKCWTACATGCACATC-3' and (reverse primer 5'-CAGCACSCCICGRATGTCAAA-3') were used to amplify a 301 bp fragment of the adenoviral hexone gene [ 21 ]. The PCR assay was performed with 10 μl of the extracted DNA in a 50 μl total volume, containing 0.5 μM of forward and reverse primer, 200 μM nucleotides, 2.5 mM MgCl 2 , and 1 unit Taq polymerase (Applied Biosystems, Foster City, CA). The PCR was conducted in a Geneamp PCR System 9600 thermal cycler (Applied Biosystems). The thermocycling conditions consisted of denaturation at 94°C for 3 min, followed by 35 cycles of 30 s at 94°C, 30 s at 55°C and 1 min at 72°C and 5 min of final elongation at 72°C. PCR products were visualized using polyacrylamide gel electrophoresis and ethidium bromide staining. Biosafety test for adenovirus A biosafety experiment was performed to check if adenoviral particles are still infectious after contact with the SDS/EDTS-pretreated chromatography paper strips. Since Ad40 and Ad41 grow poorly in cell culture it is difficult to detect these viruses in vivo . Therefore adenovirus type 1 was used for the biosafety experiments. The SDS/EDTA-pretreated filter stips were first infected with 100 μl of the HeLa cell cultured adenovirus type 1 (10 6 TCID 50/ml) and allowed to dry at room temperature for 60 min. The strips were then placed into an eppendorf tube containing 500 μl Dulbecco's Modified Eadle Medium (DMEM) (Invitrogen, Merelbeke, Belgium) supplemented with 200 mM L-glutamine (Sigma-Aldricht, Bornem, Belgium). The strips were thoroughly squeezed in the medium and the eluate was dialyzed using 3,500-Da Slide-A-Lyzer dialysis cassettes (Pierce Biotechnology, Rockford, IL, USA) to remove the cytotoxic SDS. The dialyzed eluate was inoculated on a confluent monolayer of HeLa cells and was incubated at 37°C in a humified incubator with a 5% CO 2 environment. Untreated strips infected with adenovirus type 1 and noninfected SDS/EDTA-pretreated strips were used as positive and negative controls respectively. The presence of cytopathic effect indicated the presence of live replicating virus on the strip. Cytopathic effects were monitored up to the third passage of the tissue culture supernatant. List of abbreviations Eads – enteric adenoviruses EDTA – ethylenediamine tetra-acetic acid SDS – sodium dodecyl sulphate Competing interests The author(s) declare that they have no competing interests. Authors' contributions KZ conducted the study, carried out the experiments and wrote the manuscript. PM carried out the biosafety experiments. MR developed the filter paper strip sampling method. MVR supervised the study and revised the manuscript. All authors read and approved the manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549523.xml
547917
Adaptogenic and nootropic activities of aqueous extract of Vitis vinifera (grape seed): an experimental study in rat model
Background The aerial parts of Vitis vinifera (common grape or European grape) have been widely used in Ayurveda to treat a variety of common and stress related disorders. In the present investigation, the seed extract of V. vinifera was evaluated for antistress activity in normal and stress induced rats. Furthermore, the extract was studied for nootropic activity in rats and in-vitro antioxidant potential to correlate its antistress activity. Methods For the evaluation of antistress activity, groups of rats (n = 6) were subjected to forced swim stress one hour after daily treatment of V. vinifera extract. Urinary vanillylmandelic acid (VMA) and ascorbic acid were selected as non-invasive biomarkers to assess the antistress activity. The 24 h urinary excretion of vanillylmandelic acid (VMA) and ascorbic acid were determined by spectrophotometric methods in all groups under normal and stressed conditions. The nootropic activity of the extract as determined from acquisition, retention and retrieval in rats was studied by conditioned avoidance response using Cook's pole climbing apparatus. The in vitro antioxidant activity was determined based on the ability of V. vinifera to scavenge hydroxyl radicals. Results Daily administration of V. vinifera at doses of 100, 200 and 300 mg/kg body weight one hour prior to induction of stress inhibited the stress induced urinary biochemical changes in a dose dependent manner. However, no change in the urinary excretion of VMA and ascorbic acid was observed in normal animals at all the doses studied. The cognition, as determined by the acquisition, retention and recovery in rats was observed to be dose dependent. The extract also produced significant inhibition of hydroxyl radicals in comparison to ascorbic acid in a dose dependent manner. Conclusion The present study provides scientific support for the antistress (adaptogenic), antioxidant and nootropic activities of V. vinifera seed extract and substantiate the traditional claims for the usage of grape fruits and seeds in stress induced disorders.
Background Stress can be described as the sum total of all the reactions of the body, which disturb the normal physiological condition and result in a state of threatened homeostasis. Stress is an internationally recognized phenomenon fortified by advancement of industrialization in a demanding civilization. Thus, every individual is likely to face stressful situations in day-to-day life. Stress represents a reaction of the body to a stimulus that tends to alter its normal physiological equilibrium or homeostasis and has been defined as a nonspecific response of the body to any demand imposed on it [ 1 ]. Since the introduction of adaptogens, several plants have been investigated, which were once used as tonics due to their adaptogenic and rejuvenating properties in traditional medicine [ 2 ]. The drugs of plant origin are gaining increasing popularity and are being investigated for remedies of a number of disorders including antistress (adaptogenic) activity [ 3 ]. The initial studies on Ocimum sanctum [ 4 ], Withania somnifera [ 5 ] opened a vast area of research and substantial work has been carried out on plants such as Eleuthrococcus senticosus and Panax ginseng [ 6 ]. Vitis vinifera (Linn.) (Family: Vitaceae) also called as common grape or wine grape or European grape is one of the fruit crops most widely grown throughout the world [ 7 ]. In the indigenous Indian system of medicine (Ayurveda), the aerial parts of V. vinifera have been widely used to treat a variety of common and stress related disorders [ 8 ]. The composition and properties of grape seeds have been extensively investigated, and reported to have many favourable effects on human health such as the lowering of low-density lipoprotein [ 9 , 10 ], reduction of cardiovascular diseases and cancer [ 11 ]. In addition, the seed extracts of V. vinifera are reported to have antimicrobial and free radical scavenging properties [ 12 ]. In the present investigation, the antistress activity of Vitis vinifera was evaluated in-vivo , in normal and stress induced rats following a biochemical approach. The antioxidant potential of the extract was evaluated in-vitro to support the antistress activity. The plant extract was further evaluated for nootropic activity using conditioned avoidance response in rats. Methods Preparation of extract The riped fruits of Vitis vinifera were collected from the Chittor district of Andhra Pradesh, India in the month of October and the seeds were separated from the pulp and shade dried. The dried powdered seed material of Vitis vinifera (5 kg) was extracted with boiling water (25 L) for 45 minutes and the filtrate was evaporated under vacuum below 70°C in a vacuum drier to give a final yield of 50 gm. Chemicals used Vanillylmandelic acid (VMA) and scopolamine butylbromide (SBB) were purchased from Sigma-Aldrich, St. Louis, USA, while ascorbic acid was purchased from Loba Chemie, Mumbai. All other reagents used were of analytical grade. Animals All animal experiments were performed in accordance with our Institutional Animal Ethics Committee and by the animal regulatory body of the government (Regd. No. 516/01/A/CPCSEA). Albino rats of either sex obtained from Ghosh Enterprises, Kolkata were used in the study. They were housed six per cage at a temperature of 22 ± 2°C with 12 h light/ dark cycle under controlled environment. Rats were fed with standard pellet diet (Ratan Brothers, Hyderabad), and water ad libitum . Animals were kept for seven days in laboratory for habituation. Antistress activity Rats of either sex weighing between 120–150 gm were divided into four groups (I, II, III, IV) each containing six animals. The 24 h urine sample from each group was collected into two different beakers, one containing 5 ml of 10% oxalic acid for the spectrophotometric determination of ascorbic acid at 550 nm [ 13 ] and the other containing 0.5 ml of 6 N hydrochloric acid for the determination of vanillylmandellic acid (VMA) spectrophotometrically at 360 nm [ 14 ]. The experimental protocol was divided into four phases. In the first phase of the experiment, 24 h urine samples were collected in all the four groups and subjected to analysis for both VMA and ascorbic acid and the normal values were recorded for seven consecutive days. In the second phase, the animals in each group were subjected to fresh water swimming stress individually. In this method, rats were forced to swim until exhausted (three to four minutes) in a cylindrical vessel of 60 cm height and 45 cm diameter containing water at room temperature (28°C). Water depth was always maintained at 40 cm. The 24 h urinary levels of VMA and ascorbic acid under stressed conditions were determined again as described above daily for seven consecutive days. The third phase of the experiment consists of administration of V. vinifera extract to the same groups of animals after having recovered completely to normal condition. Groups II, III and IV were administered orally with V. vinifera (suspended in 2% gum acacia) at daily doses of 100, 200 and 300 mg/kg body weight respectively for seven consecutive days while group I serving as control. The 24 h urine samples were collected and the levels of both VMA and ascorbic acid were determined. The final phase of the experiment consisted of administration of V. vinifera extract to the same groups of animals after a recovery period of one week. Groups II, III and IV were administered orally with V. vinifera at doses of 100, 200 and 300 mg/kg body weight respectively, one hour prior to the daily induction of stress for seven consecutive days while group I serving as control. The 24 h urine samples were collected and analyzed for VMA and ascorbic acid for seven consecutive days to study the influence of the extract on the stress induced biochemical changes. Nootropic activity The Nootropic activity of V. vinifera was evaluated by using the conditioned avoidance response ( CAR ) in rats as described by Cook and Weidley [ 15 ]. Rats were divided into 4 groups each containing six animals. Groups II, III and IV were administered orally with 100, 200 and 300 mg/kg body weight respectively of V. vinifera (suspended in 2% gum acacia) while animals in group I were served as control. After 60 minutes, all the animals were subjected to a training schedule individually by placing inside the perspex chamber of the apparatus. After an accustomed period of five minutes to the chamber, a buzzer was given followed by a shock through the grid floor. The rat had to jump on the pole to avoid foot shock. Jumping on the pole functionally terminates the shock and this was classified as an escape while such jumping prior to the onset of the shock was considered as avoidance. The session was terminated after completion of 60 trials with an interval of 20–30 seconds given for each trial. This procedure was repeated at 24 h intervals until all groups reach 95 to 99% avoidance. After attaining complete training of a particular group, the animals were treated with a single dose of scopolamine butyl bromide (1 mg/kg body weight, i.p.), thirty minutes before the next day dosing. The training schedule was continued further with the daily doses of the extract and vehicle until they returned to normal level from scopolamine induced amnesia. Antioxidant activity The antioxidant activity of V. vinifera was determined based on its ability to scavenge the hydroxyl radicals [ 16 ]. Hydroxyl radical scavenging activity was measured by studying the competition between deoxyribose and the extract for hydroxyl radicals generated from the Fe 3+ -ascorbate-EDTA-H 2 O 2 system. The hydroxyl radicals attacks deoxyribose and eventually results in the formation of thiobarbituric acid reacting substances (TBARS). The reaction mixture containing deoxyribose (2.8 mM), ferric chloride (0.1 mM), EDTA (0.1 mM), H 2 O 2 (1 mM), ascorbate (0.1 mM) phosphate buffer (20 mM, pH 7.4) and various quantities of the extracts in a final volume of 1 mL was incubated for 1 h at 37°C. Deoxyribose degradation was measured as TBARS by the method of Ohkawa et al., 1979 and the percentage free radical inhibition was calculated from control. Data and statistical analysis The results are expressed as means ± standard error of means. Statistical analysis was done using Student's paired t -test. In all the cases, p < 0.05 was considered statistically significant. Results Antistress activity The urinary data of VMA and ascorbic acid observed in various phases of the experiment are shown in Fig. 1 and Fig. 2 respectively. Induction of forced swim stress to the animals produced a significant increase in VMA and decrease in ascorbic acid excretion compared to their respective basal excretion in normal condition. Both the parameters were found to return to their normal levels in three to four days after withdrawal of stress. Daily treatment of V. vinifera to the animals under normal condition produced no change in the excretion of VMA and ascorbic acid compared to normal basal levels indicating that V. vinifera did not alter excretion of VMA and ascorbic acid in normal condition. Daily administration of V. vinifera one hour prior to the induction of stress inhibited the increase in VMA and decrease in ascorbic acid excretion which was manifested during stress alone. The inhibition was found to be significant at all dose levels in a dose dependent manner. Figure 1 Influence of Vitis vinifera seed extract on the 24 h urinary levels of VMA in normal and stress induced rats . Each bar indicates the mean excretion of six animals. Significant difference from normal condition of the corresponding groups: *p < 0.05 Significant difference from stress condition of the corresponding groups: # p < 0.05 NS- No significant difference from normal condition of the corresponding groups. §No significant difference from stress condition. Figure 2 Effect of Vitis vinifera seed extract on the 24 h urinary levels of ascorbic acid in normal and stress induced rats . Each bar indicates the mean excretion of six animals. Significant difference from normal condition of the corresponding groups: *p < 0.05 Significant difference from stress condition of the corresponding groups: # p < 0.05 NS- No significant difference from normal condition of the corresponding groups. §No significant difference from stress condition. Nootropic activity The CAR of rats administered with the extract of Vitis vinifera or vehicle increased gradually to 95% over seven to ten days. The percent avoidance was always higher in the extract treated groups compared to vehicle treated control group. The acquisition (time to achieve 95% CAR ) for the extract treated groups was quicker and found to be dose dependent. Animals receiving 300 mg/kg body weight of the extract have taken seven days whereas, groups treated with 200 and 100 mg/ kg doses of the extract required eight and nine days respectively to reach the point of acquisition (Fig. 3 ). Administration of scopolamine produced amnesia as seen from reduction in the observed CAR. However, continued treatment of V. vinifera produced better retention and recovery in a dose dependent manner than the vehicle treated animals. Figure 3 Effect of Vitis vinifera seed extract on the mean percent of conditioned avoidance response after oral administration in rats . Scopolamine butylbromide (SBB) was administered thirty minutes before the next day dosing with the extract after attaining complete acquisition. Antioxidant activity Degradation of deoxyribose mediated by hydroxyl radicals generated by Fe 3+ /ascorbate/EDTA/H 2 O 2 system was found to be inhibited by Vitis vinifera . The extract at quantities of 100, 200, 400 and 800 μg levels scavenged the hydroxyl radicals in a dose dependent manner. Ascorbic acid at concentration of 2500, 5000 and 10000 μg was also found to produce dose dependent inhibition of hydroxyl radicals. The quantity of the extract needed for 50% inhibition was found to be 610 μg (Fig. 4 ). Similar effect was produced by ascorbic acid at a concentration of 4875 μg. (Fig. 5 ). Figure 4 Hydroxyl radical scavenging activity of Vitis vinifera in in-vitro systems . Graphical representation of the concentration of Vitis vinifera required to inhibit 50 percent of hydroxyl radicals. Each point represents the mean percentage inhibition of six experiments. Figure 5 Hydroxyl radical scavenging activity of ascorbic acid in in-vitro systems . Graphical representation of the concentration of ascorbic acid required to inhibit 50 percent of hydroxyl radicals. Each point represents the mean percentage inhibition of six experiments. Discussion Advancement in the understanding of processes leading to explore the reason for stress induced disorders cannot obscure the simple fact that the exhaustion of energy supply still forms the basis that triggers the disorders and collapse of energy metabolism following glucose deprivation in circulation [ 17 ]. The desire to control the coping mechanism has led to emergence of science of adaptation, focusing to elucidate the mechanism that may help in modification so that insufficient, excessive and unnecessary responses can be prevented. Literature reports indicate that noradrenaline is released during stressful conditions [ 18 ] and metabolized to vanillyl mandelic acid(VMA) peripherally and 3-methoxy 4-hydroxyphenyl glycol (MOPEG) centrally. In the light of such reports, VMA, the major metabolite of sympathetic amines, was taken as indirect biochemical index to represent the increase in peripheral sympathetic activity during stress. In the present study, the increase in the urinary VMA excretion during stress was used as a non-invasive biochemical marker to study the antistress activity of V. vinifera . Several studies also indicated that the tissue levels of ascorbic acid decreased on application of stress [ 19 ]. Ascorbic acid being a free radical scavenger [ 20 ], it is more likely utilized in scavenging the free radicals involved in stress resulting in its decreased urinary concentration and also it has role in the biosynthesis of noradrenaline [ 21 ] i.e ., as a cofactor in the conversion of dopamine to noradrenaline [ 22 ]. Based on the above studies ascorbic acid excretion in urine was taken as an indirect biochemical index to indicate the influence of stress on catecholamine synthesis in rats and antistress (adaptogenic) activity of the Vitis vinifera extract on prior administration of stress induction. Treatment with Vitis vinifera extract along with stress reversed the stress induced biochemical changes i.e., increase in urinary VMA levels and decrease in urinary ascorbic acid levels, in a dose dependent manner. A number of Indian medicinal plants like Ocimum sanctum , Withania somnifera , Panax ginseng etc have been identified for their antistress activity. It was concluded that the antioxidant activity of these plants was partly responsible for their antistress activity [ 23 ]. Based on these reports the antioxidant activity of Vitis vinifera extract was also done using hydroxyl radical assay method. It was found that Vitis vinifera extract has significant good antioxidant activity which was 8 fold more than that of ascorbic acid. The antistress and antioxidant activities were correlated with the nootropic activity of the extract since the role of stress and free radicals have been implicated in the loss of memory, concentration and also in Alzheimer's disease [ 24 , 25 ]. The process of nootropic activity involves acquisition, retention and retrieval and is measured using conditioned avoidance response. The acquisition was quicker in the extract treated rats (100, 200, 300 mg/kg body weight) in comparision to control, indicating the involvement of antistress activity of the extract. When challenged with scopolamine butylbromide (1 mg/kg body weight), the amnesia was less in treated group showing better retention and recovery than control group and the Vitis vinifera extract was shown to decrease memory loss which could be due to its central cholinomimetic activitiy apart from its free radical scavenging mechanisms. Furthermore, the antioxidant activity of the seed extract provide mechanistic basis in relieving stress by way of combating oxidative damage. Conclusion In conclusion, the present study provides scientific support for the antistress (adaptogenic), antioxidant and nootropic activities of V. vinifera seed extract and substantiate the traditional claims for the usage of grape fruits and seeds in stress induced disorders. Further investigations are required to characterize the active constituent(s) responsible for observed activities of the seed extract. Abbreviations VMA: Vanillylmandelic acid CAR: Conditioned avoidance response SBB: Scopolamine butylbromide EDTA: Ethylene diamine tetra acetic acid H 2 O 2 : Hydrogen peroxide Competing interests The author(s) declare that they have no competing interests. Authors' contributions SS conceived the study, made substantial contributions in data analysis, data interpretation, writing of the manuscript and in coordination of the experiments. SN made substantial contributions in conceptualization of statistical analyses, drafting the final manuscript and designing the illustrations. RK and SK helped to conceptualize the work. KMB made significant contribution in designing the studies, conducting the experiments, interpretation of the data, and drafting the final manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC547917.xml
516033
Recurrent Barrett's esophagus and adenocarcinoma after esophagectomy
Background Esophagectomy is considered the gold standard for the treatment of high-grade dysplasia in Barrett's esophagus (BE) and for noninvasive adenocarcinoma (ACA) of the distal esophagus. If all of the metaplastic epithelium is removed, the patient is considered "cured". Despite this, BE has been reported in patients who have previously undergone esophagectomy. It is often debated whether this is "new" BE or the result of an esophagectomy that did not include a sufficiently proximal margin. Our aim was to determine if BE recurred in esophagectomy patients where the entire segment of BE had been removed. Methods Records were searched for patients who had undergone esophagectomy for cure at our institution. Records were reviewed for surgical, endoscopic, and histopathologic findings. The patients in whom we have endoscopic follow-up are the subjects of this report. Results Since 1995, 45 patients have undergone esophagectomy for cure for Barrett's dysplasia or localized ACA. Thirty-six of these 45 patients underwent endoscopy after surgery including 8/45 patients (18%) with recurrent Barrett's metaplasia or neoplasia after curative resection. Conclusion Recurrent Barrett's esophagus or adenocarcinoma after esophagectomy was common in our patients who underwent at least one endoscopy after surgery. This appears to represent the development of metachronous disease after complete resection of esophageal disease. Half of these patients have required subsequent treatment thus far, either repeat surgery or photodynamic therapy. These results support the use of endoscopic surveillance in patients who have undergone "curative" esophagectomy for Barrett's dysplasia or localized cancer.
Background The incidence of esophageal adenocarcinoma has increased more rapidly than any other form of cancer since the 1970s and now represents the majority of esophageal neoplasms in the West [ 1 ]. Barrett's esophagus is the replacement of native squamous mucosa by specialized intestinal metaplasia and is known to be the major risk factor for the development of adenocarcinoma via the metaplasia-dysplasia-neoplasia sequence [ 2 ]. Other risk factors for the development of esophageal adenocarcinoma include a lengthy and severe history of gastroesophageal reflux disease (GERD), increased body mass index, male gender and Caucasian race [ 3 - 5 ]. Recent studies, however, have detected Barrett's esophagus in nearly equal numbers of older white men whether or not they reported GERD symptoms [ 6 , 7 ]. Recommendations, therefore, regarding screening and surveillance of patients at risk for esophageal adenocarcinoma are controversial [ 2 , 8 ]. Detection of dysplastic Barrett's esophagus or mucosal adenocarcinoma is important because it allows the opportunity to intervene prior to the development of invasive neoplasia. Unfortunately, no medical or surgical GERD treatment has been consistently and convincingly demonstrated to prevent the development of adenocarcinoma [ 5 ]. Traditionally, high grade Barrett's dysplasia and mucosal adenocarcinoma are treated by surgical resection of the esophagus [ 9 ]. As endoscopic methods of screening and surveillance have become more widely applied, curative esophagectomy has produced increasing numbers of long-term survivors [ 10 , 11 ]. Coincident with this trend has been the appearance of several reports describing the development of metachronous esophageal adenocarcinoma after undergoing curative esophagectomy for the primary tumor [ 12 , 13 ]. These reports raise questions regarding the role of endoscopic surveillance in these patients. The aim of the current study was to evaluate patients who had undergone complete, presumably curative, esophageal resection for early Barrett's adenocarcinoma or high grade dysplasia in order to determine how frequently, and over what time period, they developed recurrent Barrett's esophagus or adenocarcinoma. Methods After approval by the Mayo Foundation's Institutional Review Board for Research, the electronic medical records of Mayo Clinic patients in Jacksonville, Florida, were searched to find all patients who had undergone esophagectomy for cure at the Mayo Clinic surgical facility, St. Luke's Hospital, Jacksonville, Florida, since 1995. This time period was chosen to coincide with the routine availability and clinical use of pre-operative staging with endosonography in our institution. The records of these patients were reviewed for pre-operative and post-operative staging results including computed tomography and endosonography studies. In addition, endoscopic, surgical studies and histopathological studies were studied. Specifically, the surgical specimens were reviewed to ensure that the esophagectomy specimen, including lymph node sampling, was adequate and the proximal margin was completely free of Barrett's metaplasia, dysplasia or carcinoma. The patients, in whom we have at least one follow-up endoscopy study, with biopsies obtained for histologic confirmation of mucosal disease, are the subjects of this report. Esophageal disease was staged according to the Tumor-Lymph node-Metastasis (TNM) criteria [ 14 ]. Results Since 1995, 45 patients have undergone esophagectomy for Barrett's dysplasia or localized adenocarcinoma with curative intent in our institution. At operation, none of these patients were found to have extension of malignant disease to paraesophageal lymph nodes and all esophageal glandular mucosa was resected with only normal squamous mucosa remaining at the proximal surgical margin. Subsequently, 36 of these patients (80%) have undergone endoscopy after surgery including 8/45 patients (18%) who were found to have recurrent Barrett's glandular mucosa after curative resection and are described in the table. Open transthoracic esophagectomy (Ivor Lewis procedure) with pyloroplasty was performed in most patients (39/45) including the patients diagnosed with recurrent Barrett's disease. Five different surgeons performed these operations. Most patients had evidence of gastric stasis (retained food) at endoscopy (30/36; 83%). Anastomotic dilation was performed at endoscopy in 16/36 patients (range of dilation procedures 1–10; median 3). It is possible that patients with anastomotic strictures may be at increased risk of recurrent Barrett's esophagus because of worse reflux although their swallowing symptoms may, alternatively, be related to other factors such as anastomotic ischemia or surgical sutures. Patients frequently used aspirin (42%) or COX-2 specific non-steroidal anti-inflammatory agents (25%). Twice-daily proton pump inhibitors were routinely prescribed for these patients although patient compliance is difficult to assess because of high drug costs and limited symptomatic improvement. While the small number of patients limits our analysis, these factors were found to occur in a proportional number of patients with Barrett's disease and no clear trends could be identified. Five of these patients have been diagnosed with Barrett's metaplasia or low-grade dysplasia have been followed for more than 12 months in surveillance endoscopy programs monitoring the stability of the glandular epithelium. Two of these 5 patients have been found to have short segment Barrett's metaplasia with lengths of 10 mm and 15 mm, after complete esophageal resection for Barrett's high-grade dysplasia in a 72-year-old man and Barrett's adenocarcinoma T 2 N 0 M 0 in a 78-year-old man. This metaplastic glandular epithelium was detected at a follow up of 90 months and 17 months, respectively. In the other 3 patients, short segment Barrett's low-grade dysplasia has been found in lengths between 10–25 mm after complete resection of the esophagus for Barrett's high-grade dysplasia (1 patient) and Barrett's T 3 N 0 M 0 carcinoma (2 patients). This recurrent Barrett's glandular dysplasia was detected at a follow up of 42–47 months. Erosive esophagitis was also noted in 4 of 5 patients indicating uncontrolled reflux disease. Subsequently all patients have been treated with high doses of proton pump inhibitors (such as esomeprazole 80 mg twice a day or 40 mg three or four times per day) in an attempt to maximally control reflux of acid and digestive juices from the stomach into the cervical esophagus. Three other patients developed recurrent Barrett's disease after curative resection of esophageal T 2 or T 3 N 0 M 0 adenocarcinoma. These patients varied in age from 58–80 years of age. These patients were found to have more severe erosive esophagitis suggesting worse acid reflux and mucosal injury compared to the non-carcinoma recurrent Barrett's patients. Recurrent Barrett's multi-focal high-grade dysplasia, over a 10 mm segment length was detected 88 months after esophagectomy in one patient and was successfully ablated with porfimer sodium photodynamic therapy using the methods described elsewhere [ 15 ]. High-grade dysplasia with features of invasive adenocarcinoma was noted in a 20 mm Barrett's segment diagnosed 18 months after esophageal resection in another patient. Endosonography detected focal mucosal expansion and the Barrett's mucosal adenocarcinoma T 1 N 0 M 0 was confirmed at repeat esophagectomy. Finally, a diminutive polypoid mass proximal to the surgical anastomosis was found in a 58-year-old woman who had 7 months previously undergone esophagectomy for Barrett's mucosal adenocarcinoma. Computed tomography with contrast enhancement noted esophageal wall thickening and suspicious lymphadenopathy. Repeat resection confirmed the tumor histologic grade of T 2 N 1 M 0 adenocarcinoma. Discussion Over the past four decades, the incidence of esophageal adenocarcinoma has risen dramatically, particularly in older white men [ 16 ]. Previous studies have documented the association of esophageal cancer with the severity and duration of gastroesophageal reflux [ 3 ]. The most important risk factor for the development of esophageal adenocarcinoma, however, is the development of specialized intestinal metaplasia in the lower esophagus (Barrett's esophagus) [ 5 ]. It is recommended that patients with Barrett's mucosa undergo periodic surveillance endoscopy with systematic biopsies to detect the presence of cancerous or pre-cancerous changes (dysplasia or neoplasia) [ 2 ]. In these patients, especially, Barrett's adenocarcinoma or high-grade dysplasia is found at an early stage when it is increasingly possible to undergo esophagectomy with curative intent [ 17 ]. After undergoing esophageal resection, the native squamous mucosa of the cervical esophagus will be brought into contact with the acid-secreting mucosa of the gastric body. This reconstruction allows acid and duodenal juice to reflux from the gastric conduit to the remaining cervical esophagus. Reflux of gastric and duodenal content is an important factor in the pathogenesis of Barrett's metaplasia, dysplasia and esophageal adenocarcinoma [ 18 - 20 ]. It is well known that recurrent Barrett's glandular mucosa is frequently found in the cervical esophagus after esophageal resection [ 21 - 23 ]. Recently, Oberg et al performed endoscopy, esophageal manometry and ambulatory pH studies in 32 patients who had undergone esophagectomy for a variety of diagnoses including 16 patients (50%) with adenocarcinoma associated with Barrett's metaplasia [ 24 ]. These studies were performed between 3–10 years after the surgery and most of the patients (70%) were receiving at least once daily proton pump inhibitor therapy for chronic reflux symptoms. At endoscopy, Barrett's glandular mucosa was histologically documented in 15 patients (47%) ranging in segment lengths from 0.5–4.0 cm. Importantly, recurrent Barrett's glandular mucosa was significantly more likely to occur in patients with a preoperative diagnosis of Barrett's epithelium suggesting that the esophageal mucosa of these patients may be pathogenetically predisposed to develop metaplastic changes in response to gastroesophageal reflux. In the study of Oberg et al, despite the use of potent acid-suppressing medications, severe esophageal acid exposure was noted in most patients. Patients with recurrent Barrett's epithelium were found to have significantly more severe acid exposure that occurred predominantly in the supine position [ 25 ]. There was also a direct correlation between the length of the metaplastic segment and the severity of acid reflux. Similar findings were reported by da Rocha, et al., who studied 48 patients after esophagectomy where 4 of these patents (8%) were found to have pathological changes of Barrett's metaplasia in the cervical esophageal stump [ 26 ]. In both of these studies, the authors concluded that the finding of recurrent Barrett's esophagus was related to reflux esophagitis that resulted from the action of acid-peptic and biliary secretions. These studies, however, did not detect dysplasia or neoplasia and did not address the issue of metachronous cancers and the role of endoscopic surveillance for these patients. Murata and colleagues recently reported the diagnosis of metachronous squamous cell carcinomas in five of 253 patients (2%) who had undergone esophagectomy for thoracic esophageal squamous cell carcinoma more than two years previously. These superficial carcinomas (T is or T 1 ) were detected at surveillance endoscopy and were treated with endoscopic laser ablation, mucosal resection or surgical resection. While squamous cell carcinomas are not related to gastroesophageal reflux, this paper also suggests that esophageal cancer patients (squamous or adenocarcinoma) are predisposed to the development of metachronous carcinomas in the remnant cervical esophagus. This is consistent with DeMeester's experimental model of Barrett's dysplasia and adenocarcinoma occurring after complete gastrectomy with esophago-jejunostomy and reflux of bile and digestive enzymes into the cervical esophagus [ 27 ]. Also, it is likely that these patients had other important risk factors, such as tobacco and alcohol use, that predispose to both squamous and adenocarcinomas [ 5 ]. Konishi et al reported finding an adenocarcinoma in Barrett's esophagus following a total resection of the gastric remnant in a 52-year-old man who had undergone distal gastrectomy for gastric cancer nearly twenty years previously [ 28 ]. The Barrett's mucosa associated with the adenocarcinoma contained high grade dysplasia supporting the acquired theory of pathogenesis for Barrett's esophagus that suggests that reflux esophagitis after gastrectomy may result in the dysplasia-carcinoma sequence. In addition, Streitz et al retrospectively reviewed long-term survivors after esophagectomy for adenocarcinoma [ 13 ]. With a follow-up as long as 14 years, they found 4 patients who subsequently were diagnosed with esophageal adenocarcinoma. However, the time period between the development of the first and second tumors was not specified making it not possible to determine if these were recurrent tumors or new, metachronous lesions. Finally, Riben et al have reported the development of a secondary Barrett's adenocarcinoma in a patient who had 19 years previously undergone esophagectomy for a stage IIb Barrett's adenocarcinoma [ 12 ]. These studies have demonstrated that the cervical esophagus is exposed to high amounts of acid and refluxate despite the use of proton inhibitor medications and often in the absence of severe reflux symptoms. Although our group of patients has been observed for only a median of 2 years after esophagectomy, our study confirms that the development of metaplastic columnar mucosa in the cervical esophagus is a common complication related to reflux associated injury to the squamous epithelium. Further, our findings suggest that this recurrent glandular mucosa is unstable and predisposed to the development of dysplasia and invasive carcinoma, as has already developed in most of patients. The early detection of this recurrent disease remains vitally important to preserve all possible treatment options including surveillance endoscopy follow-up, endoscopic ablation with porfimer sodium photodynamic therapy, and if necessary repeat esophagus resection surgery. Our specific recommendations include surveillance endoscopy every 6–12 months for patients who have undergone "curative" esophagectomy for Barrett's dysplasia or adenocarcinoma. In addition, we also routinely recommend indefinite use of proton pump inhibitors, regardless of symptom status, starting at twice daily dosing and increasing as necessary to control reflux symptoms and mucosal damage due to acid, bile and digestive enzymes. Whether these drug doses should be titrated based on ambulatory pH and impedance test results remains to be determined. We have generally been disappointed by prokinetic agents such as metoclopromide in improving reflux symptoms in these patients. For esophagectomy patients who develop recurrent Barrett's metaplasia we recommend the use of COX-2 inhibitors or aspirin chemoprevention to protect against the development of metachronous Barrett's carcinoma. [ 29 , 30 ]. Competing interests None declared. Author's contributions All authors participated in the study design and coordination as well as case collection and review of histopathologic and endoscopic results. All authors read and approved the final manuscript. Table 1 Recurrent Barrett's Disease after Esophagectomy with Curative Intent Pre-operative diagnosis Sex Age F/u diagnosis Time to F/U Tx Barrett's T3 N0 ACA F 58 Barrett's T2 N1 7 mos Surgery Barrett's T3 N0 ACA F 64 Barrett's LGD 42 mos PPI Barrett's T3 N0 ACA F 64 Barrett's LGD 42 mos PPI Barrett's HGD M 64 Barrett's LGD 47 mos PPI Barrett's HGD M 72 Barrett's Metaplasia 90 mos PPI Barrett's T3 N0 ACA M 69 Barrett's T1 N0 18 mos Surgery Barrett's T2 N0 ACA M 78 Barrett's metaplasia 17 mos PPI Barrett's T2 N0 ACA M 80 BE+HGD 88 mos PDT ACA = Adenocarcinoma HGD = High-grade dysplasia LGD = Low-grade dysplasia PDT = Photodynamic therapy PPI = Proton pump inhibitor medical therapy Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516033.xml
516769
Gene expression variation between mouse inbred strains
Background In this study, we investigated the effect of genetic background on expression profiles. We analysed the transcriptome of mouse hindlimb muscle of five frequently used mouse inbred strains using spotted oligonucleotide microarrays. Results Through ANOVA analysis with a false discovery rate of 10%, we show that 1.4% of the analysed genes is significantly differentially expressed between these mouse strains. Differential expression of several of these genes has been confirmed by quantitative RT-PCR. The number of genes affected by genetic background is approximately ten-fold lower than the number of differentially expressed genes caused by a dystrophic genetic defect. Conclusions We conclude that evaluation of the effect of background on gene expression profiles in the tissue under study is an effective and sensible approach when comparing expression patterns in animal models with heterogeneous genetic backgrounds. Genes affected by the genetic background can be excluded in subsequent analyses of the disease-related changes in expression profiles. This is often a more effective strategy than backcrossing and inbreeding to obtain isogenic backgrounds.
Background Due to their isogenicity, inbred mouse strains demonstrate low biological variability within each strain[ 1 , 2 ]. Genetic variation between inbred strains is considerable and has recently been characterized in detail using single nucleotide polymorphisms[ 3 ]. Differences in genetic background between strains affect the gene expression levels of a subset of genes, which probably explains phenotypic differences. Indeed, several reports have been published in which gene expression profiles have been used as QTLs in genetic mapping studies to identify complex traits [ 4 - 6 ]. From literature [ 7 - 9 ], it appears that the subset of genes for which expression is significantly affected by genetic background is small. However, this has never been related to the extent of gene expression changes observed due to disease-causing mutations. We are studying differential gene expression between affected and healthy muscle in a range of murine models for neuromuscular disorders with different genetic backgrounds (Turk et al ., manuscript in preparation). We, therefore, determined gene expression levels in hindlimb muscles from five frequently used wildtype mouse inbred strains, and compared these to the differential gene expression levels in affected muscle tissue from a mouse model ( mdx ) for Duchenne muscular dystrophy with healthy muscle tissue. Both the number of differentially expressed genes between strains as well as the fold-change levels are lower when compared to the differences found in affected versus healthy muscle tissue. Results Gene expression levels in hindlimb muscle tissue from five different inbred strains (CBA, BALB, BL6, DBA, and BL10) were determined. Total RNA from two individuals per strain was isolated, reversed transcribed, and subsequently labelled according to a recently developed protocol (adapted from Xiang et al. , 2002), which requires an input of only 1 μg total RNA. Labelled cDNA was hybridised to murine microarrays containing 7,776 65-mer oligonucleotides spotted in duplicate. Significance levels (p-values) between the five mouse inbred strains were calculated using analysis of variance[ 10 ]. Significance levels among two individual mice within each strain were determined using a hierarchical t -test providing higher statistical power than conservative methods for low (2–4) replicate numbers[ 11 ]. The higher power is yielded by borrowing information across genes to produce a better expression variance estimator. The gain in power is reported via an increase in the degrees of freedom associated with the t-test. Differentially expressed genes for both computations were selected by controlling the false discovery rate (FDR), as suggested by Benjamini and Hochberg (1995), rather than using pre-defined cut-offs for p-values or corrections for multiple testing. The FDR represents an expectation of the proportion of false positives among the selected differentially expressed genes, which increases dramatically during multiple testing, inherent in microarray experiments[ 12 ]. Using an FDR of 10% we selected 88 out of 6144 (1.4%) expressed genes that are differentially expressed between strains (Fig. 1 ). A lower number of differentially expressed genes was found in the analysis of variation within strains with identical FDRs of 10% (Table 1 ). Results with other FDR levels are available online as additional file. Correlation between gene expression levels of the two samples from each strain was high (Pearson correlation coefficient ranging from 0.87 to 0.95), also indicating low internal variation (Table 1 ). A considerable amount of differentially expressed genes (718 genes) were selected when pre-defined cut-off values (p < 0.05) were used to determine the differential gene expression between strains. However, adjusted FDR levels indicated a proportion of false positives equal to 42%. On the other hand, adjusting for multiple testing using Bonferroni correction proved to be too stringent, leaving no or few differentially expressed genes. Controlling the FDR, therefore, appears to be an optimal method for both selecting differential gene expression and simultaneously determining the validity of the experimental outcome. Figure 1 Differentially expressed genes between mouse inbred strains Relative expression levels of differentially expressed genes between mouse inbred strains are depicted in colour as relative intensity levels. Shown for each gene are GenBank accession number, description, functional annotation according to Gene Ontology, and UniGene cluster IDs. Relative expression levels are calculated by subtracting the average intensity value per gene from the strain-dependent intensity values. Differential expression was determined by selecting p-values from analysis of variance based on a false discovery rate of 10%. Table 1 Number of differentially expressed genes using several cut-off strategies Between strains Within strains MA-ANOVA Hierarchical t -test CBA BL10 BL6 DBA BALB Correlation 0.95 0.95 0.87 0.87 0.92 Naive (p < 0.05) 718 737 610 963 1043 483 Bonferroni 0 2 4 1 0 3 FDR 10% 88 2 4 14 0 16 Correlation between two individuals per strain was calculated using Pearson's correlation coefficient. Significance levels (p-values) between strains were calculated with MA-ANOVA, and within strains using the hierarchical t -test. Differential gene expression was determined by selecting genes with p-values lower than a specified threshold. Thresholds were selected using three different strategies; naive, Bonferroni corrected, and False Discovery Rate (10%), and resulted in different numbers of significantly differentially expressed genes. To put the influence of differential gene expression due to genetic background in perspective, we studied gene expression between affected and healthy tissue from hindlimb muscle derived from mdx mice, and from control mice with identical genetic backgrounds. Selection with an FDR of 10% resulted in 1298 differentially expressed genes. Differential gene expression between the two most divergent mouse inbred strains (BL6 and CBA, data not shown) was determined to allow a direct comparison with identical statistical methods. Selection with an FDR of 10% showed an approximately ten-fold decrease in the number of differentially expressed genes (126). Absolute fold changes were calculated and subsequently a comparison of the distribution was made (Fig. 2 ). Median gene expression levels are equal between affected/control and inbred/inbred. However, the number of large fold changes (>3) between affected/healthy (221) is much higher than between inbred/inbred (7), consistent with low contribution of differential expression due to genetic background. Figure 2 Effect of different genetic background on differential gene expression The distribution of absolute fold changes of differentially expressed genes (n = 1298) between affected ( mdx ) and healthy (WT) muscle were compared to the distribution of absolute fold changes of differentially expressed genes (n = 126) between two mouse inbred strains (CBA and BL6). Selections were based on a FDR of 10%. Although overall expression levels are similar between strains, a relatively high number of differentially expressed genes was due to deviating gene expression levels in BL6. We performed quantitative real-time RT-PCR (qPCR) on five genes to verify our microarray data. Two genes myomesin 1 and tropomodulin 1, which were 2.2-fold and 1.8-fold lower expressed in BL6 compared to the other strains on our microarrays, were also found to be lower expressed (2.0-fold and 2.2-fold respectively) in our qPCR assay (Fig. 3 ). Three other genes (dysferlin, cystatin B, and thrombospondin 4) showed no differential expression between any strains. Figure 3 Validation of BL6-dependent gene expression with qPCR Relative gene expression levels between mouse inbred strains of tropomodulin 1 ( Tmod1 ) and myomesin 1 ( Myom1 ) as determined by quantitative RT-PCR. Significantly lower expression (p < 0.01, marked by *) for both genes was shown in BL6 compared to other strains. Discussion This study shows that variation in overall gene expression levels between mouse inbred strains is relatively low in hindlimb muscle tissue. This is particularly evident when the number of differentially expressed genes between two mouse inbred strains (C57 vs. Bl6, 126 genes with 7 genes having a fold-change > 3) is compared to that between diseased and healthy muscle tissue ( mdx vs. wild-type, 1298 genes with 221 genes having a fold-change >3). Therefore, the use of mice with deviating genetic background may be justified in disease-related studies. Alternatively, strain-dependent gene expression differences may be evaluated in the initial study phase of gene targeting experiments, although the effect of hybrid backgrounds is difficult to assess. Gene expression studies in the brain revealed that approximately 1% of expressed genes differ between two mouse strains[ 8 ]. Application of alternative statistical methods, similar to those used in our study, on this dataset resulted in an increase in the number of differentially expressed genes (approx. 3%) between the two mouse strains[ 7 ], demonstrating that the number of differentially expressed genes is highly dependent on the statistical criteria used. A similar number of differentially expressed genes was found in a comparison of hippocampal gene expression between 8 different mouse strains[ 9 ]. The results of our study in muscle tissue demonstrated that approximately 1.4% of the expressed genes show differential expression between mouse strains. Based on these results, strain differences in gene expression seem to have a similar magnitude across different tissues. Genomic variability could be correlated with high levels of single nucleotide polymorphisms (SNPs) occurring in specific blocks between mouse inbred strains. The presence of cis-acting (single nucleotide) polymorphisms may be associated with regulatory variation affecting gene expression levels. It was estimated that probably a consistent amount (up to 6%) of the roughly estimated 35,000 mouse genes contain such functional regulatory variants[ 13 ]. We investigated if differentially expressed genes were localized in blocks with high genomic variability, but our number of differentially expressed genes was too low to obtain statistically significant answers (data not shown). This study suggests an additional method for phenotyping mouse inbred strains and provides a list of genes with significant differential expression based upon false discovery rate selection. Although overall gene expression profiles are highly similar, most significant differences are determined by low gene expression levels of BL6 compared to the other strains. A large proportion of these BL6-specific genes function as structural muscle proteins (i.e. nebulin, alpha-actinin 2, myomesin 1 and radixin). To date, however, no major differences in muscle physiology in BL6-mice have been described which can be attributed to these reduced gene expression levels. Perfectly isogenic backgrounds are sometimes difficult to obtain. This explorative study demonstrates that the effect of genetic background on muscle expression profiles is significant but rather limited compared to other effects, e.g. the dystrophic genetic defect ( mdx ) we study. As such, the genetic background will only marginally interfere with data analysis. Determination of gene expression profiles between mouse strains enables flagging a modest number of differentially expressed genes, and is an efficient and sensible approach to circumvent tedious backcrossings, necessary to obtain isogenic animals. Methods Mouse breeding, tissue preparation and total RNA isolation We obtained CBA/CaOlaHsd (CBA), BALB/cOlaHsd (BALB), C57Bl/6JOlaHsd (BL6), DBA/2OlaHsd (DBA), and C57Bl/10ScSnOlaHsd (BL10) mice from Harland Laboratories, and C57Bl/10ScSn-Dmd mdx/J ( mdx ) mice from Jackson Laboratory at the age of 6 weeks. Mice were kept under standard conditions and were sacrificed by cervical dislocation when 8 weeks old. Hindlimb muscles (m. quadriceps femoris) were dissected and promptly snap-frozen in isopentane at -80°C. Total RNA was prepared by disrupting tissue using mortar and pestle and subsequent homogenisation by a rotor-stator homogenizor (Ultra-Turrax T25, Janke & Kunkel IKA-Labortechnik) in RNA-Bee (Campro Scientific) until uniformly homogenous (15–45 sec). Total RNA was isolated according to manufacturer's instructions followed by purification using RN-easy columns (Qiagen). Quality and yield was determined using Lab-on-a-chip (BioAnalyzer, Agilent). Target preparation and hybridisation Aminoallyl labelled cDNA (aa-cDNA) was prepared based on a previously described protocol[ 14 ]. Aliquots of 1 μg of total RNA in the presence of 2 μg amino-TN 6 primer (5'-NH 2 -(CH 2 ) 6 -TN 6 , Eurogentec) were adjusted to a volume of 21 μl with DEPC-treated H 2 O (diethyl pyrocarbonate, Sigma), heated for 10 minutes at 70°C and chilled on ice for 10 minutes. Reverse transcription mastermix (1.8 μl RevertAid RNaseH-M-MuLV reverse transcriptase (200 U/μl, MBI Fermentas), 6 μl 5x first-strand buffer (MBI Fermentas), and 1.2 μl 25x aa-dUTP / dNTP solution (2 μl 50 mM dATP, 2 μl 50 mM dCTP, 2 μl 50 mM dGTP, 1.2 μl 50 mM dTTP, 0.8 μl 50 mM aminoallyl-dUTP (Ambion)) was added per reaction and incubated at room temperature for 10 minutes followed by 2 hours at 42°C. RNA was hydrolysed by addition of 10 μl 0.5 M EDTA and 10 μl 1 M NaOH and incubation at 65°C for 30 minutes followed by neutralization by addition of 10 μl 1 M HCl. Aminoallyl labelled cDNA was then purified by combining 300 μl of PB-buffer (Qiagen) to 60 μl of the neutralized sample and centrifuged through a Qiaquick column (Qiagen) at 13000 rpm for 1 minute. Two washing steps were performed by spinning 500 μl of 75% EtOH at 13000 rpm for 1 minute while discarding the flow-through. To remove ethanol-traces the columns were centrifuged for an additional minute. cDNA was recovered by eluting three times using 30 μl basic H 2 O (3.3 mM NaHCO 3 buffer, pH 9.0) and concentrated to a volume of 6.66 μl using a speedvac. Aliquots of Cy3 and Cy5 reactive dyes (PA23001, PA25001, Amersham) were prepared by dissolving each vial of monoreactive dye in 40 μl fresh anhydrous DMSO (Sigma) and dividing into aliquots of 2 μl followed by vaccuumdrying until dry and subsequent storage at 4°C in the presence of silica. Fluorescent dyes were coupled by adding 3.33 μl of bicarbonate buffer (1 M NaHCO 3 buffer, pH 9.0) to the aa-cDNA sample and dissolving the dried aliquot of reactive dye, followed by incubation at room temperature for 1 hour in the dark. To the samples 4.5 μl 4 M hydroxylamine (Sigma) was added and incubated at room temperature in the dark for 15 minutes, followed by addition of 186 μl TE -3 -buffer. Hybridisation mixtures were prepared by combining a Cy3-labeled cDNA sample with a Cy5-labeled cDNA sample and 10 μl Mouse-Hybloc (1 μg/μl, Applied Genetics Laboratories) followed by removing uncoupled dyes by spinning through a pre-wetted Microcon column (YM30, Amicon) for 8 minutes at 13000 rpm. Hybridisation mixture was washed by spinning 500 μl TE -3 -buffer through the column and discarding the flow-through. This step was repeated two times as 2 μl yeast-tRNA (10 μg/μl, Sigma) and 2 μl polyA-RNA (10 μg/μl, Sigma) were added during the last step. Mixture was collected by inverting the column and spinning for 1 minute at 13000 rpm. Hybridisation mixture was finalized by adding TE -3 -buffer to 84 μl together with 17 μl 20x SSC and 3 μl 10% SDS followed by denaturing at 100°C for 2 minutes, renaturing at room temperature for 15 minutes and spinning at 13000 rpm for 10 minutes. Labelled target was hybridised overnight on murine oligonucleotide microarrays (65-mer with 5'-hexylaminolinker, Sigma-Genosys mouse 7.5 K oligonucleotide library, spotted in duplicate). Hybridisation occurred in a automatic hybridisation station (GeneTac, Perkin Elmer) and was followed by washing with 5x 2xSSC + 0.1% SDS at 30°C, 5x 1xSSC at 30°C, 3x 0.2xSSC at 30°C, 1x 0.2xSSC at 65°C, 2x 0.2xSSC at 30°C, and subsequently scanned as described previously[ 15 ]. Experimental design, data extraction and analysis Gene expression profiles from hindlimb muscle derived from 2 male animals of each strain were generated using dye-swap experiments. Subsequent duplicate spots on each array resulted in 8 replicate measurements per gene. Targets were assigned at random to the arrays, while avoiding co-hybridisation of samples from the same strain. GenePix Pro 3.0 (Axon) was used for feature extraction and quantification. Genes were considered as being expressed when the corresponding feature was not flagged by the algorithm provided by GenePix. Local background corrected spot intensities were normalized using Variance Stabilization and Normalization (VSN) in R [ 16 ]. Array data has been made available through the GEO data repository of the National Center for Biotechnology Information under series GSE662. Correlation between individuals was calculated using Pearson's correlation coefficient. Significantly differential expression levels were determined using MA-ANOVA ( MAANOVA2.0 The Jackson Laboratory ), hierarchical t -test [ 11 ] and the False Discovery Rate [ 17 ] selection procedure. Quantitative Reverse Transcription Polymerase Chain Reaction qPCR was performed in duplicate for each individual resulting in four measurements per strain per gene. cDNA was prepared by reverse transcription using 1 μg total RNA as template. Random hexamers (40 ng) were used to prime the transcription after heating 10 minutes at 70°C followed by chilling on ice for 10 minutes. cDNA was synthesized by RevertAid RNaseH - MuLV reverse transcriptase and accompanying buffer (MBI-Fermentas) using 1 mM dNTPs. The mixture was incubated at room temperature for 10 minutes before a 2 hour incubation step at 42°C, followed by 10 minutes at 70°C. Quantitative PCR was performed using the Lightcycler (Roche). PCR mixture was prepared by combining cDNA dilution, 10 pmol forward and reverse primer, MgCl 2 (4 mM) with 4x home-made LC mastermix (0.9 mM dNTPs, BSA (1 μl/μl, Pharmacia Biotech), Taq polymerase (0.8 U/μl), 4x SYBR Green I (Molecular Probes), 4x AmpliTaq Reaction Buffer (Perkin Elmer)) to a total volume of 20 μl. Amplicons were generated during 45 cycles with annealing temperature set at 55°C. Optimal cDNA dilutions and relative concentrations were determined using a dilution series per gene. Replicate experiments (n = 4) were normalized to 1 and relative expression values were determined by calculating the ratio per gene over the average relative expression of genes, which show no differential expression on both microarray and qPCR (dysferlin, cystatin B, and thrombospondin 4). Significance levels were calculated with a one-sample t -test. PCR primer pairs were designed using the Primer3 search engine, available at: Primer3 Software Distribution . The screened genes and the oligonucleotide primer pairs used for each of the genes in this study correspond to the following nucleotides: myomesin1, 4761–4780 and 4865–4884 (NM_010867); tropomodulin1, 670–689 and 878–897 (NM_021883); dysferlin, 4218–4237 and 4353–4372 (AF188290); cystatinB, 3–22 and 151–170 (NM_007793); thrombospondin4, 2167–2186 and 2289–2308 (NM_011582). Authors' contributions RT carried out the tissue preparation, total RNA isolation, target preparation, hybridisations, experimental design, data extraction, data analysis, rt-PCR, and the drafting of the manuscript. PH participated in the experimental design, analysis, rt-PCR, and study coordination. ES participated in the experimental design and analysis. RM provided statistical support. EM was responsible for mouse breeding and tissue preparation. JM participated in experimental design. GO and JD coordinated the study. All authors read the final manuscript. Supplementary Material Additional File 1 Differentially expressed genes between mouse inbred strains are selected with a false discovery rate of 10, 15, and 20%. Selected genes are indicated with 1, genes not selected by the specified criteria are indicated with 0. The mean of the relative gene expression levels of each of the five mouse strains is shown. For each gene the GenBank accession number is shown as well as the UniGene ID, gene description and the gene ontology description. The additional file is formatted as Comma Separated Values (CSV) file, and is named Turketal2004_Additional_File.csv. Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516769.xml
544588
The Depression Network (DeNT) Study: methodology and sociodemographic characteristics of the first 470 affected sibling pairs from a large multi-site linkage genetic study
Background The Depression Network Study (DeNt) is a multicentre study designed to identify genes and/or loci linked to and/or associated with susceptibility to unipolar depression in Caucasian families. This study presents the method and socio-demographic details of the first 470 affected sibling pairs recruited from 8 different sites in Europe and the United States of America. Methods Probands fulfilling either the Diagnostic and Statistical Manual 4 th edition (DSM-IV) or the International Classification of Diseases 10 th edition (ICD-10) criteria for recurrent unipolar depression of moderate or severe degree and who had at least one similarly affected sibling were eligible for the study. Detailed clinical and psychological assessments were undertaken on all subjects including an interview using the Schedules for Clinical Assessment in Neuropsychiatry. Blood samples were collected from all participants to extract DNA for linkage analysis. Results The different sites used different recruitment strategies depending on local health care organisation but despite this there was remarkable similarity across sites for the subjects recruited. Although the Bonn site had significantly older subjects both for age of onset and age at interview, for the sample as a whole, subjects were interviewed in their mid-40s and had experienced the onset of their recurrent depression in their 20s. Preliminary genome screening was able to include 929 out of the 944 subjects (98.4%) typed at 932 autosomal and 544 X chromosome markers Conclusions This paper describes the methodology and the characteristics of the subjects from the 414 families included in the first wave of genotyping from the multi-site DeNT study. Ultimately the study aims to collect affected sibling pairs from approximately 1200 families.
Background Genetic risk factors are well established for major affective disorders and a recent twin study has suggested that unipolar depression has a stronger genetic influence than was previously thought. McGuffin and colleagues [ 1 ] have estimated that the heritability (i.e. the proportion of liability explained by genetic risk factors) may be over 70% in a clinically ascertained twin sample while a population based twin study resulted in a very similar estimate using a re-test method of assessing lifetime diagnosis [ 2 ]. The majority of studies suggest a relative risk to siblings (λs) of affective disorder is in the region of 3 [ 3 ]. However, a recent study comparing the siblings of unipolar depressives with the siblings of healthy controls using strict definitions of both depression and health found a substantially higher λs of over 9 [ 4 ]. The inheritance of unipolar depression is complex and involves an inter-play of genetic and environmental factors. For unipolar depression these include certain types of severe and threatening life events such as events associated with humiliation or loss [ 5 , 6 ]. Despite an excess of females to males of about 2 to 1 for unipolar depression, the heritability in a clinically ascertained sample was the same in men and women [ 1 ]. Some population based twin studies suggest at least some of the genes conferring liability differ between the sexes [ 7 ] while others do not [ 8 ]. Although it has been suggested that early onset depression is more clearly familial than later onset, this is not supported by a meta-analysis [ 9 ]. The only characteristics of probands associated consistently with higher familiality or heritability are recurrence of episodes and severity of disorder [ 1 , 9 ]. Previous linkage studies of unipolar depression Most previous linkage studies have been carried out in families identified by a bipolar proband and where unipolar and bipolar relatives are frequently grouped together into a broad definition of affective disorders. Most such studies have focussed on multiple affected extended pedigrees on the assumption that there may be a sub-set segregating a gene of major effect. This approach has been successful in complex disorders such as early onset Alzheimer's disease and breast cancer. However, consistent evidence of major gene effects in bipolar disorder has not been forth-coming[ 10 ]. In addition, the unknown mode of inheritance creates inherent difficulties in classic linkage approaches and consequently sib-pair methods are attractive in the study of complex familial disorder. An affected sib pair genome scan study of recurrent depression has now been published suggesting that there is a depression susceptibility locus on chromosome 15q [ 11 ]. Another genome scan focusing on multiply affected families found the strongest evidence for linkage on chromosome 12q [ 12 ]. In addition a genome scan of multiply affected families with alcoholism and in whom some individuals had depression or co-morbid alcoholism and depression found evidence of a depression linked locus on chromosome 1p. Clearly these results require further scrutiny and replication. Methods Subjects Sibling pairs affected with recurrent unipolar depression were recruited from 8 clinical sites: Aarhus, Denmark; Bonn, Germany; Dublin, Ireland; Lausanne, Switzerland; St Louis, USA and London, Cardiff and Birmingham, UK. In addition, where available, parents of the affected sibling pairs were also included in the study. Subjects were identified from psychiatric clinics, hospitals, general medical practices and from volunteers responding to media advertisements. Caucasian subjects over the age of 18 were included if they had experienced 2 or more episodes of unipolar depression of at least moderate severity separated by at least 2 months of remission as defined by the Diagnostic and Statistical Manual 4 th edition operational criteria (DSMIV) [ 13 ] or the International Classification of Diseases 10 th edition operational criteria (ICD10), for unipolar depression [ 14 ]. Probands were all white and of white European parentage. They were included in the study if they had at least one biological sibling, not a monozygotic twin, over the age of 18 years meeting the same diagnostic criteria. Subjects were excluded if either sibling had ever fulfilled criteria for mania, hypomania or schizophrenia. Subjects were also excluded if they experienced psychotic symptoms that were mood incongruent or present when there was no evidence of a mood disturbance. Other exclusion criteria were intravenous drug use with a lifetime diagnosis of dependency; depression occurring solely in relation to alcohol or substance abuse or depression only secondary to medical illness or medication, and a clear diagnosis of bipolar disorder, schizophrenia, schizo-affective disorder or acute or transient psychotic disorders in first or second-degree relatives. Clinical assessment All subjects were interviewed using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) [ 15 , 16 ]. Items of psychopathology in the SCAN interview were rated for presence and severity according to the worst and second worst episodes of depression identified by the subjects. For the purposes of rating severity, subjects were asked to identify within each of these episodes of depression a 4–6 week period when their symptoms were at their worst (peak intensity). The majority of the SCAN items were coded as follows; 0 – indicates absence of the item, 1 – the item was present but to a mild degree or intermittently throughout the peak intensity 4–6 weeks, 2 – item moderately severe and present for more the 50% of the peak intensity period or severe but present for less than 50% of the peak intensity period, 3 – item severe for more than 50% of the peak intensity period. The computerised version of the SCAN2.1 is built on top of the IShell system, which is a computer aided personal interviewing tool produced by the World Health Organisation [ 17 ] and which provides diagnoses according to DSMIV and ICD10 operational definitions. Interviewer training and reliability across sites All interviewers from each site attended a 4-day SCAN training course in the UK. Each site also undertook further inter-rater reliability meetings regularly and annually all interviewers from all sites took part in a joint inter-rater reliability exercise. Ethical approval All sites obtained ethical approval for the DENT study within their own countries and institutions. All study participants gave written informed consent for participation in the study. Self-report questionnaires and other information collected from participants at interview In addition to the SCAN interview all study participants completed the Eysenck Personality Questionnaire [ 18 ] and a detailed family history of psychiatric and physical disorders. For the 6 months prior to the worst episode and 6 months prior to the second worst episode as well as the 6 months prior to interview the Brief Life Events Questionnaire (BLEQ) identified which of 12 types of severe life events had occurred. These were based on the list proposed by Brugha and colleagues [ 19 ] to which childbirth was added. If such an event had occurred the subject was also asked to rate the impact of the event as; very distressing (a score of 3), moderately distressing (scored 2) or not very distressing (scored 1). The BLEQs therefore gave a summated impact score out of 39 for each 6-month time frame. Blood samples At the time of the SCAN interview interviewers obtained 25 ml of whole blood that was collected in 37.5 ml (EDTA containing) monovettes. In addition drops of blood were placed on a Guthrie blood spot card. The blood samples were labelled with a bar code, gently mixed and stored frozen upright in a -20 degree centigrade freezer pending DNA extraction. Phenotypic data analysis All phenotypic information from interviews and questionnaires was coded by assigning a number to each subject, and removing any personal identifying information. The same codes were used on the blood sample tubes using a bar code system. The phenotypic information was first entered on an EXCEL spread sheet after which a data file was created using Statistical Procedures for the Social Sciences (SPSS) version 10 for Windows for the statistical analyses. Results Inter-rater agreement for SCAN interview All the interviewers from all sites took part in a joint inter-rater reliability exercise (in English) involving both audio-taped interviews with study subjects and videotaped interviews with actors, role-playing a depressed subject. Item by item kappa statistics for SCAN items, were calculated comparing each interviewer's ratings against AF's "master" rating. A mean item by item kappa coefficient across all the sites of 0.77 (range 0.63 – 0.89) was obtained indicating a substantial level of inter-rater agreement. Number of subjects, age at interview, age at illness onset and gender by site For inclusion in the first part of the linkage analysis, 944 affected subjects were genotyped from the 8 study sites as follows: Aarhus 48, Birmingham 146, Bonn 110, Cardiff 126, Dublin 154, Lausanne, 56, London 111 and St Louis 193. The age at interview and age of illness onset by gender of the subjects recruited at each site are shown in Table 1 . Table 1 Numbers of male and female subjects, age at interview and age of illness onset by site Site Gender Number Age at interview (SEM) Age at illness onset (SEM) Aarhus Female 30 43.53 (2.1) 21.39 (1.8) Male 18 45.11 (2.0) 24.67 (1.9) Birmingham Female 104 48.12 (1.3) 23.44 (1.1) Male 42 49.24 (2.2) 26.19 (1.7) Bonn Female 86 51.67 (1.3) 28.10 (1.9) Male 24 50.67 (2.9) 24.24 (4.1) Cardiff Female 85 43.55 (1.1) 23.76 (1.3) Male 41 45.12(1.9) 22.92 (1.6) Dublin Female 110 42.86 (1.2) 21.12 (1.0) Male 44 44.43 (0.9) 24.97 (2.1) Lausanne Female 43 48.67 (1.5) 24.69 (2.0) Male 13 42.39 (2.0) 25.3 (2.1) London Female 80 45.04 (1.1) 22.63 (1.4) Male 31 47.29 (2.1) 20.71 (2.0) St Louis Female 132 47.14 (1.0) 18.44 (0.9) Male 61 47.15 (1.6) 17.58 (1.5) SEM = Standard Error of the Mean Mean age at interview for both sexes combined for each site were as follows: Aarhus 44.13 years (standard error of the mean (SEM)1.5), Birmingham 48.44 years (SEM 1.1), Bonn 51.46 years (SEM 1.2), Cardiff 44.06 years (SEM 0.9), Dublin 43.32 years (SEM 1.0), Lausanne 47.21 years (SEM 1.3), London 45.67 years (SEM 1.0), St Louis 47.14 years (SEM 0.9). These mean age differences were statistically significant (Analysis of variance(ANOVA): F = 6.26 degrees of freedom (df) 7, 936, p < 0.001. Tukey Post hoc test: Dublin, Cardiff, Aarhus, London, St Louis, Lausanne, Birmingham < St Louis, Lausanne, Birmingham, Bonn). Mean age at illness onset for both sexes combined per site were as follows: Aarhus 22.67 years (SEM 1.3), Birmingham 24.30 years (SEM 0.8), Bonn 27.28 years (SEM 1.2), Cardiff 23.47 years (SEM 0.9), Dublin 22.27 years (SEM 0.8), Lausanne 24.85 years (SEM 1.4), London 22.08 years (SEM 1.1), St Louis 18.17 years (SEM 0.8). These mean age differences were statistically significant (ANOVA: F = 9.82 df 7, 841 p < 0.001. Tukey Post hoc test: St Louis, London, Dublin, Aarhus, < London, Dublin, Aarhus Cardiff, Birmingham, Lausanne < Cardiff, Birmingham, Lausanne, Bonn). However, there were no significant differences between sites for the numbers of men and women recruited (see Table 1 ) (chi squared test = 6.83 df 7 p = non significant (ns)). Number of probands, siblings and other relatives recruited by site Although study participants were mainly affected proband/sibling pairs, there were a few families where parents were also included. The numbers of probands, siblings and parents recruited per site is shown in Table 2 . Table 2 Number of probands and siblings recruited from each site Site Gender Number of probands Number of siblings Number of parents Aarhus Female 11 17 2 Male 12 6 0 Birmingham Female 42 53 9 Male 18 21 3 Bonn Female 38 43 5 Male 12 12 0 Cardiff Female 41 42 2 Male 17 23 1 Dublin Female 54 51 5 Male 14 28 2 Lausanne Female 18 25 0 Male 10 3 0 London Female 37 41 2 Male 13 18 0 St Louis Female 54 63 15 Male 21 33 7 In total there were 369 families with 2 affected siblings, 36 families with 3 affected siblings, 7 families with 4 affected siblings, and 2 families with 5 affected siblings. In addition 53 parents were also interviewed and provided blood for DNA extraction. Thus there were 470 affected sibling pairs (calculated as number of pairs per family equals number of affected siblings minus 1). Gender, age at interview, age at illness onset and marital status for all sites combined Of the 944 subjects, 670 (71%) were female and 274 (29%) were males and hence, the female/male ratio was 2.45:1. Mean age at interview for all female subjects was 45.40 years (SEM 0.5) and for all males subjects was 45.69 (SEM 0.8). There were no significant gender differences for age at interview (t = -0.33, df = 477.58, p = ns) The mean age of illness onset for depressed male subjects was 22.61 years (SEM 0.7) compared to 22.52 years (SEM 0.4) for depressed female subjects. There was no significant sex difference for age of onset.(t = -0.11, df = 443.55 p = ns). Fifty five percent of male subjects and 61 % female subjects were living with a partner (married or cohabiting), while 45 % male subjects and 39% female subjects were living alone (ie separated, widowed, divorced or never married). Female subjects were significantly more likely to be living with a partner compared to male subjects. (Chi squared test = 26.89 df = 1 p < 0.001). Gender, age at interview, age of illness onset and marital status for probands, siblings and parents There were 295 female and 117 male probands, 335 female and 144 male siblings and 40 female and 13 male parents included in the total sample. There were no significantly differences for the gender of probands, siblings or parents (chi squared test = 0.85, df = 2 p = ns). The mean age at interview for probands was 45.94 years (SEM 0.6) and for siblings was 45.80 years (SEM 0.5). There were no significant differences for age at interview between probands and their siblings (t = 0.17, df = 872.95 p = ns) Probands gave a mean age of illness onset of 20.22 years (SEM 0.6) while siblings reported a mean age of onset of 21.04 years (SEM 0.6). Again these differences were not statistically significant (t = -0.98, df = 882.93, p = ns) There were also no significant differences between probands and their siblings for marital status; 161 probands and 170 siblings were living alone while 242 probands and 290 sibings were living with a partner (chi squared test = 0.81 df = 1 p = ns). Genotyping checking Genotyping was carried out by DeCode and the results checked for mis-specified relationships by the programs RELPAIR and Graphical Representation of Relationships (GRR) at the Institute of Psychiatry. RELPAIR compares the multipoint probability of the data conditional on the possible relationships, while GRR calculates the IBS mean and SD for each pair and plots these values, representing each type of relative pair using a different colour. Decisions about each problem pair were made on the basis of the results from both programs, although where there was discrepancy between the programs the GRR results were used. To check genotypes with Mendelian and other pedigree errors the PEDSTAT and MERLIN programs were used. These data cleaning processes resulted in 929 individuals being genotyped at 932 autosomal markers and 44 X chromosome markers. Success rates for the autosomal markers were above 61% and for 90% were above 86%. For the X chromosome the success rate was above 66%. For individuals the genotyping success rate was above 73% for autosomal markers and 61% for the X chromosome. Discussion Inter-site differences and similarities The Depression Network study has recruited affected sibling pairs and some of their parents from 7 European and 1 North American site for a linkage analysis of recurrent unipolar depression. Because of differences in local service organisation, different recruitment strategies have been employed at the different sites. This may account for the significant differences for age at interview and age at illness onset between sites. The Bonn site recruited the oldest sibling pairs, both in terms of when subjects were interviewed and also when their illnesses had commenced. The Bonn subjects had a mean age at interview of 51.46 years compared to the Dublin subjects whose mean age at interview of 43.31 years was the youngest. Similarly the Bonn subjects mean age at illness onset was 27.28 years compared to a mean age of illness onset nearly a decade earlier for the St Louis subjects (18.17 years). It is noteworthy however that the St Louis sample included several large affected sibships. Subjects from families where there are many affected relatives may have a more genetic form of the disorder that might be contributing to an earlier age of onset. Despite these inter-site differences, the results show that there are also considerable similarities across the sites for the subjects recruited. Subjects have been mainly interviewed in their mid 40s and have experienced the onset of their recurrent depression in their early to mid 20s. Consequently subjects had on average around 20 years of history of episodes of depression when interviewed. Gender ratio and similarlities between probands and siblings As expected the study has shown the same preponderance of female to male subjects as many other studies with a gender ratio of around 2.45:1 [ 4 ]. However compared to male subjects, female subjects were significantly more likely to be living with a partner rather than alone. We would also not expect to find any significant differences between probands and siblings in terms of gender, age at interview, age at illness onset or marital status, which the results show is the case. Indeed for the purpose of finding genes for depression we would require siblings to have experienced similar forms of the illness. Genotyping checking Although some subjects were excluded following genotyping due to errors that could not be reconciled, this preliminary genome linkage screen was able to include 929 subjects (98.4%) genotyped at 932 autosomal and 544 X chromosome markers. The results of the whole genome screen will be presented in due course. Conclusions The Depression network study is the first co-ordinated international collaboration of its kind on the genetics of depression and one of the largest ever neuropsychiatric linkage study collection to use a uniform methodology to define and describe the phenotype. Despite taking place across eight sites and in six different countries good inter-rater agreement has been achievable as has good comparability of data. The study has been designed to overcome the difficulties that have been encountered in linkage studies of other psychiatric disorders such as schizophrenia and bipolar disorder. These started out optimistically with the assumption that genes of large effect would exist in at least some multiply affected families. However, after over a decade of contradictory findings and non replications, there is now consensus that such families are very rare or perhaps nonexistent. Rather it seems likely that common familial psychiatric disorders result from the combined effect of multiple genes none of which is either necessary or sufficient to cause the condition [ 20 ]. Consequently large samples are required to have adequate power to detect genes of comparatively small effect, typically where the "risk genotype" confers a genotype relative risk of less than two. Some order is beginning to emerge as a result of meta-analyses of schizophrenia and bipolar data [ 21 - 23 ] however meta-analyses are fraught with difficulties resulting from differences in diagnostic methods, types of the family, genetic marker sets used and methods of ascertainment, in addition to the technical problems of how best to assess statistical significance. It is far preferable to have large diagnostically and ethnically homogenous data sets such as the one described here which will ultimately contain well over 1000 families. Family samples of a comparable size are also being collected elsewhere [ 11 ]. Until now studies of the genetics of unipolar depression have lagged behind those on schizophrenia and bipolar disorder but in doing so we have been able to learn from earlier mistakes. With hope therefore, uncovering the molecular genetic basis of unipolar depression promises to throw up less uncertainties and produce more consistency than has been characteristic of linkage and association studies in psychiatry. Competing interests This study was funded by 3 year research grants to each participating site from Glaxo Wellcome Research and Development. Authors' contributions AF & PMcG were overall study Principal Investigators (PIs), conceived the study and were co-ordinators of the study design, diagnostic reliability and data analysis. AF trained the interviewers from all the sites and wrote the paper. SB, LM and JP obtained the funding from Glaxo Wellcome, recruited site PIs and oversaw the quality of data collection, handling and analysis. GB analysed the genotyping data. The following authors were the individual site PIs in charge of all aspects of subject recruitment and data quality locally: OM Aarhus site, NC, LJ and IJ (joint) Birmingham site, MR and WM (joint) Bonn site, AK and MO (joint) Cardiff site, MG Dublin site, MP Lausanne site AF and PMcG (joint) London site and TR St Louis site. All authors have read and approved the contents of the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544588.xml
539239
Regional risks and seasonality in travel-associated campylobacteriosis
Backgound The epidemiology of travel-associated campylobacteriosis is still largely unclear, and various known risk factors could only explain limited proportions of the recorded cases. Methods Using data from 28,704 notifications of travel-associated campylobacteriosis in Sweden 1997 to 2003 and travel patterns of 16,255 Swedish residents with overnight travel abroad in the same years, we analysed risks for travel-associated campylobacteriosis in 19 regions of the world, and looked into the seasonality of the disease in each of these regions. Results The highest risk was seen in returning travellers from the Indian subcontinent (1,253/100,000 travellers), and the lowest in travellers from the other Nordic countries (3/100,000 travellers). In Africa, large differences in risk between regions were noted, with 502 /100,000 in travellers from East Africa, compared to 76/100,00 from West Africa and 50/100,000 from Central Africa. A distinct seasonal pattern was seen in all temperate regions with peaks in the summer, while no or less distinct seasonality was seen in tropical regions. In travellers to the tropics, the highest risk was seen in children below the age of six. Conclusions Data on infections in returning travellers together with good denominator data could provide comparable data on travel risks in various regions of the world.
Background Campylobacter infection is a zoonotic disease, observed in most parts of the world. The disease is caused by Campylobacter jejuni , or less commonly Campylobacter coli . It is estimated to cause 5–14% of diarrhoea, worldwide [ 1 ]. The incubation period is usually 2 to 5 days (range 1 to 10 days), and persons not treated with antibiotics may excrete the organisms for as long as 7 weeks [ 2 ]. Also in the Western world Campylobacter infection has emerged to be most important bacterial cause of gastrointestinal infection. Animals (variety of fowl, swine, cattle, sheep, dogs, cats, and rodents) are the major reservoir for the bacteria. Campylobacter does not easily grow in food, but the critical infective dose is low [ 3 ]. Unlike salmonellosis with well-known routes of transmission, the epidemiology of campylobacteriosis is still largely unclear, and various known risk factors could only explain limited proportions of the recorded cases [ 4 ]. Known risk factors for the disease include ingestion of undercooked meat, contaminated food and water or raw milk, direct contact with pets, farm animals and small children, and swimming in lakes, but also travel abroad [ 3 , 5 - 7 ]. Direct person-to-person transmission between adults appears to be uncommon. In temperate regions, campylobacteriosis has a distinct seasonal pattern, with the peak incidence in the summer months [ 4 , 6 , 8 , 9 ], but seasonal data on campylobacteriosis from tropical regions are scarce. Approximately 80 million persons from industrialized countries travel every year to places in Africa, Asia, Pacific Islands, Latin America and remote areas of Eastern Europe, and between, 25 and 50 % of travellers to these areas experience travellers' diarrhoea [ 10 - 12 ]. About 80 % of all episodes of traveller's diarrhoea have a bacterial cause, and Campylobacter infection is a leading cause together with infections due to enterotoxigenic Escherichia coli (ETEC), salmonellosis, and shigellosis [ 11 , 13 ]. In this study we have used returning travellers to Sweden as sentinels to estimate the comparative risks for travel-associated campylobacteriosis in 19 regions of the world, and looked into the seasonality of the disease in each of these regions. Methods Notification data on campylobacteriosis Campylobacteriosis is a notifiable disease according to the Swedish communicable disease act. Cases are notified in parallel to the Swedish Institute for Infectious Disease Control (SMI) by the clinician having seen the patient (clinical notification) and the laboratory having diagnosed the pathogen (laboratory notification). At the SMI the notifications from the two sources are merged into case records, using a unique personal identification number issued to all Swedes, and used in all health care contacts. The clinical notifications contain epidemiological information of relevance, including country of infection. For this study we retrieved notification information (age, sex, country of infection and month of infection) from the national surveillance database [ 14 ] on all cases of campylobacteriosis notified in the period 1997–2003, with country of infection outside Sweden. All information in the database is derived from the notifications, and the data (including "country of infection") are thus based on the best judgment of the notifying clinician based on the patient history and knowledge of the characteristics of the pathogen in question. Since we focused on travel-associated infections, refugees and newly arrived immigrants (with incomplete personal identification number) were sorted out before analysis. Denominator data on travel patterns Data on travel patterns were obtained from a commercial database, the Swedish Travel and Tourist Database (TDB) [ 15 ]. This database contains data from monthly telephone interviews with 2,000 randomly selected Swedish residents, with travel-related questions. Out of the total database, containing data from almost 170,000 interviews, we used 16,255 records of persons with recent overnight travel outside Sweden. Each record included information on principal geographical country/region of travel, age, sex, and month of travel, but no data on any illness. Data from the TDB are often given as regions rather than countries, to account for low numbers of respondents outside the most popular travel destinations. Statistical methods The age, sex and geographical distribution of the interviewees in the TDB, were standardised against the total population of Sweden to give an extrapolation of the actual number of travellers to each country during the seven years. We then estimated risks per 100,000 travellers (divided on the exposures sex, age and region of travel) using notifications on Campylobacter infection (cases) as numerator and extrapolated total numbers of travellers from the TDB as denominator. The actual numbers of interviewed persons (controls) were used for the calculations of 95% confidence intervals (95% CI) for the estimates, using the formula: e In risk ± 1.96*√ (1/cases+1/controls) To adjust for possible confounding and test for interaction, we also calculated odds ratios (OR) with corresponding 95% CI for the same exposures with a logistic regression model. In an initial crude analysis, odds ratios (ORs) for all exposures (age, sex, and travel destination) on the outcome campylobacteriosis were analysed, with the lowest incidence in each category used as reference. Confounding was then assessed using Mantel-Haenszel stratification. ORs for exposures with significant association with the outcome were included in a logistic regression analysis if they were shown to contribute significantly to the model in a Wald test. The presence of significant interaction was tested with tests for homogeneity. For each region we analysed seasonality separately (OR for disease per month, adjusted for age, sex and number of cases/travellers). All analyses were done using the Stata 6.0 software (Stata Corporation, College Station, Tx, USA). Ethical considerations Notification data is regulated by the Swedish Communicable disease act, and contain full personal identification. The TDB contains aggregated data only. The Medical Ethics Committee of the Karolinska Institute, Stockholm, Sweden, approved the study. Results Of 53,223 persons notified with campylobacteriosis in the period 1997 to 2003, 28,704 (54%) were travel-associated, corresponding to 42.3 cases per 100,000 travellers (Table 1 ). The total number of infections from single countries both reflected the risk of disease in the various countries, but to a large extent also the travel pattern of Swedes. The five most commonly stated countries of infection were Thailand (n = 6,129), Spain (n = 5,646), Turkey (n = 1,812), Morocco (n = 1,501), and India (n = 1,086). The 16,255 respondents with overnight travel outside Sweden in 1997–2003 from the TDB database corresponded to almost 68 million travel episodes; 78% leisure trips and 22% business trips (Table 2 ). Travel to several countries within one region was quite common, but overnight stay in more than one region was rare (less than 0.1% of travellers). Comparing the number of cases with the projected number of travellers, we estimated the risk for Campylobacter infection in each of the 19 regions under study. The highest unadjusted risks were seen in the Indian Subcontinent (1,253 per 100 000 travellers; 95 % CI 878–1,787), East Africa (502 per 100 000; 95 % CI 314–804), East Asia (386 per 100 000; CI 353–422), North Africa (362 per 100 000; 95 % CI 313–418) and Arab countries/Iran (197 per 100 000; 95 % CI 144–268). Adjusting for age, sex, and month in the logistic regression model did not change the rank between the regions (Tables 1 and 2 , Figure 1 ). In the crude risk estimate women were at significant higher risk for campylobacteriosis than men; 44.0 cases per 100,000 (95 % CI 42.8–45.2) versus 40.8 cases per 100,000 (95 % CI 39.7–41.9). After adjusting for destination, age, and month in the multivariate logistic regression model, the risks were reversed with a significantly higher OR in males (1.17; 95 % CI 1.11–1.23). However, travel destination was an effect modifier on the association between sex and campylobacteriosis, and this higher risk for males were only significant for travellers returning from a European country (OR 1.21; 95% CI 1.15–1.27) (Table 3 ). The highest adjusted age risks were seen in young/middle-aged adults 19–45 years old (OR 2.52; 95 % CI 2.27–2.80) and in small children 0–6 years old (OR 2.34; 95 % CI 1.99–2.76). Also the association between age and campylobacteriosis was modified by travel destination, and in travellers from tropical destinations, especially from Africa and Asia/Oceania the highest risk was seen in the youngest children (Table 3 ). There was a marked seasonality in the temperate regions with peak risks mainly in the summer; Nordic countries (peak in June and nadir in March, OR 11.8; 95% CI 5.9–23.4), Western Europe (peak in June and nadir in December, OR 2.4; 95% CI 1.8–3.3), Eastern Europe (peak in June and nadir in November, OR 2.4; 95% CI 1.8–3.3), North America (peak in June and nadir in March, OR 5.8; 95% CI 1.5–23.4), Southern Europe (peak in September and nadir in January, OR 3.9; 95% CI 3.0–5.0), Northern Africa (peak in September and nadir in May, OR 4.3; 95% CI 1.7–11.0), Arab countries and Iran (peak in April and nadir in August, OR 10.1; 95% CI 1.7–26.4), and Australia/New Zealand (peak in November and Nadir in July, OR 33.1; 95% CI 2.8–394). In the Eastern Mediterranean the peak risk was seen in the spring (peak in March and nadir in January, OR 5.1; 95% CI 2.4–10.8), and in Russia and former USSR in late fall (peak in November and nadir in May, OR 6.7; 95% CI 1.3–59.2). In the tropical regions the seasonality was considerably less distinct. In East Asia the risk peak was in December with nadir in May (OR 4.5; 95% CI 2.8–7.2) and in the Caribbean in February with nadir in September (OR 7.8; 95% CI 2.2–27.7). In the Indian Subcontinent, Sub-Saharan Africa, and Central/South America no distinct, significant seasonal peaks could be identified. Discussion Methodological issues In this study we report the risks for travel-associated campylobacteriosis and seasonality of the risks in various parts of the world, based on more than 28,000 notified cases. The large number of cases gives more precise risk estimates for this disease than in previous studies, although the estimates are given for quite large regions in parts of the world with few Swedish travellers. The denominator data from the TDB has previously been used in studies on dengue fever [ 16 ] and rickettsiosis [ 17 ]. We have also tested the reliability of the TDB by comparing the TDB data with in-flight passenger data obtained from some countries with such requirements. For destinations with many travellers, the two sources of information were highly compatible, e.g. less than 5 % difference for travel to Thailand. Notification data only reflects a small (but unknown) proportion of all travel-related Campylobacter infections. One should therefore be cautious in drawing conclusions from the magnitude of the figures, and rather focus on the relative risks between the various regions, as estimated by the odds ratios. Since the data are all from the same source, the risk figures from the various regions are directly comparable. However, there may be a tendency of investigating travellers from the tropics more vigorous than travellers from e.g. the other Nordic countries, thus underestimating the risks in nearby countries. However, such selection bias could likely not explain the huge differences between say West Africa and East Africa. Since we have no comparable data on the length of stay among the cases in our study, we were not able to include length of stay in our logistic regression model. However, the TDB clearly shows a longer median stay among travellers in far-away destinations. For instance the median stay in Spain was 6 nights, while in Thailand it was 14 nights. On the other hand, only cases detected after the return to Sweden are included in the analysis. The disease data therefore mainly reflect infections contracted during the last week of stay at the travel destination. Differences in length of travel are therefore to some extent evened out. It has previously been suggested that the risk of travellers' diarrhoea is higher during the first two weeks in highly endemic areas [ 10 , 18 ]. The calculated risks in this study may therefore be underestimated in travel destinations with more prolonged stay. However, persons staying long periods abroad are also less likely to be telephone interviewed in Sweden, balancing the missed cases. Regional risks The differences in risk between various regions were considerable, not only between industrialized and developing countries, but also between different developing countries. The Indian Subcontinent, East Africa, East Asia, and North Africa stood out as special high-risk areas. In a recent Finnish study, the risk of travel associated Campylobacter jejuni infection was 10 per 100,000 travellers returning from Spain and Portugal, and 50, 60, and 80 per 100,000 returning travellers from China, Thailand and India, respectively [ 19 ]. The lower risks, and lesser differences between the countries may be explained by a much smaller number of cases (n = 205) to base the risk estimates on. East Africa and India have also previously been identified as high-risk areas for travel-associated diarrhoea of various aetiology [ 20 , 21 ], but the very large differences in the risk of campylobacteriosis between East Africa, and West, Central and Southern Africa have to our knowledge not previously been described. Age, sex and season (month of travel/infection) were identified as possible confounders for the association between travel destination and risk for campylobacteriosis and were thus included in the logistic regression model. All three variables contributed significantly to the model, but the overall effect of these confounders did not alter the rank order between the regions in the logistic regression model compared to the crude analysis. Age and gender The highest risks were seen in young adults and small children, and especially in the tropics the highest risks were seen in the youngest. This is consistent with previous findings that in developed countries the disease most of hits children below the age of 5 years and young adults, while in developing countries it is most often seen in children below the age of 2 years, with an annual incidence of 40–60 % [ 1 ]. The data are also consistent with the results from other studies on traveller's diarrhoeas [ 11 ]. De Las Casas has suggested that the high risk in young adults is due to a more adventurous lifestyle when it comes to eating habits, and the elevated risk in the youngest is due to increased faecal/oral contamination and decreased immunity [ 22 ], explanations that seem plausible. An alternative explanation put forward is that young people with a greater appetite ingest more bacteria, and thereby increasing their risk of infection. In travellers returning from Europe, male gender was an independent risk factor for Campylobacter infection. This pattern is also seen in domestically acquired Swedish campylobacteriosis cases, where 56% of the notified cases in 2003 were males, and in the US where campylobacteriosis is more common in males of all age groups [ 8 , 9 ]. The higher risk in males in the US has been attributed to sex-specific differences in food-handling practices and consumption practices as well as a higher susceptibility to gastro-intestinal infections in males [ 9 ]. For travel destinations to tropical or subtropical destinations, the risk was not influenced by gender, consistent with other studies on travel-associated diarrhoea [ 23 ]. Seasonality In each of the 19 regions in the study, we looked closely at the seasonality of the disease. As has previously been shown [ 3 , 6 - 9 ], there was a striking seasonal pattern in all temperate regions, with distinct peaks in the summer. Previously these summer peaks have been partly attributed to returning travellers [ 4 ], but obviously this could not explain the same peaks in our study. The magnitude of the summer peaks was also in the same order in domestic Swedish cases, as in the returning travellers from other temperate countries. In the tropical regions, seasonal peaks of campylobacteriosis have not previously been recognized [ 8 ]. Also in this study the seasonal pattern was much less distinct in tropical than in temperate regions, and only in East Asia (peak incidence in December) and in the Caribbean (peak incidence in February) could a seasonal pattern be discerned. In a study on US medical students in Mexico, the peak incidence of Campylobacter infection was seen between November and April [ 18 ]. With only 15 cases and 8 TDB respondents, our study did not have the power to detect any seasonality in Central America. Conclusions Data on infections in returning travellers together with good denominator data could provide comparable data on travel risks in various regions of the world. This study has revealed large and unexplained regional incidence differences, e.g. between East and Central Africa. The very distinct seasonal pattern seen in all temperate regions could not be discerned in the tropics. Competing interests The author(s) declare that they have no competing interests. Authors' contributions KE raised the original idea of the study, did the statistical analyses, and prepared the first draft of the manuscript. YA contributed with in depth knowledge of campylobacteriosis and revised the draft manuscript. Both authors have read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539239.xml
549537
Expression of human dopamine receptor in potato (Solanum tuberosum) results in altered tuber carbon metabolism
Background Even though the catecholamines (dopamine, norepinephrine and epinephrine) have been detected in plants their role is poorly documented. Correlations between norepinephrine, soluble sugars and starch concentration have been recently reported for potato plants over-expressing tyrosine decarboxylase, the enzyme mediating the first step of catecholamine synthesis. More recently norepinephrine level was shown to significantly increase after osmotic stress, abscisic acid treatment and wounding. Therefore, it is possible that catecholamines might play a role in plant stress responses by modulating primary carbon metabolism, possibly by a mechanism similar to that in animal cells. Since to date no catecholamine receptor has been identified in plants we transformed potato plants with a cDNA encoding human dopamine receptor (HD1). Results Tuber analysis of transgenic plants revealed changes in the activities of key enzymes mediating sucrose to starch conversion (ADP-glucose phosphorylase and sucrose synthase) and sucrose synthesis (sucrose phosphate synthase) leading to altered content of both soluble sugars and starch. Surprisingly the catecholamine level measured in transgenic plants was significantly increased; the reason for this is as yet unknown. However the presence of the receptor affected a broader range of enzyme activities than those affected by the massive accumulation of norepinephrine reported for plants over-expressing tyrosine decarboxylase. Therefore, it is suggested that the presence of the exogenous receptor activates catecholamine cAMP signalling in plants. Conclusions Our data support the possible involvement of catecholamines in regulating plant carbon metabolism via cAMP signalling pathway.
Background The catecholamines (dopamine, norepinephrine and epinephrine) are a group of biogenic amines possessing a substituted 3, 4-dihydroxy phenyl ring that are widespread in the animal kingdom; but they have also been detected in plants [ 1 , 2 ]. The role of catecholamines in plants is poorly documented, but it is clear that they are involved in many aspects of growth and development. They were proposed as precursors for various alkaloids [ 3 , 4 ] and to be associated with processes such as ethylene production, nitrogen fixation, defence against herbivores, flowering, prevention of 3-indole acetic acid (IAA) oxidation and gibberellin signalling [ 5 , 6 ]. Analogous with animal cells in which catecholamines stimulate glycogen mobilization a similar role for catecholamines in the regulation of plant carbohydrate metabolism was suggested. Transgenic plants over-expressing tyrosine decarboxylase, which controls an important step of catecholamine synthesis, were characterized by highly increased concentrations of norepinephrine and soluble sugars, whereas starch level was dramatically decreased. Observed changes indicated a positive correlation of norepinephrine with soluble sugars and a negative correlation with starch [ 7 ]. The physiological action of catecholamines in animal cells is mediated by their interaction with G-protein coupled receptors that stimulate or inhibit the enzyme adenylyl cyclase (AC). In most animal cells cyclic AMP (cAMP) exerts its effect by activating cAMP dependent, serine-threonine protein kinase (PKA). Recently strong evidence for the occurrence and function of cAMP in higher plants has emerged [ 8 ]. It was demonstrated that cAMP levels in tobacco bright yellow 2 (TBY-2) cells are tightly connected to cell cycle progression [ 9 ]. The involvement of cAMP in gibberellin and ABA action has also been suggested [ 10 , 11 ]. Molecular evidence has shown the existence of plant protein kinases containing a high degree of sequence homology with PKA [ 12 ]. Moreover, molecular techniques led to the identification of cAMP response element-binding proteins (CREBs) [ 13 ], cyclic nucleotide-gated cation channels [ 14 ] and cAMP binding enzymes [ 15 ]. These data strongly indicate the involvement of catecholamines in regulating plant carbohydrate metabolism, possibly by a mechanism similar to that in animal cells. However, this suggestion is limited by the fact that to date no catecholamine receptor has been identified in plants. In the present study we characterize potato plants transformed with a cDNA encoding human dopamine receptor (HD1). The receptor is a rhodopsin-like integral membrane protein of 446 amino acids, seven transmembrane domains and molecular mass of 49 kDa. Our analysis revealed a regulatory effect of HD1 on carbohydrate metabolism including changes in key enzyme activities. Results Transgenic plant selection Solanum tuberosum plants transformed with pHD1-BinAR, a plasmid carrying a cDNA for the human dopamine receptor under the control of the CaMV 35S promoter (Figure 1A ), were pre-selected by means of PCR with the primers for neomycin phosphotransferase (kanamycin) gene and then selected by northern blot analysis with a HD1 specific cDNA as a probe (Figure 1B ). Four transgenic lines showing the highest mRNA expression of the expected length (1300 bp) were chosen for further analysis by western blot (HD1.10, HD1.27, HD1.35 and HD1.36). Using commercially available monoclonal rabbit IgG anti-HD1 antibody a ~37 kDa protein band was detected in transgenic plants. It was absent from control plants (Figure 1C ). Surprisingly the protein was ~10 kDa smaller than expected which may suggest posttranslational modification. Careful inspection of the cDNA sequence revealed the presence of two translational signals (232 bp and 313 bp) that would result in 40 kDa and 37 kDa proteins, respectively. However, as the translational machinery is very similar in plants and animals we suggest that the short form of receptor resulted from proteolytic action rather than de novo synthesis. It should be pointed out that a stronger signal for HD1 expression was accompanied by a stronger protein signal (lines HD1.10; HD1.35; HD1.36). Conversely in plants with weak HD1 expression, the protein signal was comparatively weak (line HD1.27). HD1 extraction with 1% Triton was much more efficient than extraction with 0.1%Triton in agreement with the expected membrane localization of the HD1 protein. Phenotype analysis Tubers of HD1 plants grown in a field were harvested after four months and analyzed. All examined transgenic lines produced more tubers per plant. The yield was not significantly changed since increase in tuber number was accompanied by decrease in tuber weight (Table 1 ). There were no obvious morphological differences between aerial parts of wild type Desiree and HD1 plants. Catecholamine level In order to develop an easy and reliable assay for the quantitative and qualitative determination of catecholamines in plants, the suitability of gas chromatography coupled to a quadruple mass spectrometer (GCMS) was recently investigated. A sensitive GCMS method based on the analysis of the trimethylsilylated catecholamine derivatives was developed to monitor the presence and concentration of these compounds and related metabolites. Based on the retention times and the mass spectra of standards the presence of dopamine, norepinephrine and a new compound normetanephrine in potato leaves and tubers was clearly detected [ 2 ]. In contrast to the previous studies performed on plants over-expressing tyrosine decarboxylase [ 7 ], which controls an important step of catecholamine synthesis, the goal of our work was to stimulate an alternative signalling pathway by introducing human dopamine receptor. Surprisingly the expression of dopamine receptor resulted in a more than two-fold increase of dopamine, norepinephrine and normetanephrine in all transgenic lines examined (Figure 2B ). This increase of catecholamines content was accompanied by significant increase of tyramine and L-DOPA, which are direct precursors of dopamine (Figure 2A ). Despite changes in concentration of catecholamines and their precursors, the level of tyrosine, which serves as a precursor for tyramine and L-DOPA, was not altered (data not shown). Determination of carbohydrate content in tubers of transgenic plants Following our recent finding that the action of dopamine and norepinephrine in potato is on starch mobilization, we decided to analyze transgenic tubers expressing dopamine receptor for soluble sugars and starch content. All transgenic plants showed decreased starch content, with levels that ranged from 20 to 60% percent of the wild type. This was accompanied by a significant increase in soluble sugar concentration ranging from 2.7 to 1.15 fold in comparison to Desiree (Figure 3 ). Concentrations of starch and soluble sugars were highly correlated. The calculated correlation coefficients between catecholamines content and the levels of glucose, sucrose, fructose and starch were 0.38, 0.69, 0.38 and -0.95, respectively. Changes in carbohydrate are most likely responsible for the altered phenotype of transgenic HD1 tubers. Reduced tuber mass can be explained by decreased starch content whereas increased tuber number by the increase of soluble sugar concentration. Sucrose – starch metabolism Under normal growth conditions the major flux in potato tuber carbon metabolism is the conversion of sucrose through hexose phosphates to starch [ 16 ]. Since HD1 plants were characterized by changed concentrations of both soluble sugars and starch we measured the activities of enzymes involved in this pathway. Sucrose transported from leaves is symplastically unloaded from the phloem and degraded by sucrose synthase (SuSy). ADP-glucose phosphorylase (AGPase) converts glucose-1-phosphate (Glu-1-P) into ADP-glucose, an immediate precursor of starch. Both SuSy and AGPase are considered as key enzymes for starch synthesis [ 17 ]. Activities of AGPase and SuSy were significantly decreased in HD1 plants to 56% and 68% of the wild type level, respectively (Figure 4 ). In agreement with their roles in starch synthesis, and their proposed coordinated regulation, activities of both enzymes and starch content were all significantly correlated (cor >0.9). Phosphoglucomutase (PGM) catalyzes the conversion of glucose-1-phosphate to glucose-6-phosphate. Tubers are characterized by the presence of cytosolic and plastidial isoforms of phosphoglucomutase. Repression of either of them results in plants with decreased starch levels pointing out the importance of the enzyme for starch accumulation [ 18 , 19 ]. The activity of PGM was significantly decreased in all transgenic lines, most likely contributing to the reduction in starch synthesis (Figure 4 ). Activities of other enzymes involved in sucrose-starch conversions (hexokinase, UGPase and starch synthase) were not changed significantly (Figure 4 ). In most of the transgenic lines inhibition of starch synthesis was accompanied by increased hexose-6-phosphates (Table 2 ). To establish whether enhanced starch mobilization also contributed to the observed decreases in starch content we measured the activity of starch phosphorylase. In two out of the four examined transgenic lines the activity of starch phosphorylase was significantly increased, further contributing to decreased starch content of HD1 plants (Figure 4 ) Moreover HD1 expression led to activation of sucrose phosphate synthase (SPS), responsible for sucrose production. Maximum SPS activity (measured wih saturating substrates, Vmax) only changed in two lines, whilst activity of the enzyme measured in the assay that contained limiting substrate concentration (Vmax/Km) and as a consequence 1/Km increased in all the lines. 1/Km, correlated well with the sucrose content of transgenic tubers (cor -0.81) (Figure 4 ). Glycolysis/TCA cycle The high concentrations of glucose and glc-6-P measured in the HD1 plants indicated changes in the glycolytic pathway. However, activities of glycolytic enzymes (hexokinase, phosphofructokinase and enolase) were not changed. The only exception was pyruvate kinase, which showed a significant decrease of activity in all transgenic lines (Figure 5 ). To investigate if this reduction of activity led to changes in carbon metabolism via the TCA cycle we measured the content of TCA intermediates. In all transgenic lines citric acid, isocitric acid and malate were significantly reduced, while fumarate showed a significant increase (Table 2 ). Discussion In contrast to the vast knowledge concerning the role and action of catecholamines in mammals, very little is known about the physiological significance of catecholamines in plants. Since most of the components of animal catecholamine signaling pathway have been also identified in plants (G-proteins, cAMP, PKA homologs) the involvement of catecholamines in plant signalling pathways is possible. Recently, the analysis of transgenic plants over-expressing tyrosine decarboxylase, which accumulate a high quantity of catecholamines, suggested a possible signalling effect on plant primary metabolism. The increase of catecholamines resulted in decreased starch concentration but increased soluble sugars [ 7 ]. The only component of mammalian catecholamine signaling pathway that to date has not been identified in plants is the catecholamine receptor. We transformed potato plants with a cDNA encoding human dopamine receptor (HD1) in order to analyze whether the presence of a receptor affects the endogenous catecholamine action. Western blot analysis showed that the protein was produced in transgenic plants and biochemical analysis of transgenic tubers revealed vast changes in carbohydrate metabolism and carbohydrate content. Surprisingly the catecholamine level was changed as well. It has to be pointed out that in contrast to plants over-expressing tyrosine decarboxylase, those expressing human dopamine receptor are characterized by increases of all known tuber catecholamines (dopamine, norepinephrine and normetanephrine). Whereas norepinephrine content was positively correlated with soluble sugars and negatively with starch, normetanephrine was considered as the product of norepinephrine turnover. Increased catecholamine content was accompanied by an increase of their precursors, tyramine and L-DOPA, suggesting upregulation of the biosynthetic pathway, mediated by tyrosine decarboxylase and tyrosine hydroxylase, respectively. It is hard to explain how expression of a human receptor triggers a positive loop leading to enhanced catecholamine synthesis and turnover. It is interesting to compare data on tuber carbohydrate levels from plants over-expressing tyrosine decarboxylase (TD) with those expressing human dopamine receptor. In both cases starch content is strongly decreased, this decrease was larger for HD1 plants (from 40% to 80%) than for TD tubers (from 12% to 60%) although the norepinephrine content was higher in TD plants (Figure 6 ). The norepinephrine content in TD plants was about four folds higher than in HD1 plants. Therefore we suggest that the exogenous receptor activates catecholamine action in potato plants. A difference in enzyme activities involved in starch biosynthesis was noted. The sucrose level was comparable in HD1 and TD plants and consistent with enhanced activity of SPS. Activity of starch phosphorylase was significantly increased in both TD and HD1 plants but the decreases in activity for AGPase, SuSy and PGM was seen only for HD1 plants. Expression of HD1 in potato ( Solanum tuberosum ) results in altered carbon metabolism The previously reported positive correlation between catecholamine level and soluble sugars content and negative correlation with starch level for tubers of potatoes over -expressing tyrosine decarboxylase and in tubers stored at 4°C [ 2 , 7 ], was also found in our study. Expression of a dopamine receptor resulted in increased catecholamine content and was accompanied by decreased starch level and increases of glucose, fructose and sucrose content. It seems likely that the introduced dopamine receptor activates catecholamine action in carbohydrate metabolism. The question now arises whether catecholamine activates starch breakdown or inhibits its synthesis or whether both processes are affected. In mammalian systems epinephrine and norepinephrine regulate glycogen turnover by stimulating glycogen mobilization and inhibiting glycogen synthesis. This appears similar in potato, with decreased starch content in HD1 tubers being a consequence of both inhibition of starch synthesis and enhanced starch mobilization. AGPase and SuSy, two key enzymes involved in starch biosynthesis, showed 44% and 32% decreases in their activities respectively. Also the activity of PGM was significantly decreased; we have not determined the contribution of the different isoforms (cytosolic and plastidial) to the observed changes. Increased content of hexose-6-phosphates demonstrates that a direct inhibition of AGPase, rather than a substrate shortage may cause inhibition of starch synthesis. Alternatively the increased hexose phosphate levels may be due to increased starch degradation in response to elevated catecholamine levels. This is supported by the increased activity of starch phosphorylase in two of four transgenic lines. The inhibition of starch synthesis and accumulation of hexose phosphates was accompanied by an increase of sucrose synthesis. Two factors should be taken into consideration. First, that SPS is subject to allosteric activation by Glc-6-P and inhibition by Pi. Second, elevated catecholamine content led to a decrease of SPS Km suggesting increase of the enzyme catalytic activty. Sucrose phosphate synthase has many potential sites of phosphorylation and three of them were shown to influence its catalytic activity. In spinach leaf, phosphorylation of Ser 158 is responsible for enzyme downregulation in darkness, phosphorylation of Ser 229 enables binding of 14-3-3 proteins and down- regulates the enzyme whereas phosphorylation of Ser 424 under stress conditions stimulates SPS activity. There is a growing body of correlative evidence that the potato tuber SPS is regulated in an analogous manner to the leaf enzyme [ 20 ]. Since the level of 14-3-3 proteins was not changed in any of the transgenic lines (data not shown) it is thus suggested that enzyme phosphorylation targeted to the stress site is responsible for its activity enhancement in HD1 plants. In mammals the action of epinephrine and norepinephrine is mediated by phosphorylation of enzymes involved in glycogen mobilization and synthesis. Very recent studies reported direct evidence that enzymes of starch metabolism (amylopectin synthesis) are regulated by protein phosphorylation and indicate a wider role for protein phosphorylation in the control of starch anabolism and catabolism [ 21 ]. Therefore, it is possible that catecholamine action in plants could also involve phosphorylation of enzymes involved in starch metabolism. Catecholamines – the new stress hormones in plants? In mammalian systems, catecholamines serve as stress hormones increasing as a result of stress. In order to see whether or not a similar response occurs in plants, leaves of potato plants were wounded and catecholamines levels prior to and 5, 10 and 13 min after wounding were determined. Although the data varied, there was a consistent increasing trend in concentration of dopamine, norepinephrine and normetanephrine [ 2 ]. Very recently a similar increase in norepinephrine was measured in potato leaves subjected to ABA and water stress treatment. Activities of both tyrosine hydroxylase (1.5 and 1.7 fold) and tyrosine decarboxylase (2.33 and 1.2fold) were increased [ 22 ]. Under normal growth conditions the major flux in potato tuber carbon metabolism is the conversion of sucrose through hexose phosphates to starch [ 16 ]. During environmental perturbations like wounding [ 23 , 24 ] water stress [ 25 ], high temperature [ 26 ] and hypoxia [ 27 , 28 ] this balance is disturbed and, consequently, large changes in tuber metabolite levels occur. Elevated temperature or water stress leads to increased respiration, a decline in 3-phosphoglycerate (3PGA), inhibition of AGPase and consequently an inhibition of starch synthesis. Decreased starch was accompanied by a stimulation of sucrose synthesis caused by increased hexose posphate levels and activation of SPS via protein phosphorylation. The activity of SuSy was decreased whereas starch mobilization was suggested to increase. These changes in carbohydrate metabolism and carbohydrate content are very similar to those observed in HD1 plants, making it conceivable that catecholamines might play a role in plant stress responses by modulating tuber primary carbon metabolism. Conclusions Introducing humane dopamine receptor into plant cells can be considered as controversial but the obtained data would argue for the value of our approach. Vast changes in the activities of key enzymes mediating carbon metabolism of potato tuber (in HD plants) led to a dramatic reduction of starch but increased sucrose content. The relation between catecholamine, primary carbon metabolism and stress seems possible. We speculate that similarly to situation in animal cells expression of HD1 in potato resulted in activation of the cAMP mediated signalling pathway. This can be supported by the result obtained for potato plants expressing another isoform of human dopamine receptor, HD2. In contrast to HD1, HD2 receptor does not affect activity of adenylate cyclase in animal cells. Similarly plants expressing HD2 showed no changes in carbohydrate metabolism (data not shown). The obvious next step would be further investigation of our plants with respect to their kinase activity as well as cAMP levels. In parallel we have made efforts to identify a plant dopamine receptor. Methods Plant material Potato plants ( Solanum tuberosum L. cv. Desiree) obtained from "Saatzucht Fritz Lange KG" (Bad Schwartau, Germany) were cultivated in a greenhouse in soil under 16 h light (22°C) and 8 h dark (15°C) regime. Plants were grown in individual pots and watered daily. For analysis, the leaves were harvested at noon from 30-day-old greenhouse grown plants and the tubers were harvested in September, 3 months after the transfer of the tissue culture plants to the greenhouse. Construction of a transgenic plant The 1.3 kb SmaI, XbaI cDNA encoding HD1 from Homo sapiens ((kindly provided by Marc G.Caron (Duke University, Medical Center); [EMB: XX55760])), was ligated in the sense orientation into the same restriction site of the plant binary vector under the control of the 35S CaMV promoter and Nos terminator. The vector was introduced into the Agrobacterium tumefaciens strain C58C1:pGV2260 and the integrity of the plasmid was verified by restriction enzyme analysis. Young leaves of wild-type potato S. tuberosum L.(cv. Desiree) were transformed with A. tumefaciens by immersing leaf explants in bacterial suspension. A. tumefaciens inoculated leaf explants were subsequently transferred to callus induction medium and shoot regeneration medium. Transgenic plants were pre-selected by using PCR with the primers for the respective phosphotransferase (kanamycin resistance) gene and then selected by northern blot analysis with a HD1 specific cDNA fragment as probe. Northern blot analysis Total RNA was prepared from frozen plant material using the guanidinium hydrochloride method. Following electrophoresis (1.5% (w/v) agarose, 15% formaldehyde (w/v)), RNA was transferred to nylon membranes (Hybond N, Amersham, UK). Membranes were hybridised overnight at 42°C in 250 mmol sodium phosphate buffer (pH 7.2) containing 7% (w/v) SDS, 1% (w/v) bovine serum albumin (BSA) and 1 mM EDTA. Radioactively labelled full-length cDNA was used as a hybridisation probe. Filters were washed three times in 1 × SSC containing 0.5% (w/v) SDS at 65°C (highly stringent condition) or in the same buffer but at 42°C (medium stringent condition) for 30 min. Western blot analysis Proteins were extracted from frozen plant material using extraction buffer E (100 mM Hepes-NaOH, pH 7.4, 10 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 20%glycerol (v/v), 0.5 mM PMSF, 70 mM beta-mercaptoethanol) supplemented either with 0.1% or 1% TritonX- 100 (v/v). The assessment of the expression of HD1 gene by means of western blot analysis using rabbit IgG anti HD1 protein was conducted as described previously. Briefly, solubilised protein was run on 12% SDS polyacrylamide gels (w/v) and blotted electrophoretically onto nitrocellulose membranes (Schleicher and Schuell). Following transfer, the membrane was sequentially incubated with blocking buffer (5% (w/v) dry milk), and then with antibody directed against the HD1 protein (1:2000 dilution). Formation and detection of immune complexes were performed as previously described [ 28 ]. Alkaline phosphatase-conjugated goat ant rabbit IgG at a dilution of 1:1500 was used as a secondary antibody. Determination of starch and soluble sugar contents Potato tuber slices and leaf discs were extracted with 80% ethanol-50 mM HEPES KOH, pH 7.4, at 80°C. The supernatant was used for enzymatic analysis of glucose, fructose and sucrose [ 29 ]. For starch measurement, extracted plant material was homogenized in 0.2N KOH, and following incubation at 95°C was adjusted to pH 5.5 with 1N acetic acid. Starch was hydrolyzed with amyloglucosidase, and the released glucose was determined enzymatically. Tissue extraction for catecholamine content measurement Frozen plant tissue (400 mg) was powdered in liquid nitrogen and extracted with methanol (4 ml per g -1 fresh weight), heated for 15 min at 70°C, and centrifuged (5 min., 12000 g). Samples were diluted with water to 50% methanol concentration and extracted with chloroform (1:1 v/v). A portion of the water phase was dried under vacuum and used for derivatisation [ 2 ]. Ribitol was used as an internal standard added directly to the sample homogenate (30 μg g -1 fresh weight). GC – MS analysis The dried extracts were incubated in pyridine/methoxyamine (20 mg mL -1 ) at 37°C for 90 min and then acidic protons were derivatised with N-Methyl- N-trimethylsilyltrifluoroacetamide (MSTFA) at 37°C for 90 min. 2 μl of sample was used for analysis [ 2 ]. A HP quadrupole mass spectrophotometer (HP 5972A), combined with a gas chromatograph HP 6890 and autosampler (all Hewlett Packard, Germany) and equipped with 30 m HP – 5MS, fused silica capillary column, was used. Injection temperature was 230°C, with the interface set to 250°C and the ion source adjusted to 180°C. The carrier gas used was helium set at a constant flow rate of 1 ml min -1 . The temperature program was 5 min isothermal heating at 70°C, followed by a 5°C min -1 oven temperature ramp to 310°C and final 1 min heating at 310°C. The system was then temperature equilibrated for 6 min at 70°C prior to injection of the next sample. Mass spectra were recorded at 2 scan s -1 with an m/z 50 – 600 scanning range. The chromatograms and mass spectra were evaluated using the MSD ChemStation program (Hewlett Packard, Germany). As standards, dopamine, norepinephrine, normetanephrine, L-Dopa, tyramine and tyrosine (Sigma) were used. The recovery samples were spiked with dopamine, norepinephrine, normetanephrine, L-Dopa, tyramine and tyrosine; the estimated recoveries were 80, 93, 85, 95, 82, 90%, respectively. The following ions were used for quantification: ribitol 307; 319, L-Dopa 218; 267; 368, dopamine 174; 338; 426, norepinephrine 174; 355, 514, normetanephrine 174; 297; 456, tyramine 174; 264; 338, tyrosine 218; 280, 354. The amounts of catecholamines were determined from the ratio of peak areas of catecholamines to peak area of the internal standard (ribitol). Preparation and analysis of samples for enzyme activities Tissue was harvested, weighed and immediately frozen in liquid N 2 . 0.5 g ± 0.1 g tissue was homogenised in a chilled mortar in 2 ml of an extraction buffer containing 30 mM HEPES-NaOH, pH 6.9, 10 mM DTT, 1 mM MgSO 4 , 0.5 mM EDTA, 0.5% (w/v) BSA and 0.5% (w/v) PVP at 4°C. The homogenate was centrifuged at 16000 g for 10 min and the supernatant desalted using Sephadex G-25 columns. Enzyme activities were determined using the following published methods; SPS [ 30 ]; AGPase [ 31 ]; SuSy, PK [ 32 ]; PGM [ 33 ]; starch synthase [ 34 ]; PGI, enolase [ 35 ]; UGPase [ 36 ]; hexokinase [ 37 ] and starch phosphorylase as described by [ 38 ]. 3.9. Statistical analysis The t-tests were performed using the algorithm embedded into Microsoft Excel. The term significant is used when P < 0.05. Authors' contributions AS carried out the metabolic analysis of transgenic plants and drafted the manuscript. AŚ carried out the construction and selection of transgenic plants and performed the statistical analysis. JS conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549537.xml
520754
Good physicians from the perspective of their patients
Background It is not currently known what is the patient's viewpoint of a "good" physician. We set out to define patient's priorities regarding different physician's attributes in 3 domains important in medical care. Methods Patients hospitalized or attending clinics at a large teaching hospital selected the 4 attributes that they considered most important out of 21 listed arbitrarily in a questionnaire. The questionnaire included 7 items each in the domains of patient autonomy, professional expertise and humanism. Results Participating patients (n = 445, mean age 57.5 ± 16 years) selected professional expertise (50%), physician's patience and attentiveness (38% and 30%, respectively), and informing the patient, representing the patient's interests, being truthful and respecting patient's preferences (25–36% each) as the most essential attributes. Patient's selections were not significantly influenced by different demographic or clinical background. Selections of attributes in the domain of patient's autonomy were significantly more frequent and this was the preferred domain for 31% and as important as another domain for 16% – significantly more than the domain of professional expertise (P = 0.008), and much more than the domain of humanism and support (P < 0.0005). Conclusions Patients studied want their physicians to be highly professional and expert clinicians and show humaneness and support, but their first priority is for the physician to respect their autonomy.
Background An expert clinician whose choices are evidence-based; sensitive and dedicated to the patient – not just disease-oriented; who heeds and respects the patient's autonomy, striving at conveying all relevant information to the patient to enable a truly shared decision-making. That can be seen as a prototype or role model of a first class physician. However, these laudable qualities, discussed in many texts [ 1 - 3 ], were purely conceived by physicians, for physicians. How do patients value these different physicians' qualities? What would their priorities be? This unique point of view, little discussed in previous studies, is the subject of our current research. Methods Patients hospitalized at our 600-bed teaching medical center or attending one of its day clinics were randomly approached and asked to fill in a one-page questionnaire. Over a period of one month, patients were approached by one of the authors (DR) and invited to participate. Every third patient on the specific day's ambulatory clinic lists and every fifth patient on the list of patients hospitalized at the Department of Medicine were approached. If the fifth patient was found to be too ill to participate (e.g. ventilated, unstable or confused) the next name on the list was selected. These numbers were arbitrarily chosen. Following a brief explanation they were handed the questionnaire. The questionnaire listed 21 physicians' characteristics or behaviors regarding the care of patients, presented in a varying order (different order for different patients) to prevent bias due to an item's position on the list. Patients were asked to select 4 attributes that they considered the most important and would like the best in their own physician. No grading was required. The questionnaire aimed at the patient's image of an excellent physician in general, and did not specify whether the physician was hospital-based or in primary care. The 21 attributes included 7 characteristics in each one of three domains: reflecting professional expertise and high-quality care; reflecting a humanitarian, patient-centered approach; and reflecting patient's autonomy and attentiveness to the patient's preferences and rights. Then patients supplied basic demographic data and the questionnaire was collected about 20 minutes later. The selection of the 21 different physician characteristics or behaviors started from collecting pertinent articles on patient autonomy; physician's humanism and patient support; and on physicians' expertise and professionalism using the author's (AS) collection and a Medline search. Some of the articles were based on patient-derived data. The next stage involved a series of meetings and discussions at the Department of Ambulatory Care and Prevention, Harvard Medical School, Boston. Three senior researchers (2 clinicians, all with extensive experience in research and medical education) – RH Fletcher, T Peters and AS, selected and categorized the 21 items based on the literature and on personal experience. The third stage was a validation study among thirty residents in various stages of their training. They were presented with a mixed list of the 21 items and requested to categorize each one into one of the 3 domains. To rule out that patients categorized these characteristics differently, we performed an additional post-study validation on 30 outpatients that were representative of our study population. Results were similar to those obtained from the residents and confirmed that the vast majority of participants view each of the items as representative of the corresponding domain (Appendix 1) [see additional file 1 ]. No changes in classification were necessary following the validation study. We planned to a) quantify and study the most 'popular' physician's attributes selected by patients as well as those selected by only a few patients. b) find-out for each patient, whether any one domain was over-represented in the patient's selections (e.g. 2 out of 4 selections belonging to a single domain and less than 2 for each of the others) or – under-represented (no selections in a domain). The study was approved by our Institutional Review Board. Statistical analysis was done using chi-square tests to examine differences between the domains, and between preference of each domain and demographic variables. T-test or one-way ANOVA were used when appropriate. Results A total of 450 patients received the questionnaire and all but 5 consented to participate and returned the filled questionnaires (n = 445). Patient's ages varied from 18 to 89 years (mean 57.5 ± 16) and two thirds were aged 50–80 years. Other patient's characteristics are summarized in Table 1 . The top eight physician's qualities preferred by the patients participating in the study (each selected by >25% of patients) are given in Table 2 . (Top) and the five physician's qualities selected by <5% are given in the bottom of the same Table. When we transposed each attribute for the domain it represents, basically 2 types or patterns of responses were generated: the AAPH type, reflecting this patient's preference for domain A; and the AAPP type, reflecting the importance (but not dominance) of domain A (as well as P). When no selection at all was made in one domain (such as domain H in the AAPP example), this was also noted (Table 3 ). Analysis of patients' preferences of physicians' qualities according to domains (qualities within Professional, Humanitarian or patient's Autonomy domains), yielded highly significant differences. For example, 139/445 patients (31%) selected more answers in the patient's autonomy domain than in any other domain, and 69 more patients gave equal importance to patient's autonomy and to one other domain (mostly professionalism). In contrast, humanistic qualities of physicians were selected as the most important by 76/445 patients (17%) only, and just 48 additional patients gave an equal importance to the humanistic and one other domain. Altogether, 90/445 patients clearly selected more characteristics in the professional domain than in any other domain (20%), and 86 other patients gave equal importance in their choices of qualities to the professional and one other domain. Thus, 69% of participating patients gave clear indication as to their preferred domain of physician's characteristics in administering medical care (Table 3 ). When these preferences and those of patients who entirely disregarded the domain were analyzed and compared, significant differences were found favoring the domain of patient's autonomy (P = 0.008 vs. professional expertise and P < 0.0005 vs. humane attitude) (Table 3 ). About 10% (38/445), selected equally between the three domains (i.e. one quality of each domain). The remaining 102/445 (22%) were 'indeterminate' in that their choices gave equal importance to two domains. When we analyzed patient's responses to the most essential questions in each of the domains, a similar pattern emerged. For example, 34% selected 2 or more qualities pertaining to patient's autonomy, and 21% selected none. In comparison, <10% named = 2 'humanistic' qualities and over 50% selected none (P < 0.0005). When patient's preferences of the different physicians' characteristics or domains were further analyzed according to the patient's age, gender, origin, income, being hospitalized or main diagnosis – no statistically significant differences or associations could be identified (not shown). Discussion Our study population of 445 patients was heterogeneous (Table 1 ), yet no statistically significant relationship could be demonstrated between demographic or clinical variables and patient's choices or priorities. These facts lend more impact to our findings. Patient's preferences of physician's attributes were found to be as unique and individual as the patients themselves. It is remarkable that among 445 patients, only 6 made identical choices (1%). Not surprisingly, patients want their physicians to be experienced and highly professional. This was the physician's characteristic that was chosen by 50% of the patients (Table 2 , Top). However, 4 of the other 7 most frequently selected attributes, each selected by 25–38% of the patients, were in the domain of patient's autonomy (523/1172 responses, 45%). The distinct priority accorded by patients to attributes in the realm of patient's autonomy, overrides even the domain of professionalism, and certainly that of physician's humanism and support. Attributes belonging to the domain of patient's autonomy were uncommon among the least demanded attributes (1 of 5, Table 2 , Bottom), and were significantly more often selected and less often disregarded than any other domain (Table 3 ). In contrast, attributes of humanism were not selected at all by almost 30% of participating patients. An overview of the results reveals that patients studied want their physicians to show professional expertise and provide humane personal care (a preferred or important domain for 39% or 28% of the patients, respectively) (Table 3 ) – however, it is even more important to them to be well informed and participate in decisions (Tables 2 , Top, and 3 ). On the other hand, 'humane' qualities that are traditionally considered important, such as showing empathy or being friendly with the patient were surprisingly found to be among the least selected attributes (3–4%, Table 2 , Bottom). Moreover, patients participating in this study seldom selected professional qualities such as research and teaching abilities, although they often go hand in hand with appointments at academic medical centers and high quality medical care [ 4 ]. Patients also seemed unaware of the unequivocal power of the prevention of "accidents waiting to happen" [ 5 ] that can be offered to them by professional physician's counsel. Only 69 patients (15.5%) selected this option as one of four of their priorities. What patients did value, in addition to clinical experience and being up-to-date (Table 2 , Top), was for their physicians not to be impatient (38%) or distracted (30%). These two attributes, perhaps the most vulnerable to current time constraints in clinical practice, may reflect patient's needs of a more relaxed, leisurely communication with their physician, be it at the hospital or in primary care. This issue may be more pertinent with the recent changes in context of the consultation, mandating more informed patients and shared decision-making [ 6 , 7 ]. The reason why patient's autonomy features so highly in our study, is not entirely clear. The patient population was heterogeneous (Table 1 ) so that factors related to origin, income or education are unlikely to be operative. Our medical center is serving a large (about 250,000) population in central Israel which is mostly urban but also rural, and has nothing unique regarding physicians or health delivery system. It is conceivable that the growing emphasis in recent years on patient's rights and patient's autonomy which is prominently reflected in the media and draws much public attention, affected patient's preferences. Further studies will be needed to determine whether, as we believe, our results reflect a novel worldwide trend in patient's preferences. Several limitations of our study ought to be considered. The study was performed in Israel, on a predominantly Jewish population. Thus, its generalizability is open to question. However, Jewish medical ethics put no special value on patient's autonomy so that the findings are likely to apply to other Western affluent societies as well. Also, intentionally, our methodology does not address the relative weight of respondent opinion. Thus, it is likely – but uncertain – that choosing more attributes in a domain is the crucial measure of the relative importance of a domain. Finally, we cannot be entirely certain that the selection process did not lead to some bias, although we consider this contingency to be unlikely. The age of paternalism in medical care has come to an end and few are sorry for its demise. Most patients want to be informed about their health even if the news are bad [ 8 ], and to be involved with their care plans [ 9 ]. To do that, patients must have clear information, which takes into account their unique circumstances [ 10 ], and there is no better source for that than the patient's physician. Our results strongly suggest that patients expect their physicians to heed these needs and prefer physicians who are sensitive to the varied aspects of patient's autonomy and patient's rights (Appendix 1) [see additional file 1 ]. The studies of Thom et al. have already indicated that certain physician's behaviors were important for patient's trust. Prominent among them were discussing options with the patient and finding out preferences – essential components of patient autonomy [ 11 ]. Similar to the present study, measures of professional competence and humanism were also required. Patient trust was significantly correlated with compliance and with clinical improvement [ 12 ]. Since patient trust appears to be such a crucial component of the patient-physician relationship, and since trust is dependent on patient's preferences being met [ 12 ] it is mandatory to establish what these preferences are today, and this is where our study comes in. Previous research has already identified the complexity of patient's needs in the modern era. While older studies make no mention of issues of patient autonomy [ 13 ], a relatively recent systematic review of the literature on patient's priorities found "humaneness" to be the most highly rated aspect of care, followed by clinical competence and patient's participation in decisions [ 14 ]. Another study from Scotland identified physicians' attentiveness and patience at the top of the list, and patients in the Netherlands mostly desired sufficient consultation time [ 15 , 16 ]. However, a strong desire for information and participation in decision-making already features in these and other studies [ 14 - 18 ]. Among our patients it came out for the first time as a top priority, second to none (Table 3 ). Failure of physicians to provide a patient-centered approach may therefore seriously undermine patient's expectations and satisfaction. Recent data suggest that this may be associated with significant adverse outcomes [ 19 , 20 ]. Physicians can be effectively trained to listen to the patient's narrative [ 21 ], recognize the patient's perspective [ 7 , 22 ] and adopt a more patient-centered approach [ 23 ]. However, besides skills in communication [ 21 , 24 , 25 ], this requires time. Time management can also be effectively taught [ 25 ] but it is still unclear whether decreasing hospital length of stay and consultation time in primary care will not adversely affect physicians' performance and patient's expectations [ 3 ]. Nevertheless, patient's preferences remain integral to modern evidence-based practice [ 26 ], and our study should provide a poignant reminder that autonomy is nowadays what the patients want most. Conclusions Our study focused on the selections of several hundred outpatients and inpatients, regarding their preferences for different attributes of their physicians. We found heterogeneous preferences, but attributes in the domain of patient autonomy and physicians' expertise (in that order), headed the list. Competing interests None declared. Authors' contributions AS initiated, designed and supervised the study, analyzed the results and wrote the manuscript. DR performed the actual study. NJ provided statistical advice and analysis. All authors read, discussed, contributed to and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 The patient's questionnaire including the 21 attributes in the three domains. Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC520754.xml
524254
Measuring Global Trends in the Status of Biodiversity: Red List Indices for Birds
The rapid destruction of the planet's biodiversity has prompted the nations of the world to set a target of achieving a significant reduction in the rate of loss of biodiversity by 2010. However, we do not yet have an adequate way of monitoring progress towards achieving this target. Here we present a method for producing indices based on the IUCN Red List to chart the overall threat status (projected relative extinction risk) of all the world's bird species from 1988 to 2004. Red List Indices (RLIs) are based on the number of species in each Red List category, and on the number changing categories between assessments as a result of genuine improvement or deterioration in status. The RLI for all bird species shows that their overall threat status has continued to deteriorate since 1988. Disaggregated indices show that deteriorations have occurred worldwide and in all major ecosystems, but with particularly steep declines in the indices for Indo-Malayan birds (driven by intensifying deforestation of the Sundaic lowlands) and for albatrosses and petrels (driven by incidental mortality in commercial longline fisheries). RLIs complement indicators based on species population trends and habitat extent for quantifying global trends in the status of biodiversity. Their main weaknesses are that the resolution of status changes is fairly coarse and that delays may occur before some status changes are detected. Their greatest strength is that they are based on information from nearly all species in a taxonomic group worldwide, rather than a potentially biased subset. At present, suitable data are only available for birds, but indices for other taxonomic groups are in development, as is a sampled index based on a stratified sample from all major taxonomic groups.
Introduction The world's biodiversity is diminishing rapidly ( Balmford et al. 2003 ; Jenkins et al. 2003 ). At the 2002 World Summit on Sustainable Development, the nations of the world agreed to pursue more effective implementation of the objectives of the Convention on Biological Diversity (CBD) in order to achieve a significant reduction in the current rate of loss of biological diversity by 2010 ( Secretariat of the Convention on Biological Diversity 2003 ). The European Union has adopted the more ambitious target of halting the loss of biodiversity by 2010 ( European Union 2001 ). We do not yet have an adequate way of monitoring progress towards achieving this target. However, the CBD Conference of the Parties at its Seventh Meeting adopted Decision VII/8, which included recommendations to develop indicators for measuring trends in the components of biodiversity based on (1) trends in extent of selected biomes, ecosystems, and habitats; (2) trends in abundance and distribution of selected species; (3) change in status of threatened species; (4) trends in genetic diversity of domesticated animals and cultivated plants; and (5) coverage of protected areas ( CBD 2004 ). None of these alone is adequate, but together they provide powerful measures of global trends in the status of biodiversity. Here we address the third of this suite of indicators, using changes in the threat status of species as measured by the categories of the World Conservation Union (IUCN) Red List. The IUCN Red List is widely recognised as the most objective and authoritative listing of species that are globally at risk of extinction ( Lamoreux et al. 2003 ; Hambler 2004 ). Species are assigned to Red List categories (see Abbreviations section) through detailed assessment of information against a set of objective, standard, quantitative criteria ( IUCN 2001 ; Table 1 ). Thousands of scientists, many of whom are members of IUCN Specialist Groups and the IUCN Species Survival Commission network, provide extensive information for assessments. Over the last few years, the IUCN Red List has been developed into a global programme to monitor the extent and rate of biodiversity degradation. The programme is currently overseen by four partner organisations: the IUCN Species Survival Commission, BirdLife International, NatureServe, and the Center for Applied Biodiversity Science at Conservation International, with additional partners being recruited, in particular to provide plant and marine expertise. Red List Authorities are appointed to ensure consistent categorisation between species and groups and for organising independent scientific review. A Red List ‘Standards and petitions' subcommittee monitors the process and resolves challenges and disputes to listings. Table 1 Simplified Overview of Thresholds for the IUCN Red List Criteria NA, not applicable One of the goals of the programme is to provide a global index of the changing state of biodiversity ( IUCN 2004 ). However, previous attempts to use the IUCN Red List to provide answers about the rate of loss of biodiversity suffered from many limitations (Cuarón 1993; Smith et al. 1993 ), and a number of problems with the general approach have been identified ( Burgman 2002 ; Possingham et al. 2002 ; but see Lamoreux et al. 2003 ). We show, using data from birds, how the limitations can be overcome, and we present for the first time Red List Indices (RLIs) that are robust, temporally sensitive, representative, and comprehensive. These provide unique baseline data on the rate of loss of biodiversity against which progress towards meeting the CBD 2010 target can be judged. They also allow finer-scale resolution of trends in particular biogeographic realms, ecosystems, and habitats. Results The total number of extant threatened and Near Threatened birds listed on the IUCN Red List has changed relatively little over the four complete assessments of all the world's birds, increasing from 1,664 species in 1988 to 1,990 species in 2004. However, large numbers of species have moved between categories, particularly in the earlier assessments ( Table 2 ). Most of these category changes have been a consequence of improved knowledge (including improved consistency of interpretation of information against the Red List criteria) or revised taxonomy. However, a significant proportion of species (equating to 2.4%–7.3% of threatened or Near Threatened species in each assessment) have moved between categories because of genuine improvement or deterioration in status. The RLI for birds illustrates the combined effect of these genuine status changes, to provide a simple metric of the changing overall status of the world's birds, in terms of their relative projected extinction risk as estimated using the categories of the IUCN Red List. This shows that there has been a steady and continuing deterioration in the threat status of the world's birds between 1988 and 2004, with an overall change in the index value of −6.90% over this period ( Figure 1 ; see Table 2 ). No change would indicate that the average status of all bird species was the same as in 1988. To put this into context, if 10% of species in the categories from Near Threatened to Critically Endangered had deteriorated in status sufficiently to be uplisted one category to a higher category of threat between 1988 and 2004, the index would have changed by −7.8%, and if 50% of such species had deteriorated by one category the index would have changed by −27.4% ( Figure 2 ). The error bars for the 2004 RLI value (based on the projected number of genuine status changes for the 2000−2004 period yet to be detected owing to information time lags; see Discussion ) show that the estimated recent RLI trends are likely to be fairly robust. Figure 1 The RLI for All Bird Species Sample size: 250 genuine status changes/2,469 species in categories Extinct in the Wild to Near Threatened in at least one assessment. Error bars for 2004 RLI value based on estimated number of genuine status changes for 2000–2004 not yet detected owing to information time lags (see Materials and Methods for further details). Figure 2 The RLI for All Bird Species in 1988–2004 Compared to Hypothetical Indices Hypothetical indices show trends if no species had changed category, and if 10% or 50% of species in the categories from Near Threatened to Critically Endangered had been uplisted to a higher category of threat or downlisted to a lower category of threat over the period. Table 2 Number of Species in Each IUCN Red List Category as Published in Collar and Andrew (1988) , Collar et al. (1994) , BirdLife International. (2000 , 2004c ), and the Number of Species Undergoing Genuine Status Changes in Each Period Table includes categories for Houbara bustard (C. undulata) and Saker falcon (F. cherrug) that were revised by BirdLife International for the 2004 IUCN Red List but subsequent to BirdLife International. (2004c ); see Materials and Methods a Includes categories retrospectively adjusted owing to category revisions coded as ‘genuine status change since first assessment' CD, Conservation Dependent; CR, Critically Endangered; DD, Data Deficient; EN, Endangered; EW, Extinct in the Wild; EX, Extinct; LC, Least Concern; NE, Not Evaluated; NR, Not Recognised; NT, Near Threatened; PE, Possibly Extinct; VU, Vulnerable To examine trends in the status of the most threatened species (i.e., those closest to extinction), the index was calculated using weights for each category related to the relative extinction risk associated with them ( Figure 3 ; see Materials and Methods ). This shows a levelling out of the decline in the index value during 2000–2004 (although the error bars indicate that in the next few years the belated discovery of genuine status changes for this period could reduce this apparent levelling out). This was because for those species that underwent genuine status changes in the categories of highest extinction risk (those that have the greatest influence on the index value when calculated in this way), the number of species that deteriorated in status during this period was balanced by the number that improved in status owing to conservation action. Specifically, two Critically Endangered species became Extinct (or Possibly Extinct) in the wild (Hawaiian crow [ Corvus hawaiiensis ] and Spix's macaw [ Cyanopsitta spixii ]), and five Endangered species became Critically Endangered, but this was offset by seven species that improved in status as a result of conservation efforts (including, e.g., Polynesian megapode [ Megapodius pritchardii ], Christmas Island hawk-owl [ Ninox natalis ], and Christmas Island white-eye [ Zosterops natalis ]; BirdLife International 2004c ). Figure 3 The RLI for All Bird Species with Categories Weighted by Relative Extinction Risk Sample size: 250 genuine status changes/2,469 species in categories Extinct in the Wild to Near Threatened in at least one assessment. Error bars for 2004 RLI value based on estimated number of genuine status changes for 2000–2004 not yet detected owing to information time lags (see Materials and Methods for further details). The RLI can be broken down by biogeographic realm ( Figure 4 ), ecosystem, habitat type ( Figure 5 ), and for particular species groups ( Figures 6 and 7 ). These show that the threat status of birds has deteriorated worldwide, with a more-or-less similar rate and proportional extent of deterioration in the Nearctic, Neotropical, Palearctic, Afrotropical, and Australasian/Oceanic realms. The RLI for the Indo-Malayan realm showed a steeper rate of decline during the 1990s (see Figure 4 ). Declines in the index for three major ecosystems (terrestrial, freshwater, and marine) and two terrestrial habitat types (forest and shrubland/grassland) all show a broadly similar pattern (see Figure 5 ), although the declines in the freshwater environment appear to have been the most severe. Finally, RLIs for selected subsets of species (including those relevant to particular international treaties) highlight the severity of the worsening situation of the world's albatrosses and large petrels in recent years (see Figures 6 and 7 ). Figure 4 RLIs for Birds in Different Biogeographic Realms Sample sizes: Neotropical, 49 genuine status changes/834 species in categories Extinct in the Wild to Near Threatened in at least one assessment; Afrotropical, 41/394; Australasian/Oceanic, 53/614; Palearctic, 34/238; Nearctic, 9/92; Indo-Malayan, 100/585. Figure 5 RLIs for Birds in Different Habitats Sample sizes: terrestrial, 206 genuine status changes/2,329 species in categories Extinct in the Wild to Near Threatened in at least one assessment; freshwater, 31/226; marine, 12/133; shrubland/grassland, 45/481; forest, 169/1,513. Figure 6 RLIs for Three Bird Families with High Conservation Profiles Sample sizes: game birds, 15 genuine status changes/123 species in categories Extinct in the Wild to Near Threatened in at least one assessment; raptors, 10/93; parrots, 19/148. Figure 7 RLIs for Three Species Groups Targeted by Particular International Conservation Treaties: The Ramsar Convention on Wetlands, the CMS, and the ACAP under the CMS Sample sizes: waterbirds, 36 genuine status changes/238 species in categories Extinct in the Wild to Near Threatened in at least one assessment; albatrosses and petrels, 6/28; migrants, 50/313. Discussion How Fast Are We Losing Avian Biodiversity? Bird species are being driven Extinct by increasing human impacts on the planet. In total, 129 bird species have been classified as Extinct since 1500, with an additional four species listed as Extinct in the Wild, but surviving in captive populations ( BirdLife International 2004b , 2004c ). Additionally, 18 Critically Endangered species are considered Possibly Extinct by BirdLife International. (2004b , 2004c ; see Materials and Methods ). Of these confirmed and likely extinctions, nine have occurred during the period 1988–2004 (BirdLife International, unpublished data). However, it is very difficult to produce robust estimates of recent extinction rates and to quantify how they have changed over short timescales. This is because extinction is difficult to detect once species become very rare ( Diamond 1987 ; Reed 1996 ), and tiny populations of species potentially doomed by habitat loss or other threats may persist for many decades ( Turner 1994 ; Brooks et al. 1997 , 1999 ). For these reasons, extinction rates perform weakly as indicators of the current state of biodiversity ( Balmford et al. 2003 ). By contrast, the RLIs presented here provide a robust, sensitive measure of the rate at which the world's birds are changing in relative projected extinction risk, as classified using the categories of the IUCN Red List. The indices show that the overall threat status of the world's birds has deteriorated steadily over the last 16 y. The RLI value has changed by −6.90% over this period. However, it should be noted that owing to the arbitrary nature of the weights applied to each category to calculate the score, this percentage decline is not directly comparable with percentage declines reported for population-based indices such as the Living Planet Index ( Loh 2002 ) or the United Kingdom headline indicator for wild bird populations ( Gregory et al.2003 ). When the RLI is weighted by the relative extinction risk associated with each category in order to emphasise trends in the status of the most threatened species, the rate of decline of the index value appears to have levelled off in recent years (see Figure 3 ), owing to the number of such species deteriorating in status being balanced by the number improving. Nevertheless, it should be emphasised that one Critically Endangered species went Extinct in the Wild in the wild during the period (Hawaiian crow [ C. hawaiiensis ]), and another is highly likely to have done so (Spix's macaw [ C. spixii ]; BirdLife International 2004c ). These are potentially irretrievable losses to genetic diversity. How can we interpret the RLI in relation to the CBD's target of reducing the rate of loss of biodiversity by 2010? The interpretation is different for measures of the state of biodiversity (e.g., total area of remaining forest) and measures of the rate of change in this state (e.g., annual percentage forest loss). For indices based on proportional change in a measure (plotted on a negative scale as with the RLI), if the measure is one of state, a significant diminution in downward trend would show that the target has been met. If the measure is one of rate of change of state, however, the target is not met until we see a positive trend, not just a decelerating decline. Some of the Red List criteria are based on absolute population size or range size, while others are based on rates of decline in these values or combinations of absolute size and rates of decline. These criteria are used to assign species to Red List categories that can be ranked according to relative projected extinction risk, and the RLI is calculated from changes between these categories. Hence an RLI value relates to the rate at which species are slipping towards extinction at a particular time. To show that the 2010 target has been met, the RLI must therefore show a positive trend. A downward trend, even if diminishing, shows that the slide of species towards extinction is accelerating, not slowing down. The negative trends in the RLI values (see Figure 1 ) thus show that in 2004 we are losing biodiversity at an increasing rate. The RLIs show some interesting regional variations. The index for birds in the Indo-Malayan realm shows a sharp decline during the 1990s (see Figure 4 ). This was a result of the intensifying destruction of forests in the Sundaic lowlands of Indonesia, which escalated particularly in the late 1990s and led to predictions of almost total loss of lowland forest in Sumatra by 2005 and in Kalimantan by 2010 ( Holmes 2000 ; BirdLife International 2001 ). As a consequence of these increasing rates of habitat loss, many species were uplisted to higher categories of threat under criterion A (rapid population declines). However, it is notable that there has been a significant deterioration in the threat status of birds of shrubland/grassland habitats as well as forest, and in the two other major ecosystems (freshwater and marine), indicating that birds in a broad spectrum of environments are under threat. RLIs can be calculated for particular species groups that have specific conservation or policy significance. For example, there are particularly active conservation networks for game birds (e.g., World Pheasant Association), raptors (e.g., World Working Group on Birds of Prey), and parrots (e.g., Loro Parque and World Parrot Trust), and the threat status of all three of these species groups is deteriorating, with steeper declines in the index value for parrots in the earlier part of the period (see Figure 6 ). In addition, there are several international conservation treaties targeting particular suites of species (the Ramsar Convention on Wetlands, the Convention on Migratory Species [CMS], and the Agreement on the Conservation of Albatrosses and Petrels [ACAP] under the CMS) for which disaggregated RLIs provide a metric against which to judge their success in improving the fortunes of the species involved. The RLI for albatrosses and large petrels shows how dramatically their threat status has deteriorated in recent years (see Figure 7 ). This is closely linked to the expansion of commercial longline fisheries (both legal and illegal), which causes incidental mortality of albatrosses and other seabirds when they get caught on baited hooks and drown ( Tuck et al. 2001 ; 2003 ; BirdLife International 2004b ). The total reported effort from fleets in the southern oceans has been well over 250 million hooks per year since the early 1990s, with some fleets expanding rapidly in the last decade ( Tuck et al. 2003 ). Models for at least some albatross species show clear links between population declines and these increases in longline fishing effort ( Tuck et al. 2001 ). Mitigation measures are effective ( Løkkeborg 2001 ), and the RLI will provide a useful measure by which to judge the effectiveness of the implementation of ACAP, following its entry into force in 2004. It should be noted that setting all disaggregated index values to a common baseline in 1988 obscures any changes prior to this period (see, e.g., Pauly 1995 ). For example, although the Indo-Malayan realm has shown the most severe recent index declines, ‘only' six extinctions occurred there between 1500 and 1988, whereas at least 62 bird species are known to have gone Extinct in the Australasian/Oceanic realm during the same period, and 40 in the Afrotropical realm ( BirdLife International 2000 ). Similarly, the terrestrial ecosystem has suffered far more extinctions since 1500 (115) than the freshwater (17) or marine ecosystems (five), but all are set to a common baseline in 1988. Category Weights The RLI is based on the number of species in each Red List category. In order to make the index sensitive, not just to the total number of threatened species, but also to the changes in category assigned to each species, each category was given a weighting. We used an ‘equal-steps' approach (with incremental increases from one for Near Threatened through to five for Extinct) to reflect the ordinal ranks of the categories, whereby each step from Least Concern to Extinct indicates that at least one measure of extinction risk has become worse. The advantage of this approach is that it is simple, and the trends in the resulting index are driven by a relatively large number of species (hence producing a more robust and representative index). This is because a species moving from Least Concern to Near Threatened contributes just as much to the changing score as a Critically Endangered species going Extinct, and the numbers of species in each category (and the number moving in and out of each category) increases disproportionately from Critically Endangered to Least Concern (see Table 2 ). However, steps between lower categories of threat represent smaller increases in extinction risk than steps between higher categories. Therefore we also tested an ‘extinction risk' approach, with each category weighted according to its relative extinction risk based on the quantitative thresholds for each of its criteria ( Table 3 ). Although this approach also relies on some assumptions (e.g., about the type of extinction risk curve, and the extinction risk associated with Near Threatened), it is based on the principles of extinction dynamics, in contrast to the equal-steps approach. Table 3 Weights for Red List Categories Critically Endangered, Endangered, and Vulnerable, Based on Relative Extinction Risk Associated with Various Red List Criteria The most important difference between the two approaches is the effect of status changes in less-threatened or nonthreatened species. The equal-steps approach gives an index that is heavily influenced by movements of species among the lower categories of threat. The extinction risk approach gives an index that is largely influenced by movements of species among the higher threat categories. For example, if a Vulnerable species improves in status and becomes Near Threatened, and at the same time, a Critically Endangered species goes Extinct, the RLI based on equal-steps weights registers no change, but the index based on extinction risk weights shows a substantial decrease. Downlisting of a Vulnerable species to Near Threatened might represent a very substantial population increase, whereas extinction of a Critically Endangered species might represent the loss of very few individuals. The latter is arguably more significant in terms of genetic diversity, but the former might be more important as an indicator of wider biodiversity trends. Thus, the extinction risk weights emphasise the loss of biodiversity owing to imminent or potential extinctions of species, whereas the equal-steps weights allow the index to capture large changes in the populations of less-threatened species. For the RLI for complete taxonomic groups, and for disaggregating the index to show trends for subsets of species, for example, in particular realms or ecosystems we used the equal-steps approach because the number of species moving between the higher threat categories (those effectively driving trends when an ‘extinction risk' weighting is used) was too small to be meaningful in disaggregated indices. Only 23% of all genuine status changes (58 species in total) involved moves in or out of the highest threat categories. However, for examining trends in the species closest to extinction, we used the weights based on relative extinction risk. Weaknesses of RLIs The usefulness of the IUCN Red List as an indicator of trends in the status of biodiversity (e.g., Smith et al. 1993 ) has been previously questioned on the grounds that (1) the categories are subjective; (2) taxonomic treatment is uneven, and listings are biased towards attractive, spectacular, high-profile, or better-known species; and (3) most species move between categories because of changes in knowledge or taxonomy, not as a consequence of genuine improvement or deterioration in status ( Cuarón 1993 ; Burgman 2002 ; Possingham et al. 2002 ; but see Lamoreux et al. 2003 ). The first of these problems has been addressed since 1994, when quantitative and objective categories and criteria for the IUCN Red List were introduced ( IUCN 1994 , 2001 ). The second problem can be overcome by calculating indices only for taxonomic groups in which all species have been comprehensively assessed and reassessed (as shown here) or by developing indices based on a stratified sample from diverse taxonomic groups (see below). The third problem has already been addressed because since 2001 the IUCN Red List has required clear documentation of the reason for any reclassification ( IUCN 2001 ). Hence, movements of species between categories owing to knowledge, taxonomy, or other ‘nongenuine' reasons can be easily excluded when calculating the index. RLIs have a fairly coarse level of resolution of status changes because of the broad nature of Red List categories. Populations in the wild may have to undergo quite significant changes in size, trend, or distribution before crossing the thresholds to qualify for a higher or lower Red List category and, hence, before changing the RLI value. This is inherent in using the Red List categories rather than more precise parameters such as estimates of population size. It is not always true, however: The Red List criteria allow for species to be assessed as threatened on the basis of projected declines, and thus changes in status can reflect new or emerging threats in anticipation of population or range changes. We suggest that the disadvantage of coarse resolution is outweighed by the advantage of using a system that allows all the world's species in a taxonomic group to be assessed, rather than just a (potentially biased) subset for which adequately detailed information is available. Insensitivity of the index to status changes may also arise from time lags between changes in a species' population or range and changes in the RLI value, because of delays before detection of the status change, and/or before this knowledge becomes available to assessors. This is potentially more problematic, but several factors act to mitigate it. The Red List Programme partners have a large and expanding network of scientists across the world providing detailed and up-to-date information for an increasing number of species. Furthermore, with improving channels of communication (in particular, the increasing use of the World Wide Web to solicit information, for example, BirdLife's Web-based Globally Threatened Bird discussion forums; BirdLife International 2004a ), we expect that such delays will diminish. For birds, the data support this: whereas just 42% of genuine status changes between 1988 and 1994 were detected in 1994 (with 43% detected during 1994–2000 and 15% detected during 2000–2004), 88% of changes during 1994–2000 were detected in 2000, and just 12% were detected in the subsequent 4 y. Using the data from the 1994–2000 period (because information gathering has improved considerably since 1988–1994), we can estimate the likely number of genuine status changes for 2000–2004 that have not yet been detected (six; see Materials and Methods ) and, hence, estimate the possible degree of error associated with the 2004 RLI value. The results show that it may be an under- or overestimate by 0.21%–0.37% (see Figure 1 ): a small and acceptable margin of error. In future, we anticipate smaller retrospective adjustments to the index values, and a smaller and more predictable error associated with the most recent index value. The major advantage of backtracking status changes to the appropriate time period is that the index trends do not get distorted by the belated discovery of genuine status changes, which, for example, might result from the exhaustive research that takes place when a Red Data Book is published (e.g., BirdLife International 2001 ). This arguably outweighs the disadvantage that the slope of the index between two particular dates may change slightly in future releases of the index. How robust are RLIs? A potential criticism is that they are based on status changes in small numbers of species. However, between 1988 and 2004 the RLI declined by a degree equivalent to almost 10% of species in the categories Near Threatened to Critically Endangered deteriorating in status sufficiently to be uplisted by one category to a higher category of threat. Although relatively few in number (250), these status changes are the most important among the world's birds in terms of changes in projected extinction risk. We therefore suggest that the declines shown by the RLI since 1988 represent very significant losses to global biodiversity. Relatively large numbers of species changed categories in 1994 and 2000 owing to improvements in knowledge and improved consistency of interpretation of information against the Red List criteria (see Table 2 ). This was because of the introduction of quantitative criteria for assigning species to categories in 1994 ( Collar et al. 1994 ; IUCN 1994 ) and the mapping of all threatened species and more rigorous justification for Near Threatened status in 2000 ( BirdLife International 2000 ). By 2000–2004, only 6.7% of threatened and Near Threatened species changed category owing to improved knowledge (see Table 2 ). Nevertheless, it is true that a small proportion of species may be sufficiently poorly known that there is uncertainty over their status and whether this has changed over time. If this introduces any bias, it may be towards an overoptimistic RLI trend. This is because well-studied species (with better data and hence more certain Red List assessments) may be more likely to be those receiving conservation attention and, hence, improving in status (or at least deteriorating less rapidly). All data used in Red List assessments for birds (e.g., population size, trends, etc.) are coded for data quality, and in future the RLI will also be calculated separately for species with high-quality data, in order to test whether such biases exist. Strengths of RLIs The greatest strength of the RLIs presented here is that they are based on comprehensive and complete assessments of nearly all species in a taxonomic group across the world (just 0.8% of birds are listed as Data Deficient and hence excluded from the calculation of the RLI). Most other global indicators based on, for example, population estimates, are derived from data biased towards common, well-studied species in the developed world, particularly Europe and North America. For example, the Living Planet Index ( Loh 2002 ) is based on indices for populations in marine, freshwater, and forest ecosystems. However, 70% of the 195 populations contributing to the freshwater ecosystem are in Europe or North America, while just 18% of the 282 populations contributing to the forest ecosystem index are in the tropics, where the greatest biodiversity is found ( Loh 2002 ). Similarly, in a global index based on data from 936 amphibian populations from 37 countries around the world, 89% of populations (835) were from Europe or North America, and just 2.2% (21) were from Asia and 5.5% (51) from South/Central America ( Houlahan et al. 2000 ). By contrast, the RLI for birds is based on trends for nearly all the world's 10,000 bird species. RLIs for other completely assessed taxonomic groups are in development (see below). At present, indicators based on more representative suites of species are only available for particular countries or regions, such as the United Kingdom headline indicator for birds ( Gregory et al. 2003 ) and the Pan-European Common Bird Index ( BirdLife International 2004b ; Gregory et al. 2004 ). Indices based on population trends (particularly at the regional scale) generally include few species that are rare, localised, or difficult to survey, including those most susceptible to extinction. RLIs can incorporate status changes in such species because the Red List process is an effective system for making meaningful inferences from data that are imprecise or incomplete. Species-based indicators such as the RLI arguably provide far more powerful measures of biodiversity loss than other indicators proposed for measuring progress towards the 2010 target ( CBD 2004 ). Trends in the extent of biomes and habitats are of necessarily coarse resolution and take no account of the distribution of biodiversity within and between habitats; trends in the genetic diversity of domesticated animals and cultivated plants provide measures related to only a tiny proportion of biodiversity; and trends in the coverage of protected areas are a measure of responses to biodiversity loss rather than a measure of the state of biodiversity. Future Steps At present, data are only available for birds to produce the sorts of indices shown here. By 2010 at least two complete global assessments will also be available, and RLIs calculated for all the world's mammals (about 5,000 species), amphibians (about 5,700 species), and hopefully some plant and marine groups. Additional indices, and an aggregation of RLI trends in multiple groups, will provide a more representative picture of the changing state of biodiversity. In recognition that this will take some time to implement, the IUCN Red List Programme is also developing a sampled RLI based on a stratified sample of about 3,000 species from all major taxonomic groups, biogeographic realms, ecosystems, and Red List categories. This will provide an index that may be more representative of trends in the threat status of all biodiversity. We suggest that RLIs will have a key role to play alongside other types of indicators in assessing progress towards reducing the rate of, or halting, the loss of biodiversity. Materials and Methods IUCN Red List assessments for birds BirdLife International (formerly the International Council for Bird Preservation) has been responsible for providing the assessments of the world's 10,000 or so species of birds for the IUCN Red List since 1963. Since 1988, BirdLife has assessed every species of bird on a regular basis, and birds are regarded as the most comprehensively documented class of organisms on the Red List. BirdLife is the official Red Listing Authority for birds, and assessments are based on data gathered from the BirdLife Partnership of organisations in over 100 countries around the world, from published and unpublished literature, and from information provided by a worldwide network of over 1,000 species experts ( BirdLife International 2000 , 2004c ). The principal categories on the IUCN Red List are: Extinct, Extinct in the Wild, Critically Endangered, Endangered, Vulnerable, Near Threatened, and Least Concern ( IUCN 2001 ). Since all bird species have been assessed, none is listed as Not Evaluated, and only 78 (0.8%) are listed as Data Deficient. In addition, two terms used by BirdLife have not yet been adopted for more general application in the IUCN Red List. Possibly Extinct is a tag applied to those Critically Endangered species that are, on the balance of evidence, ‘likely to be Extinct, but for which there is a small chance that they may still be extant, and hence they should not be listed as Extinct until local or unconfirmed reports have been discounted, and adequate surveys have failed to find the species' ( BirdLife International 2004c ). As there are taxonomic revisions between assessments, ‘Not Recognised' is applied to taxa in those assessments when they were not treated as full species ( BirdLife International 2004c ). Data from which to calculate the indices were derived from four complete assessments of the status of the world's birds by Collar and Andrew (1988) , Collar et al. (1994) , and BirdLife International (2000 , 2004c ), plus from reviews of two species (Houbara bustard [ Chlamydotis undulata ] and Saker falcon [ Falco cherrug ]) whose categories were revised for the 2004 IUCN Red List too late to be included in BirdLife International (2004c ). Information was also taken from partial assessments submitted by BirdLife to the 2002 and 2003 IUCN Red Lists (IUCN 2002 , 2003 ). The 1988 assessment predated quantitative Red List criteria ( IUCN 1994 ), and only the qualitative categories ‘threatened' and Near Threatened were used. Therefore, for species categorised as threatened in 1988, the category assigned in 1994 was applied to the 1988 assessment, with an appropriate category assigned for species that underwent genuine status changes during the period (see below). Identifying genuine status changes between Red List assessments Published lists of numbers of species in different Red List categories cannot simply be used to calculate the index, for several different reasons. For example, changing knowledge and taxonomy result in many category changes, but such revisions are not indicative of changes in the conservation status of species' populations. Hence, to identify those species changing categories between assessments for relevant reasons, a ‘reason for change' code was assigned for each recategorisation. These mutually exclusive codes were (1) recent genuine status change, (2) genuine status change since first assessment, (3) knowledge, (4) criteria revision, and (5) taxonomy. These last three codes were used for changes not relevant for calculating the indices. ‘Recent genuine status change' was applied to species that had undergone a genuine improvement or deterioration in status in the period since the previous assessment. This included species qualifying because of population declines (under IUCN Red List criterion A), particularly those with long generation times, for which more than half of the period during which the change occurred was subsequent to the previous assessment. This code was also applied in a few cases where species entered the Red List as a result of their elevation from subspecies to species level, but for which there had been a genuine status change when the taxon was compared to the equivalent population of the original species. For species categorised as threatened in 1988 and Critically Endangered, Endangered, or Vulnerable in 1994, genuine changes in status between these two assessments were identified by searching the account in Collar et al. (1994) and associated unpublished information for evidence of genuine status changes that had occurred in the previous 6 y. ‘Genuine status change since first assessment' was applied to species that had undergone a genuine improvement or deterioration in status in the period since the first complete assessment, but prior to the last assessment. This code denoted genuine changes in status that were not detected at the time they occurred. For example, Syrian serin ( Serinus syriacus ) was uplisted from Near Threatened to Vulnerable in 2004 because of the discovery that populations had declined during a drought in 1998–1999. This information was unavailable during the 2000 assessment, so the species was recategorised in 2004 and given this ‘reason for change' code ( BirdLife International 2004c ). For cases such as these, the Red List category for earlier assessments was back-cast using the improved understanding of earlier population sizes, trends, and ranges. This also applied to (a) extinctions that occurred after 1988; (b) species where no (or an incorrect) status change was recorded, but subsequent knowledge indicated that a genuine status change had occurred; (c) species for which an improvement in status did not immediately lead to a category revision because of the application of the ‘5-y rule' (whereby under the IUCN Red List guidelines downlisting to a lower category of threat should not occur until all of the criteria of the higher category have ceased to apply for 5 y or more; IUCN 2001 ; Red List Standards and Petitions Subcommittee 2003 ). ‘Knowledge' was applied to species recategorised owing to new information on, for example, population and range size, declines, ecological attributes, threats, or conservation efforts. This included information published or known before the last assessment, but only made available to, or discovered by, assessors since the last assessment. For example, Madagascar plover ( Charadrius thoracicus ) was uplisted from Near Threatened to Vulnerable in 2004 because its population was estimated to number as few as 750 individuals. However, the evidence suggests that the population may have been this small since before 1988 ( BirdLife International 2004c ). This code was also applied in cases where a species changed category owing to improved consistency of interpretation of information against the Red List criteria. ‘Criteria revision' was applied in cases when species changed category owing to revisions to the definitions of the IUCN Red List criteria ( IUCN 2001 ). ‘Taxonomy' was applied in cases when species changed category owing to taxonomic ‘lumping' or ‘splitting' or for newly described species. Calculating index values The number of species in each Red List category for each complete assessment and the number of species that changed categories as a result of genuine status changes were used to determine the index value in the following way: (1) For species assessed in two consecutive assessments (i.e., excluding any listed as Not Recognised, Not Evaluated, or Data Deficient in either or both assessments), the total numbers of species in each category in the earlier assessment (excluding Extinct and Possibly Extinct, but including those species retrospectively reassigned categories owing to genuine status changes that were identified subsequently; see above) were multiplied by a weight, and these were summed to give a total score, T, for each assessment. The weights were set as Near Threatened = 1, Vulnerable = 2, Endangered = 3, Critically Endangered = 4, Extinct in the Wild = 5 (see below). (2) Over each period between complete assessments (1988–1994, 1994–2000, and 2000–2004) the net number of genuine changes to the total in each category was calculated, multiplied by the weights above (with Possibly Extinct and Extinct = 5), and summed to calculate the proportional change in the total score, P. (3) The value of the index (I) was set to 100 in 1988. For subsequent assessments I was calculated by multiplying − P for the previous period by the previous index value (see Table S1 for values for T, P, and I for each index). Mathematically, the method can be described as follows, where T is total score; N c (t i ) is the number of species in category c at time t i , where t i is the year of the i th assessment (assessments are not necessarily made every year); W c is the weight for category c; P is proportional genuine change; I ti is the value of the index at time t i ; Cat(t i , s) is the category of species s at time t i ; W c is the weight for category c; G s = 1 if change (from t ( i −1) to t i ) in category of species s is genuine (otherwise G s = 0). where I ti−1 = 100 for the first year of assessment. The Red List categories are ordinal ranks, whereby each step from Least Concern to Extinct indicates that at least one measure of extinction risk has become worse. The ‘equal-steps' weights listed above reflect the ordinal ranks of the categories. However, the steps between lower categories (e.g., Near Threatened to Vulnerable) translate to smaller increases in extinction risk than steps between higher categories (e.g., Endangered to Critically Endangered). Therefore we also calculated the aggregated RLI using weights based on the relative extinction risk associated with each category. Several of the quantitative thresholds in the Red List criteria can be used to obtain approximate values for the relative risk of extinction (for species at the lower boundary of that category). The most obvious is criterion E, which is based on quantitative analysis of extinction probability. The quantitative thresholds in criterion E change for both extinction probability and time frame for the three categories, and depend on generation length (e.g., the threshold for Endangered is a probability of extinction in the wild greater than 20% within 20 y or 5 generations). Taking a 3-generation time frame, a generation length set arbitrarily at 5 y, and assuming a constant annual risk of extinction, the 3-generation probabilities are approximately 0.5, 0.13, and 0.016 for Critically Endangered, Endangered, and Vulnerable, respectively (H. R. Akçakaya, unpublished data). However, most extinctions do not occur as a result of random catastrophes, as implied by the assumption of the constant annual risk of extinction. Many are preceded by declines, resulting in sigmoid extinction risk curves (with probability of extinction as a function of time). For such cases, the 3-generation probabilities are approximately 0.5, 0.1, and 0.025 for Critically Endangered, Endangered, and Vulnerable, respectively (see Table 3 ). Comparable extinction risks can also be calculated based on other Red List criteria (except A, for which there is no obvious method). Assuming that the number of mature individuals (in criteria C1 and D), range or extent of occurrence (criterion B1), and area of occupancy (criterion B2) are inversely related to risk of extinction, and fixing the risk of extinction for Critically Endangered at 0.5, it is possible to calculate the probabilities for categories Endangered and Vulnerable (see Table 3 ). Based on the geometric average of these estimates, the weights for Critically Endangered, Endangered, and Vulnerable are determined as 0.5, 0.05, and 0.005 (see Table 3 ). The weight for Extinct (and hence Extinct in the Wild and Possibly Extinct) is by definition 1.0. The weight for Near Threatened is set at 0.0005, keeping the same proportion as among the weights for the three threatened categories. Calculating error bars We calculated, using the following method, the possible range of error associated with the latest (2004) RLI value owing to time lags before genuine status changes are detected ( see Discussion ). We estimated how many such undetected category changes there may be for 2000–2004 using the 1994–2000 data (information gathering has improved considerably in recent years, so comparisons with time lags for the 1988–1994 period are not meaningful). In total, 128 genuine changes for 1994–2000 were identified in 2000, and an additional 17 (13.3%) were identified in the subsequent 4 y. This suggests that an additional six category changes (13.3% of 45 genuine status changes identified in 2004) may be belatedly detected for 2000–2004. We randomly selected six species from the 9,453 species that did not undergo category changes from 2000 to 2004, with a maximum of two species per category. We ran 10,000 simulations of six species moving to categories of higher extinction risk, with probabilities for each number of category steps set by the distribution of category changes for 35 ‘uplisted' species in 2000–2004. The maximum value for P (proportional genuine change) from these simulations gave the lower error bar for the 2004 RLI value. Similarly, we ran 10,000 simulations of six species moving to categories of lower extinction risk (with probabilities for each number of category steps set by the distribution of category changes for ten ‘downlisted' species in 2000–2004), and took the minimum value for P to give the upper error bar. These are very similar to the minimum and maximum values for P derived by simulating an additional six species moving in a direction (and by a number of categories) in proportion to the distribution of these values for all 45 species that underwent genuine status changes in 2000–2004 (see Table S2 ). Disaggregating indices One of the purposes of the RLIs is to illustrate trends over time in the threat status of species in different biogeographic realms, ecosystems and families or species groups. Species were assigned (based on native distributions, excluding cases of vagrancy) to one or more biogeographic realms (Palearctic, Afrotropical, Indo-Malayan, Nearctic, Neotropical, and Australasian/Oceanic) following the boundaries mapped by Newton (2003) except that Australasian was pooled with Oceanic, and Antarctic was excluded. Where a species was assigned to more than one realm, it was included in calculating the score (T) for each of those realms. This is because a species could potentially undergo genuine changes in status in any or all realms in which it occurs. However, so that trends in indices for particular realms reflect changes in the threatening processes operating within each particular realm (rather than threats operating elsewhere in the range of the species), species were only included in the calculation of P for a particular realm if the genuine status change had been driven by factors operating in that realm. For example, Saker falcon (F. cherrug) occurs in the Palearctic, Indo-Malayan, and Afrotropical realms and was included in the score calculations for each of these. However, recent declines have been driven by factors operating on the breeding grounds in Central Asia (Environmental Research and Wildlife Development Agency, unpublished data; BirdLife International 2004c ), so the genuine change was only calculated for the Palearctic realm. By contrast, the black-browed albatross (Thalassarche melanophrys) has declined as a result of incidental capture in commercial longline fisheries in oceans in the Afrotropical, Neotropical, and Australasian/Oceanic realms ( Robertson and Gales 1998 ; BirdLife International 2004c ), and so this genuine change was incorporated into the calculation of P for all three realms. The index was disaggregated for ecosystem (terrestrial, marine, and freshwater) and for two terrestrial habitat types (forest and shrubland/grassland; see below) in a similar way. Species were included in the calculation of T for all ecosystems and habitats for which they were scored, but only included in the calculation of P for a particular ecosystem or habitat if the genuine status change had been driven by threatening processes operating in that ecosystem or habitat. Species were only assigned to a habitat type if this was of critical or major importance (i.e., the species typically occurs in no other habitat, or just one other habitat at some point in its life cycle). To exemplify how the approach can be used for taxonomic subsets of species, RLIs were also calculated for several high-profile species groups with specific conservation interest groups: raptors (Falconiformes), game birds (Galliformes), and parrots (Psittaciformes); and for species groups relevant to particular international conservation treaties: waterbirds (as listed in Wetlands International 2003 ) covered by the Ramsar convention, migrant species covered by the CMS, and albatrosses (Diomedeidae) and large petrels ( Macronectes spp. and Procellaria spp.) covered by the ACAP under the CMS. Sample sizes in the figure legends give, for the subset of species plotted, the total number of category changes owing to genuine status changes during 1988–2004 (but note that a small number of species underwent genuine status changes in more than one period between assessments) and the total number of species in categories Extinct in the Wild, Critically Endangered, Endangered, Vulnerable, and Near Threatened in at least one assessment during the period (and that are taxonomically recognised at present). Supporting Information Table S1 Index Values: Values for T, P, and I for Each Period and for Each Index (66 KB DOC). Click here for additional data file. Table S2 Calculating Error Bars: Simulated P- Values (Proportional Change in the Index Score T ) Used to Determine Error Bars for 2004 RLI Value Lists the simulated p- values (proportional change in the index score T ) based on the assumption that an additional six genuine changes occurred from 2000 to 2004 but have not yet been identified owing to time lags in knowledge (see Materials and Methods ). Six species were randomly selected from those that did not change category from 2000 to 2004 ( n = 9,453 species), with a maximum of two species from each category. For each species, the number and direction of category changes were randomly assigned with probabilities (1) based on the change in categories for the ten species that underwent genuine status changes and were downlisted to a lower category of threat during 2000–2004 (‘only down'); (2) based only on the 35 species that were uplisted to a higher category of threat (‘only up'); (3) based on all 45 species (‘both up and down'); and (4) based on all 45 species with probabilities that were set individually for each threat category (‘category dependent,' so that, e.g., the probability of an Least Concern species being downlisted to a lower category of threat was zero). In each case, the procedure was repeated 10,000 times to calculate the minimum, maximum, mean, and standard deviation of the simulated p- value. The upper error bars for the RLI were determined by the minimum simulated p- value for cases when all six species were downlisted to lower categories of threat (shown in red in the table). The lower error bars were determined by the maximum simulated p- value for cases when all six species were uplisted to higher categories of threat (shown in blue in the table). In both cases the values are close to those calculated if the six species changed category, with probabilities based on the direction and number of category changes for all 45 species, and encompass those derived using the method based on category-dependent probabilities. (35 KB DOC). Click here for additional data file.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524254.xml
506787
Tumor necrosis factor-mediated inhibition of interleukin-18 in the brain: a clinical and experimental study in head-injured patients and in a murine model of closed head injury.
Tumor necrosis factor (TNF) and interleukin-(IL)-18 are important mediators of neuroinflammation after closed head injury (CHI). Both mediators have been previously found to be significantly elevated in the intracranial compartment after brain injury, both in patients as well as in experimental model systems. However, the interrelation and regulation of these crucial cytokines within the injured brain has not yet been investigated. The present study was designed to assess a potential regulation of intracranial IL-18 levels by TNF based on a clinical study in head-injured patients and an experimental model in mice. In the first part, we investigated the interrelationship between the daily TNF and IL-18 cerebrospinal fluid levels in 10 patients with severe CHI for up to 14 days after trauma. In the second part of the study, the potential TNF-dependent regulation of intracerebral IL-18 levels was further characterized in an experimental set-up in mice: (1) in a standardized model of CHI in TNF/lymphotoxin-α gene-deficient mice and wild-type (WT) littermates, and (2) by intracerebro-ventricular injection of mouse recombinant TNF in WT C57BL/6 mice. The results demonstrate an inverse correlation of intrathecal TNF and IL-18 levels in head-injured patients and a TNF-dependent inhibition of IL-18 after intracerebral injection in mice. These findings imply a potential new anti-inflammatory mechanism of TNF by attenuation of IL-18, thus confirming the proposed "dual" function of this cytokine in the pathophysiology of traumatic brain injury.
Findings Closed head injury (CHI) is the leading cause of mortality and persisting neurological impairment in young people in industrialized countries [ 1 , 2 ]. The neuropathological sequelae of brain injury are mediated in large part by a profound host-mediated intracranial inflammatory response [ 3 - 5 ]. The pro-inflammatory cytokines tumor necrosis factor (TNF) and interleukin (IL)-18 have been identified as crucial mediators of neuroinflammation after brain injury [ 6 - 9 ]. This notion has been supported by experimental studies in rodents which demonstrated neuroprotective effects by pharmacological inhibition of either TNF or IL-18 after CHI [ 9 - 11 ]. In recent years, the concept of a "dual role" evolved with regard to concomitant beneficial and adverse effects of pro-inflammatory mediators, depending on the kinetic of their expression and posttraumatic regulation in the injured brain [ 3 , 12 , 13 ]. However, the TNF-dependent regulation of IL-18 in the injured brain has not yet been investigated. We sought to determine the interrelationship between intracranial TNF and IL-18 levels in a clinical study on patients with severe CHI and in an experimental model in mice. Patients with isolated severe closed head injury ( n = 10, Glasgow Coma Scale score ≤ 8) and indication for intraventricular catheters for cerebrospinal fluid (CSF) drainage due to increased intracranial pressure (ICP > 15 mm Hg) were included in this study. Drained CSF was collected daily for up to 14 days after trauma or until catheters were removed. The patient characteristics are shown in Table 1 . No patient was treated with steroids. The protocol for daily CSF collection is in compliance with the Helsinki Declaration and was approved by the University's Ethics Board Committee. Control CSF was collected from patients undergoing diagnostic spinal tap ( n = 10) and revealed no inflammatory CNS disease, based on normal CSF protein and glucose levels and normal white cell counts. All samples were kept on ice at 4°C and immediately centrifuged after collection, aliquoted, and frozen at -70°C until analysis. Table 1 Clinical data and intrathecal cytokine levels in patients with severe closed head injury. Patient Age (years) / Gender Type of brain injury (Marshall score) Outcome (GOS) TNF in CSF (pg/mL) IL-18 in CSF (pg/mL) Correlation r S Mean Range Mean Range 1 38 / M EML 4 6.4 1.0 – 11.5 40.6 6.5 – 155.2 - 0.804 ** 2 30 / M DI II° 3 3.6 1.0 – 7.7 114.3 29.7 – 286.4 - 0.580 * 3 56 / M EML 4 6.3 1.0 – 10.0 35.1 11.2 – 100.3 - 0.530 4 57 / F DI II° 5 6.0 1.0 – 11.7 20.1 5.0 – 168.8 - 0.761 ** 5 44 / M EML 4 1.6 1.0 – 3.4 39.8 22.6 – 74.5 - 0.751 * 6 26 / M EML 4 3.2 1.0 – 10.3 108.5 5.0 – 328.6 - 0.832 ** 7 47 / M EML 1 1.1 1.0 – 1.4 268.5 78.3 – 462.0 - 0.372 8 25 / M EML 4 2.2 1.0 – 4.0 91.6 10.3 – 290.0 - 0.195 9 37 / F DI III° 3 1.6 1.0 – 2.7 183.7 21.5 – 382.2 - 0.844 ** 10 35 / M DI II° 4 2.0 1.0 – 5.8 209.4 19.9 – 391.8 - 0.772 * Controls ( n = 10) 1.0 1.0 – 7.1 5.0 5.0 – 8.4 Statistical analysis for assessment of the correlation between tumor necrosis factor (TNF) and interleukin (IL)-18 levels in serial cerebrospinal fluid samples for up to 14 days after trauma was performed by Spearman's rank correlation (* P < 0.05, ** P < 0.01). The patients' outcome was determined at 3 months after injury by the Glasgow Outcome Scale (GOS) score [33]: 5 = asymptomatic, 4 = moderate disability, 3 = severe disability, 2 = persisting vegetative state, 1 = death. The type of brain injury was classified by the CT-scan criteria established by Marshall et al. [34] into diffuse injury (DI) grade I-III and evacuated vs. non-evacuated mass lesions (EML, NEML). Quantification of IL-18 and TNF levels in human CSF samples and in murine brain homogenates was performed by species-specific commercially available ELISA (R&D Systems, Abingdon, UK). According to the information provided by the manufacturer, the IL-18 assay recognizes both the mature and the pro-form of IL-18. All concentrations below the detection limit of 5 pg/mL (IL-18) or 1 pg/ml (TNF) were assigned a value of 5 pg/mL, and 1 pg/ml respectively. All samples were run undiluted in duplicate wells. The concentrations were calculated from the mean OD of duplicate samples, determined by spectrophotometer (Dynatech Laboratories Inc., Chantilly, VA, U.S.A.) at an extinction wavelength of 405 nm. The experimental part of the study was set-up on two different protocols with the aim to assess the TNF-dependent regulation of IL-18 in the murine brain: (1) The first part of the experimental study was designed to investigate a potential role of TNF-dependent regulation of intracranial IL-18 expression in a standardized model of CHI, using mice double-deficient in genes for TNF and lymphotoxin-α (TNF/LT-α-/-) [ 14 ]. These knockout mice were selected in order to compensate for potential redundant functions of TNF by LT-α which binds to the identical common receptors (i.e. TNF receptors p55 and p75) [ 14 - 16 ]. The generation and development of the TNF/LT-α-/- mice on a mixed C57BL/6 × 129Sv/Ev (B6 × 129) genetic background has been previously described [ 14 ]. Knockout mice and wild-type (WT) littermates of the B6 × 129 strain were subjected to a CHI ( n = 10 per group) using a standardized weight-drop model, as previously described [ 9 , 17 ]. In brief, following ether anesthesia, a midline longitudinal scalp incision was performed. Trauma was applied to the left anterior frontal area of the exposed skull by a 330 g weight with a silicon tip dropped from a height of 2 cm, resulting in a focal closed injury to the left hemisphere. Mice received supporting oxygenation with 100% O 2 until they awakened and were then brought back to their cages. Control animals were subjected to anesthesia and sham operation only ( n = 10 per group). In addition, mice with anesthesia alone ( n = 8) were used as internal control and untreated control animals ( n = 10) were analyzed for baseline evaluation of intracerebral cytokine profiles in these mice. (2) In the 2 nd part of the experimental study, mice of the C57BL/6 strain ( n = 10 per group) were treated by stereotactic intracerebroventricular (i.c.v.) injection of either 200 ng mouse recombinant TNF in 10 μl PBS, or vehicle solution only (10 μl PBS), into the left hemisphere using a sterile 27-gauge syringe, under ether anesthesia. According to data from previously published studies [ 18 - 20 ], as well as based on titration studies from our own laboratory, the i.c.v. injection of 200 ng mouse-recombinant TNF (R&D Systems) elicited an evident induction of inflammatory changes in the murine CNS, such as intracranial recruitment of leukocytes and development of brain edema in the injected hemisphere within 24 hours (data not shown). Animals from all groups (CHI and i.c.v. injection) were sacrificed at 24 h after the respective procedure, which corresponds to the time-point of maximal extent of intracerebral inflammation in the model of CHI used in this study [ 17 ]. For assessment of intracerebral IL-18 levels, the murine brains were immediately removed after decapitation. Tissue homogenization was performed using a Polytron homogenizer (Kinematica, Kriens, Switzerland) with a dilution of 1:4 in ice cold extraction buffer containing 50 mmol/L Tris buffer (pH 7.2), NaCl 150 mmol/L, Triton-X-100 1%, and protease inhibitor cocktail (Roche, Mannheim, Germany). The homogenates were shaken on ice for 90 minutes, centrifuged for 15 minutes at 3,000 g and 4°C. Thereafter, the supernatants were aliquoted and stored at -70°C until analysis. The concentrations of total protein in the brain extracts were measured by Bradford assay (Bio Rad Laboratories, Munich, Germany). The intracerebral protein concentrations were in a similar range in all mice assessed (11.7 ± 2.4 mg/mL; mean ± SD). The quantification of IL-18 levels in murine brain homogenates was performed as described above for the human samples. All mice used in this study were males, in order to avoid a bias in gender, age 12 to 16 weeks with body weights of 25 to 32 g. The animal experiments were performed in compliance with the guidelines of the Federation of European Laboratory Animal Science Association (FELASA) and approval was granted by the Institutional Animal Care Committee of the University of Zurich and of the Hebrew University of Jerusalem. In the clinical part of the study on CHI patients, the mean IL-18 concentrations in ventricular CSF collected up to 14 days after trauma were significantly higher than in control lumbar CSF from patients undergoing diagnostic spinal tap ( P < 0.05, repeated measures ANOVA; Table 1 ). These findings are coherent with data from previously published studies which demonstrated significantly elevated intracranial IL-18 levels after brain injury, both in humans as well as in experimental model systems [ 8 , 9 ]. With regard to TNF, the mean levels in individual serial cerebrospinal fluid samples were significantly elevated in 50% of all head-injured patients, compared to controls (patients #1,2,3,4,6; Table 1 ). Elevated intracranial TNF levels after traumatic brain injury have been previously reported in various clinical and experimental studies [ 21 - 26 ]. In the present study, the individual daily TNF levels in CSF were up to 10- to 100-fold lower than the corresponding IL-18 levels (Table 1 ). Interestingly, despite the fairly low TNF levels in CSF we found an inverse correlation between the daily individual intrathecal TNF and IL-18 levels in all trauma patients, as demonstrated by a negative Spearman's rank correlation coefficient ( r = -0.195 to -0.844). In 7 of 10 patients, this inverse correlation was statistically significant, with a P -value < 0.05 in three patients (# 2,5,10) and P < 0.01 in four patients (# 1,4,6,9; Table 1 ). Since the quality of the blood-brain barrier has not been determined in this cohort of head-injured patients, due to the lack of matching serum samples for assessment of albumin levels, the source of elevated cytokines (intrathecal compartment vs. peripheral serum) remains unclear. In the experimental study, IL-18 was found to be constitutively expressed in the brain of untreated mice (27.7 ± 5.4 ng/mL, mean ± SEM; "baseline", Fig. 1 ), which is in accordance with data from previous studies revealing constitutive IL-18 expression in the CNS of normal rats and mice [ 8 , 9 , 27 , 28 ]. Microglia may represent the cellular source of constitutive intracranial IL-18 levels in these mice, since Prinz and colleagues have previously shown that microglial cells, but not astrocytes, produce IL-18 in the murine CNS [ 28 ]. The intracerebral IL-18 levels increased significantly in the head-injured group by 24 hours after experimental CHI (56.9 ± 4.7 ng/mL, P < 0.01 vs. baseline, Fig. 1 ). Knockout mice lacking TNF and LT-α genes also showed significantly elevated IL-18 concentrations at 24 h after CHI (58.4 ± 7.8 ng/mL) which were in a similar range as in head-injured WT mice, as shown in Fig. 1 . This lack of differences in mice deficient in the ligands for TNF receptors may be explained by alternatively expressed inflammatory mediators or modified pathways of IL-18 regulation in these genetically engineered mice which have been shown to have significantly altered immune responses [ 14 ]. Interestingly, the intracerebral injection of vehicle alone (10 μl PBS i.c.v.) induced significantly elevated IL-18 levels in brains of normal WT mice, compared to normal untreated mice, which were in a similar range as in the brain-injured groups (53.6 ± 9.6 ng/ml, Fig. 1 ). These data imply that a minor penetrating injury by i.c.v. injection of a small volume of buffer solution represents a procedure which is sensitive enough to induce significant IL-18 production in murine brains within 24 hours. In contrast, the intracerebral injection of murine recombinant TNF (in 10 μl PBS) reduced the elevated IL-18 levels in murine brains significantly to levels than were even lower than baseline concentrations (22.13 ± 7.1, P < 0.01 vs. vehicle-injected mice), as shown in Fig. 1 . Figure 1 Intracerebral IL-18 concentrations in mice, as determined by ELISA in murine brain homogenates ( n = 10 animals per group). Untreated normal mice were used for determination of baseline IL-18 levels in this study. The four treatment groups were sacrificed after 24 hours for assessment of intracerebral IL-18 levels, as described in the text. Mice deficient in genes for TNF and lymphotoxin-α (TNF/LT-α-/-) and wild-type (WT) littermates were subjected to focal closed head injury (CHI) and sacrificed after 24 hours. Two additional groups of WT mice were given an intracerebro-ventricular (i.c.v.) injection of 200 ng mouse-recombinant TNF in 10 μl PBS or by vehicle alone (10 μl PBS i.c.v.). The data are presented as means ± SEM. * P < 0.01 vs. baseline / TNF-injection (unpaired Student's t -test). The findings from these experimental investigations corroborate the data from the clinical study, where an inverse correlation of intrathecal TNF and IL-18 levels during the first 14 days after severe CHI was found, suggesting that the inhibition of IL-18 may represent a new potential anti-inflammatory mechanism after CHI. Such a "dual role" of TNF has been suggested previously in terms of concomitant pro- and anti-inflammatory effects and detrimental as well as beneficial neuroprotective properties after brain injury [ 12 ]. While the pro-inflammatory effects mediated by TNF in the CNS have been thoroughly investigated in the past two decades [ 15 , 29 , 30 ], the concept of anti-inflammatory effects mediated by cytokines which have been formerly designated as "pro-"inflammatory mediators, is still challenging and novel. Scherbel and colleagues were the first to provide evidence of beneficial effects of TNF in the later phase (i.e. 4 weeks) after traumatic brain injury, based on studies in TNF-/- mice subjected to controlled cortical impact brain injury [ 31 ]. In these experiments, the TNF-deficient mice showed a significantly attenuated neurobehavioral impairment than WT littermates in the first 48 hours post trauma, in terms of early detrimental effects mediated by TNF [ 31 ]. However, the expected neurobehavioral recovery was absent in TNF-/- mice after 4 weeks and cortical tissue loss was significantly increased at this time-point, compared to WT littermates, implying that at later time-points the lack of TNF leads to adverse outcome after brain injury [ 31 ]. Barger and colleagues have previously shown in an in vitro model of amyloid β-mediated neurotoxicity that both TNF and LT-α (formerly designated TNF-β) can significantly attenuate neuronal degeneration by induction of antioxidative pathways through activation of the transcription factor NFκb, thus strengthening the notion of neuroprotective effects mediated by these cytokines [ 32 ]. This assumption was further corroborated in the setting of experimental CHI, where mice lacking both TNF and LT-α genes were shown to have a significantly increased mortality within one week after trauma, compared to WT littermates [ 17 ]. Overall, these data support the notion that TNF exerts detrimental effects in the early phase and beneficial neuroprotective effects in the later phase after head injury [ 12 ]. However, the assumptive underlying regulatory mechanisms of TNF-mediated neuroprotection and of TNF-mediated suppression of IL-18 in the injured brain remain unclear and have to be investigated in future experimental studies. List of abbreviations Central nervous system (CNS), cerebrospinal fluid (CSF), closed head injury (CHI), intracerebroventricular (i.c.v.), interleukin (IL), lymphotoxin-α (LT-α / TNF-β), nuclear factor κB (NFκB), tumor necrosis factor (TNF), phosphate-buffered saline (PBS), intracranial pressure (ICP), wild-type (WT), enzyme-linked immunosorbent assay (ELISA). Competing interests There are no financial interests by any of the authors with regard to the present project. Authors' contributions OIS, MCMK, CEH, ES, and PFS were responsible for conception and planning of the experiments, as well as for performing the animal experiments, collection of the human cerebrospinal fluid samples and cytokine measurements in human and murine tissue samples, as well as for writing of the manuscript. DP and IY performed the experimental i.c.v. injection experiments. WE contributed to the interpretation of the results and writing of the manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC506787.xml
522816
Norovirus gastroenteritis general outbreak associated with raw shellfish consumption in South Italy
Background Despite Noroviruses (NV, previously "Norwalk-like viruses") being a leading cause of acute gastroenteritis outbreaks, the impact of NV infection is at present unknown and little information is available about strains circulating in Italy. In April 2002 an outbreak of gastroenteritis occurred in the province of Bari (South-east Italy), involving several households. Methods A retrospective cohort study was performed in order to assess risk factors associated with illness. All households where a case occurred were included in the study. Faecal specimens were collected from ill individuals. NV-specific RT-PCR was performed. Eleven samples of mussels were collected from fish-markets involved in the outbreak. A nested PCR was used for mussel samples. Results One hundred and three cases, detected by means of active surveillance, met the case definition. Raw shellfish eating was the principal risk factor for the disease, as indicated by the analytic issues (Risk Ratio: 1.50; IC 95%: 1.18 – 1.89; p < 0.001). NVs were found by means of RT-PCR of all the stool specimens from the 24 patients tested. Eleven samples of shellfish from local markets were tested for the presence or NVs; six were positive by nested PCR and genotypes were related to that found in patients' stools. Conclusion This is the first community outbreak caused by NVs related to sea-food consumption described in Italy. The study confirms that the present standards for human faecal contamination do not seem to be a reliable indicator of viral contaminants in mussels.
Background Norovirus (NV, previously "Norwalk-like viruses"), one of four genera in the Caliciviridae family, includes a group of morphologically similar but genetically different single-stranded RNA viruses. NVs represent the most important cause of non-bacterial gastroenteritis worldwide. In industrialised countries NVs may be responsible for up to 80% of all outbreaks of gastroenteritis [ 1 ]. Outbreaks may affect all age groups and generally occur in crowded communities such as restaurants, tourist resorts, hospitals, schools and nursing homes. Contaminated food or water commonly represents the main source of infection. Epidemics spread by the faecal-oral route, even if transmission may also occur through direct person-to-person contact or aerosolised viral particles. The incubation period of NV gastroenteritis is 24–48 hours and symptoms include vomiting, diarrhoea, abdominal pain, low-grade fever, headache and myalgia. Italy has no specific surveillance system for viral gastroenteritis and laboratory diagnosis is only carried out in a few cases. Therefore the impact of NV infection is currently unknown and little information is available about circulating strains. Outbreak investigations are usually performed by local health units. Gastroenteritis notifications are often too delayed to identify the etiologic agent and source of infection correctly. Puglia, a region in the South-East of Italy, has about four million inhabitants. Bari (about 1,500,000 people in the province) is the capital. Raw mussels are largely consumed in Bari and, in general, in Puglia, especially at Christmas and Easter. In 1994 raw seafood consumption was the source of a small cholera outbreak. Moreover, hepatitis A is endemic in this region. This report describes a large outbreak of NV gastroenteritis that involved several households in Bari province. Methods Descriptive and analytical studies The outbreak occurred during the Easter holidays, between March 31 st and April 7 th 2002, in the province of Bari (Puglia region, South Italy). In Italy, at Easter many people eat in restaurants and the day after Easter most people go picnicking. In the province of Bari seafood (especially mussels) are often consumed on these occasions. The Regional Epidemiological Office, alerted by the Local Health Unit, performed a field study, collecting data from all General Practitioners and all Emergency Units in the province. Active surveillance was conducted until April 15 th 2002 (one week after the last case onset). The aim of the investigation was to describe the outbreak and to identify the etiologic agent, the source of infection and the means of transmission. A case was defined as illness in a resident in the area during the period March 31 st – April 7 th , who had diarrhoea (three or more loose stools in any 24 hour period) or vomiting (at least one episode). Fever greater than or equal to 38°C, abdominal pain or nausea were considered additional symptoms. A "probable" secondary case was defined as illness in a household with onset of symptoms more than 24 hours after the primary case. A retrospective cohort study was performed in order to assess risk factors associated with illness. All households where a case occurred were included in the study. Information, collected by means of a standardised questionnaire, included: 1) demographic individual characteristics such as age, gender, occupation; 2) type of food consumed during the last 72 hours before onset of symptoms; 3) if ill, type, date and time of onset of symptoms. To assess the association between food consumption and disease, relative risks (RR) and 95% confidence intervals (95% CI) were calculated. Age and gender were compared between ill and unaffected individuals by the chi squared and Mann-Whitney tests. A p-value less than 0.05 was considered significant. Variables associated to the illness in the univariate analysis were included in the stratified analysis; summary RR and 95% CI were calculated from the formulas of Greenland and Robins [ 2 ]. Collected data were analysed by using Epi Info 6.04 (Centres for Disease Control and Prevention, Atlanta, GA) and Statview 4.0 (Sas Institute Inc., Cary, NC) software. Laboratory investigations Faecal specimens were collected from ill individuals. Part of the specimens was refrigerated and processed within 12 hours to detect ova and parasites by direct microscopy and common bacteria by standard methods. The rest was stored at -20°C until examination by NV-specific reverse transcription-polymerase chain reaction (RT-PCR). Viral nucleic acid extraction and purification from stool specimens was performed as previously reported [ 3 ]. RT-PCR was carried out with the primers JV12-SM31 specific for the polymerase gene of NV [ 4 ]. Eleven samples of mussels were collected from two fish-markets from which the cases had bought the shellfish they consumed. Mussel samples were also processed within 12 hours to detect common bacteria by standard methods. Then a nested PCR was used. The procedure for mussel processing as well as viral RNA extraction and purification has been previously described in full [ 5 ]. The primers used for first round PCR were JV12 and SM31. Nested PCR was carried out with the use of the primers SR33 for negative strand DNA synthesis and SR48 and SR46 for positive strand synthesis of genogroup I (GI) and genogroup II (GII) sequences, respectively [ 6 ]. The 333-bp or 123-bp amplification products from cases and from mussels were subjected to sequencing with PCR primers. When required, cloning was carried out on PCR products. Sequences obtained were aligned with those available in the GenBank. Results Epidemiological issues One hundred and three subjects met the case definition; 22 of them were defined as "probable" secondary cases. Fifty eight (56.3%) of the 103 cases were female; the mean age was 42.6 years (table 1 ). The clinical pattern of the disease was characterised by the presence of vomiting (84.5%), nausea (58.3%), diarrhoea (53.4%), abdominal pain (47.6%), fever (16.5%). No difference was observed in the clinical pattern by gender. Age was significantly higher in ill individuals with severe diarrhoea (lasting more than 24 hours or undergoing specific treatment) (46.3 vs 38.3 years; p = 0.036). No difference in age was found for the other symptoms. The outbreak started on March 31 st at 5,00 pm and lasted eight days. The epidemic curve shows a single peak on April 2 nd (35 cases between 1:00 and 12:00 AM) and a right tail probably due to secondary cases (figure 1 ). Incubation time was not calculated because it was not possible to state a single exposure time. In fact this outbreak could represent the result of multiple small household outbreaks, the source of which could have been either a restaurant meal (on Easter Sunday) or a picnic the day after Easter. All cases belonged to 30 households that included 139 individuals in all. All members of each household ate together at least once (at the same restaurant or during the same picnic) in the days before the onset of symptoms of relatives. All 36 healthy subjects were interviewed. Nineteen (52.8%) of them were female, the mean age was 42.5 years. Gender and age were not significantly different between healthy and ill individuals (table 1 ). An attack rate of 74.1% (103/139) was observed. In the univariate analysis, the association between raw mussel consumption and illness was significant (RR: 1.50; 95% CI: 1.18 – 1.89; p < 0.001). The attack rate according to raw mussel consumption was 86.3% (69/80). Even cooked mussel consumption was associated with illness in the univariate analysis (RR: 1.53; 95% CI: 1.05–2.23; p = 0.003). No other food was associated with illness (table 2 ). Stratified analysis showed that cooked mussel consumption was significantly associated with illness only among those who did not eat raw seafood (RR: 3.04; 95% CI: 1.26 – 7.30; p < 0.001). Such association was not significant in the "ate raw mussels" stratum (RR: 0.84; 95% CI: 0.76 – 0.93; p > 0.05). The relative risk according to cooked mussel consumption was 1.38 (95% CI: 1.00 – 1.91) after correction for raw mussels consumption (table 3 ). Subjects defined "probable secondary cases" were considered ill according to the risk assessment. In fact, we could not be certain that "probable secondary cases" were actually co-primary cases; on the other hand, the association between raw mussels and illness was stronger excluding from analysis such cases (RR: 1.63; 95% CI: 1.22–2.18; p < 0.001). Laboratory issues Both stool and mussel samples were negative for parasites and bacterial enteropathogens. All stool samples from 24 cases were positive for NV by RT-PCR and 5 of 11 mussel samples by nested PCR. The sequences of strains revealed great heterogeneity. In fact, the simultaneous occurrence of GI and GII viruses within the same outbreak was observed. One sample of mussels showed the presence of a mixed genogroup. Sequence analysis showed that strains from 19 cases and 3 mussel samples had identical sequences and belonged to GII, clustering narrowly to the strain NLV/Tarrag/238/2001/Sp, thus strengthening the epidemiologic link of the mussels to the cases. The nucleotide sequences from three cases and 2 mussel samples formed two other distinct clusters showing the best fit with the strains Saitama U25 and Khs1-1997-JP. Sequences from two further cases and one mussel sample were assigned to genogroup I and showed a high degree of identity with the strain NLV/Steinbach/EG/2001/CA. Three familial groups consisting of 7 cases and three members of a further family group from the outbreak showed identical sequences and were classified as NV GII (the main cluster of 19 cases). A fourth member of the latter family group was positive for one of the two GI strains. The other strains investigated by sequence analysis were unrelated to those from these familial groups. Discussion In Italy a national NV outbreaks database does not exist, although NVs are the leading cause of gastroenteritis outbreaks in Europe as in the rest of the western countries [ 7 ]. Moreover, to our knowledge, this is the first NV community outbreak to be confirmed in Italy [ 8 ]. The actual number of NV gastroenteritis cases is underestimated since the illness is often mild and diagnostics are provided only in a few laboratories. Despite these factors, the knowledge of circulating strains and the recognition of the common vehicles of infection are of primary importance for prevention of such disease. This investigation showed the causative role of mussel consumption in the outbreak, confirmed by laboratory tests on both stools and food samples. The heterogeneity of viral strains associated with the consumption of contaminated mussels is no surprise due to the peculiar features of shellfish. In fact, bivalve shellfish are filter feeders and tend to accumulate whatever pollutants are in the water which can result in viral contamination from a multitude of possible sources affecting many individuals. In any case, further studies should be carried out to clarify this issue [ 9 ]. The low relative risks we found could be due to the study design. In fact, the selection of households where a case occurred could underestimate the association. However this study design became necessary because of the unfeasibility of an open cohort study. In the 1990s, NVs were identified as the primary pathogen associated with shellfish-borne gastroenteritis in the United States [ 10 ]. Since then many studies have confirmed the role of shellfish in the spread of NV infection. Raw shellfish consumption is very frequent among the people of Puglia. Mussels sold in Puglia come from a large number of suppliers: from countries within the European Union (EU) and from other Mediterranean countries. The frequent shellfish consumption had already been blamed for the cases of cholera that occurred in Puglia in 1994, during the Christmas period, when a major epidemic was reported in Albania, on the opposite coast of the Adriatic sea [ 11 ]; moreover, mussels have been identified as the principal vehicle during recurring hepatitis A epidemics [ 5 , 12 , 13 ]. A significant presence of HAV in mussels sold in Puglia has recently been demonstrated by experimental evidence [ 3 , 5 ], even in samples negative for standard microbiological controls. In Puglia informative campaigns and routine microbiological controls on seafood are currently performed. The effectiveness of these interventions seems to be scarce. Moreover, our data showed cooked mussels also played a significant role. In fact cooking might not completely inactivate NVs and because of the low infectious dose, even a limited residual contamination can result in illness [ 14 ]. The epidemic curve shows a usual pattern consistent with a common point source. However, person to person transmission may have played a role in the last phase of the outbreak. On the other hand, such a means of transmission of NV is well documented. This investigation confirms the importance of field study in gastroenteritis outbreaks, as well as obtaining stool specimens from patients and food samples for laboratory analysis in a short time [ 15 ]. The present standards for human faecal contamination do not seem to be a reliable indicator of viral contaminants in mussels [ 5 ]. Seafood samples analysed during this investigation were all negative for the presence of common bacteria. On the other hand, to protect consumers it would be necessary to use a molecular index of the human contamination. In such cases reference laboratories with high-technology facilities would be required. The lack of such laboratories could be an obstacle to implementing a routine control system based on molecular tests. Conclusions This episode confirms that large NV outbreaks occur in Italy but only an accurate investigation can recognise this pathogen and that current regulations and commercial practices need to be revised to assure the safety of shellfish consumption and to improve control of future outbreaks. Competing interests None declared. Authors' contributions MC and GB carried out microbiological assays and drafted the manuscript. CG participated in the design of the study and performed the statistical analysis. MQ conceived the study as well as participating in its design and coordination. All authors have read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522816.xml
553989
A cost-minimization analysis of diuretic-based antihypertensive therapy reducing cardiovascular events in older adults with isolated systolic hypertension
Background Hypertension is among the most common chronic condition in middle-aged and older adults. Approximately 50 million Americans are currently diagnosed with this condition, and more than $18.7 billion is spent on hypertension management, including $3.8 billion for medications. There are numerous pharmacological agents that can be chosen to treat hypertension by physicians in clinical practices. The purpose of this study was to assess the cost of alternative antihypertensive treatments in older adults with isolated systolic hypertension (ISH). Method Using the Systolic Hypertension in the Elderly Program (SHEP) and other data, a cost-minimization analysis was performed. The cost was presented as the cost of number-needed-to treat (NNT) of patients for 5 years to prevent one adverse event associated with cardiovascular disease (CVD). Result It was found that the cost of 5 year NNT to prevent one adverse CVD event ranged widely from $6,843 to $37,408 in older patients with ISH. The incremental cost of the 5 year NNT was lower to treat older patients in the very high CVD risk group relative to patients in the lower CVD risk group, ranging from $456 to $15,511. Compared to the cost of the 5 year NNT of other commonly prescribed antihypertensive drugs, the cost of SHEP-based therapy is the lowest. The incremental costs of the 5 year NNT would be higher if other agents were used, ranging from $6,372 to $38,667 to prevent one CVD event relative to SHEP-based drug therapy. Conclusion Antihypertensive therapy that is diuretic-based and that includes either low-dose reserpine or atenolol is an effective and relatively inexpensive strategy to prevent cardiovascular events in older adults with isolated systolic hypertension. Use of the diuretic-based therapy is the most cost-effective in patients at high risk for developing cardiovascular disease.
Background Hypertension is among the most common chronic conditions in middle-aged and older adults. Approximately 50 million Americans are currently diagnosed with this condition, and more than $18.7 billion is spent on hypertension management, including $3.8 billion for medications[ 1 ]. Treatment of hypertension can significantly decrease the risk of developing CVD [ 2 , 3 ]. The SHEP and other studies have demonstrated the great potential of antihypertensive treatments to significantly reduce the number of cardiovascular events in elderly patients [ 4 - 10 ]. This, in turn, may reduce the costs associated with this chronic condition. Based on the SHEP study, it is estimated that 24,000 strokes, 44,000 major cardiovascular events, and 84,000 admissions to the hospital could be prevented over a 5-year period [ 7 ]. Currently, primary care physicians can choose from numerous pharmacological agents to treat hypertension. The commonly used antihypertensive drug classes include diuretics, beta-blockers, angiotensin-converting enzyme (ACE) inhibitors, alpha-blockers, and calcium channel blockers. Selection of an evidence-based therapy with demonstrated efficacy, safety, and low cost has important economic implications. The purpose of this study was to: 1) assess cost of the SHEP-based antihypertensive treatment to prevent adverse events associated with CVD, including death, stroke, myocardial infarction, and heart failure; and 2) to compare cost of the SHEP-based treatment to the costs of other commonly used antihypertensive agent treatments. Method The SHEP trial is a randomized, double-blind, placebo-controlled clinical trial sponsored by the National Heart, Lung, and Blood Institute and the National Institute on Aging that tested the efficacy of diuretic-based stepped-care antihypertensive drug treatment of isolated systolic hypertension (ISH) to prevent strokes [ 4 ]. Study Population The study subjects consisted of community-dwelling men and women 60 years and older who had isolated systolic hypertension, defined as an average systolic blood pressure (SBP) ≥ 160 mm Hg and an average diastolic blood pressure (DBP) < 90 mm Hg over 2 baseline visits. The primary endpoint of the trial was combined nonfatal and fatal stroke over a 5-year period. Secondary endpoints included nonfatal myocardial infarction (MI) plus fatal coronary heart disease (CHD) and major cardiovascular disease (CVD) morbidity and mortality. A total of 2,365 and 2,371 persons were randomized into the treatment and placebo group of the study respectively. Subjects who met the preliminary blood pressure (BP) eligibility criteria at the initial contact visit were referred to SHEP clinics for the baseline visits. At the baseline visits, subject's demographics, medical conditions, health behaviors, and cardiovascular risk factors were obtained. Methods of these measurements have been reported 4 . Fasting blood samples were analyzed at a central laboratory, including serum glucose, lipid levels, creatinine, uric acid, sodium, and potassium. Of the 4,736 SHEP participants, 4,189 were included in this analysis. The 547 participants were excluded either because of missing data concerning CVD risk factors (n = 283) or with previous CHD or stroke (n = 264). These 547 excluded subjects had similar age, sex, race, and other characteristics as those who were included in this analysis. Intervention A stepped-care treatment approach was used, with the goal for individuals with SBP >180 mm Hg to reduce to < 160 mm Hg and for those with SBP between 160 and 179 mm Hg to have a reduction of at least 20 mm Hg. All participants were given chlorthalidone, 12.5 mg/d, or matching placebo (step 1 and dose 1 medication). Drug dosage (step 1 and dose 2 medication) was doubled, 25 mg/d, for participants failing to achieve the SBP goal at the follow-up visits. If the SBP goal was not reached at the maximal dose of step 1 medication, atenolol, 25 mg/d, or matching placebo was added (step 2 and dose 1 medication). When atenolol was contraindicated, reserpine, 0.05 mg/d, or matching placebo could be substituted. When required to reach the blood pressure goal, the dosage of the step 2 drug could be doubled (atenolol 50 mg/d or reserpine 0.10 mg/d, step 2 and dose 2 medication). Potassium supplements were given to all participants who had serum concentration below 3.5 mm0l/L at two consecutive visits. The SHEP participants were followed up monthly until SBP reached the goal or until the maximum level of stepped-care treatment was reached [ 4 , 7 ] Ascertainment of Outcome Events The present analysis focused on five types of events: 1) death; 2) first-occurring major cardiovascular event, including stroke, MI, or heart failure; 3) first-occurring stroke; 4) first-occurring MI; and 5) first clinical diagnosis of congestive heart failure (CHF). The adjudication and clarification of the events was done by a panel of three physicians blinded to treatment assignment and blood pressure status. Members of the panel reviewed the documentation of new cardiovascular events over the study period and adjudicated outcome events according to predetermined criteria. [ 4 ] Calculation of Global CVD Risk Scores Information at the baseline on age, sex, total cholesterol, high density lipid (HDL) cholesterol, systolic blood pressure, diabetes (diabetic vs. non-diabetic), and smoking (current vs. never or past smoking) were used to calculate an a priori global score for the risk of developing future cardiovascular events, according to the Multiple Risk Factor Assessment Equation jointly proposed by the American Heart Association and the American College of Cardiology.[ 11 ] The equation assigns scores to major risk factors, using cut points that were originally developed using data on incident CHD from the Framingham study. A global CVD risk score ranging from -17 to +22 was obtained by adding the subscores. Higher values reflect a more unfavorable risk profile. Because the equation does not provide the age score for persons ≥ 75 years of age (28.5% of the SHEP study population), one additional point was assigned to men and women in this age group. Based on the global cardiovascular risk score, participants were classified into one of four CVD risk groups: low, medium, high and very high. Calculation of Costs The methods of economic evaluation include cost-effectiveness analysis, cost-utility analysis, and cost-benefit analysis, which can be used to assess the trade-off between costs and benefits in choices of antihypertensive treatment regimens. The primary aim of this analysis was to examine cost of the diuretic-based antihypertensive drug intervention in the SHEP trial. A cost-minimization analysis is a special type of cost-effectiveness analysis. It can be used to compare cost difference among competing alternatives of antihypertensive drug treatments when these treatments are medically equivalent. In this study, we used cost-minimization analyses to compare costs and incremental costs of NNT for 5 years to prevent one adverse event related to CVD among antihypertensive treatment regimens. The perspective of this economic evaluation was that of a national health insurance system. We used the number-needed-to-treat as an unit of common outcome measure in the analysis. The number-needed-to-treat to prevent one adverse outcome has become a widely used measure of treatment benefits derived from the results of clinical trials. The NNT is the reciprocal of the absolute risk reduction (ARR) which is the difference between the proportions with the adverse event in the treatment and placebo groups. The 95% confidence interval of NNT was calculated based on the regression-based method described by Laupacis et al. [ 12 ] The cost specified in the analysis includes the drug acquisition cost of SHEP treatment from the perspective of a national health insurance system. According to the SHEP treatment protocol, the stepped-care was classified into four types of drug treatments: 1) the Step 1 and Dose 1: chlorthalidone 12.5 mg/d; 2) the Step and Dose 2: chlorthalidone 25 mg/d; 3) the Step 2 and Dose 1: chlorthalidone 25 mg/d plus atenolol 25 mg/d or reserpine 0.05 mg/d; and 4) the Step 2 and Dose 2: chlorthalidone 25 mg/d plus atenolol 50 mg/d or reserpine 0.1 mg/d. Direct drug acquisition costs were calculated based on the minimum average wholesale prices (AWP) within drug manufacturers in the year 2000.[ 13 ] All drug costs were based on the AWP per unit dose. The expected cost (EC) of the SHEP drug acquisition per patient in 1 year was calculated as follows: EC = W 1 × C 1 + W 2 × C 2 + W 3 × C 3 + W 4 × C 4 The W1, W2, W3, and W4 represent proportions of the participants using the Step 1 and Dose 1, the Step 2 and Dose 2, the Step 2 and Dose 1, and the Step 2 and Dose 2 medication, respectively. C1, C2, C3, and C4 represent the drug acquisition cost of the Step 1 and Dose 1, the Step 2 and Dose 2, the Step 2 and Dose 1, and the Step 2 and Dose 2 medication, respectively. A Monte Carlo method was performed to estimate the average cost and its standard deviation. To compare the cost of the SHEP-based therapy to other antihypertensive drugs, it was assumed that all antihypertensive drugs in the comparisons have equal efficacy in terms of the NNT for 5 years to prevent one CVD related event. The NNT was calculated based on the method. [ 12 ] All drug costs were expressed as dose-specific cost per patient in 1-year and/or 5-year. Using the approach, costs were calculated for each representative drug based on equipotent doses in terms of blood pressure reduction. [ 14 ] The non-SHEP based drugs, including beta-blockers (Atenolol), ACE inhibitors (Enalapril), and calcium channel blockers (Nifedipine), were selected in the analysis according to antihypertensive drug class. These drugs were considered commonly prescribed antihypertensive medications in clinical practices. [ 14 ] All costs were adjusted in 2000 constant U.S. dollars using the Consumer Price Index. In this analysis, we focused on the drug acquisition cost for antihypertensive management. Therefore, the monitoring cost for antihypertensive treatment was not included. Total treatment cost includes antihypertensive drug cost and monitoring cost. The monitoring of treatment in ambulatory care settings including physician visits and laboratory tests have an estimated cost of $284 per patient per year. [ 14 ] Total cost of the NNT for 5 years of each drug therapy was calculated by multiplying the NNT for 5 years with the drug acquisition cost for 5 years per patient. The incremental cost is the cost of NNT for 5 years to prevent one adverse event of one alternative less the cost of the base case. In calculations of the incremental costs of the NNT for 5 years by types of outcome, the cost to prevent one stroke which was used as a base case. In calculations of the incremental costs of the NNT for 5 years by risk levels of CVD, the cost to prevent one adverse event of the very high risk level being used as a base case. Result Table 1 shows the expected acquisition cost of the diuretic-based antihypertensive therapies. The step 1 and dose 1 medication was the most used therapy and followed by the step 1 and dose 2 medication. The annual drug acquisition costs of the step 1 and dose 1, the step 1 and dose 2, the step 2 and dose 1 and the step 2 and dose 2 were $10.24, $20.48, and $222.45 respectively. The expected annual drug acquisition cost per patient of the SHEP treatment without potassium supplements was $83 and with potassium supplements was $91. The 5 year annual drug acquisition cost with potassium supplements per patient was $456. Table 1 Estimated Drug Acquisition Costs of The SHEP Treatment Protocol Drug Category Drug Cost Per Patient in 1 Year Proportion Drug Cost Per Patient in 5 years step1 dose1 (chlorthalidone 12.5 mg/d) $10.24 0.43 step1 dose2 (chlorthalidone 25 mg/d) $20.48 0.23 step2 dose1 (chlorthalidone 25 mg/d plus atenolol 25 mg/d or reserpine 0.05 mg/d) $222.45 0.16 step2 dose2 (chlorthalidone 25 mg/d plus atenolol 50 mg/d or reserpine 0.1 mg/d) $221.93 0.17 Weighted SHEP Rx $83.29 0.91 KCL $88.33 0.09 Weighted SHEP Rx including KCL $91.24 $456 SD $101.78 Results of the 5 year NNT to prevent one adverse event and its associated cost by event type are shown in Table 2 . To prevent one death, the cost for the 5 year NNT was $28,284. In other words, we need to treat 62 patients for 5 years in order to prevent one of them from death and the expected drug acquisition cost for the benefit is $28,284. The cost for the 5 year NNT to prevent one patient from one CVD event of any type is about four times lower than that of death. The cost for the 5 year NNT to prevent one MI is much higher than the cost for preventing one stroke or one CHF. Using the cost to prevent one stroke as the base amount, the incremental cost for the NNT for 5 years to prevent one MI or one CHF was $22,354 and $5,474, respectively. Table 2 NNT and Drug Costs by Adverse Events Event Placebo risk Treatment risk ARR NNT (95% CI) 5-Year NNT 5-year Rx Cost Per Patient Total Cost Incremental Cost Death 0.1002 0.0858 0.0144 69 (31 – 319) 62 $456 $28,284 $13,230 CVD 0.1746 0.1147 0.0599 17 (12 – 26) 15 $456 $6,843 - Stroke 0.0705 0.0433 0.0272 37 (24 – 76) 33 $456 $15,055 $0 (base) MI 0.0312 0.0202 0.011 91 (48 – 740) 82 $456 $37,408 $22,354 CHF 0.0397 0.0198 0.0199 50 (33 – 103) 45 $456 $20,529 $5,474 Table 3 presents costs of the NNT for 5 years to prevent one CVD event of any type by CVD risk strata. The cost for the 5 year NNT increases as the CVD risk level decreases. It costs $20,529 for the 5 year NNT to prevent one of any type of CVD adverse events among patients in the low CVD risk group. In contrast, it only costs $5,018 for the same effect among patients in the very high CVD risk group. Using the cost of the very high CVD level as a base, if 12 patients in the high CVD level are treated, the extra cost to prevent one patient out of 12 from one CVD event is $456. The extra cost for patients in the low CVD risk group to receive the same effect is $15,511 relative to the patients in the very high CVD risk group. Table 3 NNT and Drug Costs by CVD Risk Profile Risk Category Placebo Risk Treatment Risk ARR NNT (95% CI) 5-year NNT 5-year Drug Cost Per Patient Total Cost Incremental Cost 1 (low) 0.1013 0.0814 0.0199 50 (18 – 59) 45 $456 $20,529 $15,511 2 (medium) 0.1476 0.0912 0.0564 18 (11 – 53) 16 $456 $7,299 $2,281 3 (high) 0.2044 0.1265 0.0779 13 (8 – 26) 12 $456 $5,474 $456 4 (very high) 0.2526 0.1699 0.0827 12 (7 – 38) 11 $456 $5,018 $0 (base) In Table 4 , the comparisons of the incremental drug acquisition cost for the 5 year NNT of the SHEP-based antihypertensive therapy to other commonly prescribed antihypertensive drugs. This analysis assumes that alternative drugs have equal efficacy to prevent CVD events. The estimated incremental net cost of the 5 year NNT to prevent one CVD event associated with use of atenolol (beta-blocker), enalapril (ACE inhibitor), terazosin (alpha-blocker), and nifedipine (calcium channel blocker) relative to the SHEP-based drug therapy ranged from $6,372 to $38,667 in older adults with isolated systolic hypertension. According to the cost ratio, it indicates that the costs of the 5 year NNT of using enalapril, terazosin, and nifedipine were up to 6.6 times more expensive compared to the SHEP-based drug therapy. Table 4 Comparisons of Drug Acquisition Costs of 5-Year NNT Among Antihypertensive Drug Classes Drug Class Commonly Prescribed 5-year Cost Per Patient 5-Year NNT Total Cost Incremental Cost Cost Ratio SHEP-based drug therapy $456 15 $6,843 $0 (base) 1 (base) Beta-Blocker Atenolol 25 mg daily $1,255 15 $18,825 $11,982 2.75 50 mg daily $1,245 15 $18,675 $11,832 2.73 100 mg daily $1,792 15 $26,880 $20,037 3.93 ACE inhibitor Enalapril 5 mg daily $2,031 15 $30,465 $23,622 4.45 10 mg daily $2,132 15 $31,980 $25,137 4.67 20 mg daily $3,034 15 $45,510 $38,667 6.65 Alpha-Blocker Terazosin 2 mg daily $2,984 15 $44,760 $37,917 6.54 5 mg daily $2,984 15 $44,760 $37,917 6.54 10 mg daily $2,984 15 $44,760 $37,917 6.54 Calcium channel blocker Nifedipine 30 mg daily $881 15 $13,215 $6,372 1.93 60 mg daily $1,762 15 $26,430 $19,587 3.86 90 mg daily $2,644 15 $39,660 $32,817 5.8 Discussion The result of an economic evaluation essentially shows the cost per benefit gained from adapting a specific treatment. The effective and efficient use of resources has been increasingly emphasized from society, health plans, and health care providers. This cost-minimization analysis incorporating outcome data from the SHEP trial presents information treatment cost for older patients with ISH. We found that a long-term, low-dose and diuretic-based antihypertensive therapy is relatively inexpensive and effectively prevents adverse events associated with cardiovascular diseases, especially in older patients who had a high CVD risk profile. Our findings indicate that the total and incremental treatment costs of antihypertensive drugs in ambulatory care settings range widely among drug classes as well as within drug classes. This analysis suggests that diuretic-based antihypertensive treatments are the least expensive, whereas atenolol (beta-blocker) is less costly than enalapril (ACE inhibitor) and nifedipine (calcium channel blocker), and terazosin (alpha-blocker) is the most expensive drugs in terms of the 5 year NNT to prevent one CVD event. It appears that use of the SHEP-based drug therapy offers greater economic benefits for controlling isolated systolic hypertension in the elderly than other antihypertensive drug treatments. Using a decision analysis model that simulated clinical decisions and outcomes that would occur when primary care physicians follow the JNC IV hypertension management guidelines, it was found that a newer class of calcium channel blockers can slightly increase the proportion of patients who achieve and maintain hypertension control, but at a substantially higher cost than with a generic diuretic drug. [ 15 ] For our analyses, we presumed that all drugs offer equivalent therapeutic benefits. This assumption may have introduced a conservative bias into our primary findings. In fact, randomized controlled trials directly comparing active treatments for hypertension reported that calcium antagonists and doxazosin were inferior to low-dose diuretics or other agents in preventing cardiovascular events, suggesting that the cost-effectiveness of diuretic-based treatments may be even more favorable than estimated in the present study. [ 15 - 17 ] Further, in a meta-analysis of over 27,000 patients, those randomized to calcium antagonists as first-line therapy ran a greater risk of experiencing a myocardial infarction (26% higher risk), congestive heart failure (25% higher risk), and all cardiovascular events combined (10% higher) as compared to those randomized primarily to low-dose diuretics, beta-blockers and ACE inhibitors.[ 16 ] Finally, the Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trial (ALLHAT) recently reported a significantly higher risk of congestive heart failure, stroke, and major cardiovascular events in the doxazosin group than in the chlorthalidone group.[ 17 ] It is noteworthy that in this trial, only minimal differences in blood pressure control occurred between treatment groups, suggesting that the magnitude of blood pressure control represents an inadequate marker for comparing the therapeutic benefits of antihypertensive therapies. With regard to costs projected in our study, it is noteworthy to consider that compared to the SHEP treatments, costs of treatments based on more recently developed antihypertensive agents (than reported here) are likely to be even higher than estimated in the present analyses. The results of this study are limited to men and women 60 years and older who have isolated systolic hypertension and no presumed contraindication to any one class of antihypertensive medications. One limitation to our study relates to the fact that comparisons were based on costs of monotherapies, while combination therapies are frequently needed to control blood pressure. The number-needed-to-treat to prevent one adverse outcome has become a widely used measure of treatment benefits in medical community, which is easy for physicians to understand. The shortcomings of NNT are that the outcome measure of an effect is with one dimension- survival probability and that it measures the specified outcome at a single point in time. Therefore, a measure of NNT can not capture an outcome in effectiveness of the intervention with two dimensions: time and survival probability. These limitations may not allow us to take time and discounting on cost and effect into account in this study. Conclusion Based on our findings, antihypertensive therapy that is diuretic-based and that includes either low-dose reserpine or atenolol represents a cost-effective regimen in preventing or delaying cardiovascular events in older adults. Use of the diuretic-based therapy is the most cost-effective in patients at high risk for developing cardiovascular disease. These results suggest that clinicians should consider using diuretics plus low-dose reserpine or atenolol as first-line therapy in patients with isolated systolic hypertension who are greater than 60 years old when there are no contraindications among these patients. List of Abbreviations Used ACE: angiotensin-converting enzyme ALLHAT: Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trial ARR: absolute risk reduction AWP: average wholesale price BP: blood pressure CHD: coronary heart disease CHF: congestive heart failure CVD: cardiovascular disease DBP: diastolic blood pressure HDL: high density lipid ISH: isolated systolic hypertension JNC IV The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. NNT: number-needed to treat SBP: systolic blood pressure SHEP: Systolic Hypertension in the Elderly Program Conflict of Interest The author(s) declare that they have no competing interests. Authors' contributions GC, LF, WM and MP participated the development of the analytic framework. GC performed all data analyses. GC, LF, WM and MP drafted and revised the manuscript. All authors approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553989.xml
553976
The case for well-conducted experiments to validate statistical protocols for 2D gels: different pre-processing = different lists of significant proteins
Background The proteomics literature has seen a proliferation of publications that seek to apply the rapidly improving technology of 2D gels to study various biological systems. However, there is a dearth of systematic studies that have investigated appropriate statistical approaches to analyse the data from these experiments. Results Comparison of the effects of statistical pre-processing on the results of two sample t-tests suggests that the results of 2D gel experiments and by extension the conclusions derived from these experiments are not independent of the statistical protocol used. Conclusions This study suggests that there is a need for well-conducted validation studies to establish optimal statistical techniques to be used on such data sets.
Background The effort to produce an index of all human proteins (the human protein index, or HPI) began over twenty years ago. This project pre-dates the human genome project by more than a decade. However, the complexity of the task of creating this index was underestimated and the relative simplicity of the human genome with four known nucleic acids arranged in a linear coding order allowed the process of the sequencing of the human genome to progress exponentially [ 1 ]. The successful completion of the human genome project is now putting the focus back on proteins. The emergence of new and improved protein technologies from re-engineered two-dimensional (2D) gel systems to mass spectrometry has made the mapping and identification of the entire proteome of a cell (tissues) a much more accessible goal. Over the past few years a number of databases documenting the protein content of a single organism, organ or organelle have been created [ 2 - 6 ], and a number of papers describing results of experiments using these new and improved techniques have been published. The advantages of 2D gel technology Two-dimensional electrophoresis is an extremely powerful tool for the analysis of complex protein mixtures. Proteins carry both positive and negative charges. The pH of the medium they are in determines their net charge. The pH that gives a zero net charge is the isoelectric point of the protein (pI). In Isoelectric Focusing (IEF), protein mixtures are electrophoresed in a gel containing a pH gradient. The proteins in the mixture migrate according to charge density until they reach the part of the gel that corresponds to their pI. At this point, their net charge is zero, and migration stops. This is the first dimension of separation in a 2D gel experiment. The electrophoresed gel is then layered on top of a polyacrylamide gel and electrophoresed once again. The proteins now move from top to bottom depending on molecular weight. The distance covered by a protein is inversely proportional to its size. This is the second dimension in a 2D gel. 2D is an effective method for identifying qualitative and quantitative differences between proteins expressed in various tissues or between tissues exposed to different experimental treatments. Although the number of proteins displayed by 2D is much lower than the estimated number of genes in a particular tissue, 2D is currently the only available technique that enables the isolation and separation of thousands of the individual proteins that constitute a tissue proteome [ 7 ]. Anderson et al [ 1 ] point out that although writing obituaries for 2D gels has become a popular past time, the supply of unrelated parameters applicable to protein separation is limited and nearly all other combinations have been explored in the past. Database for statistical analysis Images of 2D gels are acquired into a database using an image scanner. Image analysis software converts the gel image into a digitised image in a computer, matches gels and spots on gels across the different groups and creates a database with information about spot intensity and spot location. As mentioned above, the two variables – the pI representing net charge of the protein and the molecular weight of the protein – are not correlated. In geometric terms this suggests that the two dimensions are orthogonal to each other. The two dimensions in a two-dimensional gel thus can be thought of as the two axes in a two dimensional graph. The coordinate on the x-axis is a measure of the isoelectric point (pI) of the protein, and the coordinate on the y-axis is a measure of the molecular weight of the protein The information in the database includes a gel identification variable, a spot identification variable, the x and y coordinates of a protein spot and its intensity measured by the amount of light transmitted by the spot. Depending on the software package, one can obtain other parameters in the database, including a measure of the quality of the spot to various measures associated with spot intensity, such as volume, area, peak height, etc. The rationale for this study The intensity of a protein spot is assumed to be directly related to the amount of protein in the particular tissue under investigation at that given time point. Changes in protein intensity are therefore approximated by changes in the intensities of protein spots in gel images. Changes in protein structure associated with post-translational modifications such as phosphorylation, oxidative modification or glycosylations may result in changes in the pI or molecular weight of the protein and are manifested in the gel by a change in the vertical or horizontal position. The object of 2D gel experiments is to detect differences in protein intensity/complexity between two groups of gels. A number of recent publications [ 8 - 10 ] have used statistical models generally known as classifiers to detect differences in protein intensity/complexity between two groups of gels. Classifiers are increasingly being used in the analysis of hi-dimensional data sets derived from gene and protein expression experiments. These models help one to determine if the changes in protein intensity /complexity are specific enough to enable a clear separation of the gels into the right groups. They can also be used to provide a good visual demonstration of the differences between groups. Classifiers Classification is the process of assigning objects to a category. An interest in classification permeates many scientific studies [ 11 ]. There are two broad categories of classification problems. In the first, e.g. discriminant analysis, one has data from known groups. Information that distinguishes these groups (i.e. differences in protein intensity/complexity) collected from an experiment is used to assign samples (gels) to these known groups. In the second case, e.g., cluster analysis, one has the information but no preset classification. The data is mined to see if there are naturally occurring clusters. These clusters are then investigated to identify commonalities within and differences between clusters. Stein and Zvelebil (12) and Patel et al (13) describe using 2D gel data sets to build supervised and unsupervised classifiers Both types of classification problems have three stages, input, algorithm and output. Most published literature concentrates on the second of these. However, careful thought about what variables to use and how to characterize or summarize them as inputs into an algorithm are very important issues [ 11 ]. It is evident that the reliability and reproducibility of a classification is a function of the input, which in turn depends upon the process of data normalization, data reduction, and variable selection, i.e. the pre-processing of data. This paper focuses on the effects of preprocessing on the selection of variables that enters a classifier. Pre-processing In order to conduct a systematic analysis of 2D gel data, one has to pre-process the data set. Pre-processing in the case of 2D gel analysis includes: 1) normalizing intensities to remove effects of differential loading and staining; 2) transformation of outcome variables to normally distributed variables; and 3) imputing values for missing spot intensities. We are not aware of any other study that has looked systematically at the effect of pre-processing 2D gel data on the results of subsequent statistical analysis of the data. In this paper, we present a protocol for the analysis of 2D gel data and examine the effect of statistical pre-processing of 2D gel datasets. The effect of using two different formulas for normalizing versus not normalizing, log-transformation versus no log-transformation and single value imputation versus multiple value imputation, and averaging spot intensities across replicates versus keeping the replicate information separate are compared using the results of two sample t-tests. We also compare the results of two-sample t-tests provided by the image analysis software PDQUEST to results obtained after following the protocol described in Figure 1 . PDQUEST allows the user to normalize the data but has no facility for testing the distribution of the outcome variable and transforming it to fit a normal distribution. To the best of our knowledge, PDQUEST replaces missing spot intensities with a zero. The experiment The data used in this study is from an experiment that looked at the effect of a diet enhanced with grape seed extract on the proteome of whole brain homogenates of Sprague-Dawley rats [ 10 ]. There were five treated animals and five control animals. Due to sample availability and other issues related to the creation of 2D gels, each biological replicate had different numbers of technical replicates. The maximum number of replicates was four, and the minimum was two. A number of changes in proteins that were attributed to treatment differences in this study have been identified with Matrix Assisted Laser Desorption Ionization – Time Of Flight (MALDI-TOF) Mass Spectrometry. These changes have also been confirmed in later experiments with transgenic mice. Thus the protein changes detected by the statistical protocol used in this study have been shown to be biologically valid and relevant to the systems being studied. Results Variability in the resolution of protein spots in 2D gels The resolution of protein spots in a 2D gel is highly variable. It can differ considerably between technical replicates of the same biological sample. Samples 6, 7, 8, 9, and 10 were the five biological replicates in the treatment group. Samples 22, 23, 24, 25, and 26 were the five biological replicates in the control group. Table 1 demonstrates the breakdown of the resolved protein spots in the different samples and its replicates. Biological sample 7, for instance, had 546 protein spots resolved in at least one of its four technical replicates. 169 (31%) proteins occurred in all the four replicates. An additional 130 (26%) were present in at least three replicates out of four. A further 97(18%) were present only in two replicates out of four, and 150 (27%) were present only in one of the four replicates. Low correlation between technical replicates Table 2 displays the range of Pearson's correlation coefficients and Kappa coefficients between the technical replicates of the same sample. The correlation coefficient here is a measure of the association between the spot intensities on technical replicates. The correlation between technical replicates is not very high. The correlation coefficients range from a high of 0.93 to a low of 0.47. The r-square ranges from 87% to 22%. The Kappa coefficient measures the degree of agreement between the spots present on two replicate gels of the same sample. A zero indicates no agreement and one indicates perfect agreement. If the confidence interval spans zero then the hypothesis that there is no agreement between the replicate gels cannot be rejected. Ten of the sixteen separate confidence intervals in Columns 5 and 6 of Table 2 include zero. This suggests no agreement between the replicate gels of most samples. Assessing the quality of a pre-processing technique In Tables 3 to 10 the spot identification numbers in bold represent proteins that were subsequently identified and found to be biologically relevant to the system being studied (10). In all, eleven proteins that had significantly different intensities at alpha = 0.05 were identified. The measure of the quality of a particular pre-processing technique in this study was the proportion of these eleven proteins identified as statistically significant in a two-sample t-test after the particular technique was used. The effect of log-transformation and minimum values substitution on the distribution of intensities Log – Transformation Figures 3a and 3b are the QQ plots for the raw spot intensity and normalized spot intensities for the 201 protein spots on a representative gel from the control group. These plots demonstrate that the normalization technique used does not alter the basic distribution of the raw data. They also demonstrate the highly non-gaussian distribution of the spot intensities. Figure 3c demonstrates that the log transformation converts the distribution of the intensities from a very non-normal distribution to a normal distribution. The points lie very close to the straight line that represents a normal distribution. We have found that the log transformation reduces the skew in the distribution of the spot intensities if the image analysis is done in PDQUEST ® and most data sets produced by the software PROGENESIS ® . Only one out of seven 2D gel data sets analysed by us so far did not respond well to this transformation. A closer examination of this data showed it had a large number of saturated spots, and thus needed to be rerun. The transformation of the distribution of the 201 spots also worked reasonably well at the level of individual spot intensities. This was important to confirm since the two sample tests were done at individual spot level. In cases where there was a considerable skew in the distribution, e.g., SSP 1733, the log transformation made the distribution of spot intensities normal. The Anderson-Darling test for normality for SSP 1733 in the control group, has a p-value = 0.03 before the transformation, and p-value = 0.456 after the transformation. In general, this was true of most spots we examined. Table 3 compares the t-test results of non-log-transformed normalized spot intensities where the missing spot intensities were replaced with a zero; to the results using normalized spot intensities that were log-transformed. In the non-log transformed data without normalization, six of the eleven spots known to be significantly different were picked up. Six of the eleven proteins were picked after log transformation of the non-normalized data as well. It is important to note that the two lists of six proteins were not identical. After normalization, in the log-transformed data the t-test picks up all the eleven proteins, whereas, in the non-log-transformed data seven proteins are picked up (Table 4 ). Effect of normalization Table 5 compares the proteins that were found to be significant (p = 0.05) in a t-test when the spot intensities were not normalized, not log transformed, and the missing intensities were replaced with a zero (Column 1); proteins that were found to be significant in a t-test (p = 0.05) when the spot intensities were normalized using normalization 1 (see methods), not log transformed, and the missing intensities were replaced with a zero (Column 2); proteins that were found to be significant in a t-test (p = 0.05) when the spot intensities were normalized using normalization 2 (see Methods), not log transformed, and the missing intensities were replaced with a zero (Column 3); and the proteins that were found to have significantly different intensities by the image analysis software PDQUEST (Column 3), which normalizes (each intensity divided by the total intensity of the gel) the data but does not use a log transformation. As is evident from comparisons of Column 1 with Columns 2 and 3 in Table 1 , normalization has an effect on the number of proteins that are detected. The difference in Columns 2, 3 and 4 suggests that the protocol used for normalizations also has an impact on the proteins that are picked up as significant. In non-log transformed data, data sets with no normalization, normalization 1 and normalization 2 picked up six of the eleven proteins, and PDQUEST method picked up seven of the eleven proteins. Over fifty percent of the proteins picked up by PDQUEST occurred in very small numbers of gels, and hence did not meet our selection criteria for inclusion into the analysis data set. A number of the other proteins picked up by PDQUEST had skewed distribution. After log-transformations these proteins were no longer statistically significantly different. Tables 6 , 7 and 8 again demonstrate that normalization 2 has a significant impact on the number of proteins identified as significantly different in intensity. In all of these tables, it is important to note once again that the highlighted proteins in each column vary with the pre-processing technique. Effect of imputation of missing spots In Table 9 , Columns 1, 2, and 3 offer a comparison of the spots identified as significantly different (alpha = 0.05) when the three different kinds of imputation of missing spots were used. The three different kinds of imputation did not make much difference to the spots identified as significantly different in intensities at alpha = 0.05. In fact in this case we have identical lists in all three columns of table 9 . The purpose of using multiple imputation instead of single value imputation, however, has more to do with getting a better estimate of the variance of a quantity than with the correct estimation of the mean. Since the t-statistic is a function of both the difference in means as well as the variance of a variable, given a constant mean, underestimating variance would lead to false positives, and overestimating variance would lead to false negatives. This data set shows a good example of the degree of variability on the intensities. Table 10 demonstrates that single value imputations tend to either underestimate (var = 0 for SSP 6452), or over-inflate the estimates of variance (e.g. SSP 1509). The variance of SSP 1733, which has no missing spot intensities, gives us a rough idea of the degree of variability expected in spot intensities when there are no intensities missing. The estimates of variance for the proteins with missing intensities are much closer to the values seen in SSP 1733 when one uses a multiple imputation technique. Averaging across replicates versus keeping replicates separate Given the lack of association between technical replicates, we used replicate information in two ways: 1) Spot intensities were averaged across the replicates so that the t-tests compared average spot intensities in the five treatment samples versus five control samples, and 2) The replicate gels of each sample were treated as independent gels, and the t-tests compared spot intensities in the fifteen treatment gels to the intensities in the fifteen control gels. Columns 1 and 2 in Table 11 compare Method 1 above to Method 2. Six out of eleven proteins are picked up using Method 1, whereas all ten are identified as significantly different if the replicates are kept separate. Discussion Normalization Differential sample loading and stain absorption and other process variables can contribute to variability in measured protein intensity. In order to ensure that the detected differences in protein intensity are not due to a "technical" variability introduced by the process of gel creation, spot intensities are "normalized." Dividing the intensity of each protein on a gel by the total protein intensity of that gel is a widely used technique to reduce the "individual gel effect" on protein intensities [ 14 ]. Normalizing the data is an important step in many datasets, but it becomes especially important in proteomics experiments, which in general have many more variables than samples. In this case a systemic error in processing samples or gels that affects only one or two gels can have a huge impact on the results. This study has demonstrated that the results of statistical tests are not independent of the normalization technique. Testing for normality and transforming data To the best of our knowledge, none of the image analysis software packages available to date provide the tools necessary test for the distribution of the data. All of them provide t-tests or the non-parametric Wilcoxon rank sum test or ANOVA to test for differences in individual spot intensities. The probability values (p-values) for differences between groups are based on the assumptions of the normality of the distribution of spot intensities and equal variances. In order to make an informed judgement about the validity of the p-values of the tests above, it is important to know if these assumptions are met [ 15 , 16 ]. If one uses the averaged spot intensities across gels, the argument could be made that the central limit theorem obviates the need for a log transformation. Given the highly skewed nature of the raw spot intensities, the dependence of the mean and the variance of intensities, and the fact that even the non-parametric Wilcoxon test assumes symmetry in the outcome variable [ 17 ], log transformation of the data is still advisable. Missing spot intensities Despite the fact that 2D-gel technology offers many advantages, one of the pitfalls associated with this technology is the need for several replicates for proper validation of results. There are instances where one does not observe reproducible spot patterns or individual proteins even in replicate gels of the same sample. Missing spot intensities are commonly observed in 2D gel datasets. Multivariate techniques such as Principal Component Analysis and Discriminant Analysis (DA) are ideal tools to use on databases that have multiple outcome variables (protein intensities). However, SAS or any other statistical software that is used to analyse the data using multivariate techniques such as PCA and DA requires data sets with non-missing values. Gels with missing spot information will thus be dropped from the analysis. Since all gels will have some spot information missing, this will result in no gels being available for data analysis. To the best of our knowledge, all image analysis software packages substitute zeroes for missing intensity values. Missing intensities may be caused by the fact that a protein spot truly does not exist in one group compared to the other or because the spot intensity is so low that it is not detected by the image analysis software. In this study we treated all missing spots as undetectable spots. The question we were trying to answer was "If a spot exists but is 'undetectable', what is the best 'detectable' value to substitute as its intensity?" The most intuitive value was the smallest "detectable" intensity in the experiment hence we used the lowest intensity value in the experiment for the single value imputation. However, substituting a single value for all missing spot intensities would skew the distribution of spot intensities considerably. Thus the second option was to use a random process to substitute missing spot intensities with a plausible set of "detectable" intensities. We created a set of lowest 'detectable' intensities from the lowest intensity on each of the thirty gels. These values were used to impute missing intensity values as described in the methods section. Although we have used the terms "imputation" and "multiple imputation", these terms are not to be equated with multiple imputation advocated by Rubin [ 18 ]. Rubin's techniques assume the missing intensities to be Missing at Random (MAR). By our assumption, the spots are undetectable because of the intensity level, or the probability of missing is a function of the intensity. This by Rubin's definition would make the missing spot intensities Non-Ignorable Non-Response. His multiple imputation techniques thus would not be valid in this context. The issue of missing protein intensities is one that has not been addressed at all in the literature describing 2D gel studies. In this study we treated all missing intensities as the same. However, all missing intensities in 2D gels are not equal. Some missing intensities are missing because they truly do not exist in one group versus the other, whereas others are missing because of the inherent variability in the process of creating 2D gels. This suggests that one needs to approach the filling of missing values differently based on the probability of a spot being a truly missing protein or one that is missing due to the process of gel creation. One way to do this would be to assume that spots that occur in two out of three replicates (or three out of four replicates) of a sample are true spots. These missing intensities would then be replaced by the mean value of the remaining two (three) spots. On the other hand, the assumption that those proteins that are missing in all control gels or all treatment gels are proteins that are turned on or off is justifiable. This in turn suggests that for this set of proteins the random imputation of missing values with a set of plausible minimum intensity values, acting as placeholders so that the non-missing data can be used in analyses, can also be justified. Which method should one use? In the last few years, published reports of 2D gel analysis have concluded that heart failure was associated with protein modifications in three cellular systems [ 19 ], identified proteins expressed in six different regions of brains of Alzheimer's disease patients [ 7 ], had been used to establish genetic relationships in the Brasscacae family [ 20 ], to classify human ovarian tumours as malignant and benign [ 8 ], in the detection of polypeptides associated with the histo-pathological differentiation of primary lung cancer [ 21 ], and to identify eight protein feature changes that differentiated breast cancer cell lines that did or did not form tumours in nude mice [ 9 ]. These are all important studies that will be used as springboards to launch ever more expensive and sophisticated experiments. We have demonstrated that protein changes that are large (e.g. SSP 6452 present in controls and absent in all treatment gels) are independent of the statistical protocol used. The identification of more subtle changes can vary widely depending on the statistical algorithm used to pre-process and analyze the data. Our experience with the GSE data and a couple of subsequent experiments we have been involved in suggests that the algorithm we have developed is more sensitive with respect to identifying biologically relevant proteins that image analysis software might miss. However, it is a fact that pre-processing could also give rise to false positive results. It is important to establish the best statistical protocol for analysing the data from these studies. One way to get around the issue of the effect of pre-processing is to restrict a study to only those proteins that are picked up as significant by image analysis software. As we have mentioned above, in this study a large number of the spots selected by PDQUEST were poor quality spots either in terms of protein quality or consistency. Another option is to consider only those spots that appear in two sets of analyses (e.g. image analyses, and the protocol described here) as true changes. One is thus restricting oneself to gross changes. As the use of proteomic techniques moves forward, however, we think it will be important to identify more subtle changes in proteins. A number of studies have suggested that change in protein expression that starts the cascade of changes that leads to a diseased tissue need not always be gross or dramatic. Subtle changes in expression early in a pathway can cause significant changes downstream. Shapiro et al [ 22 ] suggest that subtle changes in the spatial or temporal expression of the patterning molecule Sonic Hedgehog (SHH) is linked to the proliferation and patterning of developing limbs. Similarly, a disease condition could be caused by small changes in expression in a number of proteins. Reneiri et al [ 23 ] suggest that the phenotypic expression of Retts in some but not all girls with the MECP2 mutation suggests that MECP2 causes deregulation of a very small subset of genes that have not yet been detected or that very subtle changes in many genes (by extension proteins) may cause the neuronal phenotype. The importance of picking up subtle changes in protein expression suggested by studies cited above point to a need to establish a way to identify the optimal statistical pre-processing techniques for 2D gel datasets. An intuitively appealing way to do this is to create 2D gels with serially diluted quantities of commercially available proteins, establishing the relationship of protein quantity to spot intensity and then proceeding to compare different statistical pre-processing techniques to the datasets acquired from these gels. Since it is a known fact that commercially bought proteins may not necessarily be pure and may be present in multiple modified forms due to the process of isolation, the experiment described above may be enhanced by using controlled biological samples from a cell culture with at least a hundred resolved spots. Once again, the sample could be loaded in on gels in known concentrations and the process described above would be repeated. Clearly, the set of pre-processing techniques that picks up differences that come closest to the true differences would be chosen as the optimal techniques. Some recent publications have used similar techniques to establish the validity of emerging proteomic technology. Alban et al [ 24 ] used an Escherichia coli lysate "spiked" with varying amounts of four different known proteins to test a novel experimental design that exploits the sample multiplexing capabilities of DIGE by including a standard sample in each gel. Rabilloud et al [ 25 ] compared the staining sensitivity of RuBPS and Sypro Ruby of serial dilutions of molecular weight markers. However, there are no designed experimental studies that have looked at the impact of statistical pre-processing or the effectiveness of various statistical techniques on the conclusions drawn from 2D gel experiments. Given the proliferation and promise of 2D gel experiments, we suggest that the need to conduct these validation experiments is urgent. In this study we have used the proteins that were subsequently identified by MALDI-TOF spectroscopy as a measure of how well particular statistical protocols perform. We concede that there is an inherent bias in that the spots that were identified by MALDI-TOF were selected on the basis of the protocol described in this paper. The described protocol therefore will seem to perform much better than others in this comparison. This however does not diminish the main thrust of this paper, which is that statistical protocols affect the conclusions drawn from a 2D gel experiment. Conclusions This study has demonstrated that the pre-processing of the data from 2D gel experiments can have a significant impact on the results of statistical tests. The purpose of the study was not to identify the particular statistical protocol used in this study as the optimal protocol, but rather to demonstrate that the results and conclusions from a biological experiment are not independent of the statistical protocol. The study has in effect looked at three different statistical protocols. The one described in Figure 1 , the one described in Figure 2 , in which the averaging of intensities across replicates allows one to proceed directly on to the t-tests, without the steps of testing distributions, or imputation of missing spots, and the protocol used by PDQUEST. Allowing for the inherent bias we have described above, this study shows that the protocol described in Figure 1 , with normalization technique 2, and multiple imputations, is superior to the method used by PDQUEST, or the one in Figure 2 . Given the possible bias in this study, the larger conclusion from the study is that there is a great need for research into developing optimal statistical methodology to analyze data from 2D gel experiments. Methods We used data from an experiment that compared the protein expression in the whole brain homogenate of rats that were fed a diet with 5% of grape seed extract to that of a normal rat diet. This experiment will be described in a separate article. The data from this experiment was exported to a database, which was saved as a text file. The database was imported into SAS V-9.0 (Statistical Analysis Software, 2003 Cary NC, USA), which was then used to do all the statistical analysis. Normalization 1 The intensity of each protein spot was divided by the total protein content (total intensity) in the experiment. Normalization 2 The intensity of each spot in each group was then normalized to the median intensity of its group, i.e. the intensity of each spot was converted into the intensity it would have had if the gel it was in had a total intensity equal to the median total intensity of the group. This is described by the following formula: This normalization reduces intra group variability, but maintains the inter group variability. Fold change is a common metric used in articles describing gene array and proteomic experiments. Fold change measures the degree of change in protein intensity in the treatment group, compared to the control group. This is measured by dividing the average spot intensity in the treatment group by the average spot intensity in the control group. In order for this ratio to be a true comparison of the intensities in the two groups, the intensities need to be expressed as proportions of the same quantity, i.e. they need to be divided by the same quantity. Normalization 1 has this property built into the formula. In normalization technique 2, the normalized intensities were divided by the total protein content in the experiment in order to make a fold change comparison meaningful. Subset of spots included in analysis A spot was considered present only if it was present in both replicates of a sample if there were only two replicates, in two out of three replicates if there were three replicates of a sample, and three out of four replicates if there were four replicates of a sample. We also included spots that were present in all controls and absent from all treatments, or present in all treatment gels but absent from all controls. Using these criteria, we had 201 spots that were available for statistical analysis from the GSE database. Missing spot intensities Given the criteria used to subset spots for the final analysis, there was a large number of missing spot intensities. We examined the effect of filling in the missing spots in three different ways. 1. Missing spot intensities were first replaced with the lowest value of log-transformed intensities. In this case the value was -17.28. The effect of the replacement of this single value for all missing spot intensities is the same as the effect of replacing non-log transformed spot intensities with zero. 2. We created 1000 separate data sets that randomly selected one of the 15 lowest spot intensity values from the 15 gels in the control group to replace each of the missing spot intensities in the controls, and repeated the process for the 15 treatment gels. 3. In the third imputation method, each missing spot intensity was replaced with a randomly selected value from the 30 lowest values from each gel in the experiment without regard to whether the value chosen was from a control or treatment gel. The two methods of random imputation of spot intensities were replicated 1000 times. two sample t-tests were repeated with each replication. Authors' contributions HK is the PI of the project that provided the data for this experiment. JD, HK's Research Associate, created the 2D gels, conducted the image analysis and created the database. SM designed this study and performed the statistical analyses.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553976.xml
506778
The "Goldilocks Effect" in Cystic Fibrosis: identification of a lung phenotype in the cftr knockout and heterozygous mouse
Background Cystic Fibrosis is a pleiotropic disease in humans with primary morbidity and mortality associated with a lung disease phenotype. However, knockout in the mouse of cftr , the gene whose mutant alleles are responsible for cystic fibrosis, has previously failed to produce a readily, quantifiable lung phenotype. Results Using measurements of pulmonary mechanics, a definitive lung phenotype was demonstrated in the cftr -/- mouse. Lungs showed decreased compliance and increased airway resistance in young animals as compared to cftr +/+ littermates. These changes were noted in animals less than 60 days old, prior to any long term inflammatory effects that might occur, and are consistent with structural differences in the cftr -/- lungs. Surprisingly, the cftr +/- animals exhibited a lung phenotype distinct from either the homozygous normal or knockout genotypes. The heterozygous mice showed increased lung compliance and decreased airway resistance when compared to either homozygous phenotype, suggesting a heterozygous advantage that might explain the high frequency of this mutation in certain populations. Conclusions In the mouse the gene dosage of cftr results in distinct differences in pulmonary mechanics of the adult. Distinct phenotypes were demonstrated in each genotype, cftr -/-, cftr +/-, and cftr +/+. These results are consistent with a developmental role for CFTR in the lung.
Background Cystic fibrosis (CF) is a progressive disease primarily affecting the intestines, lungs, and pancreas. The gene responsible for CF was identified in 1989 [ 1 ] as coding for the cystic fibrosis transmembrane conductance regulator ( cftr ), a membrane chloride channel. CF is one of the most common autosomal recessive diseases in Caucasians with a carrier rate of 3–4% [ 2 ], and is characterized by recurrent infection and chronic inflammation. Recently it was found that infants with CF demonstrate changes in forced expiratory volume in 1 second (FEV 1 ), functional residual capacity (FRC), and other parameters of lung function prior to the onset of recurrent infection [ 3 - 5 ]. Soon after the CF gene was discovered, a knockout mouse was developed. This mouse demonstrates subtle changes in epithelial cell phenotype, including alterations in secretory glycoconjugates and changes in secretory vesicles [ 6 ]. Monocytic infiltrates and altered lung mechanics have also been found [ 7 ]. Unfortunately, the cftr knockout mouse does not develop overt lung disease, which has severely limited its usefulness. However, the availability of new methods for pulmonary testing in rodents [ 8 , 9 ] now presents the opportunity to re-examine the cftr knockout mouse for functional lung changes. In the present study, therefore, we examined pulmonary function in young adult cftr -/-, cftr +/-, and cftr +/+ S489x mice in an effort to establish a lung phenotype. Results Effect of cftr gene dosage on pressure-volume (PV) curves Routine evaluation of dynamic lung function employs the stepwise variation in air volume on both the inflation and deflation phases of a single breath. Measurement of airway pressures at each step results in the classic pressure-volume ( PV ) curve which is dependent upon both lung structure and interfering pathology. PV curves were measured in triplicate, starting from positive end-expiratory pressure (PEEP) values of 0, 3, and 6 cmH 2 O in S489X mice at 30–60 days of age following genotyping for the normal and mutant cftr alleles. The 3 cmH 2 O PEEP curves obtained for each genotype are presented in Figure 1A . Note that these PV curves all begin at V = 0 ml, which is the FRC defined by the 3 cmH 2 O PEEP. The PV curves obtained at PEEP levels of 0 and 6 cmH 2 O were similar. Figure 1 Relationship between cftr genotypes and PV curves . Littermates from 30–60 days of age were genotyped and individuals with cftr +/+ (Black), cftr +/- (Green), and cftr -/-(Red) were evaluated using pressure volume curve analysis at Peeps of 0, 3, and 6. A: PV curve at PEEP 3; B: Calculated Cst for all PV curves; C: Calculated hysteresis for PV curves in A. All measures were corrected individually for lung weight. Error bars are +/- standard deviation. *p < 0.05 when compared to cftr +/+ and **p < 0.05 when compared to cftr+/-. The static compliance ( Cst ) of the lungs, which reflects elastic recoil at a given pressure, was calculated from the slopes of the PV curves. As shown in Figure 1B , the static compliance of the homozygous knockout lung was significantly decreased compared to the homozygous normal lung (p < 0.01)). Furthermore, Cst was significantly reduced in both cftr +/+ (p < 0.001) and cftr -/- (p < 0.001) as compared to age-matched cftr +/- mice. Hysteresis was altered among the 3 genotypes (Figure 1C ). A statistically significant increase in hysteresis was observed in both cftr +/- (P < 0.0001) and cftr +/+ mice (p < 0.05) compared to cftr -/- mice. These data suggest the presence of a gene dosage effect in which an altered lung structure in the heterozygous animals leads to an elevated compliance relative to the two homozygous animals. Lung weights were measured and no statistically significant differences were observed among the three genotypes (Figure 1D ). Airway mechanics of cftr deficient lungs We applied the forced oscillation method to the mice and determined respiratory mechanical input impedance [ 10 , 11 ]. We fit the constant-phase model of respiratory mechanics [ 12 ] to impedance ( Zrs ) and determined values for airway resistance ( Raw ), tissue damping ( G ), and tissue elastance ( H ). Figure 2 shows that Raw , G , and H were significantly reduced in the cftr +/- mice as compared to both cftr +/+ and cftr -/- animals (Panels A, C, & D) at all PEEP levels. Raw also decreased with PEEP in a similar fashion in all three genotypes. In contrast, Raw , H and G were significantly increased in the cftr -/- mice compared with cftr +/+ and cftr +/-, and showed a greater dependence on PEEP (Panels A, C & D). The ratio G / H , termed hysteresivity (both the low and high frequency), was not significantly affected by either genotype or PEEP (Panel E). Thus, the absence of either 1 or 2 copies of the cftr gene had significantly different effects on the phenotype of the lung. Paradoxically, the absence of only one cftr copy resulted in a greater lung compliance (lower elastance) than if neither or both copies were present. Figure 2 Variation in respiratory mechanic among cftr genotypes. Values for Raw (A), G (B), H (C), and η (D) were determined by fitting the constant-phase model to measurements of Zrs from cftr +/+ (Black), cftr +/- (Green) and cftr-/- (red) genotypes. All measures were normalized by multiplication by lung weight. Error bars are +/- standard deviation. *p < 0.05 when compared to cftr +/+ and **p < 0.05 when compared to cftr+/-. Discussion The usefulness of the cftr knockout mouse as a model of cystic fibrosis has been severely limited by its failure to demonstrate readily measurable lung disease, the primary cause of morbidity and mortality in humans [ 13 ]. However, in the present study use of sophisticated measurements of lung function revealed a functional lung phenotype in the knockout mouse (Table 1 ); the complete absence of cftr in the lung of young adult animals resulted in decreased Cst and η and increased Raw , G and H as compared to normal littermate controls. Table 1 Relative effect of genotype on lung function compared to that observed in the cftr +/+ mouse. Genotype PV curve position Cst PV Hysteresis Raw G H η cftr +/- ← ↑ ↑ ↓ ↓ ↓ ↔ cftr -/- → ↓ ↓ ↑ ↑ ↑ ↔ A particularly intriguing further observation was that Cst and hysteresis in cftr +/- mice was significantly higher than in cftr +/+ animals while G and H were decreased. As this was not associated with any pathology such as emphysema, we conclude that it represents a functionally different lung from that of the cftr +/+. Our data thus reveal a remarkable inverse correlation between the effect of one and two non-functional copies of the cftr gene. What do these data mean in terms of lung structure? The knockout animal has cystic fibrosis by definition, and our data now show it to also have lung disease manifest as a reduced compliance and increased resistance. These changes could reflect changes in the intrinsic mechanical properties of the parenchyma, or simply a reduction in lung volume. The former effect could include alterations in the biophysical properties of the air-liquid interface in the lungs, and would be expected to result in a change in η [ 14 ]. Indeed, because cftr is a chloride channel and is thought to be involved in water balance, a change in surface tension in the lung, and consequently in η , might be expected. However, as shown in Figure 2 and Table 1 , although G and H both increase, they do so in the same proportion so there is no significant change in η between the three cftr genotypes. On the other hand, lung weights were not different among the different groups of mice, so the decreased compliance and increased resistance of the cftr -/- animals was not simply due to their having smaller lungs than control animals. This suggests that the parenchymal structure in the lungs of the homozygous and heterozygous animals were organized differently, in a manner that affected G and H similarly. As documented in numerous publications, the mouse strain used in the present study does not develop chronic inflammatory disease up to the age (30–60 days) used in this study (for review see [ 15 ]). On the other hand, Broaches-Carter et al. [ 16 ] have shown that cftr levels are highest in the developing lung and decrease 75-fold at birth. In utero over-expression of cftr has also been shown to affect lung growth and development[ 17 ], and the severity of disease in the knockout mouse has been shown to be influenced by genetic background [ 18 ]. These data thus suggest that cftr may affect the early development of the lung in a manner that is affected by the interaction of other genes. Are there any functional consequences for increased lung compliance in the heterozygous cftr animals? Interestingly, there is no decrement in lung function in human heterozygotes [ 19 - 21 ]. Also, the heterozygote frequency for CF in humans is higher than expected, likely reflecting a selective advantage because there is no evidence for genetic drift [ 22 , 23 ]. Indeed, selective advantage in CF has been proposed to reflect resistance to tuberculosis, influenza and cholera [ 24 ]. When one looks in nature for other examples of heterozygous advantage, the sickle cell trait which confers resistance to malaria [ 25 ] is perhaps the only such recognized genetic trait in humans. In Norway rats, a single Mendelian gene controls resistance to Warfarin, an anticoagulant used to control rat populations; homozygous wild-type rats are killed by Warfarin and homozygous mutant allele animals are highly susceptible to vitamin K deficiency [ 26 ]. The results of the present study indicate that a similar selective advantage may pertain to cftr , something we term a "Goldilocks Effect". That is, while two defective copies of the gene are detrimental and two normal copies are satisfactory, one normal and one defective gene may results in an optimal dosage for lung development. Further studies of pulmonary mechanics in cftr knockout mice should reveal additional genetic loci that modulate the influence of cftr on lung growth and development. Corresponding studies in humans should be useful in evaluating the effect of therapies on reversing altered pulmonary function in the CF patient. Conclusions Using sophisticated techniques to a evaluate rodent pulmonary function; a distinct, readily quantifiable lung phenotype was identified in the cftr knockout mouse. In addition, the cftr +/- mouse had a distinguishable pulmonary function phenotype from that observed in either the homozygous normal or mutant genotype mice. These data are consistent with CFTR-dependent, physiologic changes in the structure and function of the lung. Methods Mouse strain The S489X mouse 5 th generation backcross to C57Bl/6 has been maintained by random mating for the past 8 years. This colony has a 100% mortality rate among cftr knockouts by 45 days of age unless the animals are placed on an elemental liquid diet and corncob bedding upon weaning [ 27 ]. Mice 30–60 days of age from our S489X mouse colony were genotyped for the normal and mutant cftr alleles. Age and litter matched cftr +/+ and cftr +/- were used for each cftr -/- mouse examined. Six animals were included in each group. All experiments were approved by the animal care and use committee. Pulmonary function tests The mice were anesthetized with intra-peritoneal pentobarbital (90 mg/kg) and the trachea was dissected free of surrounding tissue and cannulated with a 20-gauge cannula. The animals were then connected to a small animal ventilator ( flexiVent , SCIREQ Inc. Montreal, PQ, Canada) and ventilated with a tidal volume of 10 ml/kg; inspiratory:expiratory ratio of 66.67%, respiratory rate of 150 breaths/minute, and maximum pressure of 30 cmH 2 0. PEEP was controlled by submerging the expiratory limb from the ventilator into a water trap. Each animal was paralyzed with pancuronium bromide (0.5 mg/kg) and allowed to equilibrate on the ventilator until spontaneous breathing ceased (5 minutes). Respiratory mechanics To measure Zrs , mechanical ventilation was interrupted and the animal was allowed to expire against the set level of PEEP for 1 s. We then applied an 8 s broad-band volume perturbation signal to the lungs with the flexiVent , after which ventilated was resumed. This was repeated at PEEP levels of 0, 3 and 6 cmH 2 O. The volume perturbation signal consisted of the superposition of 18 sine waves having frequency spaced roughly evenly over the range 0.25 Hz to 19.625 Hz. Zrs was calculated from the displacement of the ventilator's piston and the pressure in its cylinder as described previously [ 10 , 11 ]. Correction for gas compressibility as well as resistive and accelerative losses in the flexiVent , connecting tubing and the tracheal cannula were performed as described previously [ 28 ] using dynamic calibration data obtained by applying volume perturbations through the tubing and tracheal cannula first when it was completely closed and then when it was open to the atmosphere. We interpreted the measurement of Zrs in terms of the constant phase model [ 12 ] where Raw is a frequency independent Newtonian resistance reflecting that of the conducting airways[ 29 ], Iaw is airway gas inertance, G characterizes tissue damping, H characterizes tissue stiffness (elastance), i is the imaginary unit, α links G and H , and f is frequency. We also calculated a quantity known as hysteresivity ( η = G/H) , which is believed to increase when regional heterogeneities develop in the lung [ 30 ]. Raw , G and H were all normalized by multiplication by lung weight. PV curves Starting at the FRC defined by the PEEP, the flexiVent was programmed to deliver seven inspiratory volume steps for a total volume of 0.8 ml followed by seven expiratory steps, pausing at each step for 1 s. Plateau pressure ( P ) at each step was recorded and related to the total volume ( V ) delivered to produce a quasi-static PV curve. Cst was calculated from the slope of each curve [ 31 ], and was normalized by division by lung weight. Zrs measurements at each PEEP level and PV curves were obtained in triplicate. Data were statistically evaluated using paired t-test with p < 0.05 being taken as significant
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC506778.xml
547903
Molecular models of NS3 protease variants of the Hepatitis C virus
Background Hepatitis C virus (HCV) currently infects approximately three percent of the world population. In view of the lack of vaccines against HCV, there is an urgent need for an efficient treatment of the disease by an effective antiviral drug. Rational drug design has not been the primary way for discovering major therapeutics. Nevertheless, there are reports of success in the development of inhibitor using a structure-based approach. One of the possible targets for drug development against HCV is the NS3 protease variants. Based on the three-dimensional structure of these variants we expect to identify new NS3 protease inhibitors. In order to speed up the modeling process all NS3 protease variant models were generated in a Beowulf cluster. The potential of the structural bioinformatics for development of new antiviral drugs is discussed. Results The atomic coordinates of crystallographic structure 1CU1 and 1DY9 were used as starting model for modeling of the NS3 protease variant structures. The NS3 protease variant structures are composed of six subdomains, which occur in sequence along the polypeptide chain. The protease domain exhibits the dual beta-barrel fold that is common among members of the chymotrypsin serine protease family. The helicase domain contains two structurally related beta-alpha-beta subdomains and a third subdomain of seven helices and three short beta strands. The latter domain is usually referred to as the helicase alpha-helical subdomain. The rmsd value of bond lengths and bond angles, the average G-factor and Verify 3D values are presented for NS3 protease variant structures. Conclusions This project increases the certainty that homology modeling is an useful tool in structural biology and that it can be very valuable in annotating genome sequence information and contributing to structural and functional genomics from virus. The structural models will be used to guide future efforts in the structure-based drug design of a new generation of NS3 protease variants inhibitors. All models in the database are publicly accessible via our interactive website, providing us with large amount of structural models for use in protein-ligand docking analysis.
Background After the development of serological tests for hepatitis A and B viruses in the 1970s it became clear that an additional agent accounted for approximately 90% of transfusion-associated hepatitis (non-A non-B hepatitis, NANBH) [ 1 ]. The novel agent, hence termed hepatitis C virus (HCV), currently infects approximately 3% of the world's population and it was classified within the Flavivirideae family. Diagnostic tests for anti-HCV antibodies developed thereafter proved that HCV was indeed the predominant cause of NANBH [ 2 ]. In view of the lack of vaccines against HCV, there is an urgent need for a treatment of the disease by an effective antiviral drug. This necessity has boosted research on the structural biology of HCV with the primary focus being to identify possible targets for pharmaceutical intervention [ 3 ]. Rational drug design has not been the primary way for discovering major therapeutics. However, recent successes in the area give reason to expect that drug discovery projects will increasingly be structure based. One of the possible targets for drug development against HCV is the NS3 protease variants. HCV RNA is translated into a polyprotein that during maturation is cleaved into functional components. One component, nonstructural protein 3 (NS3), is a 631-residue bifunctional enzyme with protease and helicase activities. The N-terminal portion of the NS3 protein was predicted to contain a serine protease domain as judged from conserved sequence patterns and by homology to Flavi- and Pestiviruses [ 4 - 6 ]. The NS3 serine protease processes the HCV polyprotein by both cis and trans mechanisms. The interative refinement and optimization of drug leads is an effective strategy for generating potent preclinical candidate [ 7 , 8 ]. Ongoing genome sequencing efforts have led to the identification of hundreds of potential therapeutic targets, many of which represent possible sources of crossover pharmacology. Homology or comparative modeling is a key feature of an integrated drug discovery effort because it allows this genomics information to be utilized early in the development of target ligands or in the engineering of ligand specificity [ 9 ]. Genome sequencing efforts are providing us with complete genetic blueprints for hundreds of organisms, including humans. We are now faced with assigning, understanding and modifying the functions of proteins encoded by these genomes. This task is generally facilitated by 3D structures [ 10 ], which are best determined by experimental methods such as X-ray crystallography and NMR spectroscopy. The theoretical approaches [ 11 ] can be divided into physical and empirical methods. The physical prediction methods are based on interactions between atoms and include molecular dynamics and energy minimization [ 12 ], whereas the empirical methods depend on the protein structures that have been already determined by experiment. They include combinatorial [ 13 ] and comparative modeling [ 14 , 15 ]. Comparative modeling uses experimentally determined protein structures to predict conformation of other proteins with similar amino acid sequences. For modeling of proteins was used restrained-based modeling implemented in the program MODELLER [ 16 ]. The models consist of coordinates for all non-hydrogen atoms in the modeled part of a protein. Models are generated entirely automatically in a four-step procedure [ 17 ]: (i) fold assignment, (ii) sequence-structure alignment, (iii) model building, and (iv) model evaluation. This procedure was applied to variants of NS3 protease using Perl-CGI, C and MPI programming. We modeled the structure of variants of NS3 protease variants available in the National Center for Biotchnology Information (Genbank), using structural bioinformatics tools. Knowledge of the three-dimensional structure variants will undoubtedly aid the design of useful inhibitors that may be used as a drug against hepatitis C virus. In order to speed up the modeling process all NS3 models were generated in a Beowulf cluster (BioComp, S.J. Rio Preto, Brazil). The potential of the structural bioinformatics for development of new antiviral drugs is discussed. Results and discussion Primary sequence comparasion The identity between the sequences of a bifunctional protease structure (PDB access codes:1CU1, 1DY9) [ 31 , 38 ] (templates) and NS3 protease variants (targets) is shown in Table 1 . The secondary structural elements are indicated in the Figure 2 without inhibitor and in the Figure 3 with inhibitor. The sequence from crystallographic structure 1CU1 shows more than 79.1% identity with the sequences of NS3 protease variants, which provide high accuracy for the models (Table 1 ). Quality of the models The atomic coordinates of crystallographic structure 1CU1 solved to resolution of the 2.5 Å were used as starting model for modeling of the NS3 protease variant structures, and the structure of NS3 complexed with an inhibitor (PDB access code: 1DY9) was used to generate homology models for docking studies. Binding of an inhibitor to the active site of an enzyme is typically connected with local and possibly also global structural rearrangement of the enzyme (induced-fit mechanism). Therefore structure-based drug design preferentially relies on the crystal structures of enzyme-inhibitor complexes containing bound inhibitors of similar chemical structures to the compounds being designed. Such complexes offer more detailed and accurate picture of the inhibitor-enzyme interactions and structural complementarity between the inhibitor and the active site. The homology models of the variants of NS3 protease which used the NS3 complexed with an inhibitor are more adequate to docking simulations. The atomic coordinates of all water molecules were removed from the templates. The analysis of the Ramachandran diagram φ-ψ plots of the 1CU1 structure (template) were used to compare the overall stereochemical quality of the NS3 protease variants structures against template solved by biocrystallography (Table 1 ). They present over 94.0% of the residues in the most favorable regions. The same analysis for crystallographic structure (1CU1) present 88.9% of residues in the most favorable, 10.5% additional allowed regions, 0.6% generously allowed regions, and 0.0% disallowed regions, which strongly indicates that the molecular models present good overall stereochemical quality. Overall description The NS3 protease variant structures are composed of six subdomains, which occur in sequence along the polypeptide chain (Figure 2 and 3 ). The protease domain exhibits the dual β-barrel fold that is common among members of the chymotrypsin serine protease family. The helicase domain contains two structurally related β-α-β subdomains and a third subdomain of seven helices and three short β strands. The latter domain is usually referred to as the helicase α-helical subdomain. The 13-residue protease activation domain of NS4A contributes one strand to the N-terminal protease β-barrel and is considered to be the sixth subdomain [ 31 ]. Differences in subdomain structure in the NS3 protease variant molecule and in the structures of the isolated protease and helicase domains were assessed in several ways. Inspection of the molecule revealed that the subdomain folds are similar. Overall preservation of structure is also apparent when the subdomains from the various structures are superposed [ 31 ]. The rmsd value of bond lengths and bond angles, the average G-factor and Verify 3D values are shown in Table 2 for NS3 protease variants structures. The same analysis for crystallographic structure (1CU1) present rmsd values of the 0.013Å bond lengths and 1.67°, the average G-fator values of the 0.14 torsion angles and 0.28 covalent geometry, and Verify 3D values of the 321.53 score total and 1.09S quality. Database design, access, and interface A MySQL database based on relational database management system (RDBMS) was developed to archive protein structure identified in infectious agents such as NS3 protease variants from hepatitis C virus. All supporting data related to the 3D structures modeling, such as protein codes, atomic coordinates in PDB format from modeled proteins, fasta sequence, links to others databases and various information about the protein were arranged in the MySQL [ 32 ] database under a master table. The aim this database is to provide access to a collection of annotated models generated by automated homology modelling of NS3 protease variants from hepatitis C virus. All models in the database are publicly accessible via our interactive website (Figure 1 ) [ 33 ]. The database user interface provides user friendly menus, so that all information can be printed in one step from any standard web browser. A small ribbon representation is included to obtain a first impression of the model structure (Figure 2 and 3 ). Atomic coordinates for the homology models can be downloaded in PDB format and their primary sequence in fasta format. The fields are defined with links to the target sequence, the template structure entries in PDB [ 34 ], structural information and analysis. There are two homology models for each sequence in the database, one obtained using 1CU1 as template and other using 1DY9 as template. The second model is adequade for docking simulation, since it was used as template a structure complexed with an inhibitor (PDB access code: 1DY9). Conclusions Large scale protein homology modeling, in which whole sequence databases or whole genomes are used as input into automated modeling algorithms, have been reported by several groups [ 35 ]. By utilizing powerful computer systems with multiple processors, these efforts have allowed the creation of large databases of homology models of proteins. This project increases the certainty that homology modeling is an useful tool in structural biology and that it can be very valuable in annotating genome sequence information and contributing to structural and functional genomics from virus, bacteria and other organisms. Inhibition studies have shown that NS3 is only modestly inactivated by classic serine protease inhibitors such as chloromethylketones or phenylmethyl sylfonylfluoride [ 36 ]. The structural models will be used to guide future efforts in the structure based design of a new generation of NS3 protease variants inhibitors. This database is freelly available for all users on the Web, providing us with large amount of structural models for use in protein-ligand docking analysis. Methods Molecular modeling Molecular modeling is usually the method of choice when there is a clear relationship of homology between the sequence of a target protein and at least one known structure. This computational technique is based on the assumption that tertiary structures of two proteins will be similar if their sequences are related, and it is the approach most likely to give accurate results [ 18 ]. There are two main approaches to homology modeling: (1) fragment-based comparative modeling [ 14 , 19 ] and (2) restrained-based modeling [ 16 ]. For modeling of NS3 protease variants from hepatitis C virus we used the second approach. Model building of NS3 protease variants was carried out using the program MODELLER [ 16 ]. MODELLER is an implementation of an automated approach to comparative modeling by satisfaction of spatial restraints [ 20 - 22 ]. The modeling procedure begins with an alingment of the sequence to be modeled (target) with related known three-dimensional structure (templates). This alignment is usually the input to the program. The output is a three-dimensional model for the target sequence containing all main-chain and sidechain non-hydrogen atoms. Next, the spatial restraints and CHARMM energy terms enforcing proper stereochemistry [ 23 ] were combined into an objective function. Finally, the model is obtained by optimizing the objective function in Cartesian space. The optimization is carried out by the use of the variable target function method [ 24 ] employing methods of conjugate gradients and molecular dynamics with simulated anneling. Several slightly different models can be calculated by varying the initial structure. A total of 1000 models were generated for each enzyme and the final models were selected based on stereochemical quality. All optimization process was performed on a Beowulf cluster with 16 nodes (BioComp, AMD Athlon XP 2100+). Analysis of the models The overall stereochemical quality of the final models for each NS3 protease variants from hepatitis C virus was acessed by the program PROCHECK [ 25 ]. The root mean square deviations (rmsd) differences from ideal geometries for bond lengths and bond angles were calculated with X-PLOR [ 26 , 27 ]. G-factor value is essentially just log-odds score based on the observed distributions of the stereochemical parameters. It was computed for the following properties: torsion angles (the analyses provided the observed distributions of φ-δ, χ 1 -χ 2 , χ -1 , χ -3 , χ -4 and ω values for each of the 20 amino acid types) and covalent geometry (for the main-chain, bond lengths and bond angles) these average values were calculated using PROCHECK [ 25 ]. The Verify-3D measures the compatibility of a protein model with its sequence, these values were calculated using 3D profile [ 28 - 30 ]. Authors' contributions NJFS carried out the molecular modeling and participated in the building and design of the database tables, HAA and CEB carried of analysis of the models and created database interface, FPS selected the primary sequences of NS3 protease variants of the hepatitis C virus, IMVGCM participated of the study, PR conceived the study and suggestions in this manuscript, JRRP participated in the study and suggestions in this manuscript, WFA conceived the study, and participated in its analysis and coordination. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC547903.xml
516027
Selection and validation of endogenous reference genes using a high throughput approach
Background Endogenous reference genes are commonly used to normalize expression levels of other genes with the assumption that the expression of the former is constant in different tissues and in different physiopathological conditions. Whether this assumption is correct it is, however, still matter of debate. In this study, we searched for stably expressed genes in 384 cDNA array hybridization experiments encompassing different tissues and cell lines. Results Several genes were identified whose expression was highly stable across all samples studied. The usefulness of 8 genes among them was tested by normalizing the relative gene expression against test genes whose expression pattern was known. The range of accuracy of individual endogenous reference genes was wide whereas consistent information could be obtained when information pooled from different endogenous reference genes was used. Conclusions This study suggests that even when the most stably expressed genes in array experiments are used as endogenous reference, significant variation in test gene expression estimates may occur and the best normalization is achieved when data from several endogenous reference genes are pooled together to minimize minimal but significant variation among samples. We are presently optimizing strategies for the preparation of endogenous reference gene mixtures that could yield information comparable to that of data pooled from individual endogenous reference gene normalizations.
Background Endogenous reference alse referred to as house keeping genes defines in biology the theoretical assumption that certain genes are ubiquitously expressed in nucleated cells possibly because their stable expression is essential for cell survival and welfare in all physio-pathological circumstances. In practical terms, endogenous reference genes provide a useful constant reference to normalize the expression of test genes in different tissues and in different conditions. This is obviously important when estimates of gene expression are provided in relative terms rather than absolute units of measurement. Thus, endogenous reference genes are used as common denominator in biological fractions where the expression of a test gene is described as the relative ratio over an arbitrarily selected internal control presumed to be stably expressed in all circumstances relevant to the experiment [ 1 - 3 ]. Most frequently, glyceraldehydes-3-phosphate dehydrogenase (GAPDH) [ 4 , 5 ], albumin (for hepatocytes) [ 6 ], β-, γ-actins [ 7 , 8 ], cyclophilin [ 9 , 10 ], α-, β-tubulins [ 7 , 11 ], hypoxantine phosphoribosyltransferase (HRPT) [ 12 , 13 ], L32 [ 14 , 15 ] and 18S, 28S ribosomal RNA (rRNA) [ 16 - 18 ] have been used as endogenous reference genes. Depending upon the experimental design, endogenous reference genes have been used individually or in combination for Northern blot analysis, reverse transcription polymerase chain reaction (RT-PCR) and quantitative real-time PCR (qRT-PCR) analysis [ 19 , 20 ]. With the development cDNA microarray technology endogenous reference genes have been used for array data normalization. However, accumulation of extensive data bases suggests that the expression of frequently used endogenous reference genes can vary substantially according to materials and conditions studied [ 1 , 2 , 6 , 14 , 17 , 18 , 20 - 27 ]. Powerful insights in patterns of gene expression could be attained recently through cDNA or oligonucleotide-based global transcript analysis tools that apply a constant reference system to determine ratios of gene expression across large data sets [ 28 , 29 ]. The constant reference is provided for each gene in question by consistently co-hybridizing individual test samples with a differentially labeled reference sample maintained identical throughout all the hybridization experiments. Gene expression data are then expressed as the ratio of expression between test and reference samples for each gene. By keeping the reference sample identical the resulting ratio represents a precise estimate of the relative expression of each gene across the various conditions tested bypassing the need to normalize with endogenous reference genes. This holds true if the hybridization kinetics between test and reference sample are accurately reproducible. We will refer to this concept as "reference concordance" and in the results we will discuss how reference concordance was used to validate the reproducibility of cDNA array data from which putative candidate endogenous reference genes were identified. In the present study, we tested a set of 419 consecutive experiments performed on a 17,000 gene cDNA array platform to which RNA from neoplastic or normal tissues were consistently co-hybridized with a differentially-labeled reference RNA derived from peripheral blood mononuclear cells (PBMC) pooled from six normal donors. The following steps were pursued: 1) Reproducibility assessment of the data set through determination of reference concordance . This was achieved by repeating 14 reference experiments using the melanoma cell line A375 as test sample (Cy5) co-hybridized with pooled PBMC as reference (Cy3). To test for inter-array and printing variation, slide number one and every other 25 slides in sequential order of printing (100 slides per printing set) were used for the repetitive A375 / pooled PBMC hybridizations. In addition, to assess labeling bias, reciprocal labeling was alternatively applied as previously described [ 30 ]. In this fashion a pool of genes expressed with high level of reference concordance was selected. 2) Identification of putative endogenous reference genes was performed on 384 array experiments of unequivocal quality by selecting genes that had demonstrated high reference concordance (>90% of the genes in the arrays) and ranking them from the lowest to the highest variance of Log2 test / reference ratios across all array experiments. 3) Validation of the candidate endogenous reference genes as predictor of relative gene expression in large data series . For this purpose, we tested the relative estimates of expression of the melanocytic lineage-restricted melanoma differentiation antigen gp100/Pmel17 (gp100) [ 31 ] in melanocytic and non-melanocytic tissues. Estimates of expression of gp100 were compared after normalization with different endogenous reference genes. For this analysis, we used a new tool that spots cDNA libraries from different tissues on an array platform to allow high-throughput evaluation of individual gene expression across broad tissue collections. We termed this tool: "transcriptome array". 4) Validation of endogenous reference gene-based normalization of gene expression by qRT-PCR and protein expression . Results Data processing The GenePix Pro 4.0 software was used for array image analysis and calibration. All statistical analyses were performed with the SPLUS package . Thirteen arrays with missing data in more than 30% of the spots and 8 arrays with irregular distribution according to M-A plots (M and A representing respectively log-ratios and average intensities) [ 32 ] were excluded. Spots in which > 50% of the pixels reached saturation in either channel, flagged spots and spots with intensity ≤ 200 in one channel and ≤ 500 in the other channel were filtered out. If the intensity in one channel was < 200 but that of the other channel was > 500, we arbitrarily assigned an intensity of 200 to the channel with the lowest intensity. The log2 Cy5/Cy3 ratios were normalized by approximating median values to zero. Spots with ratios > 100 or < 1/100 were truncated at 100 and 1/100 respectively. Data normalization was done by median centering Log2 ratio of all genes for each array. For data normalization we used certain arbitrary cutoff criteria to remove spots with weak signal. Weak signal approximating background fluorescence is not reliable and the corresponding log ratios are often distorted resulting in disproportionately high or low values that would bias the statistical results. Because the interpretation of spots with low signal is difficult to make we have adopted the policy of excluding them to focus the analysis on genes whose expression pattern is more reliably tracked by the array tool. In regard to spots with a ratio of >100 or <-1/100, most of them are due to extremely weak signal in one channel which generates a disproportionately extreme value. Although there is no published "gold standard" for the selection of such cutoffs, in practice the parameters that we used for this study are most commonly accepted by investigators and reasonable because allow retention of most of the data in the array excluding only the most extreme and least reliable information. Validation of the data set through analysis of "reference concordance" To select from the data base the most reliable genes, we first assessed reproducibility by testing the level of concordance in 14 reference experiments. Reference concordance relies on the expectation that results obtained by repeatedly hybridizing the same test and reference material should perfectly collimate and the degree of deviation from such prediction estimates experimental variance. Concordance can be easily measured by periodically re-hybridizing the same test sample with the constant reference sample. We analyzed a matrix of 7 forward and 7 reciprocally-labeled replicate array experiments hybridized periodically every other 25 cDNA array slides. Reciprocal labeling was applied to measure labeling bias [ 30 ]. Data generated by reciprocal experiments were mathematically reversed into the same labeling direction for data analysis. Genes that were discordant due to labeling bias were identified by student's t test as those with ratios significantly different between the 7 forward and 7 reciprocal experiments after reversal of the reciprocal values (P < 0.05) and the median ratio difference between the 7 forward and 7 reciprocal experiments was larger than 1.5 fold. The genes whose variances of the log ratio across all experiments were among the top 1 percentile of variance of all genes were identified as discordant due to hybridization bias. In addition, genes with more than 50% missing values were excluded. Overall 1,343 out of 16,738 were judged potentially discordant and were excluded from further analysis. The remaining 15,395 concordant genes were used for subsequent analyses. Endogenous reference gene identification We studied 384 of 419 consecutive array experiments remaining after the exclusion of the 14 reference concordance experiments and 21 experiments judged of poor quality due to missing data or irregular distribution by M-A plot analysis. Selection of candidate endogenous reference genes followed these steps. First, the seven experiments used for the analysis of reference concordance in which labeling was done identically to the rest of the experimental samples (test labeled with Cy5 and reference with Cy3) were used to calculate experimental variance (Table 1 ). Median of variances across the seven replicate arrays was calculated for all genes with average intensity in both channels > 2,000. This parameter provides an estimate of the variance due to experimental variation (background variance) since theoretically no differences in gene expression should be noted by using the same material. Based on the assessment of the seven repeat experiments, we defined differential expression as > 2 standard deviations (SD) from the mean of the 384 arrays which is equivalent to 1.46-fold change (± 0.549 in log2 ratio). Consequently, genes with fold changes < 1.46 from zero across all 384 arrays were categorized as candidate endogenous reference genes and were ranked according to ascending values of SD of Log ratio (Table 1 ). Table 1 384 experiments 7 replicates Gene Name Image ID Unique ID Mean Intensity Mean of logRatio SD of logRatio Mean of logRatio SD of logRatio cDNA FLJ40458 1571492 Hs.181346 2835 0.19 0.27 0.39 0.15 Unknown 301067 2571 0.1 0.27 0 0.21 PTH 322051 Hs.37045 3380 0 0.28 -0.02 0.27 KIAA1935 39938 Hs.300776 2359 0.16 0.28 0.01 0.19 ESTs 49313 Hs.395460 2079 0.15 0.28 0.08 0.03 cDNA FLJ30539 fis 49463 Hs.21489 2561 0.09 0.28 0.17 0.13 SDCCAG16 1576228 Hs.271926 2111 0.11 0.29 0.28 0.14 clone IMAGE:41799 278673 Hs.271721 3132 0.16 0.3 -0.04 0.1 PSIP2 289945 Hs.82110 2521 -0.07 0.3 -0.06 0.19 ETV2 1468722 Hs.194061 2646 0.09 0.31 0 0.28 ESTs 1571401 Hs.126999 2659 0.05 0.31 -0.04 0.08 NEDD8 277660 Hs.75512 3875 0.08 0.31 0.1 0.2 HIP2 486259 Hs.155485 2511 0.11 0.31 0.12 0.25 HIRIP5 745314 Hs.430439 3407 -0.08 0.31 0.07 0.31 ACTR8 156363 Hs.124219 2731 0.08 0.32 0.14 0.16 GPLD1 293696 Hs.272529 3110 -0.03 0.32 -0.05 0.28 KIAA0769 299128 Hs.19056 2148 0.14 0.32 0.05 0.2 CSNK1G2 346031 Hs.181390 2320 -0.03 0.32 -0.24 0.28 RAB31 784150 Hs.223025 3479 -0.06 0.32 -0.36 0.36 C14orf117 796100 Hs.103189 2008 0.12 0.32 -0.01 0.16 ATP6IP2 825077 Hs.183434 2159 0.02 0.32 0.28 0.32 SNRPD3 897099 Hs.1575 3031 -0.04 0.32 -0.48 0.25 Unknown 1292073 2745 -0.13 0.33 0.05 0.28 ESTs 22137 Hs.187406 3133 0.07 0.33 -0.01 0.21 ESTs 32782 Hs.443140 2464 -0.02 0.33 0.09 0.26 L3MBTL 43090 Hs.300863 2581 -0.08 0.33 -0.41 0.31 GNRH1 487071 Hs.82963 3294 0.05 0.33 0.23 0.23 clone 24629 mRNA 746258 Hs.142570 2645 0.05 0.33 -0.21 0.14 FLJ12998 826367 Hs.343627 2267 0.14 0.33 0.03 0.3 cDNA: FLJ23477 fis 1492329 Hs.145362 2142 0.09 0.34 0.35 0.34 HOXC4 1756945 Hs.50895 2581 0.03 0.34 -0.02 0.22 WSX1 1855887 Hs.132781 3259 -0.01 0.34 0.12 0.3 HIST1H1A 1872543 Hs.150206 2523 0.07 0.34 -0.17 0.14 TFAP2B 363144 Hs.33102 2510 0.05 0.34 0.67 0.35 TRIM31 509760 Hs.91096 2753 0.06 0.34 -0.08 0.2 FLJ00166 713191 Hs.43213 3134 0.08 0.34 0 0.22 CDK5R1 757873 Hs.93597 2965 -0.12 0.34 -0.22 0.15 PPP3CA 796730 Hs.272458 2605 -0.07 0.34 -0.36 0.16 PMS2L4 161373 Hs.278468 2534 -0.14 0.35 -0.21 0.22 DLGAP1 1758491 Hs.75814 2143 0.02 0.35 -0.24 0.2 ESTs 177884 Hs.14613 2054 0.01 0.35 -0.22 0.25 NRL 2364249 Hs.89606 2300 -0.04 0.35 -0.09 0.14 RAB36 281489 Hs.38772 3143 -0.08 0.35 0 0.12 RPC8 323603 Hs.353192 2367 -0.06 0.35 -0.33 0.24 ZNF177 33294 Hs.172979 2988 0.01 0.35 -0.21 0.25 LY6G5C 448136 Hs.246845 3591 -0.02 0.35 -0.44 0.28 RAD17 586844 Hs.16184 3797 -0.01 0.35 0.12 0.15 LOC253039 1033388 Hs.41181 3612 -0.02 0.36 0.31 0.08 IMPDH2 1582050 Hs.75432 2709 -0.01 0.36 -0.08 0.18 PRMT3 2074202 Hs.152337 2042 0.07 0.36 -0.06 0.21 FZD6 214916 Hs.114218 3884 0.03 0.36 0.02 0.26 AP3B2 47510 Hs.21022 2350 0.03 0.36 -0.23 0.31 NF2 769716 Hs.902 2887 -0.08 0.36 -0.11 0.31 SCP2 855395 Hs.75760 3218 -0.12 0.36 -0.1 0.19 COL8A1 1472775 Hs.114599 4178 -0.06 0.37 0.05 0.18 AKAP4 1643144 Hs.97633 2557 -0.03 0.37 -0.07 0.15 Unknown 1671546 3386 0.03 0.37 1 0.42 TCF8 178463 Hs.232068 2726 -0.04 0.37 -0.01 0.06 TNP2 1839038 Hs.2748 2737 0.08 0.37 0.11 0.24 PIWIL1 2329739 Hs.194712 2573 -0.08 0.37 -0.14 0.22 DBCCR1 47037 Hs.6090 2481 -0.01 0.37 -0.18 0.24 SLC21A3 289706 Hs.46440 2708 0 0.38 -0.11 0.27 ESTs 35105 Hs.403854 3126 0.03 0.38 0.23 0.18 cDNA FLJ34400 fis 511835 Hs.380035 3058 0 0.38 0.02 0.37 CNTN1 51640 Hs.143434 2479 -0.05 0.38 -0.11 0.3 TFCP2 843067 Hs.154970 3728 -0.06 0.38 0.16 0.34 C17orf35 510032 Hs.15196 2778 -0.02 0.39 0.09 0.28 AFAP 488062 Hs.80306 3142 0 0.42 0.12 0.15 MTP 731054 Hs.195799 2764 -0.01 0.42 -0.09 0.4 Among these genes, we further selected candidate endogenous reference genes according to the following criteria: in at least 95% of the 384 arrays the Log2 ratio had to be within 2 SD from zero and the average intensities of both channels across all samples need to be higher than 2,000. The intensity parameter was added to ensure that the selected endogenous reference genes were expressed at a relatively high level in most tissues to mitigate excessive fluctuations in Log2 ratios occurring when a low value is applied as denominator. This is important when a reference gene is applied as a denominator in the equation used to normalize the ratio of other tests genes; in such cases robust gene expression decreases the range of ratios resulting from the analysis decreasing, therefore, the experimental variance. Sixty-nine genes fit these criteria. The range of mean Log2 ratio was from 0 to 0.15 and the SD of mean Log2 ratio was from 0.27 to 0.42. Analysis of 7 replicate array based on the same group of candidate endogenous reference genes demonstrated a similar distribution of mean Log2 ratio and SD of mean Log2 ratios (with exception of TFAP2B) suggesting that this variance could be attributed to predominantly experimental noise. Among these genes we further selected for validation 11 that had high mean fluorescence intensity (underlined in Table 1 ). However, probe preparation and other technical considerations limited to analysis to only 8 of these genes which included: NEED8, HIRIP5, GPLD1, RAB31, SNRPD3, FZD6, COL8A1 and AFAP. "Leave-one-out cross-validations" were used to estimate the error [ 33 , 34 ]. In each validation experiment one array was left out. The remaining 383 arrays were used to identify endogenous reference genes using the same parameters used for the original analysis (test/reference ratios < 2 SD in at least 95% of the experiments and average intensity > 2,000 in both channels). The median Log2 ratio of the endogenous reference gene in the left out array was then compared with the "real" normalization factor consisting of the median Log2 ratio for all genes in that array supposed to approximate a balanced expression of test and reference genes for that array. This procedure was repeated for each array 384 times. Only in 3 arrays out of 384 (2.3%) were found errors in endogenous reference gene selection (errors were considered median Log2 ratio of endogenous reference genes in the left out array > 2 standard deviations apart from the "real" normalization factor in the same array (which is zero for normalized arrays). The expression of each candidate endogenous reference gene identified in this study was then compared with available information about the expression of the same genes in 12 normal human tissues reported by the Affymetrix HG-U95A-E probe sets . We found a good correlation between the expression patterns observed by us and that reported by the Affymetrix GeenChip array (data not shown). In addition, a cluster of endogenous reference genes suggested by Applied Biosystems for standardization during qRT-PCR was analyzed by evaluating the SD of their Log2 ratios across the 384 arrays. As data indicated in Table 2 , although some genes (bold) demonstrated mean SD Log2 ratios comparable to that of the endogenous reference genes identified in this study, the corresponding percentage of the arrays in which Log2 ratios were within 2 SD from zero ranged between 0.68–0.85 which is significantly lower than the frequency in which the same parameter fell for the 69 candidate house keeping genes identified here (>95%). In addition, we looked at the variation in our arrays of genes classically used as endogenous reference genes including: ribosomal protein L32, HRPT, β-actin and tubulin-α3. In all cases the mean SD of Log2 Ratio was relatively larger than for the endogenous reference genes identified in Table 1 with the following respective values: 0.88 for Ribosomal protein L32, 1.13 for HRPT, 1.32 for β-actin and 1.42 for laminin-α3. Interestingly, however, the mean intensities were relatively higher for all of these genes compared with the ones identified in this study (7,369; 12,778, 43,794 and 42,241 for the four genes respectively) suggesting that these genes are expressed probably at higher concentrations in tissue and, therefore, although relatively variable in expression in different tissues, they have been a useful marker of RNA abundance when only rough estimates are required like, for instance, for Northern Blotting. Table 2 Gene Name Image ID Description Mean Int Mean logR SD logR % of arrays NKTR 712460 NK-tumor recognition protein=cy 7319.55 0.62 0.79 0.52 NKTR 712460 NK-tumor recognition protein=cy 3048.28 0.19 0.43 0.84 NKTR 712460 NK-tumor recognition protein=cy 8637.15 0.99 0.99 0.53 NKTR 2064497 natural killer-tumor recognitio 959.97 0.72 0.85 0.7 PPIC 882459 peptidylprolyl isomerase C (cyc 5034.82 2.11 1.45 0.44 PPIB 756600 peptidylprolyl isomerase B (cyc 6606.09 1.08 1.04 0.39 PPIL2 450661 peptidylprolyl isomerase (cyclo 3035.17 0.16 0.4 0.85 PPIL2 2017652 peptidylprolyl isomerase (cyclo 850.93 0.23 0.48 0.76 PPIG 809621 peptidyl-prolyl isomerase G (cy 3991.93 0.28 0.53 0.67 PPID 884500 peptidylprolyl isomerase D (cyc 5805.51 0.33 0.57 0.66 PPID 71154 peptidylprolyl isomerase D (cyc 5652.12 0.27 0.52 0.63 PPIH 767277 peptidyl prolyl isomerase H (cy 3277.26 0.28 0.53 0.73 PPIF 758343 peptidylprolyl isomerase F (cyc 15900.53 3.18 1.78 0.13 CYP2J2 454580 cytochrome P450, family 2, subf 2302.76 0.71 0.84 0.65 PPIA 2580290 peptidylprolyl isomerase A (cyc 25478.36 1.32 1.15 0.36 PPIA 1570861 peptidylprolyl isomerase A (cyc 5028.34 0.76 0.87 0.46 GAPD 50117 glyceraldehyde-3-phosphate dehy 38011.52 1.66 1.29 0.4 GAPD 530934 glyceraldehyde-3-phosphate dehy 10526.82 1.14 1.07 0.37 GAPD 530868 glyceraldehyde-3-phosphate dehy 5052.59 0.81 0.9 0.25 GAPD 755641 glyceraldehyde-3-phosphate dehy 1693.23 0.17 0.42 0.84 PGK1 949939 Phosphoglycerate kinase 1 4452.92 0.16 0.4 0.85 B2M 878798 beta-2-microglobulin 34211.05 2.08 1.44 0.43 AMBP 2063390 alpha-1-microglobulin/bikunin p 1813.96 0.15 0.39 0.86 GUSB 2273001 glucuronidase, beta 6205.14 0.81 0.9 0.41 GUSB 276449 glucuronidase, beta 2241.73 0.22 0.47 0.68 HPRT1 280507 hypoxanthine phosphoribosyltran 12778.31 1.29 1.14 0.35 TBP 280735 TFIID=TATA box binding protein 4065.4 0.24 0.49 0.73 TFR2 461750 transferrin receptor 2 2700.22 0.19 0.44 0.84 TFR2 2408681 transferrin receptor 2 2958.88 0.21 0.46 0.82 Test gene expression estimates according to house-keeping gene selection: the transcriptome array To validate the usefulness of candidate endogenous reference genes in large sample populations we developed a new tool that we are planning to use in the future for validation of gene expression across extensive data sets. This tool displays cDNA libraries originated from different tissues or cell lines individually spotted on a solid surface. The principle of this technology is similar to RNA dot blot which uses RNA isolated from samples and transferred to membrane making, therefore, the transcriptome array a high-density dot blot. The labeled DNA probe of interest is hybridized to the immobilized complementary strain of mRNA. A reference gene hybridization will carried out simultaneously to estimate the relative expression of the gene of interest compared with the reference gene. The new tool we describe here, termed transcriptome array, utilizes cDNA generated from source mRNA for target immobilization to improve the stability of the immobilized targets and differentially fluorescence-labeled test and reference probes (RNA or double strained DNA) then can co-hybridize on to the same spots. Using a validated RNA amplification technology [ 30 ], large quantity of pure amplified RNA with proportional representation of source mRNA species could be generated from which cDNA could be obtained through a reverse transcription reaction. Because of the minimum amount of cDNA used for fabricating each transcriptome array (<5 nano gram cDNA/spot) and the size of spots (100 um), the expression of a large number of genes can be theoretically analyzed on thousands of different samples simultaneously. Since the amount of cDNA spotted may vary according to the quality of the starting material and the efficiency of RNA preparation for each sample, absolute estimation of fluorescence from the hybridized probe is not informative of the expression of the given gene in each sample. Therefore, a reference system is necessary so that comparative expression of the test gene can be presented proportional to that of a consistently expressed gene across all samples. Therefore, interpretation of data derived from the transcriptome array relies on endogenous reference gene normalization. To test the usefulness of various endogenous reference gene we resorted to the well characterized expression of the melanocytic lineage-specific gene gp100/Pmel17 [ 31 ] that is expressed exclusively though heterogeneously in cells of melanocytic lineage [ 35 , 36 ]. The assumption of this experiment was that in spite of its heterogeneity of expression in samples of melanocytic origin, overall the expression of gp100 should be higher in meloanocytic compared with non melanocytic samples with a high degree of significance. Rho-C has been associated with the metastatization process in melanoma but its specificity for melanocytic lineage is unknown [ 37 ]. The differentiation control element DICE is found in the 3'-UTR of numerous eukaryotic mRNAs and there is no solid association between its pattern of expression and specific physio-pathological or developmental conditions [ 38 ]. Therefore, we compared the expression profile of these three genes in 829 cDNA libraries that included 106 melanoma cell lines, 127 melanoma metastases, 2 benign nevi (total of 235 melanocytic samples) and 593 miscellaneous samples containing a predominance of primary esophageal, renal and colon cancers paired with normal tissues from the same organ and a large number of circulating mononuclear cells. In addition, samples from most other normal and cancerous tissues were included although in smaller number (complete list of samples available upon request). As endogenous reference genes we chose β-Actin and new candidate genes identified by this study (see previous section). The test genes (gp100, Rho-C and DICE) were separately hybridized to the transcriptome array. Each gene labeled with Cy3 was co-hybridized with individual endogenous reference genes labeled with Cy5. We then compared Log2 Cy5 / Cy3 ratios between melanocytic and non-melanocytic tissues. gp100 was, as expected, expressed more in melanocytic lesions with high degree of significance no matter what gp100 / endogenous reference gene combination was used. The range of significance, however, varied extensively depending upon the gp100 / endogenous reference gene combination. This difference was considered representative of the ability of different endogenous reference genes to normalize for tissue-specific gene expression patterns (Figure 1 ). In details, all endogenous reference genes could segregate melanocytic from non melanocytic lesions with a high degree of significance (Figure 2 ; unpaired two sample t test p 2 -values ranged from 6 × 10 -8 to 3 × 10 -36 ). There was, however, a big range in the discriminatory capacity among endogenous reference genes with NEDD8, RAB3 and FZD6 providing the highest predictive value ( t test p 2 -value = 3 × 10 -36 , 1 × 10 -32 and 1 × 10 -21 respectively) and β-actin being one of the least reliable ( t test p 2 -value = 1 × 10 -8 ). Figure 1 Comparison of gp100 expression normalized by different endogenous reference genes. Candidate endogenous reference genes were selected among those in Table 1. The gp100 probe (Cy3) was co-hybridized with one endogenous reference gene probe (Cy5) at the time to the transcriptome array. The transcriptome array included 235 cDNA libraries derived from samples of melanocytic lineage (maroon bar) and 594 cDNA libraries from samples of non melanocytic lineage (green bar). Melanocytic samples consisting in the overwhelming majority of melanoma metastases or melanoma cell lines while non-melanocytic libraries included a large collection of esophageal, kidney, colon and other cancers and several normal tissues or circulating mononuclear cells. The complete list of samples is available upon request. The expression of gp100 normalized with different endogenous reference genes was compared by unpaired two-tailed student t test and the endogenous reference genes were ranked according to the level of significance in their ability to discriminate between melanocytic and non-melanocytic lesions (data presented as the Log10 p 2 -value (shown in the boxes associated with individual graphs; for details see Figure 2). The distribution of the Log2 ratios for each individual cDNA library is shown for each gp100 / endogenous reference gene combination. In addition, results compiled using the average of the Log2 ratios for all the endogenous reference genes, the 5 and the 3 with the lowest p 2 -value are presented (orange bar graphs). Figure 2 Expression of gp100 by melanocytic and non melanocytic lesions normalized with different endogenous reference genes. Average gp100 / endogenous reference gene Log2 ratios for melanocytic and non-melanocytic lesions are shown together with the standard deviation (SD) standard error from the mean (SEM) and the Log 10 of the t -test p 2 -values when melanocytic and non-melanocytic lesions were compared. Also data derived from mathematically averaging the results obtained with all the endogenous reference genes, the ones yielding the best 5 individual p 2 -values and the best three are shown. The same data are presented visually in the bar graph below; in black are data derived with individual endogenous reference gene normalizations, in orange data derived by averaging results from different endogenous reference genes. Filled bares portrait data from melanocytic lesions, empty bars from non-melanocytic lesions. The large variation in the results obtained using different endogenous reference genes to normalize gp100 expression could have been due to a higher stability of the expression of some genes across all samples or to a differential expression of the endogenous reference genes themselves in melanocytic lesion. In the latter case, the better results obtained could be coincidental and not useful in other experimental situations. We, therefore, tested whether pooling results obtained with all endogenous reference genes independently of each predictive value in this controlled experimental situation could yield results as informative about gp100 pattern of expression as those obtainable with individual endogenous reference genes, particularly those that provided the best prediction. The advantage of this strategy is that it does not depend on previous knowledge of gene expression patterns for the selection of individual endogenous reference genes not applicable in conditions, in which the suitability of a gene for a given experimental situation is, contrary to gp100, is unknown. Thus, we compared the predictive value of the average of gp100 / endogenous reference gene Log2 ratios obtained with all endogenous reference genes and those obtained using the genes that provided the best five or the best three results (Figure 2 ). The use of pooled information from various endogenous reference genes appeared to stabilize the non-melanocytic Log2 ratios gp100 / endogenous reference. In fact, the SD of the Log2 ratios gp100 / endogenous reference among non-melanocytic samples derived by pooling endogenous reference gene results were significantly lower than the SD obtainable with any individual endogenous reference gene (F-test). The basis for this test was that non-melanocytic lesions uniformly should not express gp100 and, therefore, the Log2 ratios for any gp100 / endogenous reference gene combination should be constant resulting in very low SD. SD for non-melanocytic lesions were much lower using pooled endogenous reference genes for the normalization (Figure 2 ) and this was true at extreme levels of significance (F-test value for SD when all endogenous reference genes were used for normalization compared with individual endogenous reference were 0, 1.5 × 10 -9 , 0, 5.6 × 10 -7 , 9.0 × 10 -5 , 0, 0, 0 and 0 for NEDD8, RAB3, FZD6, AFAP, COL8A, SNRPD, HIRIP5, β-actin and GPLD1 respectively). Similar results were obtained when results from the best 5 and best 3 endogenous reference genes were pooled together. In addition, even if pooling data together did not provide the highest level of discrimination compared to the best results obtained with selected individual endogenous reference genes, the capacity of pooled data sets to discriminate gp100 expression between melanocytic and non melanocytic samples was still very high (2 × 10 -22 ; 4 × 10 -29 and 1 × 10 -34 respectively when all, the best 5 or the best 3 endogenous reference gene results were pooled together). Interestingly, different endogenous reference genes provided not only different levels of discrimination but also provided different estimates of gp100 expression in melanocytic lesions with gp100 / endogenous reference gene Log2 ratio above (NEED8, RAB3, FZD6, AFAP, SNRPD3, HIRIP5, GPLD1) or below (COL8A1, β-actin) 0 (Bottom graph, Figure 2 ). This, of course, has no impact on the ability to characterize patterns of expression of different genes in different tissues but rather suggests that the latter group of endogenous reference genes is expressed at higher copy number, therefore, biasing the data toward its own fluorescence channel. Simple adjustment in probe concentration could easily solve this problem. We then analyzed the expression of Rho-C and DICE whose pattern of expression in the transcriptome array cannot be predicted based on available information save for the notion that of the expression of Rho-C is related to the metastatic process of melanoma [ 37 ]. The pattern of Rho-C was characterized with all the endogenous reference genes and representative information is presented in Figure 3 . Most endogenous reference genes yielded a pattern suggestive of a preferential expression of Rho-C in melanoma metastases. In addition, averages of Rho-C / endogenous reference gene Log2 ratios demonstrated clearly a specific expression of this gene in melanoma metastases. This pattern was not necessarily specific for melanoma as the other cancerous tissues spotted in the transcriptome array were obtained from primary lesions. Thus, it is possible that the preferential expression in melanoma metastases was due to the metastatic process rather than the melanocytic lineage. Interestingly, the melanoma cell line Mel-A375 did not constitutively express Rho-C as previously observed by Clark et al. [ 37 ]. In addition, none of the melanoma cell lines mostly derived from melanoma metastases demonstrated expression of Rho-C. This information suggests that Rho-C may be involved in the natural metastatic process in vivo but is not constitutively expressed in vitro . Finally, DICE did not demonstrate any specificity of expression and appeared variably expressed in all specimens independently of the endogenous reference gene used for normalization. Figure 3 Relative expression of gp100, Rho-C and DICE in melanocytic and non-melanocytic samples. Test samples (Cy3) and endogenous reference genes (Cy5) were hybridized to the transcriptome array as described in Figure 1. Test gene / endogenous reference gene Log2 ratios are displayed as a bar graph for NEED8, β-actin and for the averages of test / endogenous reference gene Log2 ratios when the three endogenous reference genes with the best discriminating power in separating gp100 expression between melanocytic (maroon bar) and non-melanocytic (green bar) lesions (NEDD8, RAB3 and FZD6) were used. Samples derived from melanoma metastases are shown by the orange bar. In the boxes the Log10 p 2 -value of significance of differences between relevant samples (se text) are shown; ns = non significant). Validation of data by protein analysis and quantitative real-time PCR (qRT-PCR) We tested whether gp100 / endogenous reference gene Log2 ratios correlated with gp100 protein expression level as measured by intra-cellular FACS analysis (Figure 4 ). The gp100 / endogenous reference gene Log2 ratios were derived from the transcriptome arrays while protein expression was based on previous characterizations of the same cell lines [ 35 , 39 ]. Overall, a good correlation (Pearson's correlation) was noted for gp100 values normalized with most endogenous reference genes. However, with the exception of NEDD8, the best correlation was obtained when pooled information was used from the best 5 and 3 candidate endogenous reference genes as defined before. Interestingly, RAB3 that scored very high as a predictor of gp100 transcriptional expression in melanocytic lesions and β-actin provided the worst correlation with protein expression. This data suggest that, although different endogenous reference genes may yield better predictive value than others in large data sets, it is likely that their reliability varies in different conditions and possibly the best results can be obtained by pooling information from several of them. Figure 4 Correlation of gp100 expression estimated obtained by intracellular FACS analysis and by gene profiling with the transcriptome array. Cell lines with different level of gp100 expression by intra-cellular FACS analysis using the gp100-specific mAb HMB45 [35] were used for the analysis. The compiled information from the transcriptome array in which cDNAs from each cell line was spotted are shown. The Cy5/Cy3 (endogenous reference gene / gp100) Log2 ratios are shown for each endogenous reference gene evaluated. In addition, ratios derived by using the data from all the endogenous reference genes, the best 5 and best 3 according to figure 2 are also presented. Pearson's correlation for the data set is shown for each endogenous reference gene and the same information is summarized graphically at the bottom of the figure. MEF = mean equivalent of fluorescence [35]. We also tested by quantitative real-time PCR (qRT-PCR) the discriminatory capacity of endogenous reference genes in samples with known patterns of expression of gp100. We selected 6 previously characterized [ 35 , 36 ] melanoma cell lines four of which were known to express (MEL-553B, MEL-1317, MEL-526, MEL-1102) and two known not to express (MEL-836, MEL-1195) gp100 at the transcriptional and protein level (Figure 5 ) [ 36 , 39 ]. In addition, 4 fine needle aspirate (FNA) samples from melanoma metastases expressing gp100 [ 40 ] and a series of tumors (2 renal cell, 2 esophageal one gastric carcinomas and HELA cells line) known not to express gp100 were tested. Absolute expression of gp100 was measured and although the lesions supposed to be gp100 expressing appeared more positive than those supposed to be gp100 negative there was no absolute demarcation among the two groups (Figure 5a ). A better differentiation between positive and negative lesions could be obtained when the various endogenous reference genes were used for the analysis or the combined values from all of them (Figure 5b ). Figure 5 Absolute and normalized expression of gp100 based on qRT-PCR. Four melanoma cell lines expressing gp100 (MEL-553B, MEL 1317, MEL 526 and MEL-1102) and 2 not expressing gp100 (MEL-836 and MEL-1195) protein by FACS analysis [35] (maroon horizontal bar) were tested for gp100 expression by qRT-PCR as previously described [40]. In addition four fine need aspirates of melanoma metastases all expressing GP100 by immunohistochemistry (orange horizontal bars) and two renal cell, two esophageal and one gastric cancer specimens all not expressing gp100 (green horizontal bar) were tested. Similarly the gp100 not expressing HeLa cell line was tested. Absolute gp100 copy number is shown in (a); normalized expression of gp100 is shown using the average of all endogenous reference gene results (b). The protein expression for the melanoma cell lines is presented as mean equivalent of fluorescence (MEF) [35] as well as immunohistochemical evaluation while the expression by FNA is shown in (c) based on immunohistochemical (HIC) evaluation as previously described [46]. Absolute copy numbers of gp100 estimated by qRT-PCR are also shown. Discussion Although the concept of endogenous reference genes is appealing it may be unrealistic to expect that a gene could be equally expressed in all eukaryotic cells in all physiopathological conditions. Thus, the endogenous reference gene concept can only approximate such ideal biological behavior. In fact, numerous studies noted that the expression of the ostensible endogenous reference genes varies according to distinct physio-pathological situations and can be affected by experimental manipulation [ 2 , 20 , 24 , 26 , 27 , 41 ]. For instance, the levels of expression of GAPDH, cyclophilin and β-actin fluctuate in different tissues, disease stage and are affected by cell behavior like proliferation [ 6 , 15 , 18 , 20 , 42 ]. In addition, endogenous reference genes expression varies markedly in cancer compared to normal tissue [ 17 , 26 , 41 ] suggesting that oncogenesis may influence their expression [ 25 ]. In this study, therefore, we looked for relative stable endogenous reference gene challenging a 17,000 cDNA clone data base against approximately 400 specimens that included a broad variety of normal and neoplastic tissues as well as cell lines. Indeed, few previously unnamed genes approximated the ideal endogenous reference gene by being expressed consistently in more than 95% of the specimens analyzed within a variance that was barely and insignificantly different from what could be expected from intrinsic experimental fluctuation (Table 1 ). In fact, the threshold that we selected based on the variance observed by periodically repeating identical experiments was consistent with criteria suggested by others [ 43 ] who claimed that genes with fold changes greater than 1.5 can be classified as differentially expressed at ~95% confidence interval. This is consistent to our selection criterion that contained endogenous reference genes within a 1.46 fold maximal variation. Although these genes have not been proposed before as candidate endogenous reference genes, analysis of available data bases suggested that they are relatively stably expressed in various normal tissues (i.e. Affymetrix HG-U95A-E probe sets). Selection of these candidate endogenous reference genes provided useful information when the expression of a lineage specific gene such as gp100 was compared between relevant and irrelevant samples. In addition, gp100 expression normalized according to some of the new endogenous reference genes more closely aligned to lineage specificity than normalization performed using β-actin. This data, however, confirm that there is no such thing as a perfect endogenous reference gene. Although some endogenous reference genes demonstrated high power of discrimination their ranking was dictated by previous knowledge of the expected experimental results (gp100 being melanocytic lineage specific). However, gene expression analysis is done to determine such specificity and in most cases no previous knowledge of tissue specificity is truly available. One may argue that, in spite of the extreme variability in normalization power demonstrated by various endogenous reference genes, still significant differences could extracted between the melanocytic and non-melanocytic samples no matter what endogenous reference genes were used. This statement is correct in this experimental situation where hundreds of samples could be compared with the transcriptome array. However, in cases when only few samples are analyzed the difference in discriminatory capacity becomes critical. For instance, when randomly selected gp100 / endogenous reference gene Log2 ratios from 10 melanoma metastases were compared with those from 10 kidney specimens only NEDD8, RAb3, FZD6, AFAP, COL8, SNRPD and pooled endogenous reference gene-based normalizations yielded significant differences between the two groups while non significant values were obtained for normalizations based on HIRIP5, β-actin and GPLD1. The lowest limit of detection in the transcriptome array of specific probes in term of copy numbers is currently under investigation. Decreasing concentrations of β-actin alone printed the array spots as internal controls suggest that the detection range is from 30 ng / spot to 3 pg / spot when standard fluorescent intensity range parameters are applied. The higher limit of detection before saturation can be easily adjusted to avoid under estimates of highly expressed genes, therefore, increasing the dynamic range of detection. This is achieved by adjusting the fluorescent intensity right below the saturation for both channel and subsequently calculating the abundance of reference gene expression using serially diluted internal control spots as standard curve. With few exceptions, the expression of most genes can be estimated with relative accuracy using this semi-quantitative approach. Since in most cases endogenous reference genes are selected based on previous knowledge of their usefulness for a particular experimental situations, it is likely that the best consistency could be achieved in such circumstances when averages of several endogenous reference gene normalizations are used. This may represent a reasonable solution particularly if a cocktail of endogenous reference genes identically labeled could be used as reference within a single experiment. We are presently testing various combinations of lineage specific genes and pools of endogenous reference genes for hybridization to the transcriptome array and for qRT-PCR. Such attempts, however, have been quite disappointing so far as results obtained using mixture of endogenous reference genes do not yield information comparable to that achievable when results obtained with the same endogenous reference genes used individually are pooled together. This is possibly due to imbalanced density of various endogenous reference genes in different samples with predominant effects of some over others. Therefore, extensive work in the future will be aimed at equilibrating the interactions among endogenous reference genes in various experimental conditions to test whether mixtures of them could be used in multiplex for normalization of large expression analysis studies. Conclusions Our observations, together with previous work by others indicate that much caution should be taken when using endogenous reference to normalize the expression of test genes. It is likely that previous analyses based on single endogenous reference genes might have been strongly biased by the individual selections and the interpretation of the results should looked at with caution. In particular, this study suggests that even when the most stably expressed genes in array experiments are used as endogenous reference genes, significant variation in test gene expression estimates may occur and the best normalization is achieved when data from several endogenous reference genes are pooled together to minimize minimal but significant variation among samples. We are presently optimizing strategies for the preparation of endogenous reference gene mixtures that could yield information comparable to that of data pooled from individual endogenous reference gene normalizations. Methods The list of samples used for cDNA microarray hybridization or for preparation of the transcriptome array is available upon request. Most melanoma and renal carcinoma cell lines were generated at the Surgery Branch, National Cancer Institute (NCI), National Institutes of Health (NIH), Bethesda, MD and maintained in RPMI (Biosource International, Camarillo, CA) supplemented with 10% fetal calf serum (Biosource). Cell lines were harvested by Trypsin /Versene (Biosource) digestion. PBMC were collected by leukapheresis from six unrelated normal donors in the Department of Transfusion Medicine, Clinical Center, NIH and purified by Ficoll gradient separation. Surgically removed tumor tissues were collected from the Tissue Network (Philadelphia, PA), Surgery Branch and CPSB specimen bank, NCI, (Bethesda, MD). Fine needle aspiration (FNA) biopsies were obtained from patients with metastatic melanoma referred to the Surgery Branch, NCI for immunotherapy. Total RNA was isolated using Qiagen RNeasy kit and its quality and quantity was estimated using Agilent Bioanalyzer 2000 (Agilent Technologies, Palo Alto, CA). Target preparation for cDNA microarray hybridization Total RNA extracted from test and reference samples was transcribed in vitro into anti-sense RNA (aRNA) and reverse-transcribed into fluorescence labeled cDNA for hybridization to 17,000 gene cDNA-based microarrays [ 30 , 44 ]. Pooled PBMC were used to prepare reference aRNA to be co-hybridized in all experiments with test aRNA. cDNA targets were labeled with Cy3 (green) for reference material and Cy5 (red) for test material with the exception of reciprocal experiments. cDNA Microarrays cDNA (UniGene cluster) microarrays were printed at the Immunogenetics Section, DTM, CC, NIH with a configuration of 32 × 24 × 23 containing 17,500 elements. Clones used for printing were selected from the Research Genetics RG_HsKG_031901 8 k clone set and the 9k RG_Hs_seq_ver_070700 40 k clone set. The 17,500 spots included 12,072 uniquely named genes, 875 duplicate genes and about 4,000 expression sequence tags. For a complete list of genes included in the Hs-CCDTM-17.5k-1px printing please visit our web site at . Transcriptome array A collection of aRNA-based libraries was prepared from 960 frozen tissue samples or cell lines and individual aRNAs were reverse transcribed into cDNA in the presence of 1 μl of dN6 primer (8 μg/μl), 6 μl of first strand buffer, 3 μl of 10 mM dNTP, 3 ul of 50 mM dTT, 1.5 μl of Rnasin and 10 μl of aRNA (6–12 ug) with addition of 3 μl of Superscript II (BRL) in 5.5 μl volume after heating to 65°C for 5 minutes. Reactions were carried out in 96-well plates at 42°C for 90 minutes followed with addition of 1 μl Rnase H and heating for 20 min at 37°C. cDNA were further purified by utilizing CentriCept gelfitration plates to remove non incorporated primer and dNTP. Purified cDNA were transferred to 384-well plates and dried by speedvac. Samples were re-suspended in 13 μl of 3 × SSC followed by shaking at 3,000 rpm for 15 min. Reconstituted cDNA libraries from individual samples and spiked reference gene at different concentration were printed on to poly-L-lysine coated glass slides at the concentration of 0.5–1 μg/μl using the OmniGrid (GeneMachine. San Carlos, CA) printer and Telechem printing pins (TeleChem International, Inc. Sunnyvale, CA). Each 100 μm diameter spot was duplicated at a 250 μm distance. After complete exsiccation, slides were post-processed as described at . Probe design Transcriptome array hybridization experiments Specific PCR primers for the amplification of each gene probe were designed using Primer 3 program . In order to generate single strand fluorescence labeled cDNA for hybridization, modified specific 5' primers with an extension of the T7 promoter region were used for PCR amplification. Double strand PCR fragments (300–800 bp) were used as template for in vitro transcription to generate single stranded mRNA. 3 μg of amplified RNA were then reverse transcribed into cDNA in present of 4 μl of first strand buffer, 1 μl dN6 primer (8μg/μl; Boehringer Mannheim), 2 μl 10 X low T-dNTP (5 mM A, C and GTP, 2 mM dTTP), 2 μl Cy-dUTP (1 mM, Cy3 or Cy5), 2 μl 0.1 M DTT, 1 μl RNasin, 3 μg amplified mRNA in 8 μl DEPC H 2 O. After heating to 65°C for 5 minutes and cooling to 42°C, 2 μl of SSII was added and the labeling reaction was carried out at 42°C for 90 min. Probe purification and hybridization were performed as previously described [ 30 , 44 ]. Quantitative real-time polymerase chain reaction (RT-PCR) Primer and probe sets from each candidate genes were designed using the Primer Express 2.0 program (Applied Biosystems, Foster City, CA) and synthesized by BioSource. Taqman probes were labeled at the 5' end with the reporter dye molecule 6-carboxy-fluorescein (FAM; emission λ max = 518 nm) and at the 3' end with the quencher dye molecule 6-carboxytetramethyl-rhodamine (TAMRA; emission λ max = 582 nm). Total RNA was used for gene expression validation. Measurement of gene expression was performed using the ABI 7900HT sequence detection system (Perkin Elmer, Foster City, CA) as previously described [ 40 , 45 ]. Messenger RNA from test samples were reverse transcribed into cDNA in the presence of oligo dT (12–18 mer). The qRT-PCR procedure was performed by alternating 2 minute cycles at 50°C and 10 minute at 95°C. Forty cycles involving denaturation at 95°C for 15 seconds and annealing/extension at 60°C for 1 minute in 50 μl volume with 1 × TaqMan Master Mix (PerkinElmer). Standard curves were generated for each gene with high and accurate PCR amplification efficiency as determined by the slope of the standard curves. Linear regression analysis of all standard curves documented in all cases an R value ≥ 0.99. The sequence of primer sets and probes used for each gene are shown in Table 3 with statistics for each standard curve. Standard curve extrapolation of copy number was performed for the gene of interest as well as an endogenous reference gene for each sample. To correct for concentration of starting material, normalization of samples was performed by dividing the copies of the gene of interest by copies of the reference gene as previously described [ 40 ]. Table 3 Quantitative RT-PCR primers and probes AFAP (+)-TGTCAAGTTAAACCACTAATGTGTTGGT Y = -3.105x + 43.726 R2 = 0.99 (-)-GGCATCCAAATTCTCCAAGAAA FAM-TGCTGCCTCTCCTGAGTAGGGTGGGT-TAMRA CDK5R1 (+)-TCCTACATGGGCAACGAGATC Y = -3.315x + 45.983 R2 = 0.99 (-)-CCAAAAGGCCTCCTTGCA FAM -TACCCGCTCAAGCCCTTCCTGGTG -TAMRA COL8A1 (+)-CCGAGCTAACCGCACCTTT Y = -3.143 x +43.294 R2 = 0.99 (-)-GTCTGCGGGTTGTAGTTCTGTCT FAM -AGTGAAGTTTAACAAACTGCTGTATAACGG -TAMRA FZD (+)-AGCCTCAAAGGTTCCACATCTC Y = -3.515 x + 47.092 R2 = 0.99 (-)-AGGTCACTTCCAGTGTAACACAAATT FAM -TGAGAAAAGAGCAGGGAGGTGGTTGTCA -TAMRA GPLD1-1 (+)-CAGATTGAAGATTTCACTGCATTTC Y = -3.653 x + 49.086 R2 = 0.99 (-)-CATCAAAATGCTCACCATGGA FAM -TCTGCCCACCTCTCTCATGCTGAATCAC -TAMRA GPLD1-2 (+)-CTTTTGCCTGTAGTAGTAAATTGCTTTTA Y = -3.168x + 46.599 R2 = 0.99 (-)-AACTGGCCATATAAC CAAAGGTGTT FAM -TGAATGGTGTTTATTAAACCCTTATGGTCGATATTTCC-TAMRA HIRIP5 (+)-GCTGCCCTAGTTCAATCATTACTCT Y = -3.160x + 43.313 R2 = 0.99 (-)-CGCCTTCTACCTCCGGAATAT FAM – AAAAATGGAATTCAGAACATGCTGCA-TAMRA HIST1HIA (+)-AGGCGTCCTCCGTGGAA Y = -3.143x + 43.294 R2 = 0.99 (-)-ATGCACCCGTTGCCTTAGTT FAM – AGCCCGGCGCCTCAAAGGTG-TAMRA NEDD8 (+)-TGACCGGAAAGGAGATTGAGA Y = -3.458x + 48.686 R2 = 0.99 (-)-CCACACGCTCCTTGATTCG FAM – TGACATTGAACCTACAGACAAGGTGGA-TAMRA PTH (+)-GCATAACCTGGGAAAACATCTGA Y = -3.063x + 45.215 R2 = 0.99 (-)-TGCACATCCTGCAGCTTCTT FAM -TCGATGGAGAGAGTAGAATGGCTGC-TAMRA RAB31 (+)-CCCCTGAAGGATGCTAAGGAA Y = -3.303x + 48.517 R2 = 0.99 (-)-AGCATTTTTTGCACTTGTCTCAAC FAM – ACGCTGAATCCATAGGTGCCATC-TAMRA SNRPD3 (+)-GGCACAGCTGGAGCAGGTAT Y = -3.177x + 47.200 R2 = 0.99 (-)-CTTCAGCATGTCAGGCAAAATC FAM – ATCCGTGGCAGCAAAATCCGC -TAMRA β-actin (+)-GGCACCCAGCACAATGAAG Y = -3.127x + 44.733 R2 = 0.99 (-)-GCCGATCCACACGGAGTACT FAM-TCAAGATCATTGCTCCTCCTGAGCGC-TAMRA gp100 (+)GGTTCCTTTTCCGTCACCCT Y = -3.282x + 45.987 R2 = 0.99 (-)CTCACCGGACGGCACAG FAM-ACATTGTCCAGGGTATTGAAAGTGCCGAGAT-TAMRA Authors' contributions PJ validated the system combining high-throughput data with gene-directed analysis by quantitative real-time PCR. YZ performed statistical analysis. YN, MCP, DN, VM and KS participated in the preparation and utilization of array experiments used for the identification of stably expressed genes. NH, HS and PRT provided a large library of samples necessary for the development of the transcriptome array. FMM supervise the project as principal investigator. EW provided leadership for the development of cDNA arrays, transcriptome array and for the overall conduct of the experiments.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516027.xml
554980
The health workforce crisis in TB control: a report from high-burden countries
Background Human resources (HR) constraints have been reported as one of the main barriers to achieving the 2005 global tuberculosis (TB) control targets in 18 of the 22 TB high-burden countries (HBCs); consequently we try to assess the current HR available for TB control in HBCs. Methods A standard questionnaire designed to collect information on staff numbers, skills, training activities and current staff shortages at different health service levels was sent to national TB control programme managers in all HBCs. Results Nineteen HBCs (86%) replied, and 17 (77%) followed the questionnaire format to provide data. Complete information on staff numbers at all service levels was available from nine countries and data on skill levels and training were complete in six countries. Data showed considerable variations in staff numbers, proportions of trained staff, length of courses and quality of training activities. Eleven HBCs had developed training materials, many used implementation guidelines for training and only three used participatory educational methods. Two countries reported shortages of staff at district health facility level, whereas 14 reported shortages at central level. There was no apparent association between reported staff numbers (and skills) and the country's TB burden or current case detection rates (CDR). Conclusion There were few readily available data on HR for TB control in HBCs, particularly in the larger ones. The great variations in staff numbers and the poor association between information on workforce, proportion of trained staff, and length and quality of courses suggested a lack of valid information and/or poor data reliability. There is urgent need to support HBCs to develop a comprehensive HR strategy involving short-term and long-term HR development plans and strengthening their HR planning and management capabilities.
Background The performance of health care systems is closely related to the numbers, distribution, knowledge, skills and motivation of its workforce, particularly of those individuals delivering the services [ 1 ]. Improvements in global health are greatly dependent on how well health systems can meet the demands placed on them by governments, programmes, communities and ultimately individuals. Human resources for health (HRH), all categories of clinical and non-clinical staff who make each individual and public health intervention happen, constitute a sine qua non of health systems. Therefore, developing HRH and fostering appropriate HR management are crucial steps towards achieving and sustaining improved and equitable health. Tuberculosis (TB) constitutes the third most important cause of death and disability [ 2 ] among infectious diseases. It is estimated that in 2002, there were 8.8 million new TB cases worldwide (141/100 000) of which 3.9 million were sputum-smear positive (SS+). Despite control efforts, the global incidence of TB continues to grow in some regions, particularly in sub-Saharan Africa [ 3 ]. The emergence of AIDS and multidrug-resistant TB (MDR-TB) pose further challenges for global TB control; in 2000 there were 1.82 million TB-related deaths, of which 226 000 (12%) were attributable to HIV [ 2 , 4 ]. The World Health Assembly (WHA), in 1991, pledged countries to achieve detection of at least 70% of estimated infectious TB cases (SS+) and to cure 85% of them by the year 2000 [ 5 ]. Slow progress resulted in the deferral of these targets until 2005 [ 6 ]. Likewise, the United Nations commitment to sustaining development and eliminating poverty throughout the world led world leaders to formulate the eight Millennium Development Goals (MDGs), among them to halt, and begin to reverse, the global incidence of TB by 2015 [ 7 ]. The creation of the Stop TB Partnership in 1998 [ 8 ] and the Global Fund to fight AIDS, TB and Malaria (GFATM) in 2001 [ 9 ] represented significant developments in the fight against TB, thanks to increased financing and technical assistance made available to endemic countries. However, disease-specific programmes (including TB) are still struggling to meet their targets, and governments and their financial/technical partners have finally recognized this is largely due to shortcomings in the health care workforce [ 3 , 10 ]. Low-income and middle-income countries (LMIC) urgently need a sufficiently large health care staff with appropriate expertise, experience and motivation, working at the right places. Lack of HR strategies, inadequate HR planning and management, poor deployment practices, inflexible contracting arrangements and inability to create new posts or increase salaries resulting from international regulations capping social sector spending have contributed to the global HRH crisis [ 10 ]. In addition, poor salaries, low morale and worsening local economic circumstances result in low recruitment/retention, internal and external migration of trained staff and attrition of the health care workforce. The HIV/AIDS epidemic has increased pressure on health systems and causes death and disability of the workforce itself [ 11 ]. In 2003, national TB programme (NTP) managers from 18 of the 22 TB high-burden countries (HBCs, countries that together account for approximately 80% of the global TB burden), ranked inadequate HR first within the top five constraints to reaching the WHA global TB control targets [ 3 ]. Insufficient numbers, lack of adequately qualified or trained staff at different service levels, inadequate distribution, low motivation and poor staff retention were commonly described. Although some components of disease control programmes still remain vertical (e.g. drug procurement or laboratory QA) the majority of front-line services are now partially or totally integrated to primary health care and therefore a shortage of HR for TB control represents a shortage of HR for health delivery. There is very limited published literature on HRH issues, particularly in LMIC and there are both scanty information on methods to assess HR capacity and lack of evidence on how best to evaluate interventions to strengthen and/or build HR capacity [ 12 ]. The dearth of published information probably stems from the persistent neglect of HRH development issues; in addition, research in HRH involves a broad scope of disciplines and often different research methodologies than those recognized in clinical medicine [ 13 ]. Currently there is increasing awareness that HRH concerns must be addressed in order to reach the MDGs, to expand access to priority interventions, to promote health systems development and to achieve global health equity [ 14 , 15 ]. However, without reasonably accurate information on numbers, location, qualifications and activities of staff, it is not possible to effectively manage or plan HR for the country's health services or for specific programmes [ 16 ]. This paper reports the results of a questionnaire sent to NTP staff in the 22 HBCs to assess the workforce available for TB control (staff numbers, cadres and skills), as well as the estimated HR requirements for appropriate TB control in HBCs. It aims to inform the development of more reliable methods of gathering qualitative and quantitative information on HR and to stimulate a long-overdue discussion on HR for TB control issues, so that governments of endemic countries and technical and financial partners can finally begin to address them jointly. Methods Participants NTP managers and country-based WHO staff in the 22 HBCs: Afghanistan, Bangladesh, Brazil, Cambodia, China, Democratic Republic of the Congo, Ethiopia, India, Indonesia, Kenya, Mozambique, Myanmar, Nigeria, Pakistan, Philippines, the Russian Federation, South Africa, Tanzania, Thailand, Uganda, Viet Nam and Zimbabwe. These countries account for more than half of the world's population and approximately 80% of the global TB burden Questionnaire Discussions with NTP managers and a literature review on HRH informed the development of a questionnaire to ascertain current staff provision; quality and intensity of training; time and type of personnel involved in performing different TB control activities; and estimated staffing needs at different health service levels. The questionnaire was pilot-tested internally within the Stop TB Department (STB), WHO headquarters, Geneva; and externally with staff from the NTP in Indonesia. The first section of the questionnaire assessed staff numbers and skills. It included open questions about absolute numbers of staff involved in delivering TB control activities at each service level (provincial, district and health facility and laboratory personnel). Skills were assessed using a composite of the proportion of staff receiving training in the previous three years (at each level) and the quality of the training provided (length of courses, development of educational materials and/or use of standard -WHO or International Union Against Tuberculosis and Lung Diseases (IUATLD) – materials) as proxy measures. Section two of the questionnaire addressed estimated HR gap. Since this paper specifically refers to TB control, we estimated the workload involved in adequate TB control in HBCs. Workloads were assessed by means of open questions on the different processes involved in the management of new SS+ patients (diagnosis and administration of a short course chemotherapy regimen under proper case management conditions including directly observed therapy, DOT), the estimated duration of each task and the type of staff involved in delivering them. A series of worksheets containing the different tasks required for TB case management were designed to assist respondents completing the questionnaire and to improve standardization. Capacity of health services at two different CDRs were assessed (current CDR and at the target 70% CDR). Workloads at current CDRs were calculated by multiplying the current numbers of new SS+ patients by the estimated time needed to treat a new SS+ patient; workloads at the 70% CDR were calculated by multiplying the figure corresponding to 70% of the estimated SS+ TB cases for each country by the time needed to treat a new SS+ patient. Eleven hours was used as the time needed to treat a new SS+ patient; this had been estimated previously by a TB experts' consensus (WHO, unpublished data). Survey In March 2003, the questionnaire, together with detailed instructions in English, was e-mailed to NTP managers and country-based WHO staff in the 22 HBCs. Three e-mail reminders were sent monthly after the return deadline; follow-up with several countries included e-mail and telephone communications to clarify responses. Data were entered into spreadsheet software for the analysis; qualitative answers were discussed by the authors. Results Response rate Nineteen of the 22 HBCs returned the questionnaire (86% response rate). Two countries provided information on their HR for TB control but did not use the questionnaire format and were thus excluded from the analysis; one country reported that it would take a substantial amount of time to get reliable information from such a large country and provided other readily available HR information; another reported that given the country's current staffing deficits and competing priorities they were unable to complete the survey. Further discussions revealed that some HR information was available in this country, although it was scattered in different sources and lengthy to compile. Both countries requested technical support to assess staffing needs at different levels and to assist their government's HR development programmes. Despite e-mail and telephone remainders, three countries failed to respond. Numbers of staff Information on estimated numbers of staff was available from all 17 countries but complete in nine (53%) of them. Staff numbers within the same service level varied considerably between countries; i.e. numbers of staff at provincial level varied from 8704 in one country to 6 staff in two countries (Table 1 ); similar variations were seen at other service levels. More countries provided information on staff numbers at provincial than at district or health facility levels.; only nine countries (53%) provided numbers of staff at laboratory level. Table 1 Staff numbers at each level and estimated numbers of trained staff in the previous three years (2000–2002) Country Health service Level Provincial District Health facility Laboratory Total Trained (%) Total Trained (%) Total Trained (%) Total Trained (%) Afghanistan 40 22 (55) 360 - - - - 30 (--) Bangladesh 460 150 (33) 460 120 (26) 39329 750 (2) 1015 450 (44) Brazil 27 19 (70) - 20 (--) - 6379 (--) - - Cambodia - 72 (--) - 236 (--) 1120 705 (63) - - DR Congo 56 19 (34) 306 - 4306 - 1000 - Ethiopia 12 4 (33) - - - - - - Indonesia* 70 70 (100) 420 420 (100) 1256 1256 (100) 405 405 (100) Kenya 10 ¶ 7 (70) 94 - 45900 1148 (2) 2121 350 (16) Myanmar - 8 (--) - 276 (--) - 18056 (--) - 327 (--) Nigeria 37 19 (51) 664 149 (22) 27000 295 (1) 3000 160 (5) Pakistan 6 0 (0) 60 60 (100) 21000 500 (2) 600 200 (33) Philippines 156 - 748 - 13900 - 2200 - Russian Fed. § - - - 4062 (--) - 4197 (--) - 1167 (--) South Africa 9 9 (100) - (70–80) 202265 (70–80) - - UR Tanzania 25 2 (8) 156 58 (37) - 50 (--) - - Uganda 6 3 (50) 55 - 1050 - 350 - Viet Nam 8704 2205 (25) 2705 1248 (46) 10510 10510 (100) 804 624 (77) * Percentage of training according to 2002 training plan ¶At regional level §Federation Skills Complete information on numbers of staff trained in the previous three years was available in only six countries (35%). There was also great variation in the numbers (or proportions) of trained staff (Table 1 ). Indonesia reported it trained 100% of staff at all levels, but this was based on its target for 2002; Viet Nam reported high levels of training at health facility and laboratory levels and South Africa reported high levels of training (70% to 80%) at provincial, district and health facility levels (although absolute numbers were missing). Other countries either reported very low levels of training or had missing information. Only six countries (35%) reported accurate information on training of laboratory staff. Information on length of training courses and development of training materials was also incomplete. Length of courses for staff at the same service level varied greatly between countries; for example, courses for TB coordinators at provincial level (data from 11 countries) varied from 4 to up to 60 days, whereas at district level (information from 10 countries) courses varied from 4 days to 4 months (Table 2 ). Similar variations were reported at other levels. Development of training materials was reported by eight (47%); a further four (23%) reported developing lectures and exercises modules for training at all levels, whereas only three (18%) described using participatory educational methods. Nine countries (53%) had not developed specific training modules; five of them used WHO/IUATLD materials and the remaining countries used programme implementation manuals and guidelines as training tools. Altogether eight countries (47%) reported using WHO/IUATLD training materials. The WHO/STB training modules for health facility staff take normally five days to complete if all its tasks and units are included; only one of the countries that reported using it had a five-day course. Training at this level took three days or less in six of the 12 countries with available information. Table 2 Length of training courses at different training levels and development of training materials by country Country Length of training in days TB coordinator Staff health facility level Laboratory staff Training materials* Provincial level District level Afghanistan 10 - - 7 No materials developed Bangladesh - 4–6 2–3 6 Developed lectures and exercises, also uses WHO materials for training at provincial and district levels Brazil 5 5 5 5 Developed manuals & guidelines for all levels Cambodia - 5 3 - Developed training modules at all levels DR Congo 21 - - - No materials developed, uses WHO materials at all levels Ethiopia - 7–10 5 5 Ad hoc handouts Indonesia 12 12 6 8 Developed training modules for all levels Kenya 15 - 2 3 No specific training materials developed, uses national guidelines and WHO & IUATLD training materials Myanmar 5 5 1 5 TB manual for health facility staff and lab. technicians developed in 2002 Nigeria 21 120 3 6 Developed training materials at all levels, did not specify Pakistan - 10 variable § 10 Developed training modules, translated WHO materials for training lab. staff Philippines - - - - Use modified WHO materials at all levels Russian Fed. - - - - WHO materials were developed for Russia in 2002, use manuals and guidelines for staff at different levels South Africa 4 - 3 - No specific materials developed, uses WHO & IUATLD materials for different levels UR Tanzania 14 30 5 5 Developed a TB Manual but no specific training materials Uganda 15 - - - Developed training materials but did not specify Viet Nam 10 or 60 5 or 10 3 or 5 15–20 Uses national guidelines for training, for health facility staff it uses WHO & IUATLD training materials * Development of specific training materials for each country or use of standard WHO / IUATLD modules and courses. §Variable: Medical staff 6 days, paramedical 3 days, health workers 1 day. Performance The estimated time needed to treat a new SS+ patient is shown in Figure 1 . Times were positively skewed and ranged from 5 to 36 (mode 10) hours. Eight countries (47%) needed from 9 to 12 hours; Brazil was the only country requiring less than 9 hours; Pakistan, Uganda, Nigeria, Kenya, Russia and Philippines were out-layers requiring between 22 to 36 hours. The Russian Federation (31 hours) and the Philippines (36 hours) took the longest, due to the policy to hospitalize all patients during the intensive phase of treatment in the former and the tendency to perform strict DOT during the continuation phase in the latter. Figure 1 Average time spent to treat one new sputum-smear positive TB patient Estimated HR gap Sixteen countries (Table 3 ) reported information on estimated shortages of staff at peripheral level (health facility) at current or target CDRs. Two countries (12%) reported shortages of staff at current CDR, whereas five countries (29%) predicted health facility staff shortages at the 70% CDR. Estimated numbers of staff needed varied from 1009 more nurses in Afghanistan to 8981 health care workers in Myanmar; Uganda reported shortages of staff but did not estimate the numbers needed. Eleven countries (65%) reported no shortages of health facility staff at current or at 70% CDRs; these included countries with low current CDR and no data on staff numbers. Table 3 Perceived staff needs at different service levels. NTP managers were asked to report perceived staff needs at different service levels. Staff needs at Health Facility level were evaluated both at current and at the 70% target case detection rates Country CDR* Perceived staff shortages Comments Health facility District Provincial Central At current CDR At 70% CDR Afghanistan 19% a No 1009 No No 17 Poor distribution of staff, staff needed to run new facilities Bangladesh 33% No No No No 1 – 5 Brazil 84% No No Yes Capacity building of existing staff is a priority Cambodia 52% a No No No No No Poor distribution and training of staff at district and central levels DR Congo 52% a No No No 52 >5 Capacity building of existing staff a priority Ethiopia 33% a Yes 5 (regional) 2 Lack of data, poor distribution and training Indonesia 30% a No 3670 Yes Yes Yes Kenya 49% a No No No No 8 Poor distribution of existing staff Myanmar 73% a Yes 8981 6 4 4 Nigeria 14% No No No 111 25 Capacity building of existing staff is a priority Pakistan 13% No 2981 Yes Yes >6 Philippines 58% a No No >3 Yes >10 Russian Fed. § 34% No No No No Approx. 48 Not enough data available South Africa 97% No No Yes Yes Yes Lack of funds for recruiting new staff UR Tanzania 43% a No No >364 >88 >11 Staff retention and deployment problems Uganda 47% a Yes Yes Yes (Zonal) 4 Inconsistent data provided Viet Nam 82% a No No No No No Poor training of existing staff at district and provincial levels * Case Detection Rate for 2002, Ref 3. a No data available for the whole country; case detection rate for DOTS programmes §Federation All countries provided information on their estimated HR gap at district level; nine (53%) countries reported existing staff needs at district level (Table 3 ); nine countries reported existing needs at provincial level and all but two countries reported staff needs at central level. Some countries answered qualitatively while others estimated numbers needed, but it is not clear how these estimates were reached. Cambodia and Viet Nam were the only HBCs reporting no staff shortages at any service level. Afghanistan, Cambodia, Ethiopia, Kenya and Uganda reported poor distribution of staff. Brazil, Cambodia, Democratic Republic of the Congo, Ethiopia, Nigeria and Viet Nam reported training and capacity-building shortfalls. Discussion In many HBCs, NTP managers do not have access to accurate information on numbers, types and distribution of staff involved in TB control activities. As previously stated, TB control programmes are partially or fully integrated into health care systems; this lack of HR information thus could at best indicate poor communication and coordination between NTP managers and HR planners or, at worst, a general lack of information on HR in the health care system. In order to adequately manage and plan HR it is important to have up-to-date information on the quantity, distribution and skills of the existing health care workforce [ 16 ]. It is therefore imperative to assist HBCs to develop and maintain appropriate HR databases so that the necessary information for planning and managing their health care workforce can be readily accessible. Accreditation is often used as a proxy for competence for professional groups (doctors and nurses); for other health care staff, adequate competence is ensured through regular training and supervision. Apart from Indonesia and Viet Nam, there was no correlation between information on numbers of staff, attendance of training courses, length of courses and development of training materials. In general NTP managers in HBCs had limited information on staff attendance of training courses and on the characteristics, duration and intensity of training activities. HBCs need to develop needs-based comprehensive training policies and training strategies for health care staff at all service levels, incorporating pre-service training, re-training, in-service support and continuous professional/career development. Similarly, it is important to emphasize the role of periodic monitoring and supervisory visits as part of the staff continuous education and support processes. Appropriately designed strategies fostering career paths could increase staff motivation and performance; improve staff recruitment, retention and distribution; and even have a positive effect on enlisting of students into training programmes [ 17 ] and expansion of the health care workforce. The quality of training materials was also of concern: many countries used inadequate training tools, few had developed specific modules and even fewer used problem-based learning or participatory methodology to facilitate adequate skill development. The length of courses also varied greatly: while very short courses do not allow the development of skills and competences needed to improve performance, lengthy courses have economic and logistic implications. Differences in the quality and intensity of training could translate into service or performance variations with detrimental effects to programmes and ultimately to patients. There is a need to standardize training in terms of contents, competences, methodologies, course duration and quality of training materials. Improvement and standardization of training curricula and methodologies will facilitate the adoption of a universal standard of care for TB patients in HBCs, leading to a more rational use of the health care workforce, improved staff motivation and productivity and better outcomes for TB patients. Although the majority of countries described deficiencies of staff at central level, few reported shortages in the actual numbers of posts. However, there were concerns about the distribution and/or the skill mix (competences and efficiency) of staff. There was no clear relation between reported data on staff needs and either the TB burden in the country or the actual performance of the NTP. It is of concern that only Uganda and Myanmar reported shortages of staff at current CDRs, whereas countries with very low current CDRs did not. Only five countries projected shortages of staff at the 70% CDR; four of them estimated large numbers of staff required. If the assessment of the HR needs was based on the actual workload, it is possible no substantial gap was determined because most of the current diagnostic and treatment tasks are carried out satisfactorily by existing staff, but an unrecognized HR gap could hamper improvement in case detection rates. The inability of some countries to increase the number of posts because of recruitment ceilings imposed under structural adjustment programmes could lead to countries underreporting their HR gap. Further analysis and recommendations on staff needs at different levels is hindered by the lack of accurate information on numbers, quality and distribution of existing staff. This survey is, to our knowledge, the first attempt to ascertain HR for TB control in HBCs. The response rate was high and responses were analysed in the context of the countries' estimated TB burden and current CDRs. Limitations of the study are evidenced by the great variation in reported staff numbers and the lack of correlation between information on numbers of trained staff, length of courses, development of materials and performance, suggesting a shortage of valid information, poor standardization and/or data reliability. The fact that the questionnaire was in English could have generated some inaccurate answers. Indonesia, for example, reported 100% training completion based on its 2002 training goals rather than in absolute numbers (the actual proportion of staff trained at health centre level was 35%), but countries were re-contacted (by e-mail or telephone) to clarify inconsistent replies. Responses could have been biased towards greater HR needs if countries perceived this as an opportunity to request increased support; however, this did not seem to happen, since few countries reported staff shortages. The performance assessment component of the questionnaire assumed NTP managers could assess the HR gap by comparing existing staff numbers (at different service levels) with the product of the additional SS+ cases by the time spent in diagnostic/treatment activities. Although all the tasks were listed in the questionnaire, the interpretation of what constituted achieving each task was left to NTP managers; this could have affected their estimation of the time required. We are currently developing a more detailed task analysis tool that includes the breakdown of each TB control activity into specific single tasks and a description of what constitutes each task; this will facilitate future studies. Furthermore, this method accounted only for new SS+ cases: in some countries, health facility staff will spend more time dealing with relapses, treatment failures or other TB patients: On the other hand, in some HBCs some tasks are performed by NGOs or the private sector and this was not discounted. Constraints to HR development such as poor HR data quality, lack of a comprehensive national HR plan/strategy and little attention to continuing education programmes [ 18 ] were common to HBCs. This study evidences the huge gap in HR data (more apparent in large countries such as China and India) and the variability in quality and validity of available information. This is not an isolated problem of TB control programmes, given the dearth of published data on HRH in LMICs. Decentralization, ongoing in many HBCs, could have contributed to worsening HR information at central level. Developing HR planning and management capacity at district level and generating HR information systems should accompany decentralization processes [ 19 - 21 ] so that decentralization does not result in decentralized chaos [ 19 ]. A comprehensive HRH database facilitates HR surveillance and the management and planning of HR development in the health system [ 20 ]. The 2nd ad hoc committee on the TB epidemic recommended addressing the health-workforce crisis by collaboratively developing policies to reduce barriers to creating and filling posts in HBCs; increasing staff recruitment and retention by improving working conditions in the health sector; promoting task analysis and HR needs assessments, HR planning and training at country level; and working together with other stakeholders to develop strategies to further mobilize HR for TB control [ 22 ]. In general there is poor communication between HR planning units in the MoH and other technical programmes. A starting point for many HBCs will be a rigorous appraisal of their current HRH through in-depth assessments. Calculating programme-specific HR requirements, and informing the HR planners in the MoH, are two steps that are not consistently performed, resulting in a health care workforce often unaware of its own capacities and limitations. Conclusion (See Table 4 .) There is urgent need to assist HBCs developing HR information systems so that the up-to-date information required for appropriate planning, managing and supporting their health care workforce can be available. NTP managers in HBCs were generally aware of the need for appropriately trained staff at different service levels and, in some countries, the need to redistribute existing staff. Many HBCs require support in developing HR planning and management capabilities; however, more information is needed from countries to understand what factors most influence HR capacity so that country-specific plans can be developed. Table 4 Conclusions 1. A paradigm shift in our approach to HR is needed. The HR impact of health initiatives must be conveyed in an explicit, open and unambiguous way so that governments, planners, and financial and technical partners will have a clearer understanding of the urgency of the HR crisis and will have to take a stand on addressing it. 2. HR information systems in HBCs must be developed/strengthened. Without some reasonably accurate information on the numbers, location, qualifications and skills of staff it is impossible to administer, manage or plan the health workforce in any effective manner. 3. There is a dearth of information on HR for disease control programmes in LIMC. There is a need to develop HR assessment tools allowing for the different disciplines involved in HR issues and to conduct in-depth studies using validated methodology. 4. It is important to improve the communication link between technical programmes and HR planning at central level. There is a need to support some HBCs in developing HR management skills and in capacitating personnel in the area of HR management and planning. 5. Training is an important component of HRH development; there is a need to identify the minimum requirements of training at different service levels required to obtain a universal standard of care for TB patients and to better standardize training materials, methodologies and courses. 6. There must be a twin-track approach to addressing the HRH crisis. Current shortages must be addressed with short-term interventions in line with medium/long-term solutions developed within the context of poverty reduction strategies and national medium-term expenditure frameworks. Human resources constraints in TB will not be solved by NTPs in isolation; they are and will remain a subset of the general health workforce. The degree of integration of the HR for TB within the health workforce will vary from partial to total, depending on local conditions. The same holds true for the supply side: different health programmes compete for finite HR, posing a strain on local health systems. The health workforce crisis for TB control must be addressed within the broader HR context. Without creative solutions there will not be enough trained health professionals to implement the strategies proposed by the priority disease control programmes. Finally, a paradigm shift in the way we approach HR issues is needed. Up until now, the HR implications of public health interventions have been tacitly understood; the impact of healthcare initiatives on the limited health workforce has always been implicit, unspoken and often underrated. There is now a need to clearly convey the HR impact of health initiatives in an open, unambiguous manner. By endeavouring to make the HR implications of each existing and new public health intervention explicit, we hope governments, planners, and financial and technical partners will have a clearer understanding of the urgency of the HR situation and will have to take a stand on addressing the health care workforce crisis. Conflict of interests The author(s) declare that they have no competing interests.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554980.xml
552329
Recombinant Tula hantavirus shows reduced fitness but is able to survive in the presence of a parental virus: analysis of consecutive passages in a cell culture
Tula hantavirus carrying recombinant S RNA segment (recTULV) grew in a cell culture to the same titers as the original cell adapted variant but presented no real match to the parental virus. Our data showed that the lower competitiveness of recTULV could not be increased by pre-passaging in the cell culture. Nevertheless, the recombinant virus was able to survive in the presence of the parental virus during five consecutive passages. The observed survival time seems to be sufficient for transmission of newly formed recombinant hantaviruses in nature.
Background Recombination in RNA viruses serves two main purposes: (i) it generates and spreads advantageous genetic combinations; and (ii) it counters the deleterious effect of mutations that, due to the low fidelity of viral RNA polymerases and lack of proofreading, occur with high frequency [ 1 ]. The purging function is, naturally, attributed to the homologous recombination (HRec), i.e. recombination between homologous parental molecules through crossover at homologous sites. HRec was first described for the positive-sense RNA viruses [ 2 , 3 ] and subsequent studies lead to the widely accepted copy-choice model [ 4 ]. HRec was later shown to occur in rotaviruses thus adding double-stranded RNA viruses to the list of viruses capable of recombination [ 5 ]. Negative-sense RNA viruses that occupy the largest domain in the virus kingdom until recently were known to undergo non-homologous recombination only, forming either defective genomes, like polymerase "mosaics" of influenza A virus DI-particles [ 6 ] and "copy-backs" of parainfluenza virus [ 7 ] or hybrids between viral and cellular genes [ 8 ] or between different viral genes [ 9 ]. The first evidence for HRec in a negative-sense RNA virus has been obtained on hantaviruses [ 10 , 11 ]. Hantaviruses (genus Hantavirus , family Bunyaviridae ) have a tripartite genome comprising the L segment encoding the RNA-polymerase, the M segment encoding two external glycoproteins, and the S segment encoding the nucleocapsid (N) protein [ 12 ]. Hantaviruses are maintained in nature in persistently infected rodents, each hantavirus type being predominantly associated with a distinct rodent host species [ 13 ]. When transmitted to humans, some hantaviruses cause hemorrhagic fever with renal syndrome or hantavirus pulmonary syndrome, whereas other hantaviruses are apathogenic [ 14 , 15 ]. Persistent infection in natural hosts allows for the simultaneous presence of more than one genetically distinct hantavirus variant in the same rodent. This may result in hantavirus genome reassortment [ 16 , 17 ] or recombination, as proposed in the above-mentioned study of Sibold et al [ 10 ] who showed a mosaic-like structure of the S RNA segment and the N protein of Tula hantavirus (TULV). Most recently, we have shown transfection-mediated rescue of TULV with recombinant S segment, in which nt 1–332 originate from the cell culture isolate Moravia/Ma5302V/94 (or TULV02, for short) [ 18 ], nt 369–1853 originate from the strain Tula/Ma23/87 [ 19 ], and nt 333–368, that are identical in both variants, can be of either origin. Both M and L segments of the recombinant virus (recTULV) originate from TULV02 [ 11 ]. RecTULV was functionally competent but less competitive than TULV02. One reason for the observed lower fitness of the recTULV might be that it was generated in the presence of the wt variant, with which it has to compete, and thus not given enough time to to establish a well balanced, mature quasi-species population. We, therefore, decided to compare fitness of TULV02 with that of recTULV that underwent several passages in cell culture. Results and discussion First, we designed RT-PCR primers able to discriminate between non-recombinant (V-type) and recombinant (REC-type) types of TULV S RNA. The resullts presented in Fig. 1 show that the primer pairs designed to generate the 118 bp- long products from either V-type or REC-type S RNA amplified, indeed, homologous sequences only, whether these were taken along (lines 1 and 6) or mixed with the heterologous sequences (lines 3 and 7). Using the two specific RT-PCR conditions, the presence of V-type and REC-type S RNA was monitored on ten sequential passages of the mixture of TULV02 and RecTULV5 variants (Fig. 2 ). S RNA of V-type was seen on all passages (Fig. 2A , lines 1–10). In contrast, S RNA of REC-type, was detected up to the fifth passage (Fig. 2B , lines 1–5), and then disappeared (Fig. 2B , lines 6–10). An alternative approach to check the presence of the two different types of S RNA using specific primer pairs at the stage of nested PCR gave exactly the same result. The V-type S RNA was detected during all ten passages while the REC-type totally disappeared after the 5 th passage (data not shown). These data confirmed our earlier observation [ 11 ] that the transfection-mediated HRec yields functionally competent and stable virus, recTULV. The purified and pre-passaged recombinant virus, however, presented no real match to the original cell adapted variant, TUL02, it terms of fitness. Taking into account that the in situ formed recombinant S RNA disappeared from the mixture after four passages [ 11 ], one should conclude that the lower competitiveness of the recombinant virus seen earlier did not result from its "immature" status. When, under similar experimental settings, TUL02 has been passaging in the presence of another isolate, TULV/Lodz, none of the two viruses was able to establish a dominance during ten consecutive passages (Plyusnin et al., unpublished data). Figure 1 Checking of specificity of RT-PCRs for the wt and the recombinant S RNA segments. Lines 1–3: products of RT-PCR with primers VF738 and VR855 on RNA from cells infected with TULV02 (line 1), on RNA from cells infected with the recTULV (line 2) and on the mechanical mixture of both RNA preparations (line 3). Lines 5–7: the corresponding products of RT-PCR with primers RECF738 and RECR855. Lines 4 and 8 show negative controls. M, molecular weight marker; bands of 147 and 110 bp are indicated by arrows. Figure 2 Monitoring of wt and recS-RNA during sequential passages of the mixture of TUL02 and recTULV. A. PCR-amplicons (118 bp), obtained in RT- PCR with the primers VF738 and VR855 (specific for the wt virus) on RNA from infected cells collected on passages 1 to 10. B. PCR-amplicons (118 bp), obtained in RT- PCR with the primers RECF738 and RECR855 (specific for the recombinant virus) on RNA from infected cells collected on passages 1 to 10. NC, negative controls. M, molecular weight markers; bands of 147 and 110 bp are indicated by arrows. Although relatively short, the observed survival time of the recTULV in the presence of the original variant TUL02 seems to be sufficient for transmission of a recombinant virus, in a hypothetical in vivo situation, from one rodent to another. If transmission is performed in a sampling-like fashion – and this seems to be the case for hantaviruses [ 13 ] – the recombinant would have fair chances to survive. The existence of wt recombinant strains of TULV [ 10 ] supports this way of reasoning. Evidence for the recombination in the hantavirus evolution continues to accumulate [ 20 , 21 ]. The genetic swarm of S RNA molecules from the recTULV is represented almost exclusively by the variant with a single break point located between nt332 and nt368. The proportion of the dominant variant is larger in the passaged recTULV (13 of 14 cDNA clones analyzed, or 93%) than in the freshly formed mixture of recS RNAs (12 of 20 cDNA clones, or 60%) [ 11 ]. Thus, recTULV already represents a product of a micro-evolutionary play, in which the best-fit variant has been selected from the initial mixture of recS RNA. Whether this resulted from higher frequency of recombination through the "hot-spot" located between nt332 and nt368 or from the swift elimination of all other products of random recombination due to their lower fitness (the situation reported for polio- and coronaviruses [ 22 , 23 ]), or both, remains unclear. We favor the first explanation as the modeling of the S RNA folding suggests formation of a relatively long hairpin-like structure within the recombination "hot-spot" (Fig. 3 ). Secondary structure elements of this kind, which might present obstacles for sliding of the viral RNA polymerase along the template, were suggested as promoters for the template-switching in the early studies on polioviruses [ 22 ] and considered a crucial prerequisite for recombination [ 25 , 24 ]. The hairpin in TULV plus-sense S RNA (Fig. 3 ) is formed by the almost perfect inverted repeat that includes nt 344 to 374. In the minus-sense RNA, the structure is slightly weaker due to the fact that two non-canonical G:U base pairs presented in the plus-sense RNA occur as non-pairing C/A bases in the minus-sense RNA. Interestingly, in Puumala hantavirus, a hairpin-like structure formed by a highly conserved inverted repeat in the 3'-noncoding region of the S segment seems to be involved in recombination events, leading, however, to the deletion of the hairpin-forming sequences (A. Plyusnin, unpublished observations). The role of RNA folding in hantavirus recombination awaits further investigation. Figure 3 Hairpin-like structures predicted for the recombination "hot-spot" in the plus- and minus- sense S RNA of TULV. Conclusion The data presented in this paper show that the recTULV presents no real match to the original cell adapted variant and that the lower fitness of the recombinant virus can not be increased by pre-passaging in cell culture. The observed survival time of the recTULV in the presence of the parental virus seems to be sufficient for transmission of newly formed recombinant hantaviruses in nature. Methods Recombinant TULV RecTULV (clone 5) was purified from the mixture it formed with the original variant, TULV02, using two consequent passages under terminal dilutions [ 11 ]. After the purification, recTULV underwent three more passages, performed under standard conditions, i.e. without dilution. The presence of recS-RNA on the passages was monitored by RT-PCR and the isolate appeared to have a stable genotype (data not shown). RecTULV formed foci similar in size to those of the original variant and grew to the titers 5 × 10 3 – 10 4 FFU/ml. Competition experiments Vero E6 cells (5 × 10 6 cells) were infected with the 1:1 mixture of recTULV and TULV02, approximately 10 4 FFU altogether. After 7–12 days the supernatant (~20 ml) was collected and RNA was extracted from the cells with TriPure™ isolation reagent, Boehringer Mannheim. Aliquots (2 ml) of the supernatant were used to infect fresh cells; the rest was kept at -70°C. The following nine passages were performed in the same way. Reverse transcription (RT), polymerase chain reaction (PCR) and sequencing RT was performed with MuLV reverse transcriptase (New England Biolabs); for PCR, AmpliTaq DNA polymerase (Perkin Elmer, Roche Molecular Systems) was used. To monitor the presence of TULV S RNA on passages, RT-PCR was performed with primers VF738 (5'GCCTGAAAAGATTGAGGAGTTCC3'; nt 738–760) and VR855 (5'TTCACGTCCTAAAAGGTAAGCATCA3'; nt 831–855). To monitor the presence of recTULV S RNA, RT-PCR was performed with primers RECF738 (5'GCCAGAGAAGATTGAGGCATTTC3'; nt 738–760) and RECR855 (5'TTCTCTCCCAATTAGGTAAGCATCA3'; nt 831–855). All four primers were perfect matches to the homologous sequences; to the heterologous sequences, the forward primers have five mismatches while the reverse primers have six. Alternatively, complete S segment sequences of both variants of TULV were amplified using a single universal primer [ 19 ] and then either of the two pairs of primers was used in nested PCR. Authenticity of the PCR amplicons was confirmed by direct sequencing using the ABI PRISM Dye Terminator Sequencing kit (Perkin Elmer Applied Biosystems Division). Competing interests The author(s) declare that they have no competing interests. Authors' contributions AngP participated in the design of the study, carried out the experiments and helped to draft the manuscript. AlexP participated in the design of the study and drafted the manuscript. Both authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552329.xml
551608
parkin mutation dosage and the phenomenon of anticipation: a molecular genetic study of familial parkinsonism
Background parkin mutations are a common cause of parkinsonism. Possessing two parkin mutations leads to early-onset parkinsonism, while having one mutation may predispose to late-onset disease. This dosage pattern suggests that some parkin families should exhibit intergenerational variation in age at onset resembling anticipation. A subset of familial PD exhibits anticipation, the cause of which is unknown. The aim of this study was to determine if anticipation was due to parkin mutation dosage. Methods We studied 19 kindreds that had early-onset parkinsonism in the offspring generation, late-onset parkinsonism in the parent generation, and ≥ 20 years of anticipation. We also studied 28 early-onset parkinsonism cases without anticipation. Patients were diagnosed by neurologists at a movement disorder clinic. parkin analysis included sequencing and dosage analysis of all 12 exons. Results Only one of 19 cases had compound parkin mutations, but contrary to our postulate, the affected relative with late-onset parkinsonism did not have a parkin mutation. In effect, none of the anticipation cases could be attributed to parkin . In contrast, 21% of early-onset parkinsonism patients without anticipation had parkin mutations. Conclusion Anticipation is not linked to parkin , and may signify a distinct disease entity.
Background Mutations in the parkin gene are a common cause of parkinsonism. parkin was originally discovered as the cause of autosomal recessive juvenile parkinsonism [ 1 ]. However, recent reports suggest that not all parkin mutations are recessive [ 2 - 5 ], nor is age at onset always early [ 6 - 9 ]. Several studies have found heterozygous mutations in patients with late onset parkinsonism, suggesting that a single parkin mutation predisposes to later disease onset [ 8 - 10 ]. Collectively, these reports imply that parkin may exert a dosage effect in which possession of two mutations (homozygous or compound heterozygous) leads to early-onset parkinsonism, while possession of one normal and one mutant parkin (heterozygous) increases the risk for late-onset parkinsonism. Several groups have reported what appears to be genetic anticipation in parkinsonism [ 11 - 13 ]. In the kindreds studied by them, the parent generation had typical late-onset parkinsonism and the individuals in the offspring generation developed the disease at much earlier ages. Barring the possibility that this pattern is an artifact of ascertainment bias (which cannot be ruled out until a biological mechanism is found), the intergenerational difference in onset age may be indicative of one of several mechanisms. The most common cause of anticipation is triplet repeat expansions. In a few parkinsonism families, disease has been shown to segregate with pathogenic expansions in SCA loci, but the search for other expansion loci in familial parkinsonism has been unsuccessful [ 14 - 16 ]. Other mechanisms that could result in anticipation include: change in mitochondrial heteroplasmy, which can affect disease severity and age at onset; a modifier gene that segregates independently of the disease gene; and parent/child exposure to a toxin. Gene dosage for dominant mutations can also mimic anticipation. For example, Familial Hypercholestrolemia is an autosomal dominant disorder in which heterozygotes develop the disease after the 4 th decade of life, whereas homozygotes show symptoms at much younger ages, sometimes at birth (OMIM *143890). Is anticipation in parkinsonism related to parkin mutation dosage? We hypothesized that some parkin mutations are dominant: heterozygotes have incomplete penetrance and may develop late-onset parkinsonism, whereas homozygotes and compound heterozygotes have accelerated disease leading to early-onset parkinsonism. This unifying hypothesis was attractive because if true, it could explain the reported variations in mode of inheritance of parkin (recessive in some families, dominant in others), the range in age at onset from juvenile to late onset, and the significantly earlier onset in some of the children of affected parents. To test this hypothesis, we studied parkin in 19 kindreds with early-onset parkinsonism in the index generation, late-onset parkinsonism in the parent generation, and exhibited ≥ 20 years of anticipation. Methods Families were recruited from a movement disorder clinic. The probands and affected family members had the clinical diagnosis of idiopathic Parkinson's disease (here referred to as parkinsonism due to lack of autopsy confirmation). Diagnosis was made by a neurologist according to the British Parkinson's Disease Brain Bank criteria except that family history was not an exclusion criterion [ 17 ]. These families were identified through an index case (a clinic patient) who reported a family history of parkinsonism. Positive family history was defined as patient reporting a first or second degree relative with parkinsonism, although we did inquire about more distant relatives as well. Clinic patients were enrolled sequentially and their affected relatives were subsequently identified and enrolled. For affected relatives we obtained medical records or personally examined them when possible. In some families the index case had early onset and the relative had late onset, in others, the index case had late onset and the relative with early onset was subsequently identified. The study was approved by the Institutional Review Boards at the participating institutions. DNA was extracted from blood using standard protocol. Genotyping was blind to phenotype. To identify point mutations, we sequenced both DNA strands of all 12 exons. Exons and 50–100 bp of flanking intronic sequences were PCR-amplified [ 1 ], agarose gel-purified (Gene-clean III, Bio101), and directly sequenced by dye-terminator cycle sequencing (ABI, Big-Dye) using an ABI377 sequencer. To identify exon deletions and duplications, we analyzed gene dosage using real-time fluorescence-based PCR (ABI 7700 Sequence Detector). Amplification of subject genomic DNA was performed using fluorescently labeled probes (5' FAM or VIC, 3' TAMRA) and Taqman Universal PCR Mix (ABI) [ 4 , 18 ]. parkin exon amplifications were multiplexed under standard conditions with an 84-bp fragment of a single-copy human β-actin gene (Genbank accession number XM_004814) as an internal control. A standard curve was generated for each parkin exon and for β-actin using 0, 5, 15, 55 and 220 ng of control human genomic DNA. The number of PCR cycles required before the ABI 7700 detects each parkin exon product (CT value) was plotted against the corresponding exon standard curve, thus calculating the relative parkin copy number. The copy number for each exon was normalized to the single-copy actin gene within each multiplexed reaction and to a normal control reference individual, allowing an estimate of the number of copies of parkin . Optimal threshold levels for each primer set were maintained between plate analyses. All samples were analyzed in triplicate. Results Nineteen kindreds were chosen for parkin analysis based on having parkinsonism in two consecutive generations, late-onset parkinsonism in the parent generation (onset ages 59 to 89 years, mean 71.0 ± 8.5), early-onset parkinsonism in the offspring generation (onset ages 8 to 47 years, mean 37.2 ± 9.8), and ≥ 20 years of anticipation as measured by the difference in mean ages at onset in two generations. In the younger generation, 2 probands had onset before age 20, 1 was in his twenties, and 16 were over 30 years old. It was hypothesized that the affected individuals with early-onset parkinsonism are homozygous or compound heterozygous, and the parents and other relatives with late-onset parkinsonism are heterozygous. For parkin analysis, we began with the family member from the generation with the earlier onset, so as to enrich for parkin mutations. We expected the majority of the index cases to be compound heterozygous. At the minimum, 16–49% of index cases should have had mutations, since this is the range reported for early-onset sporadic and familial parkinsonism [ 7 , 9 , 19 ]. However, only one case had parkin mutations. This patient had compound deletions in exon 3 with onset at age 8. Her sister, onset at age 15, also had the compound mutation. The relative with late-onset was an uncle with onset age of 64. Contrary to our hypothesis, the uncle did not have a parkin mutation. The cause of late-onset parkinsonism in the uncle was different from the nieces. In effect, we did not find any cases where we could attribute the intergenerational difference in age at onset to parkin dosage. We also analyzed parkin in 28 additional patients with early-onset parkinsonism (onset age 14 to 40 years, mean 32.8 ± 7.0), either with (n = 4) or without (n = 24) family history, but with no evidence of anticipation. In this group, 3 probands had onset at or before age 20, 5 were in their twenties, and 20 were over 30 years old. Nine mutations were found in six individuals. Two subjects were compound heterozygous (onset ages 31 and 37) and four were heterozygous (onset ages 14, 25, 37, 37). The frequency of parkin carriers in early-onset parkinsonism without anticipation was 21% which is in the range 16% – 49% reported in the literature (the lower range represents population and clinic based studies similar to ours, while the higher frequencies were found in highly selected autosomal recessive families). It was of interest to determine if any of the parents of the three early-onset parkinsonism patients with compound mutations had developed late-onset parkinsonism. All six parents were heterozygous (5 were confirmed by genotyping, one was inferred), and all had remained free of parkinsonism to the ages of 53, 60, 63, 74, 76 and 78 yrs. These individuals may still develop parkinsonism. Discussion Familial parkinsonism with anticipation may be more common than classical dominant and recessive subtypes combined. In our clinic population, among 487 patients studied, 145 had a first or second degree relative with PD; that is 30% which is in line with the published figures for other referral clinics. Among the 145 familial cases, 110 had parkinsonism in consecutive generations, which is compatible with autosomal dominant inheritance, and 35 had an affected sibling or cousin, which is suggestive of recessive inheritance. However, among the 110, only 26 were compatible with a classical dominant pattern (i.e., <10 years intergenerational variation in onset); while 63 exhibited 10–68 years of anticipation, and 7 had 10–17 years of reverse anticipation (14/110 had unknown onset ages). Our clinic is a referral center, which explains the relatively high proportions of early-onset and familial cases. The interesting finding was the relative proportions of autosomal dominant, autosomal recessive and anticipation cases within the familial subtype. Despite its relatively high prevalence, at least in our clinic, familial parkinsonism with anticipation has been largely overlooked in genetic research. A few parkinsonism families have been attributed to expansions in known SCA loci, but in the majority of the kindreds, the cause remains unknown. The more obvious possibilities are mitochondrial inheritance, modifier genes, parent-child exposure to environmental triggers, as yet unidentified triplet repeats or dominant genes with dosage effect, and in some cases, artifactual appearance of anticipation due to ascertainment bias. None of these are mutually exclusive; more than one may be true and operative in different families. The current parkin literature suggest that possessing two parkin mutations is fully penetrant and leads to early-onset parkinsonism, whereas having only one mutation may be incompletely penetrant and lead to later disease onset. While the causative link to early-onset parkinsonism is widely accepted, the association of parkin with late-onset parkinsonism remains controversial [ 9 , 10 , 20 , 21 ]. We postulated that, if parkin heterozygotes are at risk for late-onset disease, then some parkin families should exhibit intergenerational variation in age at onset resembling anticipation, where heterozygous parents develop late-onset parkinsonism and children who inherit two mutations develop early-onset parkinsonism. We hoped to explain the appearance of anticipation in relation to parkin dosage, but the findings do not support this postulate. While the results rule out a link between parkin and anticipation in these families, they do not negate the association of parkin with late-onset parkinsonism. Conclusion The phenomenon of anticipation is not due to parkin mutation dosage. The underlying mechanism for anticipation may be genetic or environmental. Identification and a-priori classification of pedigrees that exhibit significant intergenerational age at onset variation, as being distinct from families that display classical dominant pattern, may facilitate gene mapping studies by reducing heterogeneity. Anticipation in parkinsonism merits investigation in its own right, not only because it is a common phenomenon and may account for a large subset of familial parkinsonism, but it may also uncover a novel mechanism in parkinsonism. Competing interests The author(s) declare that they have no competing interests. Authors' contributions PP carried out the molecular genetic studies. LM interviewed the subjects and gathered family histories and medical records. JN performed the neurological examinations. GS participated in study design and supervised molecular genetic studies. JM performed molecular analysis of one large pedigree. HP conceived the study, participated in its design and coordination and drafted the manuscript. All authors read and approved the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551608.xml
549047
TPO/Mpl Studies in Agnogenic Myeloid Metaplasia
Background Agnogenic myeloid metaplasia (AMM) is one of the Philadelphia chromosome negative myeloproliferative disorder and is diagnosed by hyperplasia of atypical megakaryocytes, hepatosplenomegaly, extramedullary hematopoiesis and bone marrow fibrosis. Fibrosis is considered to be a secondary consequence of enhanced levels of fibrogenic growth factors such as TGF β1, bFGF and PDGF produced by enhanced numbers of megakaryocytes, while the primary cause is considered to be the enhanced proliferation of a defective stem cell. We have previously reported that thrombopoietin (TPO) is elevated in patients with AMM. Others have reported that Mpl protein is decreased in these patients. Since TPO is essential for the development of megakaryocytes, and Mpl protein is the receptor for TPO, we extended the study of TPO/Mpl to in vitro and in vivo cell culture systems to better understand the mechanism that leads to reduced Mpl protein in AMM patients. Results Plasma TPO levels were significantly elevated and Mpl protein levels were significantly reduced in AMM patients in concordance with previous studies. Platelet Mpl transcripts in AMM were however similar to those in controls. We also cloned Mpl cDNA from AMM patients and tested for their ability to make functional proteins in vitro and in the in vivo system of 293 T human embryonic kidney cells. Their expression including the glycosylated forms was similar to those from the controls. We also measured the level of translation initiation factor, eIF4E and found it to be increased in patients with AMM demonstrating that the reduced Mpl protein may not be due to translation defects. Conclusions Our studies using the in vitro and in vivo systems further confirm that reduced Mpl protein levels are not due to defects in its transcription/translation. Reduced Mpl protein could be due to its increased internalisation owing to enhanced plasma TPO or in vivo intrinsic defects in patients with AMM.
Background Agnogenic myeloid metaplasia (AMM), polycythemia vera (PV), essential thrombocythemia (ET) and chronic myelogenous leukemia (CML) are characterized as myeloproliferative disorders (MPDs) arising due to exponential amplification of a haematopoietic stem cell. A typical feature of these disorders is the presence of high number of circulating progenitor cells and their cytokine independent growth in culture. So far, the molecular basis for the disorder has been recognized only for CML and is attributed to the Bcr/Abl or Philadelphia (Ph) chromosome arising due to a translocation event. The aetiology for the other three disorders referred to as Ph negative MPDs is unknown. AMM, one of the Ph negative MPD is diagnosed by hyperplasia of atypical megakaryocytes, hepatosplenomegaly, extramedullary haematopoiesis and bone marrow fibrosis. Fibrosis is considered to be a secondary consequence of enhanced levels of fibrogenic growth factors such as TGF β1, bFGF and PDGF produced by enhanced numbers of megakaryocytes (MKs), while the primary cause is considered to be the enhanced proliferation of a defective stem cell. What leads to generation of such a defective cell with enhanced proliferation is unknown. We have earlier demonstrated [ 1 ], which has later been confirmed [ 2 ] that thrombopoietin (TPO) is one of the growth factors whose level is enhanced in patients with AMM. TPO is a haematopoietic growth factor that is essential for megakaryocytopoiesis and thrombocytopoiesis [ 3 ]. TPO binds to its receptor Mpl (Myeloproliferative leukaemia), gets internalised [ 4 ], and initiates a STAT5 signalling cascade [ 5 ], which results in thrombopoiesis. Studies with TPO knockout mice have shown reduced numbers of platelets and myeloid progenitor cells confirming the role of TPO. Enhanced expression of TPO in mouse by Retroviral/Adenoviral gene transfers and in mice transgenic for TPO , the animals developed extramedullary haematopoiesis and splenomegaly due to MK and granulocytic hyperplasia. The severity of the conditions was related to the expression levels of TPO , resulting in myelofibrosis under very high expression level [ 6 - 8 ]. These results of TPO over expression are similar to the symptoms of AMM and hence indicate that high levels of TPO found in AMM patients could be responsible for the clinical features of AMM. In the mouse knockouts for Mpl receptor, there is about 85% reduction in the number of platelets, MKs and other haematopoietic cell types [ 9 ] and Mpl over expression leads to myeloproliferation [ 10 ]. In vitro, Mpl antisense oligonucleotides inhibited megakaryocytopoiesis [ 11 ]. These reports augment the role of Mpl in megakaryopoiesis. But in contrast to TPO, Mpl protein level is reduced in AMM patients [ 12 ]. This abnormality in the TPO/Mpl pathway may depict the clinical features, but is it the cause for MK hyperplasia and the disease? Studies so far have not been conclusive. According to Taksin et al. [ 13 ] defects in the TPO/Mpl pathway are unlikely to be primarily responsible for AMM. Autonomously growing MKs in AMM did not have enhanced TPO transcripts and they were not inhibited by TPO antibody. Also no mutations were detected in the Mpl gene in these patients, to account for its reduced protein level. The observation of Li et al. [ 14 ] that Mpl anti-sense oligonucleotide treatment could inhibit the autonomous growth of MKs from ET/PV/AMM in vitro suggests that expression of Mpl is essential for MK autonomy in these disorders. However Moliterno et al. [ 12 , 15 ] have reported reduced Mpl protein level with glycosylation defects in the platelets of patients with these disorders. Although this could be responsible for the impaired TPO-mediated platelet protein tyrosine phosphorylation seen in PV and AMM patients [ 16 ], it does not confer with MK autonomy/hypersensitivity for growth in vitro. This reduced expression of Mpl in a disorder characterized by MK hyperplasia whose development and differentiation is dependent on this cytokine is paradoxical and needs further investigation. Hence to better understand the role of TPO/Mpl in AMM, we have extended the study of this growth factor and its receptor gene in these patients. Our studies reveal enhanced plasma TPO level, reduced endogenous level of Mpl protein in the platelets of patients with AMM and absence of any co-relation between TPO level and Mpl expression. We have cloned Mpl cDNA from 15 AMM patients and have found that it has full potential for expression similar to controls both in vitro and in the in vivo heterologous system of 293 T cells. We also assayed for the level of eIF4E, the translation initiation factor that under normal limiting level favours translation of strong mRNAs such as those of house keeping genes and at higher level, enhances the translation of weak mRNAs such as Mpl . AMM patients were found to have elevated eIF4E level. Results TPO ELISA Undiluted platelet poor plasma (PPP) samples from 20 AMM patients and 10 controls were assayed for their TPO levels. Although some patients exhibited low TPO levels, overall, TPO level was enhanced in AMM patients as compared to controls [0–684 pg/ml in AMM patients (with a mean of 229 pg/ml) as compared to 0–71 pg/ml in controls (with a mean of 22.8 pg/ml)]. This increase in TPO was significant (P < 0.001) as determined by t-test (Fig. 1 ). Minimum detectable dose with the TPO ELISA kit is 2.78–18.5 pg/ml according to the manufacturer. The Optical Density (OD) values for samples were within the range obtained with the highest standard (2000 pg/ml) used in the ELISA. Some samples from controls and patients as well failed to register any OD probably due to very low TPO levels. Figure 1 TPO is elevated in AMM patients. TPO level in the undiluted Platelet Poor Plasma samples of 10 controls and 20 AMM patients was tested by ELISA. While TPO level was in the range of 0–71 pg/ml in controls (Mean 22.8 pg/ml), it ranged between 0–684 pg/ml in AMM patients (Mean 229 pg/ml). This elevation of TPO level in patients was significant (P < 0.001) compared to controls. Mpl Expression is reduced in platelets of patients with AMM Platelets obtained from 12 AMM patients and 12 healthy controls were lysed in 1X Laemelli buffer and analysed on SDS PAGE. Mpl antibody that was raised against the full length Mpl protein was used for western blot analysis. This antibody recognizes both glycosylated and non-glycosylated forms of Mpl. The blots were re-probed with β Actin antibody for normalization. The western blots on chemiluminescence detection showed an overall reduction in the expression of Mpl in the platelets of patients with AMM and there was no co-relation (r 2 = 0.0038) between TPO level and Mpl levels (Fig. 2 ). While there was high expression of glycosylated form of Mpl in the controls, the expression of Mpl in the platelets of patients with AMM varied from none to having reduced amounts of both glycosylated and faster moving non-glycosylated forms. A representative blot with results from 4 controls and 7 patients is shown in Fig. 3 . The developed blots were analysed using the Image J program to quantitate the intensity of the bands and the values for Mpl expression after normalization against β Actin expression are represented in Fig. 4 . Statistical analysis of the values by t-test revealed the reduced Mpl expression to be significant (P < 0.05). Figure 2 TPO level is not co-related to platelet Mpl protein level. Plasma TPO levels and platelet Mpl protein levels from 12 AMM patients were compared for co-relation. Statistical analysis revealed absence of any co-relation between the levels of these two proteins (r 2 = 0.0038). Figure 3 Mpl protein is reduced in AMM patients. Platelet lysates from 12 controls and 12 AMM patients in 1X Laemelli buffer were analysed by 6% SDS PAGE. Mpl antibody raised against the full-length protein was used for western blot analysis. The blots were re-probed with β Actin Antibody for normalization. This representative autoradiogram shows the results for 4 controls and 7 AMM patients. Figure 4 Densitometric scanning results of Mpl expression in platelets. The intensity of the bands on the Mpl western blot autoradiograph was quantitated using Image J program. The values represent Mpl expression after normalization with β Actin expression assayed for 12 controls and 12 AMM patients. The reduction in Mpl levels in the platelets of AMM patients was statistically significant (P < 0.05). Mpl RNA level is similar in controls and AMM patients Since Mpl protein level was reduced in AMM patients, to know if it was due to reduced transcription of the Mpl gene, we compared the levels of Mpl RNA in AMM patients and controls by Real-Time RT-PCR. Total RNA from the platelets of 12 controls and 12 AMM patients were used as templates for PCR with gene specific primers. The values were normalized against GAPDH expression and are represented in Fig. 5 . On statistical analysis (t-test), the difference in Mpl expression was found not to be significant (P > 0.05), indicating absence of any transcriptional inhibition of Mpl in AMM patients. Figure 5 Transcription of Mpl is similar in AMM and controls. Total RNA isolated from platelets of 12 AMM patients and 12 controls were subjected to Real-Time RT-PCR using Mpl gene specific primers. The values for Mpl expression after normalization against GAPDH values were found not to be significant (P > 0.05) between the two groups. AMM patients have high eIF4E levels We assayed for the expression of eIF4E in the platelets of 12 AMM patients to see if reduced Mpl protein level could be due to its reduced translation owing to limiting levels of eIF4E, a translation initiation factor that could favour translation of weaker mRNA such as Mpl at higher amounts. At limiting levels it favours translation of stronger mRNA over the weaker mRNA. The western blots on chemiluminescence detection showed an overall increase in the expression of eIF4E in the platelets of patients with AMM. A representative blot shown in Fig. 6 includes 5 controls and 12 AMM patients (Platelet sample from Patient # 7 has inadequate protein). The developed blots were analysed using the Image J program to quantitate the intensity of the bands and the values for eIF4E expression after normalization against β Actin expression are represented in Fig. 7 . Statistical analysis of the values by t-test revealed the elevated eIF4E expression to be significant (P < 0.0001) and absence of any significant co-relation between the expression levels of Mpl and eIF4E (data not shown). Figure 6 AMM patients have high levels of eIF4E. Platelet lysates from 12 controls and 12 AMM patients in 1X Laemelli buffer were analysed by 10% SDS PAGE. Antibody for eIF4E was used at a dilution of 1:1000 for western blot analysis. The blots were re-probed with β Actin Antibody for normalization. A representative autoradiogram including 5 controls and all 12 patient samples is shown in this figure. Platelet sample from patient # 7 has inadequate protein. Figure 7 Densitometric scanning results of eIF4E expression in platelets. The intensity of the bands on the eIF4E western blot autoradiograph was quantitated using Image J program. The values represent eIF4E expression after normalization with β Actin expression for the 12 controls and 11 patient samples. The increase in eIF4E levels in the platelets of AMM patients was statistically significant (P < 0.0001). The coding region of Mpl from AMM patients has no mutations Mpl cDNA was amplified from the platelets of 15 AMM patients along with those from 15 normal controls by RT-PCR, cloned into the Blue Script vector and the recombinant clones sequenced completely. Base changes were detected in some patients and controls as well in the exons 2,3, 4, 6 and 12 but not at the same nucleotide (data not shown). The same changes were however not observed on re-cloning the cDNA of the samples with mutations and sequencing the corresponding genomic region. These base changes were therefore considered to be due to PCR errors and not genuine mutations. Hence Mpl gene abnormalities were not detected in AMM patients. Mpl from AMM patients is able to transcribe and translate in vitro About 1 μg of the Mpl cDNA of 15 controls and 15 AMM patients in the pcDNA3 vector were subjected to in vitro transcription/translation employing the rabbit reticulolysates to see if they were capable of translation in vitro. Mpl from AMM patients were able to translate as efficiently as the cDNA from controls revealing their full potential for translation. Also, in the presence of canine microsomal membranes, the translated protein was glycosylated similar to those from the controls. Results from 2 controls and 4 AMM patients are represented in Fig. 8 . The developed autoradiographs were analysed using the Image J program to quantitate the intensity of the bands and the values are represented in Fig. 9 . Statistical analysis of the values by t-test revealed absence of any difference (P > 0.05) between cDNA from controls and AMM either in the amount of the translation product or its glycosylation. Figure 8 Mpl from AMM patients can be translated in vitro similar to controls. Mpl cDNA clones of 15 controls and 15 Patients with AMM in pCDNA3 vector were subjected to in vitro transcription/translation using Promega's TNT Quick Coupled Transcription/Translation Systems with 35 S Methionine in the absence (-) or presence (+) of canine microsomal membranes. The entire reaction product was loaded onto 6% SDS PAGE and electrophoresed. The dried gel was exposed to X-ray film and a representative autoradiograph with 2 control and 4 patient samples is shown. Figure 9 Densitometric scanning results of Mpl in vitro transcription/translation. The intensity of the bands on the autoradiograph of Mpl in vitro transcription/translation was quantitated using Image J program. The values represent Mpl band intensity for 15 samples each of controls and patients. There was no significant (P > 0.05) difference between AMM and controls in the intensity of the non-glycosylated/glycosylated bands. Mpl from AMM patients is able to transcribe and translate in vivo Mpl cDNA was tested for in vivo expression in the 293 T human embryonic kidney cells. Equal amount of Mpl cDNA from 15 AMM/15 controls was transfected into the 293 T cells and Mpl expression analysed by western blot as described for the platelets. The expression level of Mpl appeared to be similar between controls and AMM patients. Fig. 10 is a representative blot depicting results of 5 controls and 5 AMM patients. Two bands were identified. The developed blots were analysed using the Image J program to quantitate the intensity of the bands and the values for Mpl expression are shown in Fig. 11 . Statistical analysis of the values by t-test showed no significant difference (P > 0.05) in the level of Mpl expression. Figure 10 Mpl from patients with AMM can be translated in vivo. Western blot analysis of lysates of 293 T cells transfected with pcDNA3/ Mpl constructs from 15 AMM patients and 15 controls. Transfection lysates were analysed by 6% SDS PAGE. Blots were probed with Mpl antibody raised against the full length Protein. A representative blot with data for 5 controls and 5 patients is shown in this figure. The second band probably corresponds to non-glycosylated Mpl. Figure 11 Densitometric scanning results of western blot of Mpl in vivo expression. The intensity of the bands for Mpl western blot with in vivo transfection lysates was quantitated using Image J program. No significant (P > 0.05) difference in the intensity of the Mpl bands between AMM and control's cDNA transfection lysates was observed. Discussion AMM is one of the MPDs for which the molecular defect is unknown. A characteristic feature of the disorder is the presence of atypical megakaryocytes in high numbers and their autonomy or hypersensitivity to growth factors in vitro. Since the haematopoietic growth factor TPO that is essential for the development of megakaryocytes is present at high level in AMM patients, it is of interest to study the role of TPO and its receptor in AMM. We have earlier reported that AMM patients have elevated plasma TPO level and this is not associated with its enhanced transcription in the bone marrow cells [ 1 , 17 ]. To determine if this enhanced TPO level is due to its receptor abnormalities, we have analysed 12 AMM patients for the presence of Mpl protein in their platelets and found it to be reduced in AMM relative to controls. However, this reduced level was not due to its reduced transcription since Real-Time RT-PCR revealed similar levels of transcripts from AMM patients and controls. To analyse if it could be due to its reduced translation owing to its G-C rich 5' UTR, we estimated the expression of the translation initiation factor eIF4E. The promoter of Mpl displays characteristics of weak mRNA that are not translated very efficiently due to G-C rich 5' UTR [ 18 ]. The translation initiation factor eIF4E, that is present at low level under normal conditions favours the translation of strong mRNA at these limiting levels. At higher levels, although the translation of strong mRNA is not increased greatly, the translation of weak mRNAs is highly enhanced and this is usually associated with hyper proliferation of cells as in cancers [ 19 ]. Hence we assayed for the expression of eIF4E in AMM to see if reduced Mpl protein level could be due to its reduced translation owing to limiting eIF4E level compared to controls. However eIF4E level was rather significantly higher than in controls as seen in other diseases with hyper cellular proliferation. Although this is not direct evidence, at these levels of eIF4E, normal translation of Mpl can be expected. Since there was no significant difference in the Mpl RNA level of AMM patients as shown by RT-PCR, we cloned and sequenced Mpl gene from these patients to over-rule any structural defects that could prevent translation. Although base changes were observed for some patients and controls as well, on sequencing the corresponding regions of the genomic DNA, the same base changes were not observed. Hence these were not considered to be genuine mutations. Taksin et al. [ 13 ] also have reported absence of Mpl gene mutations in AMM patients. Hence to date, there are no reports describing Mpl gene mutations in AMM patients although an activating mutation in this gene has been reported by Ding J et al. [ 20 ] in a case of familial ET, a rare hereditary MPD. Moliterno et al. [ 21 ] have recently reported a single base change polymorphism in the Mpl gene of African American patients with MPD that leads to substitution of the 39 th amino acid Lysine, with Asparagine in the extra cellular domain of the protein. Patients with this polymorphism exhibited high platelet counts and low platelet Mpl protein. African Americans were however not represented amongst our patients tested. We also performed in vitro transcription/translation with the cloned cDNA from the AMM patients. Similar levels of proteins were formed irrespective of the source of the cDNA, indicating that the cDNA from AMM patients had no structural defects preventing its translation. There was no difference in the extent of glycosylation of the protein either when the reaction was performed in the presence of microsomal membranes capable of post-translational modification in vitro. To further confirm this in vivo, these cDNAs were transfected into 293 T human embryonic kidney cells and the cell lysates analysed by western blots for Mpl expression. Irrespective of the source of the cDNA, similar level of expression of Mpl protein was observed. Two bands were however detected. Since the antibody used can react with both glycosylated and non-glycosylated forms of Mpl, it is likely that the second band corresponds to non-glycosylated form. As the in vivo system is saturated with the over-expressed protein, the system for modification of the protein could be exhausted. Overall, Mpl gene from AMM patients does not seem to have any impaired structural features to prevent its transcription/translation. Our present study shows elevated TPO level; reduced Mpl platelet protein level; absence of Mpl transcriptional inhibition and absence of any structural defects in Mpl coding region in AMM patients. However, there was no co-relation between TPO level and Mpl expression, similar to the observation of Harrison CN et al. [ 22 ] in ET patients. Plasma TPO level is regulated by platelet counts and MK mass [ 23 ]. TPO binds to its receptor Mpl on platelets/MKs and gets internalised. This receptor mediated endocytosis of TPO leads to its catabolism by the lysozymes, without any recycling of the receptors to the surface [ 24 ]. In AMM patients there is initial MK hyperplasia and thrombocytosis that leads to enhanced release of growth factors such as Platelet Derived Growth Factor, bFibroblast Growth Factor and Transforming Growth Factor β1 [ 25 - 27 ]. This may lead to increase in TPO transcription, since these growth factors have been shown to stimulate the transcription of TPO from the bone marrow stromal cells in cultures [ 28 , 29 ]. Our earlier observation that elevated plasma TPO level is not associated with its enhanced transcription in the bone marrow cells [ 17 ] could be due to the fact that our cultures predominantly consisted of fibroblasts and not stromal cells. Enhanced expression of TPO and absence of an accompanying increase in Mpl transcription or re-cycling of the internalised receptor may be responsible for persistence of high plasma TPO level in AMM patients. Reduced Mpl level seen in AMM patients may be due to exhaustion of the Mpl pool by binding to the enhanced TPO and its subsequent catabolism. This gains support from an earlier observation in thrombocythemic mice wherein very high and persistent level of TPO due to its induced over expression were found to be associated with reduced platelet Mpl protein without any reduction in Mpl transcription [ 30 ]. Also in other MPDs such as ET/PV, high plasma TPO level is associated with low Mpl protein level [ 16 , 31 ]. In vivo Mpl post-translational defect may also be responsible for impaired Mpl protein forms and levels in these patients as suggested by Moliterno et al. [ 15 ]. Cloning and sequencing of TPO 5' UTR from these patients may provide additional valuable information on elevated TPO levels since mutations in 5' UTR of TPO have been associated with elevated TPO levels without any increase in its transcription in patients with thrombocytosis [ 32 , 33 ]. Conclusions Our studies with TPO/Mpl show elevated TPO and reduced Mpl level in AMM patients, without any significant co-relation in their expression levels. Mpl cDNAs from AMM patients had full potential for transcription/translation including glycosylation, in vitro and in vivo, similar to those of normal controls. Reduced Mpl protein level may not be due to defects in its transcription/translation but could be due to increased internalisation or intrinsic defects in translation and glycosylation. Methods Patients Peripheral blood was obtained from patients with AMM after informed consent and peripheral blood was obtained from normal volunteers to serve as controls. Five volumes of peripheral blood was mixed with 1 volume of ACD buffer pH 4.5–5.5 (2.2% W/V Sodium Citrate; 0.8% Citric Acid; 2.2% Glucose; 50 ng/ml PGE-1) and centrifuged at 250 g for 10 minutes. The cell pellet was diluted 5X with Hank's buffer and subjected to Ficoll/Hypaque density gradient centrifugation. The ring of mononuclear cells was washed 2X with Hank's buffer and processed for isolation of genomic DNA using the Wizard Genomic DNA Purification kit from Promega, Madison, WI, following their instructions. The Platelet Rich Plasma (PRP) supernatant was centrifuged at 800 g for 10' to obtain the platelets. The Platelet Poor Plasma (PPP) was stored at -80°C, and the platelet pellet washed 2X with the Citrate buffer, pH 6.2 (0.05 mmol/L Sodium Citrate; 0.1 mol/L NaCl; 0.14 mol/L Glucose). They were then processed for RNA isolation using the Totally RNA isolation kit from Ambion, Texas, following their instructions or were suspended in 1X PBS and lysed with equal amount of 2X Laemelli buffer (0.125 M Tris HCl pH 6.8; 4% SDS; 0.56 M β Mercaptoethanol; 0.02% Bromophenol Blue; 20% Glycerol) for western blot analysis. TPO ELISA TPO level in the blood plasma was estimated using the Human thrombopoietin Quantikine ELISA kit from R & D Systems, Minneapolis, MN. The assay is based on the two-sided sandwich principle. Standards and undiluted PPP samples from 10 normal controls and 20 AMM patients were incubated in micro-titer Plates whose wells are pre-coated with monoclonal TPO antibody. After washing off the unbound substances, the wells were incubated with polyclonal TPO antibody conjugated with HRP enzyme. After subsequent washes to remove any unbound antibody-enzyme, a substrate solution for HRP was added to the wells. The color developed that is proportional to the amount of TPO from plasma bound to the plate was quantitated using the Opsys MR plate reader. Real-Time RT-PCR Real-Time RT-PCR was performed to determine the expression of Mpl in platelets of 12 controls and 12 AMM patients, using the Geneamp 5700 SDS Instrument from Applied Bio-systems, Foster City, CA. Reagent for the assay was the Quantitect RT-PCR kit (Qiagen, Valencia, CA). Mpl mRNA in the platelets was reverse transcribed and PCR amplified from 10 ng of total RNA with gene specific primers and a florescent reporter probe that hybridises to the gene in between the two primers. The sequences of the primers were: MPL FP 5'-AAGTCCTCAGAGAGGACTCCTTTG-3 MPL RP 5'-CAGGCAAGAAGGCTGCAATC-3' MPL PROBE 5'-FAM CCTCCCAGGCCCAGATGGACTAC-3' BHQ1 The reaction conditions were reverse transcription at 50°C for 30'; 40 cycles of amplification with de-naturation at 94°C for 15", annealing and extension at 60°C for 10". Sequence Analysis of Mpl Total RNA extracted from platelets obtained from 15 patients with AMM and 15 normal controls was used to clone the full length Mpl cDNA. First strand cDNA was synthesized using the SuperScript First Strand Synthesis System for RT-PCR from Invitrogen, Carlsbad, CA, following their specifications. About 3 μg of total RNA was mixed with 50 ng of random hexamers, and 1 mM dNTPs in 10 μl reaction volume and incubated at 65°C for 5' and left on ice for 1'. A cocktail mix of RT buffer, MgCl 2 and DTT was added to a final concentration of 1X, 5 mM and 10 mM respectively along with 40 Units of RNase OUT Ribonuclease Inhibitor and incubated at 25°C for 2'. 50 Units of SuperScript II Reverse Transcriptase enzyme was added and incubated at 25°C for 10' and 42°C for 50'. The reaction was terminated by heating at 70°C for 15' and chilled on ice. 2 Units of RNase H was added and incubated at 37°C for 20' to remove the RNA from cDNA: RNA hybrids to increase the sensitivity of PCR from cDNA. Mpl was amplified using the Expand High Fidelity PCR kit from Roche, Indianapolis, IN, with gene specific primers containing Eco R 1 (FP) and Xho 1 (RP) Restriction Enzyme sites to enable directional cloning. The Restriction Enzyme sites are indicated in lower case in the Primer sequences given below. FP 5'-CGgaattcGAAGGGAGGATGGGCTAAGGC-3' RP 5'-CCGctcgagAGTTTAGCTCTGTCCAGGGAA-3' The Products were purified using the Wizard PCR Preps DNA Purification System from Promega, Madison, WI, and digested with the Restriction Enzymes, Eco R 1 and Xho 1. The digested products were then cloned into Blue Script vector. The recombinants were checked by restriction analysis and the selected positive clones were sequenced completely at Northwoods DNA Inc., Becida, MN. Mpl cDNA were re-cloned from samples that had mutations and sequenced completely. The putative mutations were confirmed by sequencing the corresponding region of the Mpl gene from respective samples. For this, the genomic DNA isolated from the peripheral blood samples were PCR amplified with gene specific primers and the PCR products were directly sequenced after purification using the Wizard PCR Preps DNA Purification System from Promega, Madison, WI. In vitro transcription/translation of Mpl The Mpl cDNA clones in the Blue Script vector were digested with Eco R 1 and Xho 1 and the released Mpl cDNA was cloned into pcDNA 3 vector to obtain cDNA with 3' translational signals. These constructs were then subjected to in vitro transcription/translation using the TNT Quick Coupled Transcription/Translation Systems from Promega. Madison, WI, in the presence of 35 S Methionine from Amersham Biosciences. About 0.5 μg of the constructs were incubated with the TNT Quick-Coupled Master Mix and 35 S Methionine in the absence or presence of 2 μl of Canine microsomal membranes provided with the TNT Quick Coupled kit, at 30°C for 2 hours. The reactions were terminated by adding 100 μl of 1X SDS loading buffer (50 mM Tris. HCl pH 6.8; 100 mM DTT; 2% SDS; 0.1% Bromophenol Blue; 10% Glycerol) and stored at -20°C. Appropriately translated and processed products were then separated on a 6% polyacrylamide gel. The gel was dried using Model 583 Gel dryer from BIO-RAD, Hercules, CA. The products of translation were visualized by autoradiography of the dried gel. The bands were quantitated using the Image J program. Transfections 293 T human embryonic kidney cells were used for the transient transfection studies with pcDNA3/ Mpl constructs of controls and patients used in the in vitro transcription/translation studies. The cells were seeded onto a 6 well plate at a density of 5 × 10 4 cells/well a day before transfection with DMEM medium containing 10% FBS. On the day of the transfection, the cells were washed with DMEM medium without serum and covered with 800 μl of the same. In an eppendorf tube, 1 μg of the DNA was mixed with 100 μl of serum free DMEM medium and 100 μl of DMEM containing 2 μl of Lipofectamine transfection reagent (Invitrogen, Carlsbad, CA) was added to this DNA. The mixture was incubated for 30' at room temperature and then added to the 800 μl of medium covering the cells. After incubating for 5 hours at 37°C the transfection medium was removed and replaced with DMEM medium with 10% FBS. The cells were harvested 48 hours post-transfection. They were washed twice with 1X PBS and the cells dislodged by gentle re-suspension. The cells were collected and centrifuged at 1000 RPM for 10'. The cell pellets were re-suspended in 100 μl of 1 X PBS each and lysed with 100 μl of 2X Laemelli buffer. Western Blot Analysis The Transfection cell lysates/platelet cell lysates in 1X Laemelli buffer were boiled for 5', clarified by centrifugation at 14 000 RPM for 5', loaded onto a 6% SDS PAGE, and transferred onto immobilon nylon membrane. The membrane after transfer was rinsed with 1X PBS and incubated in the blocking solution (5% milk in 1 X PBS; 0.1 % Tween 20) for 1 hr., at room temperature. The membrane was then incubated with 1:10 000 dilutions of Rabbit anti Mpl antibody from Upstate USA Inc. VA, over night at 4°C in the blocking solution. The membrane was then washed 4 X with 1X PBST (1X PBS; 0.1% Tween 20) for 15' each and incubated with 1:5000 dilutions of HRP conjugated anti-rabbit secondary antibodies for 1 hr., at room temperature. Washes with 1X PBST were repeated and the membrane processed for detection using the Western Lightning Chemiluminescence Reagent Kit from Perkin Elmer, Boston, MA, following their recommendations. The bands were quantitated using the Image J program. Similar procedure was followed for the detection of eIF4E protein in the platelets except that a 10% SDS PAGE was used for separation of the proteins. The eIF4E antibody from Transduction Laboratories, CA, was used at 1:1000 dilutions. Competing interests The author(s) declare that they have no competing interests. Authors' contributions KCH performed most of the experiments and was responsible for acquisition, analysis and interpretation of the data in addition to preparation of the manuscript. JCW conceived, designed the study, provided patient samples for the study and critically reviewed the data. KS performed eIF4E western blot. GH performed the early work and helped in preparation of application for grant support. ADN was involved in critical revision of the article, in obtaining patients for the study and obtaining grant support. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549047.xml
497049
Low fasting low high-density lipoprotein and postprandial lipemia
Background Low levels of high density lipoprotein (HDL) cholesterol and disturbed postprandial lipemia are associated with coronary heart disease. In the present study, we evaluated the variation of triglyceride (TG) postprandially in respect to serum HDL cholesterol levels. Results Fifty two Greek men were divided into 2 main groups: a) the low HDL group (HDL < 40 mg/dl), and b) the control group. Both groups were further matched according to fasting TG (matched-low HDL, and matched-control groups). The fasting TG concentrations were higher in the low HDL group compared to controls ( p = 0.002). The low HDL group had significantly higher TG at 4, 6 and 8 h postprandially compared to the controls ( p = 0.006, p = 0.002, and p < 0.001, respectively). The matched-low HDL group revealed higher TG only at 8 h postprandially ( p = 0.017) compared to the matched-control group. ROC analysis showed that fasting TG ≥ 121 mg/dl have 100% sensitivity and 81% specificity for an abnormal TG response (auc = 0.962, p < 0.001). Conclusions The delayed TG clearance postprandially seems to result in low HDL cholesterol even in subjects with low fasting TG. The fasting TG > 121 mg/dl are predictable for abnormal response to fatty meal.
Background The hypothesis that low levels of high density lipoprotein (HDL) cholesterol is associated with coronary heart disease (CHD), raised since the 1950s [ 1 ]. Fifty years later, it was well-established [ 2 , 3 ] as it has been excellently proved after a number of large prospective studies [ 4 , 5 ]. In the PROCAM Study [ 6 ] 45% of men and women who developed CHD had an HDL cholesterol lower than 35 mg/dl. In the Framingham Heart Study, total cholesterol levels did not provide a predictive value in identifying people at risk for CHD compared to cholesterol/HDL ratio [ 7 ]. Furthermore, a change in ratio is better predictor for successful CHD risk reduction than changes in total cholesterol levels. There is no longer any doubt that HDL cholesterol is a powerful independent inverse predictor of CHD. On the other hand, the long duration of the postprandial lipemia and repetition of meals during the daytime leads to important changes of lipoproteins postprandially. Studies have shown that disturbed postprandial lipemia is found in patients with CHD [ 8 , 9 ] and other conditions [ 10 - 12 ] related to an increased risk of cardiovascular disease. Many studies comparing patients with CHD and controls have shown that postprandial triglyceride (TG) levels were an independent predictor of CHD in multivariate analysis [ 8 , 13 ]. In the present study, we evaluated the variation of TG postprandially in respect to serum HDL cholesterol levels. The delayed TG clearance postprandially seems to result in low HDL cholesterol even in subjects with low fasting TG. Results All participants ingested their fatty meal and tolerated it well. Baseline characteristics (Table 1 ) Table 1 Clinical characteristics of the two main study groups (all low HDL patients and controls). All biochemical values were obtained in the fasting state and for TG 4,6,8 h postprandially. Characteristics Low HDL n = 29 Controls n = 23 P values Age (years) 45(13) 51(9) NS BMI (kg/m 2 ) 26(2) 26(3) NS CHD -/+ 19/10 23/0 0.002 Hypertension-/+ 27/2 23/0 NS Smokers-/+ 17/12 23/0 <0.001 Diabetes mellitus 28/1 23/0 NS TC (mg/dl) 198(45) 194(38) NS HDL (mg/dl) 31(7) 53(19) <0.001 LDL (mg/dl) 141(39) 124(38) NS Apo A (mg/dl) 123(28) 153(42) 0.007 Apo B (mg/dl) 127(56) 126(30) NS Lp (a) (mg/dl) 25(25) 30(29) NS Glucose (mg/dl) 94(13) 90(11) NS TG 0 (mg/dl) 128(54) 89(30) 0.002 TG 4 (mg/dl) 207(102) 140(36) 0.006 TG 6 (mg/dl) 210(112) 135(45) 0.002 TG 8 (mg/dl) 192(92) 103(37) <0.001 AUC (mg/dl/h) 1459(631) 1007(253) <0.005 All values are presented as means (standard deviation). BMI: body mass index, CHD: coronary heart disease, TC: total cholesterol, HDL: high-density lipoprotein cholesterol, LDL: low-density lipoprotein cholesterol, Apo: apolipoprotein, Lp: lipoprotein, TG 0 : fasting plasma triglyceride concentration, TG 4, TG 6, TG 8 : plasma triglyceride concentration 4,6,8 h after the fat load, respectively, AUC: area under the curve. For TC, HDL and LDL, to convert from mg/dl to mmol/L divide by 38.7 For TG, to convert from mg/dl to mmol/L divide by 88.6 The clinical characteristics of the two main study groups (low HDL and controls) are shown in Table 1 . Renal and liver function was normal as determined by measuring plasma creatinine, urea, uric acid, alanine aminotransferase, aspartate aminotransferase and γ-glutamate transferase. Therefore, there was no biochemical evidence of a fatty liver. Baseline glucose levels were similar in both groups. None of the patients fulfilled the criteria for the metabolic syndrome according to the National Cholesterol Education Program- Adults Treatment Panel III (NCEP ATP III) guidelines [ 14 ]. The plasma HDL cholesterol and apolipoprotein A were lower in low HDL group compared to controls ( p < 0.001, and p = 0.007, respectively) by definition. The fasting TG (TG 0 ) concentrations were higher in low HDL cholesterol group compared to controls ( p = 0.002). Postprandial TG concentrations in the two main groups (low HDL subjects and controls) (Table 1 , Figure 1 ) Figure 1 Schematic representation of TG-AUC of all groups in relation to time. The highest AUC values were observed in low HDL and low HDL-A groups which are characterized by low fasting HDL levels and the highest fasting TG levels compared to the other groups. The lowest AUC values were found in matched-low HDL, matched-controls, low HDL-N and control groups. These are characterized by lower fasting TG levels but with variable fasting HDL levels, ranging from 30(9) to 55(20) mg/dl. This means that the primary determinant of the magnitude of TG postprandial response is the fasting TG concentration. TG-AUC: triglyceride-area under the curve Low HDL-A: low HDL-Abnormal group Low HDL-N: low HDL-Normal group TG levels 4 h after the fatty meal (TG 4 ) The low HDL subjects had a significantly higher ( p < 0.006) TG level compared to the controls. TG levels 6 h after the fatty meal (TG 6 ) The low HDL subjects had a significantly higher ( p < 0.002) TG level compared to controls. TG levels 8 h after the fatty meal (TG 8 ) The low HDL subjects had a significantly higher ( p < 0.001) TG level compared to the controls. Glucose did not show any change postprandially. The above groups were subdivided in matched-low HDL and matched-control group (matched for low fasting TG levels) and in low HDL-Abnormal (low HDL-A) and low HDL-Normal (low HDL-N) groups (based on their TG postprandial response). Results of the matched low HDL subjects and matched controls (Table 2 , Figure 1 ) Table 2 Clinical characteristics and biochemical parameters of the low HDL subjects matched to controls for fasting TG levels, and their response to a fatty meal. Characteristics Matched-low HDL n = 20 Matched-controls n = 20 P values Age (years) 47(13) 51(10) NS BMI (kg/m 2 ) 26(2) 25(2) NS CHD -/+ 11/9 20/0 0.001 Hypertension-/+ 18/2 20/0 NS Smokers-/+ 13/7 20/0 0.004 TC (mg/dl) 187(50) 197(35) NS HDL (mg/dl 31(8) 55(20) <0.001 LDL (mg/dl) 136(43) 128(38) NS Apo A (mg/dl) 123(31) 158(43) 0.009 Apo B (mg/dl) 123(63) 132(27) NS Lp (a) (mg/dl) 29(27) 28(23) NS Glucose (mg/dl) 95(14) 90(11) NS TG 0 (mg/dl) 100(31) 95(27) NS TG 4 (mg/dl) 153(55) 140(36) NS TG 6 (mg/dl) 172(105) 138(45) NS TG 8 (mg/dl) 154(78) 105(38) 0.017 AUC (mg/dl/h) 1142(322) 1007(253) NS All values are presented as means (standard deviation). BMI: body mass index, CHD: coronary heart disease, TC: total cholesterol, HDL: high-density lipoprotein cholesterol, LDL: low-density lipoprotein cholesterol, Apo: apolipoprotein, Lp: lipoprotein, TG 0 : fasting plasma triglyceride concentration, TG 4, TG 6, TG 8 : plasma triglyceride concentration 4,6,8 h after the fat load, respectively, AUC: area under the curve. For TC, HDL and LDL, to convert from mg/dl to mmol/L divide by 38.7 For TG, to convert from mg/dl to mmol/L divide by 88.6 Subjects with low HDL cholesterol had higher TG concentration at 8 h postprandially ( p = 0.017). Results of the low HDL subjects divided into those with (low HDL-N group) and without (low HDL-A group) a normal response to a fatty meal (Table 3 , Figure 1 ) Table 3 Clinical characteristics and biochemical parameters of the low HDL subjects with an abnormal (low HDL-A) and normal response (low HDL-N) to a fatty meal. Characteristics low HDL-A n = 13 low HDL-N n = 16 P values Age (years) 42(14) 48(13) NS BMI (kg/m 2 ) 26(2) 27(3) NS CHD-/+ 12/1 7/9 0.006 Hypertension-/+ 11/2 16/0 NS Smokers-/+ 6/7 11/5 NS TC (mg/dl) 221(16) 179(53) 0.007 HDL (mg/dl) 32(5) 30(9) NS LDL (mg/dl) 154(22) 130(46) NS Apo A (mg/dl) 131(18) 118(31) NS Apo B (mg/dl) 135(17) 122(68) NS Lp (a) (mg/dl) 10(7) 32(28) 0.012 Glucose (mg/dl) 92(11) 95(14) NS TG 0 (mg/dl) 172(43) 92(29) <0.001 TG 4 (mg/dl) 298(82) 135(35) <0.001 TG 6 (mg/dl) 321(74) 127(35) <0.001 TG 8 (mg/dl) 270(78) 129(37) < 0.001 AUC (mg/dl/h) 2074(381) 986(252) <0.012 All values are presented as means (standard deviation) BMI: body mass index, CHD: coronary heart disease, TC: total cholesterol, HDL: high-density lipoprotein cholesterol, LDL: low-density lipoprotein cholesterol, Apo: apolipoprotein, Lp: lipoprotein, TG 0 : fasting plasma triglyceride concentration, TG 4, TG 6, TG 8 : plasma triglyceride concentration 4,6,8 h after the fat load, respectively, AUC: area under the curve. For TC, HDL and LDL, to convert from mg/dl to mmol/L divide by 38.7 For TG, to convert from mg/dl to mmol/L divide by 88.6 Thirteen (45%) of the low HDL subjects (low HDL-A group) had an abnormal TG response to the fatty meal test compared to the controls. The other sixteen low HDL subjects (low HDL-N group) had a TG response to the fatty meal that was similar to that in the control subjects. The low HDL subjects with an abnormal fatty meal response (low HDL-A group) had higher plasma total cholesterol and lower lipoprotein (a) concentrations ( p = 0.007, and p < 0.012, respectively). As expected, postprandial TG levels were higher in those classified as having an abnormal response. TG levels of all groups (low HDL, controls, matched low HDL, matched controls, low HDL-A, low HDL-N) in relation to time are schematically represented in Figure 1 . Predictors of abnormal TG response In multivariate linear regression analysis, where the independent variables were age, body mass index (BMI), total cholesterol, HDL cholesterol, TG, lipoprotein (a), and areas under the curve (AUC) was the dependent variable, the TG 0 was the only predictor of high AUC values (Coefficience B = 7.15, p = 0.016). ROC analysis showed that TG 0 levels ≥ 121 mg/dl have 100% sensitivity and 81% specificity for an abnormal TG response (auc=0.962, p < 0.001). After we divided the low HDL men in low- or high- groups by using the ROC curve cut-off values of TG 0 , total cholesterol and lipoprotein (a), i.e. low/high TG 0 , low/high total cholesterol and low/high lipoprotein (a) groups, the only distinction that predicted an abnormal TG response was the high-TG 0 group ( p < 0.001) Discussion In the present study, we found that a slower clearance of TG-rich lipoproteins from circulation postprandially results in low fasting levels of HDL cholesterol. Additionally, fasting plasma TG concentration is the primary determinant of the magnitude of postprandial lipemia. Subjects with a low fasting HDL and low TG levels showed a delayed TG clearance at 8 h compared to those with a normal fasting HDL value. The increase in TGs (from 2 to 4 h) after meal consumption mainly reflects dietary TG absorption, whereas the return to fasting levels (from 6 to 9 h) is presumably a function of TG clearance [ 15 ]. In our study, the delayed response of TGs to the fatty meal in matched group for low fasting TG levels may be indicative for an HDL involvement. It has been proposed that elevated plasma TG concentrations promote the cholesterol ester exchange reactions mediated by cholesteryl ester transfer protein [ 16 ]. It is possible that in this transient hypertriglyceridemia, the HDL particles are TG-enriched via cholesteryl ester transfer protein mediated exchange with TG-rich lipoproteins. Such HDL-TG enriched particles are cleared more rapidly from the circulation [ 17 ] leading to low serum HDL cholesterol levels [ 18 , 19 ]. In subjects with initially low HDL concentrations and low fasting TG levels, it is possible that this reaction does not happen at all or happens in a less degree and the slower removal of TGs from the circulation could be explained by insufficient amount of HDL particles which are responsible for their clearance. The high TG and low HDL cholesterol phenotype has been frequently reported in "abdominal" obesity [ 20 , 21 ]. Obesity is associated with a range of metabolic abnormalities including fasting and postprandial dyslipidemia. This was clearly shown in the study of Couillard et al., whose study subjects were characterised by a BMI of 32.3 ± 4.5 kg/m 2 [ 22 ]. Our results showed that even in slightly overweight subjects with a BMI of 26 ± 2 kg/m 2 , this phenotype may provoke similar metabolic abnormalities. The HDL formation is closely associated with TG catabolism [ 23 ] as mentioned above. The deleterious effect of the high TG and low HDL phenotype on the rate of postprandial TG clearance has been also shown in children (mean age 14 years old) [ 24 ]. Patsch et al., [ 25 ] have proposed that low HDL cholesterol levels could result from an impaired TG lipolysis, a condition that would favour an exaggerated postprandial lipemia, in subject with hypoalphalipoproteinemia. In our study, no difference was observed in the early postprandial TG response (absorption phase) between normolipidemic controls and men with low fasting HDL cholesterol and low fasting plasma TG concentration. In contrast, the group with low HDL cholesterol (31 ± 7 mg/dl) and variable values of fasting TG (53–248 mg/dl, mean value 128 ± 54 mg/dl) revealed a higher TG curve postprandially compared to healthy subjects with moderate HDL cholesterol (53 ± 19 mg/dl) and low fasting TG levels (89 ± 30 mg/dl). When both groups were matched for fasting TG levels, this curve-difference was diminished, although HDL cholesterol levels remained significantly different between the two groups. When subjects with low HDL cholesterol were subdivided based on their response (normal or abnormal) to fat loading, fasting TG concentration seemed to be the critical determinant of such response. Since there are no official guidelines to determine normal postprandial TG ranges, the characterization of a TG response as "normal" or "abnormal" after fat loading was based on this report and on previous reports of ours and others. The peak TG mean value in control subjects was 139 ± 41 mg/dl (at 6 h) [ 11 ], 176 ± 17 mg/dl (at 3.5 h) [ 26 ], 177 mg/dl [ 27 ] and 186 mg/dl [ 28 ]. Higher TG concentrations at 4 h were reported by others, approximately 213 mg/dl and 248 mg/dl [ 29 - 32 ]. Therefore, we defined an abnormal postprandial TG response to the fatty meal as any postprandial TG concentration (at 4, 6 or 8 h) higher than 219 mg/dl which was the highest TG concentration in any hour in any control individual. Subjects with abnormal postprandial lipemia had fasting TG concentration twice higher than those with normal postprandial lipemia and the TG concentration remained much higher through all tests. Additionally, the ROC analysis showed that TG 0 level ≥ 121 mg/dl have 100% sensitivity and 81% specificity for an abnormal TG response. Couillard et al., [ 22 ] reported a significant association between the magnitude of the postprandial TG response and fasting plasma HDL cholesterol concentrations. However, the subjects included in their study showed a wider range of fasting TG concentrations (44–390 mg/dl) compared to ours (53–248 mg/dl). In the subgroup of isolated low fasting HDL cholesterol, fasting hypertriglyceridemia was a prerequisite in order to have an exaggerated postprandial TG response [ 22 ]. This finding was similar to ours; an abnormal response to the fatty meal was only dependent on fasting TG concentration. Cohen and co-workers [ 33 ] measured plasma TG and retinyl palmitate responses to different fat meals in endurance-trained men with a wide range of plasma HDL cholesterol concentrations (36–105 mg/dl). Their data indicated that the magnitude of postprandial lipemia is not primarily affected by the HDL cholesterol concentrations, which agrees to our results. However, they failed to show any correlation between HDL cholesterol levels and chylomicron remnant metabolism which is contradictory to our findings, although we did not directly measured chylomicron remnant particles. Postprandial hypertriglyceridemia is not a uniform abnormality. It pathophysiologic cause is not yet known. It is possible that the response to a fatty meal is gene dependent. It has been reported that a number of gene loci, such as these of apolipoprotein E, lipoprotein lipase, apolipoprotein CIII, apolipoprotein A1, apolipoprotein A4, cholesterol ester transfer protein are related to fat load response [ 34 , 35 ]. However, this gene polymorphism-dependence remains controversial, and it is more likely that the postprandial lipemia is a polygenic phenomenon, although the phenotype of the postprandial lipemia is probably one. This concept allowed us to study only the phenotypic manifestation of the postprandial lipemia, which would have immediate clinical implications. Our aim was to evaluate the postprandial response in men with isolated low fasting HDL cholesterol. Although it has been also described elsewhere that baseline TG levels impact on the postprandial response and that HDL cholesterol levels are predictors of this response, the novelty of this study lies on the fact that delayed postprandially TG clearance was observed in low HDL group and that emphasis should be given on the decrease of TG levels < 121 mg/dl, lower that those indicated by the NCEP ATP III guidelines. Conclusions The delayed TG clearance postprandially seems to result in low HDL cholesterol levels even in subjects with low fasting TG concentration. Moreover, fasting plasma TG levels appear to be the primary determinant of the magnitude of postprandial lipemia. The fasting TG levels higher than 121 mg/dl are predictable for abnormal response to a fatty meal. This should be taken into account for the management of patients with fasting low HDL cholesterol levels. Methods Subjects The study population consisted of 52 Greek men recruited from the Lipid Clinic of Onassis Cardiac Surgery Center, Athens, Greece. Heavy drinking, liver and renal disease, diabetes mellitus, metabolic syndrome, according to the NCEP ATP III guidelines [ 14 ], hypothyroidism and professional sport activity were exclusion criteria. No subject took lipid lowering drugs before entering the study. Only patients with CHD (defined by angiography) were taking soluble aspirin (100 mg) and isosorbide mononitrate (40 mg). No other patient or control subject was on medication. The study population was divided into 2 main categories: A: The low HDL group and the control group. 1. The low HDL group consisted of 29 men, mean age 45(13) years with low HDL cholesterol (< 40 mg/dl) according to NCEP ATP III guidelines [ 14 ]. 2. The control group consisted of 23 healthy men, mean age 51(9) years with no family history of premature atherosclerosis, diabetes mellitus, arterial hypertension or dyslipidemia. Their fasting TG levels were < 150 mg/dl, total cholesterol < 240 mg/dl and HDL cholesterol = 40 mg/dl. All subjects were never smokers. B: Matched groups according to fasting TG level. 1. The matched-low HDL group consisted of 20 men, mean age 47(13) years with low HDL cholesterol and low fasting TG levels [100(31) mg/dl]. 2. The matched controls consisted of 20 men, mean age 51(10) years with lasting TG levels 95(27) mg/dl. Fat-rich meal protocol and blood sample All patients were studied in the outpatient clinic between 8.00–9.00 am after 12 hours (h) overnight fast. The fatty meal was consumed within 20 min and plasma TG concentrations were measured before and 4, 6 and 8 h after the fat load. During this 8 h period, the participants did not eat; they could only drink water and they did not smoke. Blood samples were drawn at 8:00 am (before the meal), at 12:30 pm (4 h after the meal), at 2:30 pm (6 h after the meal), and at 4:30 pm (8 h after the meal). In all samples total cholesterol, TG, HDL, apolipoprotein A and B, and lipoprotein (a) were measured. BMI was calculated as weight divided by height (expressed in kg/m 2 ). Fatty meal The fatty meal has been previously described [ 11 ]. Briefly, the fatty meal was a slight modification of that described by Patsch et al [ 8 ], consisting of 83.5% fat, 14.0 % carbohydrates and 2.5 % proteins. This was administered in a dose based on the patient's body surface area (350 g to 2 m 2 ). Definition of abnormal postprandial response We defined postprandial hypertri-glyceridemia as any postprandial TG concentration (at 4, 6 or 8 h) higher than the highest TG concentration in any hour in any control individual; this TG value was 219 mg/dl. Determination of blood lipids and glucose Plasma total cholesterol, TG and HDL cholesterol were measured using enzymatic colorimetric methods on a Roche Integra Biochemical analyzer with commercially available kits (Roche). The serum low-density lipoprotein cholesterol levels were calculated using the Friedewald formula [ 36 ]. Apolipoprotein A, B and Lipoprotein (a) were measured by nephelometery (Nephelometer:BN-100, Behring, Germany). Blood glucose was measured by the hexokinase method with a Dade Behring reagent on a Dimension (Dade Behring) instrument. All samples were analyzed within 24 h. All participants gave their informed consent before the study. The ethical committee of the Onassis Cardiac Surgery Center approved the study. Statistical analysis Values of numerical characteristics were tested for normality and are presented as mean value with one standard deviation, when normally distributed, or as median and range, when they deviated from normality. A t-test for independent samples or a Mann Whitney U test was used for the comparison of numerical values between two groups and ANOVA or Kruskal-Wallis test for three groups, with Bonferoni correction for post hoc analysis. The comparison of clinical categorical variables was performed with the use of chi-square test. Multivariate linear regression analysis was performed to determine the predictors of elevated AUC levels, where age, BMI, total cholesterol, HDL cholesterol, fasting TG (TG 0 ), and lipoprotein (a) were the independent variables tested. ROC curve statistics was used to estimate the cutoff value of baseline triglyceride, cholesterol and lipoprotein (a) levels (which they were significant univariate predictors of elevated TG levels postprandially) above which the low HDL men develop an abnormal TG response to fat load. Using those cut-off values, we created dichotomous variables for each one of the univariate predictors to test their significance in discriminating patients with abnormal TG response to fat load. The level of significance was set at p < 0.05. AUC for serial measurements of TG levels at baseline and after the fatty meal were calculated using the trapezoid rule. Authors' contributions GK conceived the study and participated in the development of the hypothesis, the study design and drafting of the manuscript. KA is a research associate who participated in the development of the hypothesis, study design and drafting of the manuscript. NP participated in data analysis and in the interpretation of the findings. NK is a physician who participated in the study design and recruitment of subjects, clinical evaluation and collection of blood samples. KS is a senior nurse who participated in the study design, in the recruitment of subjects and in blood sample collection. EP is a biochemist who participated in the laboratory part of the study. DC participated in the design of the study and its coordination.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC497049.xml
554994
A normalization strategy applied to HiCEP (an AFLP-based expression profiling) analysis: Toward the strict alignment of valid fragments across electrophoretic patterns
Background Gene expression analysis based on comparison of electrophoretic patterns is strongly dependent on the accuracy of DNA fragment sizing. The current normalization strategy based on molecular weight markers has limited accuracy because marker peaks are often masked by intense peaks nearby. Cumulative errors in fragment lengths cause problems in the alignment of same-length fragments across different electropherograms, especially for small fragments (< 100 bp). For accurate comparison of electrophoretic patterns, further inspection and normalization of electrophoretic data after fragment sizing by conventional strategies is needed. Results Here we describe a method for the normalization of a set of time-course electrophoretic data to be compared. The method uses Gaussian curves fitted to the complex peak mixtures in each electropherogram. It searches for target ranges for which patterns are dissimilar to the other patterns (called "dissimilar ranges") and for references (a kind of mean or typical pattern) in the set of resultant approximate patterns. It then constructs the optimal normalized pattern whose correlation coefficient against the reference in the range achieves the highest value among various combinations of candidates. We applied the procedure to time-course electrophoretic data produced by HiCEP, an AFLP-based expression profiling method which can detect a slight expression change in DNA fragments. We obtained dissimilar ranges whose electrophoretic patterns were obviously different from the reference and as expected, most of the fragments in the detected ranges were short (< 100 bp). The normalized electrophoretic patterns also agreed well with reference patterns. Conclusion The normalization strategy presented here demonstrates the importance of pre-processing before electrophoretic signal comparison, and we anticipate its usefulness especially for temporal expression analysis by the electrophoretic method.
Background Amplified fragment length polymorphism (AFLP) is a DNA fingerprinting technique using electropherograms [ 1 ]. AFLP analysis belongs to the category of selective restriction fragment amplification techniques, which are based on the ligation of adapters to genomic restriction fragments followed by PCR-based amplification with adapter-specific primers [ 2 ]. This technique has been widely used for genotyping since it requires no prior knowledge of genomic DNA sequences and offers potentially better discriminatory power and speed than the existing techniques for fingerprinting such as random-amplified polymorphism DNA markers (RAPD) [ 3 - 8 ]. However, it has only been used to a limited extent for expression analysis [ 9 ]. The main problems with the comparison of AFLP patterns are (i) variation in peak height, and (ii) false positive peaks which often overlap with real peaks, probably due to differences in PCR efficiency [ 5 , 10 ]. There is room for tuning selective PCR amplification [ 8 ]. Recently, we developed an AFLP-based gene expression profiling method called HiCEP (High Coverage Expression Profiling) [ 11 ]. The experimental and analytical procedures are essentially the same as those of AFLP, i.e., the technique is based on the selective PCR amplification of restriction fragments from a total restriction digest of genomic DNA. Refinements of the selective PCR technique improved reproducibility and reduced the rate of false positive peaks as well as the number of peaks. They also enabled the digestion of purified genomic DNA with two four-nucleotide recognition restriction enzymes, having a higher cutting frequency, such as Msp I and Mse I. Consequently, the HiCEP method can detect a slight expression change of transcript-derived fragments (TDFs) with high coverage. The estimated 30,000 transcripts expressed in a cell are divided into 256 subgroups (16 MspI-NN primers * 16 NN-MseI primers) containing approximately 120 PCR-amplified TDFs. This number is small enough to be separated by fluorescent capillary electrophoresis using an automated DNA sequencer such as the ABI Prism 310 (Applied Biosystems). We can achieve higher throughput by using several fluorescent dyes at once [ 14 , 15 ]. Normally, digitized electropherograms are imported into image analysis software such as GeneScan (Applied Biosystems), which outputs each fragment (band) together with its length (in bp), area and height (signal intensity), carrying out accurate fragment sizing and background subtraction for most of the operations. GeneScan is capable of separating the signal from each fluorophore to provide higher throughput analysis. However, it should be noted that intense signals from abundant TDFs can breed into each other, potentially confusing the fragment sizing [ 7 , 15 ]. Furthermore, the use of a frequently matching 4-bp cutting endonuclease ( Mse I) tends to produce many small TDFs (< 100 bp) and in our experience this range is prone to errors of fragment sizing. Cumulative errors of fragment sizing interfere with normalization across different electropherograms and lead to the mis-assignment of valid TDFs. Hence, more detailed analysis such as observation of gradual expression changes in the time series of a TDF still counts in subjective visual examination [ 11 ]. Further preprocessing of the electrophoretic data to be compared, each of which is independently normalized according to molecular weight standards, is needed. The purpose of the present study is to develop a normalization method for the automated analysis of temporal electrophoretic data. We assume the samples to be compared are identical, that TDFs have similar fragment lengths across electropherograms and that expression changes can be detected as variations in peak height using the HiCEP technique. The performance of the method is demonstrated by analyzing a large set of time-course data obtained from mouse embryonic stem (ES) cells, using HiCEP. Results and discussion We analyzed a total of 2560 HiCEP electropherograms (256 sets of ten), containing time-course data of embryonic stem (ES) cells 0, 12, 24, 48, and 96 h after adding stimulation for differentiation. Reproducibility was confirmed by the duplication. We applied the current method to each of the 256 sets. Delineation of quality profiles for lanes When a set of electrophoretic data is arranged and surveyed, one can often find ranges (called 'dissimilar ranges') in which peak fragment lengths are incorrectly measured. For example, in Fig. 1a three lanes ( 0 h-1 , 12 h-1 , and 48 h-2 ) in the range (35–50 bp) appear to be compressed on the short side. This is probably because another intense peak just under 35 bp is mistaken for the 35 bp marker peak. This reduces the overall similarity between lanes and makes it difficult to recognize identical TDFs such as red filled peaks in Fig. 1a . To this end, we first developed a method for displaying dissimilar ranges. The method is based on a moving-fragment approach that continuously determines the average correlation coefficient between particular lane P target and the other lanes within a certain range using equation 3. By using the average correlation coefficients, we can make a quality score function Q k ( t ) for all lanes ( k = 1, 2, ..., 10) at arbitrary length t (see Methods). An example of the calculation for lane 0 h-1 is shown in Table 1 . The 'quality profiles' delineated from Q ( t ) take the place of detailed visual evaluation of dissimilar ranges (Fig. 1b ). Undoubtedly, false peaks must have been used incorrectly at 35 bp in three lanes ( 0 h-1 , 12 h-1 , and 48 h-2 ) and at 75 bp in two lanes ( 0 h-2 and 96 h-1 ). Detection of dissimilar ranges Next, we applied a simple method for the automated detection of dissimilar ranges to 256 sets of electrophoretic data (see Method). The method identified a total of 362 dissimilar ranges. Most (289, 79.8%) of the ranges were of 100 bp or less. This is reasonable because the main source of fragment sizing errors is the presence of intense peaks near the marker [ 7 , 11 , 15 ] and the HiCEP technique tends to produce short fragments. In fact, of a total of 222,108 detected peaks in the range (35–700 bp) analyzed by GeneScan, 58,988 (26.6%) were < 100 bp. Visual examination revealed many of those ranges to be genuine, but not all. The set of ten electropherograms shown in Fig. 1 is a good example. Our method identified seven ranges as dissimilar: five lanes ( 0 h-1 , 0 h-2 , 12 h-1 , 12 h-2 , and 48 h-2 ) in range (35–50 bp) and two lanes ( 0 h-2 and 96 h-1 ) in range (50–100 bp). Of these, we at first suspected that two lanes ( 0 h-2 and 12 h-2 ) in range (35–50 bp) were false-positives (mistakenly identified as dissimilar). However, we observed that the range in the two lanes is worthy of being normalized: the fragment lengths on the short side of the range deviate gradually from the mean lengths of lanes 24 h-1 , 24 h-2 , 96 h-1 , and 96 h-2 [see Additional file 1 ]. Visual examination of all the electropherograms did not reveal any false-negative errors (overlooked dissimilar ranges). Recall that the samples to be compared are identical and that the measure of the quality of fragment sizing is based on a calculation of the average correlation between electropherograms. These results suggest that the normalization strategy we present here is useful, especially for temporal expression analysis. The effectiveness of the method depends on the choice of the parameter T in equation 3 in the Methods section, which is the number of consecutive fragments making up the quality profile examined by the program. Quality profiles using the shortest span ( T = 1) are noisier than those using a moderate span, and runs using spans of less than four fragments were found unsatisfactory in our investigation. On the other hand, long spans ( T = 10) tended to miss small dissimilar ranges. These trends are essentially the same as those in the delineation of hydropathy plots of proteins using a moving-window approach and in the detection of transmembrane regions [ 16 ]. Although we set T = 5 throughout the analysis, further improvement in the choice of parameters as well as the method for the detection of dissimilar ranges remains to be studied. Normalization of dissimilar ranges To normalize dissimilar ranges across a set of electropherograms, it is necessary to select one as a reference. In conventional algorithms the reference is selected manually [ 17 , 18 ]. For reproducible automated normalization, it is vital that the choice be objective. Our method selects the lane (electropherogram) having the highest average quality score in a given dissimilar range. In the case of Fig. 1 , our method selects 96 h-2 as the best reference in ranges (35–50 bp) and (50–100 bp). We cannot, of course, reject the possibility that accurate fragment sizing is performed in the minority group (such as lanes 0 h-1 , 12 h-1 , and 48 h-2 in range (35–50 bp) in Fig. 1 ), but it is natural that the best reference should be selected from lanes in the majority group. We prepared two models for accurate normalization of various types of fragment sizing errors. Model 1 is the case of an incorrect fragment sizing at the shortest (or longest) marker peak. Figure 2 shows an example of normalization using Model 1. The best approximating profile (normalized profile) is determined by considering various combinations of candidates from D × 100% expansion (or - D × 100% compression) to D × 100% compression of the short side of the original profile at intervals of d bp. The best approximating profile is one of the candidate profiles with { x × d - D × ( C e - C s )} / ( C e - C s ) × 100% compression of the side in a given range ( C s - C e bp), where x = {0, 1, ..., 2 × ( C e - C s ) × D / d }. There is of course a trade-off between the computation time and the normalization accuracy in the choices of parameters. In Model 1, we set D = 0.4 and d = 0.2. We expected that the normalization would be achieved by a linear expansion of the short side of the dissimilar range (35–50 bp) by anchoring the long-side in the target lane 12 h-1 . Indeed, the best approximating profile that achieved the highest correlation coefficient against the reference 96 h-2 was the case of x = 9 (28% expansion). Figure 3 shows an example of normalization using Model 2. Model 2 is the case of an incorrect fragment sizing at the marker length M j in a dissimilar range ( M j -1 - M j +1 bp) (see Methods). Accordingly, the program can easily determine the length of 75 bp because there is only one marker length inside of the range (50–100 bp). We can directly apply the normalization procedure for Model 1 to Model 2 by considering two hypothetical dissimilar ranges, (50–75 bp) and (75–100 bp). The main difference from Model 1 is that the two ranges cannot be normalized independently in Model 2: { x × d - D × (100 - 50)}/(100 - 50) × 100% compression (resp. expansion) of the long-side of the original profile in range (50–75 bp) and { x × d - D × (100 - 50)}/(100 - 50) × 100% expansion (resp. compression) of the short side in range (75–100 bp) affect on each other. In Model 2, we set D = 0.1 and d = 0.2 as a maximal realistic displacement. The best approximating profile is the case of x = 13 and is consistent with the reference profile. Figure 4 shows the result of normalization for electrophoretic patterns in the primer combination of Fig. 1 . Seven dissimilar ranges (coloured in red; five in range (35–50 bp) and two in range (50–100 bp); 0 h-2 has two normalized ranges) are normalized nearly perfectly (Fig. 4a ). For example, the electrophoretic pattern of 0 h-2 in range (35–50 bp) which is a possible false-positive error are normalized as 2.7% compression of a short side of the range. The correlation coefficients between the target 0 h-2 and the reference 96 h-2 in the range before and after normalization are 0.674 and 0.798, respectively. A quality profile for lane 48 h-2 indicates that an incorrect normalization is performed in range (35–50 bp) of the lane. The low correlation coefficient (0.4) between the normalized profile and the reference 96 h-2 in the range, compared to values (> 0.7) between four other normalized profiles ( 0 h-1 , 0 h-2 , 12 h-1 , and 12 h-2 ) and the reference in the corresponding range, strengthens this suspicion [see Additional file 2 ]. After visual examination it was decided that the dissimilar range (35–50 bp) of lane 48 h-2 should be extended on the long side. We searched for the best range to be normalized and chose (35–53.6 bp). The correlation coefficient of the normalized profile, expanded by 26.3% on the short side in the range (35–53.6 bp), was 0.9. Undoubtedly an exhaustive search for edges in dissimilar range might yield better normalization for some cases. However, it also dramatically increases the possible combinations of normalization candidates. It is a balance between the computation time and the number of analyzable TDFs. One way to do objective evaluation of normalized electrophoretic patterns is to re-delineate the quality profiles (Fig. 4b ). Generally, a higher quality score Q k ( t ) for lane k indicates greater consistency with the other lanes around arbitrary length t if the sample is identical ( e.g ., time-course data). The quality scores after normalization overall were higher than before (Figs. 1b and 4b ). This means the assignments of the quality scores to time-course electrophoretic data are effective for evaluating reproducibility. Evaluation of the method The normalization method we propose here can be regarded as an image warping method which deforms images by mapping between image domains [ 19 ]. There are a number of reports on warping methods especially for dealing with two-dimensional (2-D) images [ 19 - 21 ]. There are also some methods for 1-D electrophoretic data [ 17 , 18 , 22 ]. Comparison with these methods might provide an objective evaluation of the current method. However, they are not directly comparable with the current method because of different frameworks such as input data format, the requirement of pre-determined parameters, and so on [ 17 , 22 ]. A critical step in the analysis of 1-D electrophoretic data is the assignment of the correct size to each TDF. In time-course data, one expects that the same TDFs should have quite close fragment lengths across electropherograms and that temporal expression changes are reflected as differences in peak height. We developed the current method aimed at temporal expression analysis by the electrophoretic method and used a scoring system for an objective evaluation of experimental reproducibility using Q k ( t ) which indicates a relative similarity at t (bp) in lane k to the other lanes. We demonstrate two other sets of electrophoretic data and discuss the feasibility of the method. Figure 5 shows a set of electrophoretic patterns and quality scores which is different from the primer combination used in Figs. 1 , 2 , 3 , 4 . This is a representative example of electrophoretic patterns with high quality scores (arbitrary defined as > 0.7). Visual evaluation confirmed the reproducibility of the set of ten electrophoretic patterns throughout the analyzed range (35–700 bp). There is, of course, no dissimilar range detected by the current method. We should demonstrate the case of normalization to dissimilar range (35–75 bp) where both Models 1 and 2 are applicable. A set of ten electrophoretic patterns and their quality scores shown in Figure 6 is the good example. There are three lanes with dissimilar range ( 24 h-2 , 48 h-2 , and 96 h-1 ) detected by the method. Of these, 24 h-2 and 96 h-1 were normalized using Model 1 and 48 h-2 was normalized using Model 2. Visual evaluation of the electrophoretic patterns and the quality scores after normalization verified the choices of the models as appropriate (Figure 7 ). The use of normalized electrophoretic patterns facilitates the identification of TDFs (e.g., red filled fragments in Fig. 7 ) having potential temporal expression change. The development of a peak alignment algorithm for multiple lanes and integration with the current method are the next challenge. We also estimated the feasibility of the method with regard to an increasing number of peaks with certain quality score or more. The minimum value of Q ( t ) necessary for the accurate alignment of valid TDFs across lanes is about 0.7 (Fig. 4b ). Accordingly, we set the threshold to be 0.7. The number of peaks with Q ( t ) ≥ 0.7 in the range (35–700 bp) before and after normalization are 202,204 (91.0% of the total number of peaks in the range detected by GeneScan) and 205,829 (92.7%), respectively. Furthermore, 3,334 (92%) of the 3,625 (= 205,829 - 202,204) new high-quality peaks were < 100 bp, which corresponds to the biased distribution of the detected dissimilar ranges (nearly 80% of which were 100 bp or less). Conclusion When we apply the method to HiCEP time-course data, we assume that the set of electrophoretic data to be compared is identical ( i.e ., corresponding TDFs across electropherograms should have nearly the same fragment lengths). The monitoring of temporal expression change by the HiCEP technique has great potential for screening of genes related to chemotherapeutic drug resistance, circadian rhythm, and so on [ 11 , 23 , 24 ]. Although the current method was developed for pre-processing HiCEP data, the algorithm is easily applicable to the processing of other 1-D electrophoretic data such as AFLP and DD if the samples are identical or nearly identical. We strongly recommend the strategy be widely used for data processing for temporal expression analysis by the electrophoretic method. Methods Samples mRNAs were prepared from mouse embryonic stem (ES) cells at 0, 12, 24, 48, and 96 h after removal of Leukemia Inhibitory Factor (LIF) from the culture medium. The samples subjected to HiCEP reaction were duplicated. We designated each sample as 0 h-1 , 0 h-2 , 12 h-1 , 12 h-2 , 24 h-1 , 24 h-2 , 48 h-1 , 48 h-2 , 96 h-1 , and 96 h-2 . HiCEP analysis mRNAs prepared from each sample were digested with two 4-bp-cutting endonucleases ( Msp I combined with Mse I) and ligated with the corresponding adaptors. The resulting HiCEP templates, MspI-MseI-poly(A) mRNAs, were amplified by fluorescently labelled primers; for labelling, FAM, HEX, and NED were used. In total, 256 primer combinations (16 MspI-NN primers combined with 16 NN-MseI primers; N = {A, C, G, T}) were used in the HiCEP analysis. For example, a primer combination of MspI-TA and GC-MseI is capable of amplifying particular transcript-derived fragments (TDFs) corresponding to that combination. The details of the protocol of the HiCEP reaction are described elsewhere [ 11 ]. An animation of the principle is provided at the following URL . Electrophoresis and image analysis The PCR products were denatured and loaded on an ABI Prism 310 (Applied Biosystems) for capillary gel electrophoresis. The digitized images were analyzed by the GeneScan software (Applied Biotech). The size of the fragments was calculated by the software, according to internal molecular size markers of 35, 50, 75, 100, 139, 150, 160, 200, 300, 340, 350, 400, 490, 500, 600, and 700 bp, on each gel. The fragment sizing and baseline subtraction were performed by the software. The software quantifies each peak by the fragment length L (in bp), peak height H , and area A (in arbitrary units). Accordingly, the subsequent normalization procedure accepts these three-tuples as input for detected TDFs between 35 bp and 700 bp. TDFs smaller than 35 bp or larger than 700 bp were omitted from the analysis because the range was outside the size calibration range. Delineation of quality scores for lanes The starting point of normalization is a set of lanes (10 time-course measurements; 0, 12, 24, 48, and 96 h, each experiment duplicated) in each of 256 primer combinations. We explain the procedure using data from the primer combination of ' Msp I-CT combined with tt- Mse I (designated as CT-tt )' because the ten electropherograms have some ranges for which fragment sizing is obviously inappropriate (we therefore designated such ranges as "dissimilar ranges"). The first step starts from the Gaussian approximation of each lane. The use of the approximating lane is the same as described in Aittokallio et al. [ 25 - 27 ]. Briefly, a fragment F i in lane P is originally characterized by the three-tuples ( L i , H i , A i ). If lane P consists of n fragments , the approximation of the lane at length t is given by: where σ i is obtained from the following equation: The approximation is performed independently for each lane. The ten approximate profiles of time-course data in the primer combination of CT-tt are shown in Fig. 1a . For the automated identification of 'dissimilar ranges' from the expression profiles of ten lanes , we next assign quality scores to each of the fragments , where the fragments are originally numbered with respect to their lengths. By using the ten approximate profiles, relative similarity scores for intervals from fragment i to fragment ( i + T -1) ( i = 1, 2,..., n - T + 1) in lane P target ( target = {1, ..., 10}) are calculated from the following equation: where is the Pearson correlation coefficient between the target lane P target and one of the other lanes P k in the interval ( start - end bp) which always includes T fragments from fragment i to fragment ( i + T -1) ( i = 1, 2, ..., n - T +1). The interval is defined as: start = L i - 2.5 σ i and end = L i + T -1 + 2.5 σ i + T -1 . In this analysis, the number of fragments T is held constant at T = 5; other numbers are of course possible. By applying a moving window of T fragments, most of the fragments ( n - T +2 fragments in this case, with the exception of F 1 , F 2 , F 3 , F 4 , F n -3 , F n -2 , F n -1 , and F n ) have T relative similarity scores. Finally, the relative quality value Q ( L i ) for fragment F i is defined as the average of the similarity scores which satisfy start ≤ L i ≤ end . An example of the calculation is given in Table 1 . Quality scores at arbitrary lengths t , Q ( t ), are interpolated by the use of cubic splines to . The procedure is applied to each of the ten lanes and then the quality profiles corresponding to the expression profiles are created (Fig. 1b ). The quality profiles delineated from Q ( t ) have a clear interpretation. The high (or low) score for Q k ( t ) in lane k indicates a high (or low) level of relative similarity between the lane and the others around the length t . Detection of dissimilar ranges Now we have information (quality profiles) for the automated detection of dissimilar ranges. Here we adopt a simple method for detecting the range. Briefly, 1) Seek 'seed' ranges ( C seed_s - C seed_e bp) which satisfy two conditions: a) Q ( t ) ≤ thres seed and b) they contain at least two peaks. 2) Seek C tmp_s which satisfies both and C tmp_s < C seed_s ; similarly, C tmp_e , and C tmp_e < C seed_e 3) Substitute the nearest marker length (in this case, M 1 = 35, M 2 = 50, ..., = 700) to C tmp_s (resp. C tmp_e ) for C s (resp. C e ); accordingly, both C s and C e = and C s < C e Aparameter thres seed is set to 0.3 empirically. Forexample, P 9 has the following parameters in Fig. 1b : C seed_s = 57.04, C seed_e = 89.98, C tmp_s = 52.60, C tmp_e = 104.60, C s = M 2 , and C e = M 4 . Although fine tuning might be necessary, the procedure enables us to display dissimilar ranges. Selection of the reference lane When we want to correct a dissimilar range ( C s - C e bp), we have to select the "reference" (a kind of mean or typical profile in the corresponding range). One method is to choose lane P reference satisfying max { }, where is the average of Q k in the range ( C s - C e bp) in lane . For example, the algorithm selects P 10 (i.e., 96 h-2 ) as a reference in a particular range ( M 1 - M 2 bp) and also in range ( M 2 - M 4 bp). Two models for the normalization of dissimilar ranges The meaning of the word "normalization" here is to correct the fragment lengths ( L ) and the areas ( A ) of peaks in a dissimilar range so that the similarity between the normalized electrophoretic pattern and the reference pattern in the corresponding range can be maximized. To normalize a particular lane P target against the reference P reference , we now consider the following two models. Model 1 is the case of an incorrect fragment sizing at the shortest (or longest) marker peak, i.e, C s = M 1 = 35 (or C e = = 700). The peak lengths deviate more and more from the reference length moving from C e to C s (or from C s to C e ). Model 2 is the case of an incorrect fragment sizing near marker length M j ( C s < M j < C e , j = { 2 , 3 ,..., n M - 1}; the inside of dissimilar range ( C s - C e bp)). Roughly, the deviation of peak lengths from the reference length gradually increases starting from C s ; the maximum deviation is reached at M j ( C s < M j < C e ); the deviation decreases gradually; and finally disappears at C e bp. Normalization is performed by either expanding or compressing. Consider, for example in Model 1, normalization for the expression profile of P 3 ( 12 h-1 ) in range ( M 1 - M 2 bp) against the reference P 10 ( 96 h-2 ). Undoubtedly, the profile displays a systematic deviation from the reference. The degree of the deviation gradually increases starting from M 2 bp to M 1 bp probably because an intense peak generated near the shortest marker peak for the correction of M 1 bp is used mistakenly. We expect the normalization will be achieved by a linear expansion of the short side ( M 1 ) of the range ( M 1 - M 2 bp) by anchoring the long side. The best approximating profile is found by considering various combinations of normalization candidates starting from D × 100% expansion to D × 100% compression of the short side at intervals of d bp. We set D = 0.4, as a maximal realistic displacement and d = 0.2. Accordingly, in practice, the number of combinations is 2 × ( C e - C s ) × D / d + 1 (for example, there are 61 combinations of normalization candidates in the range ( M 1 - M 2 bp)) in Model 1. For each combination x ( x = {0, 1, ..., 2 × ( C e - C s ) × D / d )}, we make a candidate profile P x by changing three parameters ( L i , A i , and σ i ) accompanied by fragments ( F i ) in the dissimilar range ( C s - C e bp), according to the level of correction (expansion or compression). Those parameters are calculated as follows: Candidates are made by substituting these transformed three-tuples in a given range ( C s - C e bp) into eq. (1). The best approximate profile is the one that achieves the highest correlation coefficient between P reference and P x ( x = {0, 1, ..., 2 × ( C e - C s ) × D / d }) in the range ( C s - C e bp). In the normalization for the expression profile P 3 in the range ( M 1 - M 2 bp) against the reference P 10 , the best normalized profile by our method matches well with the reference (Fig. 2 ). A good example of Model 2 is the expression profile P 2 ( 0 h-2 ) in range ( M 2 - M 4 bp) with the reference P 10 ( 96 h-2 ); there is no possibility of Model 1 ( C s ≠ M 1 ) and the number of incorrect marker lengths is only one ( M 3 ). Model 2 is a mixture of Model 1. The normalization is also done by one of the {2 × ( C e - C s ) × D / d + 1} combinations starting from D × 100% compression of the long side in ( M 2 - M 3 bp) and D × 100% expansion of the short side in ( M 3 - M 4 bp) to D × 100% expansion of the long side in ( M 2 - M 3 bp) and D × 100% compression of the short side in ( M 3 - M 4 bp) at intervals of d (= 0.2) bp. Unlike Model 1, we set D = 0.1 as a maximal realistic displacement. In the normalization for the expression profile P 2 in the range ( M 2 - M 4 bp) against the reference, the best normalized profile by our method is matches well with the reference (Fig. 3 ). It should be noted that when a dissimilar range ( M j - M j + l bp) is very wide ( j = 1, 2,..., n M - l ; l ≥ 3), there are two or more possibilities for incorrect marker lengths in Model 2. Of these cases, we only consider cases with j = 1 in Model 1 because such cases are the only realistic ones. For the remaining cases ( j = 2,..., n M - l ; l ≥ 3), the experiment should be redone rather than trying to normalize them by considering numerous possibilities. It should also be noted that there is a case of a dissimilar range ( M 1 - M 3 bp) to which both Models 1 and 2 are applicable. In this case, the best approximate profile is decided by comparing the two best possible profiles determined using Models 1 and 2. Authors' contributions KK invented the method and wrote the paper. RF made critical comments in light of the HiCEP experimental technique. JJR edited the paper. RA and MA provided critical comments and led the project. Supplementary Material Additional File 1 Magnified electrophoretic patterns and the quality profiles in range (35–50 bp) in Fig. 1 . Descriptions are the same as those in Fig. 1 . Detailed observation of the dissimilar range for two lanes ( 0 h-2 and 12 h-2 ) confirmed the identification. Click here for file Additional File 2 Magnified electrophoretic patterns and the quality profiles in range (35–50 bp) in Fig. 4 . Descriptions are the same as those in Fig. 4 . Visual evaluation confirmed the validity of the normalizations (2.7% compression of the short side of the range) for two lanes ( 0 h-2 and 12 h-2 ) which are suspected false-positive errors. Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554994.xml
554758
Plant viral intergenic DNA sequence repeats with transcription enhancing activity
Background The geminivirus and nanovirus families of DNA plant viruses have proved to be a fertile source of viral genomic sequences, clearly demonstrated by the large number of sequence entries within public DNA sequence databases. Due to considerable conservation in genome organization, these viruses contain easily identifiable intergenic regions that have been found to contain multiple DNA sequence elements important to viral replication and gene regulation. As a first step in a broad screen of geminivirus and nanovirus intergenic sequences for DNA segments important in controlling viral gene expression, we have 'mined' a large set of viral intergenic regions for transcriptional enhancers. Viral sequences that are found to act as enhancers of transcription in plants are likely to contribute to viral gene activity during infection. Results DNA sequences from the intergenic regions of 29 geminiviruses or nanoviruses were scanned for repeated sequence elements to be tested for transcription enhancing activity. 105 elements were identified and placed immediately upstream from a minimal plant-functional promoter fused to an intron-containing luciferase reporter gene. Transient luciferase activity was measured within Agrobacteria -infused Nicotiana tobacum leaf tissue. Of the 105 elements tested, 14 were found to reproducibly elevate reporter gene activity (>25% increase over that from the minimal promoter-reporter construct, p < 0.05), while 91 elements failed to increase luciferase activity. A previously described "conserved late element" (CLE) was identified within tested repeats from 5 different viral species was found to have intrinsic enhancer activity in the absence of viral gene products. The remaining 9 active elements have not been previously demonstrated to act as functional promoter components. Conclusion Biological significance for the active DNA elements identified is supported by repeated isolation of a previously defined viral element (CLE), and the finding that two of three viral enhancer elements examined were markedly enriched within both geminivirus sequences and within Arabidopsis promoter regions. These data provide a useful starting point for virologists interested in undertaking more detailed analysis of geminiviral promoter function.
Background Traditionally, analyses of viral promoter structure-function relationship have involved directed deletion or disruption of promoter structure, followed by determination of resulting changes in transcription, if any, resulting from the alterations [ 1 ]. A relatively small subset of the promoter elements identified in this way have been subsequently isolated and tested for their ability to influence transcription when inserted into alternative, well defined, basal promoters [ 2 ]. As an alternative to so-called 'promoter bashing' approaches to the study of promoter structure, we have instead chosen to 'mine' specific regions of viral DNA for sequence elements that, when combined with a minimal plant promoter, are able to enhance transcription of a reporter gene in planta . To test the enhancer mining approach we chose to examine a collection of geminivirus and nanovirus intergenic sequences obtained from GenBank. There are a relatively large number of available sequences for these DNA viruses and due to conserved genomic organization they contain easily identifiable intergenic regions [ 3 ]. Additionally, several studies have demonstrated in planta promoter activity using isolated or modified geminivirus or nanovirus intergenic sequences [ 4 - 21 ]. Although some areas of sequence similarity exist within the intergenic regions of the geminiviruses [ 22 ], very few of these common sequence elements have been experimentally shown to contribute to transcriptional activity. We specifically avoided using any test for evolutionary conservation of candidate elements, hoping to identify unique elements that may not necessarily be shared by large groups of related viruses. For this first broad screen, the experimental rational used made two basic assumptions; 1} that viral intergenic regions contain an enrichment of DNA transcriptional regulatory elements; and 2} that important regulatory sequence elements are often duplicated within promoters, either directly repeated, or as inverted copies of sequence segments [ 22 ]. The described enhancer mining of viral sequences is not intended to be a comprehensive analysis of viral promoter structure since by design it is limited to identification of promoter elements that up-regulate gene expression and that make use of endogenous plant transcription factors available within the un-infected test plant. However, based upon their iteration, location within intergenic regions, and ability to enhance transcription in planta , any elements identified using this approach are likely to contribute to regulation of in vivo viral gene expression during plant infection. By allowing relatively large numbers of viral sequences to be examined using a defined system, the approach has the potential of generating data useful in comparing positively acting viral promoter elements within and between viral families. In addition, identification of elements that are active in planta in the absence of viral infection provides results pertinent to understanding virus-host interactions at the level of gene control. Finally, the resulting list of active and inactive viral sequences provides a valuable starting points for subsequent, more detailed, analysis of transcription regulation of individual viruses. Results Search for candidate elements The initial search for sequence repeats was performed on the major intergenic regions of 29 different geminivirus or nanovirus genomic sequences (Figure 1 and Additional file 1 ). The search was arbitrarily halted after 105 candidate repeats were identified and was not intended to provide a comprehensive representation of all duplicated sequences within any of the viral sequences examined. Although generated using different search criteria than those employed by Arguello-Astorga et al [ 22 ], the resulting collection of geminivirus sequence repeats contains some sequences similar or identical to the described "iterons" (it should be noted that functional testing of nearly all of the "iterons" listed has not yet been reported in the literature). Figure 1 Viral enhancer elements. All viral repeats that produced greater than a 25% increase in 35S min activity are listed. For each active element the accession number, relative enhancement (with standard error), repeat length, repeat separation, source virus (and genus) and viral sequence are shown. Adaptor sequences are listed in the header of the sequence column and with imperfect repeats in bold and partial palindromes within repeats underlined. Functional testing of elements Of the 105 repeats tested (Figure 1 and Additional file 1 ), 14 (13%) reproducibly resulted in increases of at least 25% above that of the 35S min construct (p < 5% by Student's T-test, the T-test was used only as a guide since by the nature of the assay used, individual data sets are small) (Figure 1 and Additional file 1 ). The remaining 91 (87%) failed to produce any measurable enhancement of reporter gene activity (see Additional file 1 ). All the positive elements identified by the in vivo assay were subsequently tested using an in vitro dual-luciferase ® system from Promega Corp. and produced levels of enhancement very similar to those obtained using the in vivo assay (the enhancement values and standard error reported in Figure 1 and Additional file 1 include both in vivo and in vitro data normalized to 35S min = 1.0). The observed enhancement of promoter activity (~2 fold) is relatively modest compared to other viral transcriptional enhancers that have been isolated and tested (e.g., G-box [ 23 ] and AS-1 [ 24 ] type elements enhance 35S min activity 8–10 fold using this assay, data not shown). This outcome may reflect limitations of the original search parameters (only repeated elements were tested). However, several of the geminiviral elements identified in this screen have been subsequently found to display clear and unique synergistic effects when combined or multimerized (Cazzonelli, Burke and Velten, manuscript in preparation), supporting their potential to contribute to viral gene regulation during infection. Since all assays were performed on tobacco plants that had been neither infected with any of the viruses screened, nor transfected with any viral components, it is unlikely that elements strictly dependent upon virally encoded regulatory factors, or factors not native to N. tobacum , would be identified. In addition, the screen was limited to those elements that increase gene expression, and no effort was made to confirm data suggesting that an element might be a 'repressor' (e.g., the 11 elements that show 'enhancement' values less than, or equal to, one third of the 35S min activity, see Additional file 1 ). Considering these limitations, the finding that 13% of the sequences tested produced measurable up-regulation of transcription supports the original assumption that basic transcription regulatory elements are enriched within repeated sequences from the viral intergenic regions. Despite having tested approximately equal numbers of inverted sequence repeats (IR) and direct sequence repeats (DR), 11 of 14 active elements were members of the DR set, with the remaining 3 positives being palindromic (inverted repeats with no sequence between the repeats). This is somewhat surprising since many of the iterated DNA sequence elements within geminivirus intergenic regions are found as both direct and inverted repeats [ 22 ], and as such could have been present in either the DR or IR set of elements. Although the numbers tested are small, and the screen was performed using a single plant species, these results suggest that directly repeated sequences within geminivirus and nanovirus intergenic repeats have a higher probability of positively influencing transcription levels than do the inverted sequence structures. It is possible that this bias may reflect the presence within the intergenic region of DNA elements responsible for viral replication [ 25 ], including a conserved inverted repeat structure with a ubiquitous central-loop sequence [ 26 ]. Seven of the IR elements tested in this study are part of predicted replication hairpin structures (see Additional file 1 ) and did not, in this test system, result in any measurable enhancement of reporter gene expression. Manual alignment of all the active DR sequences produced three classes of related elements and several unique individuals (Figure 3 ). Five of the 14 positive DR elements contain an already identified geminiviral transcription control element, the "conserved late element" or CLE {GTGGTCCC, [ 22 , 27 ]}. The CLE sequence had been previously shown to affect expression from a minimal 35S promoter, and to be up-regulated by the viral AC2 gene product [ 27 ]. The two remaining grouped elements include a pair of "CT" rich repeats (DR08 and DR13) and two related, nearly-palindromic direct repeats from beet curly top virus (BCTV, elements DR19 and DR30). Despite the lack of an exact G-box core sequence {ACGT, [ 28 ]}, the nearly palindromic structure of the DR19 and DR30 elements {aaACTTc} is reminiscent of duplicated G-box type geminiviral elements noted by Arguello-Astorga et al [ 22 ] and later proposed as functional components within tomato golden mosaic virus (TGMV) and subterranean clover stunt virus (SCSV) promoters [ 11 , 20 ]. When scanned against the online PlantCARE promoter element database {[ 29 , 30 ]} no clear consensus emerges regarding similarity of the discovered viral elements with characterized plant cis regulatory elements (the most common hits were against light or stress responsive elements, although that may simply represent the distribution of plant elements contained within the database). Figure 3 Alignment of active repeat elements. Each directly repeated element is offset (at the "/") to align both copies of the repeat. Related elements are additionally aligned as paired repeat alignments. Bases that differ within paired repeats are in lowercase bold and palindromic sub-elements within the repeats are indicated by arrows. Areas of the alignments used to determine a consensus sequence are boxed. Element occurrence in viral and Arabidopisis sequence databases Short of directed mutagenesis of each identified viral element, followed by analysis of resulting 'mutant' virus function within infected plants, it is difficult to directly determine what contribution each of the identified enhancer elements makes to viral gene regulation. Computer analysis of an element's frequency of occurrence in defined DNA sequence databases provides an alternative mechanism for gaining insight into likely biological function for short sequence elements [ 31 ]. For example, the occurrence frequency of functionally important promoter elements is higher within DNA sequences upstream from gene coding regions, compared to the frequency within non-regulatory sequences [ 31 ]. Since the element enrichment approach works best when applied to relatively short, core consensus sequences [ 31 ], viral element searches were limited to those viral enhancers that showed a clear core consensus (CLE, BCTV DR19/30, CT-rich, Figure 3 ). The viral enhancers identified in this work were found to function within un-infected test plants, indicating that the viral elements can make use of intrinsic plant transcription factors (not virally encoded) and may, therefore, be similar or identical to endogenous plant promoter elements. In order to test for enhancement of viral enhancer sequences within higher plant promoters, the PatMatch page of the TAIR web site [ 32 ] was used to access sub-datasets of the A. thaliana genomic sequence that are exclusive to annotated coding sequences {CDS} and three upstream sequence lengths {-3000, -1000, -500 bp, measured from each CDS start codon}. Each of the sub-datasets was searched for the viral elements (CLE, BCTV DR19/30, CT-rich) and, as controls, several well defined plant promoter element consensus sequences (the "G-Box" {CACGTG}, a common plant promoter element that is associated with members of the pZIP family of transcription factors [ 33 , 34 ], and two less prevalent plant promoter elements, the drought response element ('DRE', RCCGAC [ 35 ]) and abscisic acid response element (ABRE-like, ACGTGKM) [ 35 ]). Performing similar oligonucleotide frequency searches for element enrichment within viral promoters was complicated by the lack of comprehensive annotation of viral sequence entries within the GenBank database. Without clear annotation of intergenic and coding sequences within the viral GenBank entries, it was impossible to directly perform the same sort of 'upstream sequence' (in this case, viral intergenic regions) versus 'coding sequence' frequency comparisons that were possible using the fully annotated Arabidopsis genome sequence and PatMatch. As an alternative, screens were performed to determine frequencies of occurrence for viral enhancers (and control plant elements) within a sequence database consisting of all geminivirus or nanovirus GenBank entries as of May 13, 2004 [ 36 ], and the results compared with those obtained scanning the same sequences against the Arabidopsis PatMatch datasets. The searched viral sequence database has the potential for bias due to the existence of a numerous entries containing only coding regions or only intergenic sequences, as well as some duplication of sequences in separate entries. Any such bias should, however, similarly affect the baseline frequency values resulting from searches using the 18 matched random oligonucleotides (in parenthesis, Table 1 ), thus all element enrichments are considered relative to the random oligo values. It was decided to perform the searches using the full geminiviral plus nanoviral database, since limiting the viral entries to only those containing fully annotated, complete viral sequences would have greatly reduced the number of different viruses examined. Table 1 Element occurrence frequencies within viral and Arabidopsis sequence databases Element Identifier Element Sequence Occurrence frequency from each database. Values are relative to Arabidopsis CDS = 1.00 (Mean of 18 matched oligomer frequencies) Arabidopsis -3000 to -1001 Arabidopsis -1000 to -501 Arabidopsis -500 to -1 Arabidopsis CDS Gemini + nanovirus Previously Identified Promoter Elements from A. thaliana (*also confirmed as geminiviral element) ABRE-like ACGTGKM 1.65 (1.72) 1.78 (1.8) 3 (1.44) 1 (1.59) 2.45 (1.28) DRE RCCGAC 1.86 (1.75) 1.81 (1.55) 2.13 (1.46) 1 (1.13) 1.09 (0.85) G-box* CACGTG 2.28 (1.79) 2.57 (1.58) 4.35 (1.47) 1 (1.41) 3.81 (1.43) Consensus Gemini/Nanoviral Sequence Elements (**not a promoter element) CLE GTGGNCCC 3.15 (3.51) 3.62 (2.99) 3.9 (2.79) 1 (1.56) 17.36 (2.81) DR08/13 TCTCTCTCTA 3.15 (0.46) 3.6 (0.4) 7.75 (0.35) 1 (0.51) 6.92 (0.53) BCTV DR19/30 AAACTTC 0.7 (0.62) 0.69 (0.64) 0.68 (0.66) 1 (0.72) 0.64 (0.52) GV rep-stem** CGCGNCCA 2.52 (3.51) 2.2 (2.99) 2.26 (2.79) 1 (1.89) 17.11 (3.42) The results of the searches are displayed in Table 1 . Each frequency value (cHits/Mbp) represents the number of hits per million base pairs, corrected for the database base composition using empirically determined G/C and A/T ratios for each of the databases examined (see Materials and Methods). To facilitate comparison, the resulting cHits/Mbp from the Arabidopsis upstream databases (-3000 to -1001, -1000 to -501, and -500 to -1 bp) were normalized relative to the value obtained for each element's occurrence within the A. thaliana coding sequence database (CDS value set to 1.0). In addition to the predicted frequency values, in each case, the element's observed frequency was also compared to a value generated using the average of 18 random oligomers having the same length and base composition as the element tested (in parenthesis, Table 1 ). The test sequences for plant ABRE-like and G-box elements showed clear enrichment within the upstream Arabidopsis sequences, especially within the -1 to -500 region (ABRE-like element = 3.0 time the CDS value, vs 1.44 for random sequences and G-box = 4.35 vs 1.47 for random sequences, all as normalized cHits/Mbp). Results for the DRE element were less convincing (2.13 vs 1.46 in the -1 to -500 dataset) and likely reflect lower functional usage of this element within the Arabidopsis genome [ 35 ]. As expected, the CLE consensus sequence (GTGGNCCC) was found to be markedly enriched within the viral database, occurring 6 times more frequently than the mean of 18 random 8-mers of identical base composition (CLE = 17.36 normalized cHits/Mbp vs 2.81 from matched random sequences). This frequency is similar to that found (17.11 vs 3.42) using a short sequence of identical base composition and length that matches a highly conserved replication stem-loop sequence (CGCGNCCA), a component that is evolutionarily conserved within the geminivirus population [ 37 ]. Enhancement of CLE within Arabidopsis promoters is less obvious (CLE = 3.9 in the -1 to -500 database vs 2.79 for random sequences). The observed relatively small CLE enrichment is consistent with reports of a low frequency of occurrence for a CLE-like "TCP domain" binding consensus sequence (Gt/cGGNCCC) within Arabidopsis promoters [ 38 ]. It is possible that TCP domain-containing transcription factors contribute to the observed CLE enhancer activity since Arabidopsis promoters containing the TCP domain consensus binding element were found to function in transgenic tobacco and to show reduced activity after mutation of the element's core sequence [ 38 ]. The test sequences for plant element occurrence within the viral database (ABRE-like = 2.4 vs 1.28 and G-box = 3.81 vs 1.43, DRE = 1.09 vs 0.85) provide further indication of the technique's utility. The G-box viral frequency is consistent with a previous report that a G-box element contributes to transcriptional regulation from the major intergenic region of Tomato Golden Mosaic Virus {TGMV, ([ 20 ]}. The ABRE-like element enrichment in the viral database may indicate that viruses make use of biotic and abiotic stress-induced up-regulation [ 39 ] of genes driven by ABRE-containing promoters, a possibility open to additional research. Of the remaining viral elements tested against the Arabidopsis and viral databases (Table 1 ), only the DR08/13 TC-rich sequence showed clear enrichment in both plant promoter and viral sequences ( Arabidopsis -1 to -500 = 7.75 vs 0.35 and viral = 6.92 vs 0.53). Similar TC-rich regions have been reported within plant promoter regions [ 40 , 41 ], but we are unaware of any published report that confirms enhancer activity associated with an isolated TC-rich element, either viral or plant in origin. Discussion Except for the CLE elements, none of the active elements identified in this work have been experimentally reported as regulatory components of viral promoters. This is likely a reflection of both the limited number of geminivirus and nanovirus promoters that have been examined in detail [ 4 , 5 , 11 , 12 , 14 , 20 , 27 , 42 , 43 ], and the alternative approach of examining individual isolated elements used in this study. The mapped promoter components within the intergenic region of Tomato golden mosaic virus (TGMV) sub-genome A (TGMV-A) [ 14 , 20 ] provide a useful benchmark for comparison of results from this enhancer screen. Application of the repeated sequence screen to the TGMV (component B) intergenic region identified a single TGMV Direct repeat, DR38, and a single palindrome (PAL20), both of which were found to be inactive in our assay. This is consistent with published work that indicates most of the defined regulatory sequences within the TGMV-A intergenic region appear to occur as single copies [ 14 , 20 ]. The screen of intergenic repeats reported in this paper did, however, identify the CLE element, one copy of which has been shown to be part of the TGMV-A rightward promoter [ 14 , 20 ]. It is clear that testing only repeated elements will not identify all components of a promoter region, and when focusing on a specific promoter, testing of non-repeated elements (perhaps identified by evolutionary conservation) should be combined with other techniques such as insertion scanning [ 44 ]. Recently a collection of plant-functional promoters and terminators were isolated from the set of 7 Subterranean clover stunt virus (SCSV1-SCSV7) sub-genomic circles. The collection of sequence repeats tested in this study included 11 inverted or direct repeats from SCSV circles, only one of which (DR08 from SCSV2) showed any enhancing activity. It will be interesting to see how these tested repeated elements behave when examined in the context of the remainder of the SCSV promoter components. Conclusion This screen of viral intergenic repeats was undertaken to specifically identify general transcriptional enhancing elements contained within intergenic regions of a subset of geminivirus and nanovirus genomes. The screen was successful in demonstrating transcriptional enhancer activity from one proven viral promoter element and several previously unidentified elements. The occurrence of the repeated elements within intergenic regions, combined with the clear enrichment within viral sequences and Arabidopsis upstream sequences for at least the CLE and TC-rich (DR08/13) classes of elements, strongly supports participation of the enhancers in viral gene expression. The technique of testing isolated elements represents an alternative to normal promoter-by-promoter dissection and provides a useful tool for screening promoter regions for potential functional elements that have been implicated by any number of possible criteria (e.g. copy number, evolutionary conservation, comparison of promoters with similar function, microarray data, etc.). Although the number of elements tested is relatively small and, so far, only representative of promoters from the geminiviruses and nanoviruses classes of plant viruses, there is a clear trend suggesting that directly repeated elements (including those containing small internal palindromic sequences) are more likely to play significant roles in the enhancement of transcription than inverted repeats. This work represents one of the first attempts to directly screen for individual plant promoter elements that are isolated from their native promoter context. It is therefore, difficult to gauge the actual contribution of any of the elements identified to viral gene regulation and biological activity. These results do, however, provide a useful starting point for more detailed analyses of not only geminivirus and nanovirus promoters, but also overall plant promoter structure-function relationships. Methods Identification of sequence repeats The search for repeated DNA sequences was performed by visual inspection of computer-generated dot matrix comparisons (criteria: ≥ 66% identity, 10 base window, GeneWorks v2.5.2, Oxford Molecular Group Inc.). Dot matrices generated using each viral plus strand plotted against itself were used to identify direct repeats while inverted repeats were found by plotting each plus strand against its complement. Production of sequence repeat test constructs The identified repeats were synthesized as DNA cassettes containing the duplicated elements in their original orientation, either directly repeated with spacer sequence ('DR', 41 elements), inversely repeated with spacer sequence ('IR', 45 elements), or palindromic inverted repeats without spacer ('PAL', 20 elements). In order to limit the tested component to only the repeated elements themselves, any sequence occurring between the viral repeats (ranging from 0 to 146 bp, median separation = 9 bp) was replaced with a 10 bp randomized stuffer sequence (GAAGATAATC). The resulting cassettes were inserted immediately upstream from a minimal promoter (-46 to +1 relative to transcription start, 35S min) reporter system derived from the cauliflower mosaic virus (CaMV) 35S promoter fused to an intron-modified firefly luciferase (FiLUC) gene (Figure 2 , [ 45 ]). The resulting test constructs were generated as part of a modified pPZP211 [ 46 ] binary plant transformation vector (Figure 2 ) and were introduced into the Agrobacteria tumefaciens strain, EHA105 [ 47 ] by electroporation [ 48 ]. The final Agrobacteria strains each contain, in addition to the test plasmids, a second, compatible, binary transformation vector expressing an intron-modified version of the Renilla reniformis luciferase gene (RiLUC) [ 49 ] under control of the constitutive Super-promoter [ 50 ]. The FiLUC and RiLUC enzymes can be independently assayed, making the co-transferred constitutive RiLUC gene a useful marker for gene transfer and for normalization of FiLUC values between individual elements [ 45 ]. Figure 2 T-DNA map of plasmid 35S min (in pPZP212). T-DNA borders: RB = right border, LB = left border, FiLUC = firefly luciferase, Nos t = nopaline synthase transcription terminator, PClSV = Peanut chlorotic streak virus promoter, Bar = phosphinothricin acetyl transferase, 35S t = transcription terminator for the Cauliflower mosaic virus (CaMV) 35S transcript. DNA sequence insert shows the minimal 35S promoter from CaMV, from -46 to +1 (transcription start). Upstream from the minimal 35S promoter are the restriction sites (underlined: HindIII; BamH, overlined: XbaI; KpnI) used to insert test sequences and downstream is the start codon from the luciferase coding region (bold ATG). Lucifrease assays Agrobacteria harboring the test and normalization binary plasmids were grown at 28°C in LB media containing the appropriate antibiotic selection (25 μg/mL kanamycin sulfate or 100 μg/mL spectinomycin) until an OD 600 of 0.8 was achieved. The resulting cultures were centrifuged at 3000 rpm for 15 minutes, washed and re-suspended in an equal volume of infiltration media (50 mM MES, 0.5% glucose, 2 mM NaPO 4 , 100 μM Acetosyringone) before being mechanically infused (5 ml syringe) into multiple individual tobacco ( N. tobacum , cv. SR1) leaves (2–4 leaves per test construct). Assays were performed in groups of 4–8 constructs and the resulting luciferase activities (both FiLUC and RiLUC) determined after 3–4 days using an in vivo floating leaf-disk assay developed for this enhancer screen [ 45 ]. Test constructs were assayed from 1 to 6 times, with each assay consisting of 2–4 disks (3 mm diameter) per infusion. The disks used in vivo assays were each measured for light production in separate wells of a white-walled 96 well microtiter plate (FLUOstar Optima luminometer ® from BMG Lab Technologies Inc.) and all elements that tested positive in the in vivo assay were subsequently confirmed using the in vitro dual-luciferase ® from Promega Corp (assays performed according to the manufacturers instructions, separate leaf disks from the same leaf infusions were used for the in vivo assays). Each test group included an infusion containing the 35S min construct (lacking any viral test element). In order to compare the various assay systems, all activities were normalized to the activity of the 35S min construct included within each assay set (35S min activity arbitrarily set to 1.0). Determining DNA sequence element frequency in viral and Arabidopsis databases Since the element enrichment approach works best when applied to relatively short, core consensus sequences [ 31 ], database searches were limited to those viral enhancers that displayed a clear core consensus (CLE, BCTV DR19/30, CT-rich, Figure 3 ). Results from the viral enhancer searches were compared to values obtained using previously reported plant promoter elements (DRE, ABRE-like, and G-box), and a short DNA sequence that is part of a highly conserved geminiviral replication loop stem sequence (CGCGNCCA) that is identical in base composition and length to the CLE consensus (Table 1 ). The short sequence elements were each tested for their frequency of occurrence within a set of DNA sequence databases. One database consists of all entries for geminiviruses plus nanoviruses ([ 36 ], as of May, 2004) and all others are from the A. thaliana genomic sequence at the TAIR, PatMatch web site [ 32 ]. The geminivirus/nanovirus BLAST searches were set for short exact matches (the statistical significance threshold set to 1000 and word size set at the element's length), returning the number of occurrences of exact matches for the full length element within the database. The TAIR PatMatch searches (default settings: Max hits, 7500; both strands; mismatch = 0; minimum hits/seq = 1; maximum hits/seq = 100) were performed against sub-datasets representing Arabidopsis coding sequences {"GI CDS (- introns, - UTRs)"}, and various lengths of upstream regions {"Locus Upstream Sequences", -1 to -500, -1 to -1000 and -1 to -3000}. Results from the -500 search were subtracted from the -1000 results, to generate hits from -501 to -1000 and -1000 results subtracted from the -3000 data to calculate hits from -1001 to -3000. In order to allow direct comparison between searches in different databases, using sequence elements of differing length and base composition, the number of database hits was corrected for the size of the database (number of hits divided by the database size in mega-base pairs {Mbp}) and base composition (hits/Mbp divided by the predicted number of hits per Mbp using upon the element sequence and base composition of each search database). The dataset base compositions were determined from downloaded sequence files and are: A. thaliana CDS: A/T = 55.8%, G/C = 44.2%; A. thaliana upstream (-1 to -500): A/T = 67.43%, G/C = 32.57%; A. thaliana upstream (-501 to -1000): A/T = 66.24%, G/C = 33.76%; viral: A/T = 56.2%, G/C = 43.8%. The resulting frequency of occurrence is a corrected number of hits per mega-base pairs (cHits/Mbp). For ease of comparison between elements, all of the cHits/Mbp values have been normalized to the corresponding cHits/Mbp number from the A. thaliana CDS database (set arbitrarily to 1.0). Correction of the element's frequency using the calculated random probability of occurrence does not account for the possible impacted by intrinsic base-order bias that may occur within each sequence database, specifically the coding region database. These biases can potentially shift cHits/Mbp numbers markedly from those calculated using simple random base composition frequencies. To help confirm the significance of any observed enhancement in an elements frequency, mean cHits/Mbp values for 18 randomly generated sequences that match each test sequence for base composition and length were determined to provide a baseline value for comparison to that of the test element (shown in parenthesis, Table 1 ). A total of 18 sequences were used to produce the reported baseline as mean cHits/Mbp values were found to routinely level off at n values of between 8–12 random sequences examined (data not shown). Competing interests A patent application is being considered for synthetic plant promoters containing some of the elements described in this article. Disclaimer Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. Authors' contributions JV conceived of the study, participated in its design and coordination and drafted the manuscript. KM performed much of the search for short repeats within viral sequences and contributed to development of protoplast-based reporter gene assays. CIC generated and tested all the elements examined and developed the in vivo assay used to quantify enhancer activity. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Excel worksheet listing viral elements that fail to enhance expression Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554758.xml
546412
Considerations for the future development of virtual technology as a rehabilitation tool
Background Virtual environments (VE) are a powerful tool for various forms of rehabilitation. Coupling VE with high-speed networking [Tele-Immersion] that approaches speeds of 100 Gb/sec can greatly expand its influence in rehabilitation. Accordingly, these new networks will permit various peripherals attached to computers on this network to be connected and to act as fast as if connected to a local PC. This innovation may soon allow the development of previously unheard of networked rehabilitation systems. Rapid advances in this technology need to be coupled with an understanding of how human behavior is affected when immersed in the VE. Methods This paper will discuss various forms of VE that are currently available for rehabilitation. The characteristic of these new networks and examine how such networks might be used for extending the rehabilitation clinic to remote areas will be explained. In addition, we will present data from an immersive dynamic virtual environment united with motion of a posture platform to record biomechanical and physiological responses to combined visual, vestibular, and proprioceptive inputs. A 6 degree-of-freedom force plate provides measurements of moments exerted on the base of support. Kinematic data from the head, trunk, and lower limb was collected using 3-D video motion analysis. Results Our data suggest that when there is a confluence of meaningful inputs, neither vision, vestibular, or proprioceptive inputs are suppressed in healthy adults; the postural response is modulated by all existing sensory signals in a non-additive fashion. Individual perception of the sensory structure appears to be a significant component of the response to these protocols and underlies much of the observed response variability. Conclusion The ability to provide new technology for rehabilitation services is emerging as an important option for clinicians and patients. The use of data mining software would help analyze the incoming data to provide both the patient and the therapist with evaluation of the current treatment and modifications needed for future therapies. Quantification of individual perceptual styles in the VE will support development of individualized treatment programs. The virtual environment can be a valuable tool for therapeutic interventions that require adaptation to complex, multimodal environments.
Background Visual imaging is one of the major technological advances of the last decade. Although its impact in medicine and research is most strongly observed in the explosion of PET and fMRI studies in recent years [ 1 ], there has been a steady emergence of studies using virtual imaging to measure and train human behavior. Virtual environments (VE) or virtual reality (VR) have taken a foot hold in rehabilitation with dramatic results in some cases. Some applications have the patient wearing VE systems to improve their ability to locomote [ 2 ]. Others bring the VE technology to the patient to improve much needed rehabilitation [ 3 ]. With either approach, there are at least two issues that need to be addressed by the clinical or basic scientist employing virtual technology to elicit natural human behaviors. One is the ability of the technology to present images in real-time. If the virtual stimulus has delays that exceed those expected by the central nervous system (CNS), then the stimulus will most likely be ignored or processed differently than inputs from the physical world. Once a response is elicited, it must be determined whether the variability observed across individuals is due to individual differences or inconsistencies between expectation and the presentation of the virtual image. Components of a virtual environment Let us first define what we consider a VE and consider the signals that need to be transmitted for such a system to operate remotely (TeleImmersion). VE is immersion of a person in a computer generated environment such that the person experiences stereovision, correct perspective for all objects regardless of their motion, and objects in the environment move in a natural fashion with subject motion. To achieve theses characteristics, certain technology must be utilized. To provide stereovision, slightly different images must be presented to the right and left eyes with little if any cross talk between the two images. In some systems this is provided by using field sequential stereo in combination with liquid crystal shutter glasses (StereoGraphics, Inc). In this system the right liquid crystal lens is clear while the left is opaque and the perspective scene generated on the screen is that for the right eye. Then the left eye lens is clear and the right is opaque and the left eye's view is displayed. This method of producing stereo has found its way into projection based systems [ 4 , 5 ] and desktop systems also known as "fish tank VR" [ 6 ]. In other systems the person wears a head mounted display (HMD) where the right and left eye each see a dedicated display so that the computer generates a left and right eye perspective image and each image is connected to the corresponding monitor. Such systems have used miniature CRTs, Liquid Crystal Displays, and Laser light directed into the eye to create the image on the retina [ 7 ]. In contrast to the above mentioned systems, an auto-stereographic system displays stereo images to the person without the aid of any visual apparatus worn by the person [ 8 ]. The person merely looks at the screen(s) and sees stereo images as one might in the natural world. Because of their ease of use by the subject and their versatility these new and experimental systems have the potential of becoming the ultimate VE display when large motions of the subject are not needed. Regardless of the system used, to keep all the stereo objects in the correct perspective and to keep them from being distorted when the person moves in the environment, it is necessary to track the movements of the person so that the computer can calculate a new perspective image given the reported location of the person's head/eyes. The tracking systems that are used to do this are varied. The most commonly used of these are the 6-degrees of freedom (DOF) magnetic tracking systems (Ascension, Inc and Polhemus, Inc.). With these systems a small sensor cube is placed on the subject and the location of the sensor within the magnetic field is detected. When the sensor is place on the head or glasses of the person the orientation of the head and therefore the location of the eyes can be presumed. Other non-magnetically based systems use a combination of acoustic location to delineate position and acceleration detection to obtain body coordinates in space. The combination results in 6 DOF for the location information (InterSense, Inc). Other systems use cameras to track the person and then transform this information to the 6-DOF needed to maintain a proper image in the VE (Motion Analysis, Inc). So far we have confined our discussion to visual objects and have not considered the use of haptic or other forms of information to be integrated into the VE system [ 9 ]. To provide a realistic haptic experience to the subject, objects must be rendered at 1000 times per second. While a local haptic system such as that produced by Sensable Inc. and others can provide such high speed communication, when such information is floated over the network the issues of bandwidth and latency of the network are paramount to consider. While experimental networks have significantly increased the bandwidth of the network, our ability to move information over these networks is currently fixed by the speed of light. Prediction and other methods can be employed to help reduce the effective latency (Handshake Technologies, Inc), but this characteristic will continue to pose a problem for many conditions that we would like to use in tele-rehabilitation. In networked VEs several types of data need to be transmitted between collaborating sites: 1. the main data-set itself (this often consists of 3D geometry); 2. the changes to the data-set (these occur when collaborating users modify the geometry in some way – perhaps by moving the object or deforming it); 3. the virtual representation of the remote collaborator (this often is referred to as an avatar); 4. the video and/or audio channel (that facilitates face-to-face conversation.) Video has limited use in stereoscopic projection-based VEs because the large shutter glasses that the viewer uses to resolve the stereo tends to hide the viewers face from the camera. Furthermore most stereoscopic projection systems operate in dimly lit rooms which are usually too dark for effective use of video. The common model for data sharing in networked VEs is to have most of the main data-set replicated across all the sites and transmit only incremental changes. Furthermore the main data-set is often cached locally at each of the collaborating sites to reduce the need for having to retransmit the entire data-set each time the application is started. Classically TCP (Transmission Control Protocol – the protocol that is widely used on the Internet for reliable data delivery) has been the default protocol used to distribute the data-sets. TCP works well in low-bandwidth (below 10 Mb/s) or short distance (local area) networks. However for high-bandwidth long-distance networks, TCP's conservative transmission policy thwarts an application's attempt to move data expediently, regardless of the amount of bandwidth available on the network. This problem is known as the Long Fat Network (LFN) problem [ 10 ]. There are a wide variety of solutions to this [ 11 ], however none of them have been universally adopted. Changes made to the 3D environment need to be propagated with absolute reliability and with minimal latency and jitter. Latency is the time it takes for a transmitted message to reach its destination. Jitter is the variation in the latency. Fully reliable protocols like TCP have too much latency and jitter because the protocol requires an acknowledgment to verify delivery. Park and Kenyon [ 12 ] have shown that jitter is far more offensive than latency. One can trade off some latency for jitter by creating a receiving buffer to smooth out the incoming data stream. UDP (User Datagram Protocol) on the other hand transmits data with low latency and jitter, but is unreliable. Forward Error Correct (FEC) is a protocol that uses UDP to attempt to correct for transmission errors without requiring the receiver to acknowledge the sender. FEC works by transmitting a number of redundant data packets so that if one is lost at the receiving end, the missing data can be reconstructed from the redundant packets [ 13 ]. FEC however is not completely reliable. Hence to achieve complete reliability (at the expense of an infrequent increase in jitter) FEC is often augmented with an acknowledgment mechanism that is only used when it is unable to reconstruct a missing packet. The virtual representation of a remote collaborator (avatar) is often captured as the position and orientation of the 3D tracking devices that are attached to the stereoscopic glasses and/or 3D input device (e.g. a wand). With simple inverse kinematics one is able to map this position and orientation information onto a 3D geometric puppet, creating lifelike movements [ 14 ]. The 3D tracking information is often transmitted using UDP to minimize latency and jitter – however since the data is mainly used to convey a user's gesture, absolute delivery of the data is not necessary. Furthermore since tracking data is transmitted as an un-ending stream, a lost packet is often followed soon after (usually within 1/30 th of a second) by a more recent update. Audio and video data are similar in property to the avatar data in that they usually comprise an unending stream that is best transmitted via UDP to minimize latency and jitter. Often video and audio packets are time stamped so that they can be synchronized on the receiving end. When more than two sites are involved in collaboration it is more economical to send audio/video via multicast. In multicast the sender sends the data to a specific device or machine that then copies the data to the various people that are subscribers to the data. For example, a user send their data to a multicast address and the routers that receive the data send copies of the data to remote sites that are subscribed to the multicast address. One drawback of multicast is that it is often disabled on routers on the Internet as one can potentially inundate the entire Internet. An alternative approach is to use dedicated computers as "repeaters" that intercept packets and transmit copies only to receivers that are specifically registered with the repeater. This broadcast method tends to increase the latency and jitter of packets, especially as the number of collaborators increases. Quality of Service (QoS) QoS refers to a network's ability to provide bandwidth and/or latency guarantees. QoS is crucial for applications such as networked VE, especially those involving haptics or tele-surgery, which are highly intolerant of latency and jitter. Early attempts to provide QoS (such as Integrated Services and Differentiated Services) have been good research prototypes but have completely failed to deploy across the wider Internet because telecommunications companies are not motivated to abide by each others QoS policies. It has been argued that QoS is unnecessary because in the future all the networks will be over-provisioned so that congestion or data loss that result in latency and jitter, will never occur. This has been found to be untrue in practice. Even with the enormous increase in bandwidth accrued during the dot-com explosion, the networks are still as unpredictable as they were a decade ago. Ample evidence is available from the online gaming community which often remarks about problems with bandwidth, latency and jitter during game sessions [ 15 ]. These games are based on the same principles that govern the design of networked VEs and therefore serve as a good metric for the current Internet's ability to support tightly coupled collaborative work. Customer Owned Networks Frustrated by the lack of QoS on the Internet, there is growing interest in bypassing the traditional routed Internet by using the available dark fiber in the ground. Dark fiber is optical fiber that has not yet been lit. Currently it is estimated that only about 5–10% of the available fiber has been lit, and each fiber has several terabits/s of capacity. The dot-com implosion has made this dark fiber and wavelengths of light in the fiber, very affordable. The newly emerging model is to construct a separate customer-owned network by purchasing or leasing the fiber from a telecommunications company, and installing one's own networking equipment at the endpoints. A number of federally supported national and international initiatives have been underway for the last few years to create customer-controlled networks explicitly for the scientific community. These include the National Lambda rail [ 16 ], StarLight [ 17 ], and the Global Lambda Integrated Facility [ 18 ]. By creating dedicated fiber networks, applications will be able to schedule dedicated and secure light paths with tens of gigabits/s of unshared, uncongested bandwidth between collaborating sites. This is the best operating environment for tightly coupled networked, haptic VEs. Connection Characteristics for Rehabilitation The ability to use virtual technology for rehabilitation is a function of cost, availability, and the kind of applications that can best utilize the network and provide rehabilitation services. Thus far, tele-rehabilitation research has focused on the use of low speed and inexpensive communication networks. While this work is important, the potential of new high-speed networks has not gathered as much attention. Consequently, we have little but imagined scenarios of how such networks might be utilized. Let us consider the case where a high-speed network connects a rehabilitation center and a remote clinic. The question is what kind of services can be provided remotely. The scenario that we envision is one where patients are required to appear at a rehabilitation center to receive therapy. Our scenario could work in several conditions. For example, a therapist at one location may want an opinion about the patient from a colleague at another location or, perhaps, the therapist can only visit the remote location once per week and with virtual technology the daily therapy could still be monitored by the therapist remotely. In our imagined condition we have a therapist at a rehabilitation center with VE, haptic and video devices and software to help analyze the incoming data (i.e., data mining) feeding to a remote clinic with identical equipment connected together through a dedicated high speed network. As displayed in Fig. 1 , the therapist station has several areas of information that connects him/her to the patient in the remote clinic. The VE (in this case Varrier) provides the therapist with a representation of the patient and the kind of trajectory that will be needed for this training session. Notice that the use of Varrier removes the need for HMD or shutter glasses to be worn by the patient or therapist. This may seem like a minor difference, but now the patient and the therapist can see each other eye to eye. The video connection allows more communication (non-verbal or bed side manner) to take place between the two linked users of this system. The haptic device serves two purposes (1) to feedback the forces from the patient's limb to the therapist and (2) to feed the forces that the therapist wishes the patient to experience. Furthermore, we could provide a task that uses the affected limb so that learning and coordination is encouraged. Other possibilities include having the robot apply forces to the patient appendage so that adaptation and recovery of function occurs [ 9 ]. In our scenario we could allow the patient to see both the virtual limb and their own limb if needed by the therapy. As can be seen from Fig. 1 , the bandwidth and latency requirements change as a function of the kind of information that is being transmitted. Figure 1 Possible tele-rehabilitation scenario facilitated by high bandwidth networking. A system as described above is possible today although expensive. The network characteristics that would be needed for each information channel would be as follows. A high-bandwidth connection would be needed for video and audio streamed to the plasma displays at each location, in addition to the high bandwidth a low latency and jitter connection would be needed for the Varrier Display system (VE). For a force feedback haptic device communicating between the patient and the therapist, a low network bandwidth could be used but the latency and jitter need to be low. Response behaviors in the virtual environment After all possible consideration of how to best construct the virtual system, the next concern is how to associate the complex stimuli with the behavior of interest. The relative influence of particular scene characteristics, namely field of view (FOV), scene resolution, and scene content, are critical to our understanding of the effects of the VE on our response behaviors [ 19 ] and the effect of these characteristics on postural stability in an immersive environment has been examined [ 20 ]. Roll oscillations of the visual scene were presented at a low frequency – 0.05 Hz to 10 healthy adult subjects. The peak angular velocity of the scene was approximately 70°/sec. Three different scenes (600 dpi fountain scene, 600 dpi simple scene, and 256 dpi fountain scene) were presented at 6 different FOVs (+/-15°, 30°, 45°, 60°, 75°, 90° from the center of the visual field) counterbalanced across subjects. Subjects stood on a force platform, one foot in front of the other, with their arms crossed behind their backs. Data collected for each trial included stance break (yes, no), latency to stance break (10 sec maximum), subjective difficulty rating (difficulty in maintaining the Romberg stance, 1–10 scale), and dispersion of center-of-balance. Postural stability was found to vary as a function of display FOV, resolution, and scene content. Subjects exhibited more balance disturbance with increasing FOVs, higher resolutions and more complex scene contents. Thus, altered scene contents, levels of interactivity, and resolution in immersive environments will interact with the FOV in creating a postural disturbance. Expectation of the visual scene characteristics will also influence responses in a VE. When subjects had some knowledge of the characteristics of a forthcoming visual displacement most reduced their postural readjustments, even when they did not exert active control over the visual motion [ 21 ]. Thus we can hypothesize that visual stimuli present an optimal pathway for central control of postural orientation as there are many cues in the visual flow field that can identified for anticipatory processing. The important parameters of the visual field on posture can be extracted from several studies. Vestibular deficient individuals who were able to stabilize sway when fixating on a stationary light [ 22 ] became unstable when an optokinetic stimulus was introduced, implying that velocity information from peripheral vision was a cause of instability. Focusing upon distant visual objects in the environment increased postural stability [ 23 , 24 ]. We have observed in the VE [ 25 , 26 ] that small physical motions combined with large visual stimuli trigger a perception of large physical movements as occurs during flight simulations [ 27 ] and gaming. We have also observed measurable increases in the variability of head and trunk coordination and increased lateral head and trunk motion when standing quietly and walking within a dynamic visual environment [ 28 ]. The challenge is to determine whether the subject has become immersed in the environment, i.e., has established a sense of presence in the environment (see paper by Riva in this issue), and then to establish the correlation between the stimulus and response properties. The experience within the VE is multimodal, requiring participation of all sensory pathways as well as anticipatory processing and higher order decision making. Consequently, it is difficult to attribute resultant behaviors to any single event in the environment and responses across participants may be very variable. We have united an immersive dynamic virtual environment with motion of a posture platform [ 25 ] to record biomechanical and physiological responses to combined visual, vestibular, and proprioceptive inputs in order to determine the relative weighting of physical and visual stimuli on the postural responses. Methods In our laboratory, a linear accelerator (sled) that could be translated in the anterior-posterior direction was controlled by D/A outputs from an on-line PC. The sled was placed 40 cm in front of a screen on which a virtual image was projected via a stereo-capable projector (Electrohome Marquis 8500) mounted behind the back-projection screen. The wall in our system consisted of back projection material measuring 1.2 m × 1.6 m. An Electrohome Marquis 8500 projector throws a full-color stereo workstation field (1024 × 768 stereo) at 200 Hz [maximum] onto the screen. A dual Pentum IV PC with a nVidia 900 graphics card created the imagery projected onto the wall. The field sequential stereo images generated by the PC were separated into right and left eye images using liquid crystal stereo shutter glasses worn by the subject (Crystal Eyes, StereoGraphics Inc.). The shutter glasses limited the subject's horizontal FOV to 100° of binocular vision and 55° for the vertical direction. The correct perspective and stereo projections for the scene were computed using values for the current orientation of the head supplied by a position sensor (Flock of Birds, Ascension Inc.) attached to the stereo shutter glasses (head). Consequently, virtual objects retained their true perspective and position in space regardless of the subjects' movement. The total display system latency from the time a subject moved to the time the new stereo image was displayed in the environment was 20–35 ms. The stereo update rate of the scene (how quickly a new image is generated by the graphics computer in the frame buffer) was 60 stereo frames/sec. Flock of birds data was sampled at 120 Hz. Scene Characteristics The scene consisted of a room containing round columns with patterned rugs and painted ceiling (Fig. 2 ). The columns were 6.1 m apart and rose 6.1 m off the floor to the ceiling. The rug patterns were texture mapped on the floor and consisted of 10 different patterns. The interior of the room measured 30.5 m wide by 6.1 m high by 30.5 m deep. The subject was placed in the center of the room between two rows of columns. Since the sled was 64.8 cm above the laboratory floor the image of the virtual room was adjusted so that its height matched the sled height (i.e., the virtual floor and the top of the sled were coincident). Beyond the virtual room was a landscape consisting of mountains, meadows, sky and clouds. The floor was the distance from the subject's eyes to the virtual floor and the nearest column was 4.6 m away. The resolution of the image was 7.4 min of arc per pixel when the subject was 40 cm from the screen. The view from the subjects' position was that objects in the room were both in front of and behind the screen. When the scene moved in fore-aft, objects moved in and out of view depending on their position in the scene. Figure 2 An illustration of the virtual environment image in our laboratory. Procedures Subjects gave informed consent according to the guidelines of the Institutional Review Board of Northwestern University Medical School to participate in this study. Subjects had no history of central or peripheral neurological disorders or problems related to movements of the spinal column (e.g., significant arthritis or musculoskeletal abnormalities) and a minimum of 20/40 corrected vision. All subjects were naive to the VE. We have tested 7 healthy young adults (aged 25–38 yrs) standing on the force platform (sled) with their hands crossed over their chest and their feet together in front of a screen on which a virtual image was projected. Either the support surface translated ± 15.7 cm/sec (± 10 cm displacement) in the a-p direction at 0.25 Hz, or the scene moved ± 3.8 m/sec (± 6.1 m displacement) fore-aft at 0.1 Hz, or both were translated at the same time for 205 sec. Trials were randomized for order. In all trials, 20 sec of data was collected before scene or sled motion began (pre-perturbation period). When only the sled was translated, the visual scene was visible but stationary, thus providing appropriate visual feedback equivalent to a stationary environment. Data Collection and Analysis Three-dimensional kinematic data from the head, trunk, and lower limb were collected at 120 Hz using video motion analysis (Optotrak, Northern Digital Inc., Ontario, Canada). Infrared markers placed near the lower border of the left eye socket and the external auditory meatus of the ear (corresponding to the relative axis of rotation between the head and the upper part of the cervical spine) were used to define the Frankfort plane and to calculate head position. Other markers were placed on the back of the neck at the level of C7, the left greater trochanter, the left lateral femoral condyle, the left lateral malleolus, and on the translated surface. Markers placed at C7 and the greater trocanter were used to calculate trunk position, and shank position was the calculated from the markers on the lateral femoral condyle and the lateral malleolus. For trials where the sled moved, sled motion was subtracted from the linear motion of each segment prior to calculating segmental motion. Motion of the three segments was presented as relative segmental angles where motion of the trunk was removed from motion of the head to determine the motion of the head with respect to the trunk. Motion of the shank was removed from motion of the trunk to reveal motion of the trunk with respect to the shank. Motion of the shank was calculated with respect to the sled. Results The response to visual information was strongly potentiated by the presence of physical motion. Either stimulus alone produced marginal responses in most subjects. When combined, the response to visual stimulation was dramatically enhanced (Fig. 3 ), perhaps because the visual inputs were incongruent with those of the physical motion. Figure 3 Relative angles of the head to trunk (blue), trunk to shank (red) and shank to sled (green) are plotted for a 60 sec period of the trial during sled motion only, scene motion only, and combined sled and scene motion (the same data are plotted against both the sled and the scene). Using Principal Component Analysis we have determined the overall weighting of the input variables. In healthy young adults, some subjects consistently responded more robustly when receiving a single input, suggesting a proprioceptive (see S3 in Fig. 4 ) or visual (S1 in Fig. 4 ) dominance. With multiple inputs, most subjects produced fluctuating behaviors so that their response was divided between both inputs. The relative weighting of each input fluctuated across a trial. When the contribution of each body segment to the overall response strategy was calculated, movement was observed primarily in the trunk and shank. Figure 4 Overall weighting of the input variables derived from the PCA for 3 subjects. The first 3 bars (blue) represent a subsequent non-overlapping 40 sec time period to sled motion only. The next 3 bars (red) represent non-overlapping 40 sec time periods to scene motion only. The last 6 bars represent non-overlapping 40 sec time periods to both sled (blue) and scene (red) motion. The direction of each bar indicates the relative phase between the response and the input signal. Discussion Results from experiments in our laboratory using this sophisticated technology revealed a non-additive effect in the energy of the response with combined inputs. With single inputs, some subjects consistently selected a single segmental strategy. With multiple inputs, most produced fluctuating behaviors. Thus, individual perception of the sensory structure was a significant component of the postural response in the VE. By quantifying the relative sensory weighting of each individual's behavior in the VE, we should be better able to design individualized treatment plans to match their particular motor learning style. Developing treatment interventions in the virtual environment should carry over into the physical world so that functional independence will be increased for many individuals with physical limitations. In fact, there is evidence that the knowledge and skills acquired by disabled individuals in simulated environments can transfer to the real world [ 29 - 31 ]. The ability for us to use this technology outside the area of research labs and bring these systems to clinics is just starting. However, the cost is high and the applications that can best be applied to rehabilitation are limited. The cost of such systems might be mitigated if this technology allowed therapists and patients to interact more frequently and/or resulted in better patient outcomes. Such issues are under study now at several institutions. This brings us to the idea of tele-rehabilitation, which would allow therapy to transcend the physical boundaries of the clinic and go wherever the communication system and the technology would allow [ 5 ]. For example, at some location remote from the clinic a patient enters a VE suitable for rehabilitation protocols connected to the clinic and a therapist. While this idea is not new, the kind of therapies that could be applied under such a condition is limited by the communication connection and facilities at both ends of the communication cable. The ability to provide rehabilitation services to locations outside the clinic will be an important option for clinicians and patients in the near future. Effective therapy may best be supplied by the use of high technology systems such as VE and video, coupled to robots, and linked between locations by high-speed, low-latency, high-bandwidth networks. The use of data mining software would help analyze the incoming data to provide both the patient and the therapist with evaluation of the current treatment and modifications needed for future therapies. Conclusions The ability to provide rehabilitation services to locations outside the clinic is emerging as an important option for clinicians and patients. Effective therapy may best be supplied by the use of high technology systems such as VE and video, coupled to robots, and linked between locations by high-speed, low-latency, high-bandwidth networks. The use of data mining software would help analyze the incoming data to provide both the patient and the therapist with evaluation of the current treatment and modifications needed for future therapies. Although responses in the VE can vary significantly between individuals, these results can actually be used to benefit patients through the development of individualized treatments programs that will raise the level of successful rehabilitative outcomes. Further funding for research in this area will be needed to answer the questions that arise from the use of these technologies.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546412.xml
509311
Regenerating Lost Muscle: Msx1 to the Rescue
null
Cell and molecular biologists have a good start at understanding the adult salamander's enviable ability to completely regrow a lost limb or jaw. (Salamanders can even regenerate portions of their eyes and heart.) Happily, mammals share many of the required cellular skills—though in an untapped form. In response to injury, fully differentiated salamander muscle cells can produce less specialized cells that are capable of multiplying and recreating lost tissue. In this month's PLoS Biology , biochemist Jeremy Brockes and his colleagues at University College London tie this muscle cell de-differentiation to expression of a single gene, Msx1 , a player in limb development and regeneration across species. An adult salamander The protein encoded by Msx1 appears at the borders of maturing tissues in salamanders, chickens, and mice, and represses differentiation of muscle cells (called myofibers) during development. Msx1 promotes amphibian limb and tail regeneration, zebrafish fin repair, and regrowth of the tips of digits in mice. Here, Anoop Kumar et al. clarify the gene's role in muscle repair by demonstrating that Msx1 expression is required for de-differentiation of salamander myofibers—mature muscle tissue without the ability to produce new cells—into cells with single nuclei (mononucleate cells) that are able to multiply. Large, elongated cells, myofibers are distinguished by a striped pattern of actin and myosin, the proteins that produce muscle contractions. They develop from precursor cells, which proliferate and fuse together into multinucleate intermediates called myotubes. As myotubes bulk up on actin and myosin, pushing their many nuclei to the cell periphery, they mature into myofibers. Once differentiated, myofibers are committed to the life of a muscle workhorse. They cannot divide; mature muscle in mammals adds strength and repairs injury by accumulating more actin and myosin, and by fusing single-nucleus “satellite” cells into new or existing myofibers. Unlike their mammalian counterparts, salamander and newt myofibers respond to injury by splitting into several multinucleated fragments, or by budding off several of their nuclei to create individual cells. These mononucleate progeny multiply and develop to replace lost tissue. Kumar et al. found that more than half of salamander myofibers spontaneously fragment and bud when dissected and placed in cell culture. Prior study has shown that mouse myofibers don't normally behave this way, although cultured mouse myotubes can imitate amphibian myofibers if given a push, in the form of Msx1 . A fraction of mouse myotubes made to express Msx1 break off mononucleate cells that can be induced to express markers for bone, cartilage, fat, or muscle. In the salamander cells, Msx1 RNA and protein appeared in actively fragmenting and budding myofibers, especially in and around their nuclei. When Kumar et al. blocked the translation of Msx1 mRNA, the myofibers did not generate new cells. The results confirm Msx1 as a pivotal regulator of muscle cell re-entry into the cycle of cell division and tissue growth. In this context, it doesn't drive cell division—on the contrary, the authors showed that Msx1 induced cell splitting without DNA replication. The gene drove committed muscle cells to donate nuclei, creating a pool of new cells to divide, differentiate, and repair damage. Given what it can do for mouse cells in culture, the question remains whether Msx1 might help awaken a latent capacity for regeneration in living mammals.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509311.xml
554770
Developmental changes in abundance of the VSPβ protein following nuclear transformation of maize with the Soybean vspβ cDNA
Background Developing monocots that accumulate more vegetative tissue protein is one strategy for improving nitrogen-sequestration and nutritive value of forage and silage crops. In soybeans (a dicotyledonous legume), the vspA and B genes encode subunits of a dimeric vegetative storage protein that plays an important role in nitrogen storage in vegetative tissues. Similar genes are found in monocots; however, they do not accumulate in leaves as storage proteins, and the ability of monocot leaves to support accumulation of an ectopically expressed soybean VSP is in question. To test this, transgenic maize ( Zea Mays L. Hi-II hybrid) lines were created expressing soybean vspB from a maize ubiquitin Ubi-1 promoter. Results From 81 bombardments, 101 plants were regenerated, and plants from five independent lines produced vspB transcripts and VSPβ polypeptides. In leaves from seven-week-old plants (prior to flowering), VSPβ accumulated to 0.5% of the soluble leaf protein in primary transgenic plants (R 0 ), but to only 0.03% in R 1 plants. During seed-filling (silage-stage) in R 1 plants, the VSPβ protein was no longer detected in leaves and stems despite continued presence of the vspB RNA. The RNA transcripts for this peptide either became less efficiently translated, or the VSPβ protein became unstable during seed-fill. Conclusion Developmental differences in the accumulation of soybean VSPβ when transgenically expressed in maize show that despite no changes in the vspB transcript level, VSPβ protein that is readily detected in leaves of preflowering plants, becomes undetectable as seeds begin to develop.
Background Although genetic variation for protein content has been found in forage plants, this variability is narrower than that observed for other traits such as digestibility [ 1 ]. Since the major protein components in monocot forage and silage crops are involved in metabolic activity, and hence are not "true" storage proteins, it has been argued that it is not feasible to make major changes in protein quality or protein composition by conventional breeding [ 1 ]. However, genetic engineering may allow improvement in protein quality and content through expression of a storage protein not found in grass vegetative tissue. Genes encoding seed storage proteins of various plant species have been transgenically expressed to test for improvement of nutritional quality. Most experiments were conducted with tobacco and legume species including alfalfa, soybean, canola, clover and lupins. For nuclear-targeted genes, accumulation of these seed storage proteins in vegetative tissue of transgenic plants was either undetectable or very low. These included pea vicilin [ 2 , 3 ], soybean conglycinin [ 4 ], sunflower seed agglutinin [ 5 , 6 ], and phaseolin [ 7 ]. The instability of seed proteins in non-seed tissues of transgenic plants was frequently attributed to protein targeting to protease-rich vacuoles in the vegetative cells, and subsequent degradation [ 5 , 7 , 8 ]. The greatest accumulation of a seed storage protein from a nuclear-targeted gene was achieved using zeins, a maize seed storage protein that is targeted to "protein bodies" directly from the endoplasmic reticulum (ER), thus avoiding the secretory route to the cellular vacuole. Transgenic expression from the CaMV 35S promoter in tobacco resulted in the formation of these protein bodies containing the zein within vegetative tissues [ 8 , 9 ]. Alternatively, "short-circuiting" the protein-targeting route by addition of an ER retention signal to the storage-protein coding region also increased protein accumulation up to 100× [ 5 , 6 , 10 , 11 ]. In many legumes, accumulation of specific vegetative storage proteins (VSPs) in leaves and stems is the main source of increased nitrogen content [ 12 - 14 ]. Use of VSPs instead of seed storage proteins to increase vegetative protein content in monocots may provide an advantage since they have evolved to function in vegetative cell types. Legume vegetative cells that accumulate VSP proteins contain multiple vacuole types and storage proteins are targeted to specific vacuoles where they are not rapidly degraded [ 15 ], and do not interfere with cellular metabolic processes. It is uncertain if monocots can produce similar vacuoles or successfully target a VSP to them. The most studied VSPs are the soybean VSPα and VSPβ proteins, which are lysine-rich glycosylated vacuolar proteins that accumulate abundantly in leaves, stems and pods, but not in seeds [ 12 - 14 ]. Recently, soybean vspA and/or vspB genes fused to the 35S promoter were expressed in transgenic tobacco to study their accumulation in a heterologous dicot plant. Nuclear targeted genes produced VSP ranging between 2 and 6% of the soluble protein in leaves of the transgenic plants [ 16 ]; whereas, targeting to both the chloroplast and the vacuole within the same plant resulted in VSP comprising greater than 10% soluble protein [ 17 , 18 ]. Soybean VSP is therefore an excellent candidate for use in transgenic improvement of plant protein status, particularly grasses that contain limited levels of lysine [ 16 ]. However, it remains to be determined if VSPs can be expressed and accumulated in monocot plants since storage protein stability is dependent on post-translational events that may differ between monocots and eudicots. This manuscript presents the transgenic expression of soybean VSPβ in the leaves of transgenic maize and discusses this expression in relation to the developmental stage of the plant. Results and Discussion Development of primary (R 0 ) transgenic maize expressing vspB Out of 101 plants, twenty, belonging to five independent lines (71-1, 45-1, 45-3, 44-1, and 4-1) were shown by Southern blot analysis to contain a 1.5 kb hybridizing band corresponding to the intact bar gene (Fig. 1 ). All 20 plants also contained the expected 1.9 kb band that hybridized to the vspB gene. The same probe detected two bands in Eco RI restricted Soybean genomic DNA, ~5.7 and 8.6 Kb, corresponding to the highly homologous genes vspA and vspB [ 13 ]. Western blot analysis was used to detect the VSPβ polypeptide in leaf extracts from fourteen primary (R 0 ) transgenic maize plants at vegetative stage (7 weeks old). A distinct VSPβ band was not visible in silver-stained SDS-PAGE separated maize extract proteins due to complexity of the total protein pattern and the relatively low level of VSPβ expression (Fig. 2a ). However, some of these plants expressed VSPβ at a level high enough to be detected by Western blot analysis (Fig. 2b ). Two plants (71-1-53 and 45-3-1) had the highest level of VSPβ, while four plants (71-1-23, 71-1-20, 45-1-4, and 45-1-7) accumulated a lower level of VSPβ. Although a faint lower MW peptide band was visible in negative control maize extracts, VSP was clearly only present in transgenic lines. Faint detection of bands in untransformed maize (none of identical size to soybean VSPβ) is consistent with previous reports of cross-hybridizing of soybean VSP antibodies with proteins from monocots [ 14 ]. In soybean extracts used as a positive control, the antibody for VSPβ recognized both VSPα and VSPβ polypeptides (Fig. 2b,c ) Computer analysis of digital images of the Western blots was used to detect differences in relative band intensity of the immunologically detected VSPβ peptide. Because the native soybean VSPs (VSPα plus VSPβ) were easily visible on total protein stained gels (Fig. 2a ), relative quantification of total stained proteins in the soybean samples indicated that the VSPs represented about 10% of the total soluble protein in young soybean leaves. This is close to the previously reported value of 15 % [ 12 ]. The intensity of the transgenic maize 45-3-1 VSPβ band on Western blots was 44% of the soybean VSP's band (digital image pixel quantification of Fig. 2b ). Accounting for the differences in total protein applied to the gel (less soybean total protein was loaded), the VSPβ protein was estimated to have accumulated to 0.5 % of the total soluble protein. This is similar to the highest level of seed storage protein accumulation observed with the ectopic expression of zein [ 8 , 9 ], but remains less than the 1% minimal expression level predicted by Wandelt et al. [ 11 ] to be needed to directly alter the nutritional quality of the leaves. Although the 0.5% of total soluble protein was too low to alter nutritional value, detection of VSPβ in 45-3-1 allowed monitoring of VSPβ level in leaves and stems during plant development. Presence of vspB in R 1 plants R 1 plants were produced by back-crossing the R 0 plants with Hi II control non-transformed pollen. Back-crossing was performed because R 0 plants directly regenerated from tissue culture did not have synchronized production of pollen and receptive female flowers. The R 1 families segregating for bar expression were analyzed for the presence of the vspB gene by Southern blot analysis. From 57 R 1 plants analyzed, Southern blot analysis showed that 35 (61%) contained the vspB gene integrated into the genome. Figure 3 shows examples for several R 1 from five R 0 parental lines. The ~1.9 kb vspB band can be seen in nine of the 16 R 1 lines. There are also fainter bands one slightly larger than the 1.9 kb band and one or two migrating between 4 and 5 kb. These are often observed as incomplete restriction of all Eco RI sites internal to plasmid DNA that is integrated into the maize genome. Similar bands are observed even with the plasmid control. Total RNA was isolated from leaves of young R 1 plants at the vegetative stage (7 weeks after planting). Eighteen transgenic R 1 plants, including the ones originating from R 0 plants representing different levels of VSPβ accumulation (high VSPβ accumulators: 45-3-1, 71-1-53; mid-level VSPβ accumulators: and 71-1-23, 71-1-20, 45-1-4, 45-1-7 and low VSPβ accumulators: 44-1-1- 4-1-2), were analyzed for vspB transcripts. Because of high sequence similarity (85%) between vspA and vspB cDNAs, the same probe hybridized in soybean with both mRNAs as demonstrated by Staswick [ 13 ]. The transgenic plants 71-1-53A, 71-1-20A, 71-1-20E, and 71-1-20I produced a hybridizing band of approximately the expected size of 1.1 kb indicating transgene expression (Fig. 4a ); however, most did not. Use of the Ubi -1 promoter for both the vspB and bar gene probably led to a high number of transgenics with the vspB silenced, and a higher level of VSPβ accumulation will likely result in future work using a combination of different promoters. Immunodetection of VSPβ protein in transgenic young plant leaves showed variation in accumulation in comparisons between the parental (R 0 ) and their R 1 progeny, with the greatest variation observed with the highest VSPβ expressing R 0 plants (Fig 4b ). The parental line 71-1-53 had a relatively high level of expression of VSPβ, but the only transgenic offspring from this line, 71-1-53A, had almost undetectable VSPβ; in contrast, a low level accumulator, 71-1-20, produced R 1 lines expressing different levels of VSPβ, although none of them expressed at higher levels than the parent. Quantification of the relative level of VSPβ expressed in the R 1 leaves showed that the highest level measured was only 0.03% of total soluble protein. Despite the overall low level of VSPβ expression, for the purposes of this work, VSPβ accumulation in several of the plants was high enough to study the relationship of plant developmental stage and vspB /VSPβ accumulation. Both the transcript abundance and VSPβ protein accumulation were determined in the R 1 lines using real-time quantitative RT-PCR and Western blot immunodetection, respectively. Real-time RT-PCR is more sensitive that Northern blot analysis and was able to detect transcripts that were not seen with standard total RNA blotting methods. The vspB transcript was quantified in leaves from immature plants (prior to tassel formation) and silage stage plants (plants with developing seeds at the 18 DAP-days after pollination stage), as well as, stems from the silage stage plants, (Figure 5 ). The vspB transcript was detected in all five transgenics (four of which had RNA not detectable using standard Northern blot methods). The relative level of RNA among the different plant samples was not consistent across the different lines with some having more transcripts in the young leaves while others had more in the older leaves and stems. When vspB RNA abundance was compared to VSPβ protein accumulation there was no correlation between the two (Fig. 5a, b, c ). Antiserum to VSPβ detected VSPβ only in leaves from young maize plants that had not yet flowered. No VSPβ was detected in leaves and stems of silage stage corn that had developing seed. The soybean VSPβ peptide was the primary band reacting with the antiVSPβ antiserum in young leaves of transgenic maize, however, in silage stage stems and to a lesser extent the silage stage leaves, there were multiple bands detected at a different size than the VSPβ. These were also detected in the non-transgenic control plant samples and therefore, were not due to different modifications of the soybean VSPβ. Although the VSPβ protein dropped below detectable levels in the silage stage 71-1-20A, 45-3-1F and 45-3-1-G plants, the vspβ transcript was still detectable. In fact, in the 71-1-20A plants the transcript level was highest in the silage stage leaves that had no detectable VSPβ. Therefore, post-transcriptional events (i.e., changes in either the translational efficiency of the vspB transcript or the protein stability of VSPβ) were altered in the silage stage leaves and stems as compared to the leaves of immature plants. Conclusion The vspB gene was successfully introduced into R 0 regenerated maize and transferred to the R 1 progeny, of which vspB transcript and VSPβ protein were detected and studied. This is the first report on introduction and expression of a legume vegetative storage protein in a monocot plant. The inability to detect VSPβ from the maize vegetative tissue at the time of seed development, even when the vspB transcript was still expressed, must have arisen from either reduced translational efficiency of the vspB transcript or a decrease in the stability of the VSPβ protein. The reduction of seed storage protein level in leaves of transgenic eudicots was also observed with expression of vicillin in alfalfa [ 11 ] and tobacco [ 2 ]; however, it was not observed with the expression of VSPα in tobacco [ 16 ] or ovalbumin in alfalfa [ 30 ]. These data suggest that factors controlling developmental change in vegetative tissue protein accumulation are a combination of host plant traits and innate characteristics of the ectopically expressed protein. It is interesting to speculate that if, in maize (a monocot), soybean VSPβ was degraded in a manner that provided amino acids that were translocated to the seed to support seed development, then development of high level VSP expressing monocots may be a way to improve nitrogen content of the seed/grain produced by the plant. Methods Plant material and tissue culture methods Plants of hybrid "Hi-II" maize were established in a greenhouse and immature tassels were used for embryogenic type II callus production, as described by Armstrong [ 19 ]. Transgenic plant development Microprojectile bombardment of callus was performed using the procedure of Somers et al. [ 20 ]. Calli were cobombarded with equal amounts of pRSVP-1 (Shatters Jr., unpublished) and pAHC25 [ 21 ]. Plasmid pRSVP-1 was constructed by restricting the soybean vspB cDNA clone (998 bp) from pKSH3 [ 22 ] with Eco RI, blunt ends of this fragment were produced using S1 nuclease and the fragment was cloned into similarly blunt ended Bam HI restricted pAHC17 [ 21 ]. As a result, the vspB coding region was inserted downstream of the Ubi-1 promoter and a 5' untranslated region (exon) and intron; and upstream of the nos terminator sequence. The plasmid pAHC25 carried the bar gene and the uidA reporter gene, both under the control of the Ubi-1 promoter. Bombardments were performed with a Biolistic ® PDS-1000/He Particle Delivery System (Bio-Rad Laboratories, Hercules, CA) and an osmotic treatment was applied to reduce the cell damage caused by the gene transfer method [ 23 ]. Putative transgenic maize were regenerated from glufosinate resistant callus as described by Armstrong [ 19 ], and grown in five-gallon pots containing sterile sand and Metromix-350 (1:1). Plants were fertilized weekly with Peter's 20-20-20 with micronutrients (Division of United Industry Corp., St. Louis, MO). Southern blot analysis One gram of frozen young leaf tissue was ground in liquid nitrogen and genomic DNA was extracted using the Dellaporta procedure [ 24 ]. Twenty micrograms of genomic DNA were digested with Eco RI, which released a 1.9 Kb fragment containing the vspB gene, the nos terminator and part of the Ubi-1 promoter. DNA was separated on a 0.8% agarose gel, blotted onto Hybond N + membrane (Amersham Pharmacia Biotech, Inc. Piscataway, NJ) by capillary blotting [ 25 ], and UV cross-linked. The non-radioactive digoxigenin system (Roche Molecular Biochemicals, Indianapolis, IN) was used for labeling and detection of the transgene. Blotted DNA was probed with either a 611 or a 843 bp of vspB gene segment amplified from pRSVP-1 and gel purified. The forward and reverse primers 5'-GTTCTTCGGAG GTAAAAT-3' and 5'-TTCGCCTCTGTGGT-3' were used, respectively, to amplify a 611 bp segment, and the primer pair 5'-GCAGGCTACCAAAGGT-3' and 5'-TAGGTGACTTACCCACAT-3' was used to amplified the product of 843 bp. For identification of bar transgenic plants, the DNA was digested with Eco RI, which released a fragment of ~1.5 Kb that contained part of the Ubi-1 promoter, the bar gene, and the nos terminator, and was identified with a 419 bp digoxigenin labeled probe produced by PCR amplification of pAHC25 using the forward and reverse primers: 5'-GGCGGTCTGCACCATCGT-3' and 5'-GCCAAGTTCCCGTGCTTGA-3', respectively. Northern blot analysis Total RNA was isolated from 2 g of tissue using acid guanidine isothiocyanate-phenol-chlorophorm extraction [ 26 ], resuspended in T 10 E 1 and treated with 2 μl RNasin ® (4U/μl) RNAse inhibitor (Promega, Madison, WI) and stored at -70°C until use. Thirty micrograms of total RNA were separated on 1.2% agarose formaldehyde gels and transferred to Hybond N + membrane by capillary blotting overnight or by pressure blotting for one hour using a PosiBlot ® 30-30 pressure blotter (Stratagene, La Jolla, CA) according to the manufacture's instructions. Membranes were UV-crosslinked, and probed with the same probes as described for Southern blot analysis. Chemilluminescence was captured using a Kodak Digital Science Image Station 440 CF (Eastman Kodak Company, Rochester, NY). Real-Time RT-PCR Total RNA extractions for real-time RT-PCR were performed using 500 mg of tissue ground to fine powder using mortar and pestle in the presence of liquid nitrogen, then processed with the RNeasy midiprep Kit (Qiagen, Germany), following the manufacturer's protocol. Trace DNA contamination was removed from total RNA by a combination of acid phenol: chloroform 5:1 pH= 4.7 extraction and Dnase I treatment (Ambion, Texas). Real-time RT-PCR was performed on a Rotor-Gene RG-3000 (Corbett Research, Australia) using the Quantitect SYBR Green real-time RT-PCR kit (Qiagen, Germany), and the manufacturers protocols with 300 ng of Dnase I treated total RNA. Primers were designed to amplify a 108 bp fragment of the soybean vspB using the following primers: 5'-TGGTTCAACGCACTCTTC-3' and 5'-GGCTATGGTGAGCGTTCTTC-3'. Reverse transcription was performed for 30 min at 50°C followed by a 15 min denaturing at 95°C, and 40 cycles of 40 s at 95°C, 40 s at 58°C and 40 s at 72°C. Quantification was based on relative abundance to maize 18S RNA by amplifying a 174 bp fragment with primers: 5'-CCTGCGGCTTAATTGACTC-3' and 5'-GTTAGCAGGCTGAGGTCTCG-3', and using the comparative quantification function of the Rotor-Gene RG-3000 software. All real-time RT-PCR experiments were conducted in triplicate and on triplicate RNA preparations for each sample. Melting curve analysis and agarose gel electrophoresis were performed to verify single product formation. Western blot analysis Protein was extracted from 100 mg of leaves and stems with 0.5 ml of phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 ) supplemented with 1 tablet/10 ml buffer of the Complete-Mini protease inhibitor cocktail (Roche Molecular Biochemicals, Indianapolis, IN) by homogenizing in the presence of zirconia/silica 0.1 mm dia. beads. Centrifuged extract supernatants were removed and used for protein concentration determination [ 27 ]. Because of the low protein yields in stem extracts, they were precipitated with 10% trichloroacetic acid (TCA), washed with ice-cold acetone and resuspended prior to SDS-PAGE analysis. Thirty micrograms of protein were separated on 12% SDS-polyacrylamide gels. Proteins in the gel were either stained with silver nitrate [ 28 ] or transferred to Hybond-P (PVDF) membranes (Amersham Pharmacia Biotech, Inc. Piscataway, NJ) using Trans-Blot SD Semi-Dry Transfer Cell blotter and recommended protocols (Bio-Rad Laboratories). Soybean VSP was immunologically detected on the PDF membranes with anti VSPβ serum (provided by P. Staswick, University of Nebraska, produced as previously described [ 29 ] used at a 1:5,000 dilution. Detection was performed using a luminol substrate and the NEN Life Science Products, Inc. (Boston, MA) Reinascence kit. Chemiluminescent signal was captured by a Kodak Digital Science Image Station 440 Cf, and analyzed with Kodak 1D Scientific Image Software. Quantification of band intensity was always compared relative to samples from the same gel. List of abbreviations RT-PCR, reverse transcriptase-polymerase chain reaction; VSP, vegetative storage protein Authors' contributions MFG participated in experimental design, carried out the transgenic plant development, plant crosses, and molecular blotting/detection methods, and participated in manuscript draft preparation. RLS participated in experimental design, and provided guidance and training in development of transgenic maize. CM performed RT-PCR experiments. BTS participated in experimental design and provided expertise and training in plant crosses. RGS conceived of the study, participated in experimental design, coordinated the experimental plan, and wrote the draft manuscript. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554770.xml
521489
Transformation and scattering activities of the receptor tyrosine kinase RON/Stk in rodent fibroblasts and lack of regulation by the jaagsiekte sheep retrovirus receptor, Hyal2
Background The envelope (Env) protein of jaagsiekte sheep retrovirus (JSRV) can transform cells in culture and is likely to be the main factor responsible for lung cancer induction by JSRV in animals. A recent report indicates that the epithelial-cell transforming activity of JSRV Env depends on activation of the cell-surface receptor tyrosine kinase Mst1r (called RON for the human and Stk for the rodent orthologs). In the immortalized line of human epithelial cells used (BEAS-2B cells), the virus receptor Hyal2 was found to bind to and suppress the activity of RON. When Env was expressed it bound to Hyal2 causing its degradation, release of RON activity from Hyal2 suppression, and activation of pathways resulting in cell transformation. Methods Due to difficulty with reproducibility of the transformation assay in BEAS-2B cells, we have used more tractable rodent fibroblast models to further study Hyal2 modulation of RON/Stk transforming activity and potential effects of Hyal2 on RON/Stk activation by its natural ligand, macrophage stimulating protein (MSP). Results We did not detect transformation of NIH 3T3 cells by plasmids expressing RON or Stk, but did detect transformation of 208F rat fibroblasts by these plasmids at a very low rate. We were able to isolate 208F cell clones that expressed RON or Stk and that showed changes in morphology indicative of transformation. The parental 208F cells did not respond to MSP but 208F cells expressing RON or Stk showed obvious increases in scattering/transformation in response to MSP. Human Hyal2 had no effect on the basal or MSP-induced phenotypes of RON-expressing 208F cells, and human, mouse or rat Hyal2 had no effect on the basal or MSP-induced phenotypes of Stk-expressing 208F cells. Conclusions We have shown that RON or Stk expression in 208F rat fibroblasts results in a transformed phenotype that is enhanced by addition of the natural ligand for these proteins, MSP. Hyal2 does not directly modulate the basal or MSP-induced RON/Stk activity, although it is possible that adaptor proteins might mediate such signaling in other cell types.
Background JSRV is an acutely oncogenic retrovirus that can induce lung tumors in newborn sheep in as little as 10 days [ 1 ]. Unlike most acutely oncogenic retroviruses, JSRV does not carry a host cell-derived oncogene. Instead, expression of the native viral Env protein is sufficient to transform cultured cells and is likely to be the active oncogene in animals [ 2 - 5 ]. The only other examples of such retroviruses are enzootic nasal tumor virus, a close relative of JSRV that encodes a similar transforming Env protein [ 6 , 7 ]; avian hemangioma virus [ 8 ], which expresses an Env protein that can induce cell proliferation in cultured cells [ 9 ]; and spleen focus-forming virus, a replication-defective virus that expresses a recombinant nonfunctional Env protein that has been shown to induce proliferation in cultured cells and in animals by activating the erythropoietin receptor and a short form of the transmembrane receptor tyrosine kinase Stk [ 10 , 11 ]. At least two mechanisms of transformation by JSRV Env have been identified in cultured cells. JSRV Env can transform immortalized mouse, rat, and chicken fibroblasts [ 2 - 4 ], and the cytoplasmic tail of Env is absolutely required for transformation [ 4 , 12 - 14 ]. Interaction of the extracellular domain of Env with the virus receptor Hyal2 appears to play no role in transformation of the rodent fibroblasts since JSRV Env does not bind to mouse Hyal2 [ 15 ], and deletion of the Hyal2-binding domain of Env did not abrogate transformation of rat fibroblasts [ 16 ]. Although there is some controversy regarding the mechanism of transformation, most studies indicate a role for the phosphatidylinositol 3-kinase (PI3K)/Akt pathway in transformation of the rodent fibroblasts [ 7 , 12 , 13 , 17 ]. In contrast, this pathway appears to play a smaller role in the transformation of chicken fibroblasts, but other more important pathways have yet to be identified [ 4 , 14 ]. A second pathway for JSRV Env transformation involves activation of the transmembrane receptor tyrosine kinase Mst1r (herein called RON for the human and Stk for the rodent forms [ 18 , 19 ]). RON/Stk activation is associated with cell scattering, cell transformation, and oncogenesis in animals [ 20 ]. RON/Stk activity is induced by binding of its only known ligand, macrophage stimulating protein (MSP) [ 21 , 22 ]. Recent evidence indicates that the JSRV receptor Hyal2 is also involved in regulating RON activity in an immortal human epithelial cell line (BEAS-2B) that naturally expresses Hyal2 and RON [ 5 ]. In its normal inactive state, RON is bound to Hyal2, and when Hyal2 binding is prevented by JSRV Env, RON becomes constitutively active in the absence of MSP stimulation. Introduction of a dominant negative form of RON into these cells blocked transformation, indicating that main transformation pathway is through RON and that JSRV might cause cancer through a novel mechanism involving RON activation [ 5 ]. Thus it appears that JSRV Env can transform cells by at least two distinct mechanisms. Furthermore, the dependence of epithelial cell transformation on activation of the RON pathway, in contrast to fibroblasts, which do not express RON and are transformed through a different pathway, indicated that the RON pathway might be more important in oncogenesis in animals because epithelial cells are the natural target for JSRV-induced cancer in sheep. We have had difficulty in further exploring transformation through the RON pathway in BEAS-2B cells due to difficulty with reproducibility of the transformation assay. Others have found that Stk can transform immortalized rodent fibroblasts [ 23 , 24 ], and we hypothesized that if Hyal2 does downregulate RON/Stk activity, we might be able to suppress transformation by RON/Stk by overexpressing Hyal2. In addition, we wanted to explore whether Hyal2 might modulate MSP induction of RON/Stk activity. Here, we have been able to generate 208F rat fibroblasts that express either RON or Stk and find that these cells exhibit a transformed phenotype in comparison to the parental cells. Treatment of the cells with MSP induced a dose-dependent increase in scattering/transformation. Using a variety of approaches we have been unable to detect an effect of Hyal2 on the constitutive or MSP-induced phenotypes of 208F cells expressing RON or Stk. These results argue against a direct role for Hyal2 in the regulation of RON/Stk activity. Methods Cell culture Cell lines used here include 208F Fischer rat embryo fibroblasts [ 25 ], a morphologically flat subclone of NIH 3T3 Swiss mouse embryo fibroblasts suitable for transformation studies (gift from Maxine Linial, Fred Hutchinson Cancer Research Center, Seattle; originally from Doug Lowy, National Cancer Institute, Bethesda), PJ4/LAPSN packaging cells [ 26 ] that produce the LAPSN retroviral vector [ 27 ] with a JSRV pseudotype, and PT67 10A1-MLV-pseudotype retrovirus packaging cells [ 28 ]. Unless otherwise stated, cells were grown in Dulbecco's modified Eagle medium (DMEM) with high glucose (4.5 g/L) and 10% fetal bovine serum at 37°C in a 10% CO 2 -air atmosphere at 100% relative humidity. Expression plasmids Mouse Stk was expressed by using a plasmid (gift from Sandra Ruscetti, National Cancer Institute, Frederick, Maryland) that contains the mouse Stk cDNA cloned into the Not I site of the pAlter-Max expression plasmid (Promega, Madison, Wisconsin). Human RON was expressed by using a plasmid (gift from Michael Lerman, National Cancer Institute, Frederick, Maryland) that contains the human RON cDNA inserted into the pCI-neo expression plasmid (Promega). The expression cassette in both of these plasmids is identical and consists of a human cytomegalovirus immediate early promoter followed by splicing signals, the cDNA, and the simian virus 40 late polyadenylation signal. Mouse Stk was also expressed using the previously described pcDNA3-based expression vector [ 23 ] (gift from Susan Waltz, University of Cincinnati, Ohio), which contains the human cytomegalovirus early promoter, no splice signals, and the bovine growth hormone polyadenylation signal. Human RON was also expressed using the previously described retroviral vector expression plasmid pMSCVpuroRON [ 5 ], in which the RON cDNA is expressed from the Moloney murine leukemia virus promoter and the puromycin resistance gene is expressed using a phosphoglycerate kinase promoter. In this vector both spliced and unspliced mRNAs are made that encode RON. Retroviral vectors The LAPSN retroviral vector encodes human placental alkaline phosphatase (AP) and neomycin phosphotransferase (Neo). LAPSN virus was made using PT67 (10A1-MLV pseudotype) or PJ4 (JSRV pseudotype) packaging cells. Retroviral vectors encoding human Hyal1, human Hyal2, mouse Hyal2 cloned from NIH 3T3 cells, and rat Hyal2 cloned from 208F rat cells (GenBank accession numbers U03056, U09577, AF535140, and AF535141, respectively) were made by insertion of the cDNAs into the LXSN vector [ 29 ] and by generation of stable PT67 retrovirus packaging cell lines that produce the vectors as described [ 3 , 15 , 29 ]. The human Hyal1 and human Hyal2 vectors present in the packaging cell clones that were used are known to be functional based on phenotypic assays of the proteins made in cells transduced with the vectors [ 3 ]. Two independent packaging cell clones that produce the rat Hyal2 vector and two independent clones that produce the mouse Hyal2 vector were used, and the correct sequence of the cDNAs in the integrated vectors in these packaging cells was confirmed by PCR amplification of the sequences followed by complete sequencing of the cDNAs. Cell transformation assay NIH 3T3 and 208F cells were transfected using calcium phosphate as previously described [ 15 ] and were monitored for appearance of transformed foci of cells for up to a month after transfection. Cellular response to MSP Recombinant human MSP activated by treatment with kallikrein (R&D Systems, Minneapolis, MN) was dissolved at a concentration of 10 μg/ml in phosphate buffered saline with calcium and magnesium containing 0.2% bovine serum albumin and was stored at -70°C. Cells were seeded at 10,000 per well (d = 16 mm) of 24-well plates. Two days after seeding, the cells were treated with 0.06 to 20 μl of MSP stock in 0.5 ml of fresh medium (MSP concentration of 1.2 to 400 ng/ml) and cellular morphology was evaluated the day after MSP addition. Treatment of cells with 20 μl buffer containing 0.2% bovine serum albumin in 0.5 ml of fresh medium had no effect on cell morphology. Results RON and Stk exhibit low transforming activity in rodent fibroblasts We tested plasmids encoding RON or Stk (pCIneoRON and pAlter-Max-Stk, respectively) for transforming activity in NIH 3T3 mouse and 208F rat fibroblasts. The cDNAs in both plasmids were driven by a strong human cytomegalovirus immediate early promoter and each contained an intron to promote high-level expression. In these assays, strong oncogenes induce transformed foci of cells in less than a week while less active oncogenes take longer to induce foci. In both cases the numbers of foci generally increase with time after transfection. Neither the RON nor the Stk expression plasmid induced transformed foci in NIH 3T3 cells (<0.2 foci per μg plasmid DNA at one month after transfection), while a plasmid expressing JSRV Env (pSX2.Jenv) induced 33 well-developed foci per μg plasmid DNA under the same conditions. This result is consistent with previous reports showing a lack of transformation by human RON [ 30 - 32 ], but is inconsistent with reports indicating efficient transformation of NIH 3T3 cells by mouse Stk (called "Ron" in these reports) [ 23 , 24 ]. We obtained the Stk expression plasmid used by the latter group (full-length mouse Ron cDNA in pcDNA3), and again could not detect transformation of NIH 3T3 cells by this plasmid (<0.2 foci per μg plasmid DNA at one month after transfection). There was some background focal growth in all plates of transfected NIH 3T3 cells one month after transfection, but this growth was identical in plates transfected with RON/Stk or control expression plasmids. In contrast, foci induced by JSRV Env were distinct and readily detected at 2 weeks after transfection. Next we tested for transformation of 208F rat fibroblasts by the RON and Stk plasmids. The Stk plasmid (pAlter-Max-Stk) induced 3 clear foci in a total of 7 dishes (0.2 foci per μg plasmid DNA), and the RON plasmid (pCIneoRON) induced 1 clear focus in a total of 5 dishes (0.1 foci per μg plasmid DNA) (foci counted 1 month after transfection, 2 μg plasmid DNA per dish, 2–3 separate experiments for each plasmid). The Stk expression plasmid pcDNA3Stk used in the earlier studies [ 23 , 24 ] did not induce transformed foci in the 208F cells (<0.1 foci per μg plasmid DNA). For comparison, the JSRV Env expression plasmid induced 300 to over 500 foci per dish when counted 2 weeks after transfection under the same conditions. Together these results document a very low transforming activity for the RON and Stk genes. Generation of cells expressing RON and Stk that respond to MSP Foci of transformed 208F cells were isolated from plates of cells transfected with the RON or Stk expression plasmids and were tested for their response to MSP. Cells transformed by the RON plasmid were not responsive to MSP and were discarded. Cells transformed by the Stk plasmid were responsive to MSP, and two clones (208F/Stk c12 and c13) were used for further study. Both of these clones made Stk protein of the appropriate molecular weight as judged by western analysis using RON/Stk-specific antibodies (data not shown). The response of 208F/Stk c13 cells to overnight treatment with MSP is shown in Fig. 1 . Without MSP the cells were elongated and formed chains of connected cells (Fig. 1 , top right panel), in contrast to the parental 208F cells which grow as flat uniformly-expanding colonies (Fig. 1 , top left panel). Treatment of the 208F/Stk c13 cells with MSP resulted in dramatic scattering (Fig. 1 , bottom right panel) while treatment of the parental 208F cells with MSP had no effect on cell morphology (Fig. 1 , bottom left panel). The MSP response was fully reversible; incubation of cells overnight in regular growth medium reversed the scattering induced by MSP and chains of cells reformed (not shown). 208F/Stk c12 cells exhibited an even more transformed appearance than did 208F/Stk c13 cells, and many of these cells would grow in suspension, often as doublets of what appeared to be recently divided cells. These floating cells could be repeatedly regrown in new plates, showing that the cells were alive and were not simply undergoing apoptosis. Passage of the floating cells provided a simple method for maintaining the transformed phenotype of these cells, which tended to decrease with time of culture. Since we were unsuccessful in generating 208F cells expressing functional RON by isolation of transformed cells following transfection, we isolated RON-expressing cells by transfection with a RON expression vector that also expressed a selectable marker (pMSCVpuroRON), grew the cells in puromycin to select for cells expressing the marker, and isolated multiple cell clones that showed a weakly-transformed phenotype. These clonal lines were screened for scattering response to MSP, and two independent clones that had the most transformed appearance and that responded to MSP (208F/RON c9 and c10) were identified and used in further analyses. Both clones produced RON protein of the appropriate size band by western analysis using a RON/Stk-specific antibodies, and both exhibited clear surface expression of RON by FACS analysis using a RON-specific antibody while control 208F cells showed no expression (data not shown). The 208F/RON clones had a less obvious transformed phenotype than the 208F/Stk clones but still showed convincing scattering in response to MSP (see below). Hyal2 expression does not affect the phenotype of cells expressing RON or Stk We tested whether overexpression of Hyal2 would reverse the altered phenotype of 208F cells expressing RON or Stk. Since the original work in BEAS-2B human cells showed inhibition of RON activity by human Hyal2 [ 5 ], we tested human Hyal2 in the 208F/RON cells. In 208F/Stk cells we tested mouse and rat Hyal2, as well as human Hyal2, in case the rodent proteins might show more effective interaction with mouse Stk in the rat cells. To quantitate the effect of Hyal2 expression on cell phenotype, cells were seeded at low density and colonies that grew out were scored for morphology, either round and flat, round and refractile, chain-like, or scattered. Fig. 2 shows examples of the phenotypes exhibited by 208F/Stk c13 cells. The top panel shows examples of chain-like (left) and flat (right) colonies, the middle panel shows an example of a chain-like colony, and the bottom panel shows an example of a scattered colony. In initial experiments we found that transduction of the cells with control vectors encoding either AP or Hyal1 had no effect on cell morphology (data not shown), and we used these vectors as negative controls for the effects of vector transduction in subsequent experiments. We exposed 208F cells that expressed RON to retroviral vectors encoding human Hyal2, AP, or human Hyal1, grew these cells in G418 to select for cells expressing the vector, and quantitated the morphologies of the cells that grew out. The experiment was repeated by replating the G418-resistant cells at low density and repeating the morphologic quantitation. The results were similar for both cell lines in both experiments so the results for the two experiments with each of the two cell lines were averaged (Fig. 3 ). There was no significant change in morphology of the cells expressing human Hyal2 in comparison to those expressing human Hyal1 or AP, arguing against the hypothesis that Hyal2 can inhibit the activity of RON. Indeed, cells expressing human Hyal2 exhibited fewer flat colonies and more chain-like colonies than the other cell types, the opposite of what the hypothesis predicts, although these differences were not statistically significant (Fig. 3 ). To confirm that Hyal2 was expressed in these cells, we measured transduction of the cells by a JSRV-pseudotype retroviral vector, which cannot infect rat cells unless they express a functional Hyal2 receptor [ 15 ]. We confirmed that only the 208F/RON cells transduced with the human Hyal2 vector, and not those transduced with the human Hyal1 vector, were susceptible to JSRV-pseudotype LAPSN vector transduction, showing that functional human Hyal2 was indeed expressed in these cells (data not shown). Next we exposed the 208F/Stk c12 and c13 clonal cell lines to retroviral vectors encoding mouse, rat or human Hyal2, or human Hyal1, grew these cells in G418 to select for cells expressing the vector, and quantitated the morphologies of colonies that grew after plating the cells at low density (Fig. 4 ). There were no consistent changes in morphology in response to any of the Hyal2 proteins by comparison to cells expressing Hyal1 or no additional protein. To test for expression of the Hyal2 proteins in these cells, we exposed the cells to JSRV-pseudotype LAPSN vector and measured the transduction rate (Table 1 ). As expected based on previous results [ 15 ], 208F/Stk cells expressing no additional protein or expressing human Hyal1 were not transduced, cells expressing human Hyal2 were transduced efficiently, cells expressing rat Hyal2 were transduced at about 1/3 the rate of cells expressing human Hyal2, and cells expressing mouse Hyal2 were not transduced. Acquisition of JSRV vector susceptibility by the normally resistant rat 208F cells following transfer of the human and rat Hyal2 vectors indicates that the Hyal2 genes are expressed in functional forms. We cannot conclude from this assay that mouse Hyal2 is expressed in cells transduced with the mouse Hyal2 vector, but the integrated vectors in both packaging lines that were used to make the mouse Hyal2 vectors were sequenced and found to have the correct mouse Hyal2 sequence, so it is likely that normal mouse Hyal2 protein was expressed in the transduced 208F/Stk cells. Together these results indicate that the Hyal2 proteins were expressed in the transduced 208F/Stk cells and argue against the hypothesis that Hyal2 can inhibit the activity of Stk. Hyal2 overexpression in RON- or Stk-expressing 208F cells does not affect their response to MSP To quantitate cellular responses to MSP, cells seeded two days earlier were exposed to MSP at concentrations from 1.2 to 400 ng/ml (in half-log intervals) and the next day the cell morphology was examined. We measured the minimum MSP concentration at which an alteration in morphology was observed and the MSP concentration above which no further morphologic changes were observed. The right panels of Fig. 1 show an example of such extremes of morphology. The original studies showing an interaction between RON/Stk and Hyal2 studied the human proteins [ 5 ], so we first examined the effects of human Hyal2 on the MSP response of 208F/RON cells. Identical responses to MSP were observed for 208F/RON c9 and 208F/RON c10 cells that expressed human Hyal2, human Hyal1, or AP. The responses were first noticeable at 4 ng/ml MSP and were maximal at 120 ng/ml MSP. These results indicate that Hyal2 does not modulate the activation of RON in response to MSP. We next explored the possible effects of Hyal2 on the MSP response of 208F/Stk cells. We tested human, mouse, and rat Hyal2 proteins to address the possibility that only the rodent Hyal2 proteins would interact with Stk. Identical responses to MSP were observed for 208F/Stk c12 cells that expressed mouse, rat, or human Hyal2, human Hyal1, or no protein. In two independent experiments, the first effects were observed at 1.2 ng/ml MSP and the maximum effect was observed at 12 ng/ml. Identical responses were also observed for 208F/Stk c13 cells that expressed mouse, rat, or human Hyal2, human Hyal1, or no protein. In this case, the responses were first noticeable at 4 ng/ml MSP and were maximal at 40 ng/ml MSP. Untreated 208F/Stk c12 cells are more refractile and express more Stk than the 208F/Stk c13 cells (data not shown), and the higher level and basal activity of Stk in these cells likely explains their higher sensitivity to MSP. Discussion The ability of JSRV Env to transform cells in culture, the identification of Hyal2 as the cell-surface receptor for the Env protein of JSRV, and the localization of Hyal2 to the chromosome 3p21.3 lung cancer tumor suppressor locus suggested the hypothesis that JSRV might cause cancer by inhibiting a tumor suppressor activity of Hyal2 [ 3 ]. Support for this hypothesis was provided by transformation studies in the human bronchial epithelial cell line BEAS-2B, which indicated that Hyal2 could suppress the transforming activity of the RON tyrosine kinase, and that JSRV Env expression caused Hyal2 degradation and RON activation [ 5 ]. The fact that a dominant-negative kinase-dead version of RON could block transformation by JSRV Env in BEAS-2B cells indicated that the RON pathway was required for transformation. Unfortunately and despite numerous attempts we have been unable to reliably measure transformation in BEAS-2B cells using the originally described expression plasmids and transfection techniques [ 5 ]. Occasionally we observe transformed foci following transfection of the JSRV Env expression plasmid, but often cells transfected with a control non-transforming plasmid exhibit similar foci. We are able to obtain reproducible and obvious transformation of BEAS-2B cells by transducing the cells with a retroviral vector encoding JSRV Env (LJeSN vector, ref. [ 6 ]), so presumably the lack of reproducibility is due to poor Env expression following transfection. To avoid these difficulties we pursued further studies of the RON pathway in rodent fibroblast cell lines that have traditionally been used in transformation assays. Based on earlier findings that Stk could transform NIH 3T3 mouse fibroblasts [ 23 , 24 ], we attempted to isolate cells expressing RON or Stk by isolating transformed foci of cells following transfection of NIH 3T3 cells. However, we were unable to detect transformation of NIH 3T3 cells by RON or Stk expression plasmids, consistent with results from other groups for RON [ 30 - 32 ] and for Stk (Kazuo Nishigaki and Sandra Ruscetti, personal communication). Furthermore, we were unable to repeat the studies showing transformation of NIH 3T3 cells by Stk when we used the same Stk expression plasmid used in the earlier reports [ 23 , 24 ]. These reports showed a very high background transformation rate in the NIH 3T3 cells (50 foci per μg control DNA, ref. [ 24 ]; see photographs of cell culture dishes in Fig. 2 of ref. [ 23 ]) casting doubt on the reliability of the transformation assay. Unless a very flat subclone of NIH 3T3 cells is used in these assays, as we used here, a high and variable background transformation rate can be observed. Even here we observed a level of background morphologic changes in the NIH 3T3 cells such that low levels of transformation would be difficult to detect. In contrast, the 208F cells showed no background transformation and allowed us to detect a low frequency of transformation by RON and Stk. This low rate of transformation could be due to a requirement for a very high level of RON/Stk expression that occurs only in a small number of cells, or a requirement for mutations that activate the RON/Stk proteins and occur at low frequency. We isolated several foci of Stk-transformed 208F cells and found that these cells expressed Stk protein of the expected size by western blotting and that the cells responded to MSP, indicating that the Stk protein was not constitutively activated by mutations or deletions, and suggesting that the protein is simply overexpressed in these cells. In contrast, cells from the one focus of cells transformed by transfection of the RON gene were unresponsive to MSP and thus might have been the result of activating mutations or deletions in RON. To generate 208F cells expressing RON we transfected RON with a selectable marker and isolated cells showing a weakly transformed phenotype. We presume that the low level phenotypic alteration observed in 208F/RON and 208F/Stk cells in the absence of MSP addition is due either to low-level constitutive activity of RON/Stk or to RON/Stk activation by low levels of MSP made by the cells or present in the culture medium. Given that the parental 208F cells show no response to MSP, we conclude that the cell scattering phenotype observed after addition of MSP is a direct measure of RON/Stk activation. The transformed phenotype of these cells in the absence of added MSP and the scattering phenotype observed after treatment of the cells with MSP seem to represent variations of the same transformation phenotype. Indeed, 208F cells transformed by other oncogenes, including JSRV Env, exhibit a scattered phenotype similar to MSP treated 208F cells expressing RON or Stk. Furthermore, 208F cells expressing RON or Stk can exhibit a scattered phenotype in the absence of added MSP, especially if the cells are selected for a highly transformed phenotype by passaging substrate-independent cells that float in the culture medium. Given this interpretation, the response of the RON/Stk-expressing cells to MSP can be considered a form of ligand-dependent transformation, similar to what has been observed previously for cells expressing the receptors for human colony-stimulating factor 1 [ 33 ], insulin-like growth factor I [ 34 ], granulocyte/macrophage colony-stimulating factor [ 35 ], and a hybrid EGF receptor/Xmrk tyrosine kinase [ 36 ]. Alternatively, it may be that the basal and MSP-induced phenotypes of the RON/Stk-expressing cells are mediated by separate signaling pathways, as has been found for the related receptor tyrosine kinase MET [ 37 ]. These investigators found that a MET oncogene could induce transformed foci and an invasion phenotype in cultured rat fibroblasts, which when injected into nude mice would form tumors and would metastasize to the lungs. Two point mutations in the MET oncogene completely abrogated the invasion and metastasis phenotypes but did not affect the transformation and tumorigenesis phenotypes of the MET oncogene, indicating that separable signaling pathways were involved. We did not find any effect of human Hyal2 on the basal phenotype or the response to MSP of the 208F/RON cells, nor did we find an effect of human, rat, or mouse Hyal2 on the basal phenotype or the MSP response of 208F/Stk cells. These results argue against direct regulation of RON activity by Hyal2 as was indicated by prior experiments using BEAS-2B cells [ 5 ]. It is possible that additional factors present in BEAS-2B but not in 208F cells are required for regulation of RON activity by Hyal2. Recently we have found that Env proteins from JSRV and from the related retrovirus enzootic nasal tumor virus (ENTV) can transform Madin-Darby canine kidney (MDCK) epithelial cells, but by a mechanism different from that observed in BEAS-2B cells and similar to that observed in rodent fibroblast cell lines [ 38 ]. In particular, the cytoplasmic tail of Env is required for transformation, the PI3K/Akt pathway is activated and inhibition of PI3K activity reverses the transformed phenotype, expression of RON (which is not normally expressed in MDCK cells) does not affect transformation, and canine Hyal2 expressed by these cells appears uninvolved. These results show that the JSRV Env protein can transform epithelial cells besides BEAS-2B cells, and argue against a model for Env transformation involving different pathways that are uniquely active in epithelial cells as compared to fibroblasts. Future work will focus on the transformation pathways that are active in lung tumors induced in animals by JSRV Env. While it is possible that the RON pathway plays a role in oncogenesis, most of the evidence points to a main role for the cytoplasmic domain of Env and activation of the PI3K/Akt pathway. Conclusions We have shown that expression of the mouse or human orthologs of the receptor tyrosine kinase Mst1r (called RON in humans and Stk in rodents) can induce phenotypic changes indicative of transformation in a rat fibroblast cell line. These changes are enhanced by treatment of the cells with MSP, the natural ligand for RON/Stk, which induces a pronounced scattering/transformation response. Transformation of the BEAS-2B immortalized human epithelial cell line by JSRV appears to require RON activation, as a result of JSRV Env binding to and degradation of the virus receptor Hyal2, which appeared to negatively regulate RON activity. However, we find no evidence for direct regulation of the basal or MSP-induced RON/Stk activity by Hyal2 in rat fibroblasts. Competing interests None declared. Authors' contributions ADM designed the study, performed most of the experiments, and drafted the manuscript. NSVH and SLL performed the western blot and FACS experiments and provided intellectual input, and SLL performed the preliminary studies on BEAS-2B cells. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521489.xml
534090
The Chernobyl childhood leukemia study: background & lessons learned
Many challenges emerged during completion of a study to examine radiation dose and acute leukemia among children in areas of the former Soviet Union. In an era of globalization, our experiences might benefit others involved in multinational investigations.
Introduction This paper identifies the major challenges faced and the lessons learned in addressing them by the collaborative research groups involved in developing and conducting a large, multi-national case-control study of acute leukemia among children in areas of the former Soviet Union (FSU) that were most heavily exposed to radioactive fallout as a result of the April 1986 accident in reactor vessel #4 of the Chernobyl Nuclear Power Plant. In this accident, a variety of radioisotopes including iodine ( 131 I), cesium ( 137 Cs, 134 Cs), and strontium ( 90 Sr), were released from the damaged reactor vessel contaminating soil, vegetation, and groundwater [ 1 ]. Fallout from the Chernobyl accident contaminated large portions of Eastern Europe, the then Union of Soviet Socialist Republics (USSR) and more distant regions. Areas of the FSU, including the now independent republics of Belarus, Russia, and Ukraine, were among the most heavily contaminated. The intent of the research project was to examine acute leukemias without specific regard to national boundaries, while recognizing the requirement to bring together investigators from these three republics in a common effort. Acute external exposures to ionizing radiation have been etiologically linked with observed increases in the risk of all types of leukemia, except chronic lymphocytic leukemia; the risk is greatest for acute myeloid leukemia [ 2 - 12 ]. The association between exposure to ionizing radiation from the Chernobyl accident and the occurrence of leukemia has been summarized in a recent review [ 10 ] which highlights mixed results from published studies to date (see the review article for a comprehensive overview of published studies). Accounting for differences between the studies in the methodologies used to assess the radiological exposures, the procedures for identifying childhood malignancies and in the lengths of follow up, these authors concluded that there is not strong evidence demonstrating increases in childhood or adult leukemia from Chernobyl exposures [ 10 ]. Nonetheless, because the link between high dose ionizing radiation exposure and the development of leukemias is widely established, and children are felt to represent a uniquely susceptible population, this issue remains of high scientific and public heath interest. The study described herein focused on this most susceptible subgroup. Collaborative studies are challenging at best; international consortia efforts are even more so. The study of border-crossing disasters such as that in Chernobyl need be investigated as a single scientific challenge to take advantage of standardizing the investigative process across national boundaries, and to afford adequate sample size and the associated statistical power to reach meaningful conclusions. This paper will not recount the events of the Chernobyl disaster or specific results of the investigation. Rather, it will share the lessons learned while conducting a collaborative, multinational study for the benefit of other investigators. Lesson #1: Developing a basis for collaboration One of our first challenges was to establish an understanding of and a presence in each of the affected areas of the FSU. Multinational/multi-site studies are labor intensive and difficult to coordinate from distant locations. Therefore, the identification of a strong, on-site research team is essential. Each group must be treated as a full-member of the larger study team and be fully engaged. The local team is primarily responsible for study implementation. It is imperative that local teams with relevant expertise be formed. Our teams employed physicians, epidemiologists, statisticians, and dosimetrists. Local teams participated in study design, as well as study management in the field and troubleshooting problems. This required teams that were extremely knowledgeable about geopolitical boundaries, location of cities/small villages, governmental structures, and the health care delivery system. The latter area was particularly important given that the entire study hinged on the identification of acute leukemias and pair matched healthy controls. The health care systems in Belarus, the Russian Federation, and Ukraine have been adapted from the FSU system. Each Oblast (an Oblast is a large administrative unit – equivalent to a state in the United States) has 1–2, and in rare cases 4, large cancer treatment and diagnostic facilities (termed oncodispensaries). There are separate facilities for childhood (i.e., <16 years of age at time of diagnosis) and adult patients. In addition to providing nearly all cancer-related medical treatment, oncodispensaries house records of all reported cancer patients within the Oblast. These treatment centers contained critical source records vital to the success of our study. On-site research teams knew how to identify cases (e.g., review of oncodispensary records), where to identify controls (e.g., polyclinics), and what agency permission was needed to obtain access to these data. This would have been impossible for an "outsider" to do. The identification of committed and skilled collaborators from each of the Republics facilitated project coordination and monitoring from outside the FSU. However, not surprisingly, there have been some difficulties related to language and culture. Although English language proficiency is progressing rapidly within the FSU, many of the individual researchers were unable to effectively communicate via English. Moreover, the collaborators based in the USA were even more limited in their Russian language fluency. This issue of language was addressed early in this collaboration through a consensus that English would be used as the common language, which was endorsed by our FSU collaborators. Certain investigators were well trained in English and the Consortium relied on simultaneous interpretation for initial meetings of collaborators, and then switched to having an interpreter present to provide general and personal assistance as needed. All printed materials, such as study protocols and survey instruments were produced in dual language versions (Russian and English) and verified via back translation. Since the late 1990s, key project staff members in the USA and each FSU Republic maintained frequent linkages through e-mail communications at intervals between site visits and periodic project reviews. However, it should be noted that prior to the late 1990s, e-mail was scarce in the FSU and somewhat unreliable. International communications were first established via facsimile transmissions then evolved to an electronic platform (e-mail) as Internet coverage in the FSU expanded. Lesson #2: Infrastructure Development This study was supported by the International Consortium for Research on the Health Effects of Radiation (ICRHER), based in the United States. (A series of articles summarizing two research conferences appears in a special supplement of the Stem Cells journal [ 13 ].) Our research objective was to examine the relationship between exposure to chronic low doses of ionizing radiation and the incidence of acute leukemias among children in the 3 FSU republics without specific regard to national boundaries (see figure 1 ), while recognizing the obligation to bring investigators from the three republics together with each other and those in the USA in a common effort. Individuals in the 3 republics exposed during childhood to radiation from Chernobyl were felt to represent a uniquely susceptible group and an appropriate population for a retrospective study. The ICRHER was incorporated in June 1993. It was born of the insight and enthusiasm of one man, the late Admiral Elmo Zumwalt, Jr., United State Navy (retired), who was concerned about environmental exposures and cancer risk; particularly, how such exposures might affect military personnel. Admiral Zumwalt was also concerned about the humanitarian aspects of accidental exposure to toxic agents, including ionizing radiation. The Chernobyl reactor accident provided an opportunity to pursue both these interests. Initially, the ICRHER (the "Consortium") was comprised of collaborative groups at institutions in the United States of America (USA) and FSU republics, specifically, the Baylor College of Medicine, Houston, Texas; the Research Center for Radiation Medicine, Kiev, Ukraine; the Fred Hutchinson Cancer Research Center (FHCRC), Seattle, Washington; the National Center for Hematology, Moscow, Russia; Hadassah Medical Organization, Jerusalem and Haifa, Israel; and the National Marrow Donor Program, Minneapolis, Minnesota. The Bridgeport Hospital/Yale University and the Research Institute of Radiation Medicine and Endocrinology, Minsk, Belarus, joined the Consortium later. The final research team for the multinational Leukemia case-control study includes the FHCRC in collaboration with the Medical Radiological Research Center in Russia and two groups from the Roswell Park Cancer Institute (one working with the Research Institute of Radiation Medicine and Endocrinology, in Minsk, Belarus and another with the National University "Kiev-Mohyla Academy", Kiev, Ukraine). The Consortium helped to bring structure and organization in a variety of areas, from assembling the principal investigators to creating working groups to empanelling external advisory boards. A corporate entity or central office was formed to coordinate these activities, and to manage common logistic requirements, to draw up contracts for U.S. and Israeli institutions, to support the FSU collaborators, and to provide an administrative office of record. Satellite offices were established in Kiev, Minsk, Moscow, Obninsk, Bryansk and Jerusalem to provide administrative, communication and logistical support to the research activities in the FSU and Israel. FSU offices, as will be seen later, were critical to the success of this study. Given the changes in leadership and make up of the ICRHER since its inception, it was imperative to maintain flexibility and responsiveness to change. Modifications in research partners may occur for many reasons. Therefore, it is critical that some type of "corporate entity" exist to aid in these transitions through detailed documentation and to provide continuity at the executive/advisory level. Consortium investigators developed a pilot study to test the feasibility of initiating a full-scale epidemiologic investigation, and to determine whether needed collaborations could be established. The primary goals were to assess radiation dose, to identify individuals at risk for radiation-related illness (including émigrés to Israel), to develop data collection instruments and common study protocols, to elucidate mechanisms of radiation damage, and to establish core support through the central office. The original group invested considerable time in reviewing exposure data, the populations involved, and possible health outcomes. Following these critical discussions, a decision was made to launch a pilot investigation of acute leukemia. This study would prove challenging, but it was facilitated by the knowledge and experiences gleaned from the feasibility study. Lesson #3: Study Design Based on the feasibility and pilot studies, a consensus decision was made to begin a multinational leukemia case-control study in Belarus, the Russian Federation, and the Ukraine, including the Israeli component (the Israeli studies concluded in August 2000). The focus has been on the most radiosensitive malignancy in the most radiosensitive at-risk age group: occurrence of acute leukemia in those 0 to <6 years at the time of the accident [ 2 , 10 ]. Initially, the aim of the Consortium was to investigate the occurrence of acute leukemia among individuals who were 0 to age 20 years at the time of the Chernobyl accident. The original population was selected due to early indications of high radiation dose exposures and easily achievable sample size estimates based on power calculations. When doses were not found to be as high as predicted, the sample size requirements became exceedingly large. Therefore, the researchers decided to narrow their focus to individuals aged 0 to <6 years at the time of the accident, since they represented a much more radiosensitive subpopulation. Hence, there is the need for well-designed feasibility studies to provide critical information in study planning and decision-making. The final study population included 421 confirmed cases of acute leukemia and 842 population controls pair-matched to cases on age, gender, and type of residence. All participants were in utero to <6 years of age at the time of the Chernobyl accident in April 1986. Case identification Case identification was often difficult, particularly when compared to studies conducted in the U.S. that typically rely on either population-based cancer registries or hospital-based rapid case identification. Many regions studied did not have an up-to-date, population-based cancer registry. And, since nearly all medical records in the FSU exist in paper form, exhaustive manual reviews of hard copy records were often required. This task was somewhat simplified by the local custom of referring cancer cases to regional oncodispensaries, thereby, restricting record reviews to a finite number of facilities. Moreover, it was also necessary to review death files to ascertain cases not previously brought to medical attention. Trained physicians reviewed the records at the oncodispensaries in Belarus, the Russian Federation, and the Ukraine. Records were retrieved for all potential patients (i.e., date of birth, date of diagnosis, and residence location). After verifying eligibility, trained interviewers contacted parents of these patients and scheduled an interview. Control selection Control selection was equally vexing. In the US, researchers typically use neighborhood controls, random digit telephone dialing methods, and/or computerized databases [ 14 - 16 ]. In the FSU, control selection began by meeting with district health officials to obtain permission to review polyclinic records. Considerable time was then spent reading through racks of paper files and recording information for potential participants. Interviews were completed with two controls pair-matched to each case based upon age at diagnosis, sex, region/district of residence, and type of settlement. The research teams visited the polyclinic and randomly selected 20 potential controls fitting the inclusion criteria. Interviews were generally scheduled and completed for the first two potential controls on the list. Other names on the list were contacted and scheduled for interviews as needed. Over 90% of potential controls contacted by the research teams agreed to participate. These high rates of case and control participation may be attributed to the dedication and resourcefulness and commitment of the research teams and the use of modest incentives (food baskets). Interview As formidable as the above tasks were, they paled in comparison to actual data collection. During the feasibility study, investigator meetings were held in the US to develop common data collection instruments. These instruments were structured to address issues of cultural and language compatibility for use in the three Republics. While Russian was adopted as the standard language for data collection, study instruments contained both Russian and English text. Instruments were back translated into English to ensure accuracy of translation. Next, a team of interviewers needed to be recruited and trained. We relied almost exclusively on recruiting physicians who were familiar with the disease process, were credible representatives, were respected by study participants, and who could be depended on to provide accurate and verifiable data. The mix of urban areas and rural settlements presented logistical challenges in terms of tracking participants, arranging appointments, and completing face-to-face interviews. To overcome this challenge, some interviewers mailed letters to introduce the project, and to request that the participant contact the interviewer. Others traveled to local communities to personally discuss participation and to schedule interviews. Still others used local contacts to identify participant places of employment for either telephone or direct contact. Traveling to interview study participants was another challenge. In contrast to conducting research in the US where there are widespread and reliable telecommunication systems, well-developed highway systems, and accurate maps, FSU travel was more problematic. For example, not all roads to small villages in the FSU are paved and most are limited to a single lane in either direction. Roads can be particularly treacherous in the winter and rainy seasons. Maps do not always show the precise location of small villages. Villages have neither street signs nor house numbers. Further, the collapse of the Soviet Union resulted in changes in the names of many streets, settlements, and villages. This arduous process of locating and interviewing controls took considerable effort and time. Where interviewers in the U.S. are generally out and back in the same day, FSU colleagues were out in the field for days at a time and would be fortunate if they could complete a handful of interviews each day. These differences in operations must also be factored in to the overall cost of the study. Finally, local residents tended to be wary of strangers and special introductions were often necessary to gain entry. One cannot overstate the importance of having a well-trained and dedicated on-site research team that is familiar with cultural norms and local "maps". The two-hour, face-to-face interviews were generally completed with the mothers of cases and controls. Interview items were developed by US and FSU scientists and included sections addressing demographics, general health status, maternal and paternal occupational history, and a detailed questionnaire for obtaining information necessary for developing estimates of individualized internal and external radiation exposures, including questions regarding consumption of locally produced milk, meats, and vegetables; residence history and type of housing structure; use of protective measures immediately after the accident; and time spent outdoors. Lesson #4: Working Groups Those involved in multinational or multi-site investigations may consider the creation of Working Groups to monitor the various components of the study. As has been utilized in other international studies [ 17 ], the lead investigators could separate the overall study into its basic components and organize working groups that represent all collaborating teams. For example, we established a leukemia diagnostic working group comprised of hematologic morphologists and hematologists from representative ICRHER institutions, and chaired by a leading pediatric hematologist who was not associated with the Consortium. Members completed blinded reviews of bone marrow pathology slides, or other information (e.g., clinical histories, laboratory data) for cases without slides, and then assigned a histopathologic diagnosis based upon group consensus. A subset of cases was randomly selected for repeat review to assess consistency. Results of this review process affirmed the accuracy of acute leukemia diagnoses made within these areas of the FSU [ 18 ]. A common methodology to assess individual absorbed radiation doses and corresponding uncertainties was developed and tested by the dosimetry intercomparison working group (DIWG) for all subjects based on interview data and available exposure data. The interview collected detailed exposure data for each subject from the time of the Chernobyl accident until the reference date. It should be noted that interview data alone were not sufficient to determine individual dose estimations. For retrospective dose estimation specialized radioecological data were necessary. These resources provided information concerning local soil types, food contamination with different radionuclides, dates of radioactive cloud arrivals to each local community, etcetera. These data were collected by the members of DIWG based on information published within the FSU. Dosimetrists are a good example of the specialized personnel that need to be a full-time, on-site presence. They have critical knowledge of the local area as well as access to primary data essential in quantifying exposure. Individual dose estimates reflected local conditions (e.g., contamination levels, soil type, soil to milk transfer coefficients) at each location where a particular subject lived during the appropriate time interval. The subject's residence history and other important personal information, such as milk consumption and food sources, were collected during the standardized interviews by trained examiners. Fieldwork was performed jointly by physicians and dosimetrists. A Data Analysis Working Group was essential for oversight of the analytic phase. Composition included representatives from each site with a strong chair to maintain focus and momentum. Also, a Data Audit Working Group was established at study inception to conduct periodic reviews to assure adherence to study protocols. And finally, a centralized Data Coordination Office served as a repository of the common dataset and oversaw periodic computerized checks for quality and completeness. Conclusions Our experiences in the organization and successful implementation of a multinational, retrospective study of acute leukemia in regions impacted by the Chernobyl disaster have been highlighted. Issues identified during the implementation of our multi-national epidemiology study, along with strategies for resolution are summarized in table 1 . While trained research teams within each Republic were responsible for collecting data, we relied on a distinctive series of working groups of collaborators from participating institutions to coordinate various aspects of the study such as case confirmation, data quality, dosimetry, and data analyses. This allowed all project teams to remain interconnected and equally involved while utilizing the unique expertise of various collaborators. Table 1 Potential issues regarding the implementation of multi-national epidemiology studies Challenge: Resolution strategy: •Language •Dual language versions (Russian & English) for all printed materials; use of interpreters •Geographic distance between collaborators •e-mail accounts for key collaborators; site visits, progress meetings •Limited experience with epidemiology •Mandatory training workshops for interviewers; audits to assure compliance with protocols •Subject ascertainment •Cases identified through manual record reviews at oncodispensaries and childhood oncology centers; controls identified from manual review of raion medical records •Limited comprehensive cancer registry data •Manual records review at oncodispensaries, childhood cancer centers and mortality files •Lack of telephone to contact participants •Mailed letters of introduction; field trips to communities •Locations of study participants •Field trips for data collection; assistance of local residents •Radiologic contamination data in multiple locations •Visits to multiple Institutes & offices; contacts of collaborators •Adequate communications •All research sites provided immediate Internet access •Timely compensation for local investigators •Direct pay facilitated by USA agencies (e.g., Civilian Research and Development Foundation) •Common research protocols and joint methodology for individual radiation assessment •Periodic meetings of all USA/FSU investigators to promote personal relationships and scientific value of combined data •Data collation and analysis •Establish Data Coordination Office in the Former Soviet Union •Transfer all data electronically •Data access and archives •Access by mutually-agreeable policy •Transfer data to USA institution for permanent storage •Multidisciplinary international study •Highly cooperative, joint international consortium with working groups The breakup of the former Soviet Union in 1991 created national autonomy in Belarus, the Russian Federation, and Ukraine. Although the study area for this project included selected regions of the FSU, the intent was to study acute leukemia without regard to specific national boundaries, while recognizing the requirement to bring together investigators from these three Republics in a common effort. Proprietary concerns and country-specific restrictions on the sharing of scientific information were thoroughly addressed to gain agreement and to facilitate the pooling of analytic information. The Consortium played an integral role in providing infrastructure support for this project through the appointment of project support administrators at each research site to oversee communications, equipment procurement/maintenance, and compensation. The fiscal aspects of supporting research are unique to each country. A careful examination of collaborating scientific institutions and the financial regulations of each participating country prior to setting up any support mechanisms is critical, and may result in country-specific arrangements. Essential computer, laboratory, and communication equipment was supplied. Computer software, which was compatible across the three research settings, was installed and upgraded periodically. Equipment was segregated and secured to insure exclusive use by project staff. It should be emphasized that this investigation represents the largest retrospective study examining the relationship between Chernobyl radiation exposures and risk of acute leukemia conducted to date and the only research effort to bring together data from the most exposed areas into a single study; results of the multinational case-control study are presented in a separate paper [ 19 ]. The conduct of multinational epidemiologic studies presents numerous challenges [ 17 , 20 ], including issues such as language, physical infrastructure, telephone coverage, and road conditions, as well as geographic distances and issues of participant ascertainment. However, as demonstrated by our experiences, these challenges can be effectively overcome through attention to organization, communication, and quality assurance. Moreover, these challenges are greatly offset by unique opportunities to yield information of great significance to science and society. List of Abbreviations FSU, former Soviet Union USSR, Union of Soviet Socialist Republics USA, United States of America ICRHER, International Consortium for Research on the Health Effects of Radiation; the "Consortium" FHCRC, Fred Hutchinson Cancer Research Center DIWG, dosimetry intercomparison working group Competing Interests The author(s) declare that they have no competing interests. Author contributions MCM and AMM were responsible for the study concept. MCM and AMM drafted the manuscript; KMM, PLM, RCM, VFS and RWD provided critical review and input. MCM, KMM, PLM, RCM, VFS, RWD and AMM participated in interpretation, as well as in data acquisition efforts. All authors read and approved this manuscript. Figure 1 Regions surrounding the Chernobyl Nuclear Power Plant Shading identifies areas included in the International Consortium for Research on the Health Effects of Radiation (ICRHER) study of acute childhood leukemia: Gomel & Mogilev Oblasts in Belarus; Cherkassy, Chernigov, Rivno, & Zhitomir Oblasts in Ukraine; and Bryansk Oblast, Russian Federation. Solid lines identify boundaries between countries/republics. Shaded square in center of figure identifies location of Chernobyl Nuclear Power Plant in northern Ukraine. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC534090.xml
544563
Lung function, asthma symptoms, and quality of life for children in public housing in Boston: a case-series analysis
Background Children in urban public housing are at high risk for asthma, given elevated environmental and social exposures and suboptimal medical care. For a multifactorial disease like asthma, design of intervention studies can be influenced by the relative prevalence of key risk factors. To better understand risk factors for asthma morbidity in the context of an environmental intervention study, we conducted a detailed baseline evaluation of 78 children (aged 4–17 years) from three public housing developments in Boston. Methods Asthmatic children and their caregivers were recruited between April 2002 and January 2003. We conducted intake interviews that captured a detailed family and medical history, including questions regarding asthma symptom severity, access to health care, medication usage, and psychological stress. Quality of life was evaluated for both the child and caregiver with an asthma-specific scale. Pulmonary function was measured with a portable spirometer, and allergy testing for common indoor and outdoor allergens was conducted with skin testing using the prick puncture method. Exploratory linear and logistic regression models evaluating predictors of respiratory symptoms, quality of life, and pulmonary function were conducted using SAS. Results We found high rates of obesity (56%) and allergies to indoor contaminants such as cockroaches (59%) and dust mites (59%). Only 36% of children with persistent asthma reported being prescribed any daily controller medication, and most did not have an asthma action plan or a peak flow meter. One-time lung function measures were poorly correlated with respiratory symptoms or quality of life, which were significantly correlated with each other. In multivariate regression models, household size, body mass index, and environmental tobacco smoke exposure were positively associated with respiratory symptom severity (p < 0.10). Symptom severity was negatively associated with asthma-related quality of life for the child and the caregiver, with caregiver (but not child) quality of life significantly influenced by caregiver stress and whether the child was in the intensive care unit at birth. Conclusion Given the elevated prevalence of multiple risk factors, coordinated improvements in the social environment, the built environment, and in medical management would likely yield the greatest health benefits in this high-risk population.
Background Asthma morbidity and mortality have been increasing in recent years, with a disproportionate impact on urban minority children [ 1 - 4 ]. Hospitalization and morbidity rates have been shown to be elevated for nonwhites versus whites [ 3 ] and in inner-city settings with low-income populations [ 4 ]. Multiple recent studies have attempted to explain these disparities by evaluating environmental exposures and housing conditions, racial/ethnic variations, poverty, and social or psychological factors, with no definitive conclusions regarding the dominant factors [ 1 , 2 , 5 - 9 ]. Regardless of the relative contributions of these and other factors, children in urban public housing are important to consider, because they likely have elevated exposures across numerous domains, some of which could be addressed through development-wide interventions. However, there has been only limited evaluation to date of asthma in this high-risk subpopulation [ 10 , 11 ]. The Healthy Public Housing Initiative (HPHI) is a collaborative effort that includes the Boston Housing Authority (BHA), West Broadway and Franklin Hill Tenant Task Forces, Committee for Boston Public Housing, Boston Public Health Commission, Boston University and Harvard University Schools of Public Health, and Tufts University School of Medicine. A primary goal of HPHI is to evaluate the effectiveness of interventions in reducing known asthma triggers and improving the health of pediatric asthmatics in public housing in Boston. The effectiveness of environmental interventions in this context will clearly depend on the prevalence of environmentally-linked risk factors within this cohort (i.e., allergy status), as well as the prevalence of other risk factors for asthma morbidity. In addition, when evaluating the efficacy of environmental interventions, numerous health endpoints may be valuable to consider. Health care utilization will inform cost-effectiveness analyses, typically driven by infrequent but severe events such as hospitalizations [ 12 , 13 ]. On the other hand, self-rated quality of life can capture a broad array of activity-based and psychosocial outcomes, and pulmonary function measures or respiratory symptoms provide more objective and sensitive markers of health improvements. Past studies have demonstrated varied relationships among these parameters. For example, percent of predicted forced expiratory volume in one second (FEV1%) has been correlated with asthma attacks [ 14 ] and symptom score but not with symptom days [ 15 ]. Asthma-related quality of life was correlated with FEV1% in a low-income adult population [ 16 ] but not in a general population sample [ 17 ]. FEV1% was correlated with asthma-related quality of life in mild asthmatics, but was a weaker predictor than symptom intensity and was not correlated significantly for more severe asthmatics [ 18 ]. Finally, pediatric asthma symptoms have been correlated with asthma-related quality of life but not FEV1% or measures of asthma control, with relationships that vary by age of the child [ 19 ]. However, none of these studies focused on low-income pediatric populations. More broadly, past studies have generally not considered the full array of risk factors and health endpoints for inner-city asthmatics. The most comprehensive assessment to date has been the National Cooperative Inner-City Asthma Study (NCICAS), which evaluated many similar endpoints as our study in a longitudinal baseline assessment [ 20 ], although self-rated quality of life was not considered in this publication. In order to understand the characteristics of asthmatic children in our longitudinal intervention study and to determine the relationships among key health endpoints, we conducted an extensive baseline assessment for all children enrolled in our study. Thus, the objective of our analysis is to characterize the baseline risk factors and health status of a cohort of asthmatic children enrolled in an intervention study based in public housing developments in Boston and to determine concordance between and risk factors for key health endpoints (e.g., pulmonary function, respiratory symptoms, and self-reported quality of life). We hypothesize that risk factors associated with housing quality and psychosocial stress will be elevated in our cohort when compared with reference groups, and that quality of life will be significantly influenced by asthma symptom severity and other caregiver characteristics. Methods We recruited asthmatic children from the Franklin Hill, West Broadway, and Washington Beech public housing developments in Boston (located in the neighborhoods of Dorchester, South Boston, and Roslindale, respectively). Recruitment was coordinated by Community Health Advocates, residents of the developments or surrounding neighborhoods who were involved in outreach and data collection, following training about asthma, its risk factors, and interviewing techniques. Recruitment methods included advertised enrollment open houses, community meetings, mailbox drops for flyer circulation, and door knocking. Any children between the ages of 4 and 17, who lived in the developments, had self-reported doctor-diagnosed asthma, and who were willing to enroll in a longitudinal intervention study, were eligible. Enrollment occurred between April 2002 and January 2003. Written informed consent was obtained from all caregivers, with assent forms completed by children above the age of 8, and the protocols were approved by the institutional review boards of all three participating universities. In the intake interview, the caregiver (defined as the individual who knows most about asthma care for the child) was asked about family demographics, child and family asthma history, access to health care, exposure to smoking, and medication usage, with questions taken from NCICAS when possible to facilitate comparability [ 20 ]. Because psychological stress has been shown elsewhere to be a strong predictor of immune function [ 21 ] and airway inflammation and obstruction [ 22 ], the caregiver was given the Cohen four-item abbreviated Perceived Stress Scale [ 23 ]. In addition, she was asked about neighborhood social cohesion and exposure to violence [ 24 ], factors that have been related to respiratory symptoms and other measures of asthma morbidity [ 25 ]. Finally, she was surveyed about the influence her child's asthma had on her quality of life, using the Paediatric Asthma Caregiver's Quality of Life Questionnaire (PACQLQ) [ 26 ]. The Paediatric Asthma Quality of Life Questionnaire (PAQLQ) [ 27 ] was used to determine the influence of asthma on the child's quality of life, administered to the caregiver for children age 7 and younger and directly to children age 8 and older. In addition, we evaluated quality of life using the EuroQol EQ5D self-report questionnaire [ 28 ] combined with previously published formulas [ 29 ], providing a comparison between an asthma-specific scale and a general health status scale. The EQ5D also included a visual analogue scale (VAS). Pulmonary function was assessed using the NDD EasyOne Diagnostic portable spirometer (NDD Medical Technologies, Andover, MA), an instrument which correlates well with office-based spirometry [ 30 ]. Although pulmonary function was measured longitudinally within the intervention study, we focus on the baseline assessment (conducted concurrently with the intake interview). To compare lung function across children, we determined the percent of predicted value for FEV1 and peak expiratory flow (PEF) using standard reference equations [ 31 ]. Spirometry results and questions regarding symptom severity and medication use were used to classify asthma severity following NHLBI guidelines [ 32 ]. In addition, given reported height and weight, we calculated body mass index (BMI) and used age-specific BMI distributions [ 33 ] to categorize children as overweight (above 95 th percentile), at risk of overweight (85 th to 95 th percentile), or not at risk (below 85 th percentile). Allergy testing was conducted using similar methods as NCICAS [ 20 ], with skin testing using the prick puncture method. Valid tests had a negative control wheal at least 1 mm smaller than the positive histamine wheal, and tests were considered positive if the wheal for a given allergen exceeded the negative control wheal by at least 2 mm. Allergens evaluated included an 11-tree mix, a 7-grass mix, ragweed, dog, cat, mouse, cockroach, D. pteronyssinus , D. farinae , Alternaria , Aspergillus fumigatus , Cladosporium , and Penicillium . For our exploratory regression models, we evaluated the relationship between our primary outcome measures (FEV1 % predicted, a respiratory symptom score, and caregiver and child asthma-related quality of life) and a subset of demographic variables, intrinsic risk factors, health care risk factors, physical risk factors, and social risk factors. Furthermore, we considered FEV1% as a potential predictor of the respiratory symptom score, and both FEV1% and the respiratory symptom score as hypothetical predictors of quality of life. We treated FEV1% in a logistic regression, using 80% of predicted FEV1 as the cutoff for low FEV1%, and evaluated other health outcomes in linear regressions. Given missing data and numerous covariates, we conducted an initial screen using univariate regressions (retaining variables for which p < 0.2), and then constructed a multivariate stepwise regression with p < 0.1 as the entry and exit criteria. Finally, for covariates with extensive missing data, we constructed multivariate regressions both with and without these terms to evaluate the sensitivity of our findings. All statistical analyses were conducted using SAS version 8.02, using PROC REG for linear regressions and PROC LOGISTIC for logistic regressions. Results Demographics and risk factors In total, 78 children from 61 households were enrolled in the HPHI intervention study. As indicated in Table 1 , 41 (53%) of these children were from Franklin Hill, with 27 (35%) from West Broadway and 10 (13%) from Washington Beech. The mean age at the time of enrollment was 8.7 years (median = 8.0), with a similar age distribution across the three developments. A majority of participants (64%) self-reported as Hispanic, with 33% self-reporting as black or African-American. Table 1 Baseline demographic characteristics of asthmatic children in three public housing developments in Boston Franklin Hill West Broadway Washington Beech Total Number of children 41 27 10 78 Age (%) < 6 27% 30% 40% 29% 6–9 32% 26% 20% 28% 10–12 22% 30% 20% 24% >= 13 20% 15% 20% 18% Race/Ethnicity (%) * Hispanic 61% 67% 70% 64% African-American 41% 22% 30% 33% Caucasian 0% 11% 0% 4% * Race/ethnicity was asked in an open-ended question, so respondents could indicate both Hispanic status and race. So, the total can be greater than 100%. Considering prominent non-environmental risk factors associated with asthma, 16% of children were in an intensive care unit upon birth and 10% were on a respirator. Seventy percent of children had a parent or grandparent with asthma, while 43% had eczema or hay fever and 34% had a family history of eczema or hay fever. For the 75 children with recorded height and weight, 56% were categorized as overweight and 9% were categorized as at risk of overweight. Forty-two percent of children lived with a smoker, and 45% of children were around smokers at least several times per month. For the subset of 46 children (59%) who underwent allergy testing, 78% were sensitized to at least one of the tested substances, with the most prevalent allergies including D. pteronyssinu s (59%), cockroach (59%), D. farinae (50%), and tree pollen (30%) (Figure 1 ). Only 44% of these allergic children were reported to have allergies by their caregivers. Figure 1 Prevalence of allergies among asthmatic children in HPHI and NCICAS [20]. Medical care For the 70 children for whom we received information about medications, 87% were taking short-acting beta-agonists at the time of the survey. Thirty-one percent of these children on beta-agonists also reported the use of any long-term asthma control medication (corticosteroids, leukotriene modifiers, or mast cell stabilizers). All children who did not report the use of beta-agonists indicated that they were using long-term control medication. Seventeen percent of children indicated that they were currently taking allergy medication as part of their asthma control. Of note, for the remaining eight children, we could not ascertain whether non-responses indicated lack of any medication or lack of recall/medication availability. There were additional factors and barriers that indicated potentially sub-optimal asthma care (Table 2 ). Although most (92%) children had current health insurance coverage that paid some portion of asthma-related medical expenses (with 97% covered by Medicaid/MassHealth during the past year), for 28% of children, the caregiver reported having no doctor to call other than the emergency room for asthma care. Only 37% of children had a written asthma action plan signed by their doctors. Fifty-four percent of children had a spacer to use with their inhalers. Furthermore, only 27% of children had a peak flow meter, with only 19% ever using it at home or school. Children with peak flow meters tended to be slightly older than those without peak flow meters, with a similar distribution of severity. As indicated in Table 2 , the asthma management practices often differed significantly across developments. Table 2 Access to medical care and asthma management practices for children in three public housing developments in Boston Franklin Hill West Broadway Washington Beech Total p-value (Wilcoxon rank-sum test) % with doctor to call other than emergency room 68% (N = 41) 89% (N = 27) 29% (N = 7) 72% (N = 75) 0.005 % with written asthma action plan 39% (N = 41) 46% (N = 24) 10% (N = 10) 37% (N = 75) 0.14 % with peak flow meter 28% (N = 40) 33% (N = 27) 10% (N = 10) 27% 0.37 % of persistent asthmatics using long-term control medication 21% (N = 19) 57% (N = 21) 14% (N = 7) 36% (N = 47) 0.03 In addition, 71% of caregivers indicated that at least one of seven barriers impeded asthma management for their children during the last six months. The most frequently cited barrier was that the pharmacy did not have their asthma medication (38%), followed by the asthmatic child either not being home when it was time to take the medicine (29%) or refusing to take the medicine (26%). The cost of the medication was cited as a barrier by only 12% of caregivers, the lowest percentage among the seven questions, an indication that MassHealth/Medicaid coverage was largely viewed as adequate. Social stressors On the five-point social cohesion scale, there was a significant difference across developments, with a lower mean value at Franklin Hill than at West Broadway or Washington Beech (Figure 2 ). There was, however, much greater variability in perceived social cohesion within rather than across developments, consistent with previous findings [ 24 ]. Figure 2 Social cohesion and perceived stress for caregivers of asthmatic children in HPHI. Social cohesion scores range from 1–5 (5 = maximum, 1 = minimum), while Cohen perceived stress scale scores range from 0–16 (16 = maximum, 0 = minimum). We also found significant differences across developments in exposure to violence. We asked caregivers whether they were afraid that they or their children would be hurt by violence in their neighborhood, whether they have had violence used against them or other household members in their neighborhood, and whether they fear letting their children play outside in their neighborhood because of community violence (Table 3 ). In all cases, the highest rates were reported at Franklin Hill. At Franklin Hill, but not at the other developments, the responses to these questions depended on the age of the asthmatic child. For example, 84% of caregivers of asthmatic children under 8 reported fear of violence, versus 29% of caregivers of asthmatic children 8 and older (with 54% of caregivers of children under 8 reporting having violence used against their household, versus 13% for caregivers of older children). In spite of these facts, there was no significant difference in the Cohen Perceived Stress Scale between developments, with more substantial within-development variability (Figure 2 ). Table 3 Exposure to violence for caregivers of asthmatic children in three public housing developments in Boston Franklin Hill West Broadway Washington Beech Total p-value (Wilcoxon rank-sum test) % afraid of violence in neighborhood 63% (N = 30) 20% (N = 20) 43% (N = 7) 46% (N = 57) 0.01 % directly impacted by violence in neighborhood 41% (N = 32) 14% (N = 22) 0% (N = 7) 26% (N = 59) 0.02 % not let children play outside due to violence in neighborhood 60% (N = 30) 23% (N = 20) 14% (N = 7) 41% (N = 59) 0.009 Asthma severity and symptoms As indicated in Table 4 , a majority of children reported having wheezing or tightness in their chest, needing to slow down or stop their activities due to their asthma, or having nighttime asthma symptoms within the last two weeks, an indication of poorly controlled asthma. In addition, six percent of the children were reported to have a severe asthma attack (unable to say more than one or two words between breaths) in the last two weeks. Three of the children (4%) reported staying overnight in the hospital for asthma during the last two months. Although a small fraction of the cohort, this would correspond (if sustained) to an annual asthma hospitalization rate of 23% in this population. Table 4 Frequency of reported asthma symptoms within two weeks prior to enrollment in intervention study Never 1–2 times/week 3–6 times/week At least daily Wheezing, tightness in the chest, or cough (N = 74) 20% 41% 24% 15% Slow down/stop play or activities (N = 74) 34% 35% 19% 12% Never 1–2 times 3–4 times At least 5 times Wake up at night (N = 76) 32% 34% 25% 9% Spirometry For the subset of 49 children age six or older able to perform spirometry, the mean FEV1% was 88% (median of 88%, standard deviation of 15%). Twenty-nine percent of children had FEV1 less than 80% of predicted, though no values were less than 60% of predicted. The mean PEF% was 97% (median of 96%, standard deviation of 17%). Twelve percent of children had PEF less than 80% of predicted, with none having PEF less than 60% of predicted. As would be anticipated, these two measures were well correlated (Spearman correlation of 0.69), with all but one of the children with PEF below 80% of predicted also having FEV1 below 80% of predicted. Of note, of the children under 6 tested, seven (30%) were able to record acceptable spirometry values. Only a child who had been recently hospitalized for asthma exacerbation had reduced FEV1 and PEF, both below 60% of predicted [ 31 ]. Given poor performance and issues in selecting appropriate reference equations for young children [ 34 ], spirometry for children under the age of 6 is not considered further in our analysis. Based on the four questions addressing recent asthma symptoms (as summarized in Table 4 ), spirometry measures, and medications prescribed and used, we determined that a majority of the children in our cohort (56% of those with complete information) would be considered to have moderate persistent asthma, with 14% considered severe persistent, 10% mild persistent, and 20% mild intermittent. Of note, of the 47 children categorized as having persistent asthma who provided information on their medications, only 36% reported being prescribed any daily controller medication (63% of severe persistent, 33% of moderate persistent, and 17% of mild persistent children). Quality of life For children, using the PAQLQ, the median score on the seven-point respiratory symptoms subscale was 4.6, with medians of 4.2 for activity limitation, and 5.1 for emotional function, with a median overall quality of life score of 4.6 (where a score of 1 indicates maximal impairment and a score of 7 indicates no impairment). The range across children was substantial, with overall quality of life scores ranging from 1.4 to 7 (Figure 3 ). Similarly, caregivers reported median PACQLQ scores of 4.3 for activity limitation and 4.6 for emotional function, with an overall median score of 4.5 but a range from 1.4 to 7. Using the EQ5D, the median health-related quality of life was 0.81 (range from 0.28 to 1), while the visual analogue scale (VAS) yields a median quality of life score of 80 (range from 30 to 100). The wide ranges indicate some of the limitations in interpreting these values cross-sectionally rather than longitudinally. Figure 3 Distribution of asthma-related quality of life scores for asthmatic children and their caregivers. Asthma-related quality of life scores range from 1–7 (7 = maximum, 1 = minimum). The quality of life questionnaires provide insight regarding the perceived burdens of asthma beyond the aggregate quality of life scores. For example, in the EQ5D responses, 31% of children were reported to be moderately or extremely worried or depressed, with 50% of children having problems doing their usual activities and 53% of children in moderate or extreme pain or discomfort. In addition, 39% of caregivers reported that their child's asthma interfered with their job or work around the house at least some of the time, with 62% reporting sleepless nights related to their child's asthma at least some of the time. Correlation and regression analyses One key question is whether any of the primary measures of asthma severity (lung function, quality of life, and respiratory symptoms) are significantly correlated with one another in a cross-sectional baseline assessment. For lung function, we consider both FEV1% and PEF%. Quality of life outcomes include the aggregate PAQLQ score for the child, aggregate EQ5D and VAS scores for the child, and the aggregate PACQLQ scores for the caregiver. For respiratory symptoms, we develop a symptom score reflecting the responses to the questions in Table 4 as well as having a severe asthma attack in the last two weeks. The symptom score ranges from zero to eight, with a maximum of two points assigned for each question and a higher score reflecting more frequent symptoms. While this scale is simple and implicitly places equal weight on respiratory outcomes of differing severity, it is similar to scales developed elsewhere and reasonably captures the gradient in symptom frequency and severity. Considering the Spearman correlations among these key covariates (Table 5 ), the lung function measures are strongly correlated with one another but are not significantly correlated with either the respiratory symptom score or quality of life. In contrast, the respiratory symptom score is significantly correlated with the VAS and both child and caregiver asthma-related quality of life, with lower quality of life given higher symptom frequency as anticipated. The various quality of life scales are generally significantly correlated with one another, with slightly weaker relationships for the EQ5D scale, which is not specific to asthma. Not surprisingly, the correlation between caregiver and child asthma-related quality of life is stronger for children age 7 and under, where the caregiver evaluates the child's quality of life (r = 0.75), than for children age 8 and older, where the child evaluates their own quality of life (r = 0.31). Table 5 Spearman correlation coefficients between respiratory symptom score, quality of life measures, and lung function Symptom score EQ5D VAS Child AQL Caregiver AQL FEV1% EQ5D -0.07 - - - - - VAS -0.29 * 0.20 - - - - Child AQL -0.43 ** 0.44 ** 0.43 ** - - - Caregiver AQL -0.46 ** 0.27 * 0.28 * 0.49 ** - - FEV1% -0.12 0.14 -0.24 -0.08 -0.07 - PEF% -0.03 0.05 -0.24 -0.09 0.08 0.65 ** *: p < 0.05, **: p < 0.01 Given our sample size, missing data for selected covariates (such as spirometry or allergy status), and the number of factors hypothesized to influence our outcome measures, we consider our regression analyses to be exploratory in nature. The goal is to better understand the above correlations and factors that might influence the relationships among our outcome measures. For each of the outcome measures, we evaluate a subset of demographic variables (age, race/ethnicity, gender, household size, housing development), intrinsic risk factors (BMI, being in the intensive care unit at birth, eczema), health care risk factors (having a doctor to call other than the emergency room), physical risk factors (allergies to roaches, dust mites, or any agents; or environmental tobacco smoke exposure), and social risk factors (social capital, perceived stress, fear of violence in the neighborhood, and not letting children play outside due to fear of violence in the neighborhood). It should be noted that some of these covariates may be direct causative agents, while others represent proxies or outcomes that could be influenced by asthma severity (such as caregiver stress). The results from this analysis are summarized in Table 6 , including significant terms from univariate and multivariate regressions. Although multiple risk factors were predictive of low FEV1% in univariate regressions (including age, BMI, cockroach allergy, environmental tobacco smoke exposure, and social capital), no terms were statistically significant in multivariate models. The respiratory symptom score was elevated for a variety of physical and social risk factors, with household size, BMI, and environmental tobacco smoke exposure remaining significant in multivariate models. Fewer factors were predictive of child asthma-related quality of life, but the respiratory symptom score was strongly and negatively associated with the PAQLQ score. Similarly, the respiratory symptom score strongly predicted caregiver asthma-related quality of life in both univariate and multivariate models, with a significant influence in multivariate models for the caregiver's perceived stress and for whether the child required NICU care at birth. Table 6 Univariate/multivariate regressions of FEV1%, respiratory symptoms, and quality of life measures on selected risk factors FEV% < 80% Respiratory symptom score Child asthma-related quality of life Caregiver asthma-related quality of life Age 0.05 (+) NS NS NS Hispanic NS NS NS NS African-American NS NS NS NS Gender NS NS NS NS Household size NS 0.0007 (+) 0.009 (+) NS 0.04 (-) Housing development NS NS NS NS BMI 0.12 (+) 0.03 (+) 0.02 (+) NS NS Born in NICU NS NS NS 0.09 (-) 0.03 (-) Eczema NS 0.13 (+) NS NS Doctor to call other than ER NS NS NS NS Allergy to roaches 0.18 (+) NS NS 0.19 (+) Allergy to dust mites NS NS NS NS Any allergies NS NS NS NS Environmental tobacco smoke exposure 0.19 (-) 0.03 (+) 0.08 (+) NS NS Social capital 0.15 (+) NS NS NS Perceived stress NS 0.03 (+) 0.04 (-) 0.001 (-) 0.004 (-) Fear of violence in neighborhood NS 0.06 (+) NS 0.01 (-) Not letting children play outside due to fear of violence in neighborhood NS 0.10 (+) NS NS Low FEV% 0.15 (+) 0.13 (+) 0.04 (+) NS Respiratory symptom score 0.0002 (-) 0.02 (-) < 0.0001 (-) 0.009 (-) NS: No statistical significance in univariate regression (p > 0.2) Value in italics : Statistically significant in univariate regression (p < 0.2), but not in multivariate regression. The value presented is the p-value for the univariate regression, and the +/- sign indicates the direction of the relationship Values in bold : Statistically significant in multivariate regression (p < 0.1). The first value presented is the p-value for the univariate regression, and the second value presented is the p-value for the multivariate regression. The +/- sign indicates the direction of the relationship. Discussion To address our first hypothesis and interpret this baseline characterization of asthmatic children in public housing enrolled in an intervention study, it is instructive to compare the prevalence of selected risk factors with those reported in previous studies [ 20 , 25 , 35 - 39 ] (Table 7 ). Within our study, there appear to be a greater percentage of overweight children and children with a family history of asthma, as compared with NCICAS, other studies of low-income asthmatics, and general population studies. The fraction of children with cockroach or dust mite allergies is also high, although Alternaria allergy prevalence is quite low, indicating that the most effective interventions might differ between HPHI and NCICAS. In addition, exposure to violence and fear of violence is slightly higher than reported in NCICAS. Although the fraction of children with persistent asthma who are adequately medicated is quite low (36%) and indicative of poorly managed asthma, this figure is actually higher than the percentages reported in other studies, indicating that this is a pervasive problem in asthma management. Table 7 Comparison of children in HPHI intervention study with children in other studies HPHI NCICAS [20] Other low-income asthmatic populations Population-based studies Family history of asthma 70% 57% - N/A % overweight 56% 19% - 15% [39] % of families with at least one smoker 39% 59% 46% [35] 41% [36] % with cockroach allergy 59% 36% 52–78% [1] 22% [36] % with European dust mite allergy 59% 31% - 27% [36] % with Alternaria allergy 7% 38% - 16% [36] % not let children play outside due to violence in neighborhood 41% 34% [25] - - % of persistent asthmatics on long-term control medication 36% 24% 1 27% [37] 26% [38] 2 1 All asthmatics 2 Moderate/severe asthmatics only More broadly, our findings regarding medication usage coupled with the prevalence of children with mild, moderate, or severe persistent asthma and the frequency of respiratory symptoms suggest that this population is being treated sub-optimally. NHLBI guidelines [ 32 ] indicate that all individuals with persistent asthma should be prescribed an inhaled steroid or other long-term control medication in addition to a fast-acting beta-agonist. In addition, for moderate or severe persistent asthmatics, it is recommended that long-acting bronchodilators be added to the medication regimen. Compounding the problem of inadequate medication is the relatively low usage of asthma action plans or peak flow meters, which are both part of recommended patient education and self-management activities. These shortfalls could be related to inadequate quality of care, limitations in access to and continuity of care, communication gulfs between caregivers and providers, or other factors, and further investigation is needed to determine the root causes of this management gap. Although exploratory in nature, our correlation and regression analyses provide some useful insights regarding the factors associated with various measures of asthma severity in our cohort. FEV1% was not strongly correlated with other health measures, and was only weakly associated with a subset of risk factors in univariate regressions. As this measurement was taken at a single point in time at varying times of day and seasons, a weak relationship is unsurprising, especially when compared with symptoms over a two week period [ 18 ]. The lack of a relationship may also be related to a relatively small sample size and a small number of participants with low FEV1%. In addition, many asthmatics may be poor perceivers of symptoms of respiratory difficulty, potentially explaining the disconnect between pulmonary function measures and asthma symptoms or quality of life. The respiratory symptom score was moderately associated with a number of risk factors in univariate regressions, with all coefficients in the anticipated directions (i.e., greater symptom frequency and severity with larger household sizes, higher BMI, reported eczema, higher environmental tobacco smoke exposure, higher psychosocial stress, higher fear of violence, and higher prevalence of low FEV1%). The robust multivariate relationship with household size and presence of smokers could be indicative of the influence of indoor air quality or respiratory infection related to unit crowding, with the association between obesity and asthma symptom frequency and severity in agreement with past studies [ 40 , 41 ]. The extremely strong relationship between respiratory symptom score and asthma-related quality of life for the child is logical and provides some indication that both measures were reasonably constructed. The relationship between perceived stress and the caregiver's asthma-related quality of life, in addition to the child's respiratory symptoms, may indicate that the acute stresses associated with lack of control over one's surroundings could adversely affect quality of life. Since the NICU term is associated with caregiver's quality of life (but not the child's quality of life), it is less likely a surrogate of early respiratory injury and could be more broadly related to chronic stress about a child's health status. It is also interesting to note that variables such as race, ethnicity, gender, and housing development were not significant predictors in any models. There are some clear limitations in interpreting the results of our investigation. First, given the numerous risk factors for asthma, both the univariate and multivariate regressions must be interpreted with caution. For example, fear of violence in the neighborhood was significantly associated with reductions in caregiver quality of life in univariate regressions (p = 0.01), but given a strong positive association between fear of violence and caregiver stress as well as the child's respiratory symptom score, this term did not enter into the final regression model. Thus, both the significant and insignificant terms from our regression models should be interpreted with caution, as they may not reflect causal relationships. In addition, the participants in our study represented a convenience sample of individuals from three selected public housing developments in Boston, who were interested in enrolling in a longitudinal intervention study. It is unclear whether our participants are representative of asthmatic children across those developments, or more broadly, asthmatic children in public housing. Families who enrolled in our study may have been more desperate for help given current asthma conditions, or may have had greater confidence in the ability of a research project to improve their child's health. A comparison between the demographics of our study population and the demographics of the developments indicates that the age distribution and racial/ethnic composition are similar, but there is no way of knowing if our cohort represents typical characteristics of all asthmatic children across the BHA. In addition, we have evaluated the prevalence of risk factors and correlations among key indicators of respiratory health in a cross-sectional survey without a defined control group. Thus, while we have found an elevated prevalence of obesity and allergies to cockroach and dust mites when compared with other populations, we cannot infer a causal relationship with asthma. However, this comparison illustrates the relative importance of various risk factors internal to our intervention study. In addition, the correlations among respiratory health measures will not necessarily be identical cross-sectionally and longitudinally. For an intervention study, the crucial question is whether observed changes in lung function would occur simultaneously with changes in quality of life or respiratory symptoms, and knowledge of the cross-sectional correlations is not necessarily informative. That being said, the correlations among health endpoints (Table 5 ) and the regression results (Table 6 ) do provide some indication that children with more frequent respiratory symptoms have lower quality of life at baseline, making these children candidates for improvements in both health endpoints. In spite of these limitations, our study provides some important and unique information. To our knowledge, this is the most comprehensive evaluation focused on asthmatic children in public housing to date. Although many attributes of public housing are similar to low-income private housing, the viability of large-scale interventions in public housing based on common indoor environmental exposures and centralized management makes it important to characterize similarities and differences from other low-income populations. More broadly, our study demonstrates that a community-based participatory research paradigm, with members of the community conducting most of the primary data collection, is able to gather valid and meaningful data. Given that recruitment and retention of our study population would have been quite difficult without involvement from community members, it is clear that community-based participatory research is essential for detailed evaluations of asthma in public housing. The findings from this cross-sectional baseline evaluation of asthmatic children in public housing in Boston have multiple policy implications. The inadequacy of medical care for a majority of asthmatic children indicates that medical interventions might yield substantial improvements in asthma status for poor children. The fact that significant differences existed in adequacy of medical care for asthma across housing developments indicates likely variability across providers (although the data indicate that this problem exists across community health centers and academic medical centers). In addition, the high prevalence of cockroach and dust mite allergies indicates that interventions aimed at reducing or eliminating these triggers are likely to provide health improvements. This reinforces the expectation that interventions in public housing, a setting with high asthma prevalence and high prevalence of allergic responses to indoor contaminants, are likely to be meaningful and effective. Finally, the responses to the psychosocial questions as well as the percentage of caregivers who feared neighborhood violence and modified their children's activities as a result indicates that violence and stress are substantial risk factors for both respiratory health and other outcomes. Our findings point to the need for coordinated improvements in the social environment, the built environment, and in medical management. Future investigations should similarly evaluate asthma risk factors and severity in other public housing settings to determine whether our conclusions provide generalizable and relevant information for regional or national housing authorities in considering intervention strategies. In addition, a longitudinal comparison of correlations among measures of asthma severity would help determine whether conclusions drawn from a cross-sectional evaluation are robust. Conclusions We conclude that asthmatic children enrolled in a public housing-based intervention study would likely benefit from a coordinated intervention focused on reduction of indoor allergens (especially cockroach and dust mites), improved medical management, and increased social support. List of Abbreviations AQL – Asthma-related quality of life BHA – Boston Housing Authority BMI – Body mass index FEV1 – Forced expiratory volume in one second FEV1% – Percent of predicted forced expiratory volume in one second HPHI – Healthy Public Housing Initiative NCICAS – National Cooperative Inner-City Asthma Study NHLBI – National Heart, Lung, and Blood Institute NICU – Neonatal intensive care unit PACQLQ – Paediatric Asthma Caregiver's Quality of Life Questionnaire PAQLQ – Paediatric Asthma Quality of Life Questionnaire PEF – Peak expiratory flow PEF% – Percent of predicted peak expiratory flow VAS – Visual analogue scale on EuroQol EQ5D quality of life questionnaire Competing interests The authors declare that they have no competing interests. Authors' contributions JIL participated in study design and implementation, conducted statistical analyses, and drafted the manuscript. LKWH designed and administered survey instruments, provided support to participants as a nurse/case manager, and contributed to analyses related to adequacy of health care. JEC contributed to survey design and analyses for psychosocial risk factors. RED completed a literature review, drafted text, and conducted statistical analyses. SS provided asthma severity classifications and related analyses and conducted allergy testing. HPH collaborated in the conception and design of the study, and co-directed data collection and study coordination. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544563.xml
532393
Replication of Norovirus in Cell Culture Reveals a Tropism for Dendritic Cells and Macrophages
Noroviruses are understudied because these important enteric pathogens have not been cultured to date. We found that the norovirus murine norovirus 1 (MNV-1) infects macrophage-like cells in vivo and replicates in cultured primary dendritic cells and macrophages. MNV-1 growth was inhibited by the interferon-αβ receptor and STAT-1, and was associated with extensive rearrangements of intracellular membranes. An amino acid substitution in the capsid protein of serially passaged MNV-1 was associated with virulence attenuation in vivo. This is the first report of replication of a norovirus in cell culture. The capacity of MNV-1 to replicate in a STAT-1-regulated fashion and the unexpected tropism of a norovirus for cells of the hematopoietic lineage provide important insights into norovirus biology.
Introduction Viruses within the genus Norovirus (formerly “Norwalk-like viruses”) of the family Caliciviridae are major agents of acute gastroenteritis ( Green et al. 2001 ). Norovirus research, including the development of prevention and control strategies, has been hampered by the failure to grow these viruses in cultured cells despite extensive efforts ( Duizer et al. 2004 ). Most noroviruses identified thus far have been associated with gastrointestinal disease in humans, but members of the genus have been found in other species as well ( Green et al. 2001 ; Karst et al. 2003 ). Our recent discovery of the first murine norovirus, murine norovirus 1 (MNV-1), and demonstration of its ability to infect the intestinal tract of mice following oral inoculation provided an opportunity to analyze the pathogenesis of this norovirus in mice ( Karst et al. 2003 ). This previous study demonstrated that the cellular transcription factor STAT-1 and interferon (IFN) receptors are critical for resistance to MNV-1 infection in vivo. The availability of MNV-1 and STAT1-deficient (STAT1 −/− ) mice ( Durbin et al. 1996 ; Meraz et al. 1996 ) that are highly susceptible to MNV-1 infection allowed us to revisit efforts to develop a cell culture system for noroviruses. Here we show for the first time that MNV-1 grows in macrophages (MΦ) and dendritic cells (DCs) and provide the first tissue culture model for a norovirus. Using this model we demonstrate that MNV-1 growth in vitro was inhibited by the IFN-αβ receptor and STAT-1. In addition, we isolated the first three-times plaque-purified strain of MNV-1 (MNV-1.CW1) and characterized it in vitro and in vivo. Sequencing of serial passages of MNV-1.CW1 indicated remarkable sequence stability over time and indicated that an amino acid substitution in the capsid protein of serially passaged MNV-1 was associated with a loss of virulence in vivo. Results/Discussion MNV-1 Replicates in Murine MΦ and DCs As part of our ongoing investigation into MNV-1 pathogenesis, STAT1 −/− mice were infected with MNV-1 by the oral route and tissue sections analyzed by immunohistochemistry for the presence of MNV-1 protein. MNV-1-specific staining was observed in spleen and liver 2 d postinfection ( Figure 1 ). Interestingly, in the liver, Kupffer cells (resident macrophages of the liver) lining the sinusoids were specifically stained by MNV-1 immune serum (compare Figure 1 A and 1 B). In the spleen, staining was found primarily in the red pulp and the marginal zone, but also in non-lymphoid cells within the white pulp ( Figure 1 C and 1 D). This pattern is consistent with staining of MΦ and DCs ( Metlay et al. 1990 ; Leenen et al. 1998 ). Furthermore, in some cases virus-antigen-positive MΦ were detected ( Figure 1 C). Figure 1 MNV-1-Specific Staining In Vivo Occurs in Cells of the MΦ Lineage Immunohistochemistry was performed on liver (A and B) and spleen (C and D) sections from STAT1 −/− mice 2 d after oral infection. MNV-1-specific staining was seen in Kupffer cells of infected livers when probed with MNV-1 immune (A) but not preimmune (B) serum. A selected Kupffer cell lining the sinusoid is indicated by an arrowhead. MNV-1-specific staining consistent with MΦ was seen in red pulp (C) and marginal zone (D) in the spleen. The arrow indicates a cell with MΦ morphology. No staining was observed in tissues from mice infected for 1 d, in infected tissues incubated with preimmune serum, or in mock-infected tissues incubated with immune serum. RP, red pulp; WP, white pulp. Because cells containing viral antigen in infected mice resembled MΦ, we examined whether cells of the hematopoietic lineage such as MΦ and DCs were permissive for MNV-1 replication in vitro. Bone-marrow-derived MΦ (BMMΦ) and bone-marrow-derived DCs (BMDCs) were inoculated with a MNV-1 stock derived from the brain of infected IFNαβγ receptor −/− (IFNαβγR −/− ) mice ( Karst et al. 2003 ). Cytopathic effect (CPE) in cell monolayers was visible within 2 d in STAT1 −/− BMMΦ and BMDCs, but not STAT1 −/− murine embryonic fibroblasts (MEFs) ( Figure 2 A). While BMDCs showed CPE even when STAT-1 was present, wild-type (wt) BMMΦ exhibited less CPE than their STAT1 −/− counterparts. These data showed that MNV-1 had a marked tropism for MΦ and DCs but not fibroblasts. Figure 2 MNV-1 from Brain Homogenate Replicates in Cells of the DC and MΦ Lineage In Vitro BMDCs and BMMΦ, as well as MEFs from wt or STAT1 −/− mice, and RAW 264.7 cells were infected with a MOI of 0.05. (A) MNV-1 causes CPE in permissive cells. MNV-1- or mock-infected cells were observed by light microscopy 2 d postinfection. The boxed area is magnified further to show the border of the plaque. (B) Infected cell lysates were analyzed in two to four independent experiments by plaque assay at various timepoints postinfection to calculate standard deviations. For wt BMMΦ, MNV-1 growth was detected in two out of four experiments. We used this information to screen available MΦ cell lines for growth of MNV-1, including the murine lines RAW 264.7 ( Raschke et al. 1978 ) and J774A.1 ( Ralph et al. 1975 ), and the human/murine hybrid line WBC264-9C ( Aksamit 1986 ). These cells also showed visible CPE when inoculated with the MNV-1 stock ( Figure 2 , data not shown). Plaques were observed when infected RAW 264.7 monolayers were maintained under agarose ( Figure 2 A), allowing us to develop a plaque assay and quantitate virus titers. STAT1 −/− BMMΦ, STAT1 −/− and wt BMDCs, and RAW 264.7 cells consistently supported the growth of MNV-1, while wt BMMΦ varied in their ability to support virus growth ( Figure 2 B). BMMΦ and BMDCs cells lacking STAT-1 always yielded higher MNV-1 titers than their wt counterparts. Furthermore, a low level of virus replication was observed in STAT1 −/− MEFs, but as reported previously, no virus growth was observed in wt MEFs ( Karst et al. 2003 ). MNV-1 replication proceeded rapidly in permissive cells, with newly synthesized infectious virions first detected in cell lysates 9 to 12 hours postinfection (h.p.i.). Taken together, these data indicated that MNV-1 could productively infect MΦ and DCs. Verification of Viral Growth In Vitro Several approaches were used to verify that the observed CPE and plaques were caused by MNV-1. We first performed a clonal selection from the MNV-1 stock (from infected brain tissue) with three rounds of plaque purification in RAW 264.7 cells to generate the MNV-1.CW1 strain. This strain was amplified in RAW 264.7 cells, after which virus particles were concentrated and subjected to purification by isopycnic centrifugation in CsCl. A distinct band was observed in CsCl gradients at a density of 1.35 ± 0.01 g/cm 3 , consistent with that described for noroviruses ( Kapikian et al. 1996 ). Examination of the material in this fraction by negative staining electron microscopy showed the presence of virus particles with calicivirus morphology ( Figure 3 A). Furthermore, SDS-PAGE analysis of this material revealed a major protein of approximately 59 kDa, consistent with the calculated mass of the MNV-1 capsid protein ( Figure 3 B,C). Western blot analysis with antibodies generated against bacterially expressed MNV-1 capsid protein ( Figure 3 B) and mass spectrometry (data not shown) confirmed its identity as the MNV-1 capsid protein. A genomic-sized RNA molecule of approximately 7.4 kb was detected in nucleic acid isolated from the purified virions with a probe specific for the MNV-1 genome in Northern blots (data not shown). Finally, a neutralization assay was performed with the monoclonal antibody (mAb) A6.2 specific for the MNV-1 capsid protein (see Materials and Methods). MAb A6.2 specifically bound to CsCl-purified MNV-1 virions in an immunoassay, while the isotype-matched mAb 10H2, an anti-reovirus μ1c mAb ( Virgin et al. 1991 ), did not bind ( Figure 3 D). MAb A6.2, but not the isotype control antibody 10H2, showed neutralization activity in a plaque reduction assay for both the virus in the original brain homogenate (MNV-1), and the three-times plaque-purified strain MNV-1.CW1 ( Figure 3 E). Together these data confirmed that MNV-1 was the infectious agent associated with viral growth observed in the infected cell cultures. Figure 3 Characterization of the Triple Plaque-Purified Strain MNV-1.CW1 (A–C) MNV-1.CW1 purified on CsCl density gradients was visualized by (A) negative staining electron microscopy, (C) Coomassie staining, and (B) Western blot analysis with a polyclonal anti-MNV-1-capsid antibody. Molecular weight markers are indicated in kiloDaltons. (D) Specific binding of mAb A6.2 to two different concentrations of CsCl-purified MNV-1 particles in an enzyme-linked immunosorbent assay. (E) Neutralization of MNV-1 from brain homogenate and MNV-1.CW1 by mAb A6.2 but not the isotype control (10H2) mAb in a plaque neutralization assay. The assay was repeated three times to calculate standard deviations. The limit of detection is indicated by the dashed line. (F) Timecourse of viral RNA synthesis in RAW 264.7 cells. Northern blot analysis of viral RNA from cells infected with MNV-1.CW1 (MOI of 2.0) or mock-infected cells. The size of RNA markers in kilobases is shown on the left. The positions of subgenomic- and genomic-length RNA are indicated on the right. This timecourse is a representative of two independent experiments. (G) Timecourse of viral protein synthesis in infected RAW 264.7 cells. MNV-1-specific proteins were precipitated from radiolabeled cell lysates of MNV-1.CW1-infected RAW 264.7 cells (MOI of 2.0) at indicated times after infection. The size of the proteins in kiloDaltons is indicated. MNV-1 RNA and Protein Production in Permissive Cells To compare MNV-1 replication in cells with that of other caliciviruses, we analyzed viral RNA and protein synthesis in MNV-1.CW1-infected RAW 264.7 cells. Northern blot analysis using a probe specific for the positive strand of the MNV-1 genome showed an increase in the accumulation of full-length (7.4 kb) and subgenomic-length (2.3 kb) MNV-1 genome over time ( Figure 3 F). Radiolabeled MNV-1-infected RAW 264.7 cell lysates were analyzed by immunoprecipitation with serum from a MNV-1 infected mouse, and a 59-kDa protein consistent with the capsid protein was detected as early as 6 h.p.i. ( Figure 3 G). Additional proteins accumulated over time that corresponded in size to expected calicivirus nonstructural proteins such as the 76-kDa proteinase-polymerase precursor and an approximately 40-kDa NTPase protein ( Sosnovtsev et al. 2002 ). These data showed that the viral RNA and proteins synthesized in infected cells were consistent with calicivirus replication ( Green et al. 2001 ). Ultrastructural Examination of MNV-1-Infected RAW 264.7 Cells Positive-strand RNA viruses ( Dales et al. 1965 ; Mackenzie et al. 1999 ; Pedersen et al. 1999 ), including caliciviruses ( Love et al. 1975 ; Studdert and O'Shea 1975 ; Green et al. 2002 ), are known to replicate in association with intracellular membranes. Therefore, we examined the ultrastructural morphology of MNV-1.CW1-infected RAW 264.7 cells ( Figure 4 ). Over time, virus-infected cells showed a striking change in overall morphology and intracellular organization ( Figure 4 D– 4 L) compared to mock-infected cells ( Figure 4 A– 4 C). Structures resembling virus particles were observed within or next to single- or double-membraned vesicles in the cytoplasm by 12 h.p.i. ( Figure 4 D). The vesiculated areas increased in size with time ( Figure 4 G– 4 I), and by 24 h.p.i., large numbers of these vesicles and viral particles occupied most of the cytoplasm, displacing the nucleus ( Figure 4 J– 4 L). In addition, a complete rearrangement of intracellular membranes with some confronting membranes occurred ( Figure 4 J), leading to a rearrangement of the endoplasmic reticulum and loss of an intact Golgi apparatus ( Figure 4 E; data not shown). Interestingly, these smooth-membraned vesicles were often surrounded by mitochondria. A small proportion of cells also showed crystalline arrays of cytoplasmic virus particles (data not shown). These observations indicate that like other positive-strand RNA viruses, norovirus RNA replication likely occurs in association with intracellular membranes. Figure 4 Ultrastructural Studies of MNV-1.CW1-Infected RAW 264.7 Cells Cells were infected with MNV-1.CW1 (P3) (MOI of 2.0) (D–L) or mock-infected (A–C) and processed for electron microscopy 12 (D–F), 18 (G–I), or 24 (A–C; J–L) h.p.i. MNV-1 particles are indicated by arrows and confronting membranes by arrowheads. VA, vesiculated areas; Nuc, nucleus; rER, rough endoplasmic reticulum. Scale bars, 200 nm for (A), (D), (G), and (J); 500 nm for (B), (E), (H), and (K); 2 μm for (C), (F), (I), and (L). Characterization of the Plaque-Purified Strain MNV-1.CW1 In Vitro To determine whether the plaque purification and sequential amplification of MNV-1 in RAW 264.7 cells had altered its growth characteristics, different cell types were infected with passage (P) 3 of MNV-1.CW1. In general, the growth of MNV-1.CW1 (P3) in wt or STAT1 −/− MΦ and MEFs ( Figure 5 A) as well as RAW 264.7 cells (data not shown) was similar to that observed for the original parental MNV-1 virus stock (compare Figure 2 B and 5 A). Virus titers were reproducibly higher in STAT1 −/− cells compared to wt cells, and MNV-1.CW1 (P3) growth was consistently observed in wt BMMΦ. These data demonstrated that our plaque purification and serial passage in RAW 264.7 cells had not changed the tropism of the virus for primary DCs and MΦ and confirmed the importance of STAT-1 in controlling MNV-1 growth at the cellular level. Figure 5 Critical Role for STAT-1 in Limiting MNV-1 Growth In Vitro (A) MNV-1.CW1 has no defect in viral growth in vitro. Growth curves (MOI of 0.05) were performed two or three times with MNV-1.CW1 (P3) on indicated cells to calculate standard deviations. (B) MNV-1 growth in MΦ is controlled by STAT-1 and Type I IFNs. BMMΦ of the indicated genotype were infected with MNV-1.CW1 (P3) at the indicated MOI. The experiment was performed twice to calculate standard deviations. The p -values for PKR versus wt infection at MOI 0.05 and 2.0, 0.8867 and 0.1616, respectively, are not significant. Statistical analysis was performed using the paired t -test (GraphPad Prism, version 3.03). Cellular Factors Controlling MNV-1 Growth In Vitro Previous studies demonstrated that a lack of STAT-1 or both IFNαβR and IFNγR increase susceptability to MNV-1 infection. Mice lacking individual IFNR, inducible nitric oxide (iNOS) −/− , or protein kinase R (PKR) −/− are not susceptible (Karst at al. 2003). Therefore, we determined whether molecules other than STAT-1 exhibited antiviral effects at the level of the infected cell. Primary BMMΦ from wt mice or mouse strains deficient in STAT-1, IFNαβR, IFNγR, IFNαβγR, iNOS, or PKR were directly compared for their ability to support virus replication at two different multiplicities of infection (MOIs) ( Figure 5 B). Again, BMMΦ cells from both wt and STAT1 −/− mice supported MNV-1 virus replication, with higher titers observed in cells deficient in STAT-1. Cells obtained from mice lacking both Type I and II IFNR (IFNαβγR −/− ) or Type I IFNR alone (IFNαβR −/− ) supported replication of virus as efficiently as STAT1 −/− cells. In addition, wt BMMΦ and wt BMDCs secrete IFNα after MNV-1-infection, as determined by IFNα enzyme-linked immunosorbent assay (ELISA) (data not shown). This is consistent with a direct role for IFN signaling in MNV-1 growth but does not rule out the possibility that effects of STAT-1 and IFNαβR occur in vivo prior to explantation of the bone marrow. Absence of IFNγR, iNOS, or PKR did not have a statistically significant effect on MNV-1 growth in BMMΦ. Together, these data demonstrate that the antiviral molecules STAT-1 and IFNαβ are part of a cellular response that limits norovirus growth. Characterization of the Plaque-Purified Strain MNV-1.CW1 In Vivo To address the effects of cell culture adaptation on virulence, STAT1 −/− mice were infected orally with MNV-1.CW1 from three successive passages (P1, P2, and P3) ( Figure 6 A). Oral administration of MNV-1.CW1 (P1) resulted in lethal infection, similar to that previously reported for the parental MNV-1 brain tissue stock ( Karst et al. 2003 ). These data fulfill a Koch's postulate with regard to MNV-1 infection and are consistent with the identification of MNV-1 as the infectious agent that was passaged in animals in our initial studies ( Karst et al. 2003 ). In contrast, MNV-1.CW1 (P3) failed to cause a lethal infection in STAT1 −/− mice after oral inoculation, even when administered a dose of 1.5 × 10 6 plaque-forming units (pfu), 5,000 times greater than the lethal dose for P1. In addition, immunohistochemical analysis of sectioned spleen and liver from STAT1 −/− mice infected orally with 1.5 × 10 6 pfu of MNV-1.CW1 (P3) did not reveal any MNV-1-specific staining, unlike the parental virus (see Figure 1 , data not shown). This striking difference in virulence and decrease of viral antigen in infected mice, coupled with an intermediate lethality phenotype of the MNV-1.CW1 (P2) virus, showed that serial passage of the virus in cell culture could attenuate MNV-1 virulence in vivo. Figure 6 Changes in Virulence of Plaque-Purified MNV-1 over Multiple Passages Are Associated with Limited Amino Acid Changes (A) Serial passage of MNV-1.CW1 in cell culture causes attenuation. STAT1 −/− mice were infected orally with the indicated virus dose. The number of mice analyzed is indicated in parentheses. (B) Summary of sequence analysis of MNV-1 over several passages. The nucleotide and amino acid differences between the indicated viruses are shown (for detail see Table 1 ). Table 1 Sequence Analysis of MNV-1 over Several Passages Nucleotides are numbered according to consensus sequence of the parental MNV-1 virus genome (in brain tissue stock) as follows: ORF1 (nt 6–5,069), ORF2 (nt 5,056–6,681), and ORF3 (nt 6,681–7,307), encoding a large polyprotein (viral nonstructural proteins), VP1 (major capsid structural protein), and VP2 (minor capsid structural protein), respectively. Amino acid residues are numbered according to location in the corresponding ORF. The nucleotide position of interest is underlined and its location in the codon of the translated ORF is shown. Sequence heterogeneity at a particular residue was determined from the sequence chromatogram, and the data shown represent direct sequence analysis of PCR-amplified cDNA products. A change in deduced amino acid sequence from the previous passage is indicated in parentheses Molecular Analysis of Serially Passaged MNV-1.CW1 To examine the molecular basis for this attenuation, consensus sequence analysis was performed on the RNA genome of MNV-1 present in the original brain tissue stock (parental virus), and in viruses from each subsequent cell culture passage of MNV-1.CW1 (P1 through P3) ( Figure 6 B; Table 1 ). Three nucleotide changes occurred between the parental virus and P1, with one of these resulting in an amino acid substitution (histidine to arginine) at residue 845, located within the predicted “3A-like” region of the nonstructural polyprotein. In the P2 virus, which retained virulence but at a reduced level compared to the parental and P1 viruses, a second nucleotide substitution within the predicted “3A-like” coding region was observed that caused an amino acid change (valine to isoleucine) at residue 716. The partial attenuation of virulence of the P2 virus in vivo is of interest since the homologous protein in poliovirus, the 3A protein, alters the amount of cytokines secreted from cells, with likely effects on viral pathogenesis ( Dodd et al. 2001 ). Of note, a mixed population of A and G nucleotides was detected at position 5,941 of the P2 viral genome that could potentially yield two populations of virus with either amino acid lysine or glutamic acid at residue 296 of the capsid protein. In the P3 virus, which was avirulent in mice, the G nucleotide sequence at position 5,941 emerged as the predominant sequence. The resulting amino acid substitution was of interest because of its location within the hypervariable P2 domain, which contains the putative receptor-binding site ( Prasad et al. 1994 , 1999 ; White et al. 1996 ). However, altered virus binding to permissive cells cannot explain the attenuated phenotype since the parental virus and MNV-1.CW1 (P3) replicate to similar levels in BMDCs and BMMΦ in vitro. Similar to our findings, the P2 domain was also implicated in attenuation of porcine enteric calicivirus virulence ( Guo et al. 2001 ). This study suggests that the norovirus capsid protein, especially the hypervariable P2 domain, and possibly the 3A-like protein, may be important sites for the development of virulence-attenuating mutations. Conclusion Detection of MNV-1-positive cells of the MΦ and DC lineage in infected organs of STAT1 −/− mice led to our finding that MNV-1 grows in these cell types in vitro. This provides the first tissue culture model for a norovirus. In addition, the antiviral Type I IFN response with signaling through STAT-1 is crucial for resistance to murine norovirus infection in vivo and in vitro. Taken together the previous in vivo data ( Karst et al. 2003 ) and the tropism of this norovirus for cells of the innate immune system, underscore the importance of the innate immune response, specifically STAT-1 and Type I IFNs, in resistance against norovirus infection. These data may aid the development of a culture system for human noroviruses since neither cells of the MΦ/DC lineage nor cells with defects in the Type I IFN/STAT-1 antiviral pathway have likely been investigated ( Duizer et al. 2004 ). Furthermore, this MNV-1 tissue culture model will help elucidate stages of the viral life cycle and cellular factors essential for norovirus replication that may provide targets for prevention or control of an important human disease. The demonstration of a tropism of MNV-1 for DCs was unexpected, as a calicivirus tropism for DCs has not been previously described. However, like MNV-1, other caliciviruses do interact with MΦ. Viral RNA from rabbit hemorrhagic disease virus, a lagovirus, was detected in splenic and alveolar MΦ by in situ hybridization ( Kimura et al. 2001 ). In addition, feline calicivirus, a vesivirus, showed a small, transient increase in viral titers in alveolar MΦ cultures, indicative of abortive infections ( Langloss et al. 1978 ). It is possible that MΦ contribute to the spread of the virus through the host, but it must be noted that MΦ supported MNV-1 growth to a lower extent than DCs unless they lacked specific immune defense molecules such as STAT-1. This argues that MΦ may be the cell through which STAT1-dependent innate immunity limits MNV-1 virulence ( Karst et al. 2003 ). In contrast to MΦ, DCs were permissive even when STAT-1 was present (see Figure 2 ). DCs are sentinels of the immune system whose function is to acquire antigens and stimulate lymphocytes. Intestinal DCs are found in the gut in specialized lymphoid tissues where they can sample enteric antigens by extending their dendrites into the gut lumen ( Stagg et al. 2003 ; Kraehenbuhl and Corbett 2004 ). We therefore speculate that DCs in humans and mice provide noroviruses access to subepithelial regions of the intestine, thereby contributing to norovirus disease pathogenesis. Further studies are in progress to address the role of MΦ and DCs in the intestine and the physiologic relevance of these cells for MNV-1 pathogenesis in general. Material and Methods Cell cultures and mice MEFs were generated and cultured as described previously ( Pollock et al. 1997 ). RAW 264.7 cells were purchased from ATCC (Manassas, Virginia, United States) and maintained in DMEM (Cellgro, Mediatech, Herndon, Virginia, United States) supplemented with 10% low-endotoxin fetal calf serum (SH30070.03, HyClone, Logan, Utah, United States), 100 U penicillin/ml, 100 μg/ml streptomycin, 10 mM HEPES (N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid), and 2 mM L-glutamine (Biosource, Camarillo, California, United States). Bone marrow was harvested and MΦ were cultured as described previously ( Heise et al. 1998 ). To culture DCs, bone marrow cells were resuspended in RPMI1640 containing 10% low endotoxin fetal calf serum, 2 mM L-glutamine, 1 mM sodium pyruvate (Biosource), 100 U penicillin/ml, 100 μg/ml streptomycin, 1% nonessential amino acids (Biosource), and 20 ng/ml recombinant mouse GM-CSF (BD Biosciences, San Jose, California, United States) and plated at a concentration of 3 × 10 5 cells/ml in six-well plates in a total volume of 3 ml per well. The percentage of CD11c-positive DCs was determined by FACS staining after culturing cells for 7 d at 37 °C and 5% CO 2 . Approximately 70% of the cells were CD11c positive. Wt 129 and STAT1 −/− mice were purchased from Taconic (Germantown, New York, United States). IFNαβR −/− , IFNγR −/− , and IFNαβγR −/− ( Muller et al. 1994 ), PKR −/− ( Yang et al. 1995 ), and iNOS −/− ( MacMicking et al. 1995 ) mice were bred and housed at Washington University in accordance with all federal and university policies. Preparation of rabbit anti-MNV-1 serum Rabbits were immunized subcutaneously with 140 μg of MNV-1 VLPs in complete Freunds adjuvant and boosted 4 or 8 wk later with 70 μg of MNV-1 VLPs or 50 μg of UV-inactivated CsCl-purified MNV-1 in incomplete Freunds adjuvant. Serum was collected two weeks after the last boost, heat inactivated, and filtered before use. Immunohistochemistry Seven-week-old STAT1 −/− mice were infected orally with 25μl of brain homogenate containing MNV-1 (6 × 10 5 pfu) or brain homogenate from uninfected mice. Organs were collected into 10% buffered formalin and embedded in paraffin for sectioning by standard methods. Immunohistochemistry was performed as described previously ( Weck et al. 1997 ) using tyramide signal amplification (NEN Life Science Products, Boston, Massachusetts, United States). Slides were blocked in tyramide signal amplification blocking reagent (NEN Life Science Products) containing 10% mouse serum (IHC blocking buffer) for 30 min before adding antibodies. Serum was diluted 1:20,000 (spleen) or 1:100,000 (liver) in IHC blocking buffer, and tissue sections were incubated overnight at 4 °C. Horseradish peroxidase–conjugated donkey anti-rabbit secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, Pennsylvania, United States) was diluted 1:250 in IHC blocking buffer and applied to tissue sections for 1 h at room temperature. Biotin-tyramide was added at a dilution of 1:50 in 1× amplification diluent (NEN Life Science Products) for 10 min, slides were washed, and horseradish peroxidase–conjugated streptavidin (NEN Life Science Products) was added at a 1:100 dilution in tyramide signal amplification blocking reagent and incubated for 30 min at room temperature before washing. Antigen was visualized by a 3-min staining with a solution of 3, 3′-diaminobenzidine (Vector Laboratories, Burlingame, California, United States). Slides were washed and lightly counterstained with hematoxylin, dehydrated, and covered with Cytoseal XYL (Richard Allan Scientific, Kalamazoo, Michigan, United States) coverslips. No staining was observed in infected tissues incubated with preimmune serum or mock-infected tissues incubated with immune serum. Infection of cells Adherent cells were plated in 12-well plates and allowed to attach for several hours. Infections were carried out at an MOI of 0.05 or 2.0 for 30 min on ice in a volume of 0.5 ml per well. DCs were infected in bulk in the same volume. Cells were then washed twice with 2 ml of ice-cold PBS per well. To allow viral entry, 1 ml of medium was added to each well, and cells were incubated at 37 °C and 5% CO 2 for different time periods. For growth curve samples, infected cells and media were subjected to two or three cycles of freezing and thawing before plaque titration. Generation of mAb A6.2 A MNV-1-seropositive 129 mouse was injected intraperitoneally with 100 μl of a brain homogenate containing MNV-1, and the spleen was harvested 3 d later. Hybridoma fusions were performed as described previously ( Virgin et al. 1991 ) with the following modifications. Hybridoma supernatants were screened for binding to recombinant MNV-1 capsids by ELISA as described ( Karst et al. 2003 ). Stable hybridomas were characterized by Western blotting and ELISA after two rounds of subcloning by limiting dilution. A6.2 was unable to detect MNV-1 capsid protein by Western blot analysis but specifically bound to recombinant MNV-1 capsids by ELISA. The A6.2 isotype is IgG2a and was determined using the mouse mAb isotyping kit (Amersham Biosciences, Amersham, United Kingdom) and following manufacturer's protocol. MNV-1 plaque assay and plaque neutralization assay RAW 264.7 cells were seeded into six-well plates at a density of 2 × 10 6 viable cells per well. On the following day, 10-fold dilutions of virus inoculum were prepared in complete DMEM medium and plated in duplicate wells. Plates were incubated for 1 h at room temperature on a rocking apparatus before aspirating the inoculum and overlaying the cells with 2 ml of 37–40 °C 1.5% SeaPlaque agarose in MEM supplemented with 10% low-endotoxin fet al.calf serum, 1% HEPES, 1% penicillin/streptomycin, and 2% glutamine (complete MEM) per well. Plates were incubated at 37 °C and 5% CO 2 for 2 d. To visualize plaques, cells were stained with 2 ml of 56 °C 1.5% SeaKem agarose in complete MEM containing 1% neutral red per well for 6–8 h. For plaque neutralization assays, differing concentrations of purified mAb (A6.2, anti-MNV-1 capsid; isotype control, 10H2, anti-reovirus μ1c) were incubated with equal plaque-forming units of either MNV-1.CW1 or MNV-1 brain homogenate for 30 min at 37 °C prior to performing the MNV-1 plaque assay. Purification of virus particles RAW 264.7 cells were infected with MNV-1.CW1 for 2 d at an MOI of 0.05. Cellular debris was removed from freeze/thaw lysates by low-speed centrifugation for 20 min at 3,000 rpm. Supernatants were layered on top of a 5-ml 30% sucrose cushion and centrifuged at 4 °C for 2.5 h at 27,000 rpm (90,000 g ) in a SW32 rotor. Cell pellets were then resuspended in PBS and mixed with CsCl to a final density of 1.335 g/cm 3 and centrifuged for at least 18 h at 35,000 rpm (115,000 g ) in a SW55 rotor. A wide lower band (1.35 ± 0.01g/cm 3 ) and narrow upper band (1.31 ± 0.01g/cm 3 ) were typically seen in the gradient. Each band was collected by puncturing the side of the tube with a needle before overnight dialysis against PBS at 4 °C. Protein analysis CsCl-purified virions were separated by SDS-PAGE gel electrophoresis using standard protocols ( Sambrook et al. 1989 ). Proteins were visualized by Coomassie blue staining using the Simply Blue safe stain (Invitrogen, Carlsbad, California, United States) according to manufacturer's instructions. For Western blot analysis, proteins were transferred to nitrocellulose membrane and incubated with an anti-MNV-1-capsid rabbit polyclonal antibody, followed by a peroxidase-labeled secondary antibody, and visualized by ECL (Amersham Biosciences) according to manufacturer's instructions. Immunoprecipitation of radiolabeled infected cell lysates was performed as described previously ( Sosnovtsev et al. 2002 ) with serum obtained from a 129 wt mouse infected orally with MNV-1. Northern blotting The region of the MNV-1 genome from nt 5,617 to 7,039 was amplified by RT-PCR and cloned into the pGEM-T Easy (Promega, Madison, Wisconsin, United States) vector between the T7 and SP6 promoters. The resulting plasmid was linearized with Bsu361 and in vitro transcribed with SP6 RNA polymerase (Roche, Indianapolis, Indiana, United States) to generate RNA transcript probes for detection of positive-sense viral RNA, or with T7 polymerase (Roche) to generate transcripts for detection of negative-sense viral RNA. To label probes, the transcription reaction was carried out in the presence of [P 32 ]-UTP according to manufacturer's recommendations. Total RNA from virus-infected or mock-infected cells was isolated using Trizol (Invitrogen) according to the manufacturer's recommendations. One microgram of total RNA from MNV-1- or mock-infected cells was subjected to electrophoresis on a 1% formaldehyde gel. RNA Millennium Size Markers (Ambion, Austin, Texas, United States) were used as size markers. Northern blotting was performed using standard protocols ( Sambrook et al. 1989 ). Probes were hybridized overnight at 68 °C in 50% formamide containing 6× SSC, 5× Denhardt's, 0.5% SDS, and 100 μg/ml ssDNA. MNV-1 ELISA The ELISA was performed as described previously ( Karst et al. 2003 ) with the following modifications. ELISA plates were coated overnight at 4 °C with CsCl-purified MNV-1 particles at 0.2 or 1.0 μg/well. Diluted purified anti-MNV-1-capsid (A6.2) and isotype control (reovirus 10H2) mAbs, as well as the peroxidase-labeled secondary antibodies, were incubated for 60 min at 37 °C. Electron microscopy Negative staining electron microscopy of CsCl-purified virions was performed as described previously ( Karst et al. 2003 ). For thin-section electron microscopy, RAW cells were infected with MNV-1.CW1 at an MOI of 2.0, as described above. At various times postinfection cells were washed with PBS and fixed with 3% glutaraldehyde diluted in PBS at room temperature for 2 h. Cells were pelleted and washed with buffer prior to incubation with 1% osmium tetroxide (in 0.1 M cacodylate buffer) for 40 min at room temperature. After washing, the cells were incubated overnight at 4 °C in 2% uranyl acetate/80% acetone. The pellets were then dehydrated with an acetone series and embedded in Epon before polymerization at 65 °C for 72 h. Ultrathin sections (60 nm) were cut with a Micro Star (Huntsville, Texas, United States) diamond knife, and the sections were stained and contrasted with uranyl acetate and lead citrate before viewing on a JOEL 1010 electron microscope at 80 kV. Images were captured on a MegaView III side-mounted CCD camera (Soft Imaging System, Lakewood, Colorado, United States), and figures were processed using Adobe Photoshop software (Adobe Systems, San Jose, California, United States). Consensus sequence analysis of viral RNA RNA was extracted from brain tissue or cell culture material with Trizol (Invitrogen) and reverse transcribed with Superscript II enzyme (Invitrogen). Genome-specific sequences were PCR-amplified with Elongase enzyme (Invitrogen) to produce seven overlapping fragments. The DNA fragments were gel-purifed and sequenced directly with reagents in the BigDye Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, California, United States) on a 3100 DNA sequencer (Applied Biosystems). Data were analyzed with the Sequencher software package (Gene Codes Corporation, Ann Arbor, Michigan, United States). Oligonucleotide primer sequences are available upon request.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC532393.xml
509310
Mutants on the Small Screen
A recent television series in the UK celebrates the genetic diversity of human life
My graduate adviser, Sydney Brenner, used to exclaim ‘Revenons à nos mutants!’ as he sat down at the bench to search for yet more genetic variants of C. elegans . Those mutants won him a Nobel Prize, some thirty years later. Armand Leroi is another aficionado of C. elegans mutants, but he decided to write a book—and then to make a television series—on mutants of humanity, not of worms. He says at the beginning and end of the series: “We are all mutants, but some of us are more mutant than others.” This is a good slogan, and very proper for embracing humanity as a whole. He backs it up with an aphorism from Etienne Geoffroy Saint-Hilaire, pioneer of teratology, who proclaimed: “There are no monsters and Nature is one.” What Geoffroy meant was that abnormalities provide clues to normal processes, and hence are invaluable to science, if they can be properly understood. But monsters and mutants are, undeniably, fascinating in their own right. Mutants , the book, is excellent: impressively researched and illustrated and extremely well written. The resulting television series on Channel 4 in the United Kingdom has a distinctly different impact. The series covers only a few of the subjects dealt with in the book, and handles the material in a different way. Inevitably, television can't include much science or scholarly detail, but it compensates with human images that are wholly gripping—both the preserved specimens and the living subjects who talk about the strange conditions that they live with. It's a freak show, but a freak show with thoughtful scientific commentary. Each of the three programmes in the series has a particular theme. The first, ‘The Mystery of Growth’, is devoted mainly to skeletal disorders. We meet Carole Ozel, who copes with extraordinary courage with a terrible disease called fibrodysplasia ossificans progressiva (FOP), in which bony tissue forms throughout the body, gradually immobilizing the body in a second skeleton. Later we encounter a crew of charming and articulate dwarfs taking time out from a disco at the Reno convention of the Little People of America. They are happy to be called dwarfs or little people, but midget is no longer an acceptable term. Being a dwarf, in fact, is sometimes a ticket to fame and fortune, as in the case of Joseph Boruwlawski, last of the court dwarfs, who enchanted European royalty, married a beautiful woman, and lived happily to the age of 98. The second part, ‘The Dangerous Womb’, is about birth defects, conjoined twins, and basic embryology. The makers of the piece went to the trouble of getting the developmental biologist Eddie de Robertis to reconstruct the classic experiment of Mangold and Spemann that revealed the underlying basis of some of these defects. The programme goes into some detail about what is now known about the molecular basis of normal and abnormal development, and how we can begin to explain such extraordinary forms as that of the Parodi twins, who had distinct heads and shoulders, but merged into a single torso and a single pair of legs. The third programme, ‘The Meaning of Beauty’, deals with lesser but still striking abnormalities such as albinism and hypertrichosis (excess hair). It also moves into contentious areas, in a frank discussion of race genetics. Leroi makes the essential point loudly and clearly: there is much more genetic variation within any village on earth than there is between different human populations. We are extraordinarily unified, from a genetic standpoint. But, as he notes, the idea of race persists, and about 7% of global genetic variation in human DNA includes AIMs, ancestry informative markers. Some of these distinguish, for example, Africans and Europeans and provide objective information about the ancestry of different human populations. This being television, the series closes with a discussion of beauty, which Leroi proposes to be simply the absence of visible mutant defects, using Saira Mohan, Newsweek 's idea of human facial perfection, as an exemplar. Yes, she looks nice, but nice can be boring. Beauty, above all other human attributes, is profoundly influenced by culture, and it is hard to take this interpretation of beauty as an adequate explanation, rather than just a pretty way to finish the series. For the most part, Leroi makes an agreeable and humane commentator, though he is not immune to the slight self-satisfaction that seems to overcome all scientists on television. The camera also spends an excessive amount of time dwelling on him, to a point where it becomes irritating to see him walking—frequently in slow motion— into yet another museum or laboratory. Sometimes the focus on the presenter pays off, as when we see the six-foot scientist looking like a small child beside Chris Greener, the tallest man in Britain, or witness Leroi's faint chagrin at discovering that his DNA is mostly European, despite his cosmopolitan family history. More questionable are the bits where he paddles casually through a tub of preserved viscera from some long-gone sufferer from situs inversus (mirror-reversed organs) and succumbs to laughter at the sight of Ditto, the amazing two-faced pig. This may be an honest attempt at portraying the conflicted reactions we all have to abnormality, but it seems bound to cause trouble. Arty camerawork is a running feature of the programmes, and there is a great deal of smoke and mirrors about the whole production. Perhaps this is deliberate, reminding us that it is hard to look directly at extreme deformity, but there is an air of ‘I wants to make your flesh creep’ about the many sidelong shots of mutant babies in bottles in the Vrolik Museum, to which we return again and again throughout the series. Viewing the already distorted fetuses through further distorting camera angles and under green lighting doesn't really achieve anything. The treatment begins to resemble the first Alien movie, in which the audience was never allowed to see the monster directly. In the end, what is most memorable about these programmes is the living people themselves, and how they have coped with their various genetic disorders. It is very touching to see the home movies of Tiffany York, born with mermaid syndrome, or sirenomelia (in which the legs are fused together), taking her first tottery steps after corrective surgery, and to listen to her talking philosophically about her life as she floats in a Florida swimming pool. It is similarly cheering to hear from the dwarfs and albinos, or from Chuy Aceves, who has hypertrichosis and looks like the original Hollywood wolfman but suffers no ill effects and is proud of his rare condition. For these sections alone, let alone the serious and well-explained scientific background, the series is well worth seeing. It makes one feel surprisingly good about the human race and the human spirit.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509310.xml
544577
Extended-Spectrum β-lactamase (ESBL) producing Enterobacter aerogenes phenotypically misidentified as Klebsiella pneumoniae or K. terrigena
Background Enterobacter aerogenes and Klebsiella pneumoniae are common isolates in clinical microbiology and important as producers of extended spectrum β-lactamases (ESBL). The discrimination between both species, which is routinely based on biochemical characteristics, is generally accepted to be straightforward. Here we report that genotypically unrelated strains of E. aerogenes can be misidentified as K. pneumoniae by routine laboratories using standard biochemical identification and using identification automates. Results Ten clinical isolates, identified as K. pneumoniae or K. terrigena with the routinely used biochemical tests and with API-20E, were identified as E. aerogenes by tDNA-PCR – an identification that was confirmed by 16S rRNA gene sequencing for five of these isolates. Misidentification also occurred when using the automated identification systems Vitek 2 and Phoenix, and was due to delayed positivity for ornithine decarboxylase and motility. Subculture and prolonged incubation resulted in positive results for ornithine decarboxylase and for motility. It could be shown by RAPD-analysis that the E. aerogenes strains belonged to different genotypes. Conclusions Clinical E. aerogenes isolates can be easily misidentified as Klebsiella due to delayed positivity for ornithine decarboxylase and motility. The phenomenon may be widespread, since it was shown to occur among genotypically unrelated strains from different hospitals and different isolation dates. A useful clue for correct identification is the presence of an inducible β-lactamase, which is highly unusual for K. pneumoniae . In several instances, the use of genotypic techniques like tDNA-PCR may circumvent problems of phenotypic identification.
Background Enterobacteriaceae with β-lactam resistance due to the production of Extended-Spectrum β-Lactamases (ESBL) were discovered in the eighties and since that time became epidemic and endemic in hospitals worldwide [ 1 ]. Since two decades, 20 to 40 ESBL-producing strains are isolated monthly in our hospital. Amongst the clinical isolates from our hospital, two new TEM-β-lactamase genes were described [ 2 ]. In Belgium, as well as in other countries, a shift occurred from Klebsiella pneumoniae isolates as the predominant ESBL-producers [ 3 ] to predominance of Enterobacter aerogenes . It is also known that most of the E. aerogenes isolates in the Belgian hospitals belong to one of two predominant Belgian clones (BEI and BEII) [ 4 ], a situation which is comparable to that in other countries [ 5 ]. During the last years however, we found that several isolates that were identified as K. pneumoniae or K. terrigena by conventional biochemical testing were in fact E. aerogenes as could be shown by the use of genotypic methods, i.e. tDNA-PCR, validated by 16S rRNA gene sequencing, and by extensive phenotypic testing (including subculture and prolonged incubation). Results For ten clinical isolates, i.e. seven ESBL-producing clinical isolates collected during 2001 and three more recently collected isolates (Table 1 ), the API20E codes 5205773 (isolates GA1, GA2, GA3, MN2, MN3 and VGM), 5205753 (isolates DHJ1 and DHJ3) or 5204673 (isolates RA and DBH) were obtained. The first code resulted in a weak identification as either E. aerogenes , K. pneumoniae or Raoultella ( Klebsiella ) planticola , the second code did not yield any identification, and the third code resulted in a weak but acceptable identification as K. terrigena . The isolates with the codes 5205773 and 5205753 were identified as K. pneumoniae by additional biochemical testing due to negative reactions for motility (tested in semi-solid agar) and ornithine decarboxylase. However, these isolates all possessed an inducible cefalosporinase, as detected on the antibiogram using a disk approximation test, a finding which strongly contradicts an identification as K. pneumoniae or Klebsiella sp. In fact, the first hint that these strains, phenotypically identified as K. pneumoniae , were actually E. aerogenes , came from tDNA-PCR based identification. Using this method, all isolates were identified as E. aerogenes , and this by comparison of the obtained fingerprint – composed of amplified intergenic tRNA spacers of 101, 106, 111, 115, 121, 189, 190, 198 or 289, and 391 bp in length – with a library containing fingerprints of more than 3000 strains belonging to hundreds of species, available at . Confirmation of this genotypic identification was obtained by 16S rRNA gene sequencing for five isolates (Table 1 ). Analysis yielded a similarity of between 99.8% and 100% to E. aerogenes Genbank entries. The presence of a genuine K. pneumoniae isolate in patient DHJ (Table 1 ) further complicated the identification. This observation lead us to carry out additional phenotypic testing. Using the hanging drop method for testing motility, a few motile cells were observed, and upon retesting in semi-solid medium, weak migration could be observed. Like most biochemical tests in the routine laboratory, ornithine decarboxylase is read after overnight or 24 hours of incubation, but when the incubation period was prolonged to up to 2–5 days, all isolates tested positive. Because at present automated systems are frequently used for routine identification, a selection of six strains, containing four isolates of the study and two controls, i.e. phenotypically correctly identified E. aerogenes (LBV268) and K. pneumoniae (BG) isolates, were tested in two different systems, i.e. Vitek 2 (bioMérieux, Marcy l'Etoile, France) and Phoenix (BD Biosciences, Sparks, Md.). Both automated systems yielded the same results as the API20E, i.e. that the E. aerogenes isolates, aberrant due to a slow reaction for motility and ornithine decarboxylase, were misidentified as K. pneumoniae (Table 1 ). This misidentification by the automated systems is not unexpected, since they are based on biochemical testing only and a reading time of 24 hours or less. The control strains were correctly identified. Disk diffusion antibiotic susceptibility testing, carried out according the NCCLS guidelines, revealed basically the same resistotype for all isolates, characterized by resistance to ceftazidime and susceptibility to ceftriaxone. Additional resistance to aztreonam was observed for some isolates, reflecting the most dominant resistance patterns for the E. aerogenes isolates in our hospital. All isolates were also found to carry high-level resistance to cefoxitin, which is highly unusual for Klebsiella spp. Furthermore, the disk-approximation test with an amoxycillin-clavulanic acid disk close to β-lactam disks on Mueller-Hinton II agar, showed a combined pattern of synergy (broadening of the inhibition zone in the direction of clavulanic acid) and antagonism (flattening of the inhibition zone), which is suggestive for a combination of an ESBL and an inducible β-lactamase. Again, inducible β-lactamases are very rare in Klebsiella spp. but typical for Enterobacter spp. It should be noticed that this phenomenon will not be detected by automated MIC-determination systems like Vitek 2 and Phoenix. Using PCR and sequencing as described previously [ 2 ], the presence of TEM-5 could be shown in the isolates of patients GA and MN, and SHV-4 in the isolate of patient DHJ. The genotypic relationship of the phenotypically aberrant isolates was investigated with AP-PCR. Isolates GA1, GA2, GA3, MN2, MN3 and DBH were corresponding to Belgian clone BEI (Figure 1 , pattern A), isolates VGM, DHJ1 and DHJ3 were closely related to Belgian clone BEII (Figure 1 , pattern B, differing from pattern C, characteristic of clone BEII, by a single extra band) while isolate RA was not related to any of the others (Figure 1 , pattern D). This genetic diversity among the phenotypically aberrant strains makes it probable that strains with this kind of aberrant phenotype are not restricted to a single clone within E. aerogenes . Discussion K. pneumoniae and E. aerogenes are taxonomically closely related species [ 6 ] which share many characteristics. Our sequencing results (unpublished) and those of others [ 6 ] confirm that the genus Enterobacter is polyphyletic and that E. aerogenes should be placed within the genus Klebsiella . However, differentiation between E. aerogenes and K. pneumoniae is usually straightforward when based on testing for ornithine decarboxylase and motility, both positive for E. aerogenes . This is also reflected in the name " Klebsiella mobilis " [ 7 ], which is known as a valid synonym for E. aerogenes . Apparently, in some E. aerogenes isolates, the expression of these characteristics can be weak and/or delayed, and these are therefore scored negative when reading is done after the incubation periods that are routinely applied (overnight – 24 hours). However, the combination of an inducible β-lactamase and/or high-level cefoxitin resistance, which are rare in Klebsiella spp., and an identification as Klebsiella sp. should warrant further investigation. The phenomenon of E. aerogenes misidentified as K. pneumoniae or K. terrigena due to delayed or negative ornithine decarboxylase and motility was reported previously, and was also discovered because of unexpected imipenem resistance of the so-called K. pneumoniae isolates [ 8 ]. Also in this case, subculture and prolonged incubation restored the positivity for these characteristics. The problem of misidentification of E. aerogenes as K. pneumoniae (or even K. terrigena ) is probably not uncommon and probably also geographically widespread. This can be deduced from the following considerations: i) this phenomenon of misidentification of E. aerogenes was already reported in 1993 [ 8 ], ii) the phenomenon occurred in genotypically different organisms, iii) the isolates were found over an extended period of time – also recently, and finally iv) we received similar strains from other Belgian hospitals (unpublished data). It should be noted that misidentification also occurred when using the newer and automated systems like Vitek2 and Phoenix. On the other hand, the problem seems to be largely unknown. In a recent study, Hansen and colleagues [ 9 ] carried out an interlaboratory comparison of the efficacy of 18 biochemical tests for the identification of 242 strains of different Klebsiella species and of Enterobacter aerogenes , but do not mention the problem of possible delayed activity, possibly also because the study started from validated strains of each species. Conclusions Identification in a routine clinical microbiology laboratory of the most commonly encountered Enterobacteriaceae is usually considered to be fast and straightforward, but apparently identification problems may occur due to diminished or delayed expression of some characteristics, even for well-established species like E. aerogenes and K. pneumoniae . Here we showed that E. aerogenes isolates exist for which ornithine decarboxylase and motility are negative or delayed positive, and that as such these isolates can be misidentified as K. pneumoniae . This phenomenon may be quite frequent and geographically widespread. Genotypic identification techniques like tDNA-PCR, which moreover are cheaper than phenotypic testing for many bacterial species, can be semi-automatized, are faster and mostly have a higher discriminatory power, which is also reflected in this study. Methods tDNA-PCR tDNA-PCR was carried out using the outwardly directed tRNA-gene consensus primers T5A (5'AGTCCGGTGCTCTAACCAACTGAG) and T3B (5'AGGTCGCGGGTTCGAATCC), thus amplifying the intergenic tRNA-spacers, as described previously [ 10 , 11 ]. Electropherograms were normalized using GeneScan Analysis software, version 2.1 (Applied Biosystems). Transformation of GeneScan tables (ABI310, McIntosh) to tables on IBM, separation into separate digital fingerprints, and comparison of the digital tDNA-PCR fingerprints with a library of tDNA-PCR-fingerprints obtained from a large collection of reference strains, was done using in house software described previously [ 10 ]. 16S rRNA gene sequencing For five of the phenotypically aberrant isolates, the complete 16S rRNA sequence was determined by amplification of the 16S rRNA-gene with the primers 5'-AGTTTGATCCTGGCTCAG and 5'-TACCTTGTTACGACTTCGTCCCA [ 12 ], and sequencing was performed as described previously [ 12 ]. Comparison of the obtained 16S rDNA-sequence with all known sequences in Genbank was carried out using the BLAST software (National Center for Biotechnology Information ). RAPD-analysis for strain typing The genotypic relationship of the isolates was investigated using arbitrarily primed PCR with RAPD Ready-to-Go beads (Amersham Pharmacia, Uppsala, Sweden) and the ERIC II primer 5'-AAGTAAGTGACTGGGGTGAGCG [ 13 ]. Analysis of the fingerprints was obtained by visual interpretation on ethidium bromide stained electrophoresis gels. Authors' contributions GC, GV and AM were responsible for sample collection and initial biochemical identification. PVD and GW carried out automated biochemical identification. GW in addition carried out extended biochemical characterization. TDB and MV carried out the molecular analysis. GC, GV, TDB and MV drafted the manuscript. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544577.xml
552315
An adaptive method for cDNA microarray normalization
Background Normalization is a critical step in analysis of gene expression profiles. For dual-labeled arrays, global normalization assumes that the majority of the genes on the array are non-differentially expressed between the two channels and that the number of over-expressed genes approximately equals the number of under-expressed genes. These assumptions can be inappropriate for custom arrays or arrays in which the reference RNA is very different from the experimental samples. Results We propose a mixture model based normalization method that adaptively identifies non-differentially expressed genes and thereby substantially improves normalization for dual-labeled arrays in settings where the assumptions of global normalization are problematic. The new method is evaluated using both simulated and real data. Conclusions The new normalization method is effective for general microarray platforms when samples with very different expression profile are co-hybridized and for custom arrays where the majority of genes are likely to be differentially expressed.
Background Microarray technology provides simultaneous measurements of expression levels for thousands of genes. Each step from sample preparation to data analysis, however, contains potential sources of bias and variability. Proper normalization adjusts for differences which interfere with the comparison of intensities of different labels at a given probe and with the comparison of intensities of corresponding probes on different arrays. Proper data normalization should allow for the comparison of expression levels across different arrays. Subsequent data analysis results are heavily dependent on effective normalization. Normalization issues differ for dual-labeled platforms compared to single labeled platforms such as the Affymetrix GeneChip arrays. In this paper we address normalization for dual-labeled arrays with either cDNA or oligonucleotide probes. The objective of normalization for dual-labeled arrays is to correct for differences in intensities for the two labels on the same array. These differences arise from factors such as differences in sample concentrations, differences in photomultiplier tube setting, and differences in the affinity of the two labels for DNA. Median or mean based global normalization methods use a single normalization factor applied to all genes on the array to adjust for labeling bias [ 1 , 2 ]. Such methods are widely used because of their simplicity. Intensity-based and location-based methods take into account intensity and spatial dependence on dye bias normalization factors [ 3 , 4 ]. Both global and intensity/location based normalization methods assume that most of the genes are not differentially expressed between the two samples hybridized on the array, and that for the differentially expressed genes, the direction of the difference is symmetric between the two samples. In many important cases, however, these assumptions are not appropriate because: 1) more than half of the genes are differentially expressed on the array; 2) the numbers of over- and under-expressed genes on the array are unequal; 3) only genes of specific biological interest are selected to make a customized array, which are highly variable across the samples. In the above cases, the global normalization methods and intensity/location based normalization methods become less accurate and a more sophisticated method is needed [ 5 , 6 ]. There are some methods which attempt to adaptively identify the subset of 'housekeeping' genes [ 6 - 8 ]. These methods require multiple arrays in order to identify the 'housekeeping' gene set, which does not always exist. Newton et al. proposed a Gamma-Gamma-Bernoulli model for identifying differentially expressed genes in dual labeled arrays [ 9 ]. We have generalized Newton's model and here propose an adaptive method based on three-component mixture model for normalization of dual labeled microarray data. Results As described in the Methods section, we have applied our adaptive method to both the simulated data and real data. We have also compared our method with the global method and the intensity-based lowess method. Results of the simulation studies are shown as bar plots in Figure 1 . Figure 1A shows the comparison of our adaptive method, the global method and the lowess method when no noise was added. When the majority of genes in the array were non-differentially expressed (Case 1), or the numbers of over- and under-expressed genes on the array were equal (Case 2), the root mean squared error (RMSE) of the adaptive method was comparable with the other two methods; all were very small. When the array contained unequal numbers of over- and under-expressed genes and when the majority of genes were differentially expressed (Cases 3–6), the RMSEs of the global normalization method and the lowess method were much larger than those of the adaptive method. The differences ranged from around a two fold difference (0.895 in log 2 scale) when the number of under-, null, and over-expressed genes were 200, 100, and 100, to more than a three fold difference (1.617 in log 2 scale) when the number of under-, null, and over-expressed genes were 200, 50, and 50. The RMSEs for the adaptive method ranged from 0.078 to 0.159 in log 2 scale. We compared the histogram of observed intensities to the fitted marginal density from the adaptive method as a simple check to see whether the proposed model and the estimation procedure are in line with available data. Figure 2 shows the histograms of log(ratio) and log intensities of red and green channels of the simulated data, and the curve in each plot is the estimated density obtained from the fitted model. It is seen the data fits to the model quite well. Gaussian noise with SD of 0.25 and 0.50 were added so that the data was not generated from the same model used for analysis with the adaptive method. The RMSEs of the global normalization method and the lowess method remained large, while the RMSEs of the adaptive method remained small, ranging from 0.083 to 0.569 on the log 2 scale (Figure 1B and 1C ). In the above simulation, no apparent groups could be seen in the histograms of log(ratio) (Figure 2A ). Better results for the adaptive method were also obtained for a simulation case where the three groups (under-expressed, non-differentially expressed, and over-expressed) are apparent in the histogram of log(ratio). The results can be seen in Figure 4 and Figure 5 [see Additional files 3 , 4 ]. Results comparing RMSEs for the adaptive method, the global method and the lowess method with real data are shown in Table 2 . The RMSEs of the adaptive method on data generated from ten different arrays ranged from 0.128 to 0.529, in comparison with RMSEs of around 1.0 using the global normalization method. The average RMSE (0.607) of the lowess method is almost two times that of our adaptive method (0.328), although the lowess method performed better than the global method (average RMSE = 1.016). Figure 3 shows the histograms of log(ratio) and log intensities of red and green channels of the real data, and the curve in each plot is the estimated density from the adaptive method. Discussion In this paper, we propose a new method for normalization of dual-labeled arrays in cases where the number of differentially expressed genes is substantial and not necessarily symmetric in direction. The method performed effectively with both simulated and real data. We started our model building initially by introducing an unknown constant c into Newton's Gamma-Gamma-Bernoulli model [ 9 ]. The mixture model consisted of two groups: non-differentially expressed genes (Equation 1A) and differentially expressed genes (Equation 1B). log( cR k ) ~ Gamma ( a , s k ) log( G k ) ~ Gamma ( a , s k )     (1A) s k ~ Gamma ( a 0 , γ ) log( cR k ) ~ Gamma ( a , ) log( G k ) ~ Gamma ( a , )     (1B) ~ Gamma ( a 0 , γ ) ~ Gamma ( a 0 , γ ) We found that when the differential expression was symmetric between the two samples, the model worked well. However, the error increased significantly when the ratio of the numbers of under- to over-expressed genes shifted from 1. In order to make the model more flexible, we modified the model by assigning different scale factors γ R and γ G for the red channel and green channel intensities. For this modified two-component mixture, the error still remained large. We then extended the model into a three-component mixture model listed as Equations 8A-8C in the additional material [see Additional file 2 ]. The model was then quite flexible but there were too many parameters that needed to be optimized. After we tested it with simulation data and real data, we found the estimated model was not stable and difficult to optimize. We finally simplified the model to our final model given by Equations 2A-2C (see Methods section). When applying it to real data or simulated data, the estimates converged well close to globe optima. When different start points were used, the optimizations remained relatively robust. Evaluation of normalization methods can be difficult since the true normalization factors are unknown with real data for custom arrays. We avoided this problem by synthesizing customized arrays based on real data for standard arrays containing thousands of genes. In order to make the distribution of each component group look smoother, we allowed certain range of overlap between the adjacent groups. Additional sampling method was tried to divide the whole distribution range into many non-overlapping intervals. In each interval the number of genes sampled increased when the absolute value of log 2 (ratio) became larger (Table 3 [see Additional file 6 ]). The model fitting results using data generated by this sampling method are listed in Table 4 [see Additional file 7 ] and Figure 6 [see Additional file 5 ]. We compared our adaptive method with the global method and the intensity-based lowess method. The lowess method assumes that in each intensity interval either the majority of genes are non-differentially expressed or the numbers of up- and down-regulated genes are equal. The global median normalization makes these assumptions only over the array as a whole. It is not surprising that our method performed much better that the above two methods, because the global median method only works well when the assumptions are valid while the intensity-based lowess method is only effective when there are intensity-dependent biases. Correlation structure is complicated for the thousands of genes on a microarray. In our model, the intensity of each channel is conditionally independent given the scale parameter, but not marginally independent.Therefore, we are not assuming the intensities in two channels are independent. Although we did not generate correlated genes in our simulated data sets, correlations of genes do exist in the real data sets we tested. Spatial correlations are also possible but our method is not designed for that purpose. Yang et al. proposed using the lowess normalization separately within each grid on the array [ 3 ]. Our algorithm could be similarly applied within each grid to control for spatial effects. Limited simulations were performed in this study. We also tried to use real data to test our method. Since an appropriate data set with known normalization factor was not available, we synthesized such data sets by sub-setting large arrays in which the true normalization factor could be accurately estimated. In the process of synthesizing such small arrays we had to choose an empirical threshold to stratify the differentially expressed genes and non-differentially expressed genes. Although we do not believe that the superiority shown for our algorithm depends critically on the threshold chosen nor on details of the synthesis, it would be preferable to evaluate the algorithm on real data sets with know normalization factors. Although our method is designed for dual-labeled cDNA array, it can be extended to single channel Affymetrix chip data. The most popular normalization method for the Affymetrix chip compares each array to a single base line array for probe set summaries. The assumptions behind the normalization method are that the majority of the genes are non-differentially expressed and the numbers of over- and under-expressed genes are roughly equal; the same assumptions as those for dual-labeled cDNA arrays. We could treat the base line array as the 'reference channel' and the other array as the 'test' channel and apply our algorithm to probe set summaries. For Affymetrix chip data, there are multiple base pairs in a probe set and each probe has an intensity measurement. Several alternative normalization methods of Affymetrix arrays utilize the probe level information. For example, method based on an 'invariant set' proposed by Li and Wong assumes that a probe of a non-differentially expressed genes in two arrays to have similar ranks and uses an iterative procedure to identify the invariant set which presumably consists of points from non-differentially expressed genes [ 10 ]. Conclusions Our new normalization method does not require that the majority of genes be non-differentially expressed, and doesn't require multiple array replicates, dye swaps, spiked controls, or housekeeping genes. It appears much more effective than standard methods when the numbers of over- and under-expressed genes are unequal, and the majority of the genes are differentially expressed. It can be very useful for general microarray platforms when samples with very different expression profile are co-hybridized and for custom arrays where the majority of genes are likely to be differentially expressed. In both of these settings, standard normalization methods are problematic. Methods Model We define true intensities for a specific gene k in two channels as R k (red) and G k (green). Let c be a positive constant which is related to the normalization constant. The observed intensities for gene k in two channels are cR k and G k . We assume the logarithm of intensity in each channel has a Gamma distribution. The genes on the array belong to three different groups: 1) non-differentially expressed; 2) under-expressed; and 3) over-expressed. The overall data will be fitted into a mixture model listed below. For a non-differentially expressed gene k , log( cR k ) ~ Gamma ( a , s k ) log( G k ) ~ Gamma ( a , s k )     (2A) s k ~ Gamma ( a 0 , γ ). For an under-expressed gene k , log( cR k ) ~ Gamma ( a , ) log( G k ) ~ Gamma ( a , )     (2B) ~ Gamma ( a 0 , γ 1 ) ~ Gamma ( a 0 , γ 2 ). For an over-expressed gene k , log( cR k ) ~ Gamma ( a , ) log( G k ) ~ Gamma ( a , )     (2C) ~ Gamma ( a 0 , γ 2 ) ~ Gamma ( a 0 , γ 1 ). In the above Gamma distributions, the parameters a and a 0 are shape factors, and the parameters s k , γ , γ 1 , γ 2 , , are scale factors. The parameters a , a 0 , γ , γ 1 , γ 2 will be estimated from the data. Let p u ( R k , G k ), p o ( R k , G k ) and p n ( R k , G k ) be the densities of ( R k , G k ) for under-expressed, over-expressed and non-differentially expressed genes, respectively. The joint distributions of ( R k , G k ) in three groups can be derived as follows [details see Additional file 1 ]: Let θ denote the unknown parameter vector ( a , a 0 , γ , γ 1 , γ 2 , c ), which can be estimated by maximizing the likelihood function of observed data. We used the EM algorithm [ 11 ] for this maximization. Let p 1 be the proportion of under-expressed genes and p 2 be the proportion of over-expressed genes. We define indicator binary variable z k 1 to be 1 if the k th gene is under expressed, 0 otherwise; and z k 2 to be 1 if the k th gene is over expressed, 0 otherwise. The complete-data loglikelihood for all spots can be derived as follows, In the M-step, we first take derivative on Equation (4) with respect to p 1 and p 2 . This yields where K is the total number of genes on the array. To maximize Equation (4), we only need to maximize Equation (6) because the left out terms do not depend on the parameter θ . In the E-step, we compute the conditional expectations of z k 1 and z k 2 given the other parameters from the M-step. Once the constant c was obtained, the normalization constant for log intensity ratio data can be calculated as log(1/ c ). Model evaluation Simulation studies were performed by generating two channel intensities from the mixture model with c = 1.5, a = 118, a 0 = 410, γ = 31, γ 1 = 23, and γ 2 = 29. Six scenarios were included using different proportions of non-differentially expressed genes and different ratios of under- to over- expressed genes, as listed in Table 1 . One hundred data sets were generated for each scenario and the RMSE between the estimated log 2 (1/ ) and the true log2(1/ c ) was calculated. The global method takes the median log 2 (ratio) of all genes in each data set as the normalization factor. The lowess method performs robust locally linear fits of M-A plot and corrects the biases that are dependent on spot intensity [ 3 ]. The RMSE between the normalized log 2 (ratio) using the lowess method and the normalized log 2 (ratio) using the global method with the true normalization factor (log 2 (1/ c )) for all genes was also calculated for the same 100 data sets. Gaussian white noise was also added when generating the simulated data. We used standard deviation of 0.25 in log 2 scale to reflect the experimental noise in inbred strains of mice or cell line data and 0.5 in log 2 scale to reflect a larger experimental noise in human tissue data [ 12 ]. In-silico studies were performed on real data. We tested the method on ten arrays from publicly available breast cancer data [ 13 ]. Each array consists of 9216 genes. The common reference sample was a pool of RNA isolated from 11 different cultured cell lines (green channel, labeled with Cy3). RNA from tissues of breast cancer patients were used in the test channel (red channel, labeled with Cy5). The array was first normalized by the global normalization method. The median log 2 (ratio) of all genes was considered as the true normalization factor c . The genes were then divided into three groups: over-expressed genes (log 2 (ratio)>1), non-differentially expressed genes(-1.5<log 2 (ratio)<1.5), and under-expressed genes (log 2 (ratio)<-1). We randomly sampled a specified number of genes from each group (100 non-differentially expressed genes, 200 under-expressed and 100 over-expressed genes) and then combined them into an in-silico array. We constructed 100 datasets for each of the 10 arrays in this way and the RMSE between the estimated log 2 (1/ ) and the true log 2 (1/ c ) was calculated. Supplementary Material Additional File 3 Figure 4: Bar plots show comparison of RMSE by using the adaptive method (black bar) and global method (grey bar) with simulated data generated from a mixture model with c = 1.5, a = 90, a 0 = 120, γ = 8, γ 1 = 6, and γ 2 = 10 at three different noise levels (A) SD = 0; (B) SD = 0.25; and (C) SD = 0.50. Click here for file Additional File 4 Figure 5: Histograms and the estimated densities of log(ratio) and log(intensity) for a simulated data of a mixture model with c = 1.5, a = 90, a 0 = 120, γ = 8, γ 1 = 6, and γ 2 = 10. The superimposed curve on each plot is generated from the fitted model. Click here for file Additional File 2 Equations 8A-8C: a three-component mixture model. Click here for file Additional File 5 Figure 6: Histograms and the estimated densities of log(ratio) and log(intensity) for a set of real data generated from array svcc109. The superimposed curve on each plot is generated from the fitted model. The procedure to generate the data was described in the paper and the sampling rate was shown in Table 3 [see Additional file 6 ]. Click here for file Additional File 6 Table 3: Different number of genes sampled in each interval. Click here for file Additional File 7 Table 4: Comparison of RMSE by using the adaptive method and global method with real data by a different sampling method. The procedure to generate the data was described in the paper and the sampling rate was shown in Table 3 [see Additional file 6 ]. Click here for file Additional File 1 Derivation of joint distributions of (R, G) for p n , p u , and p o Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552315.xml
546406
Motor rehabilitation using virtual reality
Virtual Reality (VR) provides a unique medium suited to the achievement of several requirements for effective rehabilitation intervention. Specifically, therapy can be provided within a functional, purposeful and motivating context. Many VR applications present opportunities for individuals to participate in experiences, which are engaging and rewarding. In addition to the value of the rehabilitation experience for the user, both therapists and users benefit from the ability to readily grade and document the therapeutic intervention using various systems. In VR, advanced technologies are used to produce simulated, interactive and multi-dimensional environments. Visual interfaces including desktop monitors and head-mounted displays (HMDs), haptic interfaces, and real-time motion tracking devices are used to create environments allowing users to interact with images and virtual objects in real-time through multiple sensory modalities. Opportunities for object manipulation and body movement through virtual space provide frameworks that, in varying degrees, are perceived as comparable to similar opportunities in the real world. This paper reviews current work on motor rehabilitation using virtual environments and virtual reality and where possible, compares outcomes with those achieved in real-world applications.
Introduction One of the major goals of rehabilitation is to make quantitative and qualitative improvements in daily activities in order to improve the quality of independent living. Three determinants of motor recovery are early intervention, task-oriented training, and repetition intensity [ 1 ] while a major objective of rehabilitation is to identify the means to provide repeated opportunities for tasks that involve multimodal processes (different sensory modalities including vision, haptics, proprioception, audition) and that further enable increases in function. Carr and Shepherd [ 2 ] focus on motor relearning where relearned movements are structured to be task specific. They suggest that the practice of specific motor skills leads to the ability to perform the task and that motor tasks should be practiced in the appropriate environments where sensory inputs modulate their performance. The functional relevance of the specific environmental context has been specifically addressed by Keshner and colleagues [ 3 - 5 ] as it relates to posture control. These authors have shown that specific postural responses differ between paradigms where isolated individual control pathways are manipulated (i.e., visual, vestibular, somatosensory pathway) as opposed to within a functionally relevant context where information from multiple pathways is available. The successful integration of virtual reality into multiple aspects of medicine, psychology, and rehabilitation has demonstrated the potential for the technology to present opportunities to engage in behaviors in challenging but safe, ecologically valid environments while maintaining experimental control over stimulus delivery and measurement [for review see [ 6 , 7 ]]. Moreover, in VR, the user (patient, therapist) interacts with a multidimensional, multisensory computer generated environment, a virtual environment, which can be explored in real time [ 8 ]. Virtual reality also offers the capacity to individualize treatment needs while providing increased standardization of assessment and training protocols. In fact, preliminary evidence [ 9 - 11 ] indicates that VR provides a unique medium where therapy can be provided within a functional, purposeful and motivating context and can be readily graded and documented. Several features distinguish virtual environments from other forms of visual imaging such as video and television. A key feature of all VR applications is interaction. Virtual environments (VE) are created that allow the user to interact with not only the VE but also with virtual objects within the environment. In some systems, the interaction may be achieved via a pointer operated by a mouse or joystick button. In other systems, a representation of the user's hand (or other body part) may be generated within the environment where movement of the virtual hand is "slaved" to the user's hand allowing a more natural interaction with objects. Finally, while many applications of VR allow the user to control the viewpoint on the screen, third-person views or images of the users themselves that appear as players in the environment also provide the opportunity for interaction with the VE. A broad range of visual interfaces are used to create varying degrees of immersion in a VE ranging from conventional desktop monitors to head mounted displays. Increasingly complex, fully immersive VR systems, such as the Cave Automatic Virtual Environment (CAVE) developed at the University of Illinois at Chicago, provide the illusion of immersion by projecting stereo images on the walls and floor of a room-sized cube. Several persons wearing lightweight stereo glasses can enter and walk freely inside the CAVE. A head tracking system continuously adjusts the stereo projection to the current position of the leading viewer. In order to integrate the movement of the user with that of the VE and virtual objects, user position and motion must be tracked so that virtual images can be updated in real-time. Motion tracking approaches include color subtraction technology, video frame subtraction as well as magnetic and infrared tracking devices. Technical advances in the development of these interfaces have minimized the once lengthy lag times responsible for some of the earlier reports of cybersickness. To date, rehabilitation applications have primarily used visual and auditory sensory input while the addition of haptics is less developed. Haptic interface devices including gloves, pens, joysticks and exoskeletons provide users with a sense of touch and allow the user to feel a variety of textures as well as changes in texture. There is increasing evidence that haptic information is an effective addition towards the accomplishment of certain treatment objectives such as increasing joint range of motion and force [ 12 ]. Haptic information has also been identified as a significant signal for improving a subject's performance in more difficult tasks. For example, Shing and colleagues [ 13 ] report a specific benefit of adding haptic information to an upper extremity movement when the difficulty of the task, in this case a 3D pick and place task, was high. Integration of visual and haptic interfaces with motion tracking allows the user to become immersed in three dimensional virtual environments, including three dimensional sound, and virtual objects that can be picked up, manipulated, and even felt with the fingers and hands [ 14 ]. Another cardinal feature of virtual reality is the provision of a sense of actual presence in, and control over, the simulated environment [ 15 ]. The sense of presence has been defined as the feeling of being in an environment even if one is not physically present and resulting in behavior that is congruent with the subject's situation in the environment [ 16 ]. Early studies relied on questionnaires to characterize presence within a virtual environment [ 15 ] with more recent work suggesting that physiological measures including heart rate and galvanic skin response provide important information about user immersion [ 17 ]. Movement elicited and generated in virtual reality applications One important consideration with the application of virtual reality and movement in virtual environments is the behavior or movement characteristics of subjects in virtual environments [ 8 ]. Recent work by Feldman and colleagues [ 18 ] specifically compared movements made with or to virtual objects in a VE to movements made with or to real objects in real environments. Virtual representations of the hand were obtained by combining a fiber optic glove with a prehension force feedback device. Orientation of the hand in the VE was achieved using an electromagnetic tracker while kinematic data of the arm and trunk were recorded as the participant reached separately to real and virtual targets. Minimal movement differences in spatial and temporal kinematics of reaching in healthy adults were identified and included the amount of terminal wrist and elbow extension as well as timing of maximal grip aperture. There were no differences in movement characteristics between the real and virtual task in participants with hemiparesis. The authors suggest that VR is similar enough to reality to provide an effective training environment for rehabilitation. In contrast, we have demonstrated significant differences between functional lateral reach performances when performed in the real environment versus in a virtual environment delivered on a flatscreen [ 19 ]. The VR technology, VIVID Group's IREX system, provided participants with a third-person view of the users themselves in the virtual environments where they acted on virtual objects. Both young and old adults reached significantly further when virtual objects were presented in the VE compared to when reaches were made to real objects presented in the periphery. Lateral stability is crucial for performance of many weight-bearing tasks including turning, transferring, and stepping onto a stool while controlling a reach made as far as possible to the side requires regulation of the position of the center of mass within the limits of stability. We proposed that embedding the reaching task within a VR application may have resulted in shifting attention away from the potential for loss of balance, whereas focusing attention on balance, such as in the real-environment, may have resulted in increased fear of destabilization and underestimation of true ability. Improving the functional abilities of patients is commonly achieved by using tasks of increasing difficulty in combination with physical and/or verbal guidance of the patient's movements or actions. Thus, integrating the means to modulate the level of difficulty within a VR task is of crucial importance. A virtual reality system (VIVID GX) was used to provide independent leisure opportunities to adults with cerebral palsy and severe intellectual disabilities who were non-speaking and who used wheelchairs for mobility [ 15 ]. The participants demonstrated an exceptional degree of enthusiasm during the VR experiences reacting with appropriate, goal-oriented responses. However, a small number of participants clearly displayed involuntary movement synergies, increased reflexes and maladaptive postures, which were attributed to the level of task difficulty. The ability to change the virtual environment relatively easily, to grade task difficulty and to adapt it according to the patient's capabilities are important advantages of VR, since these features are essential to cognitive and motor remediation [ 20 ]. Does the technology work? Transfer of training Central to the issue of virtual environments as a training medium is the issue of transfer of training; does task improvement or learning transfer reliably from a VE to a real environment? Virtual environments and VR interventions should not only be used to augment current ability or to provide exposure to "other" therapeutic possibilities, but importantly to demonstrate distinct carryover to real-life functional tasks. One major challenge is identifying effective and motivating intervention tools that enable transfer of the skills and abilities achieved during rehabilitation to function in the "real" world. For example, recent studies stress that simple repetitive movements of an affected limb are not productive for the reorganization process but that it is action related to skill acquisition which contribute to the desired effect [ 21 ]. Rose and colleagues studied the transfer of training of a simple sensorimotor virtual task to performance on the "real world" equivalent [ 22 ]. The real-world equivalent consisted of a curved wire suspended between two vertical supports. With the non-preferred hand, the subject held a rod with a wire loop at the end and guided the loop along the wire without touching it. Contact between loop and wire, defined as an error, produced feedback. Errors and time to complete task were recorded. The group provided with no practice did significantly worse that the two practice groups, one practicing with the virtual task and one practicing with the real task, although with no difference between the type of practice performed. In other words, within the constraints of this task, final real-world performance benefited as much from real as virtual practice. Thus, it is not sufficient simply to demonstrate that training does transfer in a given situation. It is crucial to identify whether a specific skill or a general familiarity with the training context is being transferred. If specific skills are transferred, it is important to determine whether the transferred training lasts as long and as reliably as an equivalent amount of real world training [ 22 ]. In addition, the conditions such as degree of immersiveness, overlap between real and virtual tasks, must be understood if we are to optimize or facilitate transfer. Balance and Posture Several systems have been used in studies of balance including a combined HMD display system combined with a fixed bicycle, a flatscreen VR system providing primarily 2D visual information and more recently an immersive dynamic virtual environment combined with a posture platform. Kim et al [ 23 ] reported preliminary data from healthy adults using a bicycle linked to a virtual visual environment and suggested that this training system would be beneficial for postural balance control. They described decreases in cycling path deviation and increases in cycling velocity following a short training period and suggested that these variables, in conjunction with additional parameters, may be relevant for determining a training effect on balance rehabilitation. Several problems remain to be resolved including the limited integration of bicycle motion and auditory cues. A specific concern is that a fixed bike was used which could provide the degree of safety necessary for an individual with a significant amount of balance impairment. However, a fixed bike sets up incongruence between the expectation of lean/tilt of the bike when covering a curved path and the sensory information indicating no tilt. Multiple applications of flatscreen VR for balance training have been reported that have used video capture technology from VividGroup's GX or IREX systems [see for example, [ 9 , 10 , 24 - 26 ]]. The systems take a video image of the user and use color subtraction software to remove a monochrome background and insert the user into a virtual environment. Proprietary software is used to allow the user to interact with virtual objects within the VE. Applications that have been used in various studies include: 1) a juggling task where the participant is required to reach laterally to juggle virtual balls; 2) a conveyer belt task where the participant is required to turn sideways, pick up a virtual box from a virtual conveyer belt, turn and deposit the box on a second virtual conveyer belt; and 3) a snowboard task where the user is required to lean sideways to avoid trees, rocks and other virtual objects while boarding down a hill. The applications are modifiable allowing the task difficulty to be modified by increasing the number of virtual objects to contact, increasing the speed at which the objects or environment moves, or increasing and decreasing the height of the objects requiring users to reach to the ground or to step up onto a stool. One of the earliest reports of use of the technology in rehabilitation compared therapy delivered through VR to a conventional approach in a sample of frail, older adults [ 25 ]. Greater improvements in dynamic standing tolerance were reported for a small (n = 3 to 4) group of older adults following a VR therapy than for a small group (n = 3 to 4 per group) following a standard occupational therapy program. We have used a similar approach with a significantly larger study population of community-living individuals with traumatic brain injury [see [ 9 , 10 , 26 ] for preliminary data]. A six week, three sessions per week intervention trial compared an activity-based exercise program (ABE) with a VR-based exercise program (VRE). Both exercise programs resulted in clinically significant changes on the Community Balance and Mobility Scale (CB&M) [ 27 ], used to measure functional mobility and balance, with average improvements of 6 and 10 points recorded for the ABE and VRE groups, respectively. Although not all participants involved in the exercise programs improved on their balance measures, 10 out of 14 individuals in the VRE group and 4 out of 10 individuals in the ABE group had clinically significant improvements. Most recently, we have demonstrated significant improvements in balance and functional mobility in community-living older adults following a VR exercise program. The comparison group completed a biofeedback exercise program and also demonstrated significant balance improvement [ 24 ]. Although these two studies did not demonstrate significantly greater improvements in balance outcome with the VR exercise program relative to the comparison intervention, other benefits of VR were identified. Specifically, the participants in the VR programs indicated greater enthusiasm about the exercise programs and reported greater enjoyment and improved confidence. The implications of these psychosocial benefits for long-term exercise compliance and participation have yet to be determined. More recently, Keshner and colleagues [ 4 ] have united an immersive dynamic virtual environment projected onto a wall with a linear accelerator (sled) that is translated in the anterior-posterior direction. Study participants stand on the sled in front of a screen on which a virtual image is projected. Various combinations of inputs (i.e., translating the support surface, moving the virtual scene, or combining different motions) are used to determine responses elicited when conflicts of different magnitudes between visual and vestibular/somatosensory signals are delivered. The results of initial experiments clearly demonstrate the non-linear effect in the postural response from single versus different combinations of inputs. These findings suggest that using this or similar complex, multimodal environments for rehabilitation intervention would promote ongoing recalculation of sensory inputs that would result in appropriate updates of posture within realistic environmental contexts. Locomotion Patients with Parkinson's disease akinesia have little difficulty stepping over objects in their path even when they are totally unable to initiate a step on open ground [ 28 ]. A virtual display superimposed over a user's visual field, augmented reality, has been shown to initiate and sustain walking in akinetic Parkinson's patients. Reiss and colleagues [ 28 ] reported that a stable cue appearing about six inches in front of the toes was required to initiate the first step, while cues scrolling toward the feet, as if stable on the ground as the person moves, were needed to sustain walking. The effectiveness of the visual cue was dependent on the degree and type of akinesia with, as a general rule, more realistic cues needed as the severity of akinesia increases. A locomotor interface, GaitMaster2 (GM2), intended to provide the user with the sense of forward movement while his/her actual position in space is constant, has been tested with two individuals with hemiplegia following a stroke [ 29 ]. The user stands on two footpads that move individually with each user's foot providing a sense of movement over a virtual terrain. The footpads in the GM2 follow the trajectory of a healthy individual when walking. The user thus experiences a corrected foot trajectory for each step. Modifications in gait patterns of two hemiplegic patients following gait training with the GM2 included moderate improvements in gait speed, improvements in leg muscle activity, increased symmetry during gait and improvement in QOL. A VR-enhanced orthopedic appliance for use with individuals with spinal cord injuries has also been developed and links a gait-inducing exoskeleton to a HMD providing binocular visual displays [ 30 ]. Briefly, the exoskeleton consists of a semi-rigid sling that supports the bust and lower limbs of the user. The sling is equipped with small actuators that move the lower extremities in accordance with human gait. Preliminary results from two experimental sessions with the same patient, a 26-year old with complete paraplegia, showed improvements in self-confidence, higher levels of optimism and motivation as well as increased relaxation and activity scores. A novel VR application for locomotor rehabilitation couples a three dimensional visual scene with a self-paced treadmill [ 31 ]. Briefly, both treadmill speed and scene progression are based on real-time feedback of subject position and progression with the speed of walking adjusted easily by the individual user. Preliminary trials of the system provided subjects with varying levels of interaction with the scene surface and surrounding objects with a strong sense of presence reported by users. Ongoing work by the group includes development and evaluation of a training protocol for locomotor rehabilitation in individuals with stroke. Upper and Lower Extremity Function Several upper and lower extremity VR applications have been developed using different technologies. Preliminary data suggest potential benefits of various systems. For example, a report based on two case studies using the Vivid GX video capture technology demonstrates improvements in upper extremity function [ 32 ]. The first individual had a T9 complete spinal cord injury requiring use of wheelchair for all mobility activities. His primary rehabilitation goal was to improve sitting balance in order to enable him to perform functional activities such as reaching out for a book placed on a shelf. Analysis of videotaped records of performance revealed that initially he used only one hand at a time to interact with the virtual objects while leaving the other on his lap or on the wheelchair arm rest in order to maintain balance. As sessions with the VR system progressed, he began to use both hands during the tasks relying on weak trunk muscles to maintain balance. The second individual had a right hemispheric stroke and ambulated with a cane due to poor control of foot and poor standing balance. He had functional movement in the upper extremity, suffered from mild attention deficit and required some help when dressing the lower extremity. The application he used consisted of balls appearing in the VE from all sides requiring that he pay attention to the entire visual space. After 3 minutes of interaction, he asked to get up and continue with therapy while in a standing position (although therapist behind was necessary for safety). Both participants reported enjoyment and wanted to repeat experience if possible. Importantly, they acknowledged the relevance of the experience to their rehabilitation process. Holden and colleagues [ 33 ] developed a VE training system based on the principle of learning by imitation. Pre-recorded movements of a virtual 'teacher' are displayed as either movements of the limb's endpoint or as an entire arm. Patient movements are recorded using an electromagnetic tracking device for the arm and hand segment or a CyberGlove for hand kinematics. The "teacher" shows the patient the trajectory of the end-point (hand) path for the movement to be reproduced. Frequency of visual feedback, speed of motion, degree of movement synchronization and other aspects of the teacher-patient relationship can be modulated. Data from eight chronic post-stroke patients demonstrated variable improvements on clinical measures of upper extremity function including strength. Piron et al. [ 34 ] used a virtual reality task to assess functional motor progress of a group of 20 post-stroke patients undergoing conventional rehabilitation. The patients were required to move an envelope instrumented with a magnetic receiver to a virtual mailbox slot. The participant was provided with a view of the trajectory of the corresponding virtual envelope as it moved. Patients improved on reach velocity and reach duration with the changes related to improvements on a clinical measure of upper extremity voluntary movement. The authors suggest that the reach trajectory characteristics also improved although limited data were presented. Several questions however remain. Primarily, would similar changes in movement trajectories be observed if the subject did not "see" a virtual mailbox? Moreover, in this paradigm, the trajectory to the mailbox is only one aspect of the functional task while an equally, if not more important task component is the orientation of the envelope once it reaches the mailbox slot. This emphasizes the need to adequately characterize and represent the functional task to be practiced within the VE. The Rutgers ankle and hand systems, both incorporating the haptic sense, were developed as assessment and intervention tools although there are limited clinical data available at this time regarding efficacy [see [ 35 , 36 ]]. The two systems combine force feedback with a virtual environment that requires subjects to complete various tasks such as a virtual PegBoard task as well as reach-to-grasp (hand system) or piloting a virtual airplane through loops (ankle system). Preliminary data suggest that the systems may be useful to augment rehabilitation in patients in the chronic phase following stroke. A recent study using the hand system demonstrated transfer of skills acquired with the VR system to a functional clinical outcome measure as well as improvement on a variety of movement parameters with greatest benefit recorded in the least impaired patients [ 37 ]. Exercise and pain tolerance Chuang et al [ 38 ] compared physiological responses of the cardiovascular and respiratory systems during incremental exercise testing with and without VR in healthy older adults. A mechanically braked bicycle was linked to a visual virtual scene projected on a flatscreen display. The rate of subject movement on the bicycle matched the environmental flow on the screen and included a 5 km straight or curved road bordered by grass, trees, seashore background and street lamps. No differences were observed on submaximal and peak exercise responses but the cycling with the VR scenario resulted in longer mean values for cycling duration, distance and energy consumption. It is possible that performing the exercises while immersed in a comfortable environment resulted in an increased degree of relative tolerance. Positive outcomes of virtual reality as a distractive technique have also been reported for physiotherapy treatment sessions. Hoffman and colleagues [ 39 ] report decreased anxiety and reductions in self-report of pain from a single-pediatric patient undergoing post-operative physiotherapy. The child underwent single event multi-level surgery including femoral de-rotation osteotomy, quadriceps tendon translocation and release of the Achilles and hamstring tendons. Children experience high levels of post-operative pain association with physiotherapy treatments despite standardized pharmacological analgesia. Effective use of VR as a non-pharmacological analgesia for patients post-surgery may result in greater therapy gains. Assessment Although the majority of VR environments that have been developed for assessment to date focus on daily living skills such as meal preparation [ 40 ], spatial memory [ 8 ] and cognitive function [ 41 ], specific applications have been developed for assessment of upper and lower extremity motor function, balance and locomotion. For example, two separate assessment approaches using the PHANTOM haptic interface, a 6 degree of freedom measuring device for positional input that provides feedback force in translation and rotation have been developed. Broeren et al [ 42 , 43 ] used a relatively simple task requiring the user to reach for, grasp and move the visual representation of the device from a home position to nine separate locations in the visual field. Preliminary data suggest that this is a potential tool for identifying specific deficits of movement such as timing or accuracy that vary across patients. A more complex use of the technology, labyrinth navigation, has been used to isolate more subtle aspects of movement in patients with neurological disease including tremor amplitude and frequency, movement control, and speed of advancement through the labyrinth [ 44 ]. Assessments can be developed using VR technologies that will provide objective, repeatable and quantitative results. Standardized instructions, non-varying environmental cues, tasks and feedback can be achieved. In the extreme condition, interactions are limited to those between the patient and a virtual assessor. Since the devices are programmable, varying the complexity of assessment tasks is relatively trivial allowing for batteries of simple and more complex tasks to be developed. For example, an upper extremity assessment scale may include tasks requiring self-selected motion as well as responses to force perturbations permitting assessment of feedback limb control. Access to rehabilitation The degree of functional movement outcome achieved by therapy is often sub-optimal since intensive therapy is limited by resource allocation and access. For many individuals, such as traumatic brain injury survivors, access to therapy is terminated once a level of function is achieved even if residual deficits remain. For other individuals, even when therapy is available such as during in-patient neurological rehabilitation, low levels of interaction between the patient and environment have been reported [ 45 , 46 ]. For example, Tinson [ 46 ] reported that individuals post stroke typically spent only 20–60 minutes per day in formal therapy. Common problems influencing the degree of interaction include boredom, fatigue, lack of motivation and lack of cooperation in attending therapy [ 47 ]. Clinicians agree that such problems are undesirable and restrict progress in rehabilitation. Increasing interaction is seen as vital to effective rehabilitation, a fact borne out by experimental studies of recovery after brain damage [ 48 ]. Development and incorporation of virtual reality applications in rehabilitation may increase the possibility of stimulation and interaction with the world with potentially little or no increase on the demands of staff time. Virtual reality may provide interesting and engaging tasks that are more motivating than formal repetitive therapy. In fact, our recent experience comparing participant perceptions of exercise programs strongly suggest there is added benefit with VR compared to a conventional program (M Thornton et al, unpublished data). For example, the son of a TBI survivor participating in a VR balance retraining program noted We have tried in the past to have him involved in things but he seemed uninterested. With these exercises (referring to a VR-exercise balance retraining program) he was trying to explain what he was doing, he was interested in what he was doing, he was looking forward to going . Summary An exponentially increasing number of distinct VR applications are being developed for intervention and assessment of a broad range of motor rehabilitation needs including upper and lower extremity function, balance and locomotion. Although the initial VR rehabilitation applications that were developed, in particular applications using video capture technologies and most HMDs, were subjected to relatively prohibitive entry level costs associated with the technology, recent developments in technology have made the number of low-cost multisensory VR applications increasingly available. Significant decreases in the costs associated with HMDs and motion trackers, desktop computers and certain haptic devices, are facilitating the development of low cost off-the-shelf applications. The applications reviewed in this paper have demonstrated improvements of specific motor function with certain populations. It is clear that many of the applications that have been developed, for example gait trainers, will serve a specific rehabilitation niche. These devices have the potential to significantly extend our current understanding of movement and therapy and may substantially impact delivery of rehabilitation interventions. Critical for continued successful integration of virtual reality in motor rehabilitation is the need for the ongoing development and use of the technology to be based on clear understanding of the complexity of voluntary movement [ 49 ]. Sensorimotor integration, movement production, learning and transfer as well as psychosocial benefits are critical issues to address in ongoing and future studies. Of crucial importance is the fundamental question "Can the same objective be accomplished with a simpler approach". Prior to adoption of novel rehabilitation approaches including virtual reality based applications, users must assess whether the VR technology will provide any additional benefits to that of well trained and experienced therapists.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546406.xml
554764
Efficacy of two artemisinin combination therapies for uncomplicated falciparum malaria in children under 5 years, Malakal, Upper Nile, Sudan
Background The treatment for Plasmodium falciparum malaria in Sudan has been in process of change since 2003. Preceding the change, this study aimed to determine which artemisinin-based combination therapies is more effective to treat uncomplicated malaria in Malakal, Upper Nile, Sudan. Methods Clinical trial to assess the efficacy of 2 antimalarial therapies to treat P. falciparum infections in children aged 6–59 months, in a period of 42 days after treatment. Results A total of 269 children were followed up to 42 days. Artesunate plus Sulfadoxine/Pyrimethamine (AS+SP) and Artesunate plus Amodiaquine (AS+AQ) were both found to be efficacious in curing malaria infections by rapid elimination of parasites and clearance of fever, in preventing recrudescence and suppressing gametocytaemia. The combination of AS+SP appeared slightly more efficacious than AS+AQ, with 4.4% (4/116) versus 15% (17/113) of patients returning with malaria during the 6-week period after treatment (RR = 0.9, 95% CI 0.81–0.96). PCR analysis identified only one recrudescence which, together with one other early treatment failure, gave efficacy rates of 99.0% for AS+AQ (96/97) and 99.1% for AS+SP (112/113). However, PCR results were incomplete and assuming part of the indeterminate samples were recrudescent infections leads to an estimated efficacy ranging 97–98% for AS+SP and 88–95% for AS+AQ. Conclusion These results lead to the recommendation of ACT, and specifically AS+SP, for the treatment of uncomplicated falciparum malaria in this area of Sudan. When implemented, ACT efficacy should be monitored in sentinel sites representing different areas of the country.
Introduction The health situation in Sudan continues to be affected by long-lasting conflict and related humanitarian emergencies such as food crises and epidemics. Malaria is one of the major causes of morbidity and mortality. In Sudan, an estimated 7.5 million patients suffer from malaria each year and 35,000 die from this disease, which accounts for up to 20% of hospital deaths [ 1 ]. The problem appears to have worsened in recent years due to increasing levels of Plasmodium falciparum resistance against the two most commonly used antimalarials: chloroquine (CQ) and sulfadoxine/pyrimethamine (SP). CQ resistance in the northern and central part of the Sudan is near 50% [ 2 ]. In the South, CQ resistance below 15% was found in some isolated locations [ 3 , 4 ], but was over 80% elsewhere [ 5 , 6 ]. Resistance to SP was shown by presence of SP-resistant genotypes [ 7 ] and in vivo studies documenting of 0–11% resistance in Khartoum and the eastern part of the country [ 2 , 8 ] and varying levels of 0, 16% up to 70% in different locations in the southern part of the country [ 4 - 6 ]. Amodiaquine (AQ) provided an alternative for CQ in Africa in recent years; this drug is not officially registered in the Sudan, and due to its limited use, was expected to have preserved good efficacy. The nationally recommended treatment protocol for P. falciparum malaria in Sudan at the time of this study was still CQ as first-line and SP as second-line. A change in protocol was however in progress, and the Sudanese Health authorities had indicated that artemisinin-based combination therapy (ACT) was the preferred new first-line treatment [ 9 ]. The World Health Organization (WHO), MSF and various other NGOs supported this viewpoint [ 10 , 11 ]. In line with recommendations for several other African countries, two ACTs were proposed for Sudan: Artesunate (AS) +SP or AS+AQ. The third option, used in other African countries, Artemether-Lumefantrine (Coartem) was not seen as a viable option at the time, because of its cost. To provide information on the efficacy of ACT treatment in Sudan, the in vivo therapeutic efficacy of the first two combinations were studied in Malakal town, Upper Nile State. Patients and Methods Malakal is located in southern Sudan (on the edge of the Khartoum-governed area), which is an area is of medium to high malaria endemicity (main transmission season: June-November). This open-label study was done in Malakal paediatric hospital supported by Médecins Sans Frontières, from September 2003 to January 2004. The methodology used was in accordance with standard WHO procedures for antimalarial drug efficacy assessment in high transmission settings [ 12 , 13 ]. Children aged between 6 months and 5 years with fever (axillary temperature of 37.5°C or more) and P. falciparum infection with a density of 2,000 to 200,000 parasites/μl but no signs of severe malaria (WHO criteria, [ 13 ]) or any other serious health condition were included in the study (provided written parental consent was given). One of the two therapies was allocated (by sequential alternation), and given under supervision: either (1) AS+SP, 3 days AS 4 mg/kg body weight on day 0, 1 and 2 plus SP, 25/1.25 mg/kg, single dose on day 0 or (2) AS+AQ, 3 days AS 4 mg/kg on day 0, 1 and 2 plus AQ 10 mg/kg on day 0, 1, 2. Doses were calculated for weight categories in quarter tablets. In case of vomiting within 30 minutes, the treatment was repeated. Repeated vomiting led to exclusion from study and referral to the hospital for quinine treatment. Medications were normally given in the mornings and appointments for next day roughly matched 24 hr intervals. Artesunate was obtained from Dafra (Belgium) and SP and AQ from IDA (the Netherlands). The sample size of at least 116 inclusions per arm was calculated to detect a 12% difference between the two arms, assuming a 97% efficacy at day 28 for the AS+SP arm (based on 100% efficacy of SP found in a recent study performed at a nearby location, [ 4 ]) and 85% for the AS+AQ arm (internationally recommended level for change), with 80% power and 5% risk type I error, anticipating a loss to follow-up of 10%. Patients returned daily during the first 3 days of illness and weekly thereafter for 6 weeks (42 days) after treatment. Side effects of drugs and self-medication were recorded based on patient reports and actively asked for during visits. Cases which remained negative during follow-up were considered Adequate Clinical and Parasitological Responses (ACPR). Failures were classified as (1) Early Treatment Failures (ETF) in case of significant parasitaemia at day 2 or 3 or parasites and fever at day 3, (2) Late Clinical Failures (LCF) for cases with parasites and fever during follow-up after day 3 and (3) Late Parasitological Failures (LPF) for parasite infections without fever at day 42, all in accordance with WHO guidelines [ 13 ]. Rescue treatment for failures was quinine 10 mg/kg/eight-hourly for 7 days. PCR analysis was used to distinguish new infections from recrudescences. Slides were examined by two microscopists independently and 20% of slides were cross-checked in an external laboratory (at KEMRI, Nairobi). The density of parasites was determined by simultaneous count of white blood cells and parasites, assuming a standard density of 8,000 WBCs per μl [ 13 ]. Haemoglobin levels were evaluated at day 0, 14, 28 and 42, to observe return to healthy Hb-levels after effective antimalarial treatment. PCR analysis of the msp1 and msp2 loci was used to distinguish re-infections from recrudescences [ 14 , 15 ]. In case of two or more identical alleles in pre- and post-treatment genotypes, the case was classified as recrudescent, or as a new infection if one or none of the alleles matched. Cases were excluded from the final analysis in case of loss to follow-up, self-medication with antimalarials and concomitant disease and when PCR results showed re-infections or were undetermined due to missing samples, undetermined genotypes or non-amplifiable DNA. The research protocol was reviewed and approved by the Sudan Research Directorate, the Malaria, Leishmaniasis and Schistosomiasis Directorate, as well as by the Ethical Review Board of MSF (comprising of MSF-external experts only). During recruitment of the children in study, informed written consent was obtained from their parents/guardians. Results A total of 269 children were recruited during the 5-month study period. Of these, 134 were treated with AS+AQ and 135 with AS+SP. Baseline characteristics were similar in both treatment groups (Table 1 ). Of the children included, 40 were excluded from the final analysis due to associated serious other febrile illness (n = 15), self-medication with antimalarials (n = 7), loss to follow-up (n = 13) or erroneous inclusion/classification (n = 5, i.e. three cases of disagreement over microscopy quality control, two cases retreated with parasites but no fever prior to day 42 in AS+AQ group). There were three children with Hb level of 5.0 g/dl included (limit of in/exclusion), based on their condition assessed by the medical doctor. Altogether, 113 (84%) completed the study after treatment with AS+AQ and 116 (86%) with AS+SP (Figure 1 ). Table 1 Baseline characteristics of patients at study-enrolment, per treatment arm, Malakal, Upper Nile, Sudan 2003–4. AS + AQ (n = 134) AS + SP (n = 135) Sex ratio (F/M) 68/66 64/71 Mean age (months) (SD, ranges) 33 (13, 7–59) 33 (15, 6–59) Mean weight (kg) (SD, ranges) 11.1 (2.4, 5.7–17.0) 11.3 (2.6, 5.4–18.0) Mean axillary temperature (°C) (SD, ranges) 38.7 (0.9, 37.5–40.9) 38.9 (0.9, 37.5–40.9) Mean haemoglobin value (g/dl) (SD, ranges) 7.8 (1.7, 5.0–12.2) 7.9 (1.8, 5.0–12.7) Parasitaemia geometric mean (/μl) (ranges) 20 952 (2 100–199 500) 24835 (2 000–198 000) Figure 1 Patient flow during recruitment and follow-up Both artemisinin-combitherapies were highly effective to treat P. falciparum infections and prevent parasite re-emergence. Only 1 case of ETF was found, in the AS+SP group (Table 2 ). At day 28 and day 42, the proportion of patients still parasite-free was higher in the group treated with AS+SP than in the group AS +AQ (χ 2 test, p = 0.014 for results day 28; p = 0.0049 for day 42). Table 2 Treatment endpoints on day 28, 42 after treatment and PCR adjusted day 42, Malakal, Upper Nile, Sudan, 2003–2004 AS + AQ AS + SP N % 95% CI N % 95% CI Day 28 Results 117 116 ACPR 105 89.7 82.8–94.6 114 98.3 93.9 – 99.8 ETF 0 0 1 0.9 0.0–4.7 LCF 10 8.5 4.2–15.2 1 0.9 0.0–4.7 LPF 2 1.7 0.2–6.0 0 0 Day 42 Results 113 116 ACPR 96 85.0 77.0–91.0 112 96.6 91.4–99.1 ETF 0 0 1 0.9 0.0–4.7 LCF 15 13.3 7.6–20.9 2 1.7 0.2–6.1 LPF 2 1.8 0.2–6.2 1 0.9 0.0–4.7 PCR corrected day 42 # 97 113 No PCR result (12) (2) Re-infections (4) (1) ACPR 96 99.0 94.4–100 112 99.1 95.2–100 ETF 0 0 0.0–3.7 1 0.9 0.0–4.8 LCF 1 1.0 0.0–5.6 0 0 0.0–3.2 LPF 0 0 0.0–3.7 0 0 0.0–3.2 # Cases with missing or undetermined PCR results and re-infections were excluded from the analysis. ACPR = Adequate Clinical Parasitological Response, ETF = Early Treatment Failure, LCF = Late Clinical Failure, LPF = Late Parasitological Failure. PCR analysis was unable to generate results for a large part of cases (only 6/20 available, Figure 1 ). In the AS+AQ treatment group, of 17 failures, only 1 was confirmed to be recrudescent, whereas of the other cases, 4 were re-infections and 12 samples were not analyzable. Of the 3 'failures' in the AS + SP group, one was an early treatment failure, PCR identified 1 re-infection and 2 were non-analysable. Hence, with the PCR correction and exclusions of non-analysable and re-infection cases, AS+AQ resulted in 1.0% confirmed failures (1/97) and AS+SP in 0.9% true failures (1/113) (Table 2 ). Extrapolation of the cases with known PCR results to the ones that remained undetermined (assuming 2/3 of parasitemia due to reinfection), would give a 95.3% efficacy of AS+AQ (101/106) and 98.2% of AS+SP (112/114). Alternatively, a 'worst case' scenario, assuming that all non-analysable PCRs were actually recrudescences, would lead to 88.1% efficacy in the AS+AQ group (96/109) and 97.4% for AS+SP (111/114). Parasites were rapidly cleared in both treatment groups. At day 2 only 17% and 22% of patients still had a (low) parasitaemia in the AS+AQ and AS+SP groups, respectively. Gametocytes were found in 20% (52/262) of children on inclusion. During follow-up, 78% of these carriers at enrolment showed gametocytes again on one or more follow-up days (32/41), whereas 22% of the children without gametocytes at inclusion were later found to be gametocytemic (42/192). There was no difference between the age, temperature, parasitemia, haemoglobin on admission of the two groups, neither in rates of ACPR at study-endpoint. Gametocyte carriage was similar in both treatment groups and all gametocytes disappeared gradually from the blood during the first weeks after follow-up, and cleared by day 21 in 95% of cases (Figure 2 ). Figure 2 Parasitaemia and gametocytaemia during 6-week follow-up. Symbols: AS+SP ▲ parasites △ gametocytes, AS+AQ ● parasites ○ gametocytes. Error bars indicate 95%CI. The body temperature of all patients went down quickly; fever clearance (T<37.5°C) at day 1 and day 2 was 86% and 97% in the AS + AQ group and 83% and 93% in the AS+SP, and at day 3 all but one (on AS+SP) were free from fever. Average haemoglobin levels for the AS+AQ and AS+SP treatment groups increased from under 8.0 g/dl to near 10 g/dl 6 weeks after treatment. The proportion of children classified as moderately anaemic (haemoglobin from 5 to 8 g/dl) reduced from 51% to 11% during follow-up period, similar in both arms. No serious adverse events were observed or reported on routine clinical examination during follow-up and none of the failures developed complicated malaria. Discussion This is one of the first studies into the clinical outcomes of ACT combination therapy in Sudan and provides useful information for decision-makers working to ensure effective antimalarial protocols in this part of the country. Both of the artemisinin-based combination therapies tested here were found to be highly efficacious in the treatment of uncomplicated P. falciparum malaria in this area of the Sudan. AS+SP appears to be the better treatment option on the basis of non-PCR corrected responses, showing a lower percentage of patients returning with parasitaemia. The PCR analysis indicates the true efficacy is comparable between both treatments (near 99%), but is, however, limited by a high proportion of indeterminate cases. More realistically, an efficacy between 97–98% can be expected for AS+SP and 88–95% for AS+AQ, acceptable levels after the long follow-up of 42 days after treatment. A rapid parasite clearance and fever reduction was found following treatment with both ACTs. The rise in haemoglobin values and the reduction of the proportion of (moderately) anaemic children after treatment confirms that the malaria parasites were effectively removed from the blood and red blood cell levels rose after treatment. The ACTs tested also had an effect on gametocytes. In general, the gametocidal action of AS appears to work through preventing the development of new gametocytes rather than clearance of existing ones [ 16 ]. In our study, the 20% of gametocyte-carrying infections at enrolment cleared gradually by day 21. Newly detected gametocytes developed in 22% of cases after treatment, which is lower than after monotherapy, at least for SP. A previous study in Upper Nile [ 4 ] showed that 68% of patients (of all ages) treated with SP and 28% of those treated with AQ developed gametocytaemia during the 14 days after treatment, while gametocyte prevalence at admission was only 2%. Gametocidal effect is very important since the sexual stages of parasites are essential for person-to-person transmission of malaria via mosquito vectors. The main limitations of the study were that the number of patient exclusions were higher than anticipated (15%) due to concomitant febrile illnesses and loss to follow-up, as well as the lack of results for PCR-analysis, caused by samples missing or an inability to provide sufficient DNA on amplification due to low densities in post-treatment samples. Increasing the number of inclusions compensated losses to follow-up. Repeated attempts at PCR-analysis for problematic samples were only partly successful and, therefore, lead to extrapolation about the findings for the missing samples. At time of writing, a change of national protocols towards ACT in northern Sudan is in preparation – coordinated among health authorities, NGOs and other relevant actors. The first line treatment recommended country-wide is AS+SP, based on reported high efficacy of SP. Northern Sudan is the only country in Africa, which has chosen the option AS+SP [ 17 ]. On the basis of the results of this study, this is justifiable. One other study in Sudan on AS+SP and AS+AQ efficacy was recently completed in the Nuba Mountains and also shows high efficacy of 91.2% and 92.7%, respectively, for these ACTs at 28 days after treatment [ 18 ]. ACTs with SP or AQ as companion drugs have shown to be very effective in other areas of Africa, provided that the companion drug still maintained a good level of efficacy [ 19 - 22 ]. In the Upper Nile area both SP and AQ were still an effective treatment for P. falciparum and SP has shown high efficacy in various areas of northern Sudan [ 2 , 4 , 23 ]. It remains to be seen whether AS+SP will be equally efficacious in other, e.g. more northern areas of the Sudan. In Southern Sudan AS+AQ has recently been put forward as the therapy of choice [ 17 ]. The use of two different therapies would seem a sensible option in a vast country with areas of different patterns of resistance. An advantage of AS+SP over AS+AQ is that it is more convenient to administer, as SP is given as a single dose and can be administered under observation in a health facility, whereas AQ requires 3 days to complete a course. SP tablets are also easier to take (AQ has a bitter taste). Blister packs of AS+SP, which combine the two drugs and clearly indicate daily tablets to be taken, are currently available for different age categories in Sudan. AS+SP is cheaper than AS+AQ. Implementation of the new national protocol with AS+SP will hopefully take place as quickly as possible to prevent rising morbidity and mortality. Vulnerable displaced populations in epidemic-prone areas and areas of high perennial malaria transmission should be prioritized. Introduction of ACT will have to go hand in hand with laboratory confirmed diagnosis (microscopy or rapid diagnostic tests) to prevent unnecessary use of valuable drugs (thus minimizing drug pressure) and ensure that non-malarial cases are appropriately treated. The change in guidelines should also filter through all health service providers, including the private sector and drug vendors to decrease the potential risk of SP monotherapy and incomplete dosages. The change will initially require more funds to be made available for malaria treatment. The international community – including many NGOs, the WHO, donors, and the Global Fund – has shown willingness to support countries to change antimalarial protocols [ 24 , 25 ]. At present, MSF has already started to pilot AS+SP treatment in its project areas in Northern Sudan, i.e. Darfur, Upper Nile and Gedaref, on behalf of the Ministry of Health. The implementation of treatment as well as the future efficacy of AS+SP should be monitored carefully in a number of dispersed sentinel sites, as there is a possibility that SP resistance may further rise before the combination has been made available countrywide. Authors' contributions IB was responsible for overall supervision, data analysis and writing of the paper, RA for field-implementation, communication as well as data analysis and writing, MB for preparation and set-up of study, support and advocacy. PH was responsible for microscopy laboratory work and quality. EA participated as field supervisor of clinical work and data gathering. EBH and MAF were responsible for day-to-day study procedures and supervision of clinical teams. JM did the PCR analysis and FC was involved in design of study, data analysis and writing of manuscript. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554764.xml
509305
Spontaneous Autoimmunity in 129 and C57BL/6 Mice—Implications for Autoimmunity Described in Gene-Targeted Mice
Systemic lupus erythematosus (SLE) is a multisystem autoimmune disorder in which complex genetic factors play an important role. Several strains of gene-targeted mice have been reported to develop SLE, implicating the null genes in the causation of disease. However, hybrid strains between 129 and C57BL/6 mice, widely used in the generation of gene-targeted mice, develop spontaneous autoimmunity. Furthermore, the genetic background markedly influences the autoimmune phenotype of SLE in gene-targeted mice. This suggests an important role in the expression of autoimmunity of as-yet-uncharacterised background genes originating from these parental mouse strains. Using genome-wide linkage analysis, we identified several susceptibility loci, derived from 129 and C57BL/6 mice, mapped in the lupus-prone hybrid (129 × C57BL/6) model. By creating a C57BL/6 congenic strain carrying a 129-derived Chromosome 1 segment, we found that this 129 interval was sufficient to mediate the loss of tolerance to nuclear antigens, which had previously been attributed to a disrupted gene. These results demonstrate important epistatic modifiers of autoimmunity in 129 and C57BL/6 mouse strains, widely used in gene targeting. These background gene influences may account for some, or even all, of the autoimmune traits described in some gene-targeted models of SLE.
Introduction Systemic lupus erythematosus (SLE) is a chronic autoimmune disease characterised by the production of autoantibodies (auto-Abs) against a wide spectrum of self-antigens, mainly from subcellular compartments, especially the cell nucleus. Genetic predisposition is an important contributor to susceptibility to SLE in both humans and animals ( Vyse and Todd 1996 ; Harley et al. 1998 ; Theofilopoulos and Kono 1999 ; Wakeland et al. 2001 ). Genes in multiple pathways participate in mediating disease pathogenesis, and epistatic interactions amongst these genes influence the expression of disease. In this context, both genetic linkage studies in spontaneous lupus-prone models and synthetic murine models of autoimmunity generated by targeted disruption of specific genes modulating the immune system have widely been used to investigate the complexity of SLE. The best-studied strains of mice that spontaneously develop a lupus-like pathology are the New Zealand Black/New Zealand White hybrid strain (NZB/WF1); the MRL/Mp lpr/lpr strain, which carries the lpr mutation of the FAS receptor gene; and the BXSB strain, which carries the Y chromosome autoimmune accelerator ( Yaa ) gene ( Theofilopoulos and Dixon 1985 ). Extensive genetic mapping studies in all three strains have identified multiple intervals associated with disease susceptibility. Interestingly, the majority of the intervals detected are strain-specific, confirming the genetic complexity of the disease and indicating the presence of extensive heterogeneity in the genes contributing to the pathogenesis of the disease. However, some susceptibility loci have been mapped to similar chromosome locations in different strains, suggesting that at least some susceptibility may be shared amongst lupus-prone strains. Amongst these shared susceptibility loci, the most striking are loci on distal Chromosome 1, for which important contributing genes have been found in New Zealand and BXSB models ( Theofilopoulos and Kono 1999 ; Wakeland et al. 2001 ). Although considerable efforts have been made to identify the genes responsible for the development of the disease, with the exception of the lpr mutation, none of the genetic contributions to disease in the three well-documented murine SLE strains have yet been fully resolved at the molecular or protein level. Thus, targeted genetic disruption of candidate genes encoding proteins of the immune system has been extensively used to examine their role in immune regulation. However, the most surprising result of this powerful approach has been the high frequency with which such mutations have been associated with an autoimmune phenotype. In this regard, it is of note that hybrid strains between 129 and C57BL/6 mice, widely used in the generation of gene-targeted mice, are spontaneously predisposed to development of humoral autoimmunity with low levels of glomerulonephritis ( Obata et al. 1979 ; Botto et al. 1998 ; Bickerstaff et al. 1999 ; Santiago-Raber et al. 2001 ). Furthermore, the genetic background markedly influences the autoimmune phenotype in gene-targeted mice ( Bolland and Ravetch 2000 ; Santiago-Raber et al. 2001 ; Mitchell et al. 2002 ). These observations led to the hypothesis that the autoimmune phenotype described in some gene-targeted mice might be due primarily to combinations of as-yet-uncharacterised background genes, originating from 129 and C57BL/6 mice strains, interacting or not with the mutated allele. To test this, we conducted a genome-wide scan analysis of two large cohorts of (129 × C57BL/6)F2 mice, one of which carried a mutation in the serum amyloid P component gene ( Apcs ). The Apcs -deficient mice ( Apcs −/− ) were chosen as an example of a gene-targeted strain previously reported to develop a lupus-like disease on the hybrid genetic background (129 × C57BL/6); autoimmunity in Apcs −/− mice persists even after backcrossing the mutated gene onto C57BL/6 ( Bickerstaff et al. 1999 ). We chose this targeted gene in particular to study since the Apcs gene is located on Chromosome 1, approximately 94 cM from the centromere, within a region where several lupus-susceptibility loci, designated Sle1 ( Morel et al. 2001 ), Nba2 ( Drake et al. 1995 ; Vyse et al. 1997 ), and Bxs3 ( Hogarth et al. 1998 ; Haywood et al. 2000 ), have been mapped in NZW, NZB, and BXSB lupus-prone strains, respectively. This region contains several genes, including those encoding FcγRII, the complement receptor CR1/2 (CD35/CD21), and the decay-accelerating factor CD55 ( Prodeus et al. 1998 ; Bolland and Ravetch 2000 ; Miwa et al. 2002 ; Wu et al. 2002 ), which have each been implicated in the causation of SLE when inactivated by gene-targeting in 129 embryonic stem cells. Here we show that there are multiple genetic loci, derived from both 129 and C57BL/6 mice, contributing to autoimmunity. Furthermore, a 129-derived interval on distal Chromosome 1, when transferred onto the C57BL/6 genome, a combination commonly created by backcrossing onto C57BL/6 a gene that has been inactivated in 129 embryonic stem cells, was sufficient to cause humoral autoimmunity in its own right, irrespective of the presence of the mutated Apcs gene. These results demonstrate important epistatic interactions between genes from 129 and C57BL/6 genomes on the development of autoimmunity and illustrate the important effects of background genes in the analysis and interpretation of autoimmune phenotypes associated with targeted genetic disruptions. Results Disease Traits in (129 × C57BL/6)F2 and (129 × C57BL/6)F2. Apcs −/− Mice To investigate the genetic basis of the lupus-like disease observed in the (129 × C57BL/6) hybrid mice, we generated two large cohorts of (129 × C57BL/6)F2 animals, one carrying a mutation in the Apcs gene, and monitored them for 1 y under identical environmental conditions. Since female mice in the original reports showed a higher penetrance of disease, the present study was conducted only on female mice. The results of the phenotypic analysis at 1 y of age are summarised in Tables 1 and 2 . As previously reported ( Botto et al. 1998 ), the wild-type (129 × C57BL/6)F2 mice developed lupus traits with elevated levels of auto-Abs, starting from 6 mo of age (data not shown), and histological evidence of proliferative glomerulonephritis. In agreement with our previous observations ( Bickerstaff et al. 1999 ), the titres of anti-nuclear Abs (ANAs) and anti-chromatin Ab were considerably greater in the (129 × C57BL/6)F2. Apcs −/− mice compared with the strain-matched controls. However, in contrast to our original findings, the levels of the other two disease serological markers analysed (anti-single-stranded DNA [ssDNA] and anti-double-stranded DNA [dsDNA] Abs) and the severity of the renal pathology were not different between the two experimental groups. In view of the possibility of an association between the fixed 129-derived segment flanking the mutated Apcs gene and the autoimmune traits observed, the genome-wide linkage analysis of the two experimental cohorts was carried out separately. Table 1 Spontaneous Auto-Abs in Apcs −/− and Wild-Type (129 × C57BL/6)F2 Female Mice Significant differences between Apcs −/− and wild-type controls by Mann–Whitney U test. The numbers tested for different phenotypes are not equal due to the limited amount of serum or death of the mice before the end of the experiment. NS, not significant Table 2 Histological Assessment of Kidney Sections in Apcs −/− and Wild-Type (129 × C57BL/6)F2 Female Mice The renal sections were scored on a 0–3 scale on the intensity and extent of the histopathological changes as described in Materials and Methods. The difference between the two experimental groups was not significant Mapping of Loci Predisposing to Lupus in the Hybrid Strain (129 × C57BL/6) Mice were genotyped with 143 microsatellite markers distributed throughout the autosomes such that 98% of the genes were within 20 cM of an informative marker. A summary of the genome-wide linkage analysis for each of the disease traits measured is shown in Table 3 . The areas of linkage were defined according to the parental origin, 129 or C57BL/6. Only linkages identified in both experimental groups are reported in Table 3 , with the exception of the Chromosome 1 distal segment, where the linkage analysis could not be applied to the (129 × C57BL/6)F2. Apcs −/− mice as this region was of fixed 129 origin. Chromosomes where linkages were present only in one of the two cohorts are shown in Figures 1 – 3 . Figure 1 Linkage of Chromosome 2 with ANA and Anti-Chromatin Abs in (129 × C57BL/6)F2 .Apcs −/− Mice These associations were not detected in (129 × C57BL/6)F2 animals. Centimorgan positions were deduced by interval mapping, anchoring marker locations to data from http://www.informatics.jax.org . Dotted lines and the dashed line indicate the threshold over which linkages were considered suggestive or significant, respectively, as defined in Materials and Methods. Figure 3 Linkage of Chromosome 4 with Anti-dsDNA Abs The estimated peak in (129 × C57BL/6)F2 mice was at position 51.3 cM, whilst in the (129 × C57BL/6)F2 .Apcs −/− animals it was was at position 71 cM, indicating that most likely these were two independent loci. Centimorgan positions were deduced by interval mapping, anchoring marker locations to data from http://www.informatics.jax.org . Dotted lines the indicate threshold over which linkage was considered suggestive, as defined in Materials and Methods . Table 3 Summary of Genome-Wide Linkage Analysis in Apcs −/− and Wild-Type (129 × C57BL/6)F2 Female Mice a Suggestive linkage b Significant linkage c Highly significant linkage as defined in the material and method section Chr, Chromosome; Est. Peak, Estimated Peak; LOD, logarithm of odds; N/A, not applicable Chromosomes where linkages were not present in both experimental groups are not illustrated. See Materials and Methods for details. The oligonucleotide sequences and approximate positions of the microsatellite markers used were taken from the Mouse Genome Database, Mouse Genome Informatics, Jackson Laboratory, Bar Harbor, Maine, United States ( http://www.informatics.jax.org ) The quantitative trait linkage (QTL) analysis identified several intervals on Chromosome 1 with linkage to disease serological markers, and these regions were all derived from the 129 mouse strain (see Table 3 ; Figures 4 and 5 ). Interestingly, whilst ANA and anti-chromatin Ab levels showed suggestive or significant linkages only to the telomeric region of Chromosome 1, with an estimated peak occurring at a position approximately 90 cM near the Apcs gene, anti-dsDNA or anti-ssDNA Ab production was also linked to other segments on Chromosome 1, indicating a more complex genetic contribution from the 129 mouse strain. Figure 4 Interval Mapping Scans Showing QTL on Chromosome 1 with Anti-dsDNA and Anti-ssDNA Abs Centimorgan positions were deduced by interval mapping, anchoring marker locations to data from http://www.informatics.jax.org . Dotted lines indicate the threshold over which linkage was considered suggestive, dashed lines indicate the threshold over which linkage was considered significant, and dotted/dashed lines indicate highly significant linkage, as defined in Materials and Methods. See Table 3 for additional details. Figure 5 Interval Mapping Scans Showing QTL on Chromosome 1 with ANA and Anti-Chromatin Abs Centimorgan positions were deduced by interval mapping, anchoring marker locations to data from http://www.informatics.jax.org . Dotted lines indicate the threshold over which linkage was considered suggestive, and dashed lines indicate the threshold over which linkage was considered significant, as defined in Materials and Methods. See Table 3 for additional details. Guided by these observations, we investigated whether the increased levels of ANA and anti-chromatin Ab observed in the Apcs −/− mice were caused by a gene(s) within the fixed 129 region surrounding the mutated Apcs gene, rather than caused by the mutated Apcs gene itself. We compared the levels of these auto-Abs between all (129 × C57BL/6)F2. Apcs −/− mice and a group of 33 wild-type mice that were selected for being homozygous 129 in the region of Chromosome 1 between microsatellites D1Mit105 and D1Mit 223 (80–106 cM) ( Figure 6 A– 6 D). In contrast to the results reported in Table 1 , this comparison showed no significant differences between the two experimental groups. This result demonstrates that, most likely, the 129-derived region and not the lack of Apcs was mediating the production of ANA and anti-chromatin Ab. Consistent with this explanation, we found that the 129 mice have significantly higher levels of Apcs in circulation compared with the C57BL/6 mice (median, 83 mg/l; range, 25–208; n = 16 versus median, 5 mg/l; range, 4–9; n = 10, respectively; p < 0.0001). The C57BL/6 strain has previously been reported to be one of the murine strains defined as low Apcs producers ( Pepys et al. 1979 ; Baltz et al. 1980 ). In addition, sequence analysis of the entire Apcs coding region in both strains failed to identify any coding sequence polymorphisms in the Apcs gene (data not shown), indicating that a structural variant of the protein is unlikely to be the explanation for our findings. This is consistent with a previous report by Drake et al. (1996) that showed no Apcs coding sequence differences amongst several autoimmune and nonautoimmune murine strains. Figure 6 Auto-Ab Profiles (A) ANA titres in the (129 × C57BL/6)F2. Apcs −/− mice and (129 × C57BL/6)F2 at 1 y of age. A small circle represents one mouse; a large circle, a variable number of animals, as indicated in parentheses. Serum samples were titrated to endpoint. (B) ANA titres in the (129 × C57BL/6)F2. Apcs −/− mice and a selected number of wild-type (129 × C57BL/6)F2 animals carrying the Chromosome 1 region between D1Mit105 and D1Mit223 (80–106 cM) of 129 origin. The symbols are as in (A). (C and D) Anti-chromatin Ab levels expressed in AEUs related to a standard positive sample, which was assigned a value of 100 AEU. The comparison is between the same groups of mice as in (A) and (B), respectively. The symbols are as in (A). In addition to the 129-derived segments, in both cohorts the C57BL/6 strain contributed to the autoimmune traits with one major susceptibility locus on Chromosome 3. A genomic region between D3Mit40 and D3Mit13, with an estimated peak at position approximately 51 cM, showed a significant linkage to ANA production and weaker linkages to anti-ssDNA and anti-chromatin production (see Table 3 ; Figure 7 ). The high frequency of autoimmune phenotype in the (129 × C57BL/6) hybrid genetic background and its absence in either of the inbred parental strains imply that there are essential interactions between 129- and C57BL/6-derived alleles for the expression of autoimmunity. We investigated further the effects of genes from the C57BL/6 background by repeating the linkage analysis in (129 × C57BL/6)F2 mice, whilst controlling for the very strong 129 effect on distal Chromosome 1, as previously described ( Zeng 1994 ). The results of this analysis showed that the statistical support for the linkage of the C57BL/6 locus on Chromosome 3 for ANA increased from logarithm of odds (LOD) 5.4 to LOD 6.4. Figure 7 Interval Mapping Scans Showing QTL on Chromosome 3 with ANA, Anti-Chromatin, and Anti-ssDNA Abs See Table 3 for additional details. Centimorgan positions were deduced by interval mapping, anchoring marker locations to data from http://www.informatics.jax.org . Dotted lines indicate the threshold over which linkage was considered suggestive, the dashed line indicate the threshold over which linkage was considered significant, and dotted/dashed lines indicate highly significant linkage, as defined in Materials and Methods. In contrast to these strong associations with disease serological markers, the QTL analysis identified only two potential linkages to glomerulonephritis: one in the wild-type mice on Chromosome 7 across a 10 cM region between D7Mit246 (15 cM) and D7Mit145 (26.5 cM) of 129 origin (LOD 2.86, p = 0.0013), and one on Chromosome 17 between D17Mit100 (11.7 cM) and D17Mit216 (29.4 cM) from the C57BL/6 strain (LOD 1.3, p = 0.049 and LOD 1.67, p = 0.021 in the wild-type and Apcs −/− mice, respectively). Histological evidence of glomerulonephritis was only found in approximately 20% of the mice in each cohort, which reduces the power of the QTL analysis for this disease trait. Production of a C57BL/6.129 Chromosome 1 Congenic Line and Its Phenotypic Analysis We generated a C57BL/6 congenic line carrying the telomeric region of Chromosome 1 from the 129 mouse strain, in order to dissect the complex polygenic disease phenotype of the (C57BL/6 × 129/Sv)F2 hybrid strain into its individual genetic components. The 129 interval was backcrossed seven times onto C57BL/6, and at each generation the presence or absence of the Chromosome 1 interval was determined with several microsatellite markers. Each backcrossed generation was screened with more than three markers per chromosome to facilitate the removal of unselected 129 genomic regions. At the end of the backcrossing, the 129-derived Chromosome 1 interval in the congenic mice extended from microsatellite marker D1Mit105 to D1Mit223 (80–106 cM), which encompasses the most important 129 interval identified by the linkage studies in the (C57BL/6 × 129/Sv)F2 mice. Female Chromosome 1 congenic mice (C57BL/6.129[D1Mit105–223]), together with sex-matched Apcs −/− mice backcrossed onto C57BL/6 for ten generations (C57BL/6. Apcs −/− ) and C57BL/6 controls, were monitored for the presence of lupus. In the C57BL/6. Apcs −/− mice, the 129 genome around the Apcs locus was mapped as a stretch of approximately 17 cM, positioned from 87.9 cM (D1Mit15) to 105 cM (D1Mit17). Thus, the congenic line carried a similar 129 region (80–106 cM) to the one present in the C57BL/6. Apcs −/− mice (87.9–105 cM). At 1 y of age, all animals were sacrificed, the auto-Abs assessed, and the renal histology examined. The results of this analysis are shown in Figure 8 . As previously reported ( Bickerstaff et al. 1999 ), the levels of auto-Abs were markedly increased in the C57BL/6. Apcs −/− compared to the wild-type C57BL/6 controls. However, the C57BL/6.129(D1Mit105–223) animals also expressed high levels of auto-Abs, and these titres were not different from those detected in the matched congenic mice containing a null mutation of the Apcs gene. These results clearly demonstrated that epistatic interactions between 129 loci on Chromosome 1 and C57BL/6 genes were sufficient to mediate the loss of tolerance to nuclear autoantigens. However, in contrast to the serological data, the histological assessment of the kidneys showed evidence of markedly increased glomerulonephritis in the C57BL/6. Apcs −/− compared to both control groups ( Figure 9 ), suggesting that the lack of Apcs , when combined with other C57BL/6 susceptibility alleles, can induce the development of severe renal damage. Figure 8 Auto-Ab Profiles (A) ANA titres in C57BL/6 mice, C57BL/6. Apcs −/− mice, and C57BL/6.129(D1Mit105–223) congenic mice at 1 y of age. Small symbols represent one mouse; large symbols, a variable number of animals as indicated in parentheses. Serum sample were titrated to endpoint. (B and C) Anti-ssDNA (B) and anti-chromatin (C) Ab levels in the same cohorts of mice as in (A). The Ab levels are expressed in AEUs related to a standard positive sample, which was assigned a value of 100 AEU. (D) Anti-dsDNA Ab levels. Serum samples were screened at 1:20. Samples that were positive were titrated to endpoint. The symbols are as in (A). Figure 9 Renal Histological Assessment C57BL/6 mice, C57BL/6. Apcs −/− mice, and C57BL/6.129(D1Mit105–223) congenic mice were sacrificed at 1 y of age to obtain age-matched autopsy specimens. Bouin's fixed kidney sections were scored blinded for glomerulonephritis . Glomerulonephritis was graded on a 0–3 scale (see Materials and Methods for details). Discussion There is accumulating evidence that background genes may influence the expression of autoimmunity in gene-targeted mice. Here we report what is to our knowledge the first systematic study that has examined this in the 129 and C57BL/6 mouse strains, widely used for gene targeting. Our results demonstrate interacting loci between 129 and C57BL/6 mice that can cause the expression of a powerful autoimmune phenotype in these animals, in the absence of any gene-targeted mutations. We also developed a congenic mouse strain bearing a portion of 129 Chromosome 1 on a C57BL/6 background and showed that this wild-type congenic line expressed striking anti-nuclear autoimmunity. By comparing this Chromosome 1 congenic strain with matched congenic mice lacking the Apcs gene, we demonstrated that serum amyloid P component deficiency influences the severity of glomerulonephritis, but is not the prime mover in the induction of anti-nuclear autoimmunity, contrary to our own original interpretation of our data ( Bickerstaff et al. 1999 ). The same consideration applies to other genes located in the same Chromosome 1 region that have been implicated in the development of SLE when inactivated by gene-targeting in 129 embryonic stem cells and then backcrossed onto a pure genetic background ( Bolland and Ravetch 2000 ; Miwa et al. 2002 ; Wu et al. 2002 ). For each, there has to be a question as to whether the anti-nuclear autoimmunity is due to the gene-targeted mutant gene or to the normal 129 genes expressed in the same region as the targeted gene. The influence of background genes on the development of spontaneous autoimmune disease is well known, especially with respect to the lpr and Yaa disease-susceptibility genes. In MRL/Mp mice, the presence of the lpr gene accelerates the development of high level and broad-spectrum auto-Ab production and lethal glomerulonephritis, in addition to marked lymphoproliferative disease. In contrast, homozygosity of the lpr gene in other strains such as C57BL/6, AKR, LG/J, and C3H leads only to auto-Ab production ( Izui et al. 1984 ). Similarly, the Y-chromosome-linked Yaa gene in BXSB and MRL/Mp males enhances the rapid development of auto-Abs and glomerulonephritis ( Izui et al. 1988 ; Merino et al. 1989 ). However, in the C57BL/6 background, the Yaa gene does not lead to an autoimmune phenotype ( Izui et al. 1988 ). Not surprisingly, important effects of the genetic background on the expression of autoimmunity have also been reported in gene-targeted mice ( Bolland and Ravetch 2000 ; Santiago-Raber et al. 2001 ; Mitchell et al. 2002 ). Thus, SLE exists as a complex-trait disorder in which specific combinations of susceptibility alleles are required for the expression of the full phenotype. Through the use of microsatellite marker maps, the identification of murine SLE susceptibility intervals in experimental crosses has been made possible. These mapping studies have shown that the disease expression in relation to the inheritance of the different alleles followed a threshold liability pattern in which a positive phenotype depended upon the presence of multiple discrete susceptibility loci with no single locus being a prerequisite factor. We have employed the same approach to analyse the genetic basis of disease inheritance in the (129 × C57BL/6) hybrid strain, the most common genetic background in gene-targeted mice. Although spontaneous autoimmunity has not been documented in either of the pure 129 or C57BL/6 strains, a spontaneous lupus-like phenotype has been described in (129 × C57BL/6) hybrid strains ( Obata et al. 1979 ; Botto et al. 1998 ; Bickerstaff et al. 1999 ; Santiago-Raber et al. 2001 ), suggesting that the predisposition in these hybrid mice may arise as a result of the interaction between specific combinations of alleles inherited from both the 129 and C57BL/6 parental strains. This was confirmed by the mapping study reported here. We showed that there are multiple genetic loci contributing to the disease and these are derived from both 129 and C57BL/6 mice. We demonstrated that a 129-derived segment of Chromosome 1 was strongly linked to the expression of auto-Abs. This region is probably capable of causing the initiation of a humoral autoimmune response to nuclear antigens; however, this response does not occur in the absence of C57BL/6 genes. In support of this, we identified a C57BL/6 segment on Chromosome 3, which may interact with the 129 genes on Chromosome 1 to mediate the loss of tolerance. Interestingly, although the C57BL/6 SLE-susceptibility region on Chromosome 3 is novel, disease-modifying alleles derived from C57BL/10 and C57BL/6 strains have been mapped to a portion of Chromosome 3 close to the region identified in this study ( Morel et al. 1999 ; Haywood et al. 2000 ). Furthermore, the region on Chromosome 7 associated with the development of lupus nephritis has been linked to the same trait in other murine models of SLE ( Santiago et al. 1998 ; Morel et al. 1999 ; Xie et al. 2002 ), suggesting the possibility of shared susceptibility loci. Taken together our results suggest a complex genetic contribution from the (129 × C57BL/6) hybrid background genome, with both enhancing as well as inhibitory loci from the 129 mouse, in addition to genes promoting autoimmunity from the C57BL/6 mice. The impact that these interacting loci may have on the lupus-like disease present in several gene-targeted animals was further assessed by comparing Apcs −/− mice with Chromosome 1 genetically matched controls. In the context of SLE susceptibility, one of the most consistently mapped non-MHC regions of the mouse genome is the telomeric Chromosome 1 segment, where several disease loci, designated Sle1 ( Morel et al. 2001 ), Nba2 ( Drake et al. 1995 ; Rozzo et al. 1996 ; Vyse et al. 1997 ), and Bxs3 ( Hogarth et al. 1998 ), have been mapped in lupus-prone strains. Moreover, this region of mouse Chromosome 1 is orthologous to a region on human Chromosome 1q22–1q25, which has also been linked with human SLE ( Moser et al. 1998 ). The Apcs gene is one of the candidate genes known to lie within this region. The human serum amyloid P component binds avidly to DNA, chromatin, and apoptotic cells in physiological conditions in vitro ( Pepys 1974 ; Pepys and Butler 1987 ; Butler et al. 1990 ) and also to exposed chromatin and apoptotic cells in vivo ( Hintner et al. 1988 ; Breathnach et al. 1989 ; Familian et al. 2001 ). We have previously reported that (129 × C57BL/6). Apcs −/− mice spontaneously produce a wide range of ANAs and develop significant immune complex glomerulonephritis ( Bickerstaff et al. 1999 ). On the basis of these observations, it was postulated that Apcs , by altering the clearance of chromatin, contributes to the pathogenesis of SLE. However, in this study we found that only ANA and anti-chromatin Ab levels were significantly increased in the (129 × C57BL/6)F2. Apcs −/− mice. A possible explanation for this discrepancy may lie in the fact that the Apcs −/− mice analysed in the original study were generated from a limited number of founders and that this may have caused a nonrandom inheritance of the loci from the parental strains. Furthermore, the whole-genome analysis identified the 129 region surrounding the Apcs gene as the main locus contributing to the development of ANA and anti-chromatin Ab. In agreement with this, when we carried out a selective comparison between the (129 × C57BL/6)F2. Apcs −/− mice and Chromosome 1 genetically matched controls, we failed to detect any significant difference in the levels of these two auto-Abs. These findings, taken together, indicated that the phenotype associated with Apcs deficiency was caused by the presence of unaltered 129 genes from the telomeric region of Chromosome 1 operating in the C57BL/6 genomic background. Strong supportive evidence for this was provided by the analysis of the C57BL/6 mice congenic for this 129 region. The generation and analysis of congenic strains have successfully been used to dissect the contribution of individual susceptibility alleles to a multigenic trait such as SLE. We adopted the same strategy to investigate the relative contribution of the 129 Chromosome 1 segment and the Apcs gene to each disease trait. Using this approach, we demonstrated that the 129 interval on distal Chromosome 1, when transferred onto the C57BL/6 genome, a combination commonly created by backcrossing onto C57BL/6 a mutated gene located in that chromosomal region, was sufficient to mediate the production of auto-Abs. In this context, it is of note that several strains of mice with targeted mutations of genes encoded in this region have been reported to express a lupus-like illness, including mice lacking FcγRIIB ( Bolland and Ravetch 2000 ), complement receptors (CR1/2) ( Prodeus et al. 1998 ; Wu et al. 2002 ), and decay-accelerating factor (CD55) ( Miwa et al. 2002 ). In each case, the autoimmune phenotype was described in mice in which the null mutation was generated in 129 embryonic stem cells and then backcrossed to the C57BL/6 or another genetic background. Thus, in view of our findings, one may postulate that in each of these murine models of SLE, the effects of the targeted null gene were irrelevant. Similar conclusions may apply to other gene-targeted animals carrying mutations of genes mapped in the 129-derived susceptibility allele on Chromosome 7 ( O'Keefe et al. 1996 , 1999 ). The expression of anti-nuclear autoimmunity was identical comparing the congenic with the Apcs −/− mice. The only difference in phenotype between these mice was in the expression of glomerulonephritis, which was more pronounced in the Apcs −/− mice compared with the congenic mice. Although these findings demonstrate that Apcs is not implicated in the processing of autoantigens, as it had previously been suggested, they indicate that Apcs might still play an important protective role in lupus nephritis. In support of this, the expression of the human C-reactive protein, an acute-phase protein closely related to Apcs , has been shown to delay the onset and severity of lupus nephritis in the NZB/W strain by preventing the deposition of immune complexes in the renal cortex ( Szalai et al. 2003 ). Consistent with this, a polymorphism associated with reduced basal level of C-reactive protein has been reported to be linked to SLE in humans ( Russell et al. 2004 ). However, as the congenic mice and the Apcs −/− mice carried similar but not identical 129 regions on Chromosome 1, an alternative explanation for our findings may still lay in the numerous and complex synergistic and counteractive interactions between 129 and C57BL/6 genes involved in self-tolerance and end organ damage. Thus, whilst the lack of lupus nephritis in the congenic mice is consistent with the need for multiple susceptibility genes for the full expression of lupus, further studies will be required to fully elucidate the role of Apc s in the pathogenesis of renal damage. In summary, our findings demonstrate the impact of epistatic interactions between 129 and C57BL/6 genomes on the development of SLE and illustrate how these background gene effects may lead to incorrect interpretations when analysing the autoimmune phenotype of specific genetic disruptions. Materials and Methods Mice. All the mice were females. Wild-type C57BL/6 and 129/Sv (129S6, according to the revised nomenclature) were bred and maintained in the animal care facility at Imperial College, London, United Kingdom. (129 × C57BL/6)F1 mice were generated by intercrossing the two wild-type strains and (129 × C57BL/6)F2 mice by interbreeding the (129 × C57BL/6)F1 mice. The Apcs −/− mice were generated as previously reported ( Botto et al. 1997 ), and the (129 × C57BL/6)F2. Apcs −/− mice were generated by intercrossing Apcs −/− mice on the 129 genetic background with Apcs −/− animals backcrossed onto C57BL/6 for ten generations. A total of 141 (129 × C57BL/6)F2 and 158 (129 × C57BL/6)F2. Apcs −/− female mice were produced and monitored for 1 y. Wild-type congenic C57BL/6.129(D1Mit105–223) mice were generated by backcrossing the 129 interval between microsatellites D1Mit105 and D1Mit223 (80 cM to 106 cM) onto the C57BL/6 strain. Inherited 129 regions were mapped with microsatellite markers polymorphic between 129 and C57BL/6 mice (see below). After seven generations of backcrossing, siblings were intercrossed to generate C57BL/6.129(D1Mit105–223) congenic mice homozygous for the 129 Chromosome 1 interval. Inside 129 markers at positions 81.6 cM (D1Mit159) and 105 cM (D1Mit17), respectively, and an outside C57BL/6 marker at position 74.3cM (D1Mit159) were used to further define the interval. In the C57BL/6. Apcs −/− mice (backcrossed onto C57BL/6 for ten generations), the 129 genome around the Apcs locus was mapped as a segment from 87.9 cM (D1Mit15) to 105 cM (D1Mit17). In this analysis, the inside 129 markers were at positions 93 cM (D1Mit36) and 99.7 cM (D1Mit115) and the outside C57BL/6 markers at positions 81.6 cM (D1Mit159) and 106 cM (D1Mit223). Along with 28 C57BL/6. Apcs −/− mice and 30 C57BL/6 wild-type animals, 26 C57BL/6.129(D1Mit105–223) female mice −/− were followed up to 1 y of age. Animals were maintained in specific pathogen-free conditions. All animal procedures were in accordance with institutional guidelines. Serological analyses. Sera, collected at 6 and 12 mo of age, were assayed for the presence of auto-Abs. Levels of IgG ANA were sought by indirect immunofluorescence using Hep-2 cells, and anti-dsDNA Abs were detected by indirect immunofluorescence on Crithidia luciliae as previously described ( Mitchell et al. 2002 ). Serum samples were screened at a 1:80 (ANA) or 1:20 (anti-dsDNA) dilution and the positive samples titrated to endpoint. Abs to ssDNA and anti-chromatin were measured by ELISA, as previously described ( Mitchell et al. 2002 ). Samples were screened at a 1:100 dilution, and the results were expressed in arbitrary ELISA units (AEUs) relative to a standard positive sample (derived from an MRL/Mp. lpr/lpr mouse), which was assigned a value of 100. For interplate comparison, serial dilutions of a positive control serum sample were included on each plate. Apcs levels were assessed by ELISA using sheep anti-mouse Apcs and rabbit anti-mouse Apcs Abs (Calbiochem, Nottigham, United Kingdom). Samples were screened at a 1:3,000 dilution, and the results were expressed in milligrams per liters, referring to a standard curve derived from an acute-phase serum with a known concentration of Apcs (Calbiochem). Apcs −/− mouse serum was included as a negative control. Histological analysis. All the mice, except the few that died before the end of the experiment, were sacrificed at 1 y of age, and kidney portions were fixed in Bouin's solution and paraffin embedded, and sections were stained with periodic acid–Schiff reagent. Glomerular histology was graded in a blinded fashion as follows: grade 0, normal; grade 1, hypercellularity involving greater than 50% of the glomerular tuft in 25%–50% of glomeruli; grade 2, hypercellularity involving greater than 50% of the glomerular tuft in 50%–75% of glomeruli; grade 3, glomerular hypercellularity in greater than 75% of glomeruli or crescents in greater than 25% of glomeruli. Statistical analysis Non-parametric data are expressed as median with range of values in parentheses. All statistics were calculated using GraphPad Prism TM version 3.0 for Windows (GraphPad Software, San Diego, California, United States). Non-parametric tests were applied throughout, with differences being considered significant for p values < 0.05. The Mann–Whitney test was used for comparison of two groups, whilst for analysis of three groups the Kruskal–Wallis test with Dunn's multiple comparison test was used. Genotypic analysis Genotyping was carried out by PCR of genomic DNA using 143 polymorphic markers (list available on request) distributed throughout all 19 autosomes. PCRs were performed using standard reagents containing 1.5 mM MgCl 2 and 0.4 μM primers. Microsatellite markers were screened for size polymorphisms between 129 and C57BL/6 mice. Only primers with differences detectable on ethidium bromide-stained agarose gels or on SDS-polyacrylamide gels were used. Linkage analysis The QTL program MAPMANAGER.QTL ( ftp://mcbio.med.buffalo.edu/pub/MapMgr/ ) was used, and the two experimental groups were examined independently. Only data from mice at 12 mo of age were analysed. Log transformations of auto-Abs levels resulted in more normalised distribution and were used in QTL mapping. LOD thresholds for suggestive and significant linkages were determined by using a cohort- and trait-specific permutation test (1,000 permutations). The average threshold for suggestive, significant, and highly significant linkages were LOD ≥ 2.1 ( p ≤ 7.8 × 10 −3 ), LOD ≥ 3.6 ( p ≤ 2.4 × 10 −4 ), and LOD ≥ 5 ( p ≤ 1 × 10 −5 ), respectively ( Manly and Olson 1999 ). Supporting Information Accession Numbers The LocusLink ( http://www.ncbi.nlm.nih.gov/LocusLink/ ) ID numbers for the genes and gene products discussed in this paper are Apcs (LocusLink ID 20219), CD35/CD21 (LocusLink ID 12902), CD55 (LocusLink ID 13136), the FAS receptor gene (LocusLink ID 14102), and FcγRII (LocusLink ID 14130). Figure 2 Linkage of Chromosome 8 with Anti-Chromatin and Anti-dsDNA Abs in (129 × C57BL/6)F2 Mice These linkages were not detected in (129 × C57BL/6)F2 .Apcs −/− animals. Centimorgan positions were deduced by interval mapping, anchoring marker locations to data from http://www.informatics.jax.org . Dotted lines indicate the threshold over which linkage was considered suggestive, as defined in Materials and Methods.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509305.xml
532387
In Silico Reconstitution of Listeria Propulsion Exhibits Nano-Saltation
To understand how the actin-polymerization-mediated movements in cells emerge from myriad individual protein–protein interactions, we developed a computational model of Listeria monocytogenes propulsion that explicitly simulates a large number of monomer-scale biochemical and mechanical interactions. The literature on actin networks and L. monocytogenes motility provides the foundation for a realistic mathematical/computer simulation, because most of the key rate constants governing actin network dynamics have been measured. We use a cluster of 80 Linux processors and our own suite of simulation and analysis software to characterize salient features of bacterial motion. Our “in silico reconstitution” produces qualitatively realistic bacterial motion with regard to speed and persistence of motion and actin tail morphology. The model also produces smaller scale emergent behavior; we demonstrate how the observed nano-saltatory motion of L. monocytogenes, in which runs punctuate pauses, can emerge from a cooperative binding and breaking of attachments between actin filaments and the bacterium. We describe our modeling methodology in detail, as it is likely to be useful for understanding any subcellular system in which the dynamics of many simple interactions lead to complex emergent behavior, e.g., lamellipodia and filopodia extension, cellular organization, and cytokinesis.
Introduction Cellular processes generally involve interactions among 10 1 to 10 5 gene products. These interactions can be both biochemical, as in the activation of one protein by another, and mechanical, as in the application of force between bodies. Even when each individual interaction is simple and understood in detail, neither intuition nor qualitative description can forecast the emergent behavior of the whole system. We describe a methodology to characterize such emergent behavior using a detailed computer simulation of both biochemical kinetics and mechanical dynamics. In this paper, we apply the technique to the motility of the bacteria Listeria monocytogenes, a well-studied system in which actin network growth produces a force that moves the bacterium inside of cells. We discuss the model design, compare behaviors of the computational and biological systems, use the model to explain observed features of the bacterial motion, and identify observable experimental correlates of our hypotheses through which our interpretations may be confirmed or rejected. L. monocytogenes is a pathogenic rod-shaped bacterium that invades cells, reproduces, and spreads to neighboring cells, never exposing itself to the extracellular environment, thus avoiding a humoral immune response ( Tilney and Portnoy 1989 ). By expressing the protein ActA ( Domann et al. 1992 ; Kocks et al. 1992 , 1995 ), L. monocytogenes bypasses the host cell's normal controls on actin network growth to produce a dense “comet tail” of actin. This actin tail generates a ram force, by rectifying thermal motion, to both propel the bacterium within a cell and push the bacterium into neighboring cells through distension of the cell plasma membranes. Among experimental advances thus far made to understand this motile system are identification of the purified proteins required to reconstitute motion in vitro ( Loisel et al. 1999 ), an ability to mimic this motion using polystyrene beads coated with the bacterial ActA protein ( Cameron et al. 1999 , 2001 , 2004 ), and experiments that have revealed a discrete step-like motion on the nanometer scale ( Kuo and McGrath 2000 ; McGrath et al. 2003 ). A series of complementary theoretical models have been proposed to account for some observed features of bacterial and bead motion ( Peskin et al. 1993 ; Mogilner and Oster 1996 , 2003 ; Gerbal et al. 2000a , 2000b ; van Oudenaarden and Theriot 2000 ). These studies, taken together, show that L. monocytogenes' actin structures, first described by Tilney and Portnoy (1989) , are created from the same protein components and perform a function similar to the actin machinery in the lamellipodia of motile cells. The dendritic nucleation model for actin network growth ( Mullins et al. 1998 ; Pollard et al. 2000 , 2001 ; Pollard 2003 ) offers a qualitative description of this biochemical network. In the absence of the bacterium, specific signals activate WASP/Scar proteins, and these in turn activate the Arp2/3 protein complex to provide new filamentous actin (F-actin) nucleation sites at or near the barbed (plus) end of existing filaments ( Higgs and Pollard 2001 ). These new filaments form at a characteristic 70° angle to the parent filament, creating dense, highly branched networks ( Mullins et al. 1998 ). Filament barbed ends are rapidly capped with high affinity by capping protein, making the creation/maintenance of free barbed ends critical for continued network growth. Hydrolysis of the ATP that was bound to each actin monomer favors filament disassembly, returning actin monomers to the pool of polymerization-ready G-actin. Cofilin aids in this disassembly by fragmenting F-actin, binding with much higher affinity to ADP actin than to ATP or ADP-Pi actin. The motion of L. monocytogenes exploits all of these actin network features, except that the bacteria's ActA replaces the host cell's WASP/Scar proteins and all the associated upstream signaling mechanisms that normally activate WASP/Scar to control actin polymerization ( Welch et al. 1998 ; Zalevsky et al. 2001 ). Our model differs in several ways from previous attempts to generate mathematical or physical models for L. monocytogenes motility (though see Carlsson 2001 , 2003 ). We simulate explicitly a large number of detailed interactions of both a biochemical and mechanical nature, representing all protein–protein binding interactions with on-rate and off-rate kinetic equations. The simulation of actin filament polymerization, for example, depends on the local concentration of actin monomers and the association and disassociation rate constants (which have been experimentally determined), modulated by the steric accessibility of free barbed ends. Together these factors determine the binding/dissociation probabilities for each filament at each simulation time-step. Bulk properities of our actin “gel” arise from the contributions of the many individual parts of the actin network. Our model can thus accomodate arbitrary geometries, explicit stochastic input, and specific small-scale events. Mechanical interactions, which resolve collisions and accommodate the stretching of protein–protein linkages, follow Newton's laws. We can represent any particular interaction in as coarse or detailed a fashion as desired, subject to the availability of computer resources, and each of these can be based either on experimental information or on simple postulates. We can determine the emergent behavior of the system, which is the dynamical outcome of all the particular interactions, only by running the computer program for many hours or days. In such a model it is neither possible nor desirable to include all details. If our model fails to characterize experimentally observed behavior, then something is missing. If our model does capture an emergent behavior, however, then we can study how quantitative changes in the underlying details (e.g., protein concentrations or specific rate constants) affect this larger scale behavior. The exploration of putative mechanisms is also straightforward, as it is easy to add, remove, or modify each individual interaction. With our approach, we formalize experimentally based models of specific protein–protein interactions and biochemical kinetics in a direct and flexible way, but there are drawbacks. The theoretical approaches used to analyze the Brownian ratchet model and its refinements ( Peskin et al. 1993 ; Mogilner and Oster 1996 , 2003 ) facilitate the derivation of equations that describe important system characteristics, such as force–velocity curves. No such equations are available in our stochastic, individual molecule-based model; instead, we must distill parametric relationships from ensembles of many repeated simulations. Completing these parametric studies in reasonable human time requires considerable computer resources. The biochemical and mechanical interactions near the bacterial surface are stochastic processes involving hundreds of filaments. We model dynamic processes on a per filament basis, rather than through bulk network properties and average filament growth. The growth of any particular filament depends upon that filament's precise location, orientation, and biochemical state, all of which change through time. There is no better way to simulate such a system than with a model that tracks each of these variables for each individual filament. In the future, this type of detail will be essential to capture (and thus explain) many observed biological phenomena. The trajectories generated by this model of L. monocytogenes motility display repeated runs and pauses that closely resemble the actual nanoscale measurements of bacterial motion ( Kuo and McGrath 2000 ; McGrath et al. 2003 ). Further analysis of the simulation state at the beginning and ends of simulated pauses suggest a new interpretation of the experimental results. We show that there is no characteristic step-size or pause duration in these simulated trajectories and that pauses can be caused by both correlated Brownian motion and by synchronously-strained sets of ActA–actin filament mechanical links. The Model We model the molecular mechanics of the growth/disassembly of an actin network as it interacts with a moving rod-shaped bacterium to whose surface many ActA proteins are bound at specific locations. We distinguish molecular mechanics ( Howard 2001 ) from molecular dynamics: we are not concerned with van der Waals forces and hydrogen bonds or with conformational changes during protein–protein interactions. Our model is different from a purely continuum model, in which state variables (those dependent variables which, together, fully describe the state of a system) would characterize only the bulk properties of an actin dense tail, using average compliance and polymerization values. We instead separate the cellular world into two basic classes of entities, those that are relatively large and present in small numbers (e.g., actin filaments, a bacterium) and those that are very numerous and small (e.g., actin monomers and other diffusible proteins). We simulate the former entities, which we call “explicit players,” as individuals; our state variables keep track of the position, orientation, and biochemical state of each individual and its change with time according to appropriate physical laws (e.g., Newtonian force balance laws). Those entities that are more numerous we will call “implicit players”: we represent them with continuum field state variables, i.e., molar concentrations that vary with time and place. We use standard partial differential reaction–diffusion conservation equations to express how these continuum fields change with time as the implicit players interact with each other and with the explicit (individual) players ( Figure 1 A). Dataset S1 is a simplified psuedo-code of our simulation software. Figure 1 Model Schematic (A) shows a simple cartoon of the bacterium and some actin filaments (explicit players) against a backdrop of a moving diffusion–reaction grid attached to the bacterium. We use this spherical coordinate grid, whose origin moves with the bacterium, to keep track of the diffusion and biochemical interactions of the scalar protein concentrations fields that characterize implicit players. Protein size greatly impacts diffusion of proteins in a cellular environment ( Luby-Phelps 1994 , 2000 ), so we modify the nominal diffusion coefficients of implicit players accordingly. The dotted line is the path trajectory in 3D space of the bacterium. (B) illustrates the origin of forces that act on a bacterium in our simulation. Actin filament α is bound to an ActA protein on the bacterial surface, generating a link force that acts to hold the two objects together. Filament β is shown colliding with the bacterium, generating a collision force that acts to push the two objects apart. The barbed end of filament γ is nominally too close to bacterial surface to allow addition of an actin monomer, but Brownian motion might bend the filament into the dotted configuration, thus allowing polymerization and creating a collision. We model our filaments as rigid rods, but simulate this filament flexibility with a polymerization probability function that approaches zero with the filament–bacterium gap; i.e., it is possible for a filament to add a monomer even if the nominal filament–bacterium gap is less than 2.7 nm (the filament length increase per actin monomer). The bacterium, actin filaments, and structures of actin filaments are all subject to Brownian simulation forces and torques, represented by δ. Various different forces impinge on the simulated bacterium. Forces move two objects apart if they happen to collide at the end of a time-step. Likewise, elastic bonds linking two objects (e.g., an actin filament–ActA bond) exert equal and opposite forces that hold those objects together; these links break under sufficient strain. Forces of random orientation act on every explicit player to simulate Brownian motion (i.e., the sum of all the many collisions with small molecules that, in biological reality, contribute to the Brownian motion is represented in our model by a single vector force and a single vector torque). This system never approaches an equilibrium; Brownian motion and biochemical events ceaselessly create collisions and perturb protein–protein links. Thus, we must compute new forces, exchanged between new neighbors, in each time-step. Figure 1 B illustrates the set of mechanisms that combine to generate the net force on the bacterium in our simulation. At the heart of this simulation is the dendritic nucleation model of actin dynamics ( Mullins et al. 1998 ; Pollard 2003 ). Asynchronously, each individual actin filament can grow or shrink at either end by actin monomer polymerization/depolymerization; hydrolyze the ATP bound to one or more of its actin monomers to ADP-Pi; dissociate the Pi from one or more such monomers; be severed by ADF/cofilin; bind Arp2/3 to an ATP-actin subunit in the filament; be capped or uncapped at either end; and nucleate new filaments through Arp2/3 initiated side-branches. Repeated nucleation of new branches from existing filaments leads to a dense meshwork of actin in the comet-tail. Besides Arp2/3 mediated branching, all other cross-linking and adhesions involving actin filaments are implicit in the age and length dependent anchoring of f-actin in our simulation space. All actin filaments accumulate adhesions that gradually increase drag coefficients and eventually lock each filament into a fixed position. Figure 2 , a video frame from a typical simulation, introduces some of the explicit and implicit players. Figure 2 A Simulation Video Frame Showing Actin Filament Branching near the Bacterial Surface Arp2/3 seeds branches off of existing filaments at a characteristic 70° angle. Different ATP/ADP forms of actin have differing affinities for proteins such as ADF/Cofilin. Filament barbed tips can be capped unless they are protected through interaction with an ActA molecule; as indicated, surface-bound ActA molecules elastically link the bacterium to an actin filament near the barbed end. The simulation time-step has a subtle effect on the simulation of Brownian motion for constrained objects (that is, objects linked to other objects). Applying the same forces and torques that are appropriate for free objects exaggerates the simulated Brownian motion of constrained objects since the motion restriction that results from those constraints can only be resolved over several time-steps, and those time-steps are large relative to the intrinsic time-scale of the constraints. Experimental measurements ( Kuo and McGrath 2000 ) show very little Brownian motion (relative to similarly sized nearby vesicles) of L. monocytogenes associating with their actin tails; to match the biological reality, we need to modify our simulation of Brownian motion, since we cannot yet use much smaller time-steps. We compensate for this technical problem by carrying out simulations both for the two extreme cases (with Brownian motion appropriate for a free bacterium and with no Brownian motion of the bacterium at all) and for an intermediate degree of Brownian motion. Advances in computer processing speeds will, most likely, make such attenuation unnecessary in the near future. We will henceforth use the term “Brownian simulation force” to refer to the forces and torques that we apply to the bacterium to simulate its Brownian motion. For our model, we use typical physiological concentrations for each of the proteins involved; these are listed in Table 1 . Table 2 summarizes the reaction rate constants we used. Some crucial parameters and protein functions are as yet incompletely known. These include the exact pathways and rate constants associated with the stimulation of local actin filament polymerization by the bacterial surface-bound ActA protein. This protein has binding sites for a host of proteins, including G-actin, F-actin, Arp2/3, and Ena/VASP. In addition, while the affinity between free ActA and Arp2/3 has been measured ( Zalevsky et al. 2001 ), that value (K D = 0.6 μM) does not sufficiently characterize that interaction since the interaction rates may be limited by the flux of Arp2/3 or G-actin onto the surface. We have calculated the flux of Arp2/3 and G-actin onto the bacterium's surface to determine the expected equilibrium number of ActA–Arp2/3 and ActA–actin complexes there, as explained in Dataset S2 . We tune these approximate rate constants to create actin networks with biologically representative side-branch separation and filament numbers. Because the rate constants that we have obtained in this way will depend on the concentrations of ActA, Arp2/3, and actin monomers in the model, the rates given for ActA–Arp2/3 and ActA–actin interactions in Table 2 apply to the concentrations given in Table 1 . Table 1 Values and References for the Concentrations of Proteins We Used These values are not for any specific cell type, but are typical biological concentrations and similar to those used for in vitro reconstitutions of bacterial motility Table 2 The Biochemical Rate Constants Incorporated in Our Model A hydrolysis rate is given for a vectorial ATP hydrolysis model; experimental evidence currently supports the random hydrolysis model but we have, for simplicity, implemented a vectorial scheme for this analysis. That is, we assume that there is a distinct border within each filament between the ATP actin, ADP-Pi actin, and ADP actin regions; only monomers adjacent to a these borders can transition from ATP actin to ADP-Pi actin or from ADP-Pi actin to ADP actin. We can readily switch to a random hydrolysis model in future studies. The values in angle brackets, for the interactions between ActA, Arp2/3, and actin monomers, are calculated considering the diffusive flux onto the bacterium's surface (see Dataset S2). These values are thus dependent upon ActA density, the concentrations of Arp2/3 and actin, and a heuristic adjustment of these rates to balance new filament nucleation and side-branching in order to achieve realistic tail morphologies. The on-rates in brackets listed here apply to the concentrations in Table 1 The ActA protein is distributed asymmetrically on the bacterial surface, with more ActA near the rearward tail-forming end. The unipolar shape of our distribution is based on measurements of the fluorescence signal from RFP-labeled ActA along the bacterial length for newly divided bacteria (S. Rafelski and J. A. Theriot, unpublished data). New filaments are produced by two pathways. By activating Arp2/3, ActA is thought to catalyze the creation of new actin filaments and side-branches. We simulate the co-binding of ActA, Arp2/3, and an existing actin filament, allowing binding in any order. This complex leads to a new side-branch on the existing filament. Binding of ActA to the actin filament can occur only at specific ATP or ADP-Pi actin sites and is binding site limited, meaning that each bound ActA occludes a linear region of five monomers on the filament from further binding. Creation of a new filament de novo in our model involves the co-binding by ActA of G-actin and Arp2/3, in any order ( Boujemaa-Paterski 2001 ). In conjunction with other proteins (e.g., Ena/VASP), ActA may also regulate actin dynamics in other important ways ( Goldberg 2001 ). In this version of our model, we do not explicitly simulate Ena/VASP molecules, which can regulate actin networks by binding profilin, competing with capping protein, and regulating Arp2/3 spacing ( Krause et al. 2003 ). Instead we assume that ActA itself can uncap any actin filament barbed end to which it binds closely (within one ActA length) and ignore the other possible Ena/VASP functions. We find that this uncapping function is necessary to obtain persistent motion with a low nucleation rate of new filaments. For the values in Tables 1 and 2 , our simulated bacterium do move without this uncapping, but more slowly. Loisel et al. (1999) have reported a similar dependence in their experiments, finding that Ena/VASP is not required for bacterial motion, but that it improves the efficiency of the motion observed. Results The gross behavior of our simulated bacterium is life-like; model bacteria move in a qualitatively similar way to wild-type L. monocytogenes . Average speeds of motion varied from ten to hundreds of nanometers per second, as do real observed bacterial speeds in different experiments (40 nm/s in purified proteins in Loisel et al. 1999 ; 140 nm/s in cytoplasmic extracts in Cameron et al. 2004 ; 1.4 μm/s in vivo in Dabiri et al. 1990 ). Videos at any scale may be rendered from our simulations (Videos S1 and S2 ; other full-length simulations at www.celldynamics.org ). Figure 3 merges several frames from one of those movies, showing the large-scale formation, hydrolysis, and depolymerization of the actin comet-tail. Figure 3 A Simulation Time Series Several video frames from one simulation show, on a gross scale, the de novo ActA filament nucleation and Arp2/3 branching from existing filaments that form a comet-tail, the hydrolysis of actin filaments in the comet-tail, and the subsequent pointed end depolymerization of ADP-actin (filament severing by ADF/Cofilin is not active in this simulation). The microrheology experiments of Kuo and McGrath (2000) and McGrath et al. (2003) present an opportunity to illustrate the utility of our stochastic model founded on small-scale details. They reported that L. monocytogenes motility is multiphasic; motion of the bacterium that appears smooth on the micrometer length-scale actually consists of pauses that last many milliseconds, discrete nanometer-scale steps, as well as uninterrupted runs. No current model of this bacterial motility has fully explained these discrete steps, though numerical simulations with the tethered ratchet ( Mogilner and Oster 2003 ) can exhibit saltatory motion. Possible explanations involving strained links between ActA and actin filaments and nucleotide dependent filament templates are discussed in McGrath et al. (2003) . Figure 4 shows the distribution of step sizes and pause durations that our model produces (using the values of Tables 1 and 2 and one set of mechanistic hypotheses) with three different assumptions for the Brownian motion of the bacterium (see under The Model ). In all cases, we find that there is no characteristic step-size, but rather a continuum of steps with the smaller steps being more probable than larger ones. The qualitative shapes of these histograms are insensitive to changes in all parameters we might reasonably vary, barring values that disrupt persistent bacterial motion. The parameters we have varied include link characteristics (e.g., link length, link force, and maximum link strain), Arp2/3 branching rate, and temperature. In fact, even doubling the size of each actin monomer (easily done in a simulation) does not change these histograms significantly (data not shown). Figure 4 Step-Sizes, Pause Durations, and Speeds of Motion with Different Brownian Simulation Force Attenuation The legend shows the multiplier by which the Brownian simulation force on the bacterium is attenuated, such that a multiplier of 1 corresponds to a Brownian simulation force appropriate to an unconstrained bacterium and a multiplier of 0 signifies no simulated Brownian motion of the bacterium (see discussion of the relationship between applied Brownian simulation force and numerical time-step in the model description). We line-fit and filter trajectories by slope (speed) to identify pauses in the bacterial motion. Any line with slope less than the pause threshold might indicate a pause. The distance between adjacent pause locations is the step-size, assembled into a histogram in (A). No characteristic step-size is evident; smaller steps are simply more probable than larger ones. We exclude steps shorter than 0.2 nm. (B) shows a pause duration histogram; here we exclude pauses <8 ms in duration. (C) shows run speed histograms for runs >50 ms in duration. For a Brownian multiplier of 1, we used 30 simulations with 56,239 pause events for (A) and (B) and 6,098 runs for (C) (1,929 s total simulation time, 644 s total paused time, 116 nm/s average speed, pause threshold 40 nm/s). For a multiplier of 0.5, we used 13 simulations with 49,207 pause events and 2,287 runs (1,534 s total simulation time, 700 s total paused time, 93 nm/s average speed, pause threshold 30 nm/s). For a multiplier of 0, we used 18 simulations with 9,748 pause events and 1,179 runs (940 s total simulation time, 612 s total paused time, 56 nm/s average speed, pause threshold 20 nm/s). These results suggest that our simulated bacterium does not move with steps of any prefered size, and specifically not with a step-size related to the actin monomer dimensions. The pause in forward progress might equally be considered the defining event in the bacterium's motion; a “step” in this case is just a run made along the path between adjacent pauses. But what physical process stalls the actin polymerization ram to initiate pauses, and what physical process breaks the bacterium out of each paused state into a run? To answer these questions, we need to examine how key descriptive system features vary before, during, and after pauses in our simulations. We do this by looking both at individual pauses and at the average of these system outcomes for many thousands of pauses (see Materials and Methods ). In Figure 5 , we follow actin polymerization, link formation and breakage, link number, and path-directed forces for several sequential pause events during a single simulation. Owing to the stochasticity introduced by the Brownian simulation forces in this simulation, it is difficult to find trends in such single simulation profiles. We can learn more by turning off the Brownian simulation forces on the bacterium as has been done in figure 6 . Now frequent long pauses are observed that clearly reveal the force relationships during pause initiation and termination. Both filament link force and collision force increase in magnitude synchronously during a pause, indicating resistance to forward progress by a population of ActA–actin filament links. The bacterium moves rapidly forward only when the total filament link force suddenly plummets. This sudden decrease in link force can only be attributed to a cascade of link breakage. This result indicates that the highly strained filament links that had balanced the filament collision force during a pause are rapidly exchanged for unstrained links when a pause ends. Figure 5 System Outcome Profiles for Several Adjacent Pauses Pauses are shown by red horizontal lines in a single simulation run in which the Brownian force multiplier was 1 (unconstrained bacterium). Listed for each pause are the pause duration (δt) and distance to the following pause (δs) as reported by our line-fitted analysis. The vertical line segments in the lower half of the figure show discrete events. From the bottom, these are the number of polymerization events on filaments very close to the bacterium, the number of new links formed between the bacterium and filaments, and the number of these bacterium–filament links broken. Above these are plotted the total number of bacterium–filament links. At the top, the net filament link force and the net filament collision force are given in picoNewtons (see Figure 1 B), along with the total force. Some general trends aligned with pauses are apparent, such as decreased actin and link dynamics during a pause, but any characteristic biochemical or force response is obscured by the Brownian agitation of the bacterium. Figure 6 System Outcome Profiles for Several Adjacent Pauses—No Brownian Motion of the Bacterium Long pauses during this run without simulated Brownian motion of the bacterium (Brownian force multiplier of 0) reveal the force balance/imbalance conditions that cause pause–run behavior. System outcome profiles are shown for several adjacent pauses (red horizontal lines) with their duration (δt) and distance to the following pause (δs) displayed as reported by our line-fitted analysis. For a description of each of the entities plotted here, see the legend to Figure 5 . Note that the bacterium pauses in its forward progress when the filament collision and link forces cancel each other; then, the collision forces tend to rise during a pause as filaments in the tail catch-up with the bacterium and generate new filament collisions. Runs are coincident with the breakage of many highly strained filament links which are quickly replaced by unstrained links; note that the filament link number does not change greatly during such an avalanche of link breakages and that it is steadily climbing over this time range in this particular simulation. The average of these system outcomes further clarifies pause causality and reveals differences between the two cases illustrated in Figures 5 and 6 . From Figure 7 we see that pauses can occur with or without Brownian motion of the bacterium. But when we simulate the Brownian motion of the bacterium, we observe that pauses are correlated with an accidental sequence of similarly directed Brownian simulation forces (forward to initiate a pause, backward to sustain a pause, and again forward to break out of a pause). Note that any individual pause event in our average ensemble may experience only a subset of the correlated Brownian simulation force profile presented in Figure 7 . On that figure we have partitioned this correlated Brownian simulation force into three temporal regions, labelled A, B, and C. An individual pause event might correlate with the Brownian simulation force sequence of A, or only A and C, or only B and C, or any other combination. Additionally, some pause events might be entirely uncorrelated with any Brownian simulation force trend. In other words, this averaging method reveals system trends that occur frequently, but need not be present in every contributing event. Figure 7 Anatomy of a Pause An ensemble of average system outcome analyses, with (solid lines) and without (dotted lines) simulated Brownian motion of the bacterium (see discussion on the relationship between applied Brownian simulation force and numerical time-step in the model description). These averages were compiled from pauses with duration greater than 10 ms, using 10,365 pauses with, and 4,358 pauses without, simulated Brownian motion (there are fewer, longer pauses without Brownian motion of the bacterium). Only sufficiently time-separated pauses contributed to these averages, so that the 10 ms preceding the pause start and the 10 ms following the pause stop are guaranteed not to include the effects of any adjacent pause. The Brownian simulation force trends can be read from the total force curve, which is only slightly offset by the link and collision forces in the case where Brownian movement of the bacterium is simulated (solid lines). Here, initiation of a pause follows a large forward-directed Brownian simulation force on the bacterium (segment A), which increases link turnover and produces a large population of synchronously strained links. Backward path-directed Brownian simulation forces (segment B) maintains the pause until, aided again by forward-directed Brownian simulation forces (segment C), the bacterium transitions back into a run. The Brownian simulation force trends can be read from the total force curve, which is only slightly offset by the link and collision forces. Without simulated Brownian motion of the bacterium (dotted lines), a pause is initiated and maintained when a population of ActA–actin filament links can resist the essentially constant total filament collision force. A pause terminates in this case when these links break en masse. Any individual pause in the averaged set of pauses might not demonstrate all of these response features. We contrast the Brownian/no Brownian motion cases to better understand pause initiation, maintenance, and termination. Our most realistic simulation will incorporate effects from each extreme case. With simulated Brownian motion of the bacterium, we suggest the following causal temporal sequence for pause initiation, maintenance, and termination (with the caveat that most individual pause events will experience only a subset of this sequence): A particularly large Brownian simulation force (or accidentally correlated sequence of forces) in the forward path direction causes an unusually rapid but small forward displacement of the bacterium (region A). The steady-state rate of filament link turnover increases slightly as highly strained links break and are replaced by an ensemble of new links that all form nearly simultaneously in an unstrained state, thus creating a larger than steady-state population of coordinately unstrained links. A particularly large Brownian simulation force, or correlated sequence of forces, opposite to the path direction forces the bacterium backward against the population of linked barbed end actin filaments; filament collision force increases, filament link force falls, and actin polymerization near the surface decreases. A pause ensues. During the pause new filaments form and existing but distant barbed ends “catch up” with the bacterium, thus increasing the filament collision force forward which the links restrain. The pause terminates when a particularly large Brownian simulation force (or an accidentally correlated sequence of forces) in the forward path direction is sufficient to break a few of the strain-synchronized filament links. As these links break, the force stretching each remaining link increases, setting in motion an avalanche of cooperative link breakage and initiating a run. We are justified in interpreting these correlations of Brownian simulation forces as causal because those forces are generated in our simulations so as to be random in direction and magnitude (representing a Gaussian distribution). Nothing in our model can cause such Brownian simulation force “accidents.” Their correlation with pause initiation or termination must therefore be causal. Absent Brownian simulation forces on the bacterium, the system response throughout the course of a pause is very different. In this case, a pause occurs only when a population of synchronized filament links is able to balance the filament collision forces, which on average increase only slightly during a pause, until a cascade of breaking links allows the bacterium to run again. Judging from the shape of the step-size histograms in Figure 4 , the generation of a set of strain-synchronized links that initiate a pause is likely a random event. That figure reveals a Poisson process-like distribution of step-sizes with weak or no Brownian simulation forces; moreover, the step termination (and therefore pause initiation) appears to occur with a constant probability through time. This should be contrasted with the case of simulated Brownian motion appropriate for an unconstrained bacterium, in which step-termination (pause initiation) is correlated with forward path-directed Brownian simulation forces. The small amplitude of experimentally measured fluctuations of bacteria in vivo ( Kuo and McGrath 2000 ) suggest that the simulations absent Brownian motion of the bacterium come closest to representing the biological reality; the coincidence of similarly directed Brownian movements is probably less important than the balance between filament–bacterium collision and link forces. Discussion We have used our model to ask how L. monocytogenes motility is mediated by actin-mediated forces. Building a simulation from basic, well-understood structures and interactions, we have reconstituted bacterial motility in silico. Appropriate speeds and persistence of motion emerge, reproducing experimentally observed values. Additionally, our simulation yields as an emergent behavior the nanometer-scale saltatory motion reported by experimentalists. We can analyze details of the simulated bacterial trajectories to investigate characteristics of this saltation: what is the mechanism behind the stepping, and is there a favored step-size? Our computational experiments lead us to conclude that the tethered-ratchet model is an inherent “pauser” with several important attributes. First, there is no characteristic step-size or pause length; shorter steps and pauses are more frequent than longer ones. Second, the intensity of Brownian agitation of the bacterium influences average pause duration and frequency, but this agitation is not necessary for persistent saltatory motion. Third, pauses start when a population of filament links happen to form nearly simultaneously with low strain to balance filament collision forces. Pauses end when those links catastrophically break. To produce nanometer-scale pauses and runs, no special function need be attributed to the bacterial bound ActA protein, beyond an elastic linkage to actin filaments and some mechanism that prevents barbed end filament capping. Specifically, the ActA protein does not need a motor-protein-like stepping ability, nor need it act as a clamped-filament elongation motor ( Dickinson and Purich 2002 ). Given the large number of filaments near the surface of the bacterium and the wide variation in angle of those filaments, it is not clear that we would expect any step-size, even if ActA were motor-like with a discrete working stroke. The speed of motion during a run is variable, but it is mostly confined to a narrow range of speeds that depends on the parameter set (see Figure 4 C). Pauses are a significant feature in our simulated trajectories; bacteria spend large fractions of their time paused (from 33% to 65% in the simulations presented here). Lastly, we have explored the variation of key biochemical events and mechanical interactions during a typical nanometer-scale saltation, looking at both individual pause events and averages of many such events, in an effort to uncover the causal factors. We conclude that the tethered-Brownian ratchet model is an inherent pauser; forward motion is temporarily halted whenever a population of synchronously strained filament links balances filament collision forces. Different mechanisms cause pause initiation/termination with and without simulated Brownian motion of the bacterium. With simulated Brownian motion of the bacterium, we find that pauses events are largely driven by coordinated Brownian simulation forces: a series of forces in the forward direction helps establish a set of coordinately-strained links, forces in the backward direction can help maintain a pause, and lastly forces in the forward direction help break links to terminate a pause. Without simulated Brownian motion of the bacterium, we find that a coordinately strained set of filament links balances the filament collision forces and that a pause will ensue until those links break en masse. Formation of such a set of filament links is an accidental, but frequent, occurrence, explaining the shape of the step-size and pause duration histograms (see Figure 4 ). That we find no characteristic step-size in our simulated nanoscale stepping constrasts with experimental results ( Kuo and McGrath 2000 ; McGrath et al. 2003 ). Restricted by experimental noise, those researchers cannot see steps smaller than about 2.5 nm, if they indeed exist. Without those steps, about 30% of our simulated steps would be between 4 nm and 6 nm. We are presently sharing trajectories with the Kuo laboratory to directly compare model and experiment. Mogilner and Oster (2003) explore stepping behavior of their elastic tethered Brownian ratchet model and, for low filament tether numbers and particular capping rates and tether stiffness, observe step sizes similar to those reported by Kuo and McGrath ( Kuo and McGrath 2000 ; McGrath et al. 2003 ). A step in that model occurs when one filament tether breaks and the remaining tethers all stretch in response to the new force balance. By constrast, steps in our model occur following a catastrophic breakage of many coordinately strained tethered filaments and are not highly dependent upon tethered filament number or capping rate. Because of the method by which we resolve collisions and strained links (see Figure 8 ), we do not prescribe the elastic properties of the filament–ActA links, and so our stepping is also independent of those values. Dickinson and Purich (2002) proposed a completely different mechanism for nanoscale stepping, involving a putative elongation motor that demonstrates approximately 5.4-nm stepping in simulations. In their model, most filaments are in compression while a few lagging filaments prevent the bacterium from moving forward. It is the “release and relocking” of a single lagging actin filament by the elongation motor that allows a 5.4-nm step. While the mechanisms are severely different, the basic behavior of this elongation motor model is similar to ours. Of the population of filaments interacting with the bacterium in our simulations prior to a step, a subset generates collision forces and a coordinately strained subset attached to ActA proteins balances those collision forces, resisting forward motion. The concurrent breakage of this linked subset allows a step forward, analogous to the “release and relocking” in the elongation motor model, though many more filaments are involved in our “release,” and the step distance before these filaments rebind to ActA, and thus “relock”, is not prescribed. Figure 8 A Simple Collision Rule We calculate the magnitude, F, of equal and opposite forces applied to colliding objects such that they no longer collide after a time-step of δt. This force is calculated by considering the maximal distance, δ, of object intersection and the shape-based viscous drag, γ, for each object. In this example, we use Stokes' law for the viscous drag on spheres. We are assuming that the time-scale for a collision to resolve itself is much shorter than the discrete time-step used in the computation. We make similar calculations for more complex shapes and collisions. The computational analysis of L. monocytogenes motility described here represents a new tool that should be useful for understanding many complex subcellular systems. The construction of this computational model requires experimental measurements of the biological details in L. monocytogenes propulsion and actin dynamics in general. Only in the last several years have crucial biological details come to light, e.g., the role of Arp2/3 in filament branching, or the binding sites and functionality of the ActA protein. Additionally, the implementation of the model in silico requires significant computational power, now affordable in the form of clusters of “off-the-shelf” machines (we estimate use of 30 cpu years on a 2.8 Ghz Pentium 4 in the development and exploration of this model, 3.5 cpu years of which directly contribute to this report). Powerful object-oriented languages are also recently mature (we use Java™), making it possible to write computer code to implement such models. We believe that the confluence of detailed biological information and computational power/software heralds a new approach for understanding subcellular systems in which many thousands of simple biochemical and mechanical interactions lead to complex emergent behavior. The biological systems in which this approach will be useful are, by definition, rich in detail. This complexity favors collaborations between modelers and the experimentalists who discover and quantify the molecular details without which this study would be specious. Creating a simulation environment that makes intuitive sense to experimentalists, i.e., one in which there is clear correspondence between biological entities and their modeled counterparts, greatly facilitates communication between modelers and biologists, and it ensures appropriate refinement of the model as new biological facts are uncovered. There are many future refinements and research directions for this model. We can incorporate a more sophisticated representation of the actin hydrolysis cycle ( Bindschadler et al. 2004 ) and include specifics of the interactions between ActA and proteins such as Ena/VASP. Recent work with ActA-coated beads ( Cameron et al. 2004 ) characterizes the relationships between several biophysical parameters and motion initiation, speed, and persistence; these experimental findings can also be explored in silico. Extended with additional cellular components (e.g., a dynamic cortex, microtubules, motor proteins), the model might also be used to explore any of a number of cellular phenomena, including whole cell motility and cytokinesis. Our model, encoded in an object-oriented manner, is structured in a way that is strongly delimited by nature—so while we must still embrace approximation, we can minimize abstraction. Materials and Methods A large set of differential equations determine how our state variables change with time. We solve these equations numerically, but not in a standard way because discontinuities in time occur frequently as objects collide suddenly and as objects suddenly spring into existence or disappear (due to new filament nucleation and depolymerization). To solve these thousands of differential equations, we divide time into discrete steps (typically tens of microseconds) balancing the necessities of capturing the system dynamics and accomplishing the simulation in reasonable human time (typically 3–5 d). At the beginning of each time-step, the biochemical events and forces experienced during the last time-step will have changed the state of the system. New collisions and link forces may have arisen, as well as new objects. Existing links may break if they experience excessive strain for several consecutive time-steps. Each explicit player thus experiences a net force vector; in the next step, we move each explicit player in this vector direction so as to reduce or eliminate the strain energy associated with its collisions and links. To accomplish this practically, we calculate the forces required to resolve each individual collision (or strained linkage) in a single time-step. Figure 8 demonstrates this calculation for a collision between two spherical bodies; a similar approach is taken for all pair-wise collisions or links. In brief, we sum all forces, attenuating all their magnitudes by the same factor without changing their directions, so that acting during the time-step they produce just enough displacement to separate objects that collided in the prior time-step. This process avoids prescription of elastic constants and is equivalent to proceeding through a series of quasi-static equilibria, a formally valid approach if the biochemical dynamics are slow relative to the resolution of force imbalance. Each individual computer run simulates bacterial motion for a period of up to many minutes. We run hundreds of such simulations and then statistically analyze the ensemble of runs. Fitting straight line segments to each trajectory and filtering those segments by slope (speed of motion) reveals that each simulated bacterial trajectory is composed of a sequence of pauses (of varying duration) separated by near-constant speed runs between pause locations. After we identify all the pauses, we measure the distances between adjacent pauses; these are the putative step-sizes. Histograms of pause duration and step-size, distilled from multiple simulations, then allow comparison with experimental observations and reveal whether there exists a preferred step-size or pause duration. Figure 9 shows a segment of trajectory data and the progressive stages of our line-fitting analysis. Figure 9 Slope-Based Analysis of Bacterial Trajectories A brief sample of a bacterial trajectory from one simulation run (black), a smoothed approximation to that trajectory (blue), line segments of near zero velocity fit to the smoothed trajectory (green), and the summary of a pause event assembled from those lines (red). Our analysis software seeks nearly horizontal segments of the trajectory (i.e., pauses), of maximal possible duration, in which excursions away from pause location lie with a jitter tolerance that we specify. Trajectories are curves in 3D space with curvature and torsion. To simplify the analysis, we use a path position variable (on the vertical axis), projecting each displacement onto the path tangent vector, instantaneously defined by the bacterial orientation. Then, we identify pauses in the resulting time series to specify displacement along the smoothed trajectory. Labels on each pause report pause durations, δt, and the displacement to the next pause (step-size), δs. Random thermal agitation forces act to buffet the bacterium, and every individual part, in our simulations; this is what makes the trajectories jagged. We average many thousands of pause events into portraits characterizing system behavior preceding, during, and following the typical pause. To do this, we align pauses, time and space shifting short sections of the path projected trajectories that span a single pause event so as to superimpose their starting or stopping points ( Figure 10 ). These pauses are of different duration, so our average response will be most meaningful near the alignment point. To improve this analysis, we can also select and average only pauses of similar duration or create ensemble portraits from start-aligned and stop-aligned analyses. Figure 10 Calculating an Average System Response (A) shows a segment of simulated bacterial trajectory with five identified pauses shown in red; δt is the pause duration, and δs is the distance along the path to the next pause. To obtain the average response of any system outcome, we align the pauses at their stopping (or starting) points as shown in (B). Any trend remaining after the averaging of many thousands of events will reveal significant system behavior near the alignment point. Any individual event, however, might not exhibit all the trends revealed in such an average, so that the interpretation of these average profiles should be tempered accordingly. Not all of the capabilities of our model have been enabled in the simulations contributing to this study. Our calculations show that the local depletion of the implicit players, due to their incorporation into a larger assembly, is not significant for the concentrations, rate constants, and geometries of this system (data not shown). Thus, we do not simulate the diffusion of any of the implicit players (proteins), but rather assume that each exists at a constant concentration (see Table 1 ). With this assumption, we need not accurately represent the depolymerization of the bacterium's comet tail in modeling the movement of the bacterium. (This depolymerization could otherwise have had an important role in regenerating depleted stocks of some implicit players.) We therefore depolymerize F-actin in the most computationally efficient way: we assume an artificially high pointed end depolymerization and ignore cleavage by ADF/Cofilin. In addition, F-actin interacts with cellular components in vivo that are not explicitly represented in either the dendritic nucleation model or our simulations (e.g., with other cytoskelet al.filaments). Some of these interactions have the effect of locking down actin tail in cellular space. We approximate their effect with a time- and actin length-dependent application of adhesions that eventually fix F-actin and the actin tail in our simulation space. Supporting Information Dataset S1 Psuedo Code for an Actin Filament (29 KB DOC). Click here for additional data file. Dataset S2 Steady-State Number of ActA–Arp2/3 Complexes on the Bacterium (52 KB DOC). Click here for additional data file. Video S1 One Actin Filament Interacting with the Bacterium A close-up look at the interaction of a single polymerizing filament with the bacterium. This filament has an artificially durable link with an ActA protein on the bacterium's surface; these links are typically very transient. The tip-clearance (drawn with a cyan line), the polymerization probability, the capping probability, and the Arp2/3 binding probability (set to zero for this demonstration) are reported at each simulation time-step. (1 MB MOV). Click here for additional data file. Video S2 An Animated Simulation: Motion Initiation and Persistence An animation rendered from the output of one simulation of L. monocytogenes motility. Microscale hops, as opposed to the nanoscale steps we investigate in this paper, are apparent at this scale view. The bacterium induces an actin tail of variable density and demonstrates persistent motion. (9.8 MB MOV). Click here for additional data file.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC532387.xml
552314
Clustering the annotation space of proteins
Background Current protein clustering methods rely on either sequence or functional similarities between proteins, thereby limiting inferences to one of these areas. Results Here we report a new approach, named CLAN, which clusters proteins according to both annotation and sequence similarity. This approach is extremely fast, clustering the complete SwissProt database within minutes. It is also accurate, recovering consistent protein families agreeing on average in more than 97% with sequence-based protein families from Pfam. Discrepancies between sequence- and annotation-based clusters were scrutinized and the reasons reported. We demonstrate examples for each of these cases, and thoroughly discuss an example of a propagated error in SwissProt: a vacuolar ATPase subunit M9.2 erroneously annotated as vacuolar ATP synthase subunit H. CLAN algorithm is available from the authors and the CLAN database is accessible at Conclusions CLAN creates refined function-and-sequence specific protein families that can be used for identification and annotation of unknown family members. It also allows easy identification of erroneous annotations by spotting inconsistencies between similarities on annotation and sequence levels.
Background To achieve high quality of annotation, curators are using direct evidence from additional experiments or infer functional roles by sequence similarity to experimentally characterized genes or proteins. However, erroneous annotations generated at early stages may propagate to new homologous sequences, ultimately leading to erroneous annotation of entire families. The number of database errors is known to grow with time, as the number of entries and errors tend to accumulate. A recent analysis suggested that as databases grow annotation errors may propagate at an exponential rate [ 1 ]. One of the most accurate and consistently annotated databases is SwissProt [ 2 ] – a manually curated protein sequence database which strives to provide a high level of annotations such as the description of the function of a protein, its domain structure, post-translational modifications, variants, etc. All entries in SwissProt are annotated by experts, thus reducing the amount of errors expected from fully automatic methods. This impressive collection of textual information provides excellent opportunities for natural language processing in computational biology. Text mining of the free text in biomedical literature is well established (see [ 3 ] for a recent review). These methods use standard techniques such as the TF-IDF method, which considers background frequency of terms and the frequency of terms in the documents of interest. However, protein annotations are conceptually different from the free text of biomedical articles and abstracts upon which these methods are applied. While articles normally contain thousands of words and abstracts contain a few hundred, annotations are limited to a few key terms, averaging 7 and rarely exceeding 20 words. Research articles are always unique, whereas orthologous proteins often have identical annotations. On the other hand, differences in annotations may or may not signal difference in function, depending on the context. For example, the addition of the word "precursor" to the annotation does not imply functional variation. Also, multiple occurrence of a term in an article or an abstract signals its higher relevance to the subject. In contrast, presence of a term in a protein annotation normally signals presence of a function, and its repetition is meaningless. All these differences require the development of a different method for clustering protein annotations. This work describes a method called CLAN specially developed to assess the consistency of annotations. CLAN allows the rapid comparison of protein annotations, finding all proteins that are considered to have the same (or a closely related) function, according to the functional descriptions of the corresponding database entries. To group pairwise hits between these proteins, clustering of connected components was used, and sequence similarity within the clusters was considered. Using CLAN, an exemplar error was identified in the annotation of the M9.2 (M9.7) subunit of a vacuolar H+ ATPase (V-ATPase). Results General statistics We used CLAN [see Methods] to cluster SwissProt entries based on annotation strings. We experimented with score cut-offs [see Figure 1 and Methods for derivation of scores] between 10 -2 and 10 -6 . Depending on the threshold, the output contained between 2,070,221 (for e ≤ 10 -6 ) and 3,980,763 ( e ≤ 10 -2 ) pairwise annotation similarities. The output of CLAN was subjected to clustering of connected components, resulting in 43,416 ( e ≤ 10 -2 ) up to 47,344 ( e ≤ 10 -4 ) clusters, including singletons. When a lower score threshold was used, the number of clusters in the output decreased, as more proteins were excluded from any clusters (see below). A sample of few hundred of the annotation clusters was examined manually to assess whether they contained proteins annotated as functionally unrelated. At the low score cut-offs, we failed to find any such examples, thus suggesting a high degree of specificity. Also, hypothetical proteins were correctly not grouped together, which reflects their singularity in the protein function space. However, at these cut-off values, proteins annotated with a single word (i.e. Plastocyanin) failed to pass the threshold of meaningful annotation. This effect is due to the fact that no single term has a sufficiently low frequency to provide an score below the threshold. In total, 29,845 entries were not even defined as singletons (these include both characterized and hypothetical proteins) as their annotation strings are non-descriptive, resulting in high scores above the threshold. On the other hand, at the highest score threshold the method succeeds in clustering together many single-word annotations. However, this increase in sensitivity is counter-balanced by a decrease in specificity. At the score of 10 -2 , the biggest cluster contains proteins annotated as "Hypothetical X kDa protein in Y intergenic region", where X is the calculated protein molecular weight and Y is the genomic location. These proteins are most probably unrelated, perform different functions or may not be expressed at all, and thus should not be clustered together. In fact, this cluster is eliminated at score cut-off of 10 -3 . Comparison to Pfam database To estimate how well annotation clusters correspond to the sequence-based protein families, we compared the CLAN clusters ( e ≤ 10 -2 ) to the Pfam-A database [ 4 ]. This database is a curated semi-automatic collection of protein domains, defined by experts as sharing sequence (and usually functional) similarity. To perform a meaningful comparison, we used only CLAN clusters containing more than 3 members. We found that on average 91% of CLAN cluster members belong to a single Pfam family. However, the reverse relationship does not hold: only 51% of Pfam entries correspond to a single CLAN cluster, on average. This suggests that annotation-based clusters are about twice as small and contain only parts of sequence-based families. This is partially due to inconsistent annotation of protein families discussed below. On the other hand, it also reflects multifunctionality and diversification of some protein families. For example, a single Malate and Lactate dehydrogenase (MDH/LDH) family from Pfam is divided by CLAN to distinct annotation families, thus providing a finer distinction between family members that can be used for more accurate annotation of functional specificity [ 5 ]. As many of the CLAN clusters which correspond to several Pfam families contain multi-subunit complexes or multidomain proteins, a rigorous domain detection procedure is expected to drastically improve this result. For the results of the combination of CLAN annotation clusters with sequence-based clustering, see the section describing the CLAN database. Though SwissProt annotation is not independent from Pfam assignments and often depend on them, a few clusters were found that did not correspond to Pfam families. Thus, CLAN system might facilitate identification and rapid incorporation of new families to curated protein family databases such as Pfam. Analysis of individual clusters We aimed to determine whether similarity of the annotation corresponds to the sequence similarity of the proteins. For this, we compared protein sequences within annotation clusters using BlastP program [ 6 ]. Overall, we found very consistent sequence similarity relationships within annotation clusters. The proteins participating in the annotation clusters usually formed tight protein families, with all (or most) proteins sharing detectable sequence similarity to the other cluster members. This pattern is expected, since protein annotation is usually based upon sequence similarity. We considered this pattern to be typical and examined any deviations from it. Analysis of sequence similarity within annotation-based families identified the following exceptions to the consistent similarity patterns between all family members. 1. Fragments of proteins. The fragments often fail to produce significant Blast scores, thus appearing as singletons on the similarity graph. 2. Taxa – specific forms of enzymes that form separate protein families, such as Bacterial and Archaeal adenylate kinases, forming distinctive sequence-based families. 3. Non-orthologous gene displacements, such as malate dehydrogenases. 4. Protein complexes. Annotation – derived families often included multiple chains/subunits of protein complexes, such as Glutamyl-tRNA(Gln) amidotransferase subunits A, C, D and E. 5. False positives of CLAN. The only case found so far involves the 'hypothetical protein' cluster, appearing at the permissive score cut-off 0.01 (see previous section). 6. SwissProt annotation errors. Fragments SwissProt contains 7,232 entries annotated as "fragment". These are normally short fragments of larger protein molecules. An extreme example is LUXE_VIBFI (3 residues only), a fragment of Long-chain-fatty-acid-luciferin-component ligase. The complete sequence from a related Vibrio species documented in SwissProt (LUXE_VIBHA) contains 378 residues. There are at least 71 entries annotated as "fragment", whose sequences are less then 10 residues long. As their sequence often contains a subset from full-length close homologs, the information from these fragments is often redundant. Also, these short fragments fail to produce significant scores in any sequence similarity search procedure, and thus are usually irrelevant for sequence search engines. Moreover, fragments often create noise in sequence alignments and HMM analyses creating pseudo-conserved domains, and pseudo-gaps. In summary, we question the value of the presence of short protein fragments in a manually curated database, aiming for maximum precision rather then maximum coverage. Taxa-specific forms of enzymes In some taxa, proteins performing identical functions might diverge to a point where they are defined as separate sequence-based families, sharing marginal similarity. One such example is the bacterial and archaeal adenylate kinases. Though similar at the 3-dimensional level [ 7 ], the sequences of members of these families diverged beyond recognition by BlastP, and are thus considered as separate families. Non-orthologous gene displacements Enzymes belonging to different structural families may perform identical cellular function. In this case, annotation does not provide enough information about the type of the family a particular enzyme belongs to. One such case is represented by glucokinases, which belong to different families but perform identical function. For example, GNTK_BACSU and GNTK_ECOLI are both glucokinases and are annotated as such, were thus clustered together by CLAN but are unrelated to each other [ 8 ]. Protein complexes Subunits in protein complexes often have almost identical annotation. For example, Glutamyl-tRNA(Gln) amidotransferase subunits A, C, D, and E differ only in the chain identifier. The annotation of the subunits is somewhat misleading, as subunits A, B and C form one form of a rather ubiquitous complex [ 9 ], and subunits D and E form an alternative archaeal-specific complex [ 10 ]. Each of the chains forms a tight sequence-similarity based family, with no similarity detected between chains. Entries describing chains A, C, D and E appear in a single CLAN cluster. Interestingly, a few organisms have the entire complexes annotated in SwissProt (Table 1 ), although in many cases the documentation of these protein complexes is incomplete. The CLAN approach can help to identify the organisms where the chains remain to be identified, while the phylogenetic pattern of the complex may serve as an additional evidence for further annotation. Annotation errors Finally, CLAN can serve as a powerful tool to identify annotation errors. Overall, we found SwissProt annotations very consistent and robust. Most of the annotation-based sets resulted in tight sequence-based similarity clusters. A few cases where no similarity was found between a protein and the rest of the cluster were normally due to the limits in sensitivity by BlastP, and the homology was clear when multiple alignments, profile- or HMM-based search methods were used. Nevertheless, there are cases where annotation is problematic. Below, we discuss in detail an example of inconsistency found in the annotation of a cluster containing vacuolar H-ATPase. False annotation example: Vacuolar ATP synthase subunit M9.2 When analyzing the output of CLAN with score threshold 10 -6 , we found an interesting example of a propagated error. The protein VAOH_CAEEL (Q20591) from Caenorhabditis elegans has been annotated as "Probable vacuolar ATP synthase subunit H (EC 3.6.3.14) (V-ATPase H subunit) (Vacuolar proton pump H subunit)". This protein is clustered by CLAN with other vacuolar ATP synthase subunits. However, BlastP alignment failed to produce evidence of similarity between this protein and other members of the cluster annotated as vacuolar ATP synthase subunit H. Also, this protein is much shorter then other members of the annotation cluster, comprising only of 96 residues, while other members of the cluster have more then 400 residues. In fact, C. elegans was reported to contain another two V-ATPase H subunits, with length of 451 and 470 residues that share substantial sequence similarity with other members of the family. Multiple alignments of the family suggested that VA0H_CAEEL is a distant protein not related to the rest of the family. The SwissProt entry does not contain a link to a published work. A PSI-Blast search against the NCBI NR database allowed to identify other members of the family. The family was first described in the literature as "M9.2 V-ATPase subunit" in bovine [ 11 ]. Interestingly, an appropriate sequence entry exists in SwissProt with an adequate reference (VA0H_BOVIN). However, this protein is named in SwissProt as "vacuolar ATP synthase subunit H", and the name V-ATPase M9.2 subunit is suggested as a synonym. Another SwissProt entry (VA0H_HUMAN) contains a human homolog, with identical annotation. Interestingly, the bovine and human proteins were unified into a separate annotation cluster at score 10 -6 , which was unified with other members of "vacuolar ATP synthase subunit" cluster at lower score cut-offs. The M9.2 V-ATPase subunits were identified in mammals [ 11 , 12 ], insects [ 11 , 13 ] and C. elegans [ 11 ]. The alternative names included "M9.7" or subunit 'e' in insects [ 14 , 15 ] and ATP6H in dog [ 12 ]. A potential homolog was identified in the plant Mesembryanthemum crystallinum (gi 26986112) [ 16 ]. Using PSI-BLAST searches we have identified homologs in several other species, including Arabidopsis thaliana , Anopheles gambiae and a few paralogs in Drosophila melanogaster . The multiple alignment of these proteins is provided as supplementary data. The M9.2 protein was suggested to be a homolog [ 11 ] of yeast Vma21p protein (VM21_YEAST). The Vma21p protein is required for assembly of the V-ATPase, but is not found in the mature complex [ 17 ]. The orthology of the two proteins was suggested based on the weak sequence identity, association with V-ATPase complex, and a similar hydrophobicity profile [ 11 ]. However, the doubts concerning orthology of the two proteins were already expressed [ 14 ]. While Vma21p is localized in endoplasmic reticulum and is not a part of the mature V-ATPase [ 17 ], M9.2 is a subunit of mature V-ATPase located in the vacuolar membrane [ 11 ]. In addition, the topology of the yeast protein seems to be inverse to that of the mammalian and insect proteins: glycosylation at the C-terminus of the M. sexta protein indicates that the C-terminus is exposed to the extracellular surface, whereas the C-terminus of Vma21p appears to be localized on the cytosolic side of the membrane [ 14 ]. To identify any similarity between these two protein families, we used sequence-versus-sequence (BlastP), sequence to profile (PSI-BLAST) [ 6 ], and profile-profile (LAMA) [ 18 ] search methods. Each of these methods failed to detect any significant similarity between the two families. Our sequence analysis suggests that the initial reports of such homology between these families were not substantial, and the reported marginal similarity between the two sequences was due to the extent of hydrophobic regions in the two protein families. In summary, while the exact function of the M9.2 subunit of V-ATPase is still not known, it has a turbulent history of misinterpretation. It was confused with V-ATPase H chain in SwissProt, it was misaligned with yeast Vmi21p protein, it has several names in different organisms of which the name 'subunit e' is most confusing, as there is already a 'subunit E' in the complex. The current study used (i) CLAN to detect the discrepancies for protein annotation in SwissProt and (ii) sensitive sequence searches in order to demonstrate the independent status of this protein family. The error in the annotation of M9.2 subunit of V-ATPase was reported to the SwissProt developers, and is likely to be corrected at the time of publication of this work. False negative example Another form of annotation error results in a false negative case for CLAN. It occurs when a protein clearly belongs to a certain family but is not annotated as such. This event can be detected when several CLAN families map to a single sequence similarity-based family. One example is SUCD_BACSU, annotated as " Succinyl-CoA synthetase alpha chain (EC 6.2.1.5) (SCS-alpha)". This protein belongs to the SUCD family of proteins, and should cluster with its other members. However, its annotation contains synonyms " (Vegetative protein 239) (VEG239)", which makes this protein virtually unique in annotation space. Another well-known example is the initial characterization of some general stress proteins in bacteria, later shown to be ribosomal proteins (e.g. S1, L25) [ 19 ] – the initial annotations have not been modified. This type of assignments are usually based on genetics experiments and do not reflect the precise biochemical function of these proteins. In fact, there are multiple examples of proteins having unique or phenotype-based identifiers in their names, whereas it is clear that these proteins belong to larger protein families of known biochemical function. In our opinion, these proteins are examples of conflicting annotation with other family members and their annotations should be amended. The clan database Based on the results described above, we designed a database using the CLAN output with score threshold of 10 -2 . The CLAN database contains two levels of clustering. The first level contains clusters based on comparison of protein annotations, as described in the Methods section. To ensure completeness and reliability of the results, any protein annotated as 'fragment' was excluded from further processing. The second level in the database is generated by applying sequence-similarity-based grouping to each of the clusters obtained in the first stage. All members within annotation clusters are aligned using BlastP. The output of BlastP is subjected to various clustering procedures. In order to distinguish between the annotation-based clusters of the first database level, and the sequence-based sub-clusters of the second level, we call these 'clusters' and 'sub-clusters' accordingly. When referring to both these categories simultaneously, we use the notation (sub)clusters. To estimate the contribution from the second level of clustering, we repeated the comparison to Pfam database, using sub-clusters instead of clusters. We found that on average 97% of CLAN sub-clusters members belong to a single Pfam family, (compared to the 91% for clusters). The gain in selectivity was accompanied by gain in sensitivity, and 66% of Pfam family members corresponded to a single sub-cluster (compared to 51% for clusters). Compared to annotation-based clustering alone, the dramatic increase in the sensitivity reflects the fact that in the second level of clustering, non-homologous sequences are removed and form separate sub-clusters. The CLAN database has a web-based interactive interface, located at . To facilitate database navigation, we introduced (sub)cluster names, computed as a minimal consensus string for all (sub)cluster members. The MySQL-based storage of data allows fast searches based on sequence or cluster name, SwissProt accession or cluster identifier. The sequence similarity within clusters can be examined visually using the BioLayout software [ 20 ]. To facilitate the annotation of protein complexes, phylogenetic distribution of clusters between sub-clusters (such as Table 1 ) can be automatically generated upon the user's request. For all sub-clusters full-length multiple alignments are pre-computed with the ClustalW program [ 21 ], and the alignments can be viewed via an interactive Java interface. Discussion We have presented CLAN, a rapid and powerful method to build consistent annotation-based families, and assessed the consistency of annotation in protein databases. We have also presented the CLAN database, as a collection of protein families grouped by function. We believe that these tools can be used to improve the quality of database curation and assist annotation of newly sequenced proteins. The double-step clustering presented here has resulted in the creation of a reliable database of protein families, with include proteins similar on both annotation and sequence levels. One of the potential uses of this database is an accurate assignment of function to newly sequenced proteins by protein – to – profile alignments or HMMs. Our method has the potential to distinguish between subtle changes that might reflect diversification of protein function. For example, homologous Malate and Lactate dehydrogenases are found in different (sub)clusters, allowing their characterization as separate protein families. As most protein family databases group proteins by homology, as opposed to function, to the best of our knowledge, our approach is unique in its ability to build protein sequence families distinguishing between functions. In the future, other such comparisons could include other structural classifications such as CATH [ 22 ] or TRIBEs [ 23 ] and functional classifications such as EC or GO [ 24 ]. Methods We used SwissProt release of Feb 13, 2003 that contained 121,744 protein sequences. We aimed to find the similarity between the database entries solely based on the annotation and without any reference to sequence similarity. The annotation corpus considered here consisted of description line (protein name) and synonyms from the SwissProt database. We interchangeably use the term 'function' or 'annotation' to denote the contents of this corpus, implying the use of the corresponding annotation strings from the database. Definition of distance measure between annotation fields To obtain a distance measure, we first calculated the frequency φ of each word as the number of database entries containing the word N divided by the total number of entries in the database D (Equation 1). The word frequencies served as weights, because the words describing specific functions (such as "plastocyanin") occur less frequently than non-specific words (such as "protein"). We aimed to devise a scoring scheme that would produce highly significant scores for proteins with related function, while being able to distinguish between proteins with different functions. The score for the common terms p between protein functions may be computed as the product of frequencies of n terms shared between the two annotations (Equation 2). However, sharing only part of the total number of significant words is insufficient: proteins may use similar substrates for different reactions, or perform similar chemical reactions on different substrates. There may be three proteins with annotations AB, BC, and CD respectively, and clustering by similarity only could lead to clustering together AB and CD. For example, "lactate dehydrogenase" should be differentiated from "lactate permease" or "alcohol dehydrogenase". To achieve this distinction, we computed the difference between the two annotations, as above, for the case of the similarity measure. The score for the unique terms q is the product of frequencies of m terms unique to any of the two annotations (Equation 3). Finally, the distance e between the two annotations is defined as the fraction between the score for common terms p and the score for unique terms q for the two annotation strings (Equation 4). We refer to this scoring function as 'score' throughout the manuscript. An example of calculation of the score is shown on Figure 1 . Implementation The procedure calculating pairwise distances between protein annotations described above was implemented in a perl program. Its sole input consists of SwissProt identifiers and annotation lines, and the output contains pairs of proteins with the score of annotation similarity lower than a given threshold. Compared to sequence-similarity clustering, this procedure is very rapid: an all-against-all comparison of SwissProt database containing 121,744 entries took about 26 minutes on a single processor of a Sun-Fire-480R server with 2 GB of RAM. The result of the pairwise comparisons between protein annotations were then subjected to clustering of connected components. The protein sequences of the clusters obtained were subsequently analyzed with the BlastP program [ 6 ] and the results of similarities were clustered using GeneRage algorithm [ 25 ]. The results were visualized using the BioLayout software [ 20 ] (and Goldovsky, Cases et al., submitted). To evaluate the ability of CLAN to detect genuine protein families on the basis of annotation alone, the output ( e ≤ 10 -2 ) was compared to the Pfam-A database release 9.0 [ 4 ]. For each CLAN cluster with more then 3 entries, we have identified the corresponding Pfam entry (if available), and counted the fraction of CLAN entries found in the same Pfam cluster. The reverse comparison was computed in a similar manner.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552314.xml
546407
Simulator sickness when performing gaze shifts within a wide field of view optic flow environment: preliminary evidence for using virtual reality in vestibular rehabilitation
Background Wide field of view virtual environments offer some unique features that may be beneficial for use in vestibular rehabilitation. For one, optic flow information extracted from the periphery may be critical for recalibrating the sensory processes used by people with vestibular disorders. However, wide FOV devices also have been found to result in greater simulator sickness. Before a wide FOV device can be used in a clinical setting, its safety must be demonstrated. Methods Symptoms of simulator sickness were recorded by 9 healthy adult subjects after they performed gaze shifting tasks to locate targets superimposed on an optic flow background. Subjects performed 8 trials of gaze shifting on each of the six separate visits. Results The incidence of symptoms of simulator sickness while subjects performed gaze shifts in an optic flow environment was lower than the average reported incidence for flight simulators. The incidence was greater during the first visit compared with subsequent visits. Furthermore, the incidence showed an increasing trend over the 8 trials. Conclusion The performance of head unrestrained gaze shifts in a wide FOV optic flow environment is tolerated well by healthy subjects. This finding provides rationale for testing these environments in people with vestibular disorders, and supports the concept of using wide FOV virtual reality for vestibular rehabilitation.
Background One out of three elderly persons and more than one out of five working adults report dizziness[ 1 , 2 ]. There is a growing body of literature that suggests that persons with dizziness due to vestibular disorders fall, regardless of age [ 3 ]. Falls in persons with vestibular disorders have potentially catastrophic consequences[ 4 ]. Thus, development of rehabilitation methodologies that can improve balance could have a great impact on public health. The use of virtual reality (VR) has been explored in many areas of physical and mental rehabilitation [ 5 - 8 ]. Viirre [ 9 , 10 ] and Kramer et al. [ 11 ] were the first to discuss the use of VR with persons with vestibular disorders. The theoretical basis for using VR in the treatment of vestibular disorders is two-fold. First, persons with peripheral vestibular disorders have disequilibrium and complain of visual blurring [ 12 ]. These common symptoms may be caused by abnormalities in the vestibulo-ocular reflex (VOR) gain during head movements. Functional recovery of the VOR requires both visual inputs and movements of the head and body [ 13 ]. Retinal slip, i.e. movement of a visual image across the retina, is a powerful signal that can induce adaptation of vestibular responses [ 14 ]. If care is taken to minimize delays between head tracking devices and image updates, VR-induced retinal slip can be delivered in a controlled manner in order to cause adaptation. A randomized trial has demonstrated that persons with uncompensated peripheral vestibular disorders can improve with vestibular rehabilitation directed at inducing retinal slip [ 15 ]. People can also adapt to vestibular injuries through movement. Shepard et al. were able to reduce symptoms in 87% of patients who had chronic unilateral peripheral vestibular loss for at least 2 months [ 16 ]. Therefore, exposure to visual experiences and movement are key to the functional recovery of persons with vestibular disorders. Secondly, people with vestibular disorders complain of what has been called "space and motion discomfort" (SMD) and "visual vertigo" [ 17 , 18 ]. Situations that have been reported to precipitate SMD or visual vertigo include: walking in supermarket aisles or shopping malls, movement in cars and trains, long visual distances, or complex and confusing visual stimuli. Consequently, VR could allow persons with vestibular disorders to experience graded exposure to symptom-provoking situations in a controlled environment. Using VR for exposure therapy has a well-established foundation in the treatment of specific phobias (e.g. fear of heights) [ 8 , 19 ]. Wide field of view (FOV), screen-based projection devices (or spatially immersive displays) originated with the CAVE™ at the University of Illinois in Chicago [ 20 ]. Although the space required and cost make these systems impractical for clinical use, their wide FOV allow research laboratories to investigate how different motion cues affect balance and vestibular rehabilitation. One advantage of wide FOV devices is their ability to provide motion cues in the periphery, which can result in a greater sense of vection, or self-movement, compared with more limited FOV devices [ 21 ]. We believe that these factors may provide a substantial benefit compared with narrower FOV devices such as HMDs in the treatment of vestibular disorders. However, the wide FOV devices have also been associated with greater reports of simulator sickness [ 22 ]. Thus, while a wide FOV is desirable from a theoretical standpoint because a greater perception of motion occurs in the periphery, this same factor may elevate levels of simulator sickness and may be cause for discontinuing a treatment. The primary purpose of this paper is to present preliminary evidence for the ability of subjects to tolerate gaze shifting while situated in a wide FOV optic flow environment. We will demonstrate that healthy subjects were able to tolerate the environments without having a large incidence in simulator sickness. The incidence of simulator sickness depended strongly on how much experience the subjects had in the environment, and weakly on the duration of exposure within each visit. Methods Subjects Nine adults (22–75 years, mean ± S.D. 39 ± 19 yrs) with no history of vestibular system pathology participated after providing informed consent. Subjects had a visual acuity of 0.3 LogMAR units or better without using corrective lenses, and contrast sensitivity greater than 1.8 (Pelli-Robson Contrast Sensitivity). The protocol was approved by the University of Pittsburgh Institutional Review Board. Equipment The Balance NAVE Automatic Virtual Environment (BNAVE), a wide field of view projection-based immersive display system, was developed to investigate the multi-sensory interactions in postural control [ 23 ]. Three 2.4 m × 1.8 m (vertical × horizontal) back-projected screens are arranged as shown in Figure 1 . The side screens make an included angle of 110° with the front screen. The front screen is 1.5 m from the user, and the opening of the BNAVE at the location of the subject is approximately 2.9 m. The images are displayed using Epson 810p PowerLite LCD monoscopic projectors, with a pixel resolution of 1024 × 768 for each screen. Each projector is connected to an NVIDIA GeForce4 graphics processing unit (64 MB texture memory) installed in a separate PC (Pentium, 2.2 GHz, 512 MB RAM) running Windows 2000. The movement of the images on the three PCs is synchronized and controlled by a server via a local area network. The update rate of the images is consistently at least 30 frames per second. Figure 1 Experimental set-up for Task H, Visit 1 (see Tables 1 and 2 for explanation). Subjects stood upright on force platform and performed gaze shifts while target moved on a solid background. The target moved every 3 to 6 seconds from positions 1 to 4, located 40 to 50 degrees from midline. Several environments can be used for vestibular rehabilitation. One environment produces optic flow through the use of moving geometric patterns such as stripes or squares of alternating colors. Scene characteristics such as the spatial frequency, contrast, direction and speed of movement can all be independently prescribed. Furthermore, targets can be inserted into the environment that subjects are requested to look at. Consequently, for vestibular rehabilitation, we can ask a patient to perform unrestrained head gaze shifts to acquire moving targets while a background of moving stripes is moving past the patient, simulating the functional task of looking for a product while moving down the aisle of a grocery store. A virtual grocery store has also been developed (Figure 2 ). This environment contains several aisles, each with a different product theme. The dimensions of the aisle (width and length) are adjustable. Scene complexity can be altered by increasing the number of items on the shelves. The objects within the environment have both software-generated and photographic texture maps. In both environments, the task difficulty can be modified by varying the scene characteristics, thus exposing the patient to symptom-producing situations in a controlled and graded manner. In each environment, three-dimensional models were created using 3D Studio Max. Although the projectors used were not stereoscopic, a strong illusion of depth was elicited based on monocular depth cues such as perspective projection. The location of the eyepoint used for the perspective projection was based upon a fixed stance location in the horizontal plane and the subject's eye height. In addition, although head-tracked perspective correction was not used in the current application, this capability is possible and fully functional using a Polhemus Fastrak position tracking system (Polhemus, Inc. Colchester, VT). Figure 2 Virtual Grocery Store developed for providing exposure therapy for patients with dizziness that is increased in similar environments. Aisle length, shelf height, the number of products on the shelves, and object textures can all be manipulated depending on the goal of the therapy session. Procedure The ability of subjects to perform gaze shifts in response to moving targets superimposed on both static and moving backgrounds was examined. Eight different gaze tasks are performed on each visit (Table 1 ). Each gaze task was performed for 90 s with alternating movements to the left and right every 3 to 6 seconds (except for tasks B, F, H). Each task was performed with the six different backgrounds (Table 2 ). Each background condition was performed on a different visit. Background conditions 1 and 2 did not have optic flow and thus served as a control conditions for the remaining 4 backgrounds, which were randomized over the next 4 visits. During the high contrast conditions, the luminance of the stripes was 1 and 170 cd/m 2 , respectively. During the low contrast conditions, the luminance of the stripes was 15 and 34 cd/m 2 . The low contrast condition was based on average measurements of luminance obtained from products sampled at a local grocery store, using a luminance meter (LS-100 Luminance Light Meter, Minolta Corp. Ramsey, NJ). The spatial frequencies were set according to common sizes of soup cans (high, 4.2 cycles/meter) and cereal boxes (low, 1.4 cycles/meter) found in the local grocery store. The simulated velocity of the optic flow was 0.5 m/s. The central 25° of the display was masked by a solid region with a luminance of 15 cd/m 2 in order to avoid aliasing in the display as the stripes became smaller in the distance. Table 1 Gaze tasks performed on each of the six visits. On trials 3 to 8, the order of tasks D, E, F, G, H, and I are randomized on each visit. Trial Task 0 A) Initial reading 1 B) 10 deg head saccades to the right – calibration of head sensor 2 C) Control task – no head or eye movements 3 D) Head center, Eye saccades to ± 10 deg 4 E) Head left 50 deg, Eye saccades to ± 10 deg 5 F) Head right 50 deg, Eye saccades to ± 10 deg 6 G) Gaze stabilization during sinusoidal head movement at 0.25 Hz (VOR) 7 H) Gaze saccades to ± 40 and ± 50 deg from midline 8 I) Smooth pursuit to left between 10 – 60 deg, followed by smooth pursuit to right between 10 – 60 deg Table 2 Background conditions for each of the six visits. On visits 3 to 6, the order of conditions C, D, E, and F are randomized. Visit Condition Optic Flow Contrast Spatial Frequency 1 A No None None 2 B No High Low 3 C Yes Low Low 4 D Yes Low High 5 E Yes High Low 6 F Yes High High Rests of 3 minutes were provided after each task, during which Subjective Units of Discomfort (SUDS, 0–10 range) was rated and the Simulator Sickness Questionnaire (SSQ) was completed [ 24 ]. The SSQ contains 16 items on which subjects rate the degree of severity on a 4-item scale (0 = none, 1 = slight, 2 = moderate, 3 = severe). A total score is computed along with 3 subscales – nausea (general discomfort, increased salivation, sweating, nausea, difficulty concentrating, stomach awareness, burping), oculomotor stress (general discomfort, fatigue, headache, eyestrain, difficulty focusing, difficulty concentrating, blurred vision), and disorientation (difficulty focusing, nausea, head fullness, blurred vision, dizzy: eyes open, dizzy: eyes closed, vertigo). Furthermore, pulse and blood pressure was monitored after every trial using an automatic device. Initial recordings of each of the measures were also recorded prior to exposure. During each trial, postural sway was recorded using a force platform and 6 degrees of freedom electromagnetic trackers placed on the head and waist (Polhemus Fastrak, Colchester, VT). The accuracy of the trackers is 0.8 mm in translation and 0.15° in rotation, and the stated resolution of 0.2 mm in translation and 0.025° in orientation. The trackers have a latency of 4 ms and were digitized at 20 Hz. The horizontal and vertical eye movements were measured using video-oculography (VOG, Figure 3 ). The VOG device (Micromedical, Chatham, IL) was fastened using an adjustable helmet insert and contained see-through dichroic glass that reflected images of the eyes up to infrared cameras. The accuracy of the VOG is 0.3 deg and the images are captured at 60 Hz. Using the sampling rates of the tracker and VOG, the maximum delay between recording simultaneous movements of both the head and eye would be 33 ms. The head and eye movements were calibrated using targets placed in known locations. Eye-in-head position is combined with head-in-space position to yield continuous gaze position (eye-in-space). The timing and accuracy of the head gaze movements with respect to the targets will be the subject of a future report. Figure 3 Video-oculography (VOG) see-through goggles used to measure eye-in-head position in the vertical and horizontal planes. Subject also wore 6 degrees of freedom electromagnetic tracker on top of his head to measure head in space position. Both signals are combined to obtain gaze, or eye gaze-in-space position. Data Analysis Five dependent variables of interest were examined: SUDS, total SSQ, and 3 SSQ subscales. The three subscales of the SSQ were computed by summing the scores for the component items of each subscale, and multiplying by an appropriate weighting factor (9.54 for Nausea, 7.58 for Oculomotor, and 13.92 for Disorientation) [ 24 ]. The total SSQ score was equal to the sum of the 3 subscales, multiplied by 3.7. Histograms of each dependent variable were plotted according visit number (1 to 6) and trial number (0 to 8). After observing that the data were not normally distributed due to a large majority of 0 responses and the presence of long tails, we tabulated the frequency of non-zero responses for each dependent variable. The effect of visit number and trial number on the frequency of non-zero responses was evaluated using χ 2 statistics. Because of the large number of comparisons (5 dependent variables × 2 main effects), the significance level was set at α = 0.05/10 = 0.005. Results Gaze shifts performed by one subject in response to targets moving at least 80° in the yaw plane (Task G) are shown in Figure 4 . These gaze shifts are combinations of head and eye movements. The top plot demonstrates the target-in-space yaw position (T), head-in-space yaw position (H), and eye-in-head yaw position (E). The last two quantities are combined to produce the gaze position (G) shown in the bottom plot. In this example, it can be seen that the eye and head movements effectively combine to allow the person to look at the desired target position. Analytical techniques will allow us to quantify head and eye coordination strategies in persons with vestibular disorders, both in optic flow fields and more complex environments. Figure 4 Gaze shifts during Task H obtained from 1 subject. Top: target yaw position (T), head-in-space yaw position (H), and eye-in-head yaw position (E). Bottom: target yaw position (T), and eye gaze-in-space yaw position (G). The tolerance of the subjects to the gaze shifts was assessed using Subjective Units of Discomfort (SUDS) and the Simulator Sickness Questionnaire (SSQ). For each of the measures, the majority of the responses were zero. Therefore, each of the measures was converted into a binary scale consisting of responses equal to zero or greater than zero. The SUDS rating was greater than zero only 25% of the time. For each of the SSQ subscales, a score greater than 0 was given if any of the 7 component items for the subscale was rated greater than 0. The SSQ:Oculomotor subscale had the most non-zero responses, at 29%. The SSQ:Nausea and SSQ:Disorientation subscales had 12% and 5% non-zero responses, respectively. Overall, the SSQ-Total had 31% non-zero responses. The effect of visit number and trial number on the frequency of non-zero responses was examined using χ 2 statistics. The effect of visit number was significant for SUDS, SSQ:Nausea, SSQ:Oculomotor, and SSQ:Total (Table 3 ). The most obvious finding was that the number of non-zero responses was significantly greater the first visit. The effect of trial number was not significant for all measures (Table 4 ). Table 3 Incidence of non-zero responses for the self-reported Subjective Units of Discomfort SUDS) and Simulator Sickness Questionnaire (SSQ) subscales and total score, as a function of visit number. Mean incidence, χ 2 test of association, and p value are also provided. * indicates significant effect of visit number. Visit SUDS SSQ:Nausea SSQ:Oculomotor SSQ:Disorientation SSQ:Total 1 35 27 56 9 56 2 38 11 33 0 33 3 22 11 26 5 28 4 25 16 19 10 26 5 19 0 26 1 27 6 10 8 14 8 14 Mean 25 12 29 5 31 χ 2 17.9 25.2 30.7 10.1 25.3 p-value .003* <.001* <.001* .072 <.001* Table 4 Incidence of non-zero responses for the self-reported Subjective Units of Discomfort (SUDS) and Simulator Sickness Questionnaire (SSQ) subscales and total score, as a function of trial number. Mean incidence, χ 2 test of association, and p value are also provided. No significant effect of trial number was found. Trial SUDS SSQ:Nausea SSQ:Oculomotor SSQ:Disorientation SSQ:Total 0 (initial) 4 7 11 0 11 1 17 6 15 2 15 2 24 9 24 11 31 3 28 13 28 2 30 4 30 13 31 6 33 5 26 15 35 6 37 6 30 15 39 7 41 7 32 15 38 8 40 8 34 17 40 8 40 Mean 25 12 29 5 31 χ 2 15.0 5.2 16.0 8.0 16.2 p-value .06 .73 .04 .43 .04 Discussion The ability to perform coordinated gaze movements within an optic flow environment may lead to the development of tools to improve outcomes in vestibular rehabilitation. The current research represents the first attempt to assess self-reported tolerance to these movements in a wide field of view environment. The ratings indicate that on a majority of the trials, this group of healthy subjects experienced no discomfort and simulator sickness while performing 8 different types of gaze movements under different optic flow conditions. On 75% of the trials, subjects reported no subjective discomfort. On 69% of the trials subjects reported no symptoms of simulator sickness. Although there is no data with which to compare the incidence of symptoms during performance of coordinated eye and head movements in an optic flow environment, we have chosen to review the data obtained from flight simulators and head mounted display units. The incidence of symptoms in the current study is at the lower end of the range relative to the previous flight simulator based studies. For example, the incidence of simulator sickness in this military pilots has ranged from 6–62%, depending on the type of simulator [ 25 ]. The incidence of simulator sickness after VR exposure in non-pilots is slightly greater; approximately 60 to 80% of subjects report symptoms of eyestrain, headache, nausea, and malaise after only 10–20 minutes [ 26 , 27 ]. The decreased incidence of problems may be related to several factors. For one, the within trial exposure time was short, approximately 90 seconds. Therefore the total exposure time was only about 12 minutes, which is on the lower end of reported exposures described in the literature [ 26 , 28 ]. Secondly, significant rest breaks were provided between trials. The incorporation of rest breaks to reduce simulator sickness has not been studied well. Thus, the type of display device, as well as the content and nature of the task may have an effect on the amount of sickness. Thus, although the current results are not directly comparable to the previous research, they will serve as a foundation for future work that examines the incidence of symptoms while performing coordinated eye and head movement tasks within a virtual grocery store, or using a head mounted display. There was a significant effect of visit number of the number of non-zero responses. Analysis of the data revealed that subjects appeared to have greater levels of discomfort and symptoms of simulator sickness on the first visit. It is possible that subjects had greater levels of discomfort due to their lack of prior exposure to the environment/experiment. Furthermore, our data is consistent with findings from other studies that subsequent exposures to environments result in lower simulator sickness [ 26 , 28 ]. Interpretation of this finding is clouded by the confounding effect of the background displayed during visit 1. During the first visit, the subjects always experienced movement of targets superimposed on a solid background without optic flow. The experiment was designed in this way because we assumed that this background would elicit the least amount of symptoms, and would serve as a suitable background for subjects to learn the movements. Consequently, the finding of decreased tolerance to the movements during the first visit was unexpected. Unfortunately, we are not able to distinguish if the increased discomfort and simulator sickness was due to the subject's inexperience with the environment or due to the type of background. We did not find a significant effect of trial number on the number of non-zero responses to SUDS and the SSQ. However, it was apparent that there was a trend for greater number of non-zero responses as trial number increased for the SUDS, SSQ:Oculomotor, and SSQ:Total Severity. In previous reports using flight simulators, the level of simulator sickness increased as the duration of exposure increased [ 28 ]. Moreover, symptoms tended to persist after the simulation was finished [ 25 , 29 ]. Addition of more subjects may reveal the trial effect to be significant. Nonetheless, the short duration of exposure within each trial (i.e. 90 seconds) and the amount of rest provided to the subjects between trials (i.e. 3 minutes) may have ameliorated the development of symptoms as the trials accumulated. The SSQ:Oculomotor subscale had the greatest number of non-zero responses. Usually, this subscale is elevated secondary to the effects of using a head-mounted display (HMD) device. HMD users frequently suffer from eyestrain, and blurred vision and short-term changes in binocular vision possibly due to alterations in the balance between the vergence and accommodation systems [ 30 , 31 ]. In our case, we attribute the scores to the video-oculography (VOG) device. Several subjects commented that the VOG caused eyestrain. Furthermore, some subjects reported that the dichroic lenses interfered with their viewing of the environment. We intend to examine if using electro-oculography (EOG), which measures eye movements via surface electrodes surrounding the eyes, will reduce the amount of oculomotor symptoms. The ability to move one's head and search for targets is a functional task that is often impaired in people with vestibular disorders. In vestibular rehabilitation, patients are encouraged to move their head during daily activities because movement is needed for adaptation and reweighting of the sensory signals. Using optic flow and/or virtual environments, this activity could be restored. Our preliminary analysis demonstrates that healthy people are able to coordinate their head and eye movements in the presence of stationary and moving backgrounds with few side effects. The next step is to perform similar experiments with people who have vestibular disorders. In addition, it is imperative to study if these movements can be tolerated while using narrower FOV HMDs. Head mounted displays will most likely be the desired display system of choice for vestibular rehabilitation because of the relatively modest cost and high portability. However, there are other characteristics of the display systems that need to be considered. For example, it is common for users of HMDs to complain of eyestrain, blurred vision, headache, and nausea [ 30 , 31 ]. In addition, DiZio and Lackner suggest that wearing an HMD effectively increases the mass and inertia of the head, thereby leading to a sensory rearrangement that may have some part in simulator sickness [ 32 ]. This theory is supported by the work of Howarth and Finch, who examined the amount of nausea generated while subjects wore an HMD under 2 conditions [ 33 ]. In one, subjects changed heading by using a handheld input device. In the other, subjects changed heading by rotating their head. Nausea was significantly greater when subjects navigated using their head. The lag between head movement and scene movement, and the variability in frame update rate has also been considered to play an important role in generating sickness with the use of HMDs [ 26 , 33 ]. However, as head tracking technology has improved, and update lags have been reduced, this factor is probably not as important as it once was. Thus, research on the use of HMDs in people with vestibular disorders is necessary to determine if they can be safely used in this population. Conclusion The performance of head unrestrained gaze shifts in a wide FOV optic flow environment is tolerated well by healthy subjects. This finding provides rationale for testing these environments in people with vestibular disorders, and supports the concept of using wide FOV virtual reality for vestibular rehabilitation.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546407.xml
534091
Physical function and associations with diet and exercise: Results of a cross-sectional survey among elders with breast or prostate cancer
Background Functional decline threatens independent living and is common among individuals diagnosed with cancer, especially those who are elderly. The purpose of this study was to explore whether dietary and exercise practices are associated with physical function status among older cancer survivors. Methods Mailed surveys were used to ascertain data on physical function, dietary fat, fruit and vegetable (F&V) consumption, and exercise among elderly diagnosed with early stage (I-II) breast (N = 286) or prostate cancer (N = 402) within the past 18 months. Results Sixty-one percent of respondents reported diets with <30% of energy from fat, 20.4% reported F&V intakes of 5+ daily servings, and 44.6% reported regular vigorous exercise. Significant, independent associations were found between physical functioning and reported dietary fat intake, F&V consumption, and exercise. A simultaneous multiple regression model controlled for age, race, gender, time since diagnosis and concurrent health behaviors yielded the following estimates: (1) 0.2 increase in the SF-36 physical function subscale (PFS) score with each reported 1% decrease in percent energy from fat (p < .0001); (2) 0.9 increase in the SF-36 PFS score for each reported serving of F&V/day (p = .0049); and (3) 15.4 increase in the SF-36 PFS score with a positive response for regular vigorous exercise (p < .0001). Conclusions Results of this cross-sectional survey suggest that regular vigorous exercise and consumption of diets low in fat and rich in F&Vs are associated with higher levels of physical functioning among older cancer survivors. Interventions that promote healthful lifestyle change may deliver considerable benefit within this ever increasing and vulnerable population.
Background There are 9.6 M cancer survivors in the US today, and 61% are age 65 or older [ 1 ]. Over the next 50 years, the number of elderly cancer survivors is expected to double [ 2 ]. Unfortunately, these individuals are at greater risk for other cancers, cardiovascular disease, osteoporosis, diabetes, and functional decline [ 1 - 5 ]. While all elderly are at increased risk for functional decline [ 6 ], those diagnosed with cancer are even more vulnerable and may experience functional losses that threaten independent living [ 7 - 10 ]. Adherence to healthful lifestyle behaviors may be one way that older cancer survivors can maintain or regain higher levels of physical functioning. Yet, there are little data to support this premise. Methods We explored associations between lifestyle factors and physical functioning among elderly cancer survivors who were screened for Project LEAD (Leading the Way in Exercise And Diet), a home-based, diet and exercise intervention trial [ 11 ]. Briefly, individuals (age 60+) who were no more than 18 months beyond a diagnosis of early stage (I-II) breast or prostate cancer were identified through cancer registries, private practices, and self-referral. Patients were mailed a letter of invitation, consent forms, a postage-paid return envelope, and a brief survey. The purpose of the survey was to screen-out individuals already practicing healthful behaviors [i.e., those engaging in regular, vigorous exercise or consuming a low fat diet with 5 or more daily servings of fruits and vegetables (F&V)] and individuals unlikely to benefit from the telephone counseling-mailed material intervention [i.e., those who were mentally incompetent, severely hearing impaired, or who had conditions that precluded unsupervised exercise or a high F&V diet (i.e., scheduled hip or knee replacement, walker or wheelchair use, recent stroke or heart attack, angina, congestive heart failure, chronic obstructive pulmonary disease, end-stage renal failure, and/or warfarin-use)]. The survey incorporated validated scales to assess dietary fat (Block Dietary Fat Screener) [ 12 ], and F&V consumption (5 A Day items) [ 13 ]. Given space constraints of the survey and our focus on vigorous exercise, one item was used to assess exercise ("On average, do you do continuous vigorous exercise for at least 20 minutes, 3 or more times per week?") [ 14 , 15 ]. The SF-36 Physical Function Subscale (SF-36 PFS) also was modified slightly to omit item #2 ("Does your health limit you in moderate activities") and scores were scaled 0–100 with the assumption that the average value for the missing item was the same as that for the remaining items [ 16 ]. T-tests and chi-square analyses were performed to determine if respondents differed from non-respondents on age, time since diagnosis, race and gender, and to test for differences between male and female respondents. Pearson correlation coefficients indicated associations between F&V intake and percent dietary fat, and physical function scores. T-tests were used to explore associations between exercise and function. Linear regression analyses permitted an examination of associations between physical function and health behaviors controlling for several likely confounders including gender, age, race, time since diagnosis and concurrent health behaviors. Given the homogeneity in stage at diagnosis, stage information was omitted from this analysis. Results and Discussion Between August 2000 and May 2003, 2,431 cases were identified and physician approval for contact was granted for 2,034 cases. Incomplete address information existed for 24 of these cases, yielding 2,010 viable posted surveys, of which 688 complete surveys were returned (34% response rate). Characteristics of respondents and non-respondents are provided in Table 1 . Respondents were significantly more likely than non-respondents to be younger, white, male, and more proximal to their date of diagnosis. Table 1 Demographic characteristics of respondents and non-respondents Characteristic Respondents (n = 688) Non-Respondents (n = 1322) P-value Age (years) Mean (sd) 71.4 (5.0) 73.0 (5.9) <.0001 Range 60 – 94 64 – 96 Race [% (N)] White 83.4% (574) 75.0% (991) <.0001 African American 12.4% (85) 21.0% (278) Other/Unknown 4.2% (29) 4.0% (53) Gender [% (N)] Female 41.6% (286) 52.7% (696) <.0001 Male 58.4% (402) 47.4% (626) Time Elapsed Since Diagnosed (months) Mean (sd) 10.8(4.9) 11.3 (5.8) .046 Distribution [%/(N)] 0 – 3 months post-diagnosis 9.2% (63) 11.4% (151) >3 – 6 months post-diagnosis 8.9% (61) 10.1% (134) >6 – 9 months post-diagnosis 24.0% (165) 17.4% (230) >9 – 12 months post-diagnosis 21.8% (150) 20.7% (274) >12 – 15 months post-diagnosis 18.0% (124) 15.9% (210) 15+ months post-diagnosis 18.2% (125) 24.4% (323) Data regarding physical function and lifestyle behaviors are provided in Table 2 . In bivariate analyses, elders with prostate cancer (men) have significantly higher SF-36 PFS scores than those with breast cancer (women). A majority of all respondents adhere to a low fat diet, while a minority pursue regular vigorous exercise or eat 5 or more daily servings of F&Vs. Men are significantly more likely than women to exercise and to consume 5 or more daily servings of F&Vs. However, female cancer survivors are more likely to report low fat diets. Table 2 Physical function and health behaviors of survey respondents (N = 688) Variable Total Sample (N = 688) Women with Breast Cancer (N = 286) Men with Prostate Cancer (N = 402) p-value Physical Function [SF-36 – 9 items scaled 0–100, mean (sd)] 74.3 (25.3) 68.3 (25.5) 78.5 (24.3) <.0001 Exercise (% who "vigorously exercise for at least 20 minutes, 3 or more times a week") 44.6% 36.7% 50.2% .001 % of Energy (Calories) from Fat mean (sd) 29.1 (21.8) 25.6 (16.2) 31.7 (24.7) .0003 % consuming <30% of Energy from Fat 61.1% 65.8% 57.6% .031 Fruit and Vegetable Consumption mean number of servings (sd) 4.1 (2.9) 3.7 (2.3) 4.4 (3.2) .0005 % consuming 5+ servings/day 20.4% 15.4% 24.1% .005 Modest correlations were obtained between the indicators of exercise and diet, with positive agreement noted between F&V intake and exercise (Pearson ρ = .12/p = .003) as well as F&V intake and dietary fat (ρ = .42/p < .0001). An inverse association was noted between dietary fat and exercise (ρ = -.08/p = .05). In bivariate associations, all three behaviors were significantly associated with SF-36 physical functioning scores (p < .05): F&V intake (ρ = .09), dietary fat (ρ = -.10), and vigorous exercise (ρ = .37). In multivariable linear regression analyses, with the SF-36 PFS score serving as the dependent variable and controlling for age, race, gender, time since diagnosis, and concurrent health behaviors, the simultaneous associations of the three indicators remained statistically significant (independent) and yielded the following estimates: (1) a 0.2 increase in the SF-36 PFS score with each reported 1% decrease in percent energy from fat (p < .0001); (2) a 0.9 increase in the SF-36 PFS score for each reported serving of F&V/day (p = .0049); and (3) a 15.4 increase in the SF-36 PFS score with a positive response for regular vigorous exercise (p < .0001). In comparing these cross-sectional data on lifestyle behaviors of older cancer survivors to normative data reported on healthy elders, as well as to previous data reported on general populations of breast and prostate cancer survivors (where a dearth of data have been reported on older cancer survivors per se), we find both differences and similarities. Like data that exist on healthy elders (age 65+) responding to the 2000 Behavioral Risk Factor Surveillance System (BRFSS) survey [ 17 ], or findings of a previous study of 978 breast and prostate cancer survivors [ 18 ], a minority of the respondents to this survey report consuming 5 or more daily servings of F&Vs. However, the percentage of our respondents who reportedly achieved 5 A Day guidelines was lower than that reported by these two previous studies (20.4% as compared to 34.4% and 42%, respectively) [ 17 , 18 ]. In contrast, a majority of survivors in both this study (61.1%) and the previous study on cancer survivors (69%) report adherence to a low fat diet [ 18 ], whereas higher mean intakes of fat 32–33% were noted among a general population of elders (age 60+) responding to the National Health and Nutrition Examination Survey, Phase I (1988–1991) [ 19 ]. Like the previous study on breast and prostate cancer survivors (where 58% reported routine exercise of moderate intensity or greater compared to 47% within the general population) [ 18 , 20 ], a greater proportion of these elderly cancer survivors report routine vigorous exercise when compared to BRFSS data on the general population (i.e., 44.6% vs. 11%) [ 20 ]. Findings may differ due to differences in survey instruments, populations and varying amounts of respondent bias, however at least for dietary fat and exercise, data largely support the prevalent finding that cancer survivors tend to report healthier lifestyle behaviors [ 21 ]. The fact, that we did not see this trend in F&V consumption may be due to non-penetrance of the 5 A Day message among the survivor population, or may be the result of our population being significantly older than that reported in previous studies – a population where proportionally more individuals may be edentulous or suffering from G.I. intolerances that serve as barriers to F&V consumption. Our data on physical function, however, indicate higher levels within our sample (74.3 ± 25.3) when compared to U.S. age-matched norms (69.4 ± 26.3) [ 22 ], despite previous findings, which suggest decreased physical functioning among older cancer survivors [ 7 - 10 ]. A potential explanation for the higher physical function scores exhibited within our sample may relate to the higher prevalence of healthful lifestyle practices which in turn may increase physical function. This links back to the primary thrust of this study, which was to determine evidence for a link between functional status and health behavior. Our results suggest that regular vigorous exercise and a diet rich in F&Vs and low in fat is associated with higher levels of physical functioning among elderly recently diagnosed with cancer. Although the association between exercise and improved function has been reported repeatedly in other studies among elders [ 6 , 23 ], this is the first study to show this association exclusively among elderly cancer survivors. Further, the fact that regular vigorous activity is associated with a 15.4 point differential in functional status, accounts for a magnitude of effect that is comparable to 0.61 standard deviations (sd), and far exceeds the 0.22 sd noted in a previous study by Baker et al. as being both statistically and clinically significant [ 7 ]. Furthermore, only one study to date has reported the link between diet composition and physical function. Ortega et al. found that diets low in fat and high in F&Vs were associated with higher levels of physical functioning among a sample of upper socio-economic, elderly male Spaniards at risk for cardiovascular disease [ 24 ]. This previous report however did not include defined estimates, so it is difficult to draw exact comparisons regarding the magnitude of dietary change associated with a given difference in SF-36 PFS score. Nonetheless, our data suggest that modest gains in physical functioning may be achieved with dietary changes that are both feasible and realistic [i.e., a 0.9 increase in PFS score with each additional serving of F&V or a 0.2 increase in PFS score with each 1% decrease in percent energy from fat (roughly equivalent to a half a teaspoon of butter, margarine or oil for most caloric intakes)]. Albeit, multiple servings of F&Vs (roughly 5 per day) or substantial decreases in fat (approximately 20% of total Calories) would be necessary to achieve clinical significance if dietary changes were pursued in isolation, however if taken together, as in the pursuit of a healthier overall diet, it is conceivable that appreciable functional improvement could occur. While the relationship between exercise and physical function is strong and mechanistically more direct, low fat diets (with lower amounts of saturated and trans fats) and increased F&Vs also provide theoretically viable, yet not fully elucidated pathways, to increased function [ 25 ]. It must be noted, however that our findings differ somewhat from those of a recently reported study of cancer survivors by Blanchard et al. [ 26 ], who found a significant association between Health Related Quality of Life (HRQOL) and exercise, but not between HRQOL and F&V intake. Differences between studies with regard to the sample (e.g. our sample included 688 elderly breast and prostate cancer survivors whereas the sample of Blanchard et al. [ 26 ] was comprised of 316 breast, prostate and colorectal cancer survivors of all ages), and survey items (e.g., we used validated items taken from the National 5 a Day trials [ 13 ], whereas Blanchard et al. [ 26 ] used one composite item) may account for the discrepancy in findings. Accordingly, and for the purposes of designing effective interventions to improve physical function in older cancer survivors, further research is necessary to corroborate associations between diet and physical function (if any) and to clarify responsible mechanisms. The primary limitations of this study relate to respondent bias and cross-sectional design. Our response rate of 34% is indeed less than the 59% we have achieved in previous mailed surveys among similar populations [ 18 ]. Given that this survey was linked to accrual efforts for a yearlong behavioral intervention trial [ 11 ], a response rate in this range was anticipated. Indeed, our response rate is similar to those of 35–50% cited by Martin Brown, PhD (Chief of the National Cancer Institute's Health Services and Economics Branch) in a published interview regarding survey responses rates among cancer patients [ 27 ]. In addition, in attempting to control for factors that differed between respondents and non-respondents, we acknowledge that there likely exist other important factors that were not included (e.g. items that our survey did not ascertain such as socio-economic status), or the fact that we were unable to assess and subsequently control for lifestyle behaviors among those who did not respond – non-respondents whose lifestyle behaviors may have placed them at risk (i.e., sedentary, poor diets) and less inclined to respond, or conversely those already adhering to healthy lifestyle behaviors and less compelled to participate. Another limitation of our study was the use of abbreviated scales or items to obtain health-related data. As an example, the single item used to capture vigorous exercise may have led to inflated rates of reporting due to the absence of a response category for moderate exercise. Finally, our study was cross-sectional, and cause and effect possibly are confounded. Conclusions Findings of this study support the recent American Cancer Society diet and exercise guidelines for survivors [ 25 ], which call for a physically active lifestyle and a plant-based diet. However, more research is needed to confirm associations between lifestyle factors and physical function, especially among broader populations of survivors (i.e., other types of cancer, other age groups, and among short-term versus long-term survivors). Ultimately, randomized controlled trials enrolling of older persons with cancer are needed to determine the potential benefit of diet and exercise interventions in reorienting the trajectory of functional decline. List of Abbreviations F&V: Fruits and Vegetables SF-36: Short Form-36 Health Survey PFS: Physical Function Subscale Competing Interests The author(s) declare that they have no competing interests. Authors' Contributions WDW, ECC, MCM, CFP and HJC conceived of the study, and participated in its design. DCS participation in the coordination of the study, as well as data cleaning and interpretation of findings. RS performed the statistical analysis. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC534091.xml
544562
Heritability of Stroop and flanker performance in 12-year old children
Background There is great interest in appropriate phenotypes that serve as indicator of genetically transmitted frontal (dys)function, such as ADHD. Here we investigate the ability to deal with response conflict, and we ask to what extent performance variation on response interference tasks is caused by genetic variation. We tested a large sample of 12-year old monozygotic and dizygotic twins on two well-known and closely related response interference tasks; the color Stroop task and the Eriksen flanker task. Using structural equation modelling we assessed the heritability of several performance indices derived from those tasks. Results In the Stroop task we found high heritabilities of overall reaction time and – more important – Stroop interference ( h 2 = nearly 50 %). In contrast, we found little evidence of heritability on flanker performance. For both tasks no effects of sex on performance variation were found. Conclusions These results suggest that normal variation in Stroop performance is influenced by underlying genetic variation. Given that Stroop performance is often hampered not only in people suffering from frontal dysfunction, but also in their unaffected relatives, we conclude that this variable may constitute a suitable endophenotype for future genetic studies. We discuss several reasons for the absence of genetic effects on the flanker task.
Background The Stroop test [ 1 ] is arguably the best-known neuropsychological test to tap attentional (dys)function. In the color words version of this test the instruction is to attend to the color of the ink in which a word is printed and name this color aloud. At the same time, the printed words may also read certain color names that are different from the color of the ink in which it is printed. As has been observed on numerous occasions, there is a strong tendency to respond to the content of the word, and not to the ink color. This is evidenced by an increase in response time and a decrease in accuracy relative to a neutral control condition. The Stroop test has been used both to tap fundamentals of human information processing (e.g. [ 2 ]), and as a clinical aid to assess attentional dysfunction, e.g., due to a frontal or fronto/parietal deficit. Brain imaging and neurological studies consistently point to the prefrontal cortex (PFC) as the site involved in resolving the response conflict. As a consequence, people suffering from attentional impairments, caused by prefrontal abnormalities (developmental or acquired), tend to suffer more from Stroop interference than controls. For example, the test succesfully differentiates unaffected controls from people suffering from schizophrenia (e.g., [ 3 ]). In a similar vein, people suffering from attention-deficit/hyperactivity disorder (ADHD) suffer from Stroop interference ([ 4 ]; see also [ 5 ]), although a recent meta-analysis cast some doubt about the usefulness of the Stroop task in differentiating people with ADHD from controls [ 6 ]. There now exist numerous versions of the Stroop test. For example, instead of using color words, researchers have adopted more ecologically relevant items, such as emotion words, pictures of food items or of threatening objects, etc. In addition, it is now also common to use computerized versions of the Stroop task, permitting a trial-by-trial analysis of performance. But what all these different Stroop versions have in common is that the subject is always presented with a stimulus that simultaneously activates two conflicting response channels; one response is activated by the instructions, whereas the other response is activated by elements in the array that strongly invite an alternative – yet incorrect – response. In order to resolve this response conflict the subject has to direct attention to task relevant information and ignore information from the task irrelevant channel. The time needed to resolve this conflict is derived using subtractive logic, and can be used as an index of the efficiency of the attentional system under investigation. A task that is less widely used in clinical circles, but that also indexes the efficiency of the frontal network is the Eriksen flanker task. In the arrow version of this task, subjects have to respond to the direction of a left or right pointing arrow, and ignore flanking arrows that point in the opposite direction as the target arrow [ 7 ]. Similar to the Stroop task, there is a tendency to respond to the distracting flanker elements, and subjects have to resolve this response conflict prior to emitting the designated response. It is consistently found that response times are elevated due to the target-flanker incongruity, relative to a neutral control condition where target and flankers are congruent (that is, they all point in the same direction). There is evidence that the Stroop task and the flanker task are supported by the same cognitive system. For example, using functional magnetic resonance imaging (fMRI) it was found that both tasks activated largely overlapping brain regions, viz. the anterior cingulate cortex (ACC) and the left prefrontal cortex [ 8 ]. In addition, and similar to the Stroop task, subjects with ADHD spend more time resolving the conflict between the competing responses than controls (e.g., [ 9 ]). Also, adult subjects with ADHD showed consistent underactivation in the ACC (cognitive division) during a counting version of the Stroop task, compared to controls [ 10 ]. The Stroop task (and, to a lesser extent, the flanker task) has thus acquired a strong neuropsychological validation, and is nowadays widely used in clinical settings. However, studies adopting an individual differences paradigm have revealed that the time needed to resolve the response conflict in the Stroop task does not predict the time needed to resolve the response conflict in the flanker task. In an earlier study we [ 11 ] found that the interference scores between the tasks were uncorrelated. A similar finding was reported by [ 12 ] using slightly different task versions. Further insight into the nature of these interference tasks might be gained by adopting a genetic perspective on normal and abnormal frontal functioning. A wealth of studies has now shown that many frontal psychopathologies are influenced by genes. For example, the heritability of ADHD is estimated to be around 80% (e.g.[ 13 ]). In a similar vein, the heritability of attention problems as established by questionnaires is estimated to be around 70 – 90% (e.g. [ 14 ]). However, genetic studies are often hampered by the fact that psychopathologies are multifacetted and complex. A recent line of inquiry has started to use a 'bottom-up' approach, trying to decompose the complex phenotype (behavior) into a set of variables that are thought to represent more basic processes or traits. In this so-called endophenotypic approach the search is for neuro-behavioral vulnerability markers that are somewhere intermediate the genes and the disease [ 15 , 16 ]. Endophenotypic measures gathered in children can be used to assess genetic vulnerability to adult psychiatric disorders [ 17 ]. In this paper we will try to assess whether, and to what extent Stroop and/or flanker performance can qualify as genetic indicators for frontal abnormalities. The usesfulness of measures of Stroop performance in a genetically informative design has recently been demonstrated in a study [ 18 ] that compared Stroop performance among children suffering from ADHD, their unaffected sibs, and a group of controls. It was found that not only the children with ADHD, but also their unaffected sibs suffered more from Stroop interference than the controls. In a similar vein, it was found that not only euthymic bipolar and schizophrenic patients, but also their unaffected first-degree relatives suffered from increased Stroop interference, relative to a healthy control group [ 19 , 20 ]. However, another study [ 21 ] failed to find deteriorated Stroop performance in unaffected ADHD sibs. The usefulness of measures of flanker performance in a genetically informative design was demonstrated by a series of studies conducted by Fan and co-workers. Using a sample of healthy monozygotic (MZ) and dizygotic (DZ) twins, it was tested whether genetic variation contributed to variations in performance on basic attentional tasks [ 22 ]. These tasks were designed to tap distinct attentional brain networks (see also [ 23 ]). Of interest is performance on the flanker task, which was supposed to index the efficiency of the dopamine rich frontal executive network. Performance on this task indeed showed evidence of heritability. In a follow-up study, 200 subjects were genotyped, and were tested on a range of attention tasks. Modest associations were then found between genetic polymorphisms of several genes implicated in frontal (dys)function, such as drd4 and dat1 , and the efficiency of the frontal executive attention network [ 24 ]. Using the same twin methodology, it was also found [ 25 ] that performance on the flanker task was heritable. In addition, there was a correlation between flanker performance and IQ, and this correlation was completely mediated by a common set of genes. In this paper we ask whether variation in normal Stroop and flanker performance is caused by genetic variation. By using monozygotic twins, who share all their genetic material, and dizygotic twins, who share on average half of their segregating genes, the influence of genetic factors and environmental factors can be teased apart. If genetic effects are important, then members of monozygotic twin pairs will be more similar than members of dizygotic twin pairs in test performance. Conversely, if MZ twins show the same degree of resemblance as DZ twins, influences of environmental factors that are shared by both twins (e.g., the school or family environment) will be important. The contributions of additive genetic factors, shared environmental factors and unique environmental factors for explaining the variance observed for these measures can be explored using this twin design. If Stroop and/or flanker performance is found to be heritable, we will have a further genetic (in addition to neuropsychological) validation of these tasks with respect to frontal dysfunction. Furthermore, a high heritability of performance measures may ultimately help to unravel the genetic pathways of complex psychiatric traits. This paper is a follow-up to a previous paper, where we reported behavioral data on Stroop and flanker performance [ 11 ]. The current paper extends the previous one by investigating genetic effects on variation in performance. Results For the Stroop task, the data from 5 subjects (3 first born twins and 2 second born twins) could not be analyzed because they failed to comply with the instructions. Visual inspection of the data revealed that a few subjects had an extremely high Stroop interference score. Four subjects whose interference score was larger than 120 s were excluded from the analysis. Due to technical problems the data of 24 subjects for the flanker task were not stored or collected. Furthermore, there were 2 subjects who had an extremely high error score (> 20 errors out of 80 trials). These subjects were excluded from the analyses. Descriptives Table 1 shows the time to complete each of the three cards, separate for the first-born twins and the second born twins. The table reveals a clear increase in performance time from Card 1 to Card 2 to Card 3. The analysis of variance (ANOVA) for the first-born twins showed that there was a significant effect of card type, F(2, 278) = 1351.7, p < .001. The main effect of sex was not significant (p > .1), nor its interaction with card type. For the second born twins a near-identical pattern of results was found: a main effect of card type, F(2, 276) = 1236.2, p < .001, and no effects involving sex. Thus, we obtained a robust Stroop effect, and this was not affected by the sex of the subject. Accuracy data can be found in Table 2 . For the flanker task we found the following effects: For the first born twins the main effect of stimulus type was significant, F(1, 128) = 463.00, p < .001. Congruent stimuli yielded faster RTs than incongruent ones (556 vs. 662 ms). The main effect of sex was not significant (p > .1), nor its interaction with stimulus type. For the second born twin the main effect of stimulus type was significant, F(1, 132) = 556.02, p < .001. Again, congruent stimuli yielded faster RTs than incongruent ones (551 vs. 653 ms). Also, the main effect of sex was significant, F(1, 132) = 5.38, p < .05. Boys were somewhat faster than girls (587 vs. 618 ms). The same analyses done on the error rates yielded a comparable pattern of results. For the first born twins the main effect of stimulus type was significant, F(1, 128) = 84.86, p < .001. Congruent stimuli yielded more correct responses than incongruent ones (98.1% vs. 94.9%). In addition, the main effect of sex was significant, F (1, 128) = 4.57, p < .05, as was its interaction with stimulus type, F (1, 128) = 5.03, p < .05. These effects indicate that boys tend to respond somewhat less accurate than girls on incongruent trials. For the second born twins also the main effect of stimulus type was significant, F(1, 132) = 115.18, p < .001. Again, congruent stimuli yielded more correct responses than incongruent ones (98.8% vs. 93.9%). The main effect of sex was not significant, but the interaction was significant, F(1, 132) = 4.05, p < .05. Again, boys tended to respond somewhat less accurate than girls on incongruent trials. In order to test whether for the flanker task there was a trade-off between response speed and accuracy, we simply correlated RTs with accuracy, separately for the first born and second born twins. A possible speed-accuracy trade-off would manifest itself as a significant negative correlation between mean reaction time and percentage of errors. For the first born twins we found a significant positive correlation (r = .37, p < .001). But this correlation appeared to be due to a handful of subjects who were both quite slow and error prone. For the second born twins the correlation was small (r < .1), and not significant. Thus, we conclude that in our sample there was no evidence of a speed-accuracy trade-off. A similar analysis was done for the Stroop task. We correlated the average completion time of Card 3 with the number of errors commited with Card 3. We did not correlate speed with the number of corrections, because these measures are not independent. For both the first born twins and the second born twins we found a significant positive correlation (r = .31, and r = .21, respectively. p's < .05). Thus, subjects who were slow also tended to be inaccurate. However, the distribution of the number of errors was rather skewed (most subjects made 0 or 1 errors), which makes it difficult to interpret these correlations. So, similar to the flanker task, we conclude that there was no evidence of a speed-accuracy trade-off. Genetic analyses Table 3 shows twin correlations of times to complete Card 1, 2 and 3, and of the interference effect (Stroop effect; difference between Card 3 and Card 2). For the 3 cards, a very consistent pattern is seen: MZ correlations are high, around .7, and DZ correlations are approximately half, implying the existence of genetic influences and unique influences, with a heritability of around 70%. The twin correlations of the interference effect are somewhat lower, probably because difference scores tend to have a lower reliability[ 11 ]. MZ correlations were around .5 and DZ correlations were lower, pointing to a heritability of about 50%. Table 4 shows the twin correlations for the overall RT and the flanker interference effect (RT [incongruent] minus RT [congruent]). The pattern of twin correlations is hardly indicative of genetic effects on performance. Even though the MZ twin correlations on response speed were higher than the DZ correlations, the highest twin correlation was obtained with the DOS zygosity group. Furthermore, the highest twin correlation for the flanker effect was obtained with the DZF group. In addition, the twin correlations within the monozygotic groups were low. We therefore conclude that there were no genetic effects of flanker performance. The twin correlations obtained with the Stroop task thus appeared strongly indicative of genetic effects. Using structural equation modelling, these effects were formally tested. But prior to testing we had to establish whether there were significant differences in variances across sex and zygosity, since one of the assumptions underlying structural equation modelling is the assumption of homogeneity of variances. We conducted the Levene test on all variables, separately for the first-born and the second born twins. For none of the variables the Levene test yielded a significant effect, with the possible exception of the completion time of Card 1 for the second born twins, F(3, 136) = 2.664, p = 0.05. So, we felt it was legitimate to use structural equation modelling to test for genetic effects. Additional file 1 shows the results. The full ACE model, which allowed for sex differences in parameter estimates fitted well to the data (χ 2 ranged from 7.732 to 8.883, df = 9, p's ranged from .448 to .561), with the possible exception of Card 1 (χ 2 = 15.339, df = 9, p = 0.082), although the most parsimonious model for Card 1 fitted slightly better (χ 2 = 17.240, df = 13, p = 0.189). Sex differences in parameter estimates could be discarded from the models, although they were almost significant for Card 3 (χ 2 = 7.652, df = 3, p = 0.054). Common environmental influences were not necessary to describe the data, but additive genetic influences explained a significant part of the variance in all 4 variables. Heritabilities for the 3 cards were 75%, 70% and 74% respectively, with confidence intervals indicating that these were well above half the variance. Heritability of the interference effect was 49%, with a 95% confidence interval between 29 and 64%. Discussion In this study we assessed the heritability of performance on two well-known response interference tasks: the color word Stroop task and the Eriksen flanker task, using a large sample of 12-year old twins. The aim was to test whether Stroop performance and/or flanker performance could qualify as a suitable endophenotype for genetic frontal abnormalities, such as ADHD. First, we found that the time to complete each of the three cards was highly heritable. This may represent a general factor related to processing speed and/or rapid naming speed. Of greater importance was the finding that the interference score (the difference between completion times of Card 2 and 3) was also heritable: nearly 50% of the variation in performance was due to genetic variation. Thus, the efficiency of the network that deals with response conflict is -in part- under genetic influence. For the theoretically similar flanker task, however, there was little evidence of genetic influences on performance. Even though the MZ twin correlations on response speed were higher than the DZ correlations (as in [ 25 ]), the highest correlation was observed for the DOS twins, for which we have no explanation. In addition, there was no evidence of genetic influences on variations in the size of the flanker effect. Variation in performance thus simply appeared to be due to noise. This latter finding is at odds with a previous study where it was found that variation in the size of the flanker interference effect was 89% due to variations in genes [ 22 ]. This discrepancy could of course be due to minor differences between task versions. For example, the flanker test adopted by [ 22 ] was embedded in a visual orienting paradigm. But it could also be the case that genetic effects on flanker performance are somehow age specific. Our age group was 12 years old, whereas Fan et al.'s [ 22 ] age group was between 14 and 42 years of age. It is well known that heritabilities of different traits vary with age. For example, the heritability of IQ is known to steadily increase with increasing age (e.g. [ 26 ]), and it could be the case that genetic effects on flanker performance only emerge at a later age. Finally, it could be that Fan et al. [ 22 ] have obtained a false positive result, due to their somewhat modest sample size (26 MZ pairs and 26 DZ pairs). The question now is whether Stroop and flanker performance can qualify as a suitable endophenotype of frontal pathologies. Recently a list of 5 criteria was compiled that are ideally possessed by endophenotypes [ 15 ]. Criteria 3 to 5 deal with the relationship between phenotype and endophenotype. In brief, there should be a high correlation between the phenotype and endophenotype, this correlation should be based in genetics, and the correlation should be theoretically meaningful. In our Introduction we have briefly touched upon the relationship between performance on response interference tasks and high-level phenotypes, such as the efficiency of the frontal executive network. Our selective review of the literature indicated that there was a clear genetic link between Stroop performance and frontal pathologies, whereas evidence for a the genetic link between flanker performance and frontal pathologies was less conclusive. Criteria 1 and 2 of [ 15 ] state that the endophenotype should be reliable and heritable. With respect to heritability, we have demonstrated that – at least for this age group – there is strong evidence for genetic influences on Stroop performance, but not on flanker performance. With respect to reliability, we have no test-retest data but we can assess reliability by examining the MZ correlations, because these correlations provide a lower limit to reliability [ 27 ]. Inspection of Table 2 reveals that the Stroop performance measures are characterized by high MZ correlations, which implies high reliabilities. The flanker performance measures reported in Table 3 , however, revealed quite low MZ correlations. This finding, in combination with a low split-half reliability (reported in [ 11 ]) leads us to conclude that Stroop performance provides a more reliable measure than flanker performance. Conclusions We have found evidence for the existence of strong genetic effects on conflict resolution, although the effects are task dependent. We conclude that performance on the Stroop test yields a better endophenotype for frontal (dys)function than performance on the flanker task. So, despite the overlapping regions of brain activation in the Stroop and flanker tasks, and despite their face-value similarity, we believe that these interference tasks differ in important, yet unknown ways. Methods Subjects The subject group consisted of a group of 290 12-year old twins. There were 33 monozygotic male pairs (MZM), 24 dizygotic male pairs (DZM), 45 monozygotic female pairs (MZF), 16 dizygotic female pairs (DZF), and 27 opposite sex pairs (DOS). The twins participate in a longitudinal study of attention and attention problems. The twins are registered in the Netherlands Twin Registry (NTR), which is hosted by the Vrije Universiteit of Amsterdam [ 28 ]. The twins were randomly selected from the NTR subject pool. None of the children suffered from severe mental or physical impairments. Twin pairs were first asked in writing whether they were willing to participate in the study. Permission was also asked of the parents or caretakers. If permission was granted, the families received further information on the study, and were invited to come to the campus site to do the tests. The study was approved by the local Ethics Committee, and on the day of testing the children and their parents / legal representatives signed an informed consent form. Procedure Twins were tested on the same day. The children performed a range of neuropsychological tests that lasted approximately 4 hours per child. All tests were performed in the same order. Here we focus on the Stroop Color and Word Test [ 1 ] and the Eriksen flanker test. In the Stroop test, subjects complete 3 cards, each with 10 columns of 10 items. Subjects have to name aloud the items on each card, from the top-left corner to the bottom-right corner. Card 1 involves naming the words 'red', 'green', 'yellow' and 'blue' printed in black ink. Card 2 involves naming the colors of squares that are printed in different colors. Card 3 involves naming the ink color that the words 'red', 'green', 'yellow' and 'blue' are printed in. In Card 3 word content and ink color never match, i.e., all color words are incongruent. Speed and accuracy are stressed in the instructions. Each card is scored as the time (in seconds) to complete the card. Time is recorded by the experimenter using a stopwatch. The experimenter also recorded the number of errors (the wrong item is named, or an item is skipped), and the number of corrections (the wrong item is named, but the subject immediately corrects himself afterwards). Note that in this task we do not have trial-by-trial information on speed and accuracy; we only have summary scores across the entire Stroop card. In the Eriksen flanker task subjects were presented with a horizontal array of 5 arrows. Subjects were instructed to attend to the direction of the center arrow, and ignore the 4 flanking ones. Subjects had to press the left key to a left facing center arrow, and the right key to a right facing center arrow. The flanking arrows could either all point in the same direction as the target arrow (e.g., < < < < <; congruent condition), or they all pointed in the opposite direction (e.g., < < > < <; incongruent condition). Subjects received 40 congruent and 40 incongruent trials in a random order. For each trial, the computer stored the RT and whether the correct key was pressed. The number of valid cases in the flanker task is lower than in the Stroop task. This was due either to practical problems (some data were not collected or not stored on the computer), or because some subjects made an extremely high number of errors (less than 75 % correct). Further details of the Stroop task and the flanker task can be found in [ 11 ]. Data analysis: test of means In order to test for the effects of card type and sex on Stroop performance we performed an analysis of variance on the completion times with card type (1, 2, and 3) as within-subjects factor, and sex (males and females) as between subjects factor. For the flanker task, we performed an ANOVA on the mean correct response times and on the mean percentages correct, with stimulus type (congruent vs. incongruent) as within-subjects factor, and sex (males and females) as between subjects factor. The same analysis was done for the error rates. These analyses were done separately for the first-born twins and the second born twins, because twins within a family do not yield independent data. We adopted an alpha-level of .05. From the completion times we also calculated the size of the Stroop effect (i.e., the interference score), which is simply defined as Completion time Card 3 minus Completion time Card 2. There exists another method to determine the size of Stroop effect, which also takes into account the completion times of Card 1 [ 6 , 29 ]. But a preliminary analysis revealed that this method and the method used by us yielded virtually identical results, so we present no data based on the method proposed by [ 29 ]. Data analysis: genetic analysis Data from monozygotic (MZ) and dizygotic (DZ) twins were used to decompose the variance in performance on the both tasks into a contribution of the additive effects of genes, environmental influences that are shared by twins living in the same family, and environmental influences that are not shared by twins. Resemblance between MZ twins is an effect of both their common genetic constitution and their shared environment. Because DZ twins share on average half of their segregating genes, the shared environment contributes fully, but genetic factors only partly to their resemblance. Therefore, if the degree of MZ resemblance on some measure is higher than the degree of DZ resemblance we have strong evidence for the influence of genetic effects. Pearson correlations were calculated for the different measures between first born and second born twins for all zygosity groups. A first indication of the heritability can be derived by doubling the difference between correlations for MZ twins and those for DZ twins [ h 2 = 2(r MZ -r DZ )] [ 27 ]. A structural equation modeling approach as implemented in Mx [ 30 ] was used for genetic data analysis. The dependent variables were analyzed using a model including three latent independent factors – additive genetic factors (A), shared or common environmental factors (C) and non shared or unique environmental factors (E) – that influence variation in a particular phenotypic measure of attention (P). A path diagram of an ACE model is presented in Figure 1 . Because these latent factors are standardized to have a variance of 1.0, the double-headed arrow connecting them represents the correlation among them. The correlation between genetic effects in twin 1 and twin 2 is 1.0 for MZ twins and 0.5 for DZ twins. These between-twin correlations are represented as fixed parameters in the Mx model, as is the correlation between the common environmental factors (shared by both twins of a twin pair), which is fixed to unity for both twin groups. Parameters a , c and e represent the influence of genes, common environment and unique environment on the phenotypes (P) of twin 1 and twin 2. The total variance of the phenotype (P) = a 2 + c 2 + e 2 . The heritability ( h 2 ) is calculated as a 2 /V P . To test if parameter estimates are equal for boys and girls the fit of a model with constrained parameter estimates for a , c and e to be equal across sexes was compared to one in which they were allowed to vary. After this, the significance of c and a was investigated by dropping them one by one from the model and comparing the fit of a full model to that of a reduced model. The chi-squared statistic is computed as twice the difference between the likelihood for the full model (-LL 0 ) and that for a reduced or constrained model (-LL 1 ) (χ 2 = 2 × (LL 0 -LL 1 )) and is tested against the difference in degrees of freedom between the two models. Authors' contributions JFS drafted the manuscript, and performed the descriptive analyses. GCMvB performed the genetic analyses. TJCP collected the data. FCV and DIB conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Model fitting and parameter estimates Model fitting and parameter estimates of heritability ( h 2 ), common environmentability ( c 2 ), and unique environmentability ( e 2 ), for completion times of the three cards, and of the Stroop interference score. Best fitting models are shown in bold. Parameter estimates are given in percentages (95% confidence interval in brackets). Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544562.xml
554771
Polymorphisms in immunoregulatory genes and the risk of histologic chorioamnionitis in Caucasoid women: a case control study
Background Chorioamnionitis is a common underlying cause of preterm birth (PTB). It is hypothesised that polymorphisms in immunoregulatory genes influence the host response to infection and subsequent preterm birth. The relationship between histologic chorioamnionitis and 22 single nucleotide polymorphisms in 11 immunoregulatory genes was examined in a case-control study. Methods Placentas of 181 Caucasoid women with spontaneous PTB prior to 35 weeks were examined for histologic chorioamnionitis. Polymorphisms in genes IL1A , IL1B , IL1RN , IL1R1 , tumour necrosis factor ( TNF ), IL4 , IL6 , IL10 , transforming growth factor beta-1 ( TGFB1 ), Fas ( TNFRSF6 ), and mannose-binding lectin ( MBL2 ) were genotyped by polymerase chain reaction and sequence specific primers. Multivariable logistic regression including demographic and genetic variables and Kaplan-Meier survival analyses of genotype frequencies and pregnancy outcome were performed. Results Sixty-nine (34%) women had histologic evidence of acute chorioamnionitis. Carriage of the IL10 -1082A/-819T/592A ( ATA ) haplotype [Multivariable Odds ratio (MOR) 1.9, P = 0.05] and MBL2 codon 54Asp allele (MOR 2.0, P = 0.04), were positively associated with chorioamnionitis, while the TNFRSF6 -1377A/-670G ( AG ) haplotype (MOR 0.4, P = 0.03) and homozygosity for TGFB1 -800G/509T ( GT ) haplotype (MOR 0.2, P = 0.04) were negatively associated. Conclusion These findings demonstrate that polymorphisms in immunoregulatory genes IL10 , MBL2 , TNFRSF6 and TGFB1 may influence susceptibility to chorioamnionitis.
Background Acute infection of the amniotic fluid and chorioamnion is a common cause of preterm delivery before 34 weeks gestation. Consistent with previous reports [ 1 ], an Australian study found that 31% of preterm births (PTB) before 34 weeks had histologic chorioamnionitis [ 2 ]. An inverse relationship between delivery gestation and histologic chorioamnionitis was shown with histologic chorioamnionitis identified in 66% of births before 25 weeks. Chorioamnionitis usually results from ascent of pathogenic organisms such as Group B streptococcus and Escherichia coli [ 3 ], from the lower genital tract. A T helper cell type1 (Th1) cytokine host response occurs with the production of pro-inflammatory cytokines such as (IL-1β and TNF) produced mainly by activated monocytes/macrophages in amnion, chorionic and decidual tissue. The increased levels of these cytokines trigger prostaglandin F2 and E2 biosynthesis by the decidua and amnion, inducing uterine contractions and labour [ 4 ]. In the setting of infection, IL-1 and TNF have a synergistic effect on the acute phase cytokine IL-6 with increased production in trophoblasts and chorion and to a lesser extent decidua [ 5 ]. The intensity and duration of the inflammatory response is modulated by T helper cell type 2 cytokines and anti-inflammatory mediators such as IL-4, IL-10, IL-1 receptor antagonist (IL-1ra) and transforming growth factor beta-1 (TGFβ-1). The innate host defense system is also an important factor in pregnancy and the body's response to infection. Mannose-binding lectin (MBL), a protein of the complement system, opsonises pathogens independently of antibody for phagocytosis and influences inflammatory pathways [ 6 ]. There are several common polymorphisms in the promoter and coding region of the MBL2 gene that profoundly influence circulating levels of functional, multimeric MBL. These genetic variants and low MBL levels are associated with risk and severity of infection in a variety of clinical contexts [ 6 ]. Fas, a cell surface receptor of the TNF/nerve growth factor superfamily, mediates apoptosis after binding Fas ligand [ 7 ]. Activated leucocyte populations, including T cells, express high amounts of Fas and evidence suggests that trophoblast Fas-mediated apoptosis is increased in women with chorioamnionitis [ 8 ]. The immune system operates as a complex regulatory network of cytokines and other mediators with a degree of redundancy that changes as pregnancy progresses [ 9 ]. This makes it difficult to assess the importance of single cytokines in inflammation and response to infectious agents when examined in isolation. The genes encoding these mediators contain coding and non-coding polymorphisms that influence level or function of the encoded mediators [ 10 ]. These polymorphisms may influence susceptibility to chorioamnionitis and adverse pregnancy outcome. Single nucleotide polymorphisms (SNPs) for IL1 +3953 and TNF -308 are associated with human chorioamnionitis in in vitro studies [ 11 ] and carriage of the TNF -308 A allele has been reported as a risk factor for clinical chorioamnionitis in pregnant women [ 12 ]. However these reports were restricted to small numbers of cytokine genes and did not include the analysis of polymorphisms in molecules important in apoptosis and host defence. Examination of a range of immunoregulators and SNPs concurrently is more likely to identify which genes and alleles are of primary importance in response to intra-uterine infection. We hypothesized that polymorphisms in immunoregulatory genes may influence susceptibility to chorioamnionitis and subsequent preterm labour. To examine this hypothesis, we investigated multiple polymorphisms in immunoregulatory genes in a population of Australian Caucasoid women experiencing spontaneous PTB before 35 weeks and histologic chorioamnionitis. Methods This case-control study was approved by the Women's and Children's Hospital Ethics and Research Committee. A total of 368 unrelated women of child-bearing age who had obstetric management at the Women's and Children's Hospital (WCH) and history of PTB before 35 weeks gestation were identified from the hospital data base. All women were approached in the first instant by letter of introduction and invited to participate in the study. For women living in the Adelaide metropolitan area, the study nurse made a home visit to those who volunteered for eligibility assessment. Geographical distance excluded 50, 66 did not respond, 40 were not eligible and 31 declined. Of the women enrolled some had more than one PTB. The PTB fitting the study criteria and closest to the enrolment date became the index birth. Written consent was obtained and a 9 ml peripheral blood sample was collected for isolation of leucocytes and DNA extraction. 181 women with a history of spontaneous preterm labour and subsequent PTB between 20 and 35 weeks in the index pregnancy had histological examination of the placenta performed. Histopathology staff members were unaware of the genetic results. Clinical data recorded included gestational age at delivery (calculated by ultrasonography at approximately 18 weeks). Obstetric factors known to place a woman at risk of preterm labour or preterm prelabour rupture of membranes (PPROM) (defined as prelabour rupture of membranes 3 or more hours before preterm delivery) were also recorded. Placental examination and swabbing was performed by staff histopathologists [ 13 ] and included examination of the placental tissue in blocks, extra-placental membranes in a "roll" (amnion, chorion and decidua) and the umbilical cord in cross section. Chorioamnionitis was defined as a dense polymorphonuclear leukocyte infiltration of the chorionamniotic component of the placenta and membranes (amnion, chorion, amniochorial membranes and/or chorionic plate (top part of the fetal side of the placenta). Inflammatory reaction confined to either the decidua or subchorial intervillous space of the placental disc, without infiltration of the chorion or amnion was recorded as negative chorioamnionitis [ 3 ]. Funisitis was defined as inflammation in one or more of the umbilical cord vessels (vasculitis) with or without inflammation in the Wharton's jelly (the supporting soft tissue around the vessel in the cord). The definition of acute chorioamnionitis included funisitis and/or amnionitis. In both groups a number of women had PPROM, occasional bleeding during pregnancy, recorded clinical chorioamnionitis or a history of prior term births. Tocolytic therapy and intravenous antibiotics were used according to obstetric protocols. All participants were Caucasoid, unrelated and at least 18 years of age at enrolment. Women with diagnosed autoimmune disease, pre-eclampsia, in vitro fertilisation treatment, confirmed uterine malformations, no spontaneous labour, ultrasound-confirmed fetal abnormality, insulin dependent diabetes, or multiple pregnancy were excluded. Demographic characteristics, a comprehensive medical and obstetric history, age, any smoking, alcohol consumption, substance use and clinical chorioamnionitis (defined as any three of the following – white blood cell >1500 × 10^9/L, three consecutive C reactive protein readings ≥ 15mg/L, pyrexia ≥ 38°C, uterine pain or tenderness) were verified from case records. Clinical chorioamnionitis was not used as a subgroup for polymorphic analysis because this did not include women with sub-clinical infection. Since the risk of chorioamnionitis is not as great in PTB after 35 weeks, women with PTB between 35 and 37 weeks were not enrolled [ 2 ]. Genes were chosen if reported to be associated with inflammatory disease[ 10 ] and if existing functional data suggested a role in chorioamnionitis and preterm delivery[ 5 , 11 ]. SNPs were selected where the minor allele has a frequency >10%. Several genes were known to contain polymorphisms at multiple sites that are of functional importance (e.g. the 5'and coding variants of MBL2 ) [ 14 ], thus multiple variants were genotyped and the haplotype recorded. SNPs in the IL1 gene cluster, TNF , IL4 , IL6 , IL10 , TGFB1 , TNFRSF6 , and MBL2 genes and their relationship with histologic chorioamnionitis were therefore examined in a cohort of women with spontaneous PTB. Peripheral blood samples were blinded and tested at the Australian Red Cross Blood Service, Adelaide (ARCBS). Genomic DNA was extracted from ethylenediaminetetraacetic acid anticoagulated venous blood using standard methods as previously described [ 15 ]. SNPs for TNF (+488,-238,-308) [ 16 ], IL1A (-889), IL1B (+3962, -511), IL4 (-590), IL10 (-1082, -819, -592), TGFB1 (-800, -509) [ 17 ], MBL2 [-550, -221, codon 52 (Arg → Cys), codon 54 (Gly → Asp), codon 57 (Gly → Glu)] [ 18 ], IL6 (-174), IL-1 receptor antagonist IL1RN (+11100), IL-1 type 1 receptor IL1R1 (+970) and TNFRSF6 (-1377, -670) [ 19 ] were genotyped using the polymerase chain reaction and sequence specific primers (PCR-SSP). Multiple polymorphisms were examined in the IL10 , TNFRSF6 , TNF , TGFB1 and MBL2 genes by PCR haplotyping. The term "haplotype" denotes ordered combination of alleles on a single chromosome, and is used where multiple polymorphisms are genotyped by PCR-haplotyping for a single gene. For PCR haplotyping combinations of forward and reverse allele-specific-primers were used to directly amplify alleles on the same chromosome [ 20 ]. This technique also provides a degree of redundancy, in that multiple reactions genotype a polymorphism, thus confirming results. All PCR-SSP reactions have been extensively used in the same laboratory. Internal control primers for conserved regions of the DRB1 (major histocompatibility complex, class 11 DR beta 1) and APC (adenomatous polyposis coli) genes were included in every PCR mix to verify successful amplification. All genes except the IL1A had a repeat genotyping rate of less than 1%. The IL1A had a repeat rate of 10%. The sample size was calculated by using the statistical software package EpiInfo 6 (Version 6.04d, Centre for Disease Control and Prevention, Atlanta, GA). We estimated the study had sufficient power (1-β = 80%, P = 0.05) to detect a minimum of 17% difference in genotype frequency between chorioamnionitis and no chorioamnionitis for frequencies of approximately 10% (ie. 10% versus 27%). For a genotype frequency of 3% the minimum detectable difference was 10% (3% versus 13%). Allele and haplotype frequencies for polymorphisms were determined by direct counting from each chromosome. Frequencies of genetic variants were recorded as allele (proportion of positive chromosomes), genotype (proportion possessing homozygous or heterozygous combinations of alleles), and 'carriage' of an allele or haplotype (either in homozygous or heterozygous state). Alleles were said to be in Hardy-Weinberg equilibrium if the observed genotype frequencies did not differ significantly (P > 0.05) from those expected when analysed by Chi square. Gene variant frequencies were compared using 2 × 2 contingency table analysis and the Chi square test for independence using EpiInfo 6. Mean birth weights were compared by t -test. Multivariable analysis was performed by stepwise backwards multiple logistic regression using SYSTAT 9.0 software (Systat version 1999 SPSS Inc). The basic model included genetic variables with P values < 0.15 obtained in univariate analyses. In each multivariable model, only one genotype or haplotype of each gene was selected for analysis. Selection of genotype/haplotype for the model was based on strength of association, the previously reported importance of a gene variant and homozygosity of alleles. At each backward step in the regression analysis P values ≥ 0.07 were rejected. Kaplan-Meier survival analysis (SYSTAT 9.0) was used to identify cumulative probability in maintaining pregnancy by gestation stratified by chorioamnionitis and gene variables. Breslow-Gehan Chi square was used to weight each gestational period by the total number at risk at that time so that earlier gestations receive greater weight than later gestations. Results Sixty-nine (38%) women had histologic evidence of acute chorioamnionitis, and 112 (62%) had no such evidence. There were no significant differences between those with and without histologic chorioamnionitis for maternal age, race, history of smoking, alcohol consumption, substance use, new partner, cervical cerclage, placenta praevia after 22 weeks with bleeding, history of two or more PTBs before 35 weeks, miscarriage between 12 and 20 weeks or PPROM (Table 1 ). Histologic chorioamnionitis was positively associated with clinical chorioamnionitis [Univariate Odds Ratio (OR) 6.1, 95% Confidence Intervals (CI) 2.4–17, uncorrected P ≤ 0.0001], PTB before 29 weeks (OR 7.6, CI 3.4–18, P ≤ 0.0001) and positive placental culture (OR 3.0, CI 0.9–11, P = 0.04). Table 1 Demographic and clinical characteristics of women with preterm birth before 35 weeks. Variable Histologic Chorio-amnionitis n = 69 No Histologic Chorio-amnionitis n = 112 Significant OR (95% CI), and Uncorrected P values Maternal age (years) 29 (16–41) 30 (17–44) Delivery gestation (weeks) 28 (20–34) 30 (23–34) * Mean birth weight 1399 (± 676) 1897 (± 504) < 0.0001 Maternal age ≥ 35 years 11 (16%) 18 (16%) ns Maternal age ≤ 17 years 2 (3%) 2 (2%) ns Any smoking 22 (31%) 29 (26%) ns Any alcohol 44 (64%) 73 (65%) ns Any substance use 7 (10%) 10 (9%) ns Partner change for 14 (20%) 20 (18%) ns pregnancy Cervical cerclage 3 (4%) 2 (2%) ns Previous mid-trimester 4 (6%) 12 (11%) ns miscarriage (12–20 weeks) 2 or more preterm births 11 (16%) 20 (18%) ns Placenta praevia after 22 weeks with bleeding 1 (1%) 2 (2%) ns Documented clinical chorioamnionitis 22 (32%) 8 (7%) 6.1 (2.4–17), < 0.001 PPROM ≥ 3 hours 39 (57%) 74 (66%) ns Gestation < 29 weeks 33 (48%) 12 (11%) 7.6 (3.4–18), < 0.001 Documented positive placental culture 10 (14%) 6 (5%) 3.0 (0.9–11), 0.04 Neonatal sepsis 10 (14%) 0 19 (3 – 823), < 0.001 Neonatal death 6 (9%) 3 (3%) ns Post partum endometritis 5 (7%) 3 (3%) ns Descriptive statistical results are mean (± SD) or number (%) as appropriate, and maternal age and gestation as median (range). Characteristics are from index birth. PPROM = Preterm premature rupture of membranes 3 hours or more before delivery < 35 weeks. Univariate analyses were tested using Chi square analysis of 2 × 2 table, and uncorrected P values, Odds Ratio (OR), and 95% Confidence Intervals (CI), P values ≤ 0.05 are presented. Mean birth weights were tested by two-sample t -test, assuming unequal variances. * See Figure 1, ns, not significant. All alleles were in Hardy-Weinberg equilibrium. All gene variants were tested for associations between the two groups. In preterm women without histologic chorioamnionitis frequencies for the IL10 -1082A/-819T/-592A ( ATA ) haplotype [ 17 ], MBL2 codon 54Asp (the MBL 'B' allele) [ 21 ], TNFRSF6 -1377A/-670G ( AG ) haplotype [ 22 ] and TGFB1 -800G/-509T ( GT ) homozygosity [ 17 ] were similar to reported Caucasian controls. In a comparison of women with chorioamnionitis and those without chorioamnionitis a comprehensive listing of gene carriage, alleles, genotype and haplotype frequencies is provided ( Additional file 1 ). Univariate analysis revealed that carriage of the IL10 -1082A/-819T/-592A ( ATA ) haplotype (present in 49% of women with chorioamnionitis vs 33% without chorioamnionitis, OR 2.0, P = 0.03) and the IL10 -819T and -592A alleles (50% versus 33%, OR 2.0, P = 0.03) were positively associated with histologic chorioamnionitis, as was carriage of the MBL2 54Asp (B allele) (39% versus 25%, OR 1.9, P = 0.04) (Table 2 ). Although the frequency of the common TNFRSF6 -1377 G/G genotype was higher in women with histologic chorioamnionitis (0.87 versus 0.73, OR 2.4, P = 0.03), the variant TNFRSF6 -1377A/-670G ( AG ) haplotype (OR 0.4, P = 0.03) and homozygosity for the TGFB1 haplotype -800G/-509T ( GT ) (OR 0.2, P = 0.03) were negatively associated with chorioamnionitis (Table 2 ). No other significant associations with gene variant frequencies were seen in univariate analysis. Table 2 Genes, alleles, haplotypes and genotypes in women with preterm birth before 35 weeks and chorioamnionitis. Gene Histologic Chorio-amnionitis n = 69 No Histologic Chorio-amnionitis n = 112 Univariate analysis OR (95% CI) Uncorrected P value Multivariable analysis OR (95% CI), P value IL10 Allele -819 T 34 (0.50) 37 (0.33) 2.0 (1.0–3.8), 0.03 -592 A 34 (0.50) 37 (0.33) 2.0 (1.0–3.8), 0.03 Haplotype -1082A/-819T/-592A 34 (0.49) 37 (0.33) 2.0 (1.0–3.8), 0.03 1.8 (1.0–4.0), 0.05 Genotype -819 C/T 28 (0.41) 28 (0.25) 2.1 (1.0–4-1), 0.03 -592 A/C 28 (0.41) 28 (0.25) 2.1 (1.0–4-1), 0.03 MBL2 Codon 54Asp ('B') 27 (0.39) 28 (0.25) 1.9 (1.0–3.9), 0.04 2.0 (1.0–4.0), 0.04 TNFRSF6 Allele -1377A 9 (0.13) 30 (0.27) 0.4 (0.2–1.0), 0.03 Haplotype -1377A/-670G 9 (0.13) 30 (0.27) 0.4 (0.2–1.0), 0.03 0.3 (0.2–0.9), 0.03 Genotype -1377 G/G 60 (0.87) 82 (0.73) 2.4 (1.0–6.3), 0.03 TGFB1 Haplotype -800G/-509T 2 (0.03) 14 (0.13) 0.2 (0.02–1.0), 0 .03 0.2 (0.04–0.9), 0.04 Homozygosity Data are presented as number and frequencies of alleles, genotypes and haplotypes. Univariate analyses were tested using Chi square analysis of 2 × 2 table, and uncorrected P values, Odds Ratio (OR), and 95% Confidence Intervals (CI), P values ≤ 0.07 are presented. Multivariable analyses included genetic variables with P values ≤ 0.15 using backward step multiple logistic regression. Multivariable model included carriage of IL10 ATA , MBL2 codon 54Asp, TNFRS6 -1377A/-670G ( AG ) haplotype, and homozygous TGFB1 -800G/-509T, TNF +488 G/A, and IL1R1 +970 T/T genotypes. MBL2 codon 54Asp (allele 'B') is a single nucleotide polymorphism in exon 1 codon 54 (Gly/Asp). TNFRSF6 is the gene symbol for Fas; ns, not significant. For confirmation of independence of association, a multivariable model was used including univariate variables with P values ≤ 0.15. This model included the above genetic variables (Table 2 ) plus TNF +488 G/A genotype (25% versus 15%, OR 1.8, P = 0.11) and IL1R1 +970 T/T (6% versus 14%, OR 0.4, P = 0.08) from Additional file 1 . Multivariable analysis confirmed that carriage of the variant IL-10 ATA haplotype [Multivariable Odds Ratio (MOR) 1.9, P = 0.05)] and MBL2 codon 54Asp (MOR 2.0, P = 0.04) were independently associated with histologic chorioamnionitis, while homozygosity for the variant TGFB1 -800G/-509T ( GT ) haplotype (MOR 0.2, P = 0.04), and carriage of TNFRSF6 -1377A/-670G ( AG ) haplotype (MOR 0.4, P = 0.03) remained negatively associated with histologic chorioamnionitis. The remainder of SNPs tested showed no significant differences. To examine the cumulative probability of maintaining a pregnancy and effect of chorioamnionitis on gestation at delivery in the overall cohort, Kaplan Meier survival analysis was performed. Figure 1 shows that chorioamnionitis was more common in PTB at earlier gestations (Breslow-Gehan, Chi square 36 with 1 degree of freedom, P ≤ 0.001). In addition, the cumulative probability of maintaining a pregnancy was stratified on the carriage of IL10 ATA haplotype (Figure 2 ). Women with this IL10 ATA haplotype were more likely to deliver at earlier gestations (Breslow-Gehan, Chi square 3.5 with 1 degree of freedom, P = 0.06) (Figure 2 ). There was no evidence of the effect of MBL2 , TGFB1 , or TNFRSF6 SNPs on gestation of PTB. Figure 1 Relationship of chorioamnionitis with gestation at delivery. Kaplan-Meier survival estimates in 181 women where all gestations proceeded to delivery < 35 weeks. Effect of chorioamnionitis (C) versus no chorioamnionitis (NC), P < 0.001. Significance was tested using Breslow-Gehan Chi square analysis with 1 degree of freedom. Figure 2 Relationship of IL10 ATA haplotype with gestation at delivery. Kaplan-Meier estimates in 181 women where all gestations proceeded to delivery before 35 weeks. Effect of IL10 ATA haplotype versus No ATA , P = 0.06. Significance was tested using Breslow-Gehan Chi square analysis with 1 degree of freedom. Discussion These results extend prior reports investigating gene polymorphisms and the pathogenesis of PTB and chorioamnionitis [ 11 ]. In this study we found significant positive associations between the carriage of the IL10 ATA haplotype, MBL2 codon 54Asp and histologic chorioamnionitis, and negative associations with TNFRSF6 and TGFB1 . This is the first study to investigate TNFRSF6 and TGFB1 in combination with such a wide range of cytokine SNPs in histologic acute chorioamnionitis. Also, this is one of the few studies that have used multivariable and survival analyses for investigating adverse pregnancy outcomes such as chorioamnionitis and PTB. The differences in genotype distribution for IL10 and MBL2 may be of biologic importance in the pathogenesis of histologic chorioamnionitis. The function of IL-10 in pregnancy is of particular interest because of its role in the homeostatic control of an inflammatory immune response [ 23 ]. Although there have been conflicting reports the majority suggest that the IL10 ATA promoter haplotype and alleles comprising the haplotype are associated with low protein levels in Caucasians [ 24 ]. Low levels of IL-10 have been associated with PTB in rat models [ 25 ]. In a pregnant rhesus monkey model, combined intra-amniotic and maternal systemic IL-10 administration inhibited IL-1β-induced preterm uterine contraction [ 26 ]. Survival analysis showed that the IL10 ATA haplotype may influence the timing of PTB. IL-10 antagonises the synthesis and actions of pro-inflammatory cytokines such as IL-1 [ 27 ]. In chorioamnionitis, diminished IL-10 levels may provide an ineffective counter regulatory response to elevated maternal prostaglandin stimulated by the pro-inflammatory cytokines IL-1, IL-6 and TNF. Elevated levels of these pro-inflammatory cytokines in utero may serve as a maternal [ 28 ] and fetal [ 29 ] signal for the onset of premature parturition. The higher proportion of early PTB in women with the IL10 ATA haplotype may also be a reflection of failure of tocolysis in the presence of infection [ 26 ]. Our finding that the carriage of the IL10 ATA haplotype is more common in women with histologic chorioamnionitis suggests that this haplotype may be important in the pathogenesis of histologic chorioamnionitis and subsequent PTB. MBL is an important innate defense molecule active against a broad range of bacterial, viral, fungal and protozoan pathogens [ 30 ]. The MBL2 variant alleles in exon 1 correlate with decreased circulating MBL. The single nucleotide substitution of an adenine for a guanine in codon 54 results in replacement of aspartic acid for a glycine in the MBL [ 31 ]. Functional MBL is multimeric, consisting of tetramers, pentamers and hexamers of triplets of the basic MBL peptide. This higher order structure is necessary for high avidity interactions between MBL carbohydrate recognition domains and repeated oligosaccharide moieties on pathogens. The MBL codon 54Asp 'B' allele disrupts assembly of MBL2 peptide trimers resulting in lower stability and serum levels [ 32 ]. Low serum levels of MBL and MBL2 genetic variants are associated with infection in a range of contexts including recurrent infection in children and adults [ 6 , 14 ], recurrent vulvovaginal candidiasis [ 33 ], recurrent miscarriage [ 34 ] and autoimmunity in common variable immunodeficiency disease [ 18 ]. The association between MBL2 codon 54Asp and histologic chorioamnionitis suggests that this allele may result in low MBL levels, impaired opsonisation and clearance of pathogens, thus facilitating the development of chorioamnionitis. Fas protein is expressed on the surface of many cell types, such as lymphocytes, epithelial, fibroblasts and certain endothelial cells, and cytokines [ 7 ] and hormones [ 35 ] found in the placental microenvironment may modulate the immune response by regulating the expression of Fas and Fas ligand. In this study two polymorphisms in the Fas ( TNFRSF6 ) gene promoter were studied at positions -1377 and -670, and the -1377A/-670G ( AG ) haplotype was negatively associated with histologic chorioamnionitis. The -1377G/A is situated at the transcription factor SP-1 binding site [ 22 ], and the -670A/G substitution is located in the enhancer region and abolishes the binding site of the nuclear transcription element GAS (Gamma Interferon Activation Sequence). The functional significance of these polymorphisms has not been fully elucidated but they may modulate TNFRSF6 transcription, Fas expression and thus apoptotic and inflammatory responses. Recent evidence suggests that cells are more sensitive to Fas-mediated apoptosis when levels of Fas expression increase [ 36 ]. Apoptotic cell death can modulate host containment of pathogens [ 19 ] and thus genetically determined variation in Fas-mediated cell death may influence histologic chorioamnionitis. Reported studies on gene polymorphisms and adverse pregnancy outcome have mostly examined SNPs in the genes encoding the pro-inflammatory cytokines TNF and IL-1. The TNF promoter -308A allele was associated with clinical chorioamnionitis in a small number of women [ 12 ]. The in vitro functional data of the effects of the TNF SNPs are conflicting and may be context dependent and influenced by strong linkage disequilibrium with other intra-and-extra-genic polymorphisms (e.g. with lymphotoxin alpha and classical HLA genes [ 37 ], not examined in this study). In contrast, in this study the TNF -308A could not be identified as a susceptibility or severity factor for chorioamnionitis. SNPs of the IL1 cluster have been examined from different ethnic groups, in particular IL1B and IL1RN in women with a history of recurrent pregnancy loss [ 38 ], PTB [ 39 ], and in in vitro studies[ 11 ]. We did not find associations between genes of the IL1 cluster or IL1R1 and histologic chorioamnionitis and this may partly be explained by differences in linkage disequilibrium patterns across different ethnic populations that can lead to different association results [ 40 ]. Our data support the role of IL10 , MBL2 and TNFRSF6 variants in determining the risk of histologic chorioamnionitis. However there are limitations to the study, such as small sample size. The range and number of gene polymorphisms examined and analysed by univariate and multiple logistic regression in this study is notable, but some correction for multiple analyses is required [ 41 ]. The appropriate method still remains a contentious issue. The Bonferroni method was not conducted, as this is a conservative method and leads to overcorrection [ 42 ]. Also our study is a hypothesis-generating exploratory study. SNP frequencies in women with and without chorioamnionitis are available to be included in meta-analysis to confirm or reject these findings ( Additional file 1 ). Conclusion In conclusion, the frequency differences in this study highlight the significance and biological relevance of genetic factors, IL-10, MBL and Fas regulation among women with acute chorioamnionitis and adverse pregnancy outcome. Competing interests The author(s) declare that they have no competing interests. Authors' contributions MA contributed to study design, recruited study subjects, abstracted medical records, entered data, performed preliminary statistical analyses and drafted the manuscript. PH provided design regarding immunoregulatory genes, interpretation of results and preparation of manuscript. CM conceived methodology and supervision of PCR testing, interpretation of test results and preparation of manuscript. SH supervised preparation of DNA and performed all PCR tests. JR provided clinical authorisation and support for the study. Provided input regarding clinical database design, and provided critical analysis for obstetric correctness. HM conceived and coordinated the study, conceived medical record review, reviewed statistical analyses and preparation of manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Gene alleles, haplotypes and genotypes in Caucasoid Australian women with preterm birth before 35 weeks gestation and histologic chorioamnionitis. Frequency distribution of immunoregulatory gene alleles, haplotypes and genotypes in Caucasoid Australian women with preterm birth before 35 weeks gestation stratified by histologic chorioamnionitis. Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554771.xml
554765
Vaccination of metastatic melanoma patients with autologous dendritic cell (DC) derived-exosomes: results of thefirst phase I clinical trial
Background DC derived-exosomes are nanomeric vesicles harboring functional MHC/peptide complexes capable of promoting T cell immune responses and tumor rejection. Here we report the feasability and safety of the first Phase I clinical trial using autologous exosomes pulsed with MAGE 3 peptides for the immunization of stage III/IV melanoma patients. Secondary endpoints were the monitoring of T cell responses and the clinical outcome. Patients and methods Exosomes were purified from day 7 autologous monocyte derived-DC cultures. Fifteen patients fullfilling the inclusion criteria (stage IIIB and IV, HLA-A1 + , or -B35 + and HLA-DPO4 + leukocyte phenotype, tumor expressing MAGE3 antigen) were enrolled from 2000 to 2002 and received four exosome vaccinations. Two dose levels of either MHC class II molecules (0.13 versus 0.40 × 10 14 molecules) or peptides (10 versus 100 μg/ml) were tested. Evaluations were performed before and 2 weeks after immunization. A continuation treatment was performed in 4 cases of non progression. Results The GMP process allowed to harvest about 5 × 10 14 exosomal MHC class II molecules allowing inclusion of all 15 patients. There was no grade II toxicity and the maximal tolerated dose was not achieved. One patient exhibited a partial response according to the RECIST criteria. This HLA-B35 + /A2 + patient vaccinated with A1/B35 defined CTL epitopes developed halo of depigmentation around naevi, a MART1-specific HLA-A2 restricted T cell response in the tumor bed associated with progressive loss of HLA-A2 and HLA-BC molecules on tumor cells during therapy with exosomes. In addition, one minor, two stable and one mixed responses were observed in skin and lymph node sites. MAGE3 specific CD4 + and CD8 + T cell responses could not be detected in peripheral blood. Conclusion The first exosome Phase I trial highlighted the feasibility of large scale exosome production and the safety of exosome administration.
Background Immunization protocols rely on the capacity of the vaccine design to elicit long term protective, peptide specific, MHC restricted-CD4 + and CD8 + T cell responses in cancer patients. In preclinical studies, few vaccination strategies were shown to counteract tumor induced-tolerance and to promote immunity to cancer leading to tumor eradication [ 1 ]. One of the most promising approach recently developped is based on the adoptive transfer of mature dendritic cells (DC) pulsed with tumor peptides and control antigen [ 2 , 3 ]. Dhodapkar et al. performed a pionnering clinical trial demonstrating the rapid generation of T cell immunity against recall antigens and Keyhole Limpet Hemocyanin in normal volunteers [ 4 ] and Schüler-Thurner et al. successfully promoted active, melanoma peptide-specific, IFNγ producing, effector CD8 + T cells [ 5 , 6 ] and helper CD4 + Th1 lymphocytes [ 6 ] in the majority of patients with metastatic melanoma. If a single injection of CD34+ progenitor derived-DC vaccine can lead to induction of T cell immunity [ 7 ], correlations between clinical responses and CTL Tc1/CD4 + Th1 activation have been rarely reported [ 8 - 10 ]. Therefore, it is conceivable that antigen spreading initiated by the specific T cells and/or intervention of alternate effectors also elicited by mature DC [ 11 , 12 ] might account for this apparent discrepancy. DC process exogenous antigens in endosomal compartments such as multivesicular endosomes [ 13 ] which can fuse with plasma membrane, thereby releasing antigen presenting vesicles called «exosomes» [ 14 - 16 ]. Exosomes are 50–90 nm diameter vesicles containing antigen presenting molecules (MHC class I, class II, CD1, hsp70–90) tetraspan molecules (CD9, CD63, CD81), adhesion molecules (CD11b, CD54) and CD86 costimulatory molecules [ 17 - 19 ] i.e the necessary machinery required for generating potent immune responses. Exosomal MHC class I and II /peptide complexes are functional but require to be transferred to naive DC [ 20 - 23 ] to promote T cell activation leading to tumor eradication [ 16 , 21 ]. Exosomes pulsed with tumor peptides are more efficient than peptides alone and as efficient as mature DC for the priming of MART1 -specific CTL and for tumor growth inhibition in the HLA-A2 transgenic mouse model [ 21 ]. The molecular characterization of DC derived-exosomes [ 24 , 25 ] and the definition of quality control parameters for exosome purification and storage [ 26 ] allowed to conduct a Phase I clinical trial aimed at evaluating the feasability of exosome harvesting from autologous monocyte derived-DC cultures and the safety of exosome inoculation in melanoma patients. Secondary endpoints were the immunomonitoring of peptide specific CD4 + and CD8 + T cell responses restricted by exosomal MHC class II and I molecules respectively. Up to 41 ± 6.7 (9–115) id/sc vaccinations/patient at the lowest dosage (i.e 0.13 × 10 14 exosomal MHC class II molecules /vaccine) could be generated for all patients from one leukapheresis. One partial response and some other tumor regressions at skin and lymph node sites even in tumors that did not respond to other vaccine formulations were observed in the absence of toxicity. A case report of MART1 antigen spreading and MHC class I loss variant suggested that exosomes mediated bioactivity in vivo, supporting to conduct Phase II clinical trials. Patients, Material and Methods Protocol design, Patients' characteristics and eligibility criteria We report here about the first phase I trial (Fig. 1 ) which administers 4 exosome vaccinations intradermally and subcutaneously at 1 week intervals (Fig. 2 ). The study was approved by the Ethics Committee, local IRBs and regulatory authorities, and informed written consent was given by all patients. The authors were in compliance with the Helsinki Declaration. Fifteen patients bearing melanoma fullfilling the inclusion criteria were enrolled in the study (see Additional file 1 and Table 1 ). Inclusion criteria were: biopsy-proven American Joint Committee on Cancer stage IIIB and IV metastatic melanoma, HLA-A1 + , or -B35 + and HLA-DPO4 + phenotype (as defined by serological and molecular typing), tumor expressing MAGE3 antigen (assessed by RT-PCR, as described elsewhere [ 27 , 28 ]), inclusion at least 4 weeks after cytotoxic chemotherapy, surgery, or radiation therapy, intradermal skin test positivity to one or more recall antigens using Pasteur Mérieux multi DTH test, age > 18 years, Karnofsky performans status >80%, lymphocyte counts >1000 /mm 3 . Exclusion criteria were: prior chemotherapy or biotherapies <4 weeks before trial entry, brain metastasis, pregnancy, concurrent steroids or immunosuppressive therapy, history of asthma, congestive heart failure, autoimmune disease, active infections. Three patients presenting with skin or LN involved sites declined conventional therapies and received exosomes upfront (see Additional file 1 and Table 1 ). Patients received a 4 week outpatient vaccination course with antigen loaded DC derived-exosomes given intradermally (1/10 th ) and subcutaneously (9/10 th ) every week for a total of 4 vaccinations (Fig. 2 ). The injection sites were rotating between both thighs and forearms. The study was scheduled in two steps i.e 1) MHC class I peptide loading was «indirectly» performed on the DC culture at 10 μg/ml, and 2) MHC class I peptide loading was «directly» performed onto purified and acid eluted exosomes (Fig. 1 ). Indeed, our preclinical studies aimed at comparing the exosome immunogenicity after indirect versus direct peptide loading showed a superiority of the latter process [ 20 , 22 ]. Exosomes (quantified as concentrations of MHC class II molecules, see below) or peptides that were pulsed onto exosomes were administered in a dose escalation design. In the first step, 0.13 or 0.40 × 10 14 MHC class II molecules were inoculated in a cohort of three patients each. In the second step, the concentration of peptide loading onto exosomes varied from 10 μg/ml to 100 μg/ml in a cohort of 3 and 6 patients respectively (see Additional file 1 and Table 1 ). Figure 1 Study design . Different weeks (W) of the study are described. Screening, HLA typing and tumor evaluation were performed within 2 weeks before lymphapheresis. DTH (Multi Mérieux skin test) was performed at the time of lymphapheresis, and exosome vaccine started 3 weeks later. Figure 2 Schema of exosome purification processes . Exosomes were purified from monocyte derived-DC (MD-DC) culture supernatants according to a good manufacturing process already described [26]. In the second part of the trial, indirect loading of MHC class II peptides (MAGE3 247–258 .DP04, KKLLTQHFVQENYLEY) was performed on DC cultures followed, after exosome purification by a direct loading of MHC class I (MAGE3 168–176 .A1/B35; EVDPIGHLY) peptides at 10 or 100 μg/ml at pH 4.2 [22]. Quality control parameters included dosing of exosomal MHC class II molecules, flow cytometry analyses and functional assays using superantigens as described in material and methods. Table 1 Formulation of Product Dose Groups Peptides loaded / HLA class Peptide loading method and concentration DEX dose (expressed as numbers of MHC class II molecules) A MAGE A3 (168–176) / class I Indirect (10 μg/mL) 0.13 × 10 14 MAGE A3 (247–258) / class II Indirect (10 μg/mL) tetanus toxoid / class II Indirect (10 μg/mL) B MAGE A3 (168–176) / class I Indirect (10 μg/mL) 0.4 × 10 14 MAGE A3 (247–258) / class II Indirect (10 μg/mL) C MAGE A3 (168–176) / class I Direct (10 μg/mL) 0.13 × 10 14 MAGE A3 (247–258) / class II Indirect (10 μg/mL) tetanus toxoid / class II Indirect (10 μg/mL) D MAGE A3 (168–176) / class I Direct (100 μg/mL) 0.13 × 10 14 MAGE A3 (247–258) / class II Indirect (10 μg/mL) Clinical grade exosome production MD-DC propagation Based on size and density, we could rapidly purify exosomes from monocyte derived-DC (MD-DC) culture supernatants according to a good manufacturing process already described [ 26 ]. Briefly, the adherent fraction of peripheral blood mononuclear cells (PBMC) from a leukapheresis is replenished with fresh medium (AIMV media discarded from haptoglobin and albumin through a 500 kDa MWCO hollow fiber cartridge ultrafiltration (UFP-500-C-4A from A/G Technology, Needham, MA) followed by sterile filtration through a 0.22 um Sartopore 2 membrane) containing rhuGM-CSF (50 ng/ml, Immunex, Seattle, WA) and rhu IL-4 (250 IU/ml, Schering Plough, Kennilsworth, NJ) until day 7. Clinical grade exosome purification Exosomes secreted in the supernatant of MD-DC cultures (1–4 liters) were purified to 175 fold the starting volume according to a GMP method previously described [ 26 ]. MHC class I peptide loading of exosomes This method has been previously described [ 21 , 22 ]. Briefly, the indirect loading of peptides was performed for both MHC class I (Mage3 168–176 .A1/B35; EVDPIGHLY) and MHC class II peptides (Mage3 247–258 .DP04, KKLLTQHFVQENYLEY) in pharmaceutical quality (Multiple Peptide Systems, San Diego, CA) at 10 μg/ml at day 6 of the MD-DC culture and the exosomes were purified from day 7 culture supernatant as described above. The direct loading of MHC class I peptides at pH 4.2 has been reported [ 22 ]. Quality control parameters for exosome inoculation Insurance quality criteria allowing exosome release in patients were qualitative (expression of CD81 tetraspanins), quantitative (at least 1 × 10 14 MHC class II molecules in immunocapture assays) and included a functional assay (bioactivity in the superantigen SEE test of potency). Quantitation of exosome MHC class II concentration by adsorption ELISA and immunophenotyping These methods have been previously described [ 26 ]. Test of potency The bioactivity of exosomal MHC class II molecules was tested using a superantigen bioassay. Exosomes were incubated with femtoM dosages of SEE and washed by density cushion. SEE harboring exosomes were pulsed onto Raji cells. Exosome pulsed Raji cells were subjected to Jurkat cell lines that produced IL-2 in response to SEE. IL-2 concentrations in the supernatants of the Raji/Jurkat cocultures were assessed using a commercial IL-2 ELISA kit. Clinical monitoring Adverse events were graded according to the WHO Toxicity criteria. All patients underwent assessment of tumor status at baseline and 2 weeks after the fourth exosome vaccination. Tumor evaluation was performed using the RECIST criteria. Disease progression was defined as >20% increase in target lesions and/or the appearance of new lesions, partial response as a > 30% decrease in the sum of the longest diameters of target lesions. Immunomonitoring assays Pre-, per- and post-immunization PBMCs were collected before vaccination (week 0, 5% of total leukocytes collected from the starting leukapheresis), after 2 (week 4, 50 ml heparinized blood sample) and 4 vaccines (week 7, by leukapheresis), isolated by Ficoll density gradient and conditioned either for immediat testing or kept frozen for further analyses. In patient #12, single cell suspensions from dissociated cervical lymph nodes resected at the end of the induction and retreatment phases were also collected. Immunophenotyping Determination of peripheral lymphocytes subsets were performed by four color immunostainings using combinations of FITC-, PE-, Cychrome-, APC-labeled mAbs directed against CD2, CD3, CD4, CD5, CD8, CD19, CD25, CD27, CD45RA, CD56, CD69, CD122, HLA-DR (all purchased from Becton Dickinson, Pont de Claix, FR). Controls included isotype-matched immunoglobulins. Thawed PBMCs were stained for 15 min at 4°C, washed twice and fixed in 1X PBS with 0.1% paraformaldhehyde. Ten thousand viable PBMCs were acquired on a FACSCalibur cytometer (Becton Dickinson, BDIS, San Jose, CA, USA) according to standard FSC/SSC criteria and analyzed with the Cellquest software. Peptides and recall antigens for functionnal assays Reactivity of PBMCs was assessed in response to MAGE3 168–176 .A1/B35, MAGE3 247–258 .DP04 (Multiple Peptide System), and control viral peptides (HIV/nef 113–128 .A1, HIV/nef 73–82 .B35, FLU/pb1 591–599 .A1, FLU/np 265–274 .A2, EBV/bzlf1 54–64 .B35, EBV/bmlf1 259–267 .A2, CMV/pp65 495–503 .A2 TT/p2 830–844 , Eurogentec, Seraing, BEL) used according to the HLA classe I phenotype of patients, as well as, against tetanus anatoxin (at 100, 10 and 1 μg/ml), tuberculin (at 50, 5 and 0.5 IU/ml), and phytohemaglutinin (at 2.5, 0.5, and 0.1 μg/ml, PHA HA16, Murex Biotech, Dartford, UK). Proliferation assays Fresh PBMCs (2 × 10 5 cells/well) were cultured in triplicates in presence of graded doses of peptides (at 50, 5, and 0.5 μg/ml) or control antigens or medium alone. Proliferative capacity was determined after overnight pulsing with [ 3 H]thymidine (1μCi/well, NEN, Paris, France) at day 5 of the coculture (earlier at day 3 for PHA). Cells were harvested onto 96-well Unifilter microplates, dried overnight and radioactivity counted on a microplate scintillation counter (TopCount-NXT, Packard, CA, USA). All determinations were made in triplicate wells and data were calculated as means ± SEM of cpm. Enzyme-linked Immunospot assays ELISPOT assay for the detection of antigen-specific IFNγ-producing T cells was performed as described previously [ 29 ]. Briefly, fresh or frozen PBMCs (5 × 10 5 cells/well) were cultured in triplicate in nitrocellulose-bottomed 96-well plates (MAHA S4510, Millipore, Saint Quentin- en-Yvelines, FR) precoated with 2 μg/ml of a primary anti-IFNγ mAb (1-D1K, Mabtech, Hamburg, GER) in presence of peptides (at 5 μg/ml) or control antigens in RPMI1640 medium supplemented with 8% human AB serum. After incubation for 48 hrs, wells were washed five times and incubated with a secondary biotinylated anti-IFNγ mAb (7-B6-1, Mabtech) for 2 h, washed and stained using an extravidine-alkaline phosphatase conjugate substrate kit (Biorad, Hercules, CA, USA). Spots were evaluated and counted using a computer-assisted video imaging analyser (Bioreader 2000, Biosys, Karben, GER). Microcultures for semiquantitative detection of Mage3.A1/B35 CTL precursors The microculture method developed by Coulie et al. [ 30 ] was used to assessed CTL precursors specific for MAGE3.A1/B35 peptide. Briefly, groups of 2 × 10 5 peptide-pulsed PBMC (20 μg/ml) were cultured in Iscove's medium with 10% human AB serum supplemented with IL-2 (20 IU/ml), IL-4 (10 ng/ml) and IL-7 (10 ng/ml) and restimulated at day 7 with replacement of 50% fresh medium with addition of MAGE3.A1/B35 peptide (10 μg/ml). On day 15, microcultures were screened for specific CD8 + T cells using the relevant MAGE3.A1 or B3501 tetramers coupled to PE together with a control APC labeled-HLA-A1 containing an influenza peptide (all tetramers kindly given by D. Colau, Ludwig Institut for Cancer Research, Brussels, BE). This microculture procedure was also applied on patient #12's tumor infiltrating LN to highlight CD8+ T cells specific for HLA-A2/MART1 26–35 , and HLA-B3501/MAGE3 168–176 . Flow cytometry acquisition and analysis of tetramer positive cells were performed as recommanded by P. Coulie [ 30 ]. Immunohistochemistry on lymph node tissues Immunostainings were performed on sections obtained from formalin-fixed and paraffin-embedded lymph node samples. Sections were deparaffinized, placed in 10 mmol/L Na-citrate buffer (pH.7), and heated in a microwave during 20 min. Endogenous peroxidase was blocked with 1% hydrogen peroxide in methanol for 30 min. Slides were stained using anti-CD3 and anti-CD57 mAb (Pharmingen, France) and secondary antibodies with appropriate controls. Results Feasability of exosome production in advanced melanoma patients Leukaphereses of 1.5 blood mass performed in the 15 metastatic melanoma patients enrolled in the study allowed the recovery of 9.7 × 10 9 ± 0.8 PBMC (range: 4.4–15) containing 20.7% ± 1.8 CD14 + cells (range: 12.4–33.0). These monocytes differentiated into immature MD-DC in rhu GM-CSF and rhu IL-4 (means: 313 × 10 6 ± 100, range: 50–1200) as assessed at day 7 in flow cytometry highlighting loss of CD14 molecules, acquisition of CD1a, poor cell surface expression of CD83. In 3 patients out of 15, the first leukapheresis did not allow CD14 + cells to adhere and consequently, a second leukapheresis was required to harvest exosomes. The GMP process allowed to purify a mean of about 5.22 × 10 14 ± 0.9 (range: 1.2–15.0) exosomal MHC class II molecules from supernatants of day 6 to day 7 MD-DC required for up to 41 ± 6.7 (9–115) id/sc vaccinations/patient at the lowest dosage (i.e 0.13 × 10 14 exosomal MHC class II molecules /vaccine). As for the patients who benefited from continuation treatment requiring a second leukapheresis (pt #3, #12 and #14), the second exosome recovery was not significantly different from the first one. The quality control parameters required for GMP exosome batch release are reminded in M&M. GMP exosomes were successfully produced from DC cultures derived from all melanoma patients enrolled in this phase I study, allowing at least 4 exosome inoculations. Safety and clinical outcome No major (>grade II) toxicity was observed. A grade 1 fever was recorded in 5 patients. We noted slight inflammatory reactions at vaccine sites without outward delayed type hypersensitivity reactions. Some patients (#4, #8, #11, #12 and #15) reported a transient swelling and sensitivity of cutaneous and lymph nodes metastases 48 hrs after each exosome inoculation. Among the six patients vaccinated in the first part of the study, only one patient (#3) presenting with a stage III disease exhibited a minor response (disappearance of one sc lesion out of 3). This patient benefited from a continuation therapy with exosomes every other 3 weeks for 21 months and remained stable for up to24 months. It is noteworthy that this patient was progressing despite vaccination with MAGE3 protein prior to enrollment in the Phase I exosome trial. In the second part of the study («direct loading»), responses were only observed in the second group of patients receiving the highest dosages of peptides (100 μg/ml) pulsed onto exosomes. One female patient (#12) presenting with five supraclavicular invaded lymph nodes after conventional chemotherapy (DTIC, 2 cycles) exhibited a partial response (PR) after the induction therapy. Interestingly, disappearance of arterial neovasculature concomittant with tumor shrinkage and necrosis as assessed by doppler pulsed ultrasonography could be demonstrated (Fig. 3 ). Halo of depigmentation around neavi appeared 10 months after the initiation of vaccination (Fig. 3 ). A continuation therapy with exosomes was administered for 4 months every other 3 weeks allowing stabilisation without toxicity supporting the indication of surgery that confirmed the partial response (Jul. 2002, Fig. 3 ) followed by a second leukapheresis allowing to pursue vaccination for 10 months. This patient relapsed in contralateral nodes six months after exosomes discontinuation but exhibited a slow pace of tumor growth. Two additionnal patients (#11, #14) presenting with only skin and /or LN lesions exhibited transient stabilisation and started a continuation therapy. Interestingly, patient # 14 who displayed a SD with exosomes had previously progressed despite biotherapy with IFNα and vaccination with poxviruses recombinant for the MAGE 3A1 cDNA. It is of note that patient #9 in a M1b stage exhibited a regression on a subcutaneous nodule but did progress in the pulmonary sites (mixed response). Figure 3 Clinical outcome of patient #12 during exosome-based vaccination . This patient presented with progressive supraclavicular lymph nodes containing MAGE3 expressing tumor cells in July 2001 when enrolled in the exosomes Phase I trial starting in October 2001. She underwent a first leukapheresis for exosomes production and vaccination (weekly injections in Nov. 2001 during induction therapy and from January 2002 to April 2002 on a three week basis in continuation treatment). The initial size of the target LN are indicated on the left and were followed up by doppler pulsed ultrasonography from Nov 2001 to Dec 2001 (sizes indicated on the right side at the end of the induction therapy). Continuation therapy with exosomes was indicated and maintained clinical stability until the last available exosome dose in Jul. 2002, when she underwent surgery for lymphadenectomy. Results obtained by the pathologists are indicated (N+ if node is invaded by tumor cells, N- if not, MAGE3+ as expression of MAGE3 mRNA in RT-PCR). A second leukapheresis was performed on Aug. 2002 allowing a second therapy with exosomes on a 3 week basis that was continuated until Jul 2004. Six months after exosomes discontinuation (Oct 2004), the patient relapsed in contralateral LN and presented with one lung metastasis. Exosome therapy promoted 2 stable diseases, 1 minor response, 1 partial response and 1 mixed response in skin or LN sites even in patients progressing with biotherapies or alternate vaccines. Patient #3 and #12 with respectively minor and partial responses are still alive. Neither of these patients exhibited a delayed type hypersensitivity response to the immunizing epitopes after the completion of the 4 vaccines. T cell immunomonitoring The phenotypic analyses of lymphocyte subsets did not reveal any significant changes in the percentages nor absolute counts following exosome therapy (not shown). Interestingly, the CD122 molecule i.e IL-2Rβ chain was upregulated in both CD4 and CD8 T cell subsets after exosome therapy (p < 0.05 using Student t'test, week 7 versus week 1 for CD4+ T cells). Serial ex vivo ELISPOT analyses of PBMC before (W1, Fig. 4A ) and after (W7, Fig. 4B ) exosome vaccination did not reveal any significant Th1 (using TT or DP04 MAGE 3 peptides) or Tc1 (using MAGE 3. A1/B35 peptides) type immune responses induced by exosomes. Recall responses (proliferative responses to TT protein using autologous DC or restimulation with autologous DC pulsed with MAGE 3.DP04) did not allow to conclude that exosomes boosted MHC class II specific responses at these time points (not shown). In accordance with these findings, the skin reactivities to iterative injections of id/sc exosomes did not correspond to clinically significant Delayed Type Hypersensitivity responses. Extensive immunomonitoring by screening of 15 day-microcultures of MAGE3 peptide pulsed PBMC performed in the presence of IL-2, IL-4, IL-7 [ 30 ] using the MAGE3.A1/B35 tetramers allowed to detect and clone specific CTL precursors only in 3 patients (#3, 9 and 11) but estimated at a low frequency (10 -6 - 10 -7 of the CD8 + T cells). These studies did not allow to conclude for significant anti-MAGE3A1/B35 CTL responses induced by exosome therapy nor to suggest the clonal expansion of discrete T cell specificities (not shown) in patients' peripheral blood. Figure 4 Evaluation of Tc1/Th1 immune responses to melanoma and viral/recall antigens . A. Before exosome vaccination (W1). B. After exosome inoculation (W7). PBMC obtained at baseline (W1) and after 4 exosome injections are cultured 48 h with the immunizing melanoma antigens i.e Mage 3.A1/B35 or Mage3.DP04 (5 μg/ml) or with viral/recall control antigens (FluMP.A1, HIV.A1, EBV.B35, tetanus anatoxin, tuberculin) or with PHA. The three first patients were also assayed with the universal MHC class II restricted TT peptide. The specific T cell response in each of the evaluated patients is expressed as the number of IFNγ spot forming unit/5 × 10 5 PBMC. Comprehensive analyses of a partial response Patient#12 presenting with five supraclavicular lymph nodes exhibited a partial regression after 4 exosome vaccines in November 2001 and underwent a continuation therapy from Jan. 2002 to April 2002 (Fig. 3 ). The HLA class I haplotype of the patient was A2.A26.B35.B44 and the initial tumor sites expressed both MAGE1, MAGE3 and other melanoma antigens such as Melan-A/MART1 and NA-17 (as assessed using semiquantitative RT-PCR). The antitumor effects did not correlate with an enhanced frequency of MAGE3 specific CTL precursors as demonstrated using microculture detection assays after the first 4 vaccines or after the continuation treatment (not shown). However, tumor shrinkage after the continuation therapy prompted a partial surgical resection in July 2002 and comprehensive analyses of tumor infiltrating lymphocytes (TILN) and tumor cells in 4 avalaible sites (#1, 3, 5, 6 in Fig. 3 ). Site 5 was scored tumor free by the pathologists. Site 1 and 3 maintained the transcription of Mage3 mRNA. In these TILN sites, detection of HLA-A2/MART1 26–35 , -/gp100 209M-217 -/Tyrosinase 239–251 and HLA-B3501/MAGE3 168–176 -specific CD8+ lymphocytes using soluble fluorescent tetramers were performed directly ex_vivo and after microculture restimulation assays using MART1 26–35 or MAGE3 peptides.Only high frequencies of MART1 specific CD8+ T cells were detected in TILN from site 1 and 6 (1.22 and 0.85 % respectively) and in peripheral blood (0.53% in May 2002 versus 0.15% in Oct. 2001) that were confirmed by expansion in all microculture assays (Fig. 5A and 5B ). In parallel, halo of depigmentation around naevi appeared (Fig. 3 ). Moreover, a cell line established from site 1 (CUR1) did not express the HLA-A2 allele in flow cytometry using the MA2.1mAb but maintained the expression of HLA-BC molecules (Fig. 5C ). In November 2003 while disease was slowly progressive, another LN was removed allowing to re-establish another cell line (CUR2) that no longer express HLA-BC molecules (as assessed in flow cytometry using W6.32 or anti-HLA-BC mAb, Fig. 5C ). Figure 5 Antigen spreading and MHC class I loss variant in patient#12 . Flow cytometry analyses on serial blood specimen (A) or tumor invaded lymph nodes (B) gating on CD8+ T lymphocytes stained using A2/Mart1 or A2/gag specific fluorescent tetramers. Ex vivo microcultures stimulated with Mart1 peptides and examined according to similar settings. (C) Flow cytometry analyses of two CUR tumor cell lines (pt#12) after the first exosome therapy (continuation treatment) in clinical response and after the second exosome course (second leukapheresis) at relapse for MHC class I (anti-HLA-A2 mAb, anti-HLA-BC mAb and W6.32 Ab) expression. A positive control was included which consisted of a allogeneic HLA-A*0201melanoma line FON. Immunohistochemistry performed on the primary (April 2001) and secondary (July 2002) tumor specimen revealed quantitative modifications of tumor infiltrating CD3+CD57+ T cells. While activated T cells were mostly surrounding the tumor bed initially, a two fold expansion or recruitment of activated T cells invade tumor areas after exosome vaccines (Fig. 6 ). Figure 6 Lymphocyte recruitment and activation in patient#12's melanoma . T cells that were double-stained by anti-CD3 and anti- CD57 mAb were counted in 2 sections of 6 lymph nodes available from the lymph node dissection specimen obtained before (A. April 2001) and after (B. July 2002) treatment (cf Fig. 3). Counts were performed in 12 sections in total as follows: cells that were double-stained by CD3 and CD57 were counted on the whole section including B cell areas, and the total count was reported to 1 mm 2 . Results showed an increase of the CD3+CD57+ cells after treatement (mean = 122/mm 2 ) as compared to the count before treatment (mean: 58/mm 2 ). Discussion This work is reporting for the first time the feasability and safety of DC derived-exosome-based vaccination in melanoma patients. From a single leukapheresis, up to 10 14 -10 15 exosomal MHC class II molecules could be purified in GMP conditions for all patients, allowing 9–123 sc/id vaccinations with 0.13 × 10 14 exosomal MHC class II molecules. The optimal dosages of exosomal MHC class II molecules required to trigger an efficient peptide-specific CTL response leading to tumor rejection was about 10 10 -10 11 molecules in preclinical studies in the HLA-A2 transgenic mouse model [ 21 ]. The optimal method for pulsing exosomal MHC class I molecules with peptides was a direct loading after acid elution of the exosomal pellet [ 20 , 22 ]. A dose response was observed in vitro, prompting the use of 100 ug/ml for peptide pulsing onto GMP exosomes. As to MHC class II associated-peptide loading, pulsing immature day 5 DC with 10 ug/ml was optimal [ 22 ]. However, in this Phase I study, since no untoward clinical effect nor pharmacodynamic parameter appear to be dose-dependent, neither the peptide pulsing method nor the dosage of exosomes or peptides could help defining the maximal tolerated doses of exosomes. Interestingly, in the second Phase I trial initiated at the Dukes' University in unresectable lung carcinoma, there was a trend for a better efficacy of exosomes directly pulsed with MHC class I peptides in long term survival [ 31 ]. Only one objective response (PR) according to the RECIST criteria was recorded. However, one minor response and two stabilisations prompted the continuation of treatment as well. Tumor regression was exclusively observed in skin and lymph node lesions, as already reported for peptide-based vaccines [ 32 ]. It is noteworthy that three of these patients had primarily resisted to alternate immunotherapy strategies (i.e IFN type I, ALVAC-Mage 3, Mage 3 protein) but it is also likely that such immunomodulators might have facilitated exosome efficacy. However, we were not able to detect significant Mage 3 specific CTL precursors in peripheral blood (nor lymph nodes and lesions) after exosome vaccination. Notewithstanding, MHC class I tumor loss variant and naevi depigmentation was observed in one case (pt #12), suggesting melanoma antigen spreading mediated by HLA-A2-restricted CTL cells. It is conceivable that the Mart1-specific CTL monitored in blood and in LN TIL could have mediated part of these antitumor effects. No DTH responses to iterative inoculations of exosomes or MHC class II restricted-direct or recall responses (to TT epitopes (3 pts tested), or Mage DP04 epitopes (12 patients tested)) were detected at 4 and 6 weeks of exosome vaccination. It is noteworthy that peptides and not whole proteins (such as TT or KLH) were used to pulse DC in vitro, in contrast to what has been reported in most DC trials. However, in the transgenic Marilyn mouse model [ 23 ], id inoculation of I-A b harboring exosomes pulsed with H-Y epitopes could promote expansion of peptide specific CD4 + transgenic T cells. Nevertheless, MHC class I restricted-CTL responses elicited by exosomes pulsed with Mart1 peptides in HLA-A2 transgenic mice were dramatically boosted by adjuvants [ 21 ]. Therefore, it is conceivable that injection of exosomes in the area of an inflammatory draining lymph node harboring mature APC might promote elicitation of T cell immune responses leading to tumor regressions in patients. Alternatively, we recently showed that regulatory C4+CD25+ T cells (Treg) restrict T cell responses elicited by DC derived-exosomes (Chaput et al, in preparation). Therefore, high yields of Treg at start might severely impair the efficacy of exosomes to prime or boost T cell responses. Supporting this view, our unpublished data demonstrate synergistic T cell dependent-antitumor effects between peptide pulsed-exosomes and immunopotentiating dosing of cyclophosphamide. Since no specific CD4+ or CD8+ T cells generated by the exosome vaccines could be detected and could account for the tumor regression in patient#12, what could have been the primary effectors accounting for antigen spreading? We were able to show enhanced NK cell effector functions following exosome administration in peripheral blood of 8/13 patients including patient #12 for whom CD3-CD57+ cells expressing PEN5 (mostly expressed on CD56 dim NK cells) were markedly infiltrating the tumor after therapy (N. Chaput, manuscript in preparation). Therefore, it is conceivable that DC derived-exosomes from melanoma patients are specifically endowed with NK cell stimulatory capacity in vivo. This hypothesis deserves further investigations and should be pursued in the next clinical studies. Moreover, since our preclinical data in tumor bearing mice highlighted a critical role of suppressor T cells in restricting exosome-mediated specific T cell responses, T regulatory cells should be monitored and combination of cyclophosphamide together with exosomes could be invisioned (N. Chaput, in preparation). Phase II/III trials will address the potential of exosomes to enhance time to progression in advanced non small cell lung carcinoma patients. Abbreviations DC : Dendritic cells; MD-DC : monocyte-derived DC; Dex: Dendritic cell -derived exosomes; CTL: Cytotoxic T Lymphocytes; GMP: good manufacturing processes; MHC: Major Histocompatibility Complex; LN : lymph nodes. Authors' contributions Bernard Escudier 1• was the principal investigator, Thierry Dorval 2• , Sophie Piperno 2 , Caroline Robert 3,9 were the clinicians ensuring patients' enrolment, follow up and clinical care. • BE and TD equally contributed to this work Marie-Pierre Cabi 4 , Sophie Novault 3 , Christophe Leboulaire 3 , Mojgan Movassagh 3 and Christophe Leboulaire 3 were monitoring T cell responses corrdinated by Olivier Lantz 4 and Eric Angevin 1 Nathalie Chaput 3 , Fabrice André 3 and Caroline Flament 3 monitored NK cell responses Catherine Boccaccio was in charge of the cell therapy unit at the Institut Gustave Roussy 5 Sebastian Amigorena 4 , Christian Bonnerot 4 and Laurence Zitvogel 1 were the scientists at the source of preclinical data orienting the study design. Thomas Tursz 1 was heading the Host and Tumor Development Programm of the Institut Gustave Roussy, initiating the clinical study. Vincent Serra 6 , Nancy Valente 7 , Olivier Dhellin 6 were the Director of Anosys SA, the medical director of Anosys Inc. and the Pharmacist in charge of production, regulatories and quality insurance in Anosys SA respectively. Jean-Bernard Le Pecq 6,7 was the Chief and Scientific Officer of Anosys leading the research and development of the exosome programm. Alain Spatz 8 , Christophe Borg3 were in charge of the immunohistochemitry analyses on tumor specimen. Supplementary Material Additional File 1 Dosing, demographic, base line data and clinical outcome Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554765.xml
544576
Structure and evolution of protein interaction networks: a statistical model for link dynamics and gene duplications
Background The structure of molecular networks derives from dynamical processes on evolutionary time scales. For protein interaction networks, global statistical features of their structure can now be inferred consistently from several large-throughput datasets. Understanding the underlying evolutionary dynamics is crucial for discerning random parts of the network from biologically important properties shaped by natural selection. Results We present a detailed statistical analysis of the protein interactions in Saccharomyces cerevisiae based on several large-throughput datasets. Protein pairs resulting from gene duplications are used as tracers into the evolutionary past of the network. From this analysis, we infer rate estimates for two key evolutionary processes shaping the network: (i) gene duplications and (ii) gain and loss of interactions through mutations in existing proteins, which are referred to as link dynamics. Importantly, the link dynamics is asymmetric, i.e., the evolutionary steps are mutations in just one of the binding parters. The link turnover is shown to be much faster than gene duplications. Both processes are assembled into an empirically grounded, quantitative model for the evolution of protein interaction networks. Conclusions According to this model, the link dynamics is the dominant evolutionary force shaping the statistical structure of the network, while the slower gene duplication dynamics mainly affects its size. Specifically, the model predicts (i) a broad distribution of the connectivities (i.e., the number of binding partners of a protein) and (ii) correlations between the connectivities of interacting proteins, a specific consequence of the asymmetry of the link dynamics. Both features have been observed in the protein interaction network of S. cerevisiae .
Background Molecular interaction networks are ubiquitous in biological systems. Examples include transcription control [ 1 ], signal transduction, and metabolic pathways [ 2 ]. These networks have become a focus of recent research, because of their important roles in metabolism, gene expression, and information processing. Data on such networks are rapidly accumulating, massively aided by high-throughput experiments. Some of these networks are suffciently complex that their characterization requires statistical analysis, an area of considerable recent interest [ 3 - 5 ]. One key issue in this area is the distinction between structures reflecting biological function and those arising by chance. To address this issue requires an understanding of the biological processes that shape the network on evolutionary time scales. More precisely, one has to identify the statistical observables containing specific information about the evolutionary dynamics that shape a network. In this paper we focus on protein interaction networks, whose nodes correspond to proteins, and whose links correspond to physical interactions between two proteins. Several complementary experimental techniques have been used to analyze pairwise protein and domain interactions, as well as protein complexes, in genome-scale assays [ 6 - 13 ]. Common to these approaches is a high rate of individual false negative and false positive interactions [ 14 , 15 ]. Different protein interaction data sets thus differ in many ways, but they also reveal similar aggregate (or global) network features, such as the fraction of nodes with a given connectivity. This implies that only large-scale statistical features of protein interaction networks can currently be reliably identified by high-throughput approaches. We here present an empirically grounded model that explains empirically observed statistical features of such networks. The currently best characterized protein interaction network is that of the baker's yeast Saccharomyces cerevisiae . On evolutionary time scales, this network changes through two processes, illustrated by figure 1 . These are (i) modifications of interactions between existing proteins and (ii) the introduction of new nodes and links through gene duplications . Duplications of a single gene result in a pair of nodes with initially identical binding partners. Segmental and global duplications of the genome lead to the simultaneous duplication of many genes. On the other hand, processes affecting the interactions between existing proteins are referred to as link dynamics . Link dynamics results primarily from point mutations leading to modifications of the interface between interacting proteins [ 16 ]. Both kinds of processes, link dynamics and gene duplications, can be inferred from a statistical analysis of the network data, and their rates can be estimated consistently with independent information. Figure 1 The elementary processes of protein network evolution. The progression of time is symbolized by arrows. (a) Link attachment and (b) link detachment occur through nucleotide substitutions in the gene encoding an existing protein. These processes affect the connectivities of the protein whose coding sequence undergoes mutation (shown in black) and of one of its binding partners (shown in gray). Empirical data shows that attachment occurs preferentially towards partners of high connectivity, cf. fig. 3. (c) Gene duplication usually produces a pair of nodes (shown in black) with initially identical binding partners (shown in gray). Empirical data suggests duplications occur at a much lower rate than link dynamics and that redundant links are lost subsequently (often in an asymmetric fashion), which affects the connectivities of the duplicate pair and of all its binding partners [ 22 , 25 , 38 ]. Of course, proteome function in vivo is influenced by further factors, notably gene regulation, which determines the concentrations of the proteins interacting in a living cell. The very definition of a bound state depends on the concentrations of the binding partners: A pair of proteins which binds at high concentrations may no longer form a bound state at lower concentrations. Here we concentrate on protein interactions at constant concentrations as they can be inferred from high-throughput datasets. Previous work by others [ 17 - 19 ] shows how structural features of the network can in principle be explained through mathematical models of network evolution based on gene duplications alone. (For similar duplication-based models of regulatory and metabolic networks, see [ 20 , 21 ].) However, the overall rate of link dynamics has been estimated from empirical data in [ 22 ] and is at least an order of magnitude higher than the growth rate of the network due to gene duplications. It must therefore be included in any consistent evolutionary model. In this paper, we present a model of network evolution that is based on observed rates of link and duplication dynamics. At these rates, the model predicts that important structural features of the network are shaped solely by the link dynamics. Hence, the evolutionary scenario of our model is quite different from the duplication-based models [ 17 - 19 ]. The statistical network structure predicted by the model is in accordance with empirical observations, see the discussion below. This paper has two parts. In the first part, we estimate the rates of link attachment and detachment from empirical data. Specifically, we do not just estimate average rates of link dynamics for the whole network, because this has been done previously [ 22 ], but we show how the dependence of link attachment and detachment rates depends on the connectivities of both nodes (proteins) involved. (The connectivity of a protein is defined as the number of its interaction partners). We find evidence that the basic rate of link attachment is asymmetric . That is, this rate increases with the connectivity of only one of two the nodes involved. This reflects an asymmetry in the underlying biological process: a new protein-protein interaction is typically formed through a mutation in only one of two proteins. In the second part of the paper, we assemble the estimated rates of link dynamics into a model of network evolution. Unlike for most other cases studied so far [ 3 , 4 ], the dynamics of these networks cannot be written as a closed equation dependent on the connectivity distribution , i.e. the fraction of nodes with a given number of neighbors. Instead, the analysis of networks under asymmetric link dynamics involves the link connectivity distribution , defined as the fraction of links connecting a pair of nodes with given connectivities. The model has only one free parameter, the average connectivity of nodes in the network. Its stationary solution correctly predicts statistical properties observed in the data. Central properties of this solution are connectivity correlations between neighboring vertices, in accordance with recent observations in high-throughput protein interaction data [ 23 ]. These correlations are a consequence of the asymmetric link attachment process. Results and discussion Estimates of evolutionary rates Two kinds of processes contribute to the evolutionary dynamics of protein interaction networks. The first consists of point mutations in a gene affecting the interactions of the encoded protein. As a result, the corresponding node may gain new links or loses some of the existing links to other nodes, as illustrated in fig. 1(a) and 1(b) , respectively. We refer to these attachment and detachment processes, which leave the number of nodes fixed, as link dynamics . The second kind of process consists of gene duplications followed by either silencing of one of the duplicated genes or by functional divergence of the duplicates [ 24 - 26 ]. In terms of the protein interaction network, a gene duplication corresponds to the addition of a node with links identical to the original node, followed by the divergence of some of the now redundant links between the two duplicate nodes; see fig. 1(c) . Individual yeast genes have been estimated to undergo duplication at a rate of the order of 10 -2 per gene and per million years [ 27 ]. Some 90% of single gene duplicates become silenced shortly after the duplication, leading to an effective rate g of duplications one order of magnitude lower, i.e., ~ 10 -3 per million years [ 22 , 25 , 27 , 28 ]. Only a fraction of the yeast proteome is part of the protein interaction network, and gene duplicates involving proteins that are not part of the network do not contribute to its growth. Hence, g ~ 10 -3 per million years should be considered an upper bound for the growth rate of the protein interaction network by gene duplications. A crude lower bound for the link attachment rate is a ~ 10 -1 new interaction partners per node and million years. For instance, [ 22 ] estimated the rate at which new interactions were formed as no less than 294.5 new interactions per million years and approximately 1000 proteins. (These estimates are based on the formation of physical interactions between products of duplicate genes, and the approximately known age of the duplicates [ 22 ]. Importantly, most of these new interactions form between old duplicates, duplicates that are no longer under the relaxed selection pressure that is characteristic of young duplicates.) The above estimate gives a number of new interaction partners per protein per million years of a = 2 × 294.5/1000 = 0.589, five times greater than the lower bound of 0.1. To maintain an average network connectivity at the empirically observed value κ ≈ 2.5 interaction partners per protein [ 25 , 29 ], the link detachment rate d has to be close to a , thus d ~ a ~ 10 -1 per million years. This rate of link attachment and detachment is much larger than the duplication rate of g ~ 10 -3 per protein and million years. Hence, the link dynamics is decoupled from the much slower duplication dynamics. On intermediate evolutionary time-scales, the network reaches a stationary state of the link dynamics, while its number of nodes does not change significantly. This stationary state determines the structural statistics of the network, in particular the distribution of connectivities. On long time-scales, however, the network may grow through duplications. We emphasize that all these evolutionary rates are order-of-magnitude estimates, and that such estimates are suffcient for our model and the conclusions we derive from it. One basic but important empirical observation about link dynamics is the fast loss of connectivity correlations of proteins encoded by duplicate genes. Fig. 2(a) shows this loss, as estimated from empirical data. Specifically, the figure shows the average relative connectivity difference | k - k '|/( k + k ') of duplicate protein pairs as a function of the time since duplication, parameterized by the fraction K s of synonymous (silent) nucleotide substitutions per silent site. (As an order of magnitude estimate, a value of K s = 0.1 corresponds to a duplication age of 10 million years [ 25 , 27 ].) In the shortest time interval after duplication, the connectivities are still measurably similar. Soon thereafter, however, the relative connectivity difference becomes statistically indistinguishable from that of a randomly chosen pair of nodes, indicated by the horizontal line in fig. 2(a) . Hence, diversification after duplication is a rapid process, with a time constant of the order of several 10 million years, consistent with the fast rate of link dynamics discussed above. An additional empirical observation underscores the minor importance of gene duplication in shaping the observed network structure. In models of network evolution based on gene duplication [ 17 - 19 ], a protein acquires new links through duplications of its neighbors (see, for example, the grey nodes in fig. 1(c) ), at a rate proportional to its connectivity. This mechanism would generate an abundance of high-connectivity nodes. In addition, it would also generate a high fraction of pairs of neighbors that are products of a gene duplication. This is also true for intermediate models, incorporating both gene duplications and link dynamics, provided the duplication rate is comparable to the rate of link dynamics, or exceeds it. However this prediction of models based on gene duplications is not supported by the data. Fig. 2(b) shows the fraction of duplicate protein pairs among the k ( k - 1)/2 neighbor pairs of a node of connectivity k . This fraction is small and it does not increase significantly with k . The data in this figure are also consistent with the earlier observation that the majority of duplicate pairs have few or no interaction partners in common [ 25 ]. Figure 2 (a) Duplicate protein pairs lose their connectivity correlations over time. The average relative connectivity difference | k - k '|/( k + k ') of duplicate pairs with connectivities k , k ' > 0 is plotted against the time since duplication, parameterized by the synonymous (silent) nucleotide divergence K s . The horizontal line indicates the value expected for two randomly chosen nodes. The average number of duplicate pairs per bin was 16 (from low values of K s to high ones the number of duplicate pairs per bin were 12, 5, 3, 6, 6, 8, 13, 27, 44 respectively). (b) Duplications do not strongly influence network structure. The histogram shows the fraction of duplicate pairs among the k ( k - 1)/2 neighbor pairs of a node of connectivity k plotted versus k . A high number of duplicate pairs would be expected if duplications were a significant mechanism of link gain, see text. The mean and the standard error of this fraction were determined using proteins which are products of duplicate genes with sequence similarity K a < 1. The number of vertices used per column ranges from 374 for k = 2 to 8 for k = 12. We note that in our discussion of node dynamics we have not separately considered the effects of ancient genome duplications [ 39 , 40 ]. The conclusion that gene duplications do not shape the statistical features of the protein interaction network applies both to single gene duplications and to genome duplications. Indeed, the analysis of duplicates presented in figure 2 includes both pairs of genes resulting from single duplications and those stemming from genome duplications. Furthermore, the evolutionary dynamics of individual duplicated genes is similar for the products of single genome and whole genome duplications. For example, individual gene duplicates are lost with approximately the same probability in single duplications and in whole genome duplications. For this reason we do not, at this stage, include genome duplications separately in our model. Dependency of attachment rates on connectivities The total rates a and d at which links are attached and detached in a protein interaction network allow no inference of how these processes shape the statistical properties of the network. To make such an inference, one must also know how the link dynamics depends on the connectivities of the nodes involved. The simplest possibility is that link attachment rates a and detachment rates d are functions of a node's connectivity k . The rates a k and d k at which links are attached or detached from a node of connectivity k have been estimated previously using interactions between products of duplicate genes [ 22 ]. They increase approximately linearly with k . In representing attachment and detachment rates ( a , d ) as functions of connectivity k ( a k , d k ), one assumes implicitly that that the mechanism of link attachment and detachment is identical (symmetric) for the two nodes involved in a changed link. Previous analyses of protein network evolution [ 22 ] as well as models of network evolution [ 30 ] were based on such a symmetric process. However, the biological mechanism underlying link dynamics is intrinsically asymmetric. When a new link is formed, typically only one node undergoes a mutation, whereas the other node remains unchanged. This asymmetry means that the rate of link dynamics will generically depend in one way on the connectivity of the node undergoing mutation, and in another way on that of the unchanged node. As a result the rates a k and d k of link attachment and detachment are insuffcient to describe the dynamics of the network, since these rates will be different depending on whether the node is undergoing a mutation or not. This observation motivates the following estimate of the dependency of the link dynamics rate on node connectivities. We define a k , k ' as the probability per unit time that a given non-interacting pair of proteins with respective connectivities k and k ' will acquire a link, multiplied by the number of proteins N . Analogously, we define the detachment rate d k , k ' as the probability per unit time that a given interacting pair of proteins with respective connectivities k and k ' will lose their link. The scaling convention of both rates is chosen such that the average connectivity of the network remains constant as the number of nodes N increases: the number of nodes pairs (where a link may be added) is proportional to N 2 , whereas the total number of links (which may be deleted) is proportional to N . We refer to the special case where the rates factorize, i.e. a k , k ' ~ a k a k ' , as symmetric attachment (and analogously for the detachment rates d k , k ' ). The specific form of these rates assumes that link dynamics is a local process , so the probability for the formation or destruction of a link depends on the connectivities of only the two proteins involved in this process. We now explain how one can estimate the dependency of a k , k ' on its arguments, k and k '. As described earlier [ 22 ], one can use the observed number of physical interactions among duplicate gene products (cross-interactions) to estimate attachment rates. Briefly, such cross-interactions may arise in two ways. First, a protein that forms homodimers (a self-interacting protein) may undergo duplication, leading to two identical self-interacting proteins which also interact with each other. If both self-interactions are subsequently lost independently , yet the interaction between the nodes is retained, a cross-interaction is formed. This scenario does probably not account for the majority of cross-interactions, because it is inconsistent with data suggesting that self-interactions do not get lost overly frequently after duplication [ 22 ]. The second avenue of forming interactions between duplicate gene products involves a non-homodimerizing protein that undergoes duplication. Subsequently, an interaction between the duplicate proteins may form. If this mechanism is dominant, as we argue, one may use the number of cross-interactions to obtain order-of-magnitude estimates of the attachment rate [ 22 ]. From the number of interactions that each of the two involved proteins has with other proteins, one can estimate how the attachment rate depends on k and k '. The main caveat of this approach is that the connectivity of the duplicates may have changed since the time the link between them was formed. The result of this procedure is shown in fig. 3 . The sample size of 38 cross-interactions is extremely limited, but suffcient to demonstrate an increase of the attachment rate along the diagonal k = k ', and no systematic change along other directions. A different representation of the same data in fig. 3b ) also shows an increase of the attachment rate consistent with k + k '. Figure 3 Link attachment occurs preferentially towards proteins of high connectivity. (a) The color-coded plot shows the fraction of duplicate pairs with connectivities ( k , k ') that have gained a mutual interaction (cross-interaction) since duplication, as a function of k and k '. Points where all duplicate pairs have cross-interactions are shown in white, points where none carry a cross-interactions are shown black. Points (particularly at high connectivities) where no data is available are also shown in black. The number of duplicate pairs with given connectivities ranges from 2 to 39. Points in the k , k '-plane where only a single pair of duplicates exists are excluded. (b) For this histogram the data from a) are binned for low, medium, and high k + k ' and the average for each bin is shown against k + k '. The number of k , k ' values contributing to each bin are 10, 14, and 11, from left to right. Error bars give the standard error. (c) Assuming the functional form f k + f k ' for the probability of a cross-interaction between nodes with connectivities k and k ' (asymmetric attachment), the most likely values of f k may be deduced from the data (see text). The maximum-likelihood result shows an approximately linear increase of f k with k . The alternative scenario, symmetric attachment, yields a smaller maximum likelihood. Only duplicate pairs with K a ≤ 0.4 were used in this analysis in order to avoid overcounting of cross-interactions of duplicates of even older duplicates. An attachment process where one node with connectivity k is chosen with a probability , and a second one is chosen with probability , gives an attachment rate . The attachment rate a kk ' ~ k + k ' which we observe empirically is thus explained by an asymmetric attachment process where one node is chosen uniformly at random ( = constant), and the other node is chosen with a probability proportional to its connectivity ( ~ k ). Note that the rate a k , k ' itself is symmetric under interchange of the labels k and k ', since either of the two nodes may take on the role of being preferentially chosen. However, the rate a k , k ' does not factorize, exactly as required for an asymmetric attachment process. We now present an additional, complementary approach, based on maximum likelihood analysis, which validates the functional form of a k , k ' . The probability that out of n kk ' pairs of duplicates with given connectivities k and k ', m kk ' pairs interact is , where g kk ' gives the probability for a cross-interaction. are the binomial coefficients. The probability p for observing for each pair k ≤ k ' m kk ' interactions in n kk ' pairs of duplicates is then given by . Symmetric and asymmetric attachment differ in how the probability of a cross-interaction g kk ' depends on k and k '. In the symmetric case, g kk ' = g k g k ' . In the asymmetric case where one node is chosen uniformly, the other with a probability f k , we have g kk ' = f k + f k ' . Using simulated annealing [ 31 ] we have calculated the (maximal) likelihoods p that the connectivity correlation pattern shown in fig. 3a resulted from either an asymmetric process, or a symmetric process, respectively, by maximizing p with respect to f k and g k . We find that the maximal likelihood for asymmetric attachment exceeds that for symmetric attachment by a factor p asym / p sym ~ 4. The data thus favor an asymmetric attachment process, consistently with the biological interpretation given above. In addition, in the maximum likelihood analysis of the asymmetric model, f k shows an approximately linear increase with k (see figure 3c ). Although this result is by no means conclusive, the data shows there is no reason to a priori consider only symmetric processes. Thus far, we have only discussed the link attachment rate. For the detachment of links, we analogously assume that links are lost due to mutations at one of two linked nodes, and that the rate of this process does not depend on the properties of the other node that is unaffected by a mutation. The simplest mechanism reflecting these assumptions is one where a protein loses on average d links per unit time. A protein is chosen in an equiprobable manner from all nodes for removal of one of its links. The link to be removed is chosen at random from all its links. (An alternative detachment process consists in the loss of a certain fraction of links and leads to very similar results.) The resulting detachment rate is d k , k ' ~ (1/ k ) + (1/ k '), where the inverse terms stem from nodes (rather than links) being chosen uniformly. Dynamical model of network evolution The rates of the link dynamics discussed above, together with a slow growth of the network due to duplications, define a simple model for the evolution of protein interaction networks. Unlike previous models of the evolution of protein interaction networks [ 17 - 19 ] which emphasize the role of gene duplications, our model is based on the asymmetric link dynamics deduced from empirical data in the preceding section. By analytical solution or by numerical simulation one may investigate the networks generated by our model and compare their statistical properties to those of the empirical data on protein-interaction networks. This will be done in the present section. Before analyzing this model in the limit of large networks, we discuss the specific values of model parameters we used, and present the results of numerical simulations of a finite network. We chose the initial network size such that after a suffcient waiting time, when a stationary state has been reached, the size of the simulated network matches that of the protein interaction data set we used (see methods). Duplication of nodes is modeled simply by adding new nodes with connectivity zero to the network at a rate of g = 10 -3 per node per million years, as motivated above. Using this simplistic growth mechanism is appropriate since, as shown above, the link dynamics will quickly alter the initial connectivity of a new node, as well as connectivity correlations with its neighbors. We begin with a total number of 4600 nodes, uniformly linked at random (giving a Poissonian connectivity distribution) such that the average connectivity of nodes with non-zero connectivity is κ = 2.5, the average connectivity found in the data set we used. After a waiting time of 25 million years there are 4696 nodes in total, of which 1872 nodes have non-zero connectivity. This is the size of the pooled protein interaction data set we used. The waiting time of 25 million years is of the same order of magnitude as the time scale on which connectivity correlations of duplicate nodes decay in figure 2a ) of a few 10 million years. New links are added at a rate of a = 0.59 new interactions per node per million years, using the asymmetric preferential linking rule we motivated above. Specifically, to form a new link we chose one node uniformly and a second node preferentially (i.e., with a probability proportional to its connectivity k ) and link the two nodes. We removed links at a rate that keeps the average connectivity constant. Specifically, at each time-step a link is deleted by choosing a node uniformly for link deletion if the average network connectivity exceeds κ = 2.5. The link to be deleted is chosen equiprobably from the links of this node. The connectivity distribution of a network whose evolution was simulated in this manner is shown in figure 4a ) (open circles, °). This distribution is robust with respect to changes in the ratio of duplication to link dynamics g/a over at least an order of magnitude (results not shown). Figure 4 (a) The asymmetric link dynamics produces a broad connectivity distribution. The model prediction of the connectivity distribution of nodes with non-zero connectivity agrees well with yeast protein interaction data (filled diamonds). The solution of the rate equation (4) is shown as a solid line, the result of a computer simulation emulating the link dynamics encapsulated in (4) for a network of finite size is shown as circles (°). Nodes with the highest k (lower right) occur only once in the network. (b) High-connectivity vertices are preferentially connected to low-connectivity vertices, as also observed empirically. The figure shows the relative likelihood of the link distribution and the 'null distribution' of an uncorrelated random network, see text. We now turn to the consequences of this evolutionary dynamics for the statistical properties of the network. Since the link dynamics places and removes a link with a rate depending only on the connectivities of the nodes at either end, the evolutionary dynamics of the network can be represented in terms of the link connectivity distribution q k , k ' . This distribution is defined as the fraction of network links that connect vertices of connectivities k and k ', where c ij = 1 if node i is linked to j and 0 otherwise. For convenience, a factor κ has been included in the normalization, i.e., ∑ k , k ' q k , k ' = κ . The link connectivity distribution q k , k ' captures correlations between the connectivities of neighboring vertices [ 23 , 32 - 34 ]. It is related to the single-vertex connectivity distribution by for k > 0 and p 0 = 1 - ∑ k > 0 p k . The rates a k , k ' and d k , k ' are related to the total rates a and d of link detachment per unit time by the normalization For a network of infinite size, link and growth dynamics result in a deterministic differential equation for the evolution of the link connectivity distribution q k , k ' The terms J k , k ' arise from links that are not added or removed but that change their values ( k , k '), These are the links joining a mutated protein or its binding partner with third vertices, shown as open circles in fig. 1 . The parameter g accounts for a uniform increase of the number of nodes caused by gene duplications. In writing eq. (4), we have assumed that next-nearest neighbor connectivity correlations vanish. This assumption is self-consistent since the stationary solution has indeed only nearest-neighbor correlations. Truncating all correlations and writing down an evolution equation for the connectivity distribution p k turns out to be inconsistent since asymmetric link dynamics generates non-trivial connectivity correlations. This distinguishes the present model from simpler models of network growth, which can be self-consistently formulated at the level of the distribution p k . We solved equation eq. (4), which describes the evolution of the connectivity correlations numerically for its steady state. For initial conditions we use a Poissonian connectivity distribution where the average connectivity of connected nodes is 2.5, and connectivity correlations which factorize q k , k ' ~ kk ' p k p k '. We followed the time evolution of q k , k ' defined by eq. (4) until a steady state was reached using the parameters a and g given above and choosing d such that the average connectivity of connected nodes remains at a constant κ = 2.5. This procedure leads to a stationary link connectivity distribution and a resulting connectivity distribution independent of initial conditions. Because the evolution equation is a rate-equation that applies to a network of infinite size, the parameters determining the stationary state are the ratio between growth and attachment rate, the functional form of the attachment and detachment rates, and the average connectivity. The stationary state turns out to be asymptotically independent of the duplication rate for small duplication rates. In fact, if we solve eq. (4) numerically for any ratio g / a < 10 -1 , the results are statistically indistinguishable from that for g = 0, implying great robustness against errors in the rate estimates discussed above. The statistical properties of our model in its stationary state may now be compared with the corresponding quantities in the protein-interaction network. The connectivity distribution agrees well with the empirical data as shown in fig. 4(a) along with the results of numerical simulations. The distribution is broad but not scale free. (From the empirical data with connectivities distributed over little more than a single decade the scale-free property of protein networks – meaning that connectivities are distributed according to a power law – can not be confidently ascertained. Furthermore the empirical data shown in fig. 4 distinctly deviates from a power-law.) This also holds for uniform detachment, where d kk ' = constant, and it is a crucial difference to models with symmetric attachment, where preferential attachment leads to scale-free networks, both at constant network size [ 30 ], and in growing networks [ 3 , 35 ]. For the connectivity correlations, we find that vertices of high k are more frequently linked to vertices of low k ' than in an uncorrelated random network with the same connectivity distribution . Fig. 4(b) shows the relative likelihood is the link connectivity distribution of the network with no connectivity correlations. Correlations with this property have recently been reported for the protein interaction network in yeast [ 23 ], but a quantitative comparison with the prediction of our model will have to await a greater amount of reliable protein interaction data. We note that connectivity correlations are a specific property of networks shaped by asymmetric dynamics, and are absent in the case of symmetric dynamics, as discussed in the appendix. In other words, the empirically observed non-trivial connectivity correlations require asymmetric link dynamics. This is an a posteriori reason for considering asymmetric link dynamics. A further consequence of asymmetric attachment is that our model does not obey detailed balance (as is the case of symmetric link dynamics, where attachment and detachment rates do factorize, see [ 30 ]). Asymmetric attachment or detachment rules violate the condition, necessary for detailed balance, that the product of transition probabilities along a circular trajectory in the space of networks is independent of the direction of this tour. This may be demonstrated easily by considering, e.g. four nodes labeled 1 – 4 to be connected linearly and disconnected again. Starting and ending with a single link between nodes (1, 2), say, the product of the rates of adding a link between (2, 3), then (3, 4) before removing the links between (2, 3) and then (3, 4) is , that for the same tour in reverse is , which are generally equal only if the rates facorize in their arguments. Conclusions We have presented a stochastic evolution model for protein networks, which is based on fast link dynamics due to mutations of the coding sequence of existing proteins and a slow growth dynamics through gene duplications. The crucial ingredient of the link dynamics is an asymmetric preferential attachment rule, which is supported by empirical data. The asymmetry has a simple biological interpretation, namely that mutations in one gene may lead to a new interaction of its product with that of another, unchanged, gene. Such a mechanism, where the two nodes involved in the generation of a new link play different roles, is probably the norm, rather than the exception, in biological networks. This holds particularly for regulatory networks, where a new interaction between two genes is formed by changes in the regulatory region of only one of them. Asymmetric link dynamics leads to a network model, where the aggregate variables necessary to describe network structure are the connectivity correlations q k , k ' , which give the fraction of links with connectivities k and k '. In our case, the model successfully reproduces the connectivity distribution found in empirically available protein interaction data. The asymmetry of the link dynamics also leads to connectivity correlations between interacting proteins, which have been observed empirically [ 23 ]. A model with symmetric link dynamics, on the other hand, produces no such correlations. Higher order correlations of this kind [ 33 ] are of particular interest for future work as they may be a quantitative signature of natural selection on the level of the network as a whole. Methods Data processing The protein interaction data in this paper was pooled from three sources. The first of these sources is a large-scale high-throughput experiment using the yeast two-hybrid assay [ 13 ] (data available from [ 41 ]). It comprises 899 pairwise interactions among 985 proteins. The second source is also a high-throughput two-hybrid experiment, from which we used a "core" set of 747 interactions between 780 proteins, interactions that had been confirmed through replicated experiments [ 9 , 42 ]. The third source is the public MIPS database [ 36 , 43 ] of May 2001. From this database, we included only pairwise interactions that were not produced by the two-hybrid assay, but instead by other techniques such as cross-linking or co-purification of two proteins. This resulted in 899 interactions between 680 proteins After pooling the three data-sets and eliminating redundant interactions, we were left with a network of 2463 interactions and 1893 proteins. While enormously valuable in their own right, analyses of protein complexes do not identify pairwise protein interactions, and were thus not suitable for our analysis [ 7 , 8 ]. We also excluded interaction data derived from experiments identifying domain-specific rather than whole-protein interactions [ 10 - 12 ]. For all three data sets taken separately, the connectivity distributions are statistically indistinguishable [ 22 ]. Moreover, the observations on link addition we use here [ 22 ], as well as the patterns in Fig. 2 hold qualitatively for each data set individually. Data on yeast gene duplicates, generated as described in [ 27 ], was kindly provided by John Conery (University of Oregon, Department of Computer Science). Briefly, gapped BLAST [ 37 ] was used for pairwise amino acid sequence comparisons of all yeast open reading frames as obtained from GenBank. All protein pairs with a BLAST alignment score greater than 10 -2 were retained for further analysis. Then, the following conservative approach was taken to retain only unambiguously aligned sequences: Using the protein alignment generated by BLAST as a guide, a sequence pair was scanned to the right of each alignment gap. The part of the sequence from the end of the gap to the first "anchor" pair of matched amino acids was discarded. The remaining sequence (apart from the anchor pair of amino acids) was retained if a second pair of matching amino acids was found within less than six amino acids from the first. This procedure was then repeated to the left of each alignment gap (see [ 27 ] for more detailed description and justification). The retained portion of each amino acid sequence alignment was then used jointly with DNA sequence information to generate nucleotide sequence alignments of genes. For each gene pair in this data set, the fraction K s of synonymous (silent) substitutions per silent site, as well as the fraction K a of replacement substitutions per replacement site were estimated using the method of Li [ 28 ]. Asymmetric link dynamics and connectivity correlations The existence of non-trivial correlations may be attributed directly to the asymmetry of the link dynamics. Symmetric link dynamics, which is a standard mechanism in models of networks at constant size [ 30 ], leads to networks with uncorrelated connectivities: Generalizing the approach of [ 30 ] to include connectivity-dependent detachment, one obtains for symmetric link dynamics with rates a k and d k an equilibrium distribution giving the probability of finding the network in the state given by adjacency matrix c ij of . This results in a connectivity distribution and trivial connectivity correlations , which factorize in the connectivities. This results ina constant . A model with symmetric link dynamics can thus produce any empirically observed connectivity distribution, but no networks with statistically significant connectivity correlations. Authors' contributions ML and AW contributed equally to this work. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544576.xml
497048
Multi-component based cross correlation beat detection in electrocardiogram analysis
Background The first stage in computerised processing of the electrocardiogram is beat detection. This involves identifying all cardiac cycles and locating the position of the beginning and end of each of the identifiable waveform components. The accuracy at which beat detection is performed has significant impact on the overall classification performance, hence efforts are still being made to improve this process. Methods A new beat detection approach is proposed based on the fundamentals of cross correlation and compared with two benchmarking approaches of non-syntactic and cross correlation beat detection. The new approach can be considered to be a multi-component based variant of traditional cross correlation where each of the individual inter-wave components are sought in isolation as opposed to being sought in one complete process. Each of three techniques were compared based on their performance in detecting the P wave, QRS complex and T wave in addition to onset and offset markers for 3000 cardiac cycles. Results Results indicated that the approach of multi-component based cross correlation exceeded the performance of the two benchmarking techniques by firstly correctly detecting more cardiac cycles and secondly provided the most accurate marker insertion in 7 out of the 8 categories tested. Conclusion The main benefit of the multi-component based cross correlation algorithm is seen to be firstly its ability to successfully detect cardiac cycles and secondly the accurate insertion of the beat markers based on pre-defined values as opposed to performing individual gradient searches for wave onsets and offsets following fiducial point location.
Background Computerised classification of the electrocardiogram (ECG) is a complex and multi staged process. The overall goal is to determine if the patient is 'normal' and may remain untreated, or whether the patient exhibits any cardiac abnormalities requiring treatment. The classification of the ECG by computerised techniques has been an active area of research for more than 4 decades. A plethora of algorithmic techniques have been applied and developed [ 1 , 2 ] all with the common goal of enhancing the classification accuracy and becoming as reliable and successful as expert cardiologists. The process can be divided into a number of sequential stages. Pre-processing stages of Beat Detection and Feature Extraction/Selection provide suitable information from the recorded ECG in the form of a digitised feature vector [ 3 ]. This can be considered to describe the current morphology of the recorded signal, hence, following processing by the classification algorithm [ 1 , 2 ], a set of suggestive diagnostic statements can be produced. With a multi-stage computerised approach, the overall classification capabilities of the system are highly dependent on the early stages of processing, i.e., accurate detection of each ECG complex in all recorded leads. Hence the necessity of a reliable beat detection algorithm is of paramount importance. Regardless of the approach employed to analyse and classify the ECG signal, all require accurate detection of each QRS complex. Beat detection algorithms are designed with two main objectives. Firstly, the algorithm employed should provide reliable detection of each cardiac cycle. Secondly, the temporal location of the reference points should be described accurately. In general terms detection of each cardiac cycle involves the location of a fiducial point, usually taken as the peak amplitude of the R-wave or of the QRS complex. From this it is then possible to detect markers for the other interwave components (if present) and features of the ECG; QRS onset and offset, P onset and offset and T offset (Figure 1 ). Figure 1 P wave, QRS complex and T wave beat markers inserted into ECG recording. Two of the most commonly employed approaches to beat detection involve the use of established non-syntactic algorithms [ 3 - 5 ] or cross correlation (CC) algorithms [ 3 , 6 ]. A brief overview of each approach follows. Non-syntactic beat detection A non-syntactic approach to beat detection normally involves an algorithm based on two distinct steps; a pre-processor and a decision rule. The function of the pre-processing stage is to enhance the QRS complex and suppress all other components of the recorded signal i.e. P and T waves and noise artefacts. This is achieved through firstly applying a linear filter to extract the required frequencies, followed by a non-linear transformation with the goals of providing a single positive peak for each QRS complex. The output from the non-linear transformation is then processed by a thresholding function to indicate the presence or absence of a QRS complex. Following location of a reference point remaining inter-wave components can be located to the left (P wave) and the right (T wave) of the QRS complex. A number of non-linear transformations exist which can be applied and can occur in a number of different permutations. A commonly adopted approach is conceptualised in Figure 2 . The primary stage of bandpass filtering (approximately 5–25 Hz) reduces the noise artefacts that may be present in the signal. Detection of the fiducial point need not necessarily occur on the original signal, but can be considered advantageous once some signal transformations have been applied, hence the inclusion of the differentiator, squaring function and moving window integrator. The differentiation stage acts as a highpass filter, exploiting the slope characteristics of the QRS complex. The squaring function is also favourable to the high-frequency components of the signal and serves the purpose for signal rectification. The integrator includes information relating to the duration of the QRS complex, which is recognised in electrocardiography as normally having the longest duration of any component in the ECG. QRS complexes cannot occur physiologically closer than 200 ms in succession. Therefore it is common practice once a QRS complex is detected, that a 200 ms refractory blanking period is initiated. This eliminates the condition where the same QRS complex is detected twice or even a T wave is mistaken for a QRS complex. (For a more detailed review of studies applying non-syntactic beat detection see [ 3 ].) Figure 2 Non-syntactic non-linear transformations transforming original ECG signal into a series of impulse like outputs. Cross Correlation based beat detection The CC function can be used to measure the similarities between two signals [ 7 ]. This process entails the computation of the sum of the products of corresponding pairs of points of two signals, within a specified time frame, or window. The process also takes into consideration any potential phase differences between the two signals via the inclusion of a lead or lag term. The formula for CC is represented as: where N represents the number of samples, j is the lag factor and x 1 and x 2 are the two signals. To normalise the results based on the number of sample points, the factor 1/ N is introduced. When the value of r (12) is maximal, this is considered as the point of maximal similarity between the two waveforms. As the required amount of lagging between the two signals is initially unknown, various degrees of lags within the specified correlation interval must be performed. It is possible to use CC for the purposes of beat detection, by locating the point of maximal similarity between an ECG signal and a predefined template and hence identifying the temporal location of the QRS complex. It is also necessary for the CC algorithm to store a template or a reference signal of the ECG signal. The origin of the template may be from a variety of sources. It may be of an adaptive nature, whereby a section of the patient's recorded ECG is averaged and stored prior to the analysis. Alternatively, a mathematical model may be used or a collection of ECG recordings from a database used to produce a generic template. Many studies in the past have successfully reported the use of CC as a means of automated beat detection. Abboud et al . [ 6 ] used the CC function, calculated using the cross spectrum and fast Fourier transform algorithm for extrastoyle rejection and location of fiducial points. Variations to the CC algorithm have also been successfully reported. These were considered to be more computationally efficient as they do not require the intense multiplicative processes associated with CC, but may be based on for example weighted correlation of differences methods [ 8 ] and average magnitude cross difference methods [ 9 ]. Methods of CC have generally focused on the usage of one waveform as the basis for the template, for example the QRS complex. The algorithms of Abboud et al . [ 6 ] were adapted by Govrin et al . [ 10 ] to facilitate the location via CC of both the P waves and QRS complexes. Templates of both waves were used with unknown ECG traces to identify the individual components of each cardiac cycle. It is the aim of the current study to investigate the potential of adapting the general CC based approach as an accurate and reliable means for beat detection. This approach in the past has been shown to be successful and hence is investigated with the aims of further enhancement. Individual templates for P waves, QRS complexes and T waves are generated and the CC approach applied individually to identify each of these components as opposed to a one template CC based approach. Standard approaches have also been developed for benchmarking purposes. The structure of the remainder of the paper is as follows: Section 2 describes the new beat detection method based on the fundamentals of CC. In addition a description of the benchmarking methods developed are also presented. Section 3 describes the structure of the data set and presents the results and discussions for the beat detection algorithms. Final conclusions to the study are presented in Section 4. Methods The aims of the current study were to investigate the possibilities of developing a new approach to beat detection which would offer enhanced accuracy and reliability in comparison to established techniques. Two different approaches, one based on a standard non-syntactic approach and one based on a standard CC approach were developed for the purposes of benchmarking. Developments of both of these algorithms were based on existing published approaches. A new approach, based on a multi-component based CC algorithm was additionally developed. Details of the 3 beat detection approaches are as follows. Non-syntactic approach The first stage of the non-syntactic approach is the inclusion of a bandpass filter, centred at 17 Hz. The purpose of this is to isolate the predominant QRS energy and attenuate the low-frequency characteristics of the P and T waves, any baseline drift present and the higher frequencies associated with electromyographic noise and power line interference [ 11 ]. (The passband that maximises the QRS energy is approximately in the 5–25 Hz range.) An FIR bandpass filter was implemented with the difference equation expressed in Equation 2: where y i is the filtered output sequence, x i is the input ECG signal and b q represents the coefficients calculated with the Remez exchange design algorithm [ 12 ]. To produce the necessary stage of differentiation, a five-point derivative with the second term equal to zero (as represented in Equation 3) was employed. H ( z ) = 0.1(2 + z -1 - z -3 + 2 z -4 )     (3) This has been chosen as the function H ( z ) in Equation 3 behaves in a similar manner to a parabolic smoothing filter and does not amplify any high-frequency noise. The advantage of such a filter is that it goes to zero at half of the sampling frequency. The non-linear transformation is performed by the point-by-point squaring of the signal samples. This derivative approximates an ideal derivative in the dc through to 30 Hz frequency range. This is the necessary frequency range since all higher frequencies are significantly attenuated by the bandpass filter. Finally, the squared waveform is passed through a moving window integrator. A window integrator with the difference equation as presented in Equation 4 was employed. where N is the number of samples in the width of the moving window. The integrator sums the area under the squared waveform over an approximate 180 ms interval, advances one sample interval and integrates a new 180 ms window. To locate the fiducial point, the maximum of the QRS complex is required. Following application of the aforementioned processes to the recorded signal, the maximum amplitude of the resultant signal is sought. An initial threshold value based on 80% of the maximum amplitude located during an initial training period is used to locate potential QRS complexes. Within a 125 ms window, following exceeding this value, the maximum amplitude located is considered to be the fiducial point. A refractory blanking stage of 200 ms is employed prior to repeating the process for the next cardiac cycle. The markers for the onset and offset are inserted following gradient searches to the right and left of the fiducial point within a window of 200 ms on each side. A thresholding technique was exercised, whereby following the location of 6% of the gradient, for a duration of at least 25 ms, to the left and right of the fiducial point the start and end points were identified respectively. Prior to location of the T wave offset, the original signal was low-pass filtered at 12 Hz in an effort to maximise the frequency content of the T wave whilst suppressing the remaining inter-wave components. With knowledge of the end of the QRS-complex and the beginning of the next adjacent QRS-complex a value of the peak of the T wave, i.e. the maximum value within the identified window, was calculated. The T-offset is located based on a gradient descent thresholding search method. The condition of thresholding was set at 25% of the maximum gradient of the T wave, for a duration of at least 25 ms, to the right of its peak. Following filtering and smoothing, the P wave was sought in a window of 300 ms prior to the QRS onset. Similar to the T wave offset location, the maximum (or peak) is located within the defined window prior to the location of the onsets and offsets. The onset is located following detection of 75% of the maximum gradient, for a duration of at least 25 ms, to the left of the peak and the offset is located following detection of 75% of the maximum gradient, for a duration of at least 25 ms, to the right of the peak. CC approach The necessary templates required for the CC were generated during an initial training stage prior to any CC based analysis. During this stage templates required on an individual lead-per-lead basis were generated by averaging cardiac cycles in the initial section of the patient's ECG recording. This requirement for a training stage generally results in CC based approaches being more suitable for ambulatory processing situations, for example Holter recordings, as opposed to analysis of short term rest ECGs. Given that infinite signals are to be analysed, the CC function for a lag k can be defined as in Equation 5: where the correlation coefficient is defined as: where z can be set to either x or y . The CC approach is only used to detect the fiducial point. The signal is cross correlated with a complete PQRST template. The correlation function calculates a number of correlation coefficients and returns the position of the highest value. Providing this value is greater than a predefined threshold, a QRS is deemed detected. The threshold value can be changed for each signal and template, if required. This ensures adaptability to each individual patient's recording. The QRS onset and offset as well as the P wave onset and offset and T wave offset are detected with similar rule-based gradient search techniques as described in the non-syntactic approach in the previous section. The main difference being that the search windows in which the gradient searches are performed are dependent upon the distances specified in the template with which the ECG signal is cross correlated. Multi-component based CC algorithm With the proposed multi-component based CC detection method, separate templates for each interwave component are used. Each template consists of an individual component of the complete cardiac cycle, detailing accurate marker positions for the respective inter-wave component onset and offset. In the given approach, 3 templates were employed, one for the P wave, one for the QRS complex and one for the T wave (Figure 3 ). Figure 3 An example of templates used by the multi-component based CC approach for (a) P wave (b) QRS complex (c) T wave. The detection method is based purely on CC and is structured into a number of procedural steps as shown in Figure 4 . The CC function employed in all steps is as described in Equation 5. The only difference is that for each interwave component sought, differing templates are used within the algorithm. The first step of the process is concerned with the location of the QRS complex. With this process, a QRS template is cross correlated with the ECG signal. The next step is the detection of the P wave, which is performed using the same CC function, but in this case, the template is representative of the patient's P wave. The last stage involves the detection of the T wave. In this process, a template, representative of the patient's T wave is cross correlated with the ECG signal using the same CC function as with the previous 2 steps. Figure 4 Basic structure of the multi-component based CC detection method using 3 separate templates. In each case the value indicated as being the point of maximum similarity, i.e. the highest correlation coefficient, between the ECG and the template, within the given correlation interval is compared with a pre-defined threshold. The threshold level, used to initially locate the peak amplitude can be varied in the algorithm if required and is established during a pre-learning phase. If the amplitude of the value exceeds the threshold, a waveform detection is considered as having occurred, otherwise the process is repeated with a new portion of the signal. Markers for the QRS onset, QRS offset, P wave onset and P wave offset, as well as the T wave offset form part of the templates used during the CC process. Hence as soon as the individual interwave components are detected, the markers are generated automatically based on the templates used with no requirement to perform any means of gradient based searching. Results and Discussion To examine the efficiency of the algorithms, excerpts from the already established QT database were employed [ 13 ]. This database has been designed specifically for the evaluation of algorithms which detect waveform boundaries in the ECG. The database has approximately 100 records, each record consisting of 15 minute excerpts of two-channel digitised ECGs. The recordings were chosen to include a broad variety of QRS and ST-T morphologies. The records within the QT database were chosen from the MIT-BIH database, the European ST-T database and several other ECG databases collected at Boston's Beth Israel Deaconess Medical Centre. For each record, a minimum of 30 beats have been manually annotated by clinical experts. For each annotated beat, the following markers have been inserted; P wave onset, P wave offset and P wave peak amplitude, QRS onset, QRS offset and QRS peak amplitude and T wave offset and T wave peak amplitude. For the purposes of the given study, each algorithm was exposed to 3000 beats from this database. Comparison of beat detection algorithms To quantify the accuracy of each of the three aforementioned beat detection algorithms in terms of their correct positioning of beat markers, measurements of mean error ( me ) and standard deviation ( SD ) of this error were used. The me value is used to determine how close the detector is to the annotated markers, with the SD value providing information relating to the stability of the detection criteria. For the purposes of validation, the database used for testing [ 13 ] has associated with it a set of tolerance values for each of the beat markers. These measures can be considered to be the minimum values that should be expected with any automatic algorithm. The accuracy with which the automated algorithms performed the detection was compared with manually annotated and clinical validated beat markers. Such comparisons with a gold standard and subsequently with other automated approaches operating on the same data sets adhere to recommended approaches of comparison of automated medical decision support systems [ 14 ]. Average values for the SD (Equation 6) of the me can be generated with the following equation: Where x id is the detected marker position for ECG trace i as identified by the algorithms used and x im is the original stored marker position from the database annotated by experts for ECG trace i . Table 1 shows the performance of each of the algorithms in detecting the 3000 cardiac cycles from the test set. Table 2 indicates the performance of each of the algorithms in comparison with the accepted tolerances for marker insertion. Table 1 Results of performance following exposure to 3000 cardiac cycles for each algorithm. Non-Syntactic CC Multi-component based CC Number of QRS detected out of possible 3000 2850 2799 2931 Table 2 Results of marker accuracy following exposure to 3000 cardiac cycles for each algorithm. Marker Tolerance SD ms Non-Syntactic SD ms CC SD ms Multi-component based CC SD ms P-onset 10.2 22.6 22.1 11.9 P amplitude 23.4 23.8 7.8 P-offset 12.7 16.7 19.7 11.6 QRS-onset 6.5 10.4 10.1 6.6 QRS amplitude 14.3 1.8 1.8 QRS-offset 11.6 12.8 13.1 6.9 T amplitude 19.2 21.4 8.2 T-offset 30.6 18.7 20.6 14.6 In terms of the overall accuracy in detecting cardiac cycles, as shown in Table 1 , the multi-component based CC approach provided the best results. The advantages of considering only the QRS complex during the CC process offers an improvement in detection of each cardiac cycle. It can be considered that the PQRST wave poses a more complex situation to achieve an accurate measurement of maximum similarity during the correlation process than with a template which only represents the QRS portion of the ECG. Considering the accuracy with which the three algorithms were able to detect the correct position for the marker insertion as shown in Table 2 indicates that the multi-component based CC approach outperformed the two benchmarking algorithms in 7 out of the 8 marker insertion positions. In the remaining case (QRS amplitude) the multi-component based CC achieved a similar SD of 1.8 ms with the traditional CC approach. The non-syntactic and CC approaches performed to a similar level with the former providing superior results in 5 out of the 8 marker values. The SD values, however, only differ slightly. This can be considered to be the result of both techniques using a similar approach of gradient descent searching to identify onsets and offsets. The difference in the results can be attributed to the different manner in which each technique defines the search window within which the gradient searching is performed. The increased accuracy of marker insertion of the multi-component based CC approach can be attributed to a number of factors. The major factor being the avoidance of any gradient searching techniques for the marker positioning as is required by the two benchmarking approaches. As the multi-component based CC approach has pre-defined marker positions in-built as an inherent feature of its design, the ability to detect the fiducial point accurately is the most important process the algorithm initially undertakes. Following this process, markers are inserted based on the values stored within the templates used during the correlation process. Methods used in the two benchmarking techniques of gradient searches are prone to false detection of local gradients and noise still remaining in the signal under examination. Considering the accepted tolerances for the 5 markers as identified in Table 2 , the multi-component based CC approach conformed to 4 of these (QRS onset marginally higher – 6.6 ms vs 6.5 ms) with the P onset lying outside of what was considered to be an acceptable range. Given the historical difficulty of P wave detection and accurate marker location reported by many studies [ 15 ], this is not initially considered to be a drawback of the algorithm, but more an indication of an area requiring further improvement of the approach. Overall the multi-component based CC approach outperformed the two benchmarking techniques in both accuracy of cardiac detection and marker insertions, however, 2 parameters within the algorithm were found to have a significant influence on the algorithm's performance; correlation interval and threshold parameter. Values for these parameters can be established during a training period. Correlation interval Results following the testing process indicated that care must be taken when selecting the appropriate correlation interval for CC based approaches. If this value is too large, interwave components of adjacent waves maybe considered during the correlation process. On the other hand, if the interval is too small, the templates used during the correlation process may not have the ability to discriminate between different desired portions of the underlying signal. Figure 5 shows the multi-component based CC approach processing an excerpt from one record with a correlation interval of 560 ms and 280 ms. The markers indicated at the top of the trace are those automatically inserted by the algorithm. The markers indicated at the bottom of the trace are those which have been inserted manually by clinical experts. Figure 5 Insertion of beat markers with a correlation interval of (a) 560 ms (b) 280 ms. Markers included on the bottom of the trace are those indicated and inserted by clinical experts. The notation used is as follows: t] represents the position of the markers for the t wave amplitude and t wave offset respectively; [N] represents the position of the markers for the QRS onset, peak and offset respectively, where N is used to represent a QRS complex; u] represents the position of the markers for the u wave amplitude and u wave offset respectively. Markers included on the top of the trace are those indicated following automated processing. The notation used is as follows: [Q R S] represents the position of the markers for the QRS onset, QRS peak and QRS offset respectively; T T] represents the position of the markers for the T wave peak and T wave offset respectively. In instances where no markers have been indicated, the algorithm has failed to correctly detect the waveform boundaries and peaks. In the first instance detection rates are low as the interval is too large, however, when the interval is reduced to 280 ms, the rate of detection of the markers for the algorithm increases significantly. Given that under normal conditions the duration of the QRS complex is in the region of 100 ms [ 16 ] and a second QRS complex cannot physiologically occur for a further 200 ms, this would suggest a value of the correlation interval in the region of 300 ms as a suitable choice. Threshold parameter The threshold parameter can be considered to be the minimum value for a correlation result to be considered as a true wave detection. For example, P waves can be considered to have relatively low energy content. If the threshold value is initially set at too large a value then it will not be possible for the point at which the CC function returns the point of maximum similarity to exceed this and subsequently indicate a waveform detection. Figure 6 shows the multi-component based CC algorithm with threshold values of 85% and 60% of training waveform peak amplitude averages. As can be seen it was found necessary that the final thresholding check following the CC considered a lower percentage of the average signal amplitude to ensure successful detection of the wave. Figure 6 Insertion of beat markers for threshold values of (a) 85% (b) 60% of training averages. Markers included on the bottom of the trace are those indicated and inserted by clinical experts. The notation used is as follows: t] represents the position of the markers for the t wave amplitude and t wave offset respectively; [N] represents the position of the markers for the QRS onset, peak and offset respectively, where N is used to represent a QRS complex; [p] represents the position of the markers for the p wave onset, amplitude and p wave offset respectively. Markers included on the top of the trace are those indicated following automated processing. The notation used is as follows: [Q R S] represents the position of the markers for the QRS onset, QRS peak and QRS offset respectively; T T] represents the position of the markers for the T wave peak and T wave offset respectively; [P P P] represents the position of the markers for the p wave onset, peak and offset respectively. In instances where no markers have been indicated, the algorithm has failed to correctly detect the waveform boundaries and peaks. The multi-component based CC approach is affected by both the correlation interval and threshold parameters of the algorithm, hence a training/tuning process is required. Each of the other two algorithms suffer from similar inherent algorithmic drawbacks and hence this is not considered to be a disadvantage of the approach provided it is taken into consideration during application of the algorithm. Abnormal recordings Although, as previously mentioned, the QT database has a large variety of abnormal ECG recordings, two specific examples are highlighted at this point to further compare the performance of the multi-component based CC approach and the non-syntactic based approach. In the first instance a recording where an inverted T wave is present is examined and in the second instance a recording similar to conditions exhibited by First Degree Heart Block is examined. Figure 7(a) shows the correct insertion of the markers for the multi-component based CC approach and Figure 7(b) shows the insertion of the markers with the non-syntactic based approach. As can be seen, comparing these two techniques, both have the ability to correctly detect the peak of the T wave, however, the multi-component based CC approach has the ability to detect in all instances the end of the T wave, which was not correctly detected in any of the cases analysed by the non-syntactic approach. The ability to know the shape and form of each component of the ECG waveform in advance has shown here the further benefits of the multi-component based CC approach and in addition showed its ability in instances of non-normal ECG recordings to perform successfully. Figure 7 Application of the multi-component based CC approach (a) and the non-syntactic based approach (b) to abnormal ECG waveforms. In this case, an inverted T wave is present. Markers included on the bottom of the trace are those indicated and inserted by clinical experts. The notation used is as follows: t] represents the position of the markers for the t wave amplitude and t wave offset respectively; [N] represents the position of the markers for the QRS onset, peak and offset respectively, where N is used to represent a QRS complex; [p] represents the position of the markers for the p wave onset, amplitude and p wave offset respectively. Markers included on the top of the trace are those indicated following automated processing. The notation used is as follows: [Q R S] represents the position of the markers for the QRS onset, QRS peak and QRS offset respectively; T T] represents the position of the markers for the T wave peak and T wave offset respectively; [P P P] represents the position of the markers for the p wave onset, peak and offset respectively. In instances where no markers have been indicated, the algorithm has failed to correctly detect the waveform boundaries and peaks. Figure 8(a) and 8(b) show the results of the multi-component based CC approach and the non-syntactic based approach respectively given the instance of First Degree Heart Block. As can be seen the multi-component based CC approach outperforms the non-syntactic based approach, with the ability to detect all elements of the QRS complex. The non-syntactic based approach has only the ability to detect the peak of the QRS complex. The complexity associated with the abnormality has deemed the algorithm unable to analyse any further interwave components. The benefits of matching the waveform with a template previously established for the person under examination offers the ability to specifically tailor to the recordings under investigation as opposed to generically processing the signal as with the non-syntactic approach. Figure 8 Application of the multi-component based CC approach (a) and the non-syntactic based approach (b) to abnormal ECG waveform. In this case First Degree Heart Block is present in the recording. Markers included on the bottom of the trace are those indicated and inserted by clinical experts. The notation used is as follows: t] represents the position of the markers for the t wave amplitude and t wave offset respectively; [N] represents the position of the markers for the QRS onset, peak and offset respectively, where N is used to represent a QRS complex; [p] represents the position of the markers for the p wave onset, amplitude and p wave offset respectively. Markers included on the top of the trace are those indicated following automated processing. The notation used is as follows: [Q R S] represents the position of the markers for the QRS onset, QRS peak and QRS offset respectively; T T] represents the position of the markers for the T wave peak and T wave offset respectively; [P P P] represents the position of the markers for the p wave onset, peak and offset respectively. In instances where no markers have been indicated, the algorithm has failed to correctly detect the waveform boundaries and peaks. Conclusions The accurate detection of the interwave components of the ECG can be considered to significantly effect the overall performance of the computerised classification process. Three approaches to beat detection were developed and extensively tested on 3000 cardiac cycles to assess their performance. Non-syntactic beat detection and CC algorithms were used as means of benchmarking algorithms. A new approach of multi-component based CC was proposed and results showed it to out perform the two benchmarking techniques in both accuracy of cardiac cycle detection and marker insertions. The multi-component approach identified beat markers based on 3 individual CC processes addressing the QRS complex, the P wave and the T wave individually. For each of these correlation steps, once a correlation match had been established, markers were inserted for onsets and offsets based on predefined values. It is proposed that the increase in performance of this approach in comparison with the benchmarks can be attributed to the lack of heuristic gradient searching required for marker insertion. In addition, the ability to match a portion of the ECG signal, namely the QRS, during CC as opposed to matching to the entire PQRST reduces the complexity of the overall process and subsequently enhances performance. Overall, the results have shown the benefits of employing a multi-component based CC approach. Further studies are currently underway to investigate the performance of the algorithms under instances of noise conditions and further variants of non-normal ECG recordings and in addition possible improvements to the stage of P wave detection. Authors' contributions TL carried out the development of the algorithms and performed the testing procedures. TL, CDN and FJO collaboratively designed the study. TL, CDN and FJO co-authored the paper.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC497048.xml
553988
Presence of factors that activate platelet aggregation in mitral stenotic patients' plasma
Background Although the association between mitral stenosis (MS) and increased coagulation activity is well recognized, it is unclear whether enhanced coagulation remains localized in the left atrium or whether this represents a systemic problem. To assess systemic coagulation parameters and changes in platelet aggregation, we measured fibrinogen levels and performed in vitro platelet function tests in plasma obtained from mitral stenotic patients' and from healthy control subjects' peripheral venous blood. Methods Sixteen newly diagnosed patients with rheumatic MS (Group P) and 16 healthy subjects (Group N) were enrolled in the study. Platelet-equalized plasma samples were evaluated to determine in vitro platelet function, using adenosine diphosphate (ADP), collagen and epinephrine in an automated aggregometer. In vitro platelet function tests in group N were performed twice, with and without plasma obtained from group P. Results There were no significant differences between the groups with respect to demographic variables. Peripheral venous fibrinogen levels in Group P were not significantly different from those in Group N. Adenosine diphosphate, epinephrine and collagen-induced platelet aggregation ratios were significantly higher in Group P than in Group N. When plasma obtained from Group P was added to Group N subjects' platelets, ADP and collagen-induced, but not epinephrine-induced, aggregation ratios were significantly increased compared to baseline levels in Group N. Conclusion Platelet aggregation is increased in patients with MS, while fibrinogen levels remain similar to controls. We conclude that mitral stenotic patients exhibit increased systemic coagulation activity and that plasma extracted from these patients may contain some transferable factors that activate platelet aggregation.
Introduction Systemic thromboembolism represents a major complication in patients with mitral stenosis (MS), especially in those who have atrial fibrillation [ 1 , 2 ]. A hypercoagulable state has been reported in patients with MS and sinus rhythm [ 3 , 4 ]. The association between MS and higher levels of coagulation is well-known; however, the source of increased coagulation remains unclear. Increased regional left atrial coagulation activity may be involved, even when systemic coagulation assessed by peripheral blood sampling is normal [ 5 , 6 ]. Others have reported no significant variation in thrombogenesis, platelet activation, and endothelial dysfunction between the left atrium, right atrium and peripheral arteries or veins [ 7 ]. This study assessed systemic coagulation and platelet activities by measuring fibrinogen levels and platelet aggregation in mitral stenotic patients' peripheral venous blood and compared these measures to hematologic parameters in normal controls. Methods Study groups Eighteen newly diagnosed patients with rheumatic MS were enrolled in the study. All were coded as having NYHA class I functional capacities. Two patients were excluded from this study, one due to pregnancy, the second due to current use of oral contraceptives. The study group thus consisted of 16 patients with MS (Group P, 14 female, 2 male, mean age 27.8 ± 6.5 years). Peripheral venous coagulation activity and platelet activation tests were also evaluated in 16 control patients (Group N, 13 female, 3 male, mean age 28.3 ± 6.1 years), who were normal volunteers. All participants gave informed consent. Exclusion criteria Patients were excluded from participating in the study for the following reasons: aortic and/or pulmonary valve disease, severe mitral regurgitation and/or left ventricular systolic dysfunction, atrial fibrillation, hypertension, diabetes mellitus, dyslipidemia, a history of renal or liver disease, malignancy, venous thrombosis, systemic or pulmonary embolism, congenital hemorrhagic disease, thrombocytopenia, thrombocytosis, acute or chronic inflammatory disease, autoimmune disease or current use of oral contraceptives or anticoagulant or anti-platelet drugs. Echocardiography In patients with MS, transthoracic echocardiography was performed with a 2.5-MHz transducer and a Hewlett Packard Sonos 4500 system to assess left atrial (LA) diameter, mitral valve area (MVA) and the transmitral mean pressure gradient (TMmPG). The LA anteroposterior diameter was determined using standard M-mode criteria [ 8 ], and mitral valve area was calculated according to the pressure half-time method [ 9 ]. Fibrinogen measurement Peripheral venous blood (for measurement of fibrinogen levels) was transferred into tubes containing 1:9 trisodium citrate (0.109 M, 3.2%). Blood samples were centrifuged within 1 hour of collection at 2,500 g for 10 minutes to separate out the plasma. An STA fibrinogen kit measured fibrinogen levels in plasma, using the clotting method of Clauss. The median inter-assay and intra-assay coefficients of variation for the assays were 3.57% and 3.65% for fibrinogen, respectively. Platelet activation tests Blood samples were obtained from the forearm veins of participants, using standard blood drawing procedures (normal blood flow and no pressure). Several samples were drawn for different components of the study. Nine milliliters of blood was collected into special tubes containing 1 ml of 3.8% sodium citrate. Platelet count and other parameters, including white blood cell count and hemoglobin, were measured at the Ege University Hospital Hematology Laboratory using an automated hemocytometer (Cell Dyne 4000 Abbott, USA). Patients with platelet counts less than 2 × 10 5 /ml were excluded from the study. For in vitro platelet aggregation tests, platelet-rich plasma (PRP) samples were obtained by centrifuging samples at 1,100 rpm for 10 minutes in a refrigerated centrifuge (Hettich EBA 12R, Germany). Centrifuging samples at 3,000 rpm for 10 minutes yielded platelet-poor plasma (PPP) samples. The platelet counts of both PRP and PPP samples were performed using the same hemocytometer. Platelet-poor plasma samples were added to the PRP samples to obtain a total platelet count of 2 × 10 5 /ml. These platelet-equalized plasma (PEP) samples were analyzed to determine in vitro platelet function, using ADP, Collagen and Epinephrine (all from Bio/Data Corp. Horshram, Germany) in an automated aggregometer (Bio/Data Corp. Platelet Agregation Profiler, Horshram-Germany). Results were evaluated and calculated using maximum aggregation ratios. PRP from healthy controls was centrifuged for 10 minutes, after which the supernatant was decanted, and normal saline added to the platelet fraction and centrifuged for 10 minutes. This procedure was repeated three times. Plasma obtained from patients with MS was added to the PPP from the control group, and platelet function tests were performed, using the same procedures mentioned above. Statistical analyses Chi Square tests were employed to compare categorical variables, while Mann Whitney U tests were used, where appropriate, in the univariate analysis. Platelet aggregation changes were evaluated by repeated-measures analysis of variance for intra-group comparisons. Values of p < 0.05 were considered significant. SPSS 10.0 Statistical software was used in the statistical analysis. Results There were no significant differences between Groups P and N for the following variables: age, gender, current smoker, platelet, white blood cell and hemoglobin count. Clinical characteristics of the study groups are shown in Table 1 . The peripheral venous fibrinogen levels in patients with MS (301.9 ± 53.7 mg/dL) were not significantly different from those in control subjects (284.6 ± 55.7 mg/dL). Table 1 Clinical features of the study groups and echocardiographic findings of patients with mitral stenosis. Group P (n = 16) Group N (n = 16) p Mean age (yrs) 27.8 ± 6.5 28.3 ± 6.1 NS Male/female (n) 2/14 3/13 NS Current smoker (n) 5 6 NS Platelet count × 10 3 /mm 3 253.1 ± 33.5 245.6 ± 34.3 NS WBC/mm 3 5706 ± 1266 5643 ± 1229 NS Hb (g/L) 12.4 ± 1.26 12.8 ± 1.22 NS Fibrinogen (mg/dL) 301.9 ± 53.7 284.6 ± 55.7 NS LA diameter (cm) 4.22 ± 0.19 - - MVA (cm 2 ) 1.67 ± 0.16 - - TMmPG (mmHg) 6.5 ± 0.96 - - Mitral regurgitation (n) None 4 - - Mild 11 - - Moderate 1 - - Group P: Patients with mitral stenosis, Group N: Control subjects, WBC: White blood cell, Hb: Hemoglobin, LA: Left atrium, MVA: Mitral valve area, TMmPG: Transmitral mean pressure gradient, NS: Non-significant. Platelet activation tests Adenosine diphosphate, collagen and epinephrine-induced platelet aggregation ratios were significantly higher in Group P than in Group N (84.9 ± 9.8, 76.6 ± 21.4, 80 ± 10.5 and 53.1 ± 23.2, 47.5 ± 14.5, 53.7 ± 21.3, p < 0.05, respectively). When the PPP obtained from Group P was added to Group N subjects' platelets, and platelet function tests were repeated, ADP and collagen-induced (73 ± 17.5 and 65.7 ± 17.5, respectively), but not epinephrine-induced (61.9 ± 16.2), aggregation ratios were statistically increased compared to baseline (Figure 1 ). Figure 1 Induced platelet aggregation ratios in Groups P and N. Group P: Patients with mitral stenosis, Group N: Control subjects, PPP: Platelet-poor plasma, Plt: platelet, ADP: Adenosine diphosphate, COL: Collagen, EPI: Epinephrine. Discussion Left atrial thrombus in MS is associated with increased systemic levels of peptide by-products from the coagulation cascade [ 10 , 11 ]. Increases in markers of coagulation activity may also occur in the absence of thrombus, perhaps signifying that an imbalance in hemostatic regulation favoring pro-coagulant mechanisms may precede and predispose to thrombus formation [ 12 , 13 ]. Information about coagulation activity in patients with MS could, therefore, provide a method of identifying patients at risk of developing systemic thromboembolism. A number of studies have reported that indexes of hypercoagulation derived from peripheral blood may not reflect intracardiac thrombogenesis. Yamamoto et al. [ 6 ] noted that fibrinopeptide A and thrombin/antithrombin III were higher in the left atrium compared to the right atrium and the femoral vein. Peverill et al. [ 5 ] found that levels of prothrombin fragments 1+2 increased in the left atrium in comparison to peripheral blood levels. Li-Saw-Hee et al. [ 7 ] reported no significant variation in indexes of thrombogenesis, platelet activation, and endothelial dysfunction between left atrium, right atrium, and the peripheral artery or vein. A number of differences between these studies and our study may, in part, account for these varying observations. All of our patients exhibited newly diagnosed, mild-to- moderate MS, and none was taking anticoagulants or anti-platelet drugs. Second, we did not investigate systemic levels of peptide by-products of the coagulation cascade or individual blood components that reflect coagulability. In our experimental design, we sought to examine the effect of mitral stenotic patients' plasma on control subjects' platelets. In addition to protein factors, platelets play a vital role as systemic components of the hemostatic system. They are critical to the activation of intrinsic pathway factors. Because platelet activation seems to be influenced by diabetes mellitus [ 14 ], smoking [ 15 ], hypertension and the use of oral contraceptives [ 16 ], patients with these conditions were excluded from our study. Hwang et al. [ 17 ] sought to examine a correlation among echocardiographic variables, hematologic parameters or platelet aggregability and the occurrence of spontaneous echo-contrast (SEC) in the left atrium. Platelet aggregability was evaluated with a turbidometric method, using different concentrations of activating agents. No significant difference was found in platelet aggregability between patients with left atrial SEC and patients without left atrial SEC. Neither group of patients had been receiving anti-platelet or anticoagulant therapy. Ileri et al. [ 4 ] observed hypercoagulation in patients with mitral stenosis and sinus rhythm when SEC was present. Pongratz et al. [ 18 ] investigated the activation status of platelets in the peripheral blood of patients with atrial fibrillation. A significantly higher level of circulating platelets expressing P-selectin and CD63 and more leukocyte-platelet conjugates were found in patients positive for both SEC and left atrial thrombus or embolic events. Increased spontaneous platelet aggregation in the presence of SEC was described by Rohmann et al. [ 19 ], who measured the half-maximal formation of platelet aggregates in peripheral blood, upon stimulation with ADP. In our study, all patients had mild-to-moderate MS, with functional capacities rated as being NYHA class I. Although peripheral fibrinogen levels were similar between Group P and N, platelet aggregation was significantly higher in Group P than in Group N. Moreover, alteration of platelet aggregation in Group N, triggered by the addition of plasma from mitral stenotic patients, may indicate the presence of possible transferable platelet activators in peripheral blood. In conclusion, we examined systemic coagulation activity in patients with MS by measuring plasma level of fibrinogen and by performing platelet activation tests in peripheral venous blood samples. Our examinations yielded three main findings. First, mitral stenotic patients have similar fibrinogen levels in peripheral venous blood as compared to controls. Second, platelet aggregation is higher in patients with MS than in controls. Third, in comparison to controls, patients with MS have increased systemic coagulation activity and the plasma extracted from these patients seems to have some transferable factors that activate platelets. The results from this study emphasize the complexity of the events related to platelet activation in MS. Also compounding the challenges faced in this area is uncertainty about the gold standard method for assessing platelet activation and whether platelet aggregation ( ex vivo ) truly reflects platelet activation in vivo [ 20 ].
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553988.xml
551609
Customized birth weight for gestational age standards: Perinatal mortality patterns are consistent with separate standards for males and females but not for blacks and whites
Background Some currently available birth weight for gestational age standards are customized but others are not. We carried out a study to provide empirical justification for customizing such standards by sex and for whites and blacks in the United States. Methods We studied all male and female singleton live births and stillbirths (22 or more weeks of gestation; 500 g birth weight or over) in the United States in 1997 and 1998. White and black singleton live births and stillbirths were also examined. Qualitative congruence between gestational age-specific growth restriction and perinatal mortality rates was used as the criterion for identifying the preferred standard. Results The fetuses at risk approach showed that males had higher perinatal mortality rates at all gestational ages compared with females. Gestational age-specific growth restriction rates based on a sex-specific standard were qualitatively consistent with gestational age-specific perinatal mortality rates among males and females. However, growth restriction patterns among males and females based on a unisex standard could not be reconciled with perinatal mortality patterns. Use of a single standard for whites and blacks resulted in gestational age-specific growth restriction rates that were qualitatively congruent with patterns of perinatal mortality, while use of separate race-specific standards led to growth restriction patterns that were incompatible with patterns of perinatal mortality. Conclusion Qualitative congruence between growth restriction and perinatal mortality patterns provides an outcome-based justification for sex-specific birth weight for gestational age standards but not for the available race-specific standards for blacks and whites in the United States.
Background Birth weight-specific perinatal mortality curves among male and female births intersect to produce a paradox: overall perinatal mortality rates and perinatal mortality rates at lower birth weights are relatively higher among male births, while at higher birth weights perinatal mortality rates are relatively higher among female births [ 1 ]. This puzzling observation reflects a general phenomenon that is also seen when birth weight- and gestational age-specific perinatal mortality curves are contrasted across race, plurality, maternal smoking status, parity, altitude, country, and other determinants of birth weight and gestational age [ 2 - 14 ]. We have previously presented a solution for this paradox of intersecting mortality curves that involves a reformulation of perinatal and neonatal mortality risk [ 15 - 20 ]. This reformulation, based on the fetuses at risk approach, eliminates the crossover phenomenon and provides several new insights into perinatal health issues. In this paper, we demonstrate the paradoxical crossover of birth weight-specific perinatal mortality curves among male and female births and show how this phenomenon is resolved using the fetuses at risk approach. We also explore issues related to fetal growth restriction among males and females using the same approach. This latter issue is particularly important from a conceptual and clinical standpoint because the current literature on birth weight for gestational age standards (sometimes referred to as fetal growth standards) is confusing. Some standards provide unisex reference values [ 21 - 24 ], several are sex-specific [ 1 , 25 - 34 ] and yet others provide both sex-specific and unisex reference values [ 35 - 38 ]. Of equal concern is the fact that several standards are customized for different races [ 1 , 25 , 27 - 29 ], parity [ 25 , 27 , 29 , 34 , 36 ], plurality [ 24 , 30 ] and other characteristics [ 27 ], while others are not [ 21 - 23 , 26 , 31 - 33 , 35 , 37 ]. We used the fetuses at risk approach to contrast growth restriction and perinatal mortality rates among males and females in order to provide empirical justification for sex-specific (vs unisex) birth weight for gestational age standards. We also constructed and compared gestational age-specific growth restriction and perinatal mortality curves among whites vs blacks in order to evaluate currently available birth weight for gestational age standards (single standard vs separate standards for whites and blacks in the United States). Methods We used data on all reported live births and stillbirths in the United States in 1997 and 1998 (National Center for Health Statistics perinatal mortality data file for all states and the District of Columbia for 1997 and 1998). Live births and infant death records for these years have been previously linked and gestational duration has been calculated based on the last menstrual period (LMP). Missing or inconsistent information on gestational age has been imputed or replaced in a small fraction (approximately 7 percent) of records by the National Center for Health Statistics (Hyattsville, Maryland). Gestational age was imputed from the month and year of the LMP when the exact LMP day was missing [ 39 ]. LMP-based gestational age information was replaced by the clinical estimate [ 40 ] when the former was inconsistent with birth weight or when there was no information on LMP (approximately 5 percent of births). Analyses were restricted to singleton live births and stillbirths ≥22 weeks gestational age and ≥500 g birth weight in order to eliminate potential problems arising from regional differences in birth registration. Male and females births were first contrasted in terms of their gestational age and birth weight distributions. Birth weights were categorized into 500 g intervals for this purpose (500–999 g, 1,000–1,499 g, 1,500–1,999 g and so on). Birth weight-specific perinatal mortality rates, calculated within these birth weight categories, were computed as per convention by dividing the number of stillbirths and early neonatal (0 to 6 days) deaths in any birth weight category by the number of total births (stillbirths and live births) in that birth weight category. Similarly, gestational age-specific perinatal mortality rates among male and female births were contrasted, with rates computed by dividing perinatal deaths at any given gestation by the number of total births at that gestation. The numbers of fetuses at risk for stillbirth and early neonatal death at each gestation were then used to calculate a second set of perinatal mortality rates. Under this fetuses at risk formulation, the stillbirth rate at 28 weeks gestation was computed by dividing the number of stillbirths at 28 weeks by the number of live births and stillbirths at 28 or more completed weeks of gestation. This implies that fetuses who delivered at 29, 30, 31 and 32 or more weeks gestation were also at risk of stillbirth at 28 weeks [ 15 - 19 , 41 - 44 ]. The fetuses at risk formulation applies equally to early neonatal death since a fetus (unborn) at 28 weeks gestation is at risk of birth and early neonatal death at that gestation [ 15 , 17 , 18 ]. Thus gestational age-specific perinatal/neonatal mortality rates under this formulation were calculated with perinatal/neonatal deaths at any gestational age in the numerator and the fetuses at risk of perinatal/neonatal death at that gestation in the denominator. This represents a survival analysis model with censoring of subjects (fetuses) at death or birth which ever occurs earlier (for a schematic depiction of the survival analysis model, see reference 18). In this model, neonatal death (and, in other contexts, serious pregnancy-related morbidity such as cerebral palsy [ 16 ]) is assigned to the point of birth since the responsible pathologic event/process is present at birth [ 18 ]. Gestational age-specific 'birth rates' (i.e., the number of births at any particular gestational week divided by the number of fetuses at risk of birth at that gestation) and rates of gestational age-specific labor induction/cesarean delivery were also estimated using the fetuses at risk approach [ 15 - 18 ]. We also examined gestational age-specific patterns of fetal growth restriction using the fetuses at risk approach [ 15 , 17 - 19 ]. The number of small-for-gestational age (SGA) live births at each gestation was divided by the number of fetuses at risk at that gestation in order to obtain the gestational age-specific SGA rate (or the gestational age-specific fetal growth restriction rate). SGA live births were identified using the 10 th percentile cut-off from a birth weight for gestational age standard based on live births in the United States [ 38 ]. Gestational age-specific SGA rates were calculated using both the unisex and sex-specific 10 th percentile values provided by this standard [ 38 ] to evaluate how well patterns of gestational age-specific growth restriction correspond with patterns of gestational age-specific perinatal mortality. This evaluation was premised on the belief that fetal growth restriction patterns should be qualitatively congruent with gestational age-specific perinatal mortality patterns. Such an expectation is consistent with clinical understanding and studies which show that growth restricted fetuses have a substantially higher perinatal mortality than appropriate-for-gestational age fetuses. For instance, Williams et al [ 1 ] showed that perinatal mortality at each gestational week was much higher among growth restricted births at the 10 th percentile of birth weight for gestational age (eg., perinatal mortality rate 138 per 1,000 total births at 34–35 weeks) compared with appropriate-for-gestational age births at the 50th percentile of birth weight for gestational age (eg., perinatal mortality rate 27 per 1,000 total births at 34–35 weeks). We also examined gestational age-specific growth restriction differences among males and females using rate ratios (eg., growth restriction rate among males at 35 weeks gestation divided by growth restriction rate among females at 35 weeks gestation) and contrasted these with gestational age-specific differences in stillbirth and neonatal mortality rates (also using rate ratios eg., stillbirth rate among males at 35 weeks divided by the stillbirth rate among females at 35 weeks; early neonatal death rate among males at 35 weeks divided by the early neonatal death rate among females at 35 weeks). This was done to ascertain the relationship between patterns of growth restriction and patterns in the two components of perinatal mortality (stillbirth and early neonatal death). Comparisons of male and female gestational age-specific growth restriction and gestational age-specific perinatal mortality patterns were contrasted with similar comparisons according to maternal race. Specifically, live births and stillbirths ≥22 weeks of gestational age and ≥500 g birth weight in the United States in 1997 and 1998 were used to compare gestational age-specific growth restriction and perinatal mortality rates among whites vs blacks. Identification of SGA live births among blacks and whites was carried out using a single standard for both races [ 38 ] and also a race-specific standard [ 29 ]. As with contrasts between males and females, the contrasts between whites and blacks were restricted to singleton births. Differences in rates were assessed using rate ratios and excess risks. Taylor series 95% confidence intervals were calculated on all rate ratios. All p values presented are two-sided. Sensitivity analyses were carried out to assess the potential effect of gestational age errors on patterns of growth restriction and perinatal mortality among males and females. Specifically, we reassessed growth restriction and mortality patterns among males and females after excluding all births for whom menstrual-based gestational age was either imputed or replaced by the clinical estimate of gestation. Results There were 3,905,694 singleton male births in the United States in 1997 and 1998 (≥22 weeks gestational age and ≥500 g birth weight). The low birth weight (<2,500 g) rate among male live births was 5.5%, and 10.5% of male live births were born preterm (<37 weeks). There were 3,723,153 female births in the United States during the same period and relative to males, female live births had a higher rate of low birth weight (6.4%, p < 0.0001) but a lower rate of preterm birth (9.4%, p < 0.0001). Males had a 14% (95% confidence interval 12 to 16, p < 0.0001) higher perinatal mortality than females; perinatal mortality rates among males and females were 6.78 and 5.95 per 1,000 total births, respectively. The gestational age distribution of male live births (Figure 1 ) was 'shifted to the left' relative to female live births (p < 0.0001), while the birth weight distribution of females was markedly 'shifted to the left' relative to that of male live births (p < 0.0001). Birth weight-specific perinatal mortality rates (conventional calculation, perinatal deaths per 1,000 total births in a given birth weight category) showed the crossover paradox with males having relatively higher rates of perinatal death at birth weights <4,000 g, while females had relatively higher perinatal mortality rates at higher birth weights (Figure 2a ). In contrast, gestational age-specific perinatal mortality rates (conventional calculation, perinatal deaths per 1,000 total births at any gestational week) showed similar mortality patterns among males and females (Tables 1 and 2 ), with males having a slightly higher perinatal mortality rate at some gestational ages (Figure 2b ). Figure 1 Gestational Age and Birth Weight Distributions of Male and Female Singleton Live Births. Gestational age (1a) and birth weight (1b) distributions of male and female singleton live births ≥22 weeks and ≥500 g in the United States, 1997 and 1998. Figure 2 Conventional Calculation: Birth Weight- and Gestational Age-Specific Perinatal Mortality Rates among Male and Female Births. Conventional calculation: birth weight-specific (2a) and gestational age-specific (2b) perinatal mortality rates per 1,000 total births among male and female singleton births in the United States, 1997 and 1998. Table 1 Gestational Age-Specific Numbers and Rates of Perinatal Death among Male Singleton Births, United States, 1997 and 1998. Gestational age Stillbirths Live births Early neonatal deaths Perinatal mortality rate (1)† Fetuses at risk Perinatal mortality rate (2)† 28 648 6,808 263 122.2 3,841,944 0.24 29 579 8,100 221 92.2 3,834,488 0.21 30 701 11,297 230 77.6 3,825,809 0.24 31 668 14,339 208 58.4 3,813,811 0.23 32 809 20,242 209 48.4 3,798,804 0.27 33 786 30,140 212 32.3 3,777,753 0.26 34 881 51,673 298 22.4 3,746,827 0.31 35 915 86,166 338 14.4 3,694,273 0.34 36 1,033 154,986 354 8.9 3,607,192 0.38 37 1,144 308,629 394 5.0 3,451,173 0.45 38 1,173 626,450 470 2.6 3,141,400 0.52 39 1,122 925,764 541 1.8 2,513,777 0.66 40 897 848,527 431 1.6 1,586,891 0.84 41 469 444,468 237 1.6 737,467 0.96 ≥42* 454 292,076 229 2.3 292,530 2.33 Total‡ 17,680 3,888,014 8,800 6.8 3,905,694 6.78 † Total births at each gestational week served as the denominator for perinatal mortality rates (1), while perinatal mortality rates (2) were calculated using fetuses at risk as the denominator (see text). All rates are expressed per 1,000. * Large increase in perinatal mortality (2) at ≥42 weeks is partly because the period of risk exceeds 1 week (see also Figures 3-5). ‡ All gestational ages, including those ≥22 weeks and those with missing gestational age. Table 2 Gestational Age-Specific Numbers and Rates of Perinatal Death among Female Singleton Births, United States, 1997 and 1998. Gestational age Stillbirths Live births Early neonatal deaths Perinatal mortality rate (1)† Fetuses at risk Perinatal mortality rate (2)† 28 614 5,838 184 123.7 3,665,497 0.22 29 530 7,000 158 91.4 3,659,045 0.19 30 611 9,742 179 76.3 3,651,515 0.22 31 578 12,493 173 57.5 3,641,162 0.21 32 632 17,168 195 46.5 3,628,091 0.23 33 654 25,282 187 32.4 3,610,291 0.23 34 747 44,275 221 21.5 3,584,355 0.27 35 796 75,238 234 13.5 3,539,333 0.29 36 874 133,386 286 8.6 3,463,299 0.33 37 927 267,501 337 4.7 3,329,039 0.38 38 1,071 563,676 361 2.5 3,060,611 0.47 39 1,072 885,523 419 1.7 2,495,864 0.60 40 926 851,848 376 1.5 1,609,269 0.81 41 505 455,313 222 1.6 756,495 0.96 ≥42* 443 300,234 198 2.1 300,677 2.13 Total‡ 15,537 3,707,616 6,614 6.0 3,723,153 5.95 † Total births at each gestational week served as the denominator for perinatal mortality rates (1), while perinatal mortality rates (2) were calculated using fetuses at risk as the denominator (see text). All rates are expressed per 1,000. * Large increase in perinatal mortality (2) at ≥42 weeks is partly because the period of risk exceeds 1 week (see also Figures 3-5). ‡All gestational ages, including those ≥22 weeks and those with missing gestational age. Gestational age-specific perinatal mortality rates calculated using the fetuses at risk approach showed that perinatal mortality rates increased with increasing gestational age (Figure 3 ). Males had a higher perinatal mortality than females at virtually all gestational ages (Tables 1 and 2 ). Gestational age-specific 'birth rates' (Figure 3a ), gestational age-specific labor induction rates (Figure 3b ) and gestational age-specific labour induction and/or cesarean delivery rates (data not shown) were marginally (but consistently) higher among pregnancies with males as compared with pregnancies with females (Figure 3 ). For example, the birth rate among males at 35 weeks gestation was 23.6 per 1,000 fetuses at risk, while that among females at 35 weeks was 21.5 per 1,000 fetuses at risk (rate ratio 1.10, 95% confidence interval 1.09 to 1.11, p < 0.0001). The labour induction rates at 35 weeks among males and females were 3.6/1,000 and 3.1/1,000 fetuses at risk, respectively; rate ratio 1.10, 95% confidence interval 1.07 to 1.13, p < 0.0001. Figure 3 Fetuses at Risk Approach: Gestational Age-Specific Birth, Labor Induction and Perinatal Mortality Rates among Male and Female Births. Fetuses at risk approach: Gestational age-specific birth rates (3a, primary Y-axis), labor induction rates (3b, primary Y-axis) and perinatal mortality rates (3a and 3b, secondary Y-axis) among male and female singleton births in the United States, 1997 and 1998. Figure 4 compares gestational age-specific rates of fetal growth restriction among males and females. When growth restriction was determined using a sex-specific standard, growth restriction rates among males were higher than growth restriction rates among females at all gestational ages and this pattern was qualitatively congruent with sex differences in perinatal mortality (Figure 4a ). For instance, males at 35 weeks gestation had an 8 percent (95% confidence interval 5 to 11, p < 0.0001) higher growth restriction rate than females at the same gestational week (sex-specific standard) and this was qualitatively congruent with a 17 percent (95% confidence interval 7 to 27 percent, p = 0.0003) higher perinatal death rate among males compared with females at 35 weeks gestation. On the other hand, when a unisex standard was used to identify growth restricted live births, males had a lower rate of growth restriction at all gestational ages and this was not qualitatively congruent with the higher gestational age-specific pattern of perinatal mortality among males (Figure 4b ). For instance, at 35 weeks gestation, growth restriction rates determined using a single standard for both males and females showed that males had a 20 percent (95% confidence interval 18 to 22 percent, p < 0.0001) lower rate of growth restriction compared with females (not consistent with the 17% higher perinatal mortality rate). Figure 4 Fetuses at Risk Approach: Gestational Age-Specific Growth Restriction and Perinatal Mortality Rates among Male and Female Births. Fetuses at risk approach: Gestational age-specific fetal growth restriction (primary Y-axis) and perinatal mortality rates (secondary Y-axis) among male and female singleton births, with growth restriction rates based on sex-specific (4a) and unisex (4b) birth weight for gestational age standards, United States, 1997 and 1998. Overall growth restriction rates based on a sex-specific standard showed that rates were 3% (95% CI 2 to 3) higher among males. Stillbirth and early neonatal mortality differences (rate ratios) among male vs female births both favored females (Table 3 ), although the mortality differences were much larger for early neonatal mortality (27%, 95% CI 23 to 31) than for stillbirth (8%, 95% CI 6 to 11). Gestational age-specific differences in growth restriction between males and females based on a sex-specific standard (eg., rate ratio at 35 weeks 1.08, 95% CI 1.05 to 1.11, Table 3 ) tended to be similar to gestational age-specific differences in stillbirth rates (eg., rate ratio at 35 weeks 1.10, 95% CI 1.00 to 1.21, Table 3 ), while differences in gestational age-specific early neonatal mortality tended to be larger (eg., rate ratio at 35 weeks 1.38, 95% CI 1.17 to 1.63, Table 3 ). Sensitivity analyses carried out to examine the potential effect of gestational age errors (by excluding births among whom gestational age was imputed or for whom the clinical estimate of gestation was used) showed essentially the same patterns of growth restriction and perinatal mortality among males and females. Table 3 Gestational Age-Specific Rates of Fetal Growth Restriction Based on a Sex-Specific Standard [38] and Differences in Growth Restriction, Stillbirth and Early Neonatal Mortality Among Males and Females, Singleton Births, United States, 1997 and 1998. Gestational age Fetal growth restriction Stillbirth rate ratio(males vs females) Early neonatal mortality rate ratio(males vs females) Males Females Rate ratio (males vs females) Number Rate † Number Rate † 28 631 0.2 502 0.1 1.20 1.01 1.36 29 764 0.2 660 0.2 1.10 1.04 1.33 30 1,165 0.3 923 0.3 1.20 1.10 1.23 31 1,539 0.4 1,369 0.4 1.07 1.10 1.15 32 2,141 0.6 1,866 0.5 1.09 1.22 1.02 33 3,294 0.9 2,946 0.9 1.07 1.15 1.08 34 5,691 1.6 5,098 1.5 1.07 1.13 1.29 35 8,934 2.5 7,922 2.3 1.08 1.10 1.38 36 15,813 4.6 13,910 4.2 1.09 1.13 1.19 37 30,029 9.1 25,180 7.9 1.15 1.19 1.13 38 55,401 18.5 49,599 17.1 1.09 1.07 1.27 39 84,257 35.7 79,440 33.9 1.05 1.04 1.28 40 75,983 52.8 74,504 51.1 1.03 0.98 1.16 41 36,956 62.6 34,214 56.5 1.11 0.95 1.10 ≥42 14,679 100.5 14,551 96.7 1.04 1.05 1.19 Total‡ 337,277 91.6 312,684 89.3 1.03 1.08 1.27 † Gestational age-specific growth restriction rates (based on a sex-specific standard [38]) were calculated by dividing the number of small-for-gestational age live births (<10 th percentile) at any gestational age by the number of fetuses at risk at that gestation. Stillbirth and early neonatal mortality rates were also calculated using fetuses at risk as the denominator. ‡ All gestational ages ≥22 weeks, except for growth restriction indices which were based on live births between 28 and 42 weeks. Patterns of gestational age-specific growth restriction among whites and blacks could not be reconciled with patterns of gestational age-specific perinatal mortality, when growth restriction was defined by a race-specific standard (Figure 5a ). Growth restriction rates defined using the race-specific birth weight for gestational age standard showed a crossover with blacks having significantly higher growth restriction rates than whites below 39 weeks and significantly lower growth restriction rates at 39 weeks and over. For instance, rates of growth restriction as defined by the race-specific standard were significantly lower among blacks compared with whites at 40 weeks gestation (rate ratio 0.89, 95% confidence interval 0.88 to 0.91, p < 0.0001), despite a significantly higher perinatal mortality rate among blacks at 40 weeks gestation (rate ratio 1.43, 95% confidence interval 1.29 to 1.58, p < 0.0001). On the other hand, rates of gestational age-specific growth restriction were qualitatively congruent with patterns of gestational age-specific perinatal mortality when growth restriction among blacks and whites was defined using a single birth weight for gestational age standard (Figure 5b ). For example, at 40 weeks gestation, the significantly higher rate of perinatal death among blacks was consistent with the significantly higher rate of growth restriction seen among blacks when a single standard was used to define growth restriction (rate ratio for growth restriction at 40 weeks among blacks vs whites 2.06, 95% confidence interval 2.04 to 2.09, p < 0.0001). Growth restriction (based on a single standard for both races) and perinatal mortality rates were substantially higher among births to black mothers as compared with births to white mothers at all gestational ages (Figure 5b ). Figure 5 Fetuses at Risk Approach: Gestational Age-Specific Fetal Growth Restriction and Perinatal Mortality Rates among White and Black Births. Fetuses at risk approach: Gestational age-specific fetal growth restriction (primary Y-axis) and perinatal mortality rates (secondary Y-axis) among white and black singleton births, with growth restriction rates based on a race-specific standard (5a) and on a single birth weight for gestational age standard (5b), United States, 1997 and 1998. Discussion We have confirmed previous observations that birth weight-specific perinatal mortality rates among male and female births exhibit a puzzling crossover paradox [ 1 ]. Gestational age-specific perinatal mortality rates among males and females were similar when mortality rates were calculated per convention (using total births at a particular gestation for calculating the perinatal mortality rate). On the other hand, use of the fetuses at risk formulation [ 15 - 19 , 41 - 44 ] showed that males have a consistently higher perinatal mortality rate at all gestational ages. Further, our study shows that gestational age-specific growth restriction and perinatal mortality rates both increase with advancing gestational age. Gestational age-specific rates of growth restriction among males and females are qualitatively congruent with gestational age-specific perinatal mortality patterns when growth restriction rates are based on a sex-specific birth weight for gestational age standard. Use of a single standard for males and females results in a gestational age-specific pattern of growth restriction that cannot be reconciled with gestational age-specific differences in perinatal mortality among males and females. In contradistinction, contrasts between whites vs blacks show that use of a single birth weight for gestational age standard for both races is justified, while the use of a currently available race-based standard is not defensible. Gestational age-specific growth restriction patterns among whites vs blacks based on a single standard correspond qualitatively to patterns of gestational age-specific perinatal mortality among whites and blacks (Figure 5 ). Birth weight for gestational age standards are modeled after infant and child growth standards and assume that fetal growth restriction occurs at a constant rate throughout pregnancy. This assumption is implicit in the use of the same, fixed cut-off (eg., the 3 rd percentile or the 10 th percentile cut-off of birth weight for gestational age) for identifying fetal growth restriction at all gestational ages. Our findings challenge the former assumption and show that in fact fetal growth restriction rates are better viewed as increasing with advancing gestational age (Figures 4 and 5 ). This contention is supported by the finding that gestational age-specific growth restriction rates follow the pattern of gestational age-specific perinatal mortality rates. Recent studies which show that the incidence of hypertensive disorders and chorioamnionitis increases with increasing gestational age provide at least a partial explanation for the gestational age-dependent rise in fetal growth restriction and perinatal mortality rates [ 45 , 46 ]. Table 3 shows that differences in stillbirth rates between males and females are smaller than differences in early neonatal mortality rates. The phenomenon of higher neonatal mortality differentials (relative to stillbirth differentials) between males and females has been previously noted [ 1 ] and is probably a consequence of obstetric intervention. Obstetric intervention (i.e., early delivery through labor induction and/or cesarean delivery) is typically prompted by signs of fetal compromise and will be more likely among pregnancies with male fetuses given the male fetuses' greater biological vulnerability. Such intervention leads to a reduction in the stillbirth differential, while having a smaller (or the opposite) effect on neonatal mortality differences between males and females. This explanation is supported by the higher rates of labor induction (and labour induction and/or cesarean delivery) observed among pregnancies with male fetuses (Figure 3b ). Differences in rates of congenital anomalies that are lethal after birth and more frequent in males (eg., X-linked recessive conditions) may partly contribute to this phenomenon as well. The slightly higher rate of gestational age-specific labor induction/cesarean delivery among males relative to females is encouraging since it suggests that the small mortality risk difference between males and females is already being addressed by modern obstetric practice (despite male sex not being formally identified as a factor in decision making related to obstetric intervention). This may be a consequence of the use of sex-specific birth weight for gestational age standards or sex-specific ultrasound-based fetal growth standards and, as mentioned, probably also reflects higher rates of suspected fetal compromise among pregnancies with male fetuses. Despite the marginally higher rates of labor induction among pregnancies with male fetuses, however, mortality differences persist. Research should be directed at ascertaining whether excess neonatal mortality among males can be successfully reduced through explicit recognition of male sex as a factor for altering the threshold for obstetric intervention. Although contemporary birth weight for gestational age standards have substantial face validity [ 1 , 47 , 48 ], their development would benefit from greater empirical support and validation. For instance, it should be feasible to refine standards based on empirically observed (cause-specific) patterns of birth weight-specific perinatal mortality and serious neonatal morbidity (at each gestational age). This would represent an improvement over current standards which rely heavily on theoretical assumptions (eg., normality of birth weight at any given gestational age) and insufficiently on relevant empirical information (namely, perinatal morbidity and mortality related to growth restriction). Such cross-sectional information cannot address fetal growth in continuing pregnancies, however; the latter requires longitudinal information which is ideally obtained through ultrasonographic measurements. On the other hand, estimation of fetal weight through ultrasonography [ 31 , 49 ] needs to be improved [ 50 , 51 ] and diagnostic methods for identifying fetal growth restriction have tended to rely on other indicators of growth restriction besides estimated fetal weight. Our study has limitations that are typical of studies that use large data bases. Errors in gestational age information are inevitable, although the magnitude of these errors is likely to be similar among male and female births. The overall rate of missing gestational age was low, however (0.9 percent among white live births and 0.8 percent among black live births). Our estimates of gestational age-specific fetal growth restriction rates are approximate. Ideally, estimation of the incidence of fetal growth restriction requires identification of fetal growth restriction on a longitudinal basis among continuing pregnancies [ 18 ]. The alternative measure of gestational age-specific growth restriction employed in our study represents an index of 'revealed' fetal growth restriction [ 18 ]. This approximation is unlikely to be a factor that seriously distorts patterns of gestational age-specific growth restriction since faltering of fetal growth typically leads to a spontaneous delivery or delivery following obstetric intervention. Other potential limitations of our study include the use of gestational age information on stillbirths. The gestational age at delivery of a stillbirth typically overestimates the gestational age at the time of fetal death, although this difference is unlikely to be large in recent years. Further, both male and female stillbirths would have been affected by this measurement error to a similar extent. Conclusion The fetuses at risk approach resolves the paradox of intersecting perinatal mortality curves. Male births have higher rates of gestational age-specific perinatal mortality than female births. There is empirical justification for using sex-specific standards of birth weight for gestational age since gestational age-specific growth restriction patterns based on such standards correspond qualitatively with gestational age-specific perinatal mortality patterns. On the other hand, a single birth weight for gestational age standard for whites and blacks in the United States appears more appropriate than currently available race-specific standards since gestational age-specific growth restriction patterns among blacks and whites (based on a single standard) are qualitatively congruent with gestational age-specific patterns of perinatal mortality. Competing interests The author(s) declare that they have no competing interests. Authors' contributions KSJ proposed the study, carried out the analyses and drafted the manuscript. The results of the analyses were presented and discussed at a meeting of the Fetal and Infant Health Study Group of the Canadian Perinatal Surveillance System. All authors contributed to revising the manuscript for intellectual content. All authors read and approved the final version. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551609.xml
546413
Linking biogeography to physiology: Evolutionary and acclimatory adjustments of thermal limits
Temperature-adaptive physiological variation plays important roles in latitudinal biogeographic patterning and in setting vertical distributions along subtidal-to-intertidal gradients in coastal marine ecosystems. Comparisons of congeneric marine invertebrates reveal that the most warm-adapted species may live closer to their thermal tolerance limits and have lower abilities to increase heat tolerance through acclimation than more cold-adapted species. In crabs and snails, heart function may be of critical importance in establishing thermal tolerance limits. Temperature-mediated shifts in gene expression may be critical in thermal acclimation. Transcriptional changes, monitored using cDNA microarrays, have been shown to differ between steady-state thermal acclimation and diurnal temperature cycling in a eurythermal teleost fish ( Austrofundulus limnaeus ). In stenothermal Antarctic notothenioid fish, losses in capacity for temperature-mediated gene expression, including the absence of a heat-shock response, may reduce the abilities of these species to acclimate to increased temperatures. Differences among species in thermal tolerance limits and in the capacities to adjust these limits may determine how organisms are affected by climate change.
Review Introduction Understanding the roles played by physiological adaptation to temperature in governing the distribution patterns of species has taken on a new urgency because of the potential effects of global warming on both aquatic and terrestrial ecosystems. There is already compelling evidence for widespread changes in ecosystems due to climate change. Recent meta-analyses have shown that species' distribution patterns, the structures of ecosystems, and the timing of annual events (phenology) such as migration and reproduction have changed markedly during the past century, when average global temperatures rose by approximately 0.6°C [ 1 - 4 ]. If the consensus view that average global temperatures will rise by approximately 3°C during this century is correct [ 5 ], then even more extensive changes in the biosphere are certain to occur and, in view of this five-fold increase in warming rate, to occur rapidly. Suffice it to say that analyses that can assist in predicting – and, perhaps, even in ameliorating – these biological effects are strongly needed at this time. Because most previous analyses of the biological effects of climate change have been correlative, rather than focused on the underlying causal mechanisms behind the observed effects, there is clearly a need for physiologists to contribute to this important, on-going discussion. In this brief review, I discuss studies performed with a taxonomically varied group of aquatic ectotherms that shed light on several questions related to the roles played by physiological adaptations in setting species' distribution patterns. The more clearly we understand the mechanistic basis of biogeography, the better prepared we will be to predict the effects of climate change on distribution patterns and ecosystem structure. The questions I address include the following. First, how do thermal tolerance limits differ among species adapted to different temperatures? Which species are most likely to be threatened by increases in habitat temperatures, i.e., which species currently live closest to their thermal tolerance limits? Second, what are the "weak links" in physiological systems that appear most likely to set thermal tolerance limits? Can we identify physiological, biochemical, and molecular level effects that account for thermal tolerance ranges? Third, how do capacities for acclimation to changes in habitat temperature differ among species? Which species have the greatest – or the least – ability to acclimate to increases in temperature? Mechanistically speaking, what types of acclimatory changes are needed to adapt physiological systems to permit tolerance of new thermal conditions? Fourth, what types of changes in gene expression are needed to achieve thermal acclimation? How does gene expression during acclimation to rapid, diurnal change in temperature differ from the response to slower, seasonal time-scale change in temperature? How do stenothermal and eurythermal species differ in their capacities for altering gene expression in temperature-adaptive manners? Have extreme stenotherms like Antarctic animals that have evolved for many millions of years under highly stable, cold temperatures lost key abilities to acclimate to increasing temperatures? To address these questions, I focus primarily on studies that our laboratory has performed on aquatic ectothermic species that have been chosen for study because they seemed to fit well the August Krogh Principle, which can be paraphrased as follows: For any question a biologist asks, Nature can provide a most appropriate study organism. Through study of such "Kroghian" species, it has been possible to obtain at least initial answers to the several questions raised above. Thermal tolerance relationships of porcelain crabs (genus Petrolisthes ) and turban snails (genus Tegula ) One of the challenges facing a comparative or evolutionary physiologist wishing to elucidate adaptive differences among species is the need to separate effects due to phylogeny from true adaptive variation. One means of achieving this end is to study groups of closely related organisms from widely different habitat conditions, for which a comprehensive phylogeny exists. A group of species that is especially "Kroghian" from this standpoint are porcelain crabs (genus Petrolisthes ) from the eastern Pacific Ocean [ 6 ]. Congeners of Petrolisthes are abundant (46 species occur in the eastern Pacific), found over wide ranges of latitude, and occur in subtidal and intertidal habitats. Among these congeners, body temperatures extend from slightly below 0°C to over 40°C; for some eurythermal species, the range of body temperatures across different seasons can be over 32°C [ 7 - 10 ]. The thermal tolerance limits of field-acclimatized temperate (California and Chile), subtropical (northern Gulf of California) and tropical (Panama) congeners of Petrolisthes from different heights along the subtidal to intertidal gradient are shown in Fig. 1 . Data are presented as LT 50 values, the temperatures at which fifty percent of an experimental population is killed by the heat treatment. Heating rates simulated measured rates of environmental temperature change found in the species' habitats [ 10 ]. The predicted correlation between adaptation temperature and LT 50 is found (Fig. 1A ): the most heat-tolerant intertidal species, those from subtropical and tropical habitats, had LT 50 s that were approximately 15°C higher than those of temperate sub-tidal species and 6–7°C higher than temperate intertidal species. However, the likelihood of heat death facing the different species under natural habitat conditions is not reflected by the LT 50 per se , but rather the by proximity of this trait to current extremes of habitat temperature. As shown by the alignment of LT 50 values vis à vis the line of unity (LT 50 = maximal habitat temperature) in Fig. 1B , warm-adapted intertidal species are in much greater jeopardy from heat death than their more cold-adapted subtidal relatives. Moreover, the abilities of congeners of Petrolisthes to increase thermal tolerance through acclimation to increased temperature are less in intertidal species than in more cold-adapted subtidal congeners [ 10 ]. Thus, warm-adapted intertidal species face greater current – and, most likely, future – threats from high temperatures than less heat-tolerant, subtidal congeners. Figure 1 Heat tolerance of porcelain crabs. Upper thermal tolerance limits (LT 50 (°C)) of congeneric porcelain crabs (genus Petrolisthes ) native to eastern Pacific habitats in California, Chile, the northern Gulf of California, and Panama and occurring at different heights along the subtidal-to-intertidal gradient. Each symbol represents a different species of crab. B. Thermal tolerance versus maximal habitat temperature, with a line of unity given to show the proximity of current habitat temperatures to lethal temperatures. Hatched symbol represents P. cinctipes . The R 2 value for the regression line for all species is given. Data are from [10]. One physiological system that is a "weak link" in the thermal tolerance of these species is heart function. As shown in Fig. 2A , sharp reductions in heart rate occur when a species-specific high temperature is reached. This is termed the "Arrhenius break temperature" (ABT) to denote that it is the temperature at which a sharp discontinuity in the slope of an Arrhenius plot (ln rate of heart beat versus reciprocal temperature (K)) occurs. Once the ABT is exceeded, heart function of porcelain crabs does not recover from heat stress. ABT values for the temperate congeners P. cinctipes (intertidal) and P. eriomerus (subtidal) are 31.5°C and 26.5°C, respectively. The maximal habitat temperatures for the two species are approximately 31°C and 16°C, respectively [ 9 , 10 ]. Thus, whereas the subtidal congener has an approximately 10°C range between its highest habitat temperature and the upper thermal limit of heart function, the intertidal congener's upper habitat temperature, LT 50 , and ABT for heart function are essentially the same, 31–32°C. It bears emphasizing that ABT values for other physiological processes, for example, mitochondrial respiration [ 11 , 12 ] and enzymatic activity [ 13 ], may be considerably higher than upper lethal temperatures of the whole organism. For some species of animals, then, heart function appears to be a "weak link" in the "chain" of physiological processes that govern thermal tolerance. Figure 2 The effects of temperature on cardiac activity in porcelain crabs. Upper panel. Arrhenius plots of ln heart rate (beats per minute) versus measurement temperature (1/K) for two congeners of porcelain crabs, Petrolisthes cinctipes and P. eriomerus , having maximal habitat temperatures of approximately 32°C and 16°C, respectively. Arrhenius break temperatures (ABTs) are the temperatures at which a sharp decrease in heart rate is noted (see Stillman & Somero [9] for computational methods). Lower panel. Acclimatory-induced change in upper and lower critical temperatures (CT max and CT min , respectively) of heart function for 4 congeners of Petrolisthes : P. cinctipes and P. eriomerus from temperate central California habitats, and P. gracilis and P. hirtipes from the northern Gulf of California. CT max is the ABT, and CT min is the temperature at which heart beat ceased as temperature was lowered. Decreasing temperatures did not cause a sharp break in Arrhenius plots, so ABTs could not be determined at low temperatures [8,9]. Acclimation temperatures were 8°C and 18°C for the California species and 15°C and 25°C for the Gulf of California species. The differences in CT max and CT min between the two acclimation groups of each species are shown. Each symbol represents a different species, whose maximal habitat temperature is given on the abscissa. Figure modified after [8]. The acclimatory plasticity of ABTs of heart function differs among species in a parallel fashion to acclimatory change in LT 50 (Fig. 2B [ 8 ]). In this experiment, 4 congeners were studied: the two temperate species from California discussed above, P. cinctipes and P. eriomerus , and upper intertidal and mid-intertidal species from the northern Gulf of California, P. gracilis and P. hirtipes , respectively. Acclimation temperatures were 8°C and 18°C for the temperate species and 15°C and 25°C for the sub-tropical species. ABT (= CT max ) of P. eriomerus rose by 2.3°C when acclimation temperature was increased by 10°C, but the most warm-adapted species, P. gracilis , increased ABT (CT max ) by only 0.5°C. The eurythermal intertidal species did exhibit greater capacities than the subtidal species to acclimate to lower temperatures, however, as shown by their greater ability to lower the temperature at which heart beat ceased (CT min ) (Fig. 2B ). Despite being eurythermal, the warm-adapted species are again seen to be in greatest jeopardy from further increases in maximal habitat temperature; their abilities to extend their tolerance of lower temperatures is not matched by a similar capacity for increasing tolerance of heat. The differences in thermal tolerance and capacities for warm acclimation noted with porcelain crabs were mirrored in studies done with another set of congeneric marine invertebrates found at different vertical positions along the subtidal to intertidal gradient, turban snails of the genus Tegula [ 14 ]. Intertidal and subtidal Tegula congeners exhibit significant differences in thermal limits of protein synthesis and onset temperatures for production of heat-shock proteins that reflect their vertical distributions [ 15 , 16 ]. In agreement with these data, field-acclimatized populations of three Tegula congeners had significantly different thermal limits of heart function: the low- to mid-intertidal congener, T. funebralis , had a higher ABT of heart function (31.0°C) than two subtidal congeners, T. brunnea and T. montereyi (ABTs of 25.0°C and 24.2°C, respectively) [ 17 ]. The ABT of T. funebralis approximated the highest body temperature recorded for this species in the field, 32°C [ 15 ]. Body temperatures for the subtidal species only rarely reach 20°C [ 15 ], so these species would seldom, if ever, experience heart failure in their habitats. Furthermore, in agreement with the studies of porcelain crabs [ 8 ], T. funebralis had a lower ability to increase ABT during warm acclimation than the subtidal species [ 17 ]. These interspecific differences in thermal tolerance limits for whole organism survival and maintenance of heart function provide important lessons concerning biogeographic and vertical patterning and the potential effects of climate change. First, the differences in thermal tolerance between subtidal and intertidal species indicate that the former species would be unable to persist at the upper habitat temperatures commonly found in the intertidal zone during daytime low tides in warm seasons. For example, the temperate subtidal species Petrolisthes eriomerus has an upper lethal temperature approximately 6°C below the upper limit of body temperature recorded for the intertidal species P. cinctipes . Interspecific variations in LT 50 and ABT values also mirror the biogeographic patterning found across latitude. For instance, north temperate congeners of Petrolisthes would be unable to survive under the habitat conditions found in the northern Gulf of California or Panama. The close agreement among LT 50 , ABT, and maximal habitat temperatures for intertidal species suggests that further increases in habitat temperature could have strong impacts on the persistence of these species in their habitats, especially when the limits of acclimatory ability are reached. Because the greatest thermal stress occurs during emersion periods during the warmest time of the day, and because solar heating is more critical than ambient water temperature, the precise consequences of global warming are difficult to predict. Furthermore, because of latitudinal variation in the timing of tidal cycles, these and other intertidal species may face greater threats from climate change at mid-latitudes than at lower latitudes [ 18 ]. Despite these uncertainties, however, the seemingly paradoxical conclusion reached above, to the effect that many warm-adapted species are more threatened by increases in habitat temperature than cold-adapted congeners, appears valid and should be taken as a caveat that predicting the consequences of climate change is a complex challenge. Acclimatory changes in gene expression: eurytherms versus stenotherms The abilities of ectotherms to cope satisfactorily with increases in habitat temperature are based on several factors, including the proximity of habitat temperatures to the edges of the thermal tolerance range and the abilities to shift the tolerance range through acclimatization. In the case of porcelain crabs and turban snails, acclimatory ability was shown to vary among species. This finding that even closely related congeneric species differ in acclimatory ability raises questions about the mechanistic basis of phenotypic plasticity and the roles of evolutionary thermal history in establishing this plasticity. The ability to modify the phenotype in response to a change in body temperature is certain to depend strongly on a capacity for modulating gene expression in an adaptive manner. Although comprehensive surveys of temperature-induced shifts in gene expression have not been done for any marine organisms, studies of model species, defined here as species for which the genome has been sequenced and relatively well annotated, are beginning to reveal the pervasiveness of stress-induced shifts in gene expression. The studies of Gasch and colleagues [ 19 ] on yeast, for example, have shown that a variety of physical and chemical stressors, including temperature, hypoxia, reactive oxygen species and osmotic shock, triggered relatively similar changes in expression of approximately 900 genes, which represents approximately 14% of the yeast genome. They termed this gene regulatory response the "environmental stress response (ESR)". Because the proteins, lipids and nucleic acids that form the structural foundation of all physiological processes of cells will be affected in a qualitatively similar manner by changes of temperature in all organisms [ 20 ], it is reasonable to conjecture that an ESR generally similar to that found in yeast may characterize all species. For instance, induction of stress-induced chaperones (heat-shock proteins) for repair of damaged proteins, and shifts in synthesis of the various enzymes required for modifying the properties of lipid-based systems are likely to be ubiquitous events in the thermally induced ESRs of different species [ 20 ]. Despite the likely occurrence among species of generally similar requirements for temperature-adaptive shifts in gene expression, species that fall into different regions of the stenotherm to eurytherm spectrum may have distinctly different capacities for acclimatory regulation of transcription. In the context of effects of climate change, one of the most important interspecific differences relates to the ability of extreme stenotherms, which may have evolved for millions of years under stable thermal conditions, to alter gene expression in the face of temperature change. The Antarctic notothenioid fishes are a primary case in point. These species have evolved for 10–14 million years in a thermally stable "ice bath," in which annual temperature variation is usually less than 1–2°C [ 21 ]. McMurdo Sound populations of notothenioids are unable to acclimate to temperatures above approximately 4°C [ 22 , 23 ], making these species among the most stenothermal of organisms. Antarctic marine invertebrates, too, are extreme stenotherms, with heat death occurring at temperatures only a few degrees above 0°C [ 24 ]. In the case of the notothenioid fish Trematomus bernacchii , the lack of capacity to acclimate to elevated temperatures may stem in part from deficiencies in its ability to alter gene expression as its body temperature changes. Unlike virtually all other species, T. bernacchii is unable to increase the synthesis of any class of heat-shock protein following thermal stress [ 23 ]. Lacking this ability, T. bernacchii may be unable to effect adequate chaperone-mediated restoration of the native structures of heat-denatured proteins. A build-up of aggregates of denatured proteins is likely to be cytotoxic and lead to the eventual death of cells [ 25 ]. The absence of a heat-shock response in Antarctic notothenioids, but its presence in temperate New Zealand notothenioids [ 26 ], suggests that evolution under cold, stable thermal conditions has led to a depletion of the genetic resources of Antarctic fish. The inability of some notothenioid species to synthesize hemoglobins [ 27 ] or myoglobins [ 28 ] is another illustration of the depauperate genome of these cold-adapted stenotherms. Loss of oxygen transport pigments is viewed as a reflection of the lack of need for these pigments because of the combination of high oxygen solubility at low temperatures and the generally sluggish locomotory habits of the notothenioids [ 27 , 28 ]. In general, disappearance of these diverse genetic capacities in notothenioids is consistent with relaxed selection against the loss of genes that are no longer needed under conditions of stable low temperatures. Although future work on temperature-induced changes in the transcriptomes of Antarctic stenotherms will be needed before any broad generalizations can be made about the contents of their "genetic tool kits," it seems likely that the stenothermy of these species could be due in large measure to a loss of the type of acclimatory plasticity that is found in eurytherms. Lacking this phenotypic plasticity, Antarctic stenotherms seem uniquely vulnerable to the effects of global warming. Even among eurythermal ectotherms, there may be a variety of capacities for modulating gene expression to compensate for changes in body temperature. Although few data are currently available to test this point, acclimatory responses to rapid, diurnal variations in body temperature may require different shifts in gene expression from those that characterize slower, e.g., seasonal time-scale, responses to temperature change. A recent study [ 29 ] of temperature-induced changes in the transcriptome (the population of messenger RNA (mRNA) molecules in the cell) of liver tissue of a eurythermal teleost fish, Austrofundulus limnaeus , revealed wide-scale shifts in gene expression, as noted earlier for yeast [ 19 ]. A total of 540 mRNAs out of 4,992 examined changed by two-fold or more during acclimation. Many of the mRNAs changing in response to thermal acclimation were for proteins that are well known to be key elements of acclimatory response. For example, transcripts for heat-shock proteins and for enzymes involved in temperature-compensatory shifts in membrane lipid composition showed the predicted changes. However, for both classes of proteins, distinctly different expression profiles were found under the two acclimation regimes employed in this study: (i) steady-state acclimation at 20°C or 37°C for up to two weeks, and (ii) diurnally cycling temperatures that varied from a high of 37°C near mid-day to 20°C at night, a cycle that simulates the environmental conditions the species encounters in its shallow pond habitats in South America [ 29 ]. Under steady-state acclimation, transcripts for heat-shock proteins Hsp70 and Hsp90 showed the largest amount of change among molecular chaperones. Under cycling conditions, low-molecular-mass chaperones showed the greatest change. This finding suggests that protein damage and the repair costs it entails may differ under constant versus intermittent heat stress. Although both acclimation regimes led to changes in transcript abundance for proteins associated with alterations in lipid composition, differences between cycling and steady-state acclimation were noted in the types of lipids that appear to be produced. For steady-state acclimation, the expected changes in transcripts for enzymes involved in acyl chain double bond content (saturation) were observed, consistent with many previous studies of homeoviscous acclimation in membranes [ 20 ]. During cycling conditions, changes in transcripts for proteins involved in cholesterol biosynthesis and transport were more pronounced than shifts in mRNAs for enzymes of acyl chain biosynthesis. Although the significance of this difference in adaptive response is not known, it seems possible that insertion into or removal from the membrane of cholesterol could be achieved more rapidly than changes in the composition of phospholipids. In any event, differences in the time-frame of thermal stress may lead to differences in the types of temperature-compensatory strategies used to modulate the fluidity and phase of membranes. One of the most striking changes in transcript abundance observed in the study of Austrofundulus limnaeus was for the message encoding high mobility group b1 protein (HMGB1)(Fig. 3 ). High mobility group proteins are DNA-binding proteins that exert wide-ranging effects on transcriptional activity [ 30 ]. Rather than serving as transcriptional regulators for specific genes, HMGB1 proteins are general activators of transcription that exert their effects by influencing the transcriptional competence or "openness" of DNA structure. Increases in concentrations of HMGB1 favor open DNA structures and increased transcriptional activity of many genes. The changes in mRNA for HMGB1 found during thermal cycling (Fig. 3 ) and during steady-state acclimation [ 29 ] indicate an extremely tight control of this message. The increase in content of HMGB1 message with falling temperature (steady-state or cycling) is consistent with the role of this protein in maintaining DNA structure in an open, transcriptionally competent state. Thus, decreases in temperature will enhance the stability of the non-covalent bonds that stabilize higher order of DNA structure. To offset this increased stability of DNA structure, the cell may increase the concentrations of HMGB1 proteins, thereby allowing transcriptional activity to be temperature-independent. This is not to suggest that changes in HMGB1 protein will lead to genome-wide changes in transcriptional activity, of course. Rather, by increasing the openness of DNA structure, the promoter regions of genes will be susceptible to the effects of gene-specific transcription factors whose influences will lead to temperature-specific alterations in the transcriptome. Figure 3 Temperature effects on the transcriptome of the eurythermal fish Austrofundulus limnaeus . Changes in liver tissue in the level of the mRNA encoding high mobility group B1 protein (HMGB1), during thermal cycling between 37°C and 20°C. The Y-axis plots the logarithm of the ratio of expression in experimental (thermally cycled) versus control animals. Expression of the hmgb1 gene showed no circadian rhythm in fish held at constant temperature. Data from [29]. The changes in gene expression noted in Austrofundulus limnaeus and in a recent study of carp [ 31 ] represent an initial view of the complexity of transcriptional changes that ectothermic animals may experience during thermal acclimation, either to short-term or long-term changes in body temperature. Different changes to the transcriptome may be involved in different time courses of acclimation. It will be important to establish whether extreme stenotherms like Antarctic fish are capable of adjusting their transcriptomes in temperature-adaptive manners, when body temperature is increased. The importance of maintaining DNA structure in an open configuration that allows transcription factors to effectively modulate gene expression is a phenomenon that merits additional study. This capacity may be of pivotal significance in determining the effectiveness with which ectotherms can respond to changes in temperature; it may be foundational for most, if not all, of the temperature acclimation response. How this capacity differs between eurytherms and stenotherms may determine how these two groups cope with thermal fluctuations in their present habitats and what their potentials for coping with climate change are likely to be. Conclusions Ectothermic species differ widely in thermal tolerance limits and in their abilities to adjust these limits in temperature-adaptive manners. Comparisons of congeneric species from different latitudes and different positions along the subtidal to intertidal gradient have provided important insights into adaptive variation and the threats posed by increased habitat temperature. In what at first view may seem paradoxical, warm-adapted congeners may be more threatened by increased temperatures than their cold-adapted subtidal relatives. This difference stems from two factors: the proximity of current habitat temperatures to upper lethal temperatures (LT 50 s) and the more limited abilities to acclimate to increases in temperature noted for intertidal species. The physiological determinants of upper thermal limits are certain to be multifarious, but heart function stands out as a key contributor to these limits. In porcelain crabs and turban snails, upper lethal temperatures coincide closely with temperatures of heart failure. Such cardiac effects also could contribute to a shift from aerobic to anaerobic processes of ATP generation, which has been proposed as one important mechanism of death from high and low temperatures [ 32 , 33 ]. Acclimatory response to change in body temperature is likely to be an important determinant of the effects of global warming on species. Temperature-adaptive shifts in gene expression serve as a foundation for physiological acclimation. Changes in the transcriptome differ between long-term, steady-state acclimation and responses to diurnally cycling temperatures. The ability to modulate gene expression requires an openness of DNA structure, which appears to be closely regulated during steady-state and cycling thermal regimes. Being able to modulate the transcriptional competence of DNA in the face of changing temperature, to ensure that necessary shifts in gene expression can occur, may be a fundamental requirement for thermal acclimation. Species such as Antarctic notothenioid fishes that have had a long evolutionary history at constant temperatures may have lost many of the critical gene regulatory responses needed for thermal acclimation and, as a consequence, they may be uniquely vulnerable to global warming.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546413.xml
549046
Interactions between xenoestrogens and ketoconazole on hepatic CYP1A and CYP3A, in juvenile Atlantic cod (Gadus morhua)
Background Xenoestrogens and antifungal azoles probably share a common route of metabolism, through hepatic cytochrome P450 (CYP) enzymes. Chemical interactions with metabolic pathways may affect clearance of both xenobiotics and endobiotics. This study was carried out to identify possible chemical interactions by those substances on CYP1A and CYP3A, in Atlantic cod liver. We investigated effects of two xenoestrogens (nonylphenol and ethynylestradiol) and of the model imidazole ketoconazole, alone and in combination. Results Treatment with ketoconazole resulted in 60% increase in CYP1A-mediated ethoxyresorufin- O -deethylase (EROD) activity. Treatment with nonylphenol resulted in 40% reduction of CYP1A activity. Combined exposure to ketoconazole and nonylphenol resulted in 70% induction of CYP1A activities and 93% increase in CYP1A protein levels. Ketoconazole and nonylphenol alone or in combination had no effect on CYP3A expression, as analyzed by western blots. However, 2-dimensional (2D) gel electrophoresis revealed the presence of two CYP3A-immunoreactive proteins, with a more basic isoform induced by ketoconazole. Treatment with ketoconazole and nonylphenol alone resulted in 54% and 35% reduction of the CYP3A-mediated benzyloxy-4-[trifluoromethyl]-coumarin- O -debenzyloxylase (BFCOD) activity. Combined exposure of ketoconazole and nonylphenol resulted in 98% decrease in CYP3A activity. This decrease was greater than the additive effect of each compound alone. In vitro studies revealed that ketoconazole was a potent non-competitive inhibitor of both CYP1A and CYP3A activities and that nonylphenol selectively non-competitively inhibited CYP1A activity. Treatment with ethynylestradiol resulted in 46% decrease in CYP3A activity and 22% decrease in protein expression in vivo . In vitro inhibition studies in liver microsomes showed that ethynylestradiol acted as a non-competitive inhibitor of CYP1A activity and as an uncompetitive inhibitor of CYP3A activity. Conclusions Ketoconazole, nonylphenol and ethynylestradiol all interacted with CYP1A and CYP3A activities and protein expression in Atlantic cod. However, mechanisms of interactions on CYP1A and CYP3A differ between theses substances and combined exposure had different effects than exposure to single compounds. Thus, CYP1A and CYP3A mediated clearance may be impaired in situations of mixed exposure to those types of compounds.
Background A great challenge in pharmacology and toxicology is to understand the molecular mechanisms behind how mixtures of compounds affect living organisms. This study focuses on two classes of substances, imidazoles and xenoestrogens, and how these chemicals alone and in combination affect hepatic drug-metabolizing hepatic cytochrome P450 (CYP) enzymes – specifically, CYP1A and CYP3A enzymes, in juvenile Atlantic cod ( Gadus morhua ). Imidazoles and triazoles are used as fungicides both clinically as well as in horticulture and agriculture, posing a potential threat to wildlife. The triazole propiconazole has been detected in the aquatic environment [ 1 ]. The azole antifungal effect resides in inhibition of CYP51 mediated ergosterol biosynthesis [ 2 ]. In addition to disrupting key enzymes in fungus, azoles such as the imidazoles clotrimazole, ketoconazole, miconazole and prochloraz also cause endocrine disruption in vertebrates by inhibition of key enzymes in steroid homeostasis [ 3 - 7 ]. Moreover, these fungicides inhibit drug-metabolizing CYP forms, including members of the CYP1, CYP2 and CYP3 gene families in vertebrates [ 5 , 8 - 13 ]. Effects on CYP forms may have adverse effects on metabolic clearance of endobiotics and xenobiotics. For example, in a study in fish, pre-exposure to clotrimazole resulted in increased bioaccumulation of the pro-carcinogen benzo [ a ]pyrene in gizzard shad ( Dorosoma cepedianum ) [ 14 ]. Xenoestrogens comprise a wide variety of structurally diverse chemicals such as o,p -DDT, ethynylestradiol, alkylphenols and bisphenol A. These substances are well-known or supposed to be endocrine disrupting substances in vertebrates and share in common that they activate the estrogen receptor (ER) and thereby elicit estrogenic responses [ 15 - 17 ]. In addition to being estrogenic, these xenoestrogens interact with drug-metabolizing CYP forms, including members of the CYP1A and CYP3A subfamilies in vertebrates [ 18 - 22 ]. Xenoestrogens are continuously released into the environment as a result of various anthropogenic activities. Induction of vitellogenesis in fish is a biomarker routinely used to assess the presence of estrogenic substances in the aquatic environment [ 23 , 24 ]. Induction of CYP1A-mediated ethoxyresorufin- O -deethylase (EROD) activity is another established biomarker used to assess exposure to aromatic hydrocarbons. This response proceeds through activation of the aryl hydrocarbon receptor (AHR) by aromatic hydrocarbons including polyaromatic hydrocarbons, and planar polychlorinated biphenyls and dioxins [ 25 ]. Some AHR agonists have been shown to be anti-estrogenic and cross-talk between AHR and ER has been suggested in vertebrates [ 26 - 33 ]. In addition to activation of the ER, xenoestrogens also affect other steroid receptors. Nonylphenol up-regulated CYP3A1 gene expression in rat, through activation of the pregnane X receptor (PXR) [ 34 , 35 ]. We previously reported induction of CYP3A and CYP1A protein levels in Atlantic cod exposed to alkylphenols [ 22 ]. Azole fungicides induce expression of multiple vertebrate CYP genes including members of the CYP1A, CYP2B and CYP3A subfamilies [ 8 , 9 , 13 , 36 - 38 ]. Clotrimazole activates the ligand-binding domain of the PXR, involved in CYP3A signalling, in vitro from several mammalian species and zebra fish ( Danio rerio ) [ 39 ]. Both imidazoles and xenoestrogens inhibit drug-metabolizing enzymes, including members of the CYP1A and CYP3A subfamilies in vertebrates [ 8 - 13 , 18 , 20 , 22 ]. Thus, xenoestrogens and imidazoles conceivably share common routes for biotransformation. However, there is a lack of data regarding effects of combined exposure of imidazoles and xenoestrogens on these CYP forms in wildlife. Living organisms usually are exposed to mixtures of different classes of xenobiotics. Conceivably, exposure to mixtures may be more of a health threat than exposure to single compounds, as a result of interactions. Anthropogenic compounds may enter the environment through industrial activities and through the use of pharmaceuticals [ 40 ]. Atlantic cod is an economically important species for fishery and a growing aquaculture industry, in addition to its ecological relevancy. Its distribution in the Northern Atlantic and the North Sea makes it vulnerable to effluents from on-shore and off-shore industries and from run-off entering the waters near highly industrialized and urbanized areas. The rationale of the present study was to identify possible sites of interactions between imidazoles and xenoestrogens. We hypothesise that combined exposure to these compounds may compromise the metabolic clearance not only of these xenobiotics themselves, but also of endobiotics such as circulating steroid hormones that share common routes of metabolism through hepatic CYP1A and CYP3A. Such endocrine disrupting effects may adversely affect the stability of wildlife populations. The specific aim of our study was to examine interactions between two classes of compounds in livers of Atlantic cod. Thus, we investigated the effects of the model imidazole ketoconazole and of two types of xenoestrogens (nonylphenol and ethynylestradiol), as well as of a mixed exposure to ketoconazole and nonylphenol, on hepatic CYP1A and CYP3A protein expression and catalytic activities, and also on vitellogenesis and plasma levels of sex steroid hormones. Results In vivo effects on CYP1A Exposure to ketoconazole (12 mg/kg b.w.) and/or a combination of ketoconazole and nonylphenol (12 mg/kg b.w. + 25 mg/kg b.w.) resulted, respectively, in 159 and 172% average induced increases in CYP1A-mediated EROD activities (Fig. 1A ), and in 133 and 193% increases in CYP1A protein levels in Atlantic cod (Fig. 1B ). Treatment with nonylphenol (25 mg/kg b.w.) resulted in 41% reduction and ethynylestradiol (5 mg/kg b.w.) resulted in 72% reduction, respectively, of CYP1A activities compared to vehicle treated fish (Fig. 1A ). However, when compared to fish exposed to the combination of ketoconazole and nonylphenol, exposure to nonylphenol alone and ethynylestradiol resulted in 65% and 84% decrease in CYP1A activity (Fig. 1A ). Exposure to nonylphenol and ethynylestradiol had no effect on CYP1A protein expression (Fig. 1B ). The CYP1A protein levels were elevated by 93% in fish exposed to a mixture of ketoconazole and nonylphenol (Fig. 1B ). Figure 1 A) In vivo CYP1A enzyme activities (A) and in vivo CYP1A protein expression (B) . CYP1A enzyme activities and protein expression in juvenile Atlantic cod exposed in vivo to vehicle (5 ml peanut oil/kg fish), ketoconazole (12 mg/kg fish), nonylphenol (25 mg/kg fish), ethynylestradiol (5 mg/kg fish) and ketoconazole + nonylphenol (12 + 25 mg/kg fish). A) EROD activities. B) CYP1A protein levels analyzed using PAb against rainbow trout CYP1A. Each bar represents mean values of eight to nine fish ± SD; a Significantly different from vehicle treated fish; b Significantly different from ketoconazole+nonylphenol treated fish; P < 0.05. In vivo effects on CYP3A Fish exposure to ketoconazole, ethynylestradiol and nonylphenol resulted in decreased CYP3A-mediated benzyloxy-4-[trifluoromethyl]-coumarin- O -debenzyloxylase (BFCOD) activities, when compared to vehicle treated fish (Fig. 2A ). Furthermore, mixed exposure to ketoconazole and nonylphenol resulted in a 98% decrease in CYP3A activity, which was greater than the additive effects of these two compounds administrated alone (Fig. 2A ). Fish exposed to the ketoconazole and nonylphenol mixture displayed significantly reduced CYP3A activities when compared all other treatment groups (Fig. 2A ). No effect on CYP3A protein expression was observed in fish treated with ketoconazole and nonylphenol, either alone or in combination (Fig. 2B ). However, ethynylestradiol treatment resulted in 22% decrease in CYP3A protein levels (Fig. 2B ). Figure 2 In vivo CYP3A enzyme activities (A) and in vivo CYP3A protein expression (B) . CYP3A enzyme activities and protein expression in juvenile Atlantic cod exposed in vivo to vehicle (5 ml peanut oil/kg fish), ketoconazole (12 mg/kg fish), nonylphenol (25 mg/kg fish), ethynylestradiol (5 mg/kg fish) and ketoconazole + nonylphenol (12 + 25 mg/kg fish). A) BFCOD activities. B) CYP3A protein levels analyzed using PAb against rainbow trout CYP3A. Each bar represents mean values of eight to nine fish ± SD; a Significantly different from vehicle treated fish; b Significantly different from ketoconazole+nonylphenol treated fish; P < 0.05. Western blot analyses of CYP3A proteins using PAb against rainbow trout CYP3A revealed the presence of one CYP3A immunoreactive protein band in liver microsomes, with an apparent molecular size above 50 kD, in Atlantic cod (Fig. 3A ). By using 2D gel electrophoresis followed by immunoblotting, two immunoreactive CYP3A protein spots were detected above 50 kD, with pI values around 4.8 and 5.1, respectively (Fig. 3B ). The most basic isoprotein appears to be inducible by treatment with ketoconazole (Fig. 3B ). Ethynylestradiol and nonylphenol treatment did not induce expression of the more basic isoform. Present data does not elucidate whether those two protein spots are different gene products, or if they result from post-translational modifications such as phosphorylation. Figure 3 CYP3A Western blot (A) and CYP3A 2D-immunoblots (B) . A) Western blot of hepatic microsomal CYP3A proteins in juvenile Atlantic cod treated with vehicle (5 ml peanut oil/kg fish) and ketoconazole (12 mg/kg fish) detected using PAb against rainbow trout CYP3A. B) 2D-gel electrophoresis followed by immunoblotting using PAb against rainbow trout CYP3A. Each blot represent pooled liver microsomes of eight to nine fish for each treatment; vehicle (5 ml peanut oil/kg fish), ketoconazole (12 mg/kg fish), nonylphenol (25 mg/kg fish), ethynylestradiol (5 mg/kg fish), ketoconazole + nonylphenol (12 + 25 mg/kg fish). In vitro inhibition studies In vitro inhibition studies using pooled Atlantic cod liver microsomes showed that ketoconazole, nonylphenol, ethynylestradiol and the ketoconazole:nonylphenol (1:5) mixture inhibited CYP1A (EROD) activity, with IC 50 values (inhibitor concentration required to achieve a 50% inhibition) ranging from 0.6 to 20 μM. The CYP3A-mediated BFCOD activity also was inhibited by ketoconazole (IC 50 = 0.3 μM), ethynylestradiol (IC 50 = 40 μM) and the ketoconazole:nonylphenol (1:5) mixture (IC 50 = 5:25 μM). Nonylphenol alone was an insignificant inhibitor of microsomal CYP3A activities in Atlantic cod (IC 50 = 160 μM). For comparison, IC 50 values for nonylphenol and ethynylestradiol also were determined in cDNA expressed human CYP3A4 baculovirus supersomes, compared to the prototypical CYP3A4 inhibitor ketoconazole (IC 50 = 0.4 μM). In contrast to Atlantic cod liver microsomes, nonylphenol inhibited the human CYP3A4 mediated BFCOD activity (IC 50 = 35 μM) and ethynylestradiol was a weak inhibitor (IC 50 = 50 μM) of this activity. The IC 50 values are summarized in Table 1 . Table 1 IC 50 values and inhibition constants (K i ) for ketoconazole and xenoestrogens on CYP1A- and CYP3A activities assayed in vitro . Compound(s) IC 50 (μM) 1,a K i (μM) 1,a IC 50 (μM) 2,b K i (μM) 2,b IC 50 (μM) 3,b Ketoconazole (KC) 0.6 (0.0) c 0.04 – low [S] 0.3 (0.1) c 0.2 0.4 c 0.2 – high [S] Nonylphenol (NP) 5.2 (1.1) 3.5 160 (40) Not analysed 35 Ethynylestradiol 20 (1.2) 5.4 – low [S] 40 (7.1) 54 – low [S] 50 10.3 – high [S] 95 – high [S] KC:NP (1:5) 1.3 (0.2):6.2 (1.0) Not analysed 5.3 (1.1):25.0 (5.3) Not analysed Not analysed 1 Hepatic Microsomal CYP1A activity; 2 Hepatic Microsomal CYP3A activity; 3 cDNA Expressed Human CYP3A4; a Substrate [S] = 7-Ethoxyresorufin; b [S] = 7-Benzyloxy-4-[trifluoromethyl]-coumarin; c Published in [22]. Each IC 50 value represents the mean from 2–4 separate assays, followed by the SD, in brackets. The K i values are estimated from one representative Dixon plot. The inhibitory effects of these compounds were further investigated on hepatic microsomal CYP1A and CYP3A enzyme kinetics. The K i values were determined in Dixon plots (Figs. 4 and 5 ) and summarized in Table 1 . Ketoconazole was a potent non-competitive inhibitor of both CYP1A and CYP3A activities with K i values in the sub-μM range (Fig. 4 ; Table 1 ). Ethynylestradiol was a non-competitive inhibitor of CYP1A with K i from 5.4 to 10.3 μM and an uncompetitive inhibitor of CYP3A with K i from 54 to 95 μM (Fig. 5 ; Table 1 ). Nonylphenol was a non-competitive inhibitor of CYP1A activity with K i around 3.5 μM (Table 1 ). There were no effects of pre-incubation either with ketoconazole or ethynylestradiol on hepatic microsomal CYP3A protein levels in this study (Fig. 6 ). Figure 4 Non-competitive inhibition of CYP1A by ketoconazole (A) and non-competitive inhibition of CYP3A by ketoconazole (B) . Dixon plots for ketoconazole on A) EROD activity (diamonds represent 8.2; squares represent 25 and triangles represent 677 pM ethoxyresorufin). B) BFCOD activity (diamonds represent 48; squares represent 84 and triangles represent 200 μM BFC). Figure 5 Non-competitive inhibition of CYP1A by ethynylestradiol (A) and uncompetitive inhibition of CYP3A by ethynylestradiol (B) . Dixon plots for ethynylestradiol on A) EROD activity (diamonds represent 8.2; squares represent 25 and triangles represents 677 pM ethoxyresorufin). B) BFCOD activity (diamonds represent 200; squares represent 267 and triangles represents 356 μM BFC). Figure 6 CYP3A Western blot after in vivo incubation . Western blot of CYP3A proteins in pooled liver microsomes from Atlantic cod detected using PAb against rainbow trout CYP3A. The blot illustrates representative samples after in vitro incubation with 1.0 μM ketoconazole and 50 μM ethynylestradiol for 30 or 60 min. Plasma vitellogenin- and sex steroid hormone levels Treatment with nonylphenol, ethynylestradiol and the combination of ketoconazole and nonylphenol resulted in induction of vitellogenin, whereas these treatments had no statistically significant effect on 17β-estradiol, testosterone and 11-keto-testosterone plasma levels compared to either vehicle treated fish or fish treated with each test compound alone. The results are summarized in Table 2 . Table 2 Plasma levels of vitellogenin and sex steroid hormones in juvenile Atlantic cod exposed in vivo to ketoconazole and xenoestrogens. Treatment Vitellogenin 17β-Estradiol Testosterone 11-Keto-Testosterone (μg/ml plasma) (pg/ml plasma) (pg/ml plasma) (pg/ml plasma) n = 7–8 n = 8 n = 7–8 n = 6–8 Vehicle (5 ml peanut oil/kg b.w.) 0.6 (1.0) 62 (56) 86 (68) 33 (38) Ketoconazole (KC) (12 mg/kg b.w.) 0.6 (0.9) 60 (26) 120 (78) 71 (46) Nonylphenol (NP) (25 mg/kg b.w.) 266 (199) a 76 (52) 75 (33) 42 (67) Ethynylestradiol (5 mg/kg b.w.) 4,350 (1,463) a 100 (69) 90 (40) 35 (23) KC + NP (12 + 25 mg/kg b.w.) 268 (205) a 36 (28) 61 (51) 36 (38) a Significantly different from vehicle treated fish ( P < 0.001; Kruskal-Wallis ANOVA, followed by Mann-Whitney U-test). Each value represents the mean from 6–8 fish, followed by the SD, in brackets. Discussion Effects on CYP1A For data evaluation we must bear in mind that western blot analysis of CYP1A protein levels is less sensitive than the EROD assay [ 41 ] and so densitometry analysis of western blot data fails to detect minor changes. Treatment of juvenile Atlantic cod with ketoconazole resulted in elevated EROD activities. Mixed exposure to ketoconazole and nonylphenol resulted in induced EROD activities and CYP1A protein levels. Induction of hepatic CYP1A gene expression by exposure to imidazoles and/or triazoles also has been reported in rat, bobwhite quail ( Colinus virginianus ) and rainbow trout ( Oncorhynchus mykiss ) [ 8 , 13 , 37 , 38 ]. However, it is possible that induction of EROD activity, partly or completely, is masked by CYP1A inhibition caused by ketoconazole present in the tissue. Inhibition of CYP1A is supported in the present study, showing that ketoconazole was a potent non-competitive inhibitor of EROD activity in vitro . Ketoconazole and other imidazoles also have been shown to be potent inhibitors of EROD activities in other vertebrates [ 9 , 13 , 14 , 42 ]. Treatment of Atlantic cod with nonylphenol and ethynylestradiol resulted in decreased EROD activities, whereas no effects of these substances were observed on CYP1A protein levels. This decrease in EROD activity is probably caused by nonylphenol or ethynylestradiol present in the liver microsome fraction. Nonetheless, chemical data are required, in the future, to confirm this. In vitro inhibition studies in liver microsomes confirmed that nonylphenol and ethynylestradiol acted as non-competitive inhibitors of the EROD activity. Hence, ketoconazole, nonylphenol, and ethynylestradiol interact with CYP1A enzymes, indicating a possible site for interaction of these different classes of xenobiotics. In addition, ketoconazole treatment induces CYP1A expression, which further may affect this interaction. Effects on CYP3A Atlantic cod exposed to nonylphenol, ethynylestradiol and ketoconazole displayed reduced hepatic CYP3A (BFCOD) activities. The CYP3A inhibitory effect by ketoconazole is well known and ketoconazole is the most established diagnostic inhibitor, used to assess human in vitro CYP3A4 activity [ 12 , 43 ]. Studies in fish demonstrate that ketoconazole is a potent inhibitor of hepatic BFCOD activities in killifish ( Fundulus heteroclitus ), rainbow trout and Atlantic cod with IC 50 values at 0.01, 0.1 and 0.3 μM, respectively [ 13 , 22 ]. In rainbow trout, exposure to ketoconazole resulted in elevated hepatic and intestinal CYP3A protein levels [ 13 ]. In the present study, 2D gel electrophoresis revealed the presence of two CYP3A immunoreactive spots in Atlantic cod liver microsomes with pI values around 4.8 and 5.1, respectively. The more basic isoform (pI 5.1) appeared to be responsive to ketoconazole treatment. The existence of multiple CYP3A genes has been shown in several vertebrate species, including teleosts [ 44 ]. It is conceivable that there are two different CYP3A genes in Atlantic cod and that these genes respond differently to ketoconazole treatment. Protein isoforms revealed on 2D gel electrophoresis may also be due to post-translational modifications such as phosphorylation [ 45 ]. Phosphorylation of several members of the CYP2 gene family, through phosphokinase A, resulted in immediate loss in catalytic activity [ 46 ]. The shift to a more basic form in this report could imply a dephosphorylation of CYP3A upon ketoconazole treatment. However, as these spots were not detected directly on the 2D gels by using either Coomassie blue or silver staining, no spots could be selected for sequencing to investigate whether these two immunoreactive spots represent different gene products. In juvenile Atlantic salmon ( Salmo salar ), multiple hepatic CYP3A proteins also were seen [ 19 ]. The two proteins responded differently to nonylphenol treatment. High doses of nonylphenol (125 mg/kg b.w.) suppressed the high-molecular weight CYP3A protein band, whereas lower doses of nonylphenol (25 mg/kg b.w.) resulted in induction of this isoform [ 19 ]. In the present study, exposure to nonylphenol resulted in reduced CYP3A activities in juvenile Atlantic cod liver. Nevertheless, nonylphenol did not inhibit microsomal BFCOD activities in vitro , whereas nonylphenol was a weak inhibitor of that activity using recombinant human CYP3A4. The Atlantic cod we exposed to a mixture of ketoconazole and nonylphenol displayed in vivo CYP3A activities that were lower than the additive effect of each compound administered alone. The mechanism for this possible interaction still is not known. In mammals, more than one substrate can simultaneously bind to the active site of CYP3A4 [ 11 ]. Thus, in Atlantic cod, conceivably both ketoconazole and nonylphenol might bind to CYP3A enzyme and prevent access of the diagnostic BFC substrate. The CYP3A protein levels remained unchanged in these fish suggesting that combined exposure of ketoconazole and nonylphenol selectively inhibits in vivo CYP3A activity. Ethynylestradiol has been shown to act as a mechanistic inactivator ( i.e. "suicide" substrate) of the CYP3A4 enzyme, resulting in loss of CYP3A4 protein levels [ 47 , 48 ]. In Atlantic cod, a possible mechanism-based inactivation of CYP3A by ethynylestradiol was suspected. Thus, exposure to ethynylestradiol resulted in significantly reduced CYP3A levels and ethynylestradiol acted as an uncompetitive inhibitor of microsomal CYP3A activities. However, pre-incubation of hepatic microsomes with ethynylestradiol for up to 60 min did not result in any significant loss of CYP3A protein content, which implies that, in Atlantic cod, ethynylestradiol is not acting as mechanism-based inhibitor of CYP3A. Nonetheless, further studies are needed, as for example 2D gel electrophoresis of pre-incubated liver microsomes followed by immunoblotting. Vitellogenesis and sex steroid hormones In humans, prolonged ketoconazole therapy results in decreased clearance of 17β-estradiol, which may cause gynecomastia, presumably through inhibition of hepatic CYP3A4 [ 5 ]. In the present study, nonylphenol dependent induction of vitellogenesis was not significantly affected by treatment with a single dose of ketoconazole. In the present study, exposure to xenoestrogens and ketoconazole alone had no statistically significant effect on sex steroid levels compared to control fish. In another study in first spawning Atlantic cod, exposure to alkylphenols resulted in decreased plasma levels of 17β-estradiol (in females) and testosterone and 11-keto-testosterone (in males) [ 49 ]. Additional data, including plasma levels of ethynylestradiol, and increasing the sample sizes are required to definitely elucidate whether, in Atlantic cod, exposure to xenoestrogen and ketoconazole alone or in combination may affect sex steroid homeostasis. Conclusions This study identifies, in Atlantic cod, interactions between ketoconazole and two different types of xenoestrogens on CYP1A and CYP3A. Ketoconazole acted as a non-competitive inhibitor of CYP1A and CYP3A activities and ketoconazole treatment also induced CYP1A protein expression. Ethynylestradiol acted as a non-competitive inhibitor of CYP1A and an uncompetitive inhibitor of CYP3A activities. In vitro studies revealed that nonylphenol was a non-competitive inhibitor of CYP1A; but it did not inhibit CYP3A. However, in vivo , nonylphenol synergistically impaired the ketoconazole-mediated inhibition of CYP3A activity, without affecting CYP3A protein levels. The study further illustrates that induction of CYP1A- and CYP3A gene expression can be partly or completely masked by inhibition of catalytic activities or vice versa. Taken together, the results indicate that CYP1A and CYP3A represent sites of interactions between those classes of xenobiotics. In future risk-assessment of, e.g. , municipal effluents or produced water from oil platforms, that have been shown to contain xenoestrogens, it should be considered to identify other classes of substances, for example azoles that also interact with CYP1A and CYP3A. Our data may warn for ecotoxicological implications, as induction of EROD activity as well as plasma vitellogenin routinely are used as biomarkers to assess exposure to AHR and ER agonists in various biomonitoring programs in the aquatic environment. Methods Chemicals The 4-nonylphenol and the 17α-ethynylestradiol, for the in vivo exposure experiment, were obtained from Fluka Chemie AG (Buchs, Switzerland). The 4-nonylphenol for the in vitro inhibition studies was from Berol Nobel (Stenungsund, Sweden). Dimethylsulphoxide (DMSO), 7-ethoxyresorufin, horseradish peroxidase- (HRP) conjugated goat-anti-mouse IgG, iodoacetamide, ketoconazole, ponceau- S , resorufin and tween-20 were obtained from Sigma Aldrich (Stockholm, Sweden). Reduced nicotinamide-adenine-dinucleotide-phosphate (NADPH) was from Roche Diagnostics (St Louis, MO, USA and Bromma, Sweden). Ready gels (12% continuous acrylamide in Tris:HCl), 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), precision protein standards (low range) and supported nitrocellulose membrane (0.45 μm) were purchased from BioRad (Sundbyberg, Sweden). The 17β-estradiol and testosterone enzyme immuno assay (EIA) kits were purchased from Cayman Chemical (Ann Arbor, MI, USA). The 11-keto-testosterone EIA kit and the Atlantic cod vitellogenin Enzyme Linked ImmunoSorbent Assay (ELISA) kit were obtained from Biosense Laboratories AS (Bergen, Norway). HRP-conjugated donkey-anti-rabbit IgG, the ECL™ Western blotting detection reagents and Immobiline™ DryStrip 7 cm ranging from pH 4 to 7 were from Amersham Biosciences (Uppsala, Sweden). Ampholytes for isoelectric focusing (Servalyt ® Carrier ampholyt 3–10) was purchased from Serva Feinbiochemica (Heidelberg, Germany). Dithiothreitol (DTT), Kodak X-Omat AR-ray film, X-ray developer and fix were from VWR International (Stockholm, Sweden). The 7-benzyloxy-4-[trifluoromethyl] coumarin (BFC), 7-hydroxy-4-[trifluoromethyl] coumarin (HFC) and the CYP3A4 inhibition kit were from BD Biosciences Company, Gentest™ (Woburn, MA, USA). All other chemicals used were of the purest grade available in Sweden or Norway, from Sigma-Aldrich, BioRad and VWR international. Animals and sampling Hatchery reared juvenile Atlantic cod of both sexes with an average body weight (b.w.) around 400 g were supplied by Sekkingstad, Preserving AS, Hordaland, Norway. The fish were kept in 0.5 m 3 indoor glass fibre tanks, at Industrial Laboratory (ILAB), Bergen High Technology Centre (Bergen, Norway), provided with continuously flowing seawater at a temperature of 8 ± 0.5°C and a salinity of 3.4%. Throughout the experimental period, the fish were subjected to continuous 24 h artificial light (the regime the farm used for optimal fish growth). The fish were acclimated to these conditions for five days prior to the experimental period. During the experimental period the fish were starved and i.p. injected with either 12 mg ketoconazole/kg b.w. resuspended in peanut oil (2.5 mg/ml); 25 mg nonylphenol/kg b.w. dissolved in peanut oil (5.0 mg/ml); 5 mg ethynylestradiol/kg b.w. dissolved in peanut oil (1.0 mg/ml) or a mixture of ketoconazole and nonylphenol (12 mg ketoconazole + 25 mg nonylphenol/kg b.w. in peanut oil). Control fish were injected with 5 ml peanut oil/kg b.w. (vehicle). There were eight to nine fish in each treatment group. When designing the experiment, we could only test one combination due to limited fish numbers. We selected nonylphenol to combine with ketoconazole because a previous study showed that, in Atlantic cod, alkylphenols affect CYP1A/3A more strongly than the natural estrogen 17β-estradiol [ 22 ]. The ketoconazole dose (12 mg/kg) was selected based on the results on CYP1A and CYP3A protein levels and enzyme activities from a previous dose-response study in rainbow trout [ 13 ]. The nonylphenol dose (25 mg/kg) was selected as this dose is known to induce vitellogenesis in a number of fish species. In addition, in a previous study on the Atlantic salmon, this dose of nonylphenol also had effects on CYP1A and CYP3A [ 19 ]. After five days exposure, the fish were sacrificed by a sharp blow to the head. Blood samples were collected from the dorsal vein using a heparinized syringe and the liver was quickly dissected out and placed in ice-cold homogenization buffer (0.1 M sodium phosphate buffer pH 7.4, containing 0.15 M KCl, 1 mM EDTA and 1 mM DTT). Liver microsomes were prepared according to the published protocol by Goksøyr [ 50 ], and stored at -80°C. Total microsomal protein content was measured according to a published method by Bradford [ 51 ], using bovine serum albumin as standard, and a SpectraFluor spectrophotometer from Tecan (Grödig/Salzburg, Austria). Blood plasma was isolated by centrifugation at 5,000 g for 10 min at room temperature and stored at -80°C. Ethical approval licence number of ILAB Bergen: 119. Experiment no. 0204. For in vitro inhibition studies, feral Atlantic cod of both sexes were caught off the West coast of Sweden and placed in concrete basins provided with recirculating aerated seawater at 10 ± 2°C and a salinity of 3.0% and alternative light/dark photoperiods of 12 hours. Prior to sampling, the animals were starved and acclimated to these conditions for three weeks. Eight fish were injected i.p. with β-naphthoflavone (BNF), 50 mg/kg b.w. dissolved in peanut oil (5.0 mg/ml). The fish were placed in a 100 l glass aquarium provided with aerated seawater (above) and 30% of the water volume was replaced each day. To eliminate visual stress, the sides of the aquaria were covered with black plastic sheets. After 3 days exposure, the fish were sacrificed. Livers were quickly dissected out and placed in ice-cold homogenization buffer. Livers were pooled from twenty untreated Atlantic cod and from eight BNF treated Atlantic cod, respectively. Microsomal fractions were isolated (above) and stored in aliquots at -80°C. Ethical approval from the Ethical committee of Göteborg license number (99–2003). The duration of exposure was decided according to results from previous time-course studies showing maximal CYP1A protein and EROD activities in rainbow trout and in the marine viviparous blenny ( Zoarces viviparous ), 3 days post-injection with either the prototypical CYP1A inducers BNF or 3-methylcholanthrene [ 52 - 54 ]. CYP1A- and CYP3A protein blot analyses Western blot analyses of 40 μg hepatic microsomal protein were carried out using enhanced chemoluminescence (ECL), based on the protocol previously described [ 55 ] and PAb raised against rainbow trout CYP1A and CYP3A [ 41 , 55 , 56 ]. The intensity of each protein band was determined by densitometry on scanned fluorograms using Labview 7.0 from National Instruments (Austin, TX, USA). The 2D gel electrophoresis was performed using immobilised pH gradient gels with linear gradient from pH 4 to 7. The samples were concentrated by acetone precipitation and pellets dissolved in rehydration buffer (8 M urea, 2 M thiourea, 20 mM DTT, 4% CHAPS, 0.5% Triton X-100, 0.5% ampholyte 3–10 and <0.02% bromophenolblue) to a final protein concentration of 20 μg/μl or 80 μg/μl. The samples were rehydrated overnight followed by isoelectric focusing for 2.5 h. The rehydration was passive and carried out overnight in an Immobiline Dry Strip reswelling tray (Amersham Biosciences). First-dimension isoelectric focusing (IEF) was performed on a Multiphor II unit (Amersham Biosciences) at 20°C using a MultiDrive XL power supply (Pharmacia LKB). Settings for IEF were 30 min at 100 V and 3 h at 3500 V for a total of 10,520 Vh. Amperage and wattage were set to 2 mA and 5 W. The proteins were resolved on 9% continuous acrylamide gel in Tris:HCl, including sodium dodecyl sulphate polyacrylamide using a mini-gel apparatus from BioRad at 200 V for 45 min. Each sample consisted of pooled liver microsomes from eight to nine fish from each treatment group. Gels loaded with 25 μg microsomal protein were initially stained with 0.1 % (w/v) Coomassie brilliant blue, and then destained, followed by silver staining. The latter was performed according to Heukeshoven and Dernick [ 57 ]. Stained gels were scanned and analyzed using the PDQUEST 7.1 software (BioRad). Gels loaded with 100 μg microsomal proteins were electrotransferred to nitrocellulose membrane and immunoblotted for CYP3A, as described above. Catalytic assays The CYP1A activity was determined as EROD activity, using resorufin as standard in a SpectraFluor plate reader according to the protocol provided by Nilsen et al. [ 58 ]. The CYP3A catalytic activity was measured as BFCOD activity, using HFC as standard. The BFC assay was performed based on a published protocol by Miller et al. [ 59 ] and optimized for rainbow trout liver microsomes (T. Hegelund and M. Celander, unpublished data). The reaction mixture consisted of 200 μM BFC, bovine serum albumin (1.6 mg/ml), 2 μM NADPH and 10 μl liver microsomes in a total volume of 200 μl in 0.2 M potassium phosphate buffer pH 7.4 in a 96-multiwell plate using a VICTOR™ 1420 Multilabel Counter from Wallac Sverige AB (Upplands Väsby, Sweden). In vitro inhibition of CYP1A and CYP3A In vitro inhibition studies were carried out in 96-multiwell plates using a VICTOR™ 1420 Multilabel Counter. The IC 50 values were determined for nonylphenol, ethynylestradiol, ketoconazole and the ketoconazole:nonylphenol (1:5) mixture on CYP1A and CYP3A activities. The substances were dissolved in DMSO and diluted with ethanol. The final concentrations never exceeded 0.01% (v/v) DMSO and 0.001% ethanol (v/v). For CYP1A and CYP3A inhibition studies, pooled liver microsomes from BNF-treated and from untreated Atlantic cod, respectively, were used. For comparison, the IC 50 values for ketoconazole, nonylphenol and ethynylestradiol were determined in cDNA expressed human CYP3A4 baculovirus supersomes using the CYP3A4 inhibition kit from BD Gentest. In vitro incubation studies Pooled liver microsomes from untreated Atlantic cod were pre-incubated for 10, 30 and 60 min, at room temperature, with ethynylestradiol and ketoconazole following CYP3A western blot analysis. The reaction mixture consisted of microsomes (2.5 or 5.0 mg protein/ml) and various concentrations of ethynylestradiol (35, 50 and 100 μM) or ketoconazole (0.3 and 1.0 μM) ± 3 μM NADPH in a total volume of 50 μl in homogenization buffer, containing 20% (v/v) glycerol. The CYP3A western blot analysis was performed as described above. Ethynylestradiol and ketoconazole were dissolved in acetonitrile (vehicle) and the final acetonitrile concentration in the reaction mixture was 0.02% (v/v). Plasma vitellogenin analysis Plasma levels of vitellogenin protein were determined using a non-competitive sandwich ELISA kit and employing rabbit PAb against Atlantic cod vitellogenin from Biosense Laboratories AS (Bergen, Norway) [ 58 ]. Each plasma sample was diluted (1:20, 1:15,000 and 1:50,000) and 100 μl was analyzed and compared to purified Atlantic cod vitellogenin protein standards (ranging between 0.12 and 2,000 ng/ml). The signal was detected at A 492 after 15 min incubation with substrate solution, using a VICTOR™ 1420 Multilabel Counter. Plasma sex steroid hormone analyses Plasma levels of 17β-estradiol and testosterone were determined using competitive EIA kits from Cayman Chemical (Ann Arbor, MI, USA). Plasma levels of 11-keto-testosterone were analyzed using a competitive EIA kit, from Biosense Laboratories AS (Bergen, Norway). Plasma from each fish was concentrated (2:1) by extraction once with six volumes diethyl ether and 50 μl was analyzed and compared to purified standard substances. The signals were detected at A 405 after 40 min (17β-estradiol), 60 min (testosterone) or 80 min (11-keto-testosterone) incubation with substrate solution, using a VICTOR™ 1420 Multilabel Counter. Statistics Data were tested for homogeneity of variances using the Levene's test. When there was homogeneity of variances we used a parametric one-way ANOVA, followed by Scheffé post hoc test. When there was no homogeneity of variances we used the non-parametric Kruskal-Wallis ANOVA, followed by the two-tailed Mann-Whitney U test. No values were log transformed. Data are presented as means (n = 6–9 fish) accompanied with the standard deviations (SD). The significance level (α) was set at 0.05. The statistical analyses were performed using SPSS 11.0 software, from SPSS Sweden AB (Sundbyberg, Sweden). Authors' contributions LH performed most of the analyses, participated in fish exposure, sampling, experimental design and drafted the manuscript. BEG assisted with experimental design, 2D-analysis and writing. AG participated in experimental design and writing. MCC rose funding, coordinated, participated in fish exposure, sampling, experimental design and writing.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549046.xml
522817
Neoadjuvant or adjuvant therapy for resectable esophageal cancer: a clinical practice guideline
Background Carcinoma of the esophagus is an aggressive malignancy with an increasing incidence. Its virulence, in terms of symptoms and mortality, justifies a continued search for optimal therapy. A clinical practice guideline was developed based on a systematic review investigating neoadjuvant or adjuvant therapy on resectable thoracic esophageal cancer. Methods A systematic review with meta-analysis was developed and clinical recommendations were drafted. External review of the practice guideline report by practitioners in Ontario, Canada was obtained through a mailed survey, and incorporated. Final approval of the practice guideline was obtained from the Practice Guidelines Coordinating Committee. Results The systematic review was developed and recommendations were drafted, and the report was mailed to Ontario practitioners for external review. Ninety percent of respondents agreed with both the evidence summary and the draft recommendations, while only 69% approved of the draft recommendations as a practice guideline. Based on the external review, a revised document was created. The revised practice guideline was submitted to the Practice Guidelines Coordinating Committee for review. All 11 members of the PGCC returned ballots. Eight PGCC members approved the practice guideline report as written and three members approved the guideline conditional on specific concerns being addressed. After these recommended changes were made, the final practice guideline report was approved. Conclusion In consideration of the systematic review, external review, and subsequent Practice Guidelines Coordinating Committee revision suggestions, and final approval, the Gastrointestinal Cancer Disease Site Group recommends the following: For adult patients with resectable thoracic esophageal cancer for whom surgery is considered appropriate, surgery alone (i.e., without neoadjuvant or adjuvant therapy) is recommended as the standard practice.
Background Carcinoma of the esophagus is an aggressive malignancy with an increasing incidence. Its virulence, in terms of symptoms and mortality, justifies a continued search for optimal therapy. The large and growing number of patients affected, the high mortality rates, the worldwide geographic variation in practice, and the large body of good quality research warrants a clinical practice guideline. This clinical practice guideline was developed by the Gastrointestinal Cancer Disease Site Group (DSG) of Cancer Care Ontario's Program in Evidence-based Care (PEBC), using the methods of the Practice Guidelines Development Cycle [ 1 ]. This practice guideline report is a convenient and up-to-date source of the best available evidence on neoadjuvant or adjuvant therapy for resectable esophageal cancer, developed through systematic review, evidence synthesis, and input from practitioners in Ontario. The PEBC has a formal standardized process to ensure the currency of each clinical practice guideline report. This process consists of the periodic review and evaluation of the scientific literature and, where appropriate, integration of this literature with the original clinical practice guideline information. The systematic review on neoadjuvant or adjuvant therapy for resectable esophageal cancer, which forms the basis for this clinical practice guideline, is available in a companion document [ 2 ]. Based on the systematic review, draft recommendations were developed by consensus of the Gastrointestinal Cancer DSG to create the clinical practice guideline report. The clinical practice guideline is intended to promote evidence-based practice in Ontario, Canada. As part of the PEBC's clinical Practice Guideline Development Cycle, all draft recommendations are sent to Ontario practitioners for external review. The efficacy of this external review process has been previously described [ 3 ]. The external review is a mailed survey consisting of items that address the quality of the draft practice guideline report and draft recommendations and whether the draft recommendations should serve as a practice guideline. Final approval of this practice guideline report was obtained from the Practice Guidelines Coordinating Committee (PGCC). Methods Clinical practice guideline development Systematic review A systematic review with meta-analysis on neoadjuvant or adjuvant therapy for resectable esophageal cancer was developed by the Gastrointestinal Cancer DSG of Cancer Care Ontario's Program in Evidence-based Care [ 2 ]. The evidence examined did not support the use of neoadjuvant or adjuvant chemotherapy or radiotherapy for resectable thoracic esophageal cancer. Gastrointestinal cancer disease site group consensus In discussions regarding the completed systematic review, the Gastrointestinal Cancer DSG agreed that the evidence did not support a recommendation for neoadjuvant or adjuvant chemotherapy or radiotherapy for resectable thoracic esophageal cancer. A recommendation that surgery alone should be the standard of care for this patient population was drafted, and it was recommended that the draft practice guideline be sent out to Ontario practitioners for external review. The role of radiotherapy alone and chemoradiation alone without surgery is addressed in a separate Gastrointestinal Cancer DSG Clinical Practice Guideline: Combined modality radiotherapy and chemotherapy in the non-surgical management of localized carcinoma of the esophagus [ 4 ]. Results External review Practitioner feedback was obtained through a mailed survey of 163 practitioners in Ontario (27 medical oncologists, 21 radiation oncologists, 112 surgeons, and three gastroenterologists). The survey consisted of items evaluating the methods, results, and interpretative summary used to inform the draft recommendations and whether the draft recommendations should be approved as a practice guideline. Written comments were invited. Follow-up reminders were sent at two weeks (post card) and four weeks (complete package mailed again). The Gastrointestinal Cancer DSG reviewed the results of the survey. Eighty-six surveys (58%) were returned. Twenty-nine respondents (34%) (nine medical oncologists, seven radiation oncologists, and 13 surgeons) indicated that the report was relevant to their clinical practice and completed the survey. Key results of the practitioner feedback survey are summarized below. 1. Number surveyed: 163 practitioners in Ontario, Canada involved in the care of cancer patients 2. Return rate: 58% (mean Gastrointestinal Cancer DSG return rate: 60.2%; range: 51% – 84%) 3. Written comments attached: 10% 4. Agreement with the summary of evidence: 90% 5. Agreement with the recommendation: 90% 6. Approval of the recommendation as a practice guideline: 69% Summary of main findings Three (10%) respondents provided written comments. One practitioner hypothesized that preoperative chemoradiation might have a role in adenocarcinoma of the lower third of the esophagus (as suggested by Walsh et al [ 5 ] with 100% adenocarcinoma and by Urba et al [ 6 ] with 75% adenocarcinoma), but not in squamous cell carcinoma (as suggested by Bosset et al [ 7 ] and by Le Prise et al [ 8 ]). Another respondent noted that the survival advantage at three years for combined treatment for preoperative chemoradiotherapy is discounted in the guideline report, and suggested that the guideline recommend the selection of the option preferred by informed patients. There was a request for an algorithm to help in deciding between surgical and non-surgical treatment. The same respondent commented on the limited discussion on quality of life. Two radiation oncologists disagreed with the recommendations and thought that the draft practice guideline report should not be approved as a practice guideline, but neither provided written comments. Discussion Gastrointestinal cancer disease site group modifications and actions After completion of the practitioner feedback survey, additional trials were found. The results of two randomized trials both found surgery alone to be significantly superior to radiation alone [ 9 , 10 ], which resulted in an original draft recommendation regarding radiation alone as a primary modality for localized esophageal cancer being removed from the final practice guideline. In response to this feedback, the Gastrointestinal Cancer DSG acknowledged that the majority of studies have been performed in squamous cell carcinomas. While adenocarcinomas were included in some studies, a distinction between the two histological subtypes was not made because previous studies have not consistently found that they respond differently to chemotherapy or radiation, and nine references [ 11 - 19 ] were added to support this. The Gastrointestinal Cancer DSG did not feel the evidence was compelling enough to recommend preoperative chemoradiotherapy over surgery alone based on the three-year data. After consideration, the Gastrointestinal Cancer DSG decided not to create an algorithm as suggested as a similar project is currently under development. After addressing the comments obtained from practitioners during the external review, the Gastrointestinal Cancer DSG voted that the overall guideline recommendations should be approved, and submitted the practice guideline to the Practice Guidelines Coordinating Committee for review. Practice guidelines coordinating committee approval process The practice guideline report was circulated to members of the Practice Guidelines Coordinating Committee for review and approval. All 11 members of the PGCC returned ballots. Eight PGCC members approved the practice guideline report as written and three members approved the guideline conditional on the Gastrointestinal Cancer DSG addressing specific concerns. PGCC members requested that the following issues be addressed prior to the approval of the guideline report: One member noted that although the majority of studies had been performed in squamous cell carcinomas, some studies included adenocarcinomas, and it would be helpful if the pathological subtypes were discussed. In particular, this member wanted to know if there was any difference in response or outcome for the two histological subtypes. Another member noted that although the pooled analysis for preoperative chemoradiation versus surgery alone detected no difference at one year, the pooled estimate almost reached significance. This member was concerned that the discussion may be too dismissive of the data, and suggested there be some acknowledgment that further follow-up and additional studies are needed. In response to this feedback, the Gastrointestinal Cancer DSG expanded on the earlier revisions concerning the similarities in response to treatment between squamous cell carcinomas and adenocarcinomas. Also, after the original practice guideline was submitted to the PGCC, two meta-analyses [ 20 , 21 ] both detecting a statistically significant difference in survival at three years favouring preoperative chemoradiation versus surgery alone were obtained. The Gastrointestinal Cancer DSG re-pooled the mortality data from the six trials [ 5 - 8 , 22 , 23 ] at three years and obtained similar results. Conclusions In consideration of the systematic review, external review, and subsequent Practice Guidelines Coordinating Committee revision suggestions, and final approval, the Gastrointestinal Cancer Disease Site Group developed the following Clinical Practice Guideline: Practice guideline This practice guideline reflects the most current information reviewed by the Gastrointestinal Cancer DSG. Target population These recommendations apply to adult patients with resectable and potentially curable thoracic (lower two-thirds of esophagus) esophageal cancer for whom surgery is considered appropriate. Recommendation • If surgery is considered appropriate, then surgery alone (i.e., without neoadjuvant or adjuvant therapy) is recommended as the standard practice for resectable thoracic esophageal cancer. This Clinical Practice Guideline report is based on work completed in October, 2003. All approved PEBC Clinical Practice Guideline reports are updated regularly. Please see the PEBC's web site for a complete list of current and on-going projects. Competing interests The author(s) declare that they have no competing interests. List of abbreviations used In order of appearance: DSG, Disease Site Group; PEBC, Program in Evidence-based Care; PGCC, Practice Guidelines Coordinating Committee. Authors' contributions RM, RW, and LZ created the initial drafts of this clinical practice guideline with input from other members of the Gastrointestinal Cancer DSG. RM, RW, and BR created the final draft of this clinical practice guideline. Creation of the submitted manuscript was performed by BR and RM. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522817.xml
554981
Survivin 2α: a novel Survivin splice variant expressed in human malignancies
Background Survivin and its alternative splice forms are involved in critical cellular processes, including cell division and programmed cell death. Survivin is expressed in the majority of human cancers, but minimally in differentiated normal tissues. Expression levels correlate with tumor aggressiveness and resistance to therapy. Results In the present study, we identify and characterize a novel survivin isoform that we designate survivin 2α. Structurally, the transcript consists of 2 exons: exon 1 and exon 2, as well as a 3' 197 bp region of intron 2. Acquisition of a new in-frame stop codon within intron 2 results in an open reading frame of 225 nucleotides, predicting a truncated 74 amino acid protein. Survivin 2α is expressed at high levels in several malignant cell lines and primary tumors. Functional assays show that survivin 2α attenuates the anti-apoptotic activity of survivin. Subcellular localization and immunoprecipitation of survivin 2α suggests a physical interaction with survivin. Conclusion We characterized a novel survivin splice variant that we designated survivin 2α. We hypothesize that survivin 2α can alter the anti-apoptotic functions of survivin in malignant cells. Thus survivin 2α may be useful as a therapeutic tool in sensitizing chemoresistant tumor cells to chemotherapy.
Background Alternative splicing is estimated to occur in 40–60% of all human genes, accounting for some of the discrepancies between the large number of known proteins and the three-fold lower number of human genes in the genome. Alternative splicing generates a multitude of isoforms that have overlapping but distinct functions during embryonic development and that also contribute to maintaining homeostasis in adult differentiated tissues (reviewed in [ 1 ]). Alternative splice forms of key proteins in cancer, TP53, MDM2 [ 2 ] and c-MYC [ 3 ], have been shown to play a role in oncogenesis. Survivin was originally identified by structural homology to IAPs in human B-cell lymphoma [ 4 ]. It is composed of a single BIR domain and an extended carboxy-terminal coiled coil domain [ 5 ]. Transcription from the Survivin locus gives rise to alternatively spliced transcripts identified in both human and mice [ 6 - 8 ]. To date, three alternatively spliced isoforms have been described in humans [ 6 - 8 ]. Survivin-2B is generated by the insertion of an alternative exon, exon 2B; Survivin-Δ Ex3 arises from the removal of exon 3 resulting in a frameshift and translation of part of the 3'UTR generating a unique carboxy-terminus; Survivin-3B results from the introduction of a novel exon 3B resulting in a frameshift and premature termination of the protein (Figure 1A ). Figure 1 Structural analysis of survivin 2α compared with the other human survivin isoforms. A: The survivin pre-mRNA generates at least five mature mRNA transcripts. Boxes represent exons, with the sizes indicated below. The size of the additional nucleotide sequences for survivin 3B and survivin 2α are shown. In survivin 2α, 197nt of intron 2 are added, of which 195nt are non-coding. Protein domains and motifs are indicated above the diagrams. The black arrow indicates the stop codon; the coiled-coil domain is shown by the dotted line; and the double lines represent the BIR domain. B: Survivin 2α transcript and predicted protein sequence. The sequence for the survivin 2α transcript was obtained by sequencing the insert contained within the IMAGE clone. C: Predicted 3D structural modeling of survivin 2α and survivin. The amino acid sequence of survivin 2α was used with SWISS-MODEL to predict a 3D structure by homology modeling. The resulting PDB file was visualized and manipulated using Swiss PBD-Viewer for the view presented. The 3D structure for survivin was obtained from Swiss-PROT database (PDB entry 1F3H). The yellow arrows indicate regions of differences between survivin 2α and survivin. The red arrows represent the first 2 alpha helices of the BIR domain, and the pink arrow represents the 3 rd helix, that is absent in survivin 2α. The green arrow points to the C-terminal coiled-coil domain. D: Protein Analysis of survivin 2α. Total HeLa cell lysate was loaded on a 18% SDS-PAGE and transferred into nitrocellulose membrane. The proteins were detected by immunoprobing with a polyclonal survivin antibody. A protein of approximate molecular weight 8.5 kDa, corresponding to the predicted size of survivin 2α is detected in HeLa cell lysates. Survivin has 2 main functions; one as a chromosomal passenger protein [ 9 ] and the other as an inhibitor of apoptosis [ 10 ]. Survivin 2B has been shown to be a pro-apoptotic protein that sensitizes resistant leukemia cells to chemotherapy in a p53-dependent fashion [ 11 ]. Survivin-Δ Ex3 functions as an anti-apoptotic protein and is upregulated in malignancies (Mahotka et al., 1999). No function has yet been described for survivin-3B. In this report we identify and characterize a novel isoform of survivin, survivin 2α. We show that survivin 2α is expressed at high levels in malignant cells, co-localizes with survivin and has the potential to attenuate the anti-apoptotic effect of survivin. Results and Discussion Structural Characteristics of Survivin 2α In this work, we characterized a novel isoform of the survivin locus. We surveyed the aligned survivin EST sequences available in the UCSC Human Genome Browser and identified an EST from a human breast tumor cDNA library (I.M.A.G.E. clone 1631662). We sequenced the entire cDNA and designated it Survivin 2α. The complete cDNA sequence is shown in Figure 1B . The protein contains the coding sequences from exon 1 and exon 2, and one additional amino acid before termination (Figure 1B ). This 74 amino acid protein, with a predicted molecular weight of 8.5 kDa, contains a truncated BIR domain and lacks the carboxy-terminal coiled-coil domain in its entirety (Figure 1A ). There are no defined localization signals in the protein, and PSORTII predicts localization within the nucleus and the cytoplasm (Table 1 ). Alignment with the known human survivin isoforms shows that the sequence of Survivin 2α is identical to exons 1 and 2 of the other survivin splice variants, with the exception of the last amino acid. Alignment of Survivin 2α with the three mouse survivin isoforms also reveals some similarity with survivin40, a 40-amino acid mouse splice variant (not shown). The 3D predicted structure of Survivin 2α shows the absence of the alpha-helical coiled-coil domain, present in survivin (Figure 1C ). It also shows minor predicted rearrangements in the structure that may occur to stabilize the protein (Figure 1C , yellow arrows). These re-arrangements occur within the BIR domain, and could have functional implications for the role of Survivin 2α in apoptosis. Table 1 Table of the predicted localization and structural features of survivin and the novel isoform survivin 2α. Localization Survivin Survivin 2α Cytoplasm 56.5% 39.1% Nucleus 17.4% 34.8% Cytoskeleton 0% 4.3% Golgi Apparatus 0% 0% Plasma membrane 4.3% 0% ER 4.3% 4.3% Peroxysome 0% 0% Mitochondria 13.0% 13.0% Lysosomes 4.3% 4.3% Features BIR 1 Partial Coiled-Coil 1 0 Protein Size 142 aa 74 aa Predicted Molecular Weight 16.4 kDa 8.5 kDa The BIR domain has been shown to be important for homodimerization and coordination of the zinc atom co-factor [ 12 ]. In the survivin protein, Histidine 80 (H80) is required for zinc atom coordination and homodimerization. Expression constructs containing mutations at this residue within the Survivin protein have previously been shown to accelerate PCD (Programmed Cell Death) in vitro. Similarly, mutations in Cytosine 84 (C84) enhance PCD, as a result of displacement of the wild type Survivin protein [ 13 ]. The Survivin 2α protein, truncated at amino acid 74, lacks both of these amino acid residues. Additionally, Survivin 2α lacks the third alpha helix in the BIR domain. As the anti-apoptotic function of Survivin is mediated both by the BIR domain and by the interaction of its C-terminal coiled coil domain with microtubules of the mitotic spindle [ 10 , 14 , 15 ], it would be predicted that Survivin 2α might not have anti-apoptotic properties. Survivin 2α is highly expressed in tumor cells Survivin is critical for global normal embryonic development, as demonstrated by the early embryonic lethality of mice with homozygous deletions in the survivin gene locus [ 16 ]. Survivin proteins are virtually absent from most normal differentiated tissues, however these proteins are expressed in certain highly proliferative areas within normal tissues [ 17 - 19 ]. In contrast, survivin is highly expressed in the majority of human malignancies, derived from different cell origins. We evaluated the expression of survivin 2α in 7 different cancer cell lines, 2 non-transformed tissues and 7 primary medulloblastoma tumors by quantitative PCR. We designed primers that will specifically amplify Survivin 2α after selection of polyadenylated RNA. Survivin 2α expression in tumor cells and primary medulloblastoma tumors varied from 2–100 fold above non transformed cells (Table 2 ). The levels of Survivin 2α transcripts are comparable to those of Survivin ΔExon3 (Table 3 ). Like Survivin, Survivin 2α is expressed at increased levels in transformed cells compared to non-transformed cells, and therefore it suggests that it could have a role in tumorigenesis. Additionally, we detected expression of endogenous Survivin 2α protein in HeLa cells, suggesting that the transcript is translated (Figure 1D ). Table 2 Survivin 2α expression (relative to normal tissue). Cell Type Relative Increase Normal Cerebellum 1.00 Normal Breast (MCF10A) 0.97 Breast Carcinoma (MCF7) 8.17 Osteosarcoma (U2OS) 39.06 Lung (A549) 3.03 ALL (Jurkat) 1.84 Soft Tissue Sarcoma (RH28) 94.90 Cervical Carcinoma (HeLa) 58.22 Medulloblastoma (Daoy) 34.23 Primary Tumors Medulloblastoma #1 4.68 Medulloblastoma #2 154.55 Medulloblastoma #3 93.24 Medulloblastoma #4 5.69 Medulloblastoma #5 8.54 Medulloblastoma #6 9.81 Medulloblastoma#7 75.10 Table 3 Expression of survivin splice variants in medulloblastoma (relative to survivin) Tissue Survivin 2B Survivin ΔEx3 Survivin 2α Medulloblastoma #3 37.63 0.46 0.36 Medulloblastoma #4 43.82 0.18 0.18 Medulloblastoma (Daoy) 1290.16 35.10 0.66 Functional Properties of Survivin 2α To characterize a function for Survivin 2α, we transfected Daoy cells with Survivin 2α and a combination of Survivin 2α and Survivin. To induce apoptosis in the Daoy cells, we treated them with 2 μM of the chemotherapeutic agent vincristine. Vincristine is a vinca alkaloid that binds to tubulin, inhibiting microtubule polymerization. It kills Daoy cells in culture by PCD. We analyzed early apoptotic events in vincristine-treated transfected cells by Annexin V staining. Survivin 2α antagonized the anti-apoptotic effect of Survivin in co-transfection assays with or without a cell death stimulus (not shown and Figure 2A ). As inhibition of apoptosis by Survivin involves activation of the caspase pathway [ 20 ], we assayed Survivin 2α transfected cells for caspase 3 activation. Caspase-3 was strongly activated in vector control and Survivin 2α transfected cells in the presence of vincristine. Much lower levels of caspase-3 activation were observed in Survivin-transfected cells (Figure 2B ). In the absence of an apoptotic stimulus we observed a 35% increase of caspase-3 activity in Survivin 2α cells, as well as a 46% increase in early apoptosis, as assessed by Annexin V staining. We also performed electron microscopy analysis of Survivin 2α transfected and non-transfected cells. We sorted transfected cells from non-transfected cells by FACS based on GFP fluorescence, and processed each population for EM analysis (Figure 2C ). Overall, there was a 43% increase in incidence of apoptosis in Survivin 2α-expressing cells versus non-expressing cells. Our results suggest that Survivin 2α can attenuate survivin's anti-apoptotic activity and sensitize tumor cells to chemotherapy. These findings have important therapeutic implications in the treatment of chemoresistant tumors. Figure 2 Early apoptosis in tumor cells transfected with survivin and survivin 2α. A: Daoy cells were transiently transfected with control vector, survivin, survivin 2α, or a combination of survivin and survivin 2α. Apoptosis was measured using Annexin V/PI staining 24 hours after transfection, in the absence of a cell death stimulus. The results are shown relative to empty vector control. Error bars represent standard deviations from triplicate experiments. Results were adjusted for transfection efficiency based on parallel transfection with a GFP-expressing plasmid. B: Daoy cells were transiently transfected with control, survivin or survivin 2α. Apoptosis was measured by Caspase-3 activity following treatment with vincristine. C: HeLa cells were transiently transfected with control or GFP-tagged survivin 2α. Electron Microscopy analysis of transfected cells shows a representative cell undergoing apoptosis as induced by survivin 2α. Images were taken 12,000× magnification and scale bars are shown. Survivin 2α alters the subcellular localization of survivin To characterize the subcellular localization of survivin 2α we performed direct fluorescence assays in HeLa cells transfected with a GFP- survivin 2α construct. Survivin 2α localized to the nucleus and the cytoplasm in interphase cells (Figure 3A ). In cells undergoing mitosis, survivin 2α was confined to the cytoplasmic compartment (Figure 3B ). Interestingly, when co-expressed with survivin, survivin 2α co-localized with survivin to the centromeres of the chromosomes in prometaphase (Figure 3C ) and metaphase (Figure 3D ), and at the midbody during late telophase/cytokinesis (Figure 3E ). Moreover, the normal cytoplasmic localization of survivin shifted to the nucleus in interphase cells. This data suggests a direct interaction between the two proteins, as well as a potential mechanism for the attenuation of survivin's anti-apoptotic activity by survivin 2α. Figure 3 Confocal microscopy analysis of survivin 2α sub-cellular localization. HeLa cells were transfected with GFP-survivin 2α or HcRed survivin, as detailed on the top of the figure. Green pixels correspond to GFP expression, red pixels correspond to HcRed expression and blue pixels represent DNA labeled with Hoechst dye. When co-localization of GFP and HcRed occurs the pixels are yellow. A: Expression of survivin 2α at interphase localizes to nuclear and cytoplasmic structures. B: During M-phase survivin 2α is excluded from the condensed/dividing chromosomes and is localized in the cytoplasm of the dividing cell. C, D, E, F: When co-expressed with survivin, survivin 2α localization does not change at interphase. During M-phase survivin 2α co-localizes with survivin to the centromeres of the dividing chromosomes (D and E), and in the midbody region at cytokinesis (F). Scale bar = 5 μm Survivin 2α physically interacts with survivin To further investigate the possibility that survivin 2α interacts with survivin we performed co-immunoprecipitation experiments. We co-transfected HeLa cells with constructs encoding a Flag-survivin fusion protein and a myc-survivin 2α fusion protein. We used a Flag antibody to precipitate protein complexes, and a myc antibody to detect myc-tagged survivin 2α. We detected survivin 2α-myc in the complexes precipitated with the Flag antibody, substantiating a physical interaction of survivin with survivin 2α (Figure 4 ). Figure 4 Co-immunoprecipitation of survivin-survivin 2α. HeLa cells were transfected with constructs encoding tagged forms of survivin (Flag) and survivin 2α (myc). Lysates from transfected cells were subjected to immunoprecipitation with an antibody against the Flag epitope. The resulting immunoprecipitated complexes were resolved by SDS-PAGE and subjected to Western blotting. The membrane was immunoprobed with an antibody against the myc epitope tag. Survivin 2α-myc is clearly visualized in lysates precipitated with a Flag antibody. Conclusion We characterized a novel survivin splice variant that we designated survivin 2α. We hypothesize that survivin 2α can alter the anti-apoptotic functions of survivin in malignant cells. Thus, survivin 2α may be useful as a therapeutic tool in sensitizing chemoresistant tumor cells to chemotherapy. Methods Patient samples Seven fresh frozen primary medulloblastoma tumor samples were obtained from the Cooperative Human Tumor Network (CHTN), after approval through the Columbus Children's Hospital IRB. Sequencing IMAGE clone 1631662 (Invitrogen) was sequenced using primers that flanked the multiple-cloning-site. Plasmids and Cloning The cDNA for survivin 2α was amplified from the EST clone (Invitrogen) and cloned into the Kpn I- Bam HI sites of pcDNA4/TO/myc-HisB (Invitrogen) generating an in-frame fusion with the C-terminal myc-tag, or into the Kpn I- Bam HI sites of pEGFP-N3 generating an in-frame fusion with the C-terminal GFP tag. The start codon in both constructs corresponds to the naturally occurring start codon in the cDNA transcript. The resulting clones were confirmed by sequencing. Cell Culture and Transfection HeLa (cervical adenocarcinoma), Daoy (medulloblastoma), Jurkat (acute lymphoblastic leukemia) and MCF-7 (breast adenocarcinoma) cells (ATCC) were grown in DMEM supplemented with 10% FBS at 37°C, 5% CO 2 ; U2OS osteosarcoma cells (kindly donated by Dr. Greg Otterson) were grown in McCoy's 5A medium supplemented with 10% FBS at 37°C, 5% CO 2 ; RH28 (alveolar rhabdomyosarcoma, kindly donated by Dr. Steve Qualman) and A549 (lung carcinoma) (ATCC) were grown in RPMI1640 supplemented with 10% FBS at 37°C, 5% CO 2 . MCF10-A, a non-transformed breast cell line (ATCC) was grown in MEGM, Mammary Epithelial Growth Medium, Serum-free, (Clonetics) supplemented with 100 ng/ml cholera toxin (Sigma Aldrich) at 37°C, 5% CO 2 . Transient transfections were performed using Effectene transfection reagent (Qiagen) at a DNA: Effectene ratio of 1:10. Drug Treatment Induction of apoptosis by vincristine was done by treatment of cells with complete growth medium supplemented with vincristine sulfate at a final concentration of 2 μM. RNA isolation and Real Time PCR RNA was isolated from 10 6 proliferating cells or frozen tumor tissue using TriZol reagent (Invitrogen) as recommended by the supplier. Poly(A) RNA was purified using Oligotex dT kit (Qiagen). 100 ng of poly(A) purified RNA was used as a template in a reverse transcription reaction using random hexamers and Omniscript Reverse transcriptase (Qiagen) and performed according to manufacturer's instructions. Quantitative real-time PCR reactions using Taqman probes (FAM/TAMRA) were run in triplicate on an ABI Prism 7700 Real-time PCR machine (Applied Biosystems). Control GAPDH reactions (Applied Biosystems) were run to normalize ΔCt values. Relative change was calculated by the comparative C T method, 2 (-ΔΔCt) . The survivin 2α specific primers consist of: Forward 5'GCTTTGTTTTGAACTGAGTTGTCAA; Reverse 5'GCAATGAGGGTGGAAAGCA; and Probe: 6FAM AGATTTGAGTTGCAAAGACACTTAGTATGGGAGGG TAMRA Apoptosis Assays Two apoptosis assays were performed: Caspase-3 assay and Annexin-V FLUOS. For caspase assays 2,000 cells from each experimental condition were subjected to the caspase-3 assay, Caspase 3/7 GLO (Promega) and analyzed on a Victor3 plate reader (Applied Biosystems). Experiments were performed in triplicate. Annexin V/propidium iodide staining was carried out using the Roche Annexin-V-Fluos Staining Kit following the manufacturer's instructions. Fluorescein and propidium iodide fluorescence measured with a Coulter EPICS XL flow cytometer. Experiments were performed in triplicate. Microscopy Proliferating HeLa cells, grown on glass coverslips, were transiently transfected with a GFP-tagged survivin 2α expression construct or co-transfected with GFP-tagged survivin 2α and HcRed-tagged survivin. 24 hours post-transfection the cells were fixed in 4% paraformaldehyde and stained with 50 μg/ml Hoechst dye. Cells were analyzed on a Zeiss LSM510 META confocal microscope, using a 63x PlanApochromat objective. For electron microscopy analysis, proliferating HeLa cells were transfected with GFP-tagged survivin 2α construct for 12 hours. The cells were aseptically sorted by FACS based on green fluorescence from GFP-survivin 2α for positive and negative populations. This was done in order to separate an enriched population that consisted of >90% GFP expressing cells. 10 6 cells for each condition were fixed in 2.5% gluteraldehyde for 24 hours and processed for EM. For cell analysis, 10 to 12 fields containing 8–10 cells per field at a magnification of 3500× were used. At least 100 cells were counted for each experimental condition and assigned to categories of healthy or dying based on their morphological appearance, including nuclear integrity. Image collection was performed on a Hitachi H-600 transmission electron microscope equipped with a GATAN image acquisition system. Co-Immunoprecipitation HeLa cells transfected with Flag-survivin and survivin-2α myc were collected in Cell Lysis Buffer (100 mM Tris-HCl pH8.0, 100 mM NaCl, 0.5% Triton X-100, 0.2 μM PMSF) and incubated at 4°C for 30 min. The cell lysate was clarified by centrifugation and the clarified supernatant dissolved 1:5 in Co-IP buffer (50 mM Tris-HCl pH7.5, 15 mM EGTA, 100 mM NaCl, 0.1% Triton X-100, 1x protease inhibitors cocktail, 1 mM DTT, 1 mM PMSF). The equivalent of 400 μg of lysate total protein was incubated with 2 μg of anti-Flag M2 antibody at 4°C for 1 h with constant rotation. As a control the same amount of lysate protein was incubated in the absence of antibody. Fifty microliters of agarose-conjugated protein A (Invitrogen) were added and the mixture incubated for a further hour in the same conditions. The protein-antibody-protein A complexes were pulled down by centrifugation and subjected to 3 washes with co-IP buffer. The proteins were analyzed through electrophoretic separation in a 20% SDS-PAGE, electroblotted onto nitrocellulose and immunoprobed with an antibody against myc-tag. Detection was performed using the ECL kit (Amersham). Protein standards were used for size determination. Bioinformatics Subcellular localization predicted by PSORTII program. Coiled-Coil domain predicted by Coils and PairCoil programs Authors' contributions HC performed bioinformatic analysis, subcellular localization, functional studies, co-immunoprecipitation and drafted the manuscript. LH performed quantitative real time PCR in cell lines and primary tumors. RA conceived the study and participated in its design and coordination, and was responsible for overseeing the final version of the manuscript. All authors have read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554981.xml
506779
Injuries at a Canadian National Taekwondo Championships: a prospective study
Background The purpose of this prospective study was to assess the injury rates in male and female adult Canadian Taekwondo athletes relative to total number of injuries, type and body part injured. Methods Subjects (219 males, 99 females) participated in the 1997 Canadian National Taekwondo Championships in Toronto, Canada. Injuries were recorded on an injury form to documents any injury seen and treatment provided by the health care team. These data were later used for this study. The injury form describes the athlete and nature, site, severity and mechanism of the injury. Results The overall rate of injuries was 62.9/1,000 athlete-exposures (A-E). The males (79.9/1,000 A-E) sustained significantly more injuries than the females (25.3/1,000 A-E). The lower extremities were the most commonly injured body region in the men (32.0 /1,000 A-E), followed by the head and neck (18.3/1,000 A-E). Injuries to the spine (neck, upper back, low back and coccyx) were the third most often injured body region in males (13.8/1,000 A-E). All injuries to the women were sustained to the lower extremities. The most common type of injury in women was the contusion (15.2/1,000 A-E). However, men's most common type of injury was the sprain (22.8/1,000 A-E) followed by joint dysfunction (13.7/1,000A-E). Concussions were only reported in males (6.9/1,000 A-E). Compared to international counterparts, the Canadian men and women recorded lower total injury rates. However, the males incurred more cerebral concussions than their American colleagues (4.7/1,000 A-E). Conclusions Similar to what was found in previous studies, the current investigation seems to suggest that areas of particular concern for preventive measures involve the head and neck as well as the lower extremities. This is the first paper to identify spinal joint dysfunction.
Background With the inclusion of Taekwondo as a medal sport in the 2000 Olympic Games interest has gained momentum from participants, national governments and scientists alike. With increased participation, the issue of safety becomes very important and timely. The rules and regulations with regards to these Championships were that of the World Taekwondo Federation [ 1 ]. All athletes were sixteen years of age and older and held black belt first degree as a minimum requirement to compete. At the time of data collection, punches were allowed to the front of the torso in the area covered by the chest protector worn by the athletes. No punches were allowed to the head or other parts of the body. Kicks were allowed to the torso and head, the latter of which was covered by a helmet similar to the one worn in amateur boxing. Regardless of the area of contact, only one point was granted by the referees for a successful blow. One could win the match by means of a knockout, therefore, contact was encouraged. There was a change of rules introduced in 2003, which included granting two points for head shots and an additional point for an eight-count knockout [ 1 ]. Prospective studies on Taekwondo injuries sustained at single tournaments have been conducted before [ 1 - 3 ]. For instance, Zemper and Pieter [ 2 ] found injury rates for American elite male Taekwondo athletes to be 127.4/1,000 athlete-exposures and for females, 90.1/1,000 athlete-exposures. One athlete-exposure (A-E) refers to one athlete being exposed to the possibility of being injured. Since there are always two athletes competing during any one bout, there are two athlete-exposures per bout [ 2 , 3 ]. In a later study, Pieter et al. [ 3 ] reported injury rates of 139.5/1,000 A-E and 96.5/1,000 A-E for European men and women, respectively. In neither study were the differences tested for statistical significance. However, at a recreational tournament in the United Kingdom, the men (51.3/1,000 A-E) sustained statistically significantly more injuries than the women (47.6/1,000 A-E) [ 4 ]. It is not clear why the men in the studies mentioned above sustained more injuries. Sample size may be a factor. For instance, in a prospective study covering multiple Taekwondo tournaments, Pieter and Zemper [ 5 ] reported a statistically significantly higher injury rate for the women (105.5/1,000 A-E versus 95.1/1,000). In the largest prospective judo injury study to date, young and adult females (130.6/1,000 A-E) incurred a higher injury rate than their male counterparts (122.6/1,000 A-E) [ 6 ]. In addition to sample size, the number of female competitors is also lower [ 5 , 6 ]. As expected in a collision sport, the contusion was the most frequently occurring injury type in both male and female Taekwondo athletes [ 2 - 4 ]. At the elite level, more serious injuries such as fractures and cerebral concussions also occur [ 3 , 7 ]. Men seem to incur more of these serious injuries [ 2 , 3 ]. Less information is available in karate, for the authors have not consistently reported the injuries by gender. Based on what is known from prospective studies on elite karate athletes, the men also seem to incur more serious injuries, some of which have led to time loss [ 8 , 9 ]. The body region most frequently affected in single tournaments involving both recreational and elite Taekwondo athletes is the lower extremities, especially the (instep of the) foot [ 2 - 4 ]. Since full-contact or Olympic Taekwondo is characterized by kicking, this should come as no surprise. There is a lack of data on gender differences in injuries to body region and body part in karate. Most prospective studies combined the injury rates for males and females [e.g., [ 8 ]]. Using a prospective design, Pieter [ 10 ] reported that the head and neck sustained most of the injuries in both elite male and female karate athletes. There is also scarce information on male-female comparisons on body regions injured in judo athletes. Our own studies in judo seem to indicate that in the women, the upper extremities are mostly affected, while in the men, the head and neck as well as the lower extremities are injured most often [ 11 , 12 ]. In line with the frequent use of the legs in Taekwondo, the main injury mechanism was found to be delivering or receiving a kick [ 2 , 5 ]. Further analysis revealed that the roundhouse kick was most often implicated, especially in men [ 3 , 4 , 13 ]. The fact that injuries occur as a result of receiving a kick may be partially related to unblocked attacks, which has led to the recommendation for the coaches to work on improving the blocking skills or evasive maneuvers of their athletes [ 2 ]. The purpose of this study was to identify and compare the rates of injury in Canadian male and female Taekwondo competitors relative to total number of injuries, type, body part injured and mechanism. Methods Subjects (219 males and 99 females) participated in the 1997 Canadian National Taekwondo Championships in Toronto, Canada. Injuries were recorded on an injury form to document any injury seen and treatment provided by the health care team as it was required by law. The first author was the only person who kept the injury forms and entered the data, therefore, keeping the identity of athletes confidential. Oral consent was obtained from the athletes for assessment and providing therapy. Data describe the athlete and nature, site, severity and mechanism of the injury. No reliability and validity information for the instrument is available and this study was carried out to pre-test the injury data collection form (Figure 1 ). Injuries were diagnosed by the tournament physician (MK), who has been the national team chiropractic physician for several years and is an experienced (black belt) Taekwondo athlete himself. One injury form was filled out by the attending physician for each time the athlete reported a new injury. However, at each presentation there could be more than one injury reported on the same Injury Report form. For the purposes of this study, an athlete was considered injured if any of the following conditions applied [ 14 ]: 1) any circumstance that forced the Taekwondo athlete to leave the competition; 2) any circumstance for which the referee or athlete had to stop competition; 3) any circumstance for which the athlete requested medical attention. In other words, the definition included so-called time-loss injuries (stoppage of a bout) as used in the NCAA Injury Surveillance System [ 15 ]. Injury rates were calculated from matches fought using the basic rate formula: (# injuries / # athlete-exposures) × 1,000 = # injuries per 1,000 athlete-exposures (A-E). The Colorado concussion classification was utilized in management of the concussions [ 16 , 17 ]. According to this classification, a first degree concussion is identified by confusion, no loss of memory and no loss of consciousness (LOC). A second degree involves confusion, loss of memory but no LOC and the third degree is when there is LOC [ 16 , 17 ]. Results The age range for the males was 17–34 years with a mean of 24.2 years and for the females, 16–26 years with a mean of 21 years. The age was not recorded for 3 males and 1 female. Table 1 displays the injury data and rates for the Canadian Taekwondo athletes. The lower extremities were the most commonly injured body region in the men (32.0 /1,000 A-E), followed by the face (eyes, nose, cheek, lips, jaw; 18.3/1,000 A-E), and the spine (neck, upper back, low back and coccyx) (13.8/1,000 A-E). If the head and neck (which includes the face area) are combined, as was done in other studies [ 4 , 7 ], this body region incurred the second highest injury rate: 24.9/1,000 A-E. All injuries to the women were sustained to the lower extremities with the foot incurring most of the injuries (15.2/1,000 A-E; Table 2 ). The top five injuries in the males include the sprain (22.8/1,000 A-E), followed by the joint dysfunction (13.7/1,000 A-E), contusion and laceration (11.4/1,000 A-E each), and strain (9.1/1,000 A-E). The cerebral concussion is ranked sixth (6.9/1,000 A-E). There were one third degree and two first degree concussions. In the women, the contusion was the most often occurring injury (15.2/1,000 A-E), followed by the sprain and strain (5.1/1,000 A-E each; Table 3 ). Table 4 displays the rates of the injury mechanisms by gender. Receiving a kick by the men included those connecting with the head/face (18.3/1,000 A-E), trunk (6.9/1,000 A-E), and thigh (2.3/1,000 A-E). Delivering a kick as an injury mechanism in the men included, among others, kicking to the elbow (6.9/1,000 A-E), kicking with the toes, to the trunk and with the knee (2.3/1,000 A-E each). In the women, delivering a kick (10.1/1,000 A-E) was the main injury mechanism and comprised kicks to the elbow, while the kick that was received involved a knee kick. Discussion The injuries incurred by the Canadian Taekwondo athletes compare favourably to those found by others. As mentioned above, American elite athletes recorded injury rates of 127.4/1,000 A-E (men) and 90.1/1,000 A-E (women) [ 2 ], while European colleagues had rates of 139.5/1,000 A-E (men) and 96.5/1,000 A-E (women) [ 3 ]. At one Greek national championship, the men (20.6/1,000 A-E) sustained statistically significantly fewer injuries than the women (36.4/1,000 A-E) [ 13 ]. The total injury rate of the Canadian male and female Taekwondo athletes combined, was also lower than that of their African counterparts (86.6/1,000 A-E) [ 7 ]. At an Open British tournament, injury rates of 51.3/1,000 A-E (men) and 47.6/1,000 A-E (women) were reported [ 4 ]. However, the athletes competing at this particular tournament were of sub-elite level. It is hypothesised that injuries may be related to level of skill and experience, although confirmatory research still needs to be carried out [ 18 ]. All competitors were black belts, but no information is available on their experience in Taekwondo and in competition. Future research should include general Taekwondo as well as competition-specific experience in addition to belt level. Comparative data gleaned from prospective studies on other martial arts injuries incurred at single tournaments are depicted in table 5 . In view of the nature of the sport, as alluded to above, it is not surprising to find the lower extremities to sustain most of the injuries as was found in previous studies as well [ 2 , 3 , 13 ]. Within the lower extremities, the foot (i.e., instep) was the most often injured body part, as was the case with the females in the present study, which led to the suggestion for the Taekwondo governing bodies to recommend padding to help decrease injuries to this site [ 22 ]. In karate, on the other hand, the head and neck incur most of the injuries [ 8 , 10 ], while in judo, the upper extremities are more at risk [ 6 , 11 ]. As expected, the contusion was found to be the most frequently occurring injury type in other studies on Taekwondo injuries [e.g., [ 2 , 3 , 13 ]]. The sprain ranked in the top three of most frequently occurring injuries across several tournaments [ 5 ]. The contusion was also the most often occurring injury in karate [ 9 , 10 ], while the epistaxis ranked second in Dutch men and women [ 10 ]. In Finnish elite male karate athletes the laceration was ranked second and the epistaxis in women [ 9 ]. In judo, the strain in men and the abrasion in women were sustained most often [ 11 , 12 ]. Since the sample sizes in the current investigation as well as in the aforementioned karate and judo studies are rather small, more research is needed to arrive at more definitive conclusions. However, differences between the martial arts in terms of techniques used and competition rules undoubtedly play a major role. Of more concern, however, is the occurrence of cerebral concussions. The Canadian males recorded a higher rate than found in American (4.7/1,000 A-E) [ 2 ] and Greek (1.0/1,000 A-E) [ 13 ] elite Taekwondo athletes, but lower than those competing in the 1993 European Cup (15.5/1,000 A-E) [ 3 ] and the 1991 World Championships (15.3/1,000 A-E) [ 23 ]. The Canadians also recorded lower rates than elite Dutch karate athletes (13.2/1,000 A-E) competing under semi-contact rules [ 10 ]. Given the serious implications of these injuries, preventive measures, testing of equipment and follow-up research are urgently needed [e.g., [ 18 ]]. In accordance with what was found previously, the injury mechanisms included both receiving and delivering kicks for men and women alike [ 2 , 5 ]. The men, more than the women, tended to get injured as a result of receiving a kick [ 2 , 3 ]. It is suggested that the technique most likely implicated is the roundhouse kick [ 3 , 4 , 13 ]. Kicking the elbow typically leads to injury, especially if the kick is executed with the instep of the foot, such as when using the roundhouse kick to attack or counter-attack. Yet another reason to implement foot padding, as already mentioned above. In karate, punching is the main injury mechanism for both men and women [ 8 , 10 ], which may be related to the head and neck region being most frequently injured. More research on Canadian Taekwondo athletes of different age groups and skill levels is needed. Age in the present study is not believed to have played a role in the injuries sustained. The injury profile of the Canadians is quite similar to those found in other studies with Taekwondo athletes in their early twenties [ 2 , 3 , 5 ]. Children and juniors in Taekwondo were reported as incurring higher injury rates than adults [ 4 , 13 ]. Future studies should also include time lost due to injury. Joint dysfunction was identified as the second most common injury sustained by male athletes (13.7/1,000 A-E). Haldeman [ 24 ] defines joint dysfunction quoting Drum (1973) as,"Joint mechanics showing area disturbances of function without structural change; subtle joint dysfunctions affecting quality and range of joint motion. They are diagnosed with the aid of motion palpation, as well as stress and motion radiography investigation" [p. 623]. Greenman [ 25 ] states: "Joint dysfunction is characterized by findings of misalignment, relative fixation, loss of normal range-of-motion and end-play, tenderness, and tissue texture abnormality" [p. 13–14]. Although controversial, the term has been used widely in the literature, mostly by chiropractors, physical therapists and occasionally by biomechanists and medical doctors [ 24 - 33 ]. Further studies are required to validate the current finding. Conclusion The total injury rates for the Canadian Taekwondo athletes compare favourably to those reported in the literature, which is contrary to what was expected based on such a small sample size. It is hypothesized that recent rule changes may have contributed to these relatively low rates when compared to those found for other single tournaments, although more research is indicated before a definitive conclusion may be drawn. Interestingly, joint dysfunction was identified for the first time, which warrants more study. The injury data collection form should also include the technique used as a specification of the injury mechanism. General Taekwondo and competition-specific experience in addition to belt rank should also be recorded. Competing interest None declared. Authors' contributions MK collected the data, designed the Injury Report form, and wrote the initial draft of the manuscript. WP did the result part, revised and proof read the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC506779.xml
554995
Prevalence, types and demographic features of child labour among school children in Nigeria
Background To determine the prevalence, types and demographic features of child labour among school children in Nigeria. Methods A cross-sectional interview study of 1675 randomly selected public primary and secondary school pupils aged 5 to less than 18 years was conducted in the Sagamu Local Government Area of Ogun State, Nigeria from October 1998 to September 1999. Results The overall prevalence of child labour was 64.5%: 68.6% among primary and 50.3% among secondary school pupils. Major economic activities included street trading (43.6%), selling in kiosks and shops (25.4%) and farming (23.6%). No child was involved in bonded labour or prostitution. Girls were more often involved in labour activities than boys (66.8% versus 62.1%, p = 0.048): this difference was most obvious with street trading (p = 0.0004). Most of the children (82.2%) involved in labour activities did so on the instruction of one or both parents in order to contribute to family income. Children of parents with low socio-economic status or of poorly educated parents were significantly involved in labour activities (p = 0.01 and p = 0.001 respectively). Child labour was also significantly associated with increasing number of children in the family size (p = 0.002). A higher prevalence rate of child labour was observed among children living with parents and relations than among those living with unrelated guardians. Conclusion It is concluded that smaller family size, parental education and family economic enhancement would reduce the pressure on parents to engage their children in labour activities.
Background Child labour covers all economic activities carried out by children regardless of their occupational status [ 1 ]. It has probably being in existence almost as long as the history of mankind. Economic activity is a broad concept that encompasses most productive activities of children. It includes both work that is permissible under the International Labour Organization's (ILO) conventions and that which is not [ 2 ]. Child labour is prevalent worldwide, occurring both in developing and developed countries [ 3 , 4 ]. It is estimated that about 352 million children are engaged in some form of economic activity in the world [ 2 ]. Estimates of the ILO put the number of children fully at work in developing countries at 120 million and those working and schooling at 250 million [ 5 ]. Most of child labour takes place in Asia, the Pacifics and Africa [ 6 - 9 ]. In most parts of Africa the prevalence ranges from 20% to 54% [ 10 ]. The prevalence in Nigeria, Cote d'Ivoire and Zambia ranges from 20% to 30% [ 10 ]. Very few Nigerian studies [ 6 ] provide information on children who both school and work. Some studies [ 11 ] from other parts of the world actually excluded schooling children from the definition of child labour. This study aims to document the prevalence and types of child labour among primary and junior secondary school pupils in Nigeria. Methods This presentation is part of a broader study on child labour among school children in public primary and junior secondary schools in Sagamu Local Government Area of Ogun State, Nigeria. The study was conducted between October 1998 and September 1999 (inclusive). It was approved by the Ethical Committee of the Olabisi Onabanjo University Teaching Hospital Sagamu (OOUTH). Sagamu Local Government Area (SLGA) is one of the 15 Local Government Areas in the State. It is predominantly rural and semi-urban with a land space of 68.4 sq km. The estimated population for the year 1997 is 177,514. Administratively, SLGA is made up of 11 wards. The total school enrolment in the 50 primary and 16 junior secondary schools (JSS) for the year 1997 was 30,597, comprising 21,476 primary and 9,121 JSS pupils. Excluding 6,762 new entrants who did not have academic records for the preceding year left 23,835 pupils (17,891 primary and 5,944 junior secondary school pupils). Thus the ratio of primary to junior secondary school pupil population was about 3 to 1. This ratio was considered while selecting the subjects for the study. Sampling The survey sample was drawn randomly from all 50 public primary and 16 secondary schools in the LGA. Using ballot papers, one primary and one secondary school were randomly selected from each administrative ward. However, one ward had no primary school and four wards had no secondary school. Thus, a total of ten primary and seven secondary schools were selected. There were five classes in each primary school and two in each junior secondary after excluding primary 1 and JSS1 pupils because they were new entrants in their respective schools and therefore did not have academic records for the preceding year. Each class was made up of at least two arms. One arm of each eligible class in the selected schools was randomly selected, also using ballot papers. Altogether, 64 classes were selected: 50 classes from 10 primary schools and 14 classes from 7 secondary schools. Thirty pupils were randomly selected from each class to form the study school. The pupils were interviewed one at a time between the hours of 8.00 am and 2.00 pm on Monday through Friday using a structured, close-ended questionnaire. The interviews were conducted by author MBF with the help of previously trained research assistants. Relevant background information on age, sex, family background, involvement or otherwise in after-school economic activities were obtained. Information on parental education, occupation/income was obtained through requests accompanying letters to parents and guardians asking for their consent. The occupation/income and educational attainments of the parents were used to determine socio-economic index scores of the children using modified criteria described by Oyedeji [ 12 ]. Each parent was given two index scores: one for occupation/income and the other for educational attainment. The scores were rated on a scale of 1 to 5, from the more educated and more highly placed occupation/income groups to the least advantaged. The mean of the four scores for both parents was calculated and the value became the assigned socioeconomic group of the family/child. Where the calculated mean was not a whole number, the next higher integer was used. Child labour was defined as any type of economic task, paid, unpaid, or exploitative, engaged in by a child less than 18 years of age, which places the interests of the beneficiary well above those of the child and is detrimental to the physical mental, social, educational and moral development of the child [ 1 ]. For the purpose of analysis the study children were grouped into two; those involved in child labour and those who were not. Chi-square analysis was used to compare proportions and probability; (p) values less than 5% (0.05) were accepted as statistically significant. Results A total of 1675 day pupils were recruited into the study. The male/female ratio was 1:1.02. The majority (86.4%) of respondents were of Yoruba extract. Their ages ranged between 5 and less than 18 years. Those in primary schools were 1299 while 376 were in junior secondary schools as shown in Table 2 . One thousand and eighty pupils were involved in child labour giving an overall prevalence of 64.5%. The leading occupational activity was street trading (Figure 1 ). Three hundred and eighty seven girls were engaged in street trading compared to 308 boys (45.7% versus 37.2 %, X = 0.004). No child was involved in bonded labour or prostitution. Table 1 General characteristics of the subjects (n = 1675). Subjects No. of respondents (%) Sex distribution Male 828 (49.4) Female 847 (50.6) Tribe Yoruba 1448 (86.4) Igbo 111 (6.60) Hausa 35 (2.1) Others 81 (4.9) Age distribution 5 to 6 years 50 (3.0) 7 to 16 years 1594 (95.2) 17 years 31 (1.8) Level of education Primary 1299 (77.6) Secondary 376 (22.4) Table 2 Distribution of subjects involved in child labour by gender, education, family religion, custodian and number of children in the family. (n = 1675) Involved in child labour Not involved in child labour Total Characteristics no (%) no (%) χ 2 p-value Gender Male 514 (62.1) 314 (37.9) 828 (100.0) Female 566 (66.8) 281 (33.2) 847 (100.0) 3.91 0.048 Level of schooling Primary 891 (68.6) 409 (31.4) 1299 (100.0) Secondary 189 (50.3) 187 (49.7) 376 (100.0) 42.8 < 0.0001 Religion Christianity 631 (61.9) 388 (38.1) 1019 (100.0) Islam 445 (68.3) 207 (31.7) 652 (100.0) 6.94 0.008* Others 4 (100.0) 0 (0.0) 4 (100.0) Custodian Both parents 643 (65.8) 334 (34.2) 977 (100.0) Single father 73 (54.1) 62 (45.9) 135 (100.0) Single mother 196 (65.8) 102 (34.2) 298 (100.0) Other relatives 165 (64.7) 90 (35.3) 255 (100.0) Unrelated guardian 3 (30.0) 7 (70.0) 10 (100.0) 12.74 0.014 No. of children in the family 1 13 (46.4) 14 (53.6) 27 (100.0) 2 or 3 143 (58.6) 101 (41.4) 244 (100.0) 4 or 5 352 (62.7) 209 (37.3) 561 (100.0) 6 or 7 308 (68.6) 142 (31.4) 450 (100.0) 8 or 9 264 (67.2) 129 (32.8) 393 (100.0) 9.13 0.0025# *Comparison was limited to Christians and Moslems because there were very few practitioners of other religions. # Chi-square analysis for linear trend Figure 1 Pattern of child labour among study children. As shown in Figure 2 , child labour was most prevalent among children 9 to 14 years of age (10.2% to 13.0% of all the children involved in child labour). Table 2 shows the distribution of the children according to gender, educational level, family religion, custodian and number of children in the family. Girls, primary school pupils and Moslem children were all significantly involved in child labour than their counterparts (p = 0.048, 0.001 and 0.008 respectively). Comparison was limited to Christians and Moslems because there were very few practitioners of other religions. Figure 2 Age distribution of children involved in child labour. Over 80% of the study children lived with one or both parents, 15.2% with other relatives and very few (0.3%) with unrelated guardians. Children living with parents (both or single) and relatives had higher prevalence rates of child labour. The number of children living with unrelated guardians was relatively small but this group had less than 50% the prevalence of child labour of their counterparts living with relatives (χ 2 = 5.22, p <0.025). The prevalence of child labour did not vary significantly among children living with different types of relatives (p > 0.05). A significant rising trend of involvement in child labour with increasing number of children in the family (χ 2 for trend = 9.13, p = 0.0025) was observed. As many as 82.8% of working children were assigned economic tasks by one or both parents. Also 89.8% of the children performed these tasks in order to contribute to the family economy for feeding and for school fees. In Table 3 child labour decreased with increasing parental education. This trend was clearer with respect to maternal education. However, in Table 4 , child labour rates increased with decreasing parental socio-economic status. Table 3 Distribution of subjects involved in labour activities according to parental education. Father Mother Parental level of education no (%) Odds ratio no (%) Odds ratio University graduate or equivalent 30 (35.7) 1.00 14 (45.2) 1.00 Senior secondary school certificate/Teaching or other professional certificate 107 (54.6) 2.16 72 (46.2) 1.04 Junior secondary school certificate/Teacher's Grade II certificate/Equivalent 136 (68.7) 3.95 86 (54.8) 1.47 Junior secondary school education/Modern III certificate/Primary school leaving certificate 513 (68.5) 3.91 516 (68.3) 2.62 No formal education/Quranic school/Barely literate/dead parent 227 (67.2) 3.68 329 (68.1) 2.59 χ 2 for trend 29.33 32.86 p < 0.00001 < 0.00001 Table 4 Distribution of subjects according to parental socio-economic status. Socioeconomic class Involved in child labour n (%) Not involved in child labour n (%) Total respondents Odds ratio I 18 (32.7) 37(67.3) 55 1.00 II 78 (50.0) 78 (50.0) 156 2.06 III 403 (64.5) 222(35.5) 625 3.73 IV 442 (70.4) 186(29.6) 628 4.88 V 42 (70%) 18 (30%) 60 4.8 χ 2 for trend = 40.87, p < 0.00001 OR = Odds ratio I = very high II = high III = medium IV = low V = very low Discussion Child labour is a topical issue of global concern but statistics on the subject are often underestimated partly because of practical difficulties and also because of differences in the design and implementation of surveys [ 5 ]. The observed prevalence rate of 64.5% found in this study confirms the existence of child labour as an important secondary activity of school children in Sagamu Local Government Area of Nigeria. It is in consonance with high rates reported in Nigeria [ 10 ] and other developing countries [ 7 ]. The finding is also in keeping with an earlier report from Nigeria [ 6 ] and various parts of the world, that children both school and work [ 4 , 9 , 13 ]. Results of surveys by the ILO in four developing countries found that two-thirds of children combined school with work [ 4 , 15 ]. A rate of 51.9% was found in an earlier study carried out in Nigeria [ 16 ], while the rate was about 50 % in both Cairo and Bogotá [ 17 ]. The differences in prevalence rate may reflect differences in methodology and data collection. The present study focused on children within the compulsory age of schooling, which is also probably the peak age for child labour. Also, this study unlike previous local ones was school-based. The children were the primary respondents and interviews were conducted away from the influence or interference of parents. There was a higher representation of girls than boys among working children also in accordance with previous observations [ 7 , 12 ]. It is attractive to conjecture that the finding reflects gender bias in upbringing aimed at preparing the girls for traditional roles of small-scale economic trades to boost family economy. It may also be speculated that girls in the traditional African society are more amenable and responsive to parental control and as such are assigned economic tasks more often. The present study considered the role of children in economic tasks within the family enterprise and in and around the home. These activities usually involve more girls than boys in the traditional Nigerian setting. Moslem children were more often involved in child labour than their Christian counterparts. Other studies [ 19 , 20 ] had noted the influence of religious affiliation in the Moslem states of northern Nigeria and commented that wife seclusion (purdah) often puts a great economic burden on children. However, the practice of purdah is not a feature of southern Nigeria Islamism. Thus, the findings cannot be explained on that basis. It may however be speculated that the association of Moslem religion with child labour is related to family size. There was an increasing trend of engagement in child labour with increasing number of children within the family. This observation is consistent with other studies [ 21 ]. It may be, as has been suggested by other workers [ 22 ], that the desire for large family size is based on potential economic considerations. On the other hand, it may be validly argued, that a large family size, with the attendant spread of lean resources over many dependants, would force parents to engage their children in labour activities. Street trading was the dominant economic activity of working children in the study, agreeing with the findings from other urban and semi-urban settings in Nigeria [ 16 , 17 ]. However in Enugu [ 23 ], elsewhere in the country house servants constitute about two-thirds of the work force. Enugu is a major regional headquarters and is predominantly inhabited by the Igbo tribe in contrast to Sagamu LGA with its rural/semi-urban status and predominantly Yoruba population. The social and economic pressures differ and may account in part, for the variations in major economic tasks performed by children. It is comforting that, in contrast to reports from some other developing countries [ 3 , 18 , 24 ], no child in the present series was bonded or involved in prostitution. These practices are strange to the Ijebu Yoruba, buttressing the well-known influence of culture and tradition [ 10 ] on prevalence of child labour. This influence of culture, with respect to child rearing might also partly explain the high prevalence rate of child labour herein reported. The Ijebu Yoruba are known to be very industrious. They get their children introduced to family economic activities rather early. The stress of rural and semi-urban economies might also have put pressure on parents to engage their children in some form of economic activity in order to enhance family income [ 3 , 7 , 13 , 21 , 25 - 27 ]. Most working children were assigned their roles by their parents, largely to supplement family income. The prevalence of child labour did not differ significantly among children living with one or both parents or with a related guardian, Children living with single mothers were, not surprisingly, also more affected. Harsh economic realities, the burden of paying school fees and catering for a large family size may be the underlying reason parents compel their children to engage in work activities. On the other hand, children living with non-related guardians were surprisingly significantly less involved in labour activities. This finding however should be interpreted with a lot of caution because the number of children living with non-relatives was small. It is noteworthy however, to the extent that children living with their biological parents should theoretically be better cared for than those in foster homes. It is conceivable, that affected children may have been sent to live with more affluent guardians to give such children better prospects for the future. The prevalence of child labour increased with decreasing parental education and socio-economic class. These findings agree with other studies [ 7 , 8 , 13 , 21 , 28 ]. The reasons for these observations are fairly obvious. Educated parents are more likely to understand the need for a growing child to concentrate on activities capable of enhancing realization of their full potential and the need to avoid potentially 'harmful' activities. Also, better socio-economic standing removes the pressure on parents to put their children to work. Conclusion There is a need to protect the girl child, advocate reduction in family size, and promote parental education/economic empowerment in order to reduce the urge on children to perform economic roles. There is also a need to study all the ramifications of child labour for proper planning and for the protection of school children if Nigeria is to improve on her child survival, developmental and protection indices in the near future. Competing interests The author(s) declare that they have no competing interests. Authors' contributions MBF conceived of the study, participated in its design, collection, statistical analysis and interpretation of data. FON contributed to the statistical analysis, interpretation of the data and has been involved in drafting and revising the manuscript. AOO participated in coordination of the study and revision of the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554995.xml
547902
Towards an understanding of barriers to condom use in rural Benin using the Health Belief Model: A cross sectional survey
Background HIV/AIDS is the most dramatic epidemic of the century that has claimed over two decade more than 3 million deaths. Sub Saharan Africa is heavily affected and accounts for nearly 70% of all cases. Despite awareness campaigns, prevention measures and more recently promotion of anti viral regimens, the prevalence of cases and deaths is still rising and the prevalence of systematic condom use remains low, especially in rural areas. This study identifies barriers to condom use based on the Health Belief Model (HBM) in Benin, West Africa. Methods The study was a cross-sectional survey conducted from June to July 2002. Two hundred fifty one (251) individuals were interviewed using a structured questionnaire adapted from a standardized WHO/GAP questionnaire. A logistic regression was used to identify factors associated with condom use. Results In spite of satisfactory knowledge on HIV/AIDS transmission, participants are still at high risk of contracting the infection. Sixty three (63) percents of the interviewees reported being able to recognize infected people, and condom use during the last occasional intercourse was declared by only 36.8% of males and 47.5% of females. Based on the HBM, failure to use condom was related to its perceived lack of efficacy [OR = 9.76 (3.71–30.0)] and perceived quality [OR = 3.61 (1.31–9.91)]. Conclusions This study identifies perceived efficacy (incomplete protective effect) and perceived utilization-related problem (any reported problem using condoms) as the main barriers to condom use. Hence, preventions strategies based on increasing perceived risk, perceived severity or adequate knowledge about HIV/AIDS may not be sufficient to induce condom use. These data will be useful in designing and improving HIV/AIDS prevention outreach programs in Sub Saharan Africa.
Background One of the current challenging tasks faced by health professionals and scientists worldwide is the prevention and control of HIV/AIDS. This disease claims yearly a huge toll of deaths, productivity and economic losses, especially in sub-Saharan Africa where the population is already weakened by poverty, malaria and tuberculosis [ 1 , 2 ]. Curtailing the HIV/AIDS pandemic requires a holistic approach [ 3 ]. In Benin, several programmes have been developed to target high-risk groups and to modify cultural risk factors for the transmission of the infection [ 4 - 7 ]. Nonetheless, the prevalence of HIV infection and the rate of other sexually transmissible Infections (STI), and the number of people living with HIV/AIDS (PLWHA) are still increasing [ 8 , 9 ]. In Benin (and in most of West African countries), limited accessibility to anti retroviral medication (ARV) has pushed public health authorities to focus on prevention measures. Nowadays, most of the efforts are being shifted towards access to ARV, care to PLWHA [ 10 ], and second generation surveillance. However, there is a scarcity of operational research designed to identify barriers and facilitators to behavioural change [ 11 , 12 ]. This study, which was conducted in rural Benin, identifies factors deterring condom use that could be targeted by HIV outreach programmes using the Health Belief Model (HBM) framework. The Health Belief Model postulates that an individual's actions are based on beliefs. It underlines main factors for decision making such as perceived vulnerability or susceptibility, perceived severity of the outcome or conditions, perceived efficacy or benefit of control measure and the perceived barriers to prevention. It has been extensively used in behavioural sciences to predict behaviours and to design behavioural prevention programs. [ 13 - 15 ] Methods Study location and participants Benin is a West-African country with a population of 6.2 million people. The male/female sex ratio is 0.96, 48.5% of the population is under age 15, the rate of literacy is nearly 30.0 % and farming as the main occupation [ 2 ]. The study was conducted in Toffo, a county with a population of 80,000 inhabitants, located 50 miles north of Cotonou, the economic capital of Benin. Individuals aged 15 to 55 years old living in Toffo between June 10 and July 23, 2002 were invited to participate in the study. Health professionals, individuals less than 15 years old and school teachers were excluded from the study. Authorization to conduct the study was obtained from the Ministry of Health and the National STI/HIV/AIDS programme in Benin. Sample size, sampling and data collection The sample size was calculated based on results from the Benin 2001 Demographic and Health Survey [ 2 ]. Using the proportion of condom use reported in that survey (14.6%) and to detect a difference of 10% with a power of 90% and an error of 5%, a sample size of 245 people was needed [ 16 ]. A stratified random sample was applied using the 10 villages of Toffo as strata, and an average of 25 individuals per stratum was invited to participate after a brief presentation of the study, its aims and its potential applications. Participation was voluntary and participants could withdraw anytime during the interview. 270 individuals in the catchments area were invited and 251 accepted to participate in the study. Sixteen completed forms were removed from the analysis because of incomplete data on demographic characteristics and key responses related to the Health Belief Model. The final sample size was 235. Data were collected using a questionnaire adapted from the surveillance questionnaire developed by the Global AIDS Program of the World Health Organization (see survey items in Annexes). A pretest of the questionnaire was carried out on a convenient sample of 20 people of both genders living in Toffo and interviewers (3 social workers, one female and 2 males) were trained by SHH using simulated interviewees. Data analysis Data were stored in a file using Epi info 2000 [ 3 ] and then imported in SPSS 11.0 [ 17 ] and in SAS 8.1 [ 18 ]. The analysis included descriptive statistics, cross-tabulation and logistic regression. The following variables were included in the analysis: socio-demographic characteristics, perceived vulnerability (participant feeling at risk or not), perceived severity of the disease (AIDS perceived as deadly or not), perceived efficacy (condom effective to prevent infection or not), perceived barrier (any reported problem with use of condoms) and condom use during the last occasional sexual intercourse Results Socio-demographic characteristic The age distribution of the study population was similar to that of the Benin population [ 2 ]. Table 1 shows that there were 1.9 times more males than females sampled and the former were older (difference between mean age = 3.2 years with 95%CI = 1.8–4.5). A proportion of 69% of the participants declared being married (monogamous or polygamous) and 28% single. Farming was the most common reported occupation (37.4%) followed by laborers and small business. Sixty-six percent of males compared to 30% of females declared not having had any school education. The most common reported religion was Christianity (52 %) albeit polytheism being a social fact in Benin. Nearly two-thirds of the participants (63%) declared "Fon" as their ethnic group. There was no significant difference in socio-demographic variables by gender. Table 1 Socio-demographic characteristics of 235 participants in a HIV/AIDS interview, Toffo county, Benin (June–July 2002) Characteristics Males N (%) Females N (%) Total N (%) All participants 155 (66.0) 80 (34.0) 235 Age groups 15–24 56 (36.1) 37 (46.3) 93 (39.6) 25–34 52 (33.5) 29 (36.3) 81 (34.4) 35–44 30 (19.4) 5 (6.3) 35 (14.9) 45 + 17 (11.0) 9 (11.3) 26 (11.1) Marital status Married, monogamous 87 (56.1) 27 (33.8) 114 (48.5) Married, polygamous 23 (14.8) 24 (30.0) 47 (20.0) Single 44 (28.4) 21 (26.3) 65 (27.7) Widow, Separated 1 (0.7) 8 (0.9) 09 (07.3) Occupation Farming 81 (52.3) 7 (8.8) 88 (37.4) Working class 48 (31.0) 7 (8.8) 55 (23.4) Small business 10 (6.5) 33 (41.3) 43 (18.3) Housewife NA 21 (26.3) 21 (08.9) High school student 4 (1.7) 11 (4.7) 15 (06.4) Other 12 (7.7) 1 (1.3) 13 (05.6) Educational level None 47 (30.3) 48 (60.0) 95 (40.4) Some 108 (69.7) 32 (40.0) 140 (59.6) Ethnicity Fon 104 (67.1) 46 (57.5) 149 (65.1) Aizo 26 (16.8) 24 (30.0) 49 (21.4) Other 25 (16.1) 10 (12.5) 37 (13.5) Religion Christian 42 (52.5) 80 (51.6) 122 (51.9) Muslim 3 (3.8) 2 (1.3) 5 (2.1) Traditional 21 (26.3) 54 (34.8) 75 (31.9) Other 14 (17.5) 19 (12.3) 33 (14.0) Knowledge Table 2 presents a summary of the knowledge of the participants on HIV/AIDS and crude odds ratio estimates comparing males to females. There was a high awareness of AIDS (99, 9%), and its perceived risk (97% of participants considered AIDS as a deadly disease) among participants. Females were more aware than males of the modes of transmission of HIV infection (87 % of females versus 50% reported knowing at least 2 modes of transmission). There was a difference in preventive measures for HIV/AIDS by gender: Females reported mainly fidelity and abstinence whilst males primarily reported condom use. In addition, 84% of females whereas 52% of males reported being able to identify an HIV-infected person. This indicates the need for improving knowledge of the disease in general for both genders. Education level and religion did not have a meaningful effect on knowledge across age groups. Table 2 Distribution of knowledge by gender of 235 participants in a HIV/AIDS interview, Toffo county, Benin (June–July 2002). Knowledge N (%) Difference* (%) Male Female [95% CI] Have you ever heard of HIV/AIDS? Yes 154 (99.4) 80 (100) - 0.6 [-1.8; 6.2] What is your source of information? Radio (± other sources) 146 (94.2) 58 (72.5) 21.7 [11.3; 32.2] Health professionals (only) 4 (2.6) 8 (10.0) - 7.4 [-14.4; -0.4] Friends (only) 5 (3.2) 14 (17.5) -14.3 [-23.1; -5.5] According to your knowledge, what is AIDS? Deadly disease 134 (86.5) 79 (99.8) -13.3 [-18.7; -7.8] Projected disease 17 (11.0) 1 (1.2) 9.8 [4.3; 15.3] Don't know 3 (1.9) 0 (0.0) 1.9 [-2.5; 4.0] Other 1 (0.6) 0 (0.0) 0.6 [-0.6; 1.8] Modes of transmission of HIV? Knows at least two modes of transmission 77 (49.0) 70 (87.0) -38 [-48.8; -27.2] Knows sexual transmission 68 (43.9) 6 (7.5) 36.4 [26.7; 46.1] Do not know any 9 (6.5) 4 (5.0) 1.5 [-4.6; 7.6] Prevention What are the prevention methods of getting HIV? Abstinence /fidelity 13 (8.4) 44 (55.0) -46.6 [-58.3; -34.8] Serological test. 0 (0.0) 8 (10.0) -10 [-16.6; -3.4] Condom 131 (84.5) 20 (25.0) 59.5 [48.4; 70.6] Don't know 4 (2.6) 0 (0.0) 2.6 [0.0; 5.1] Other 7 (4.5) 8 (10.0) -5.5 [-12.8;1.8] How can a HIV-infected person be identified? Symptoms 80 (51.6) 67 (83.8) -32.2 [-43.5; -21] Can not differentiate 23 (14.8) 2 (2.4) 12.4 [5.9; 19] Don't know 49 (31.6) 11 (13.8) 17.8 [7.3; 28.3] Other 3 (2.0) 0 (0.0) 2 [-0.2; 4.2] Difference* = Difference between males and females in the prevalence of knowledge with 95 % confidence interval (using a binomial distribution) Behavioural risk factors The overall condom use in this population was low (34%). Table 3 describes the distribution of frequency of not ever using condom, last occasional intercourse without condom use and median number of sexual partners during the past 12 months, according to age, gender, education and marital status. Single participants with some education declared using condom more frequently. The proportion of subjects who declared using condom decreased with age and with males being marginally more likely to declare using it than females. We did not find any significant difference about reports on the use of condom during the last occasional intercourse by age groups, gender, educational or marital status. The small number of individuals declaring condom use during the last occasional intercourse maybe the reason why no significant associations were found. Table 3 Distribution of selected behavioural risk factors of 235 participants in a HIV/AIDS interview by age groups, gender, education level and marital status in Toffo county, Benin (June–July 2002) Behavioural risk factors Number of participants who declared not using condom N (%) Last occasional intercourse without using condom N (%) Median number of sexual partners during the last 12 months N (%) Age group 15–24 56 (60.2) 54 (58.1) 1.0 25–34 50 (61.7) 49 (60.5) 1.0 35–45 26 (74.3) 23 (65.7) 1.0 45 + 22 (84.6) 14 (53.8) 1.0 Gender Male 96 (61.9) 98 (63.2) 2.0 Female 58 (72.5) 42 (52.5) 1.0 DIFF [95% CI] Female/Male 10.6 (-2.0; 23.0) - 10.7 (-24.0; 3.0) NA Education None 73 (76.8) 54 (56.8) 1.0 Some 81 (51.9) 86 (61.4) 2.0 DIFF [95% CI] None/Some 24.9 (13.0; 36.8) - 4.6 (-17; 8.0) NA Marital status Single 36 (55.4) 34 (52.3) 2.0 Other 118 (69.4) 106 (62.4) 1.0 DIFF [95%CI] Other/Single 14 (.07; 28) 10.1 (-4; 24) NA DIFF [95% CI] = Difference in Proportion with 95 % confidence interval NA: Not Applicable However, it is interesting to note that whilst females declared ever using condom less often than males, they declared having used the condom during the last occasional intercourse more often. In particular, even though 73% of females declared never using the condom, 47% declared having used it during the last occasional intercourse. None of these differences were significant but they clearly indicated contradictory tendencies and answers to the questionnaire. It may be that among females who ever used condom, they use it more frequently than males. Finally, the median number of sexual partners during the last 12 months varied by gender (2 for males versus 1 for females), by education (2 for some versus 1 for none) and by marital status (2 for single versus 1 for other) but not by age groups. Determinants of condom use behavioural change Overall, there was a high perceived risk of contracting HIV infection among interviewees: 94% considered themselves as vulnerable to HIV/AIDS. This proportion was higher in females compared to males. Similarly there was a high perceived severity of HIV/AIDS: 99% of females compared to 87% of males perceived HIV/AIDS as a severe and deadly disease. Conversely, there was a relatively low perceived efficacy of condom as a protective measure: only 37% of the interviewees perceived condom as an effective mean in protecting from getting HIV infection. We identified several socio-cultural barriers to behavioural change namely reported problems using condom (88% of the interviewees), the alleged capability to physically recognize an HIV infected person and the denial all together of the disease (only 19% participants believe HIV/AIDS exists). Also, cultural practices such as polygamy (20% of the study population), poverty, the belief that there is a cure for the disease (74%) and religion (9 % of non favorable reaction towards condom are among declared Christians) were all not favorable to HIV infection control. Logistic regression using the theoretical health belief model Table 4 describes the results of a logistic regression, fitted to assess the strength of association between perceived vulnerability (participant feeling at risk or not), perceived severity of the disease (AIDS perceived as deadly or not), perceived efficacy (condom effective to prevent infection or not), perceived barrier (problems with using condoms) and the lack of condom use. Perceiving condom as ineffective (OR = 9.8, 95%CI = 3.2–30.0) and having reported problems with using the condom (OR = 3.6, 95%CI = 1.3–9.9) were both associated with the lack of use of condom. However, perceiving oneself as vulnerable to HIV infection (OR = 6.9, 95%CI = 0.9 – 52.5) was not strictly statistically significant since most interviewees felt vulnerable, reducing the power of detecting a significant difference. This variable was also a weak confounder for the effect of perceiving the condom as ineffective. Not perceiving HIV/AIDS as a deadly disease (OR = 2.5, 95%CI = 0.3 – 19.7) was not associated with the lack of use of condoms. Table 4 Crude* and adjusted** odds ratio (OR) estimates with their 95% Confidence Interval (95% CI) of the effect of perceived efficacy of condom, barriers to condom use, vulnerability and severity on the lack of condom use for 235 participants in a HIV/AIDS interview, Toffo county, Benin (June–July 2002). Variables Crude OR 1 (95% CI) Adjusted OR 2 (95% CI) No perceived risk to HIV infection 6.9 (0.9 – 52.5) NA 3 AIDS not perceived as a deadly disease 2.5 (0.3 – 19.7) NA 3 Perceived incomplete protection using condoms 11.5 (3.8 – 34.7) 9.8 (3.2 – 30.0) Reporting any problem using condom 5.4 (2.1 – 13.7) 3.6 (1.3 – 9.9) 2 LL Chi-square: 32.2 P-value < 0.0001 Score Chi-square: 32.2 P-value < 0.0001 C-statistic = 0.81 1 Crude OR: calculated from a univariate logistic regression 2 Adjusted OR: Calculated from a multivariate logistic regression including Perceived incomplete rotection using condoms and reporting any problem using condom. 3 NA: Not Applicable because these variables did not have an important effect on the use of condom. Discussion This study was the first ever to use the Health Belief Model (HBM) to assess cultural behaviour in rural Benin towards condom use and HIV/AIDS. The HBM was reported to be one of the most widely used behavioural frameworks for more than five decades but has been criticized for its inability to efficiently predict people's behaviour [ 14 ]. There is general agreement that the components of HBM should include self-efficacy and cues to action, and that susceptibility and severity should be conditional on action or inaction [ 13 , 14 ]. The lack of generally accepted model construct also makes comparisons difficult across studies. An effort was made in this study to address these concerns by clearly defining the model's construction. Our results showed there is a high awareness on AIDS in general and that women knew more about the modes of transmission of HIV/AIDS and its impacts than men. Conversely, women were more likely to feel that they could identify HIV-infected individuals from their symptoms. In addition, females were less likely to declare using condom in general even though a higher proportion declared having used condom during the last occasional sexual intercourse. This finding is disturbing and could be explained by the difference in perception of the question "do you use condoms?" It is difficult to judge what the true answer is but it is likely that rare events are better reported, and thus women may be more prone to recall the use of condom than men during occasional intercourse given that they declared on average fewer sexual partners. It is also possible that among women who do use condom, they will use it more regularly than men. Our measure of perceived vulnerability might not be sensitive enough to capture differences in perceived risks. In fact, all women and most men felt they were at risk of acquiring the infection, yet only a small proportion were using condoms. Another explanation may be that perceived risk is not a driving force in behavioural change in this subset of the population. This is an illustration of the complexity of modeling human behaviour and can thus make a case for further cultural-specific HIV-behavioural research. When only considering the percentage of condom use by gender, females appear to be at a higher risk of acquiring HIV even though they appeared to know more about transmission routes and prevention methods. This might be due to the well established difficulty facing women in negotiating the terms of sexual intercourse. In fact, gender inequality is associated with poverty, condom with distrust and sexual economic exchange is not perceived as prostitution [ 19 ]. All these factors make women vulnerable to acquire HIV infection, and therefore it is important to consider empowerment of women, gender inequality and poverty as key strategies of HIV/AIDS prevention programmes. Despite a relatively acceptable knowledge of modes of transmission and prevention methods, only a few of participants declared using condoms, which is an indication that a relatively good knowledge about HIV/AIDS, even though necessary, may not be a key factor in behavioural change in fighting HIV epidemic in the study population. These findings also indicate that programmes which aim only at increasing awareness and knowledge may not succeed. Using the HBM to analyze the determinants of behavioural change in our study population, we can conclude that there is a high-perceived vulnerability and perceived severity, and yet this does not encourage condom use. An important proportion of participants do not believe in the efficacy of condoms and there are barriers to the use of condoms. Our results are comparable to that found in a similar study in the USA [ 20 ] and in a review of published studies using HBM [ 14 ] where perceived barriers were found to be the single most powerful predictors of the HBM. Our findings are also consistent with results of studies conducted in Kenya [ 21 ] and in Ghana [ 22 ], in which perceived barriers were found as being the strongest predictors of condom use. However, these results can not be generalized across settings For example, in a study conducted among American university students, the HBM did not significantly explain condom use but rather condom use was associated with sexual practices [ 23 ]. Perceived benefit of avoidance of pregnancy was found as one of the strongest predictors of consistent condom use in New York female adolescents [ 24 ] and in Zimbabwe social support was found to be the most consistent factor associated with sexual risk reduction [ 25 ]. These observed differences in the strongest(s) predictor(s) of the HBM can be noticed through several other works [ 26 - 30 ]. Hence, it appears important to conduct operational behavioural researches in each local setting to identify factors that influence condom use. One limitation of our study was that for ethical reasons, subjects less than 15 years old were excluded even though some may have already been sexually active. Also, there was a potential selection bias by not having equal number of interviewers by gender, which resulted in an over-sampling of males. Our results would be biased if the reason for poor recruitment of women was linked to their behaviours, which is not likely to be the case. There were three males interviewers for one female (difficulties in recruiting female educated social worker in the area), and interviewer/participants must be from same gender. For the purposes of the analysis we assumed that reported knowledge and behavioural risk factors are independent. Finally there is no evidence for the validity or reliability for the original WHO questionnaire, however its use allows for comparability of results across settings. Conclusions Condom use, in our study population, depends on its perceived quality and perceived efficacy. There is an indication that behavioral communication change strategies based on increasing perceived risk or vulnerability of the population or based on fear factor by increasing perceived severity of HIV/AIDS are less likely to be deterrent towards condom use and require more researches. HIV outreach programs must target more barriers of condoms use. Condom outreach programmes should be defined at community level and must be defined in association with the community, using problem-solving techniques and selecting the most relevant targets, based on their importance and changeability [ 13 ]. Data from this study could be useful for the design and planning of health intervention programmes, resource allocation and evaluation of condom outreach activities in Benin. Competing interests The author(s) declare that they have no competing interests. Authors' contributions SHH conceived of the study, designed the protocol, carried out and supervised the field work and data collection, performed and interpreted the statistical analysis and wrote the manuscript. HC participated in analysis and interpretation of the data, and in the writing of the manuscript. NJH contributed in earlier analysis of the data and reviewed the manuscript. All authors read and approved the final manuscript. Table 5 Survey items, HIV/AIDS and condom use survey, Toffo county, Benin (June – July 2002). SURVEY ITEMS RESPONSE CATEGORIES General knowledge on HIV/AIDS 1. Have you ever heard about HIV/AIDS Yes/No 2. In your knowledge how severe is HIV/AIDS Deadly, don't know, imaginary, other 3. How could someone get infected by HIV? At least 2, one or no correct answer(s), 4. Who you think are at risk of getting HIV? Everyone/specific groups/Don't know 5. In your knowledge, what are the prevention methods of getting HIV? Abstinence, Fidelity, Condom, Other, Don't know 6. How could you recognize a HIV-infected person? Could not, Cachexia, Other symptoms, Don't know Beliefs on HIV/AIDS 7. Do you believe HIV really exists? Yes/No/Don't know 8. Do you think you are at risk of getting HIV? Yes/No/Don't know 9. If no to 8) why? Fidelity/Condom use/Other/Don't know 10. Where you believe HIV originates from? God/Bewitchment/Other/Don't know 11. Do you think one can completely cure from HIV/AIDS? Yes/No/Don't know 12. If yes to 11) how? Medicine/Herbs/Prayers/Other/Don't know 13. How would you rate the protective effect provided by condoms? Complete/Incomplete/Useless/Don't know 14. Does your religion believe HIV exists? Yes/No/Don't know 15. What is the position of your religion towards condom use? Favorable/Unfavorable/Indifferent/Don't know Behaviors and attitudes 16. Would you mind taking a HIV screening test if you were asked? Yes / No / Don't know 17. Do you use condoms? Yes / No / No answer 18. If No or No answer to 17) why? Don't like / Only God save / Other 19. How often do you use condoms? Always / Sometimes / Never 20. Did you use condom during the last occasional intercourse? Yes / No 21. Do you encounter any problem using condoms? Yes / No 22. If yes to 21) what type of problems Less lubrificated / Less pleasure / break easily / Other 23. Numbers of sexual partners during the last 12 months Socio demographic characteristics 24. Age Full years 25. Sex Male / Female 26. Marital status Married monogamous / Married polygamous/ Single / Divorced / Widowed / Separated 27. Education (Ability to read) Fluent / With difficulty / Not able to read at all 28. Occupation Student / Farmer / 29. Religion Christianity / Islam / Animist / Other 30. Place (please give the name of your village) Village name Adapted from the World Health Organization / Global AIDS Program's questionnaire Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC547902.xml
522803
Exposure to malaria affects the regression of hepatosplenomegaly after treatment for Schistosoma mansoni infection in Kenyan children
Background Schistosoma mansoni and malaria infections are often endemic in the same communities in sub-Saharan Africa, and both have pathological effects on the liver and the spleen. Hepatosplenomegaly associated with S. mansoni is exacerbated in children with relatively high exposure to malaria. Treatment with praziquantel reduces the degree of hepatosplenomegaly, but the condition does not completely resolve in some cases. The present analysis focused on the possibility that exposure to malaria infection may have limited the resolution of hepatosplenomegaly in a cohort of Kenyan schoolchildren. Methods Ninety-six children aged 6–16, from one community in Makueni district, Kenya, were treated with praziquantel. At baseline, all children had hepatomegaly and most had splenomegaly. The source of S. mansoni infection, a river, was molluscicided regularly over the following three years to limit S. mansoni re-infection, whereas malaria exposure was uninterrupted. Hepatic and splenic enlargement was assessed annually outside the malaria transmission season. Results Children living in an area of relatively high exposure to both infections presented with the largest spleens before treatment and at each follow-up. Spleens of firm consistency were associated with proximity to the river. The regression of hepatomegaly was also affected by location, being minimal in an area with relatively low S. mansoni exposure but high exposure to malaria, and maximal in an area with relatively low exposure to both infections. Conclusions The outcome of treating cases of hepatosplenomegaly with praziquantel in this cohort of Kenyan children depended strongly on their level of exposure to malaria infection. Furthermore, a residual burden of hepatosplenic morbidity was observed, which was possibly attributable to the level of exposure to malaria. The results suggest that exposure to malaria infection may be a significant factor affecting the outcome of praziquantel treatment to reduce the level of hepatosplenic morbidity.
Background Hepatosplenomegaly is a widespread but neglected condition affecting many communities in sub-Saharan Africa where both Schistosoma mansoni and malaria infections are endemic. Severely enlarged and hard organs may be an indicator of increased portal pressure, as well as pointing to an increased risk of portal hypertension and its sequelae, including oesophageal varices and haematemesis [ 1 ]. Historically, hepatomegaly in sub-Saharan Africa has been associated with S. mansoni infection, and on a global scale up to 8.5 million individuals may be affected [ 2 ], with the prevalence of organomegaly generally highest in children [ 3 ]. Enlargement of spleens has often been associated with the level of malaria transmission [ 4 , 5 ]. One prevailing suggestion has been that hepatosplenomegaly amongst younger children is attributable to malaria infection, whereas older children have acquired some level of immunity to malaria and hence if they do have hepatosplenomegaly it is likely to be attributable to schistosomiasis [ 6 ]. However, chronic splenomegaly associated with portal hypertension has been reported in Kenyan hospitals amongst residents from areas where both S. mansoni and malaria infections are endemic [ 7 , 8 ], and it has long been suggested that the presence of both infections may confound attempts to quantify the impact of either one [ 9 , 10 ]. Recently, it was observed that IgG3 antibody responses to P. falciparum schizont antigen (Pfs Ag) were higher in Kenyan children with S. mansoni infection and hepatosplenomegaly compared with children with infection but no hepatosplenomegaly [ 11 ], further implicating exposure to malaria in childhood as a risk factor for severe hepatosplenic morbidity. Given that malaria and schistosome infections are often endemic in the same communities, and are both targets of large-scale but independent, intervention programmes, it is important to understand how continued exposure to one parasite may affect the outcome of intervention against the other. The current longitudinal study evaluated praziquantel treatment on the regression of hepatosplenomegaly amongst children living in an area with endemic S. mansoni infection and seasonal malaria infection [ 12 , 13 ]. Before treatment, there was a strong correlation between S. mansoni egg count and the degree of both hepatomegaly and splenomegaly [ 12 ]. However, spatial analysis then revealed heterogeneity in exposure to both infections, and in particular there was an exacerbation of splenomegaly amongst children living in an area of relatively high exposure to both malaria and S. mansoni infections [ 14 ]. The prevalence of hepatomegaly and splenomegaly decreased slowly during the three-year follow-up period, but had not completely abated by the end of the study [ 13 ]. Here, we report that the level of exposure to malaria infection was likely to have been an important factor in limiting the resolution of hepatosplenomegaly in these children. Methods Study area Mbeetwani is situated in the district of Makueni, half way between Nairobi and Mombasa and approximately 10 km east of the main highway. The ethnic background of the population is predominantly Akamba. During the dry seasons, the river Kambu is the main source of water. Running from west to east, the river lies to the north of the community (Figure 1 ). Surface water is present at the western end during the dry season, whereas the eastern end is dry and water is retrieved by digging wells into the riverbed. River water is drawn for drinking, cooking and other domestic purposes. Washing of clothes and bathing takes place at numerous sites, and animals are taken to the river to drink. Clinical examination The cohort and examination procedure are described in detail elsewhere [ 12 ]. Briefly, 96 children aged between 6 and 16 were selected for ultrasound detectable hepatomegaly after a community survey. They were examined before treatment with praziquantel at 40 mg/kg, and then examined again in the same month in the following three years. Treatment was given again two years into the study. All surveys took place in October of each year, towards the end of the dry season. A total of 67 children attended all examinations, whereas 71 children attended at least the baseline and third year follow-up surveys. Clinical examination consisted of liver and spleen palpation whilst children were in a supine position. Un-palpable spleens and livers were recorded as such, otherwise extensions in cm below the rib cage along the mid-clavicular line (MCL) and mid sternal line (MSL) were measured for the liver, and extension below the rib cage along the mid-clavicular and mid axillary lines (MAL) was recorded for the spleen. The firmness of each palpable organ was recorded as either soft, firm, or hard, as described previously [ 12 ]. An accompanying ultrasound examination was conducted according to World Health Organisation guidelines [ 15 ] to test for the presence of periportal fibrosis. Both the clinicians and the ultrasonographers were blinded as to the location of children's domiciles and previous examination results, and the same team performed the examinations throughout the study period. Mapping Details of the mapping procedure are given elsewhere [ 14 ]. Briefly, the longitude and latitude of each house were recorded using a Magellan GPS 315 receiver. The course of the river Kambu was determined by taking readings approximately every 100 metres along a ten kilometre stretch that extended beyond the borders of the study area. The approximately rectangular study area was divided into 4 sectors with similar numbers of houses by bisecting the north-south and east-west transects at 4 km east of the western boundary and 1 km south of the river. Assays Since the surveys were undertaken during the dry season, parasitaemia was not expected to provide a good enough estimate of exposure. Therefore, anti-Pfs IgG3 responses were used as a proxy measure of recent exposure [ 16 ], measured by ELISA as described previously [ 17 ]. Parasitology For the quantification of S. mansoni eggs, each child provided 3 stool samples before the baseline survey in October 1999. Two 50 mg Kato slides were prepared from each stool sample using the Kato-Katz procedure [ 18 ]. All slides were prepared and examined by the same team. Finger-prick blood samples were taken on the same day that clinical and ultrasound examinations were performed, for detection of malaria parasitaemia. Mollusciciding Regular mollusciciding of the river Kambu was undertaken with Bayluscide ® (Bayer CropScience AG, Alfred-Nobel-Str 50, D40789 Monheim am Rhein, Germany). Area-wide application and focal spraying using backpacks was undertaken twice yearly to coincide with periods when snail populations were reduced due to environmental conditions. Snail sampling activities at 9 sites on a monthly basis were undertaken to confirm that snail numbers were kept at very low levels (HC Kariuki, unpublished observations). Treatment and ethical considerations The objectives of the study were carefully explained to the local community including the parents and teachers before the start of any activities. Informed assent was obtained from the parents or guardians of the children. After the clinical and ultrasound examination in October 1999, all children were treated with a single dose of praziquantel (Distocide ® , Shin Poon Pharmaceuticals, Seoul, Republic of Korea), at 40 mg/kg body weight. The children were again treated in 2001 and 2002 under the same regimen. No other source of treatment was available locally during this period. Tablets were administered together with a piece of bread and a soft drink in order to minimise gastrointestinal side effects. The study was approved by the Kenya Medical Research Institute Ethical Review Committee. Analysis Baseline data for each parameter were compared with 3 year follow up data amongst 71 children who attended the first and last surveys to estimate the number of children within each sector who presented with a lower degree of organomegaly at the end of the study. Repeated-measures ANOVA of each organ enlargement parameter was also undertaken for the 67 children who attended all surveys, with age, sex and sector of residence as fixed factors. Since ANOVA does not indicate the direction of any change, a post-hoc analysis was undertaken within each sector to determine whether there was an overall increase or decrease in the extent of organomegaly. This was achieved by comparing baseline enlargement and enlargement after 3 years follow up for each parameter, using the Wilcoxon Signed Ranks test. All analyses were conducted in SPSS v11 (SPSS inc, 2001, Chicago). Results Geographical distribution of exposure Figure 1 illustrates the geographical distribution of houses within the study area, as well as the course of the river Kambu. A total of 70 houses were geo-referenced. Also given are the median and 25–75%ile range of egg counts and anti-Pfs IgG3 responses within each sector before the intervention. Details of the analysis of S. mansoni egg counts and IgG3 responses within each sector are given elsewhere [ 14 ]. Briefly, there was a significant clustering of high egg counts in children living at the western end of the study area (Sector A and Sector C). Significant clustering of the highest anti-Pfs IgG3 responses was observed amongst children living in the two northern sectors (Sector A and Sector B). Thus, sector A was an area of relatively high exposure to both infections, sector B was an area of relatively high exposure to malaria only, sector C was an area of relatively high exposure to S. mansoni only, and sector D was an area of low exposure to both infections. Temporal changes in organ consistency by sector Significant variation in the prevalence of hard spleens over time was observed within each sector (Table 1 ), attributable to an overall decline in the number of children with hard and enlarged spleens (Figure 2a ). The prevalence of hardened spleens was highest at all time points in the sector with relatively high anti-Pfs IgG3 responses and high egg counts (Sector A). By two years post-treatment, no child resident more than one kilometre from the river presented with a hard spleen. The prevalence of firm livers in all sectors varied over time, except in sector B (Table 1 ). In all other sectors there was a decrease in the prevalence of firm livers over the course of the follow-up period (Figure 2b ). At the end of the study, the lowest prevalence of firm livers was observed in children from sector D. Temporal changes in organomegaly by sector Significant variation in the degree of MAL splenomegaly was observed in each sector over the three-year follow-up (F = 20.7, p < 0.001). Figure 3 illustrates a decline in the degree of MAL splenomegaly within each sector, and this was confirmed by post-hoc Wilcoxon Rank analysis (Table 2 ). The rate of regression was affected independently neither by sector, age nor sex, and children from sector A presented with the largest spleens at each time point. At the third year post treatment follow-up, the fraction of each group presenting with no MAL splenomegaly in each sector was 10/24 (Sector A), 6/14 (Sector B), 13/15 (Sector C), and 16/20 (Sector D). MCL splenomegaly also varied significantly over time (Figure 3b , F = 14.1, p < 0.001), but the extent of variation was affected by sector (F = 2.191, p = 0.024). Post-hoc Wilcoxon rank analysis revealed that there was a significant decrease in the extent of MCL splenomegaly only in sector D (Table 2 ). MSL liver enlargement varied significantly over time (F = 3.78, p = 0.012), with the variation being affected by sector (F = 2.57, p = 0.008). There was no consistent decline in MSL hepatomegaly over time within any one sector (Figure 3c ). Although a fraction of the children in each sector presented with reduced hepatomegaly at the third year follow-up, Wilcoxon Rank analysis of the direction of change within each sector confirmed that an overall significant decrease in the extent of MSL hepatomegaly between the baseline survey and the third year follow-up occurred only amongst children from sector D (Table 2 ). Children from sector B presented with the least improvement in MSL hepatomegaly, with 10/12 exhibiting no change or an increase in this parameter when baseline data were compared with data from the third year follow-up. MCL hepatomegaly values could not be used in the repeated measures analysis due to a lack of normal variation in the data. Wilcoxon rank analysis within sectors revealed no significant change in this measurement between baseline and the third year follow-up in any sector (Table 2 ). Ultrasound results There was no evidence of periportal fibrosis in any of the children at any time point. Discussion Many clinical surveys of sub-Saharan communities have reported that organomegaly of the spleen or liver is a very common condition, but attribution to a specific aetiology has always been problematic. Distinguishing the contribution of malaria from that of schistosome infections is particularly complex since they are often endemic in the same communities. However, it is well established that exposure to both S. mansoni and malaria varies over small areas [ 19 - 21 ], and the potential therefore exists to compare morbidity in areas of overlapping exposure with areas where infections of one species are more prevalent. Here, we have exploited observations concerning micro-geographical variation in the distribution of each infection to demonstrate how the benefits of treatment with praziquantel on hepatosplenomegaly may be affected by local heterogeneity in malaria exposure. Our results further implicate both parasites as aetiological agents of chronic hepatosplenomegaly including firmness of the organs in school-aged children, and have strong implications not only for estimating the burden of S. mansoni and malaria, but also for estimating the outcome of interventions. The generation of these results was facilitated by several important features of the study design. The intermittent mollusciciding of the river Kambu was particularly important, since it reduced the potential for morbidity to rebound as a result of re-infection, as has been seen elsewhere in studies of both S. mansoni and S. haematobium [ 22 ]. The timing of the follow-up surveys outside the malaria transmission season was another important component, since it allowed an assessment of the clinical situation without the confounding effects of transient morbidity associated with acute malaria. Without these particular features, it is unlikely that we could have come to any conclusions about the effects of chronic exposure to malaria infection on the regression of hepatosplenomegaly. Another important component of the study was the mapping of both house co-ordinates and the local river. By doing so, we were able to identify spatial clustering of relatively high S. mansoni egg counts at the western end of the study area. Such clustering is likely related to variation in the amount of surface water along the river, and hence variation in transmission potential due to the strong relationship between snail abundance and water level observed during long term studies of transmission in this area [ 23 ]. The clustering of anti-Pfs IgG3 responses to schizont antigen along a tract parallel to the course of the river suggests a sharp decrease in transmission further from the river [ 14 ]. Although we observed several significant changes following praziquantel treatment of the cohort, it is important to note that this was a retrospective analysis, which carries a few limitations. The observations made in this study were based on a small, case-only cohort, and therefore do not represent the outcome of an intervention programme involving mass treatment of a population. As this was a retrospective analysis, no control was possible concerning treatment for malaria during the follow-up period. The population had access to antimalarial drugs at local shops, but so far as we are aware there was no systematic intervention against malaria infection during the follow-up period. The results therefore encompass the effects of background treatment for, as well as exposure to, malaria infections. Reliability of the clinical measurements is an important factor, especially given the size of the cohort. We have assessed this procedure elsewhere and it has been found to be satisfactory (unpublished observations). The involvement of three or four clinicians at each examination also reduced the degree of imprecision in the organomegaly measurements, and previous analysis has demonstrated that interpretable changes in the measurements occurred within this cohort [ 13 ]. The major result from this analysis is that the outcome of treatment with praziquantel, combined with very limited S. mansoni re-infection, depended strongly on where members of the cohort were resident. Previously, we demonstrated that children within the same cohort that had relatively high-level exposure to both S. mansoni and malaria presented with significantly larger spleens before treatment than children highly exposed to either parasite alone [ 14 ]. This contrasted with an observation on splenomegaly in school children exposed to malaria only in Ghana, where splenomegaly was observed to decline with distance from mosquito breeding sites [ 5 ]. Here, we report that children with relatively large spleens at baseline and living in the area with high levels of exposure to malaria still had the largest spleens 3 years after treatment, despite an overall reduction in their organomegaly. Hardness of enlarged spleens at baseline was most commonly observed amongst children living close to the river, irrespective of their level of pre-treatment intensity of S. mansoni infection [ 14 ]. In the absence of follow-up data, the inference may have been that the hard consistency of enlarged spleens was primarily attributable to chronic malaria exposure. However, during the follow-up period, we observed a gradual and monotonic decrease in the prevalence of hard spleens that suggests the removal of S. mansoni infection was the trigger for the improvement [ 13 ]. One possible explanation is that the hardening of the spleens was due to a synergistic effect of co-infection. Hyperplasia and congestion in the spleen associated with relatively high levels of exposure to malaria may have been sufficient to cause chronic enlargement of the spleen. Hardening and further enlargement may then have occurred as a secondary effect of congestion in the liver attributable to schistosome infection, leading to increased portal pressure and dilation of the splenic vasculature. Three years after treatment with praziquantel, and with very limited re-infection by S. mansoni [ 12 ], the prevalence of palpable spleens had diminished considerably. At all time points, the prevalence of hard spleens and the degree of splenomegaly were highest amongst children living in the sector with highest egg counts before intervention and highest anti-Pfs IgG3 responses. By removing S. mansoni infection and limiting re-infection, we may have abrogated any synergistic effects of co-infection on the spleen, and thereby uncovered the residual burden of chronic exposure to malaria. Importantly, the rate of regression of MAL splenomegaly was not affected significantly by location, which suggests that the effects of praziquantel treatment on splenomegaly attributable to S. mansoni infection are not dependent on the level of exposure to either malaria or S. mansoni . The effects of praziquantel treatment on hepatomegaly in this cohort were more subtle. There was gradual decline in the prevalence of hepatomegaly after praziquantel treatment [ 13 ] indicating that the removal of S. mansoni infection was a critical factor for improvement. However, upon analysing hepatomegaly data across sectors, it emerged that children from the sector with relatively high anti-Pfs IgG3 responses, but relatively low S. mansoni egg counts, had the lowest rate of regression. A possible explanation of this observation is that an aetiological agent other than S. mansoni was responsible for the observed hepatomegaly. A likely candidate is malaria infection. Enlargement of the liver in acute malaria infection is temporary and recedes rapidly after treatment [ 24 ]; however, studies of young children in Gambia have shown that repeated infection with malaria, perhaps when combined with other unidentified factors, can lead to the development of chronically enlarged livers [ 25 , 26 ]. Our observations suggest that even if school-aged children are examined outside the malaria transmission season, they may still be affected by hepatomegaly attributable to malariainfection. Children from the sector with relatively low exposure to both infections experienced a significant decrease in liver enlargement along the mid sternal line when baseline data were compared with data from the third year follow-up. It is possible that hepatomegaly in these children was attributable to neither S. mansoni nor malaria infection. Alternatively, because they were relatively lightly infected with S. mansoni , and experienced relatively low exposure to malaria infection, it is possible that they were more likely to regress in terms of hepatomegaly within the follow-up period. Conclusions In conclusion, our observations lend further support to the hypothesis that severity of hepatosplenomegaly in Kenyan school-aged children is related to their degree of exposure to both S. mansoni and malaria infections. Specifically, although the degree of hepatomegaly or splenomegaly may be correlated with S. mansoni infection, there is likely to be further exacerbation of the condition if a child is concurrently exposed to malaria. We have now observed that the apparent benefits of treating a case of hepatosplenomegaly with praziquantel may be reduced if a child lives in an area of relatively high exposure. We have therefore confirmed the long-standing, but untested, hypothesis that co-infections of S. mansoni and malaria may obscure the clinical evaluation associated with infection by either species. Our results also introduce the necessity of considering the level of exposure to malaria when evaluating the clinical outcome of praziquantel treatment for S. mansoni infection. Competing interests The author(s) declare that they have no competing interests. Authors' contributions MB conceived of, and conducted, the analysis, and drafted the manuscript. BJV and AEB performed clinical examinations and participated in the design of the study. CA and AO conducted clinical examinations. CHK, GK, JM, JHO and EM participated in the planning and execution of field activities. DWD participated in the design of the study and in fieldwork. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522803.xml
552328
Simian immunodeficiency virus (SIV) envelope quasispecies transmission and evolution in infant rhesus macaques after oral challenge with uncloned SIVmac251: increased diversity is associated with neutralizing antibodies and improved survival in previously immunized animals
Background Oral infection of infant macaques with simian immunodeficiency virus (SIV) is a useful animal model to test interventions to reduce postnatal HIV transmission via breast-feeding. We previously demonstrated that immunization of infant rhesus macaques with either modified vaccinia virus Ankara (MVA) expressing SIV Gag, Pol and Env, or live-attenuated SIVmac1A11 resulted in lower viremia and longer survival compared to unimmunized controls after oral challenge with virulent SIVmac251 (Van Rompay et al. , J. Virology 77:179–190, 2003). Here we evaluate the impact of these vaccines on oral transmission and evolution of SIV envelope variants. Results Limiting dilution analysis of SIV RNA followed by heteroduplex mobility assays of the V1–V2 envelope ( env ) region revealed two major env variants in the uncloned SIVmac251 inoculum. Plasma sampled from all infants 1 week after challenge contained heterogeneous SIV env populations including one or both of the most common env variants in the virus inoculum; no consistent differences in patterns of env variants were found between vaccinated and unvaccinated infants. However, SIV env variant populations diverged in most vaccinated monkeys 3 to 5 months after challenge, in association with the development of neutralizing antibodies. Conclusions These patterns of viral envelope diversity, immune responses and disease course in SIV-infected infant macaques are similar to observations in HIV-infected children, and underscore the relevance of this pediatric animal model. The results also support the concept that neonatal immunization with HIV vaccines might modulate disease progression in infants infected with HIV by breast-feeding.
Background The continued need for breast-feeding in developing countries due to nutritional or socio-economic reasons poses a considerable risk for postnatal mother-to-child transmission of HIV, and breastfeeding is estimated to account for 33–50% of infant HIV infections worldwide [ 1 - 5 ]. This dilemma underscores the need for a vaccine that, when administered shortly after birth to the infant, could protect against HIV transmission via breast-feeding. The ultimate goal of a neonatal HIV vaccine is to prevent infection; however, vaccination of newborns of HIV-infected women early in life may elicit HIV-specific immune responses that substantially reduce infant disease progression in the event that breast milk transmission occurs. Advances in the understanding of the mechanisms of oral transmission of HIV variants may aid the development of an effective infant HIV-1 vaccine. Recent studies have demonstrated that infants of HIV-infected women can be infected with single or multiple HIV variants [ 6 , 7 ] shortly before or during the birth process. However, little is known regarding the diversity of HIV transmitted by breastfeeding. These questions are difficult to address in human studies because the characteristics of HIV variants in breast-milk at the time of transmission are unknown. In addition, it is often difficult to obtain virus from infants at early times after HIV infection. Finally, the presence in infants of different levels of transplacentally transferred HIV-specific maternal antibodies with differing anti-viral properties complicates assessments of HIV variant transmission. Longitudinal studies of HIV-infected adults have shown that the rate of disease progression is inversely related to the rate of evolution of HIV envelope quasispecies [ 8 , 9 ]. Also, without antiviral treatment, virus-specific immune responses are directly related to HIV quasispecies evolution [ 10 ]. The reported relationship between HIV envelope variant evolution and disease progression in HIV-infected infants and children is contradictory. Some studies have found greater HIV envelope variant evolution in rapid progressors [ 11 - 13 ] while other investigations have found that slowly progressing HIV-infected children have greater HIV quasispecies divergence or diversity over time [ 14 , 15 ]. However, all of these retrospective studies necessarily evaluated HIV variant evolution in a limited number of serial blood samples during the first months of life from a small number of HIV-infected children (two to six per cohort). More recently, a longitudinal study of 10 perinatally HIV-infected children found that changes in HIV envelope quasispecies during the first year of life were associated with a better clinical outcome [ 7 ]. A few reports have described a correlation between nascent HIV-specific immune responses, the evolution of HIV variants and disease progression in HIV-infected infants [ 16 , 17 ]. Simian immunodeficiency virus (SIV) infection of infant macaques is a useful and relevant animal model of pediatric HIV infection for rapidly testing the efficacy of pediatric HIV vaccine and drug interventions [ 18 - 20 ]. This SIV/infant macaque model was previously used to assess the efficacy of two vaccines, (i) modified vaccinia virus Ankara (MVA) expressing SIV Gag, Pol and Env (MVA-SIVgpe) and (ii) live-attenuated SIVmac1A11, against oral challenge with virulent uncloned SIVmac251. We reported an improved clinical outcome (i.e., disease-free survival) for vaccinated compared with unvaccinated infants, which was associated with reduced plasma SIV RNA and sustained SIV-specific humoral immune responses [ 21 ]. Here in this report, we used a heteroduplex mobility assay (HMA) to evaluate the genetic diversity in the V1–V2 envelope ( env ) region of SIV variants present in the SIVmac251 virus inoculum and compare the transmission and evolution of the SIV env quasispecies in plasma following oral inoculation of these vaccinated and unvaccinated infant macaques. Three major questions were addressed: (i) Compared to the SIVmac251 virus inoculum, are few SIV envelope variants transmitted orally?, (ii) Is the lower viremia and better clinical outcome of vaccinated infants related to the initial genetic diversity of SIV env quasispecies?, and, (iii) Is the evolution of SIV envelope quasispecies during the course of infection associated with the development of SIV neutralizing antibody? We demonstrate that while the vaccines did not modulate oral transmission of viral variants, an association was found between vaccine-induced enhanced antiviral immune responses, increased env diversity, and a slower disease course. These findings in vaccinated infant macaques are similar to observations in HIV-infected children with slow disease progression and support the relevance of the SIV infant macaque model for developing neonatal vaccine strategies to prevent pediatric HIV infection and AIDS. Results Characterization of variants in SIVmac251-5/98 virus stock HMA analysis revealed that the undiluted SIVmac251-5/98 virus stock was comprised of a diverse population of V1–V2 env variants. To determine the most common variant(s) in the virus stock, six independent serial dilution experiments were conducted. Viral RNA was isolated from 1 ml of virus stock and 10-fold dilution series (undiluted to 10 -9 ) of the RNA were prepared from 6 separate aliquots of virus stock. The resulting RNA was analyzed by RT-PCR and HMA. Figure 1 shows the results of 4 of these 6 separate virus stock dilution/HMA experiments. The observation that multiple heteroduplex bands were observed through the 10 -5 or 10 -6 dilutions of viral RNA indicates that the undiluted SIVmac251-5/98 stock contains multiple env variants at high frequency. An RT-PCR endpoint (i.e. dilution to a single variant) was not reached in 3 of the 6 dilution experiments. An example of this is shown in dilution series A (Figure 1 ). In the other 3 dilution series (Fig. 1 , series B, C, and F), the last dilution that yielded an RT-PCR product consisted of a homogeneous population of envelope variants represented by one main variant (homoduplex band). This endpoint variant was designated the virus stock endpoint variant (VSEV). The fact that endpoint variants were reached at different dilutions for each dilution series is probably due to the variability at each step of these independently performed experiments. Figure 1 Characterization of variants in SIVmac251-5/98 virus stock. HMA analysis of four separate dilution series of viral RNA from the SIVmac251-5/98 virus stock is shown. The presence of multiple bands in the undiluted samples (lane 1 of each gel) reveals the virus stock was comprised of a diverse viral population. The last lane of each gel shows the variants in the highest dilution that yielded an RT-PCR product. Dilution series A shows an example of a dilution experiment that did not result in a virus stock endpoint (homogenous variant population); the 10 -6 dilution included more than 1 variant, while the next dilutions (10 -7 –10 -9 ) dilution did not yield RT-PCR products, and therefore no variant pattern is shown for those dilutions. This dilution pattern was observed in 3 of 6 dilution series (other 2 not shown). For the other 3 dilution series (B, C, and F), the variant (band) remaining in the highest dilution was considered to be the most common variant, and was designated the Virus Stock Endpoint Variant (VSEV). Dilution series B: no product was amplified from the 10 -7 dilution (lane 8), but a product was amplified from the 10 -8 dilution (lane 9). Dilution series C: lanes 7 and 8 show the presence of 2 different variants (VSEV-1 and VSEV-2) in the endpoint dilutions (10 -6 and 10 -7 ) of this series. Dilution series F; the 10 -6 dilution in this series harbored an endpoint variant that migrated to the same gel position as VSEV-2 in dilution series C. The VSEV in dilution series B and F had different mobilities on the HMA gel (Fig. 1 ). Dilution series C resulted in two endpoint variants, one at 10 -6 and the other at 10 -7 ; the positions of these two VSEV corresponded to one of each of the two VSEV in dilution series B and F. Thus, the dilution of the virus stock to an RT-PCR endpoint resulted in 4 independent variants (represented by homoduplex bands) that migrated to two different positions on the HMA gels. Based on these positions, the homoduplex bands that migrated furthest were referred to as VSEV-1 and the variants that migrated a shorter distance were designated VSEV-2 (Fig. 1 ). To confirm that the four endpoint homoduplexes represented only two variants, an HMA mixture experiment was performed, in which all pairwise combinations of the virus stock endpoint variants were mixed prior to HMA [ 22 ]. These experiments demonstrated that the two variants designated VSEV-1 are indeed similar (i.e., = 1–2% difference in nucleotides with no insertion/deletion), as their mixtures resulted in the formation of a single homoduplex band on an HMA gel; similarly, the two variants referred to as VSEV-2 are similar (Fig. 2 ). In contrast, the formation of heteroduplexes and two main homoduplexes in the mixtures of VSEV-1 and VSEV-2 demonstrate that these 2 variants are significantly different from each other (Fig. 2 ). Thus, VSEV-1 and VSEV-2 are 2 distinct variants that exist at similar frequencies and represent the most common variants in the undiluted SIVmac251-5/98 virus stock. These results are consistent with observations of the virus stock from which SIVmac251-5/98 was made [ 22 ]. Figure 2 Characterization of the dominant variants in SIVmac251-5/98 virus stock. HMA analysis of all four endpoint variants shown in Fig. 1 (lanes 1–4) and all possible pairwise mixtures of those variants (lanes 5-10) are shown. Letters B, C, and F refer to the dilution series shown in Fig. 1. Lane numbers refer to the lane designations of the variants that were mixed in lanes 5–10 (e.g., L1 + L2 indicates that the variants shown in lanes 1 and 2 were mixed). Lane 6 shows that the 2 endpoint variants labeled VSEV-1 (B 10 -8 and C 10 -7 ) are similar variants due to the formation of a single homoduplex and no heteroduplexes when these 2 variants were mixed. Lane 9 indicates that the 2 endpoint variants labeled VSEV-2 (C 10 -6 and F 10 -6 ) in Fig. 1 are very similar. The formation of heteroduplexes and two main homoduplexes in the mixtures shown in lanes 5, 7, 8, and 10 indicate that VSEV-1 and VSEV-2 do not share the same V1–V2 envelope sequence. Experimental design of animal experiments and summary of outcome Nineteen newborn rhesus macaques were divided into 5 experimental vaccine groups (table 1 ). Group 1 (n = 5) consisted of unimmunized control animals. Group 2 (n = 2), group 3 (n = 4) and group 4 (n = 4) were vaccinated with MVA-SIVgpe at 0 and 3 weeks of age; group 4 had maternally-derived SIV antibodies (due to immunization of their mothers with inactivated SIV). Group 5 (n = 4) was immunized with live-attenuated SIVmac1A11 at 0 and 3 weeks of age. As described elsewhere [ 21 ], except for group 2, all other groups were inoculated orally with SIVmac251-5/98 at 4 weeks of age; all these animals became persistently viremic, but the immunized animals had lower virus levels, enhanced antiviral immune responses and a delayed disease course in comparison to the unimmunized animals. Four of the 5 unimmunized infected animals developed AIDS within 14 weeks of age, while the fifth animal needed euthanasia at 28 weeks. Four MVA-SIVgpe-vaccinated SIVmac251-5/98-infected animals developed AIDS by 19 to 27 weeks of age (2 animals of groups 3 and 4 each; table 1 ). The remaining eight vaccinated SIVmac251-5/98-infected infants, including all four SIVmac1A11-vaccinated animals, were clinically stable at the end of the observation period (28 weeks of age). Table 1 Experimental design and summary of outcome. Immunization a groups and animal numbers sex MHC I alleles b Variant Pattern c Week 1 Plasma Viral RNA d Time of euthanasia (wks) e MamuA*01 MamuB*01 Group 1 Unvaccinated + SIVmac251 31319 M + + A 4.3 × 10 7 13 31321 M +/- - A 1.7 × 10 8 28 31322 F +/- +/- A 1.2 × 10 8 14 31325 M + + B 5.5 × 10 6 12 31608 f F +/- +/- C 7.5 × 10 5 11 Group 2 MVA-SIVgpe only 31480 M - - na na na 31488 M +/- +/- na na na Group 3 MVA-SIVgpe + SIVmac251 31378 M - - A 4.8 × 10 5 28 g 31533 M +/- - A 3.7 × 10 7 26 31540 M +/- - C 2.5 × 10 7 28 g 31542 M - - B 3.3 × 10 5 26 Group 4 MVA-SIVgpe with Mat. Abs. + SIVmac251 31526 M +/- +/- A 6.9 × 10 7 27 31732 F - +/- A 1.8 × 10 7 19 31833 F +/- - A/C 4.5 × 10 5 28 g 31856 F - +/- B 1.4 × 10 6 28 g Group 5 SIVmac1A11 + SIVmac251 31777 F +/- - A 6.8 × 10 7 28 g 31778 F - - B 4.7 × 10 5 28 g 31779 F - - A 2.3 × 10 8 28 g 31780 F +/- - A 9.9 × 10 7 28 g a Vaccine administered in 2 doses, at birth and 3 weeks of age. Animals of groups 1, 3, 4 and 5 were challenged orally at 4 weeks of age with SIVmac251-5/98. b The presence of the MHC type I alleles of MamuA*01 and MamuB*01 is indicated as + (present, but unknown whether homozygous or heterozygous), +/- (heterozygous based on known haplotypes of parents), and - (homozygous for absence of particular allele). c Variants in plasma at one week post-challenge with SIVmac251-5/98. d Copies of viral RNA per ml one week after challenge with SIVmac251-5/98, as measured by bDNA assay. e Age (weeks) at time of euthanasia. f Infant 31608 was born to an SIVmac251-infected macaque, and thus had maternal anti-SIV antibodies, but no virus was detected in this infant at 4 weeks of age. g indicates that animal was clinically stable at time of experimental euthanasia at 28 weeks of age; all other SIV-infected animals were euthanized due to life-threatening disease prior to or at 28 weeks of age. The animals of group 2 were not euthanized. na indicates not applicable. Detection of SIV envelope variants in plasma of neonates early after oral inoculation The genetic diversity of SIV env variant populations in the plasma of the infant monkeys one week after oral inoculation with SIVmac251 was analyzed by HMA (Fig. 3 ). Each plasma sample was analyzed in replicates (≥ 2) to assure reproducibility of the gel banding patterns. As indicated by the presence of heteroduplex bands, all infants were infected with multiple SIV env variants, indicating that the SIVmac251-5/98 virus stock contained several variants capable of establishing infection by the oral route. However, there were differences in HMA banding patterns. In each group, some animals had several strong heteroduplex bands; this pattern of variant transmission was referred to as infection pattern A (e.g. Fig. 3 , animal 31319). In contrast, one or two infants in each group were infected with a genetically more homogenous variant population, consisting of one major variant (homoduplex band), while heteroduplex bands were less pronounced. These monkeys infected with genetically more homogeneous viral populations harbored one of two main env variants, distinguished by different electrophoretic mobilities of the homoduplexes representing these variants. These more homogeneous variant populations were referred to as infection patterns B and C (e.g. Fig. 3 , animals 31325 and 31608, respectively). Infection pattern C contained a homoduplex band that migrated slightly slower than the homoduplex band characterizing infection pattern B. One newborn in each vaccine group was infected with a SIV variant of transmission pattern B. Infection pattern C was detected in one newborn of each group except the SIVmac1A11 vaccinates (table 1 , group 5). Therefore, no substantial difference was observed among the different vaccine groups in viral genetic diversity in plasma collected 1 week after virus inoculation. However, all but one infant (31540) infected with more homogenous populations of env variants (infection patterns B or C) had 10- to 100-fold lower virus levels one week after SIVmac251 challenge than all but one infant (31378) infected with more heterogeneous populations of SIV variants (transmission pattern A, Table 1 ). This association of homogeneous viral variants with reduced SIV RNA in plasma at 1 week after infection was statistically significant (P ≤ 0.05; one-sided Fisher's Exact test) but did not persist. From week 2 after challenge throughout the duration of the study, plasma SIV RNA levels showed no correlation with the initial SIV variant pattern detected in plasma. The rate of disease progression in these animals was also not associated with the initial envelope variant transmission patterns (table 1 ). Further, there was no correlation between the presence of the MHC type I alleles Mamu-A*01 or Mamu-B*01 and the viral variant infection patterns, levels of SIV RNA in plasma, or disease progression (table 1 ). Figure 3 Variant populations present in plasma of infant macaques one week after oral challenge with SIVmac251-5/98 . RT-PCR and HMA analysis was performed on replicate samples to confirm reproducibility of the results. Three main transmission patterns were observed, labeled A (multiple variants; diverse virus population), B and C (one major homoduplex (Ho) with a few faint heteroduplexes (He); relatively homogenous virus population). One infant (31833) harbored a plasma virus population that had elements of both transmission patterns A and C. SIV251 V.S. indicates the SIVmac251-5/98 virus stock. To determine which SIV envelope variant was present in the highest frequency in each infection pattern, serial end-point dilution experiments were performed with RNA isolated from plasma collected one week after SIVmac251 challenge, followed by RT-PCR and HMA. Similar to the methods described above, mixture experiments were then performed, including with VSEV-1 and VSEV-2. These experiments demonstrated that 1 week after infection, the most common variants in animals with the more homogenous transmission patterns B and C were similar (i.e., less than 1–2 % difference based on the absence of heteroduplex bands) to VSEV-1 and VSEV-2, respectively (data not shown). The most common variants by end-point dilution in the 11 monkeys with transmission pattern A and A/C were similar to VSEV-1 (5 animals), or VSEV-2 (5 animals) or both (1 animal). Greater quasispecies diversity in vaccinated compared to control infants during chronic SIV infection HMA was used to analyze the evolution of genetic diversity of V1–V2 env populations in plasma of the monkeys during the course of infection (1 week after oral SIVmac251-5/98 challenge until euthanasia) (Fig. 4 ). Results from two standard measures of the nucleotide sequence heterogeneity of V1–V2 env plasma variants derived from the HMA analyses are shown in Fig. 5 : (i) entropy (E), an estimate of the overall viral RNA sequence complexity for each sample and, (ii) median mobility shift (MMS), a measure of the SIV quasispecies sequence divergence reflected by the degree of base-pair mismatch after DNA strand re-annealing of strands of envelope variants [ 8 ]. Figure 4 Evolution of plasma variants in SIVmac251-5/98-infected infant macaques . HMA analysis was performed on sequential plasma RNA samples, and each analysis was done at least twice to assure reproducibility. Virus diversification is evidenced by the detection of additional minor heteroduplex bands, the disappearance of major heteroduplex bands, and/or the decrease in density of the homoduplex bands. V.S. indicates the SIVmac251-5/98 virus stock. The lane numbers refer to the number of weeks after SIVmac251-5/98 inoculation (which was performed at 4 weeks of age). The homoduplex band for week 0 for animal 31780 (prior to SIVmac251 challenge) represents the vaccine virus SIVmac1A11; viral RNA levels for the other SIVmac1A11-immunized animals at this time were too low to result in a detectable RT-PCR product. Figure 5 Evolution of viral diversity and SIV neutralizing antibody response. HMA data for each animal (Fig. 4) were further analyzed by calculating the entropy and the median mobility shift (MMS). Viral RNA levels were measured by bDNA. SIV neutralizing antibodies were determined as described in the Materials and Methods; neutralizing antibody titers below cut-off value (i.e., < 30) were given a value of 10 for presentation on these graphs. Dashed lines indicate a regression line for entropy, MMS or neutralizing antibody titer that is significantly different (P < 0.05) from zero (i.e. significantly increasing or decreasing values from 1 week to 24 weeks pc, with r 2 values ≥ 0.45). The diversity of SIV env quasispecies in plasma varied among animals at the first sample (1 week after challenge) as indicated by the gel banding pattern (Fig. 4 ) and entropy measures (Fig. 5 ). Entropy for SIV env populations was high (> 0.9) for all unvaccinated animals (Group 1) and for 7 of the 12 vaccinated animals (Fig. 5 ). Initial entropy < 0.9 for vaccinated animals was associated with lower SIV RNA in plasma at 1 week after challenge (P < 0.05; Fisher's Exact Test). No consistent pattern of entropy over the 24 week course of infection was observed; in two of the five controls (31321, 31608) and three of the 12 vaccinates (31533, 31732, 31780) entropy decreased near the time of euthanasia. Overall, there was no association of SIV envelope diversity as measured by entropy with either viral RNA levels or virus-specific neutralizing antibodies in plasma (see below and Fig. 5 ). The sequence divergence of SIV envelope variants in plasma of each animal over time was estimated by the MMS, shown in Fig. 5 . Four of the five unvaccinated animals had initial MMS values ≥ 0.5 which decreased at varying rates until the time of euthanasia; the remaining control animal (31325) had an initial MMS < 0.3 which did not change significantly over the course of infection (Fig. 5 ). Thus, in unvaccinated infants the population of SIV env variants in plasma exhibited either no sequence divergence or increasing sequence similarity over time; this observation is consistent with the absence of sustained SIV-specific immune responses in these animals ([ 21 ]; see below). There was no association between MMS values and SIV RNA plasma levels for these unimmunized animals. For 3 of the 4 vaccinated animals that developed AIDS within the observation period of 28 weeks (animals 31732, 31533, 31542), we also observed little change or a decrease of genetic divergence (i.e., as measured by stable or decreasing MMS values) of plasma env variant quasispecies. In contrast, diversification in plasma SIV env variant populations (i.e., a significant increase in MSS values) was observed by 3 to 5 months of infection in 4 of the 8 vaccinated monkeys (31540, 31833, 31856 and 31778) that were still relatively healthy at 28 weeks (Fig. 5 ). This diversification corresponded to the detection of different patterns of heteroduplex bands and/or fainter homoduplex bands over time (Fig. 4 ). Although increased diversification seemed to correlate with improved disease-free survival, this diversification was not associated with any obvious changes in plasma virus levels. SIV neutralizing antibodies in vaccinates correlate with evolution of SIV quasispecies diversity The possibility that SIV envelope-specific immune responses were associated with the observed plasma SIV RNA levels or evolution of SIV env quasispecies was evaluated by measuring levels of plasma antibodies that neutralized the homologous challenge virus, SIVmac251-5/98, during the course of infection (Fig. 5 ). SIV neutralizing antibodies were detected in none of the unvaccinated control animals, but in all except one (31777) of the 12 vaccinated animals within 16 to 20 weeks after infection (Fig. 5 ). Although the presence of SIV neutralizing antibodies was associated with increased survival of the vaccinated animals, no obvious relationship was detected between the SIV neutralizing antibody levels and either SIV RNA plasma levels or entropy over time in vaccinated animals. However, in the 5 animals with increasing sequence divergence (i.e., increasing MMS values; animals 31540, 31526, 31833, 31856 and 31778), neutralizing antibodies were detected around the time that MMS values increased, and the neutralizing antibody response was sustained (i.e., detectable in ≥ 3 plasma samples) in these 5 animals (Fig. 5 ). In contrast, animals with stable or declining MMS values had sustained (31533, 31542, 31779), transiently detected (31378, 31732, 31780) or undetectable (31319, 31321, 31322, 31325, 31608, 31777) anti-SIV neutralizing antibodies. Thus, a sustained SIV neutralizing antibody response was associated with increased divergence of SIV envelope variants in plasma (P = 0.009; Fisher's Exact test). Discussion The present study is among the most comprehensive longitudinal studies describing SIV envelope variation in vivo following mucosal SIV infection of infant macaques. In this study, we examined the extent of genetic diversity of the SIV envelope variant pool in the plasma of infant macaques that were inoculated orally at 4 weeks of age with an uncloned, genetically diverse virus stock SIVmac251-5/98. In addition, this is the first study to evaluate whether the transmission and evolution of viral variants was modulated by two different SIV vaccines, MVA-SIVgpe and SIVmac1A11, or the presence of maternally-derived anti-SIV antibodies. HMA analysis revealed that the animals became infected with multiple SIV envelope variant populations, but which predominantly consisted of one of two single envelope variants that were very similar to the two most common variants in the SIVmac251-5/98 stock. These results are consistent with reports of mother-to-infant HIV transmission of multiple variants [ 23 - 25 ], single variants [ 14 , 26 , 27 ] or both [ 6 , 28 - 31 ], but inconsistent with studies reporting vertical transmission of single, minor variants [ 10 , 26 , 32 - 34 ] from the mothers' virus population. This discrepancy could be explained by differences in the HIV inoculum regarding dose, virulence and genetic diversity compared to SIV. In the present study, macaques were inoculated with a relatively high dose of SIVmac251-5/98, while infection of human infants is likely to occur due to exposure to lower amounts of virus. An inherent limitation of studies of vertical transmission of HIV is that the exact timing of infection is usually unknown, and therefore the mothers' population of viral variants at the time of transmission and the source (e.g., breast-milk) and dose of virus is unknown. Our observation that oral exposure of 17 infant macaques to the same dose of the same virus stock resulted in different transmission patterns further underscores the complexity of studying variant transmission in humans, and suggests that the different outcomes observed for vertical transmission of HIV may not necessarily reflect "selection" of HIV variants but may be more a stochastic event. In this context, studies looking at the effect of heterogeneity of viral variants in the HIV-1 infected mother and the rate of vertical transmission have also shown conflicting results [ 6 , 26 , 35 ]. Also, the HIV studies mentioned focused on prenatal or intra-partum transmission, whereas our study modeled postnatal HIV transmission via breastfeeding by oral inoculation of 1-month old infant macaques with SIVmac251-5/98. The route(s) of infection in utero or during birth for individual infants and source of virus (cell-free or cell-associated) is usually unknown, and therefore different mechanisms may be responsible for viral transmission via these routes [ 6 ]. Consistent with this view, others have reported that more SIV variants were detected in orally infected newborn macaques than in infants born to SIV-infected female macaques for which transmission occurred in utero [ 36 ] or during the late breast-feeding period [ 37 ]. Neither of the SIV vaccines used in this experiment (MVA-SIVgpe and SIVmac1A11), nor the presence of maternal antibodies in one of the MVA-SIVgpe immunized groups altered which envelope variants were transmitted because in each group, some monkeys became infected with more heterogeneous and others with more homogeneous virus populations. It is possible that neither MVA-SIVgpe nor SIVmac1A11 elicited immune responses that effectively targeted the predominant SIV env variants in the SIVmac251-5/98 stock, or that anti-envelope immune responses were elicited against regions of the envelope other than V1–V2. It is also possible that vaccine-induced immune mechanisms at the time and/or site(s) of initial infection were not potent enough to modulate the variant transmission patterns. Viral levels in plasma of monkeys with more homogeneous populations of SIV env variants tended to be lower one week after oral inoculation with SIVmac251-5/98. The higher initial virus levels in infants infected with multiple variants may reflect higher replication capacities of diverse variant populations compared to those comprised of one main variant, especially in the initial target cells during the first days of infection. We have observed this previously for adult macaques inoculated intravaginally [ 22 ]. From the second week after SIVmac251-5/98 inoculation onwards, however, there was no correlation between viral genetic complexity (measured by entropy) or divergence (measured by MMS) and plasma SIV RNA levels. Thus, once systemic infection was established, virus replication attained similar levels regardless of the initial diversity, and there was no difference in AIDS-free survival times. Based on the measurement of MMS values, we observed little change or a decrease in genetic divergence of plasma SIV env variant quasispecies in all unvaccinated and most vaccinated animals that developed AIDS within the observation period of 28 weeks. In contrast, diversification in plasma SIV envelope variant populations was observed in 4 of the 8 vaccinated monkeys that were still relatively healthy at 28 weeks. This increased divergence of plasma viral variants at ~3 to 5 months after infection was generally associated with more sustained levels of SIV-specific neutralizing antibodies, and also of SIV Gag and Env-specific antibodies (measured by ELISA, as shown previously [ 21 ]). Similar associations between viral genetic divergence, immune parameters and/or disease progression have been described in HIV-infected adults and children [ 6 , 8 , 12 , 13 , 15 , 16 , 29 , 38 - 44 ], and recently also in juvenile macaques following intravenous or intra-rectal SIVsm inoculation [ 45 ]. Our studies extend these observations by demonstrating that this correlation of more sustained immune responses, enhanced viral divergence and slower disease progression is also observed in infant macaques following oral SIV infection. Together, these results suggest that the rate of virus evolution is determined by a combination of the extent of virus replication (which induces random mutations due to the error-prone reverse transcriptase) and selection pressures such as antiviral immune responses that promote the outgrowth of new variants. The generation of increasingly divergent viral variants ("immune escape mutants") reflects attempts of the immune system, albeit only partially effective, to control virus replication. In contrast, high viremia and little evolution of viral envelope variants is associated with severe immunodeficiency (and thus little immune selection pressure) and rapid disease progression. Conclusions The patterns of SIV env variant transmission and evolution in infant macaques that were inoculated orally with the same SIVmac251-5/98 stock reflect the range of results that is observed in mother-to-infant transmission of HIV, where the dose and genetic diversity of the virus at the time of transmission are unknown. While the vaccines tested here did not modulate oral transmission of viral variants, an association was found between vaccination and enhanced antiviral immune responses, increased env diversity, and a slower disease course. These findings are similar to observations in HIV-infected children with slow disease progression and underscore the relevance of the infant macaque model for developing neonatal vaccine strategies to prevent pediatric HIV infection and AIDS [ 46 ]. These results also support the concept that neonatal immunization could prevent rapid disease progression in infants who become HIV-infected by breast-feeding. Materials and Methods Infant immunizations, virus inoculations, and sample collection All newborn rhesus macaques ( Macaca mulatta ) were from the HIV-2, SIV, type D retrovirus, and simian T-cell lymphotropic virus type 1-free colony at the California National Primate Research Center. Newborn monkeys were hand-reared in a primate nursery, and all animals were housed in accordance with American Association for Accreditation of Laboratory Animal Care standards. We adhered to the "Guide for Care and Use of Laboratory Animals" [ 47 ]. When necessary, animals were immobilized with 10 mg/kg ketamine hydrochloride (Parke-Davis, Morris Plains, NJ) injected intramuscularly (IM). EDTA-anticoagulated blood samples were collected regularly for monitoring virologic and immunologic parameters as described previously [ 21 ]. Four newborn macaques had maternally derived SIV antibodies, because their mothers had been immunized and boosted during three or four consecutive pregnancies with whole-inactivated SIVmac251 plus Montanide ISA 51 adjuvant (Seppic, Fairfield, NJ), administered intramuscularly as previously described [ 21 ]. One of two SIV vaccines was administered to newborn monkeys at birth and 3 weeks of age: Modified Vaccinia virus Ankara expressing SIVmac239 gag, pol, and env (MVA-SIVgpe) was given to 8 newborn monkeys, including the 4 with maternal antibodies. SIVmac1A11 was given to 4 newborn monkeys. Details about these vaccines are described elsewhere [ 21 ]. At 4 weeks of age, these 17 monkeys were inoculated orally with 2 doses (24 hours apart) of uncloned virulent SIVmac251. Ketamine anesthesia was used for each inoculation. Each dose consisted of 1 ml of undiluted SIVmac251 of a stock designated by lot number -5/98, and was administered atraumatically by dispensing virus slowly into the mouth with a syringe. The SIVmac251-5/98 virus stock used in this study was derived from a previous SIVmac251 stock (lot 8/95) that was serially passaged intravenously in rhesus macaques as described [ 21 ]. This SIVmac251-5/98 stock contained 1 × 10 5 50% tissue culture infective doses (TCID 50 ) and 1.4 × 10 9 copies of RNA per ml (determined by bDNA assay). Quantitation of plasma viral RNA Viral RNA in plasma was quantified using a branched DNA (bDNA) signal amplification assay specific for SIV, with conditions as described previously [ 22 ]. RNA isolation and RT-PCR RNA was extracted from plasma samples (100–140 μl) using a viral RNA isolation kit (Qiagen, Inc., Valencia, CA) following the manufacturer's protocol. A 590 bp fragment encompassing the V1–V2 region of SIV env was then amplified in a nested RT-PCR assay as previously described [ 22 ]. Analysis of SIV variants by heteroduplex mobility assay (HMA) Genetic diversity in viral variant populations was analyzed using a modification of the HMA methods described elsewhere [ 22 , 48 ]. In brief, V1–V2 env fragments were generated by RT-PCR as described above, and the presence of sufficient product was confirmed on a 1.5% agarose gel. The RT-PCR products were then mixed with 1.5 μl of 10× annealing buffer (1 M NaCl, 100 mM Tris, 20 mM EDTA), denatured at 94°C for 2 minutes and placed immediately on wet ice to promote heteroduplex formation. Samples were then run on non-denaturing 5% polyacrylamide gels and stained with ethidium bromide (0.5 μg/ml). Reverse-images of the stained gels were photographed with a digital imaging system (Alpha Innotech Corporation, San Leandro, CA). All gel images were color reversed to enhance visualization of banding patterns (e.g. black bands on white background). The number of heteroduplex bands observed is a measure of SIV envelope diversity in each monkey's virus population (i.e. a large number of bands on a gel corresponds to a large number of V1–V2 variants in the sample). The RT-PCR and HMA analysis on plasma samples was performed in replicates (of at least 2) to assure reproducibility of the gel banding patterns. To further characterize SIV envelope variants, we used an additional form of HMA analysis that assesses the relative genetic similarity of specific viral variants by comparing the HMA patterns that result from combinations of these variants. Mixtures of SIV RNA from plasma samples were analyzed by HMA to allow estimation of sequence similarity between two different homogeneous virus populations (i.e. the most common inoculum variants and/or plasma variants from infected monkeys). This "mixture analysis" is based on the HMA HIV subtyping protocol developed by Delwart et al. [ 49 ] and was performed as described previously [ 22 ]. Briefly, mixtures of equal volumes of SIV V1–V2 env PCR product amplified from two different samples were mixed and subjected to HMA analysis as described above. Mixtures that result in a single homoduplex band are 98–100% identical [ 48 ] in the nucleotide sequence of the PCR fragment analyzed (V1–V2 env region). Mixtures that result in the formation of heteroduplexes are comprised of variant populations with nucleotide sequences that differ by more than 1–2% or have an insertion/deletion (i.e., a single codon length difference will also cause a gel shift) [ 48 ]. We have validated this HMA method for SIVmac251 in a previous study [ 22 ]. Calculation of entropy and median mobility shift All measures of entropy (E) and median mobility shift (MMS) were estimated according to methods described by Delwart et al. [ 8 ]. Images from HMA gels were captured with a CCD camera as binary TIFF files and color reversed to enhance visualization of banding patterns (e.g. black bands on white background). Each TIFF gel image file was then opened using Adobe Photoshop Version 6.0 (Adobe Inc., San Jose, CA) and edited to ensure that the lightest inter-lane areas of the gel image had "0" signal intensity (as read by the NIH Image Program and required for the Hdent program described below). Digitized gel lanes were scanned by using the plot profile function of the NIH Image Program (available at ). Lane scans within the same gel were of equal length (i.e. same number of pixels) and were recorded from positions immediately below the single-stranded DNA position to immediately below the homoduplex. The signal intensity at each pixel along the scan was transferred to a Microsoft Excel (Richmond, Wash.) file. Because different numbers of pixels per lane were acquired from different gels, each gel was standardized by partitioning into 191 divisions, the smallest number of pixels in the scans under study. This allowed the maximum distinction of fine banding patterns, while permitting unbiased comparison between gels. The quasispecies diversity for each sample was estimated by calculating a normalized Shannon entropy, a measure of the breadth or spread of the signal distribution in each HMA gel lane, using the HDent program (available at ) as described by Delwart et al. [ 8 ]. The Shannon entropy ( S ) is defined as: S = - Σ (from i = 1 to N ) P (i)ln [ P (i)], where N is the number of partitions in a lane, and P (i) is the fraction of the total signal in partition i . The maximum possible entropy is ln( N ), and we defined the normalized entropy as S/ln( N ). The normalized entropy has a range of 0 to 1, where 0 reflects no diversity (all of the signal is in a single partition), and 1 reflects maximum entropy, in which the signal is evenly distributed throughout all partitions in the lane. Thus, entropy is large for lanes with many, closely spaced or overlapping bands and small for lanes with only one band or a few, narrow bands. Shannon entropy estimates quasispecies genetic diversity by measuring the pattern of SIV V1–V2 env heteroduplex distribution in an HMA gel lane rather than the specific electrophoretic mobility of heteroduplexes. However, the electrophoretic mobility of heteroduplexes through a polyacrylamide gel is proportional to the sequence differences in reannealed DNA strands [ 38 , 48 , 49 ]. The degree of SIV quasispecies envelope sequence divergence among the V1–V2 env variants present in plasma samples was estimated by calculating a median mobility shift (MMS) for each HMA gel lane using the HDent program. The MMS is a measure of the midpoint of the total signal in an HMA gel lane that has values between 0 and 1, where 0 corresponds to the bottom of an HMA gel lane (i.e. nearest homoduplex bands) and 1 corresponds to the top of the lane. Thus, a MMS score of 1 reflects maximum sequence diversity (i.e. all heteroduplexes bands have maximum mobility reduction and no visible homoduplexes); a MMS value of 0 reflects maximum sequence similarity (> 98%) where all signal for a lane is in homoduplex bands and there are no visible heteroduplexes. Assessment of MHC class I alleles DNA extracted from lymphoid cells (with QIAamp ® DNA mini kit, QIAgen, Valencia, CA) was used to screen for the presence of the rhesus macaque major histocompatibility complex (MHC) class I alleles Mamu A*01 and Mamu B*01, using a PCR-based technique [ 50 , 51 ]. Neutralizing antibodies Neutralizing antibody titers in EDTA-anticoagulated plasma were measured according to methods described previously [ 52 ], except that CEM-R5 cells (i.e. CEM×174 cells expressing CCR5 by transfection; generously provided by James Robinson) were used. Neutralizing antibody titers were expressed as the reciprocal of the plasma dilution at which 50% of cells were protected from virus-induced killing as measured by neutral red uptake. The virus consisted of SIVmac251-5/98 briefly propagated in human PBMC. Statistical analysis Fisher's exact test, performed with Instat v. 2.03 (GraphPad Software, Inc., San Diego, CA), was used to evaluate possible association of SIV env V1–V2 variants detected by HMA in plasma of monkeys one week after oral SIVmac251 inoculation with the levels of plasma viral RNA at this same time point. To determine potential linear associations of Entropy values or MMS values over time and SIV neutralizing antibody levels over time, linear regression was performed using Prism v.3.0 (GraphPad Software Inc., San Diego CA). For all statistical comparisons, a P value less than 0.05 was considered significant. Competing interests The author(s) declare they have no competing interests. Authors' contributions JG carried out the HMA studies and drafted the manuscript; KVR participated in the design and coordination of the study, acquisition and analysis of data, and helped draft the manuscript; DM analyzed and interpreted all neutralizing antibody assays; PE and BM designed and provided the MVA constructs, participated in the experimental design and manuscript writing; MM designed and coordinated the study, assisted in the data analyses and helped draft the manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552328.xml
524255
Predicting Risk of Mosquito-Borne Disease in Variable Environments
null
Malaria remains one of the greatest threats to global health, infecting more people than ever before. Confined mainly to the tropical areas of Africa, Asia, and Central America, malaria hits Africa the hardest; the poverty-stricken lands of sub-Saharan Africa account for 90% of malaria infections worldwide. Despite ongoing efforts to battle the disease—by controlling mosquito populations, reducing human contact, and developing drug prevention and treatment—the crisis continues to worsen. The primary variables affecting risk of infection are the rate at which humans are bitten and the proportion of mosquitoes that are infectious. These two factors are often regarded as positively correlated, meaning that if the percentage of infectious mosquitoes increases, so will the human biting rate. But in a new study, David Smith, Jonathan Dushoff, and F. Ellis McKenzie challenge this assumption. Using a mathematical modeling approach to examine the relative contributions of the two factors across different landscapes and seasons, the authors show that the factors are not positively correlated. In fact, their calculations show that the rate humans are bitten and the proportion of infectious mosquitoes peak at different times and places. Their modeling results suggest that the standard metric to estimate risk of infection—the number of times an infectious mosquito bites a person per day, called the entomological inoculation rate (EIR)—is flawed when variable conditions are taken into account. Using the average EIR to estimate average risk of infection in variable environments generates biased estimates because there is not a direct correlation between EIR and the proportion of humans who are infected. The distribution of humans and suitable habitat for mosquito larvae varies across the landscape. And the density of mosquito populations varies seasonally, rising and falling with changes in rainfall, temperature, and humidity. Temporal and spatial variations in mosquito populations affect the rate humans get bitten, the number of infectious mosquitoes, and the risk of infection. To understand how these space- and time-induced variations in mosquito populations shape the epidemiology of human infection, Smith and colleagues developed a set of mathematical models that calculate the relative impact of different parameters, in order to determine which factors most influence where and when risk of infection is highest. First, they evaluated what factors affect the primary components of the EIR: the human biting rate and the proportion of infectious mosquitoes. As expected, the model predicts that fluctuations in mosquito density influence the EIR by changing the human biting rate. As more people are bitten, more people become infected; consequently, more mosquitoes feed on infected humans and so become infectious. Only adult mosquitoes transmit infection, so as mosquito populations age, the proportion of infectious mosquitoes increases. During the dry season, few mosquitoes are born, and so while the human biting rate and EIR decline, the proportion of infectious mosquitoes increases. Because mosquito populations are densest near breeding sites—where younger mosquitoes outnumber adults—the human biting rate and the number of bites by infectious mosquitoes per person per day reflect shifts in mosquito density, not in the proportion of infectious mosquitoes. The model predicts that human biting rate is highest shortly after mosquito population density peaks, typically either near breeding sites or where human density is highest. The proportion of infectious mosquitoes, on the other hand, reflect the age of the mosquito population: it peaks where older mosquitoes are found—farther from breeding sites—and when populations are declining. By mapping larval habitats against the local risk of mosquito-borne infections, Smith and colleagues conclude, epidemiological models can be developed to predict risk for local populations. Their results make the case that mathematical models can help public health officials calculate risk of infectious diseases in heterogeneous environments—that is, real world conditions—when vector ecology and the parameters of transmission are well characterized. Any plan to prevent and control the spread of mosquito-born infections would clearly benefit from paying attention to mosquito demography and behavior.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524255.xml
554759
Involvement of β-chemokines in the development of inflammatory demyelination
The importance of β-chemokines (or CC chemokine ligands – CCL) in the development of inflammatory lesions in the central nervous system of patients with multiple sclerosis and rodents with experimental allergic encephalomyelitis is strongly supported by descriptive studies and experimental models. Our recent genetic scans in families identified haplotypes in the genes of CCL2, CCL3 and CCL11-CCL8-CCL13 which showed association with multiple sclerosis. Complementing the genetic associations, we also detected a distinct regional expression regulation for CCL2, CCL7 and CCL8 in correlation with chronic inflammation in multiple sclerosis brains. These observations are in consensus with previous studies, and add new data to support the involvement of CCL2, CCL7, CCL8 and CCL3 in the development of inflammatory demyelination. Along with our own data, here we review the literature implicating CCLs and their receptors (CCRs) in multiple sclerosis and experimental allergic encephalomyelitis. The survey reflects that the field is in a rapid expansion, and highlights some of the pathways which might be suitable to pharmaceutical interventions.
Introduction Multiple sclerosis (MS) is a disabling disease of the central nervous system (CNS) with features of autoimmunity and neurodegeneration. Although the identity of primary antigenic determinant(s) is uncertain, an interaction between β-chemokine ligands and their receptors plays a central role in the recruitment and retention of inflammatory cells in the CNS. Thus, both the disease relevant chemokine ligands and their receptors represent potential therapeutic targets in MS. Chemokines are a group of small, structurally related chemoattractant molecules that regulate cell trafficking through interactions with a set of receptors [ 1 ]. Evidence suggests that the migration of autoreactive immune cells via the blood-brain barrier (BBB) is an early and critical process during the development of inflammatory CNS lesions of experimental allergic encephalomyelitis (EAE) and MS, and that this transmigration is regulated by chemokines produced at the blood-brain barrier (BBB) and in the CNS. Subcellular signals induced by the binding of chemokines to their G-protein-coupled receptors leads to an increased avidity of integrins on leukocytes to their corresponding receptors on endothelial cells, followed by a facilitated migration of leukocytes towards the chemokine gradient in the CNS [ 2 , 3 ]. In addition to chemotaxis, chemokines are also involved in the regulation of T cell differentiation, apoptosis, cell cycle, angiogenesis and metastatic processes. Further, chemokines can control the generation of soluble inflammatory products such as free radicals, nitric oxide, cytokines and matrix metalloproteases [ 1 , 4 ]. Considering the predominantly T helper type 1 (TH1) mediated process of inflammatory demyelination and the TH2 driven suppression of inflammation, the differential effects of various chemokines on TH1 or TH2 polarization may have particular significance. The currently known, approximately 50 chemokine genes in humans are divided into four subfamilies on the basis of characteristic patterns of cysteine residues close to the N-terminal end of the products. The CC chemokine ligand family (CCL) (also known as β-chemokines or Small Cytokine Group A – SCYA in mice) is characterized by two adjacent cysteines, while the CXC (SCYB) and CX 3 C (SCYD or fractalkine) chemokine families have one or three intervening amino acids, respectively, between the two cysteines. In the XC family (SCYC or lymphotactin), only one cysteine is present [ 1 ]. All four classes of chemokines play important roles in the immune inflammatory network, but because of the complexity of interactions, here we only discuss the CC chemokine family. In humans, there are 27 CC chemokines, most of which including CCL2, CCL7, CCL11, CCL8, CCL13, CCL1, CCL5, CCL16, CCL14, CCL15, CCL23, CCL18, CCL3 and CCL4, respectively, are encoded as a cluster within chromosome 17q11. The genes for CCL27, CCL19 and CCL21 are located within chromosome 9p13, while CCL17 and CCL22 are encoded at 16q13. The remaining CCL genes can be found on chromosome 2 and 7 [ 1 ]. A functional classification was also proposed to distinguish between lymphoid and inflammatory chemokines [ 1 , 5 ]. Lymphoid or homeostatic chemokines (e.g. CCL21, CCL25, CXCL13) are constitutively expressed and control physiologic trafficking of cells of the adoptive immune system during hematopoiesis and immunosurveillance. Inflammatory or induced chemokines (e.g. CCL2, CCL3, CCL5, CCL7, CCL8, etc...) are transcriptionally regulated during inflammation and mediate the recruitment of inflammatory cells to target tissues. The effects of chemokines are mediated by G-protein coupled receptors with seven-transmembrane-domains. Chemokine receptors tend to bind multiple chemokine ligands and vice versa. However, the biologically most efficient interaction often occurs between a receptor and its primary ligand (e.g. CCL2 – CCR2). The receptor binding involves high affinity interactions and signal transduction initiated by the dissociation of G-protein complex into Gα and Gβγ subunits. Gα induces the activation of the phosphoinositidine 3-kinase pathway, while the Gβγ subunits activate phospholipase C and induce Ca 2+ influx and protein kinase C activation. The involvement of MAP kinases as well as JAK/STAT signaling also has been shown [ 6 ]. As of today, 10 CC chemokine receptors (CCRs), 6 CXCRs, one CX 3 CR1 and one XCR1 are known [ 1 , 6 ]. This review focuses on the immunomodulatory effects of the β-chemokine or CCL family in EAE and MS. CC chemokines predominantly are involved in the recruitment of monocytes / macrophages and dendritic cells (monocyte chemoattractant proteins -MCP-1 [CCL2], MCP-2 [CCL8], MCP-3 [CCL7], MCP-4 [CCL13] and macrophage inflammatory proteins – MIP-1α [CCL3] and MIP-1β [CCL4]), and to lesser degrees, T lymphocytes and NK cells (MCP and MIP chemokines, regulated upon activation normal T cell expressed and secreted cytokine [RANTES]) or occasionally other cell types (e.g. eosinophil chemotactic protein – eotactin [CCL11]) into inflammatory lesions of MS. Genetic evidence for the involvement of β-chemokines in multiple sclerosis A meta-analysis of raw genotype data from three genome scans in MS families revealed the highest nonparametric linkage (NPL) score = 2.58 at 17q11 [ 7 ]. Among several candidate genes (e.g. NOS2A, OMG, NF1), a cluster of evolutionarily closely related β-chemokine genes [CCL2, CCL7, CCL11, CCL8, CCL13, CCL1, CCL5, CCL16, CCL14, CCL15, CCL23, CCL18, CCL3 and CCL4, respectively] is encoded within a 1.85 Mb segment of 17q11.2-q12. Our recent linkage disequilibrium mapping confined the susceptibility regions to 3–30 kb haplotypes defined by single nucleotide polymorphisms (SNP) within the genes of CCL2, CCL11-CCL8, CCL8-CCL13, CCL13 and CCL3 [ 8 ]. A second study is under way to confirm and further refine the MS relevant haplotypes, and then, to identify the specific disease causing nucleotide variants in an independent set of families. Within the orthologous mouse chromosome 11, two quantitative trait loci (QTL), eae6 and eae7 were identified. While these loci control the severity and duration of EAE, eae7 is also a susceptibility locus for the monophasic remitting / non-relapsing subtype of the disease [ 9 ]. Sequence polymorphisms within the genes of Scya1 (TCA-3 or CCL1), Scya2 (MCP-1 or CCL2) and Scya12 (MCP-5 or CCL12) in eae7 showed striking segregations among mouse strains resistant or susceptible to EAE. Using an advanced intercross line in combination with congenic strains, Jagodic et al. [ 10 ] fine mapped eae18 and identified two adjacent QTLs, eae18a and eae18b , on the rat chromosome 10 in a myelin-oligodendrocyte glycoprotein (MOG)-induced, chronic relapsing EAE. The eae18b locus is also orthologous to human chromosome 17q11 and encodes a cluster of β-chemokine genes. The recognition of β-chemokine genes as susceptibility and quantitative trait loci in mouse and rat EAE along with the human data revealing the β-chemokine gene cluster as a susceptibility locus in MS, strongly suggest the involvement of β-chemokine variants in the development of inflammatory demyelination. CCL and CCR molecules in inflammatory demyelination Experimental allergic encephalomyelitis EAE is a valuable model for studying the effector arm of immune response in inflammatory demyelination. It can be induced in susceptible strains of inbred and outbred species by active immunization with myelin related proteins and their peptides (myelin basic protein – MBP, proteolipid lipoprotein – PLP, myelin oligodendrocyte glycoprotein – MOG) emulsified in Freund's complete adjuvant along with intravenous Pertussis toxin, or with a passive transfer of myelin antigen specific T cell lines into naïve recipients. Using various immunization protocols, acute and chronic relapsing (CR-EAE) models have been developed. In both the active immunization and the passive transfer models of EAE, the efferent arm of immune response involves the migration of monocytes / macrophages, dendritic cells and activated myelin-antigen-specific T lymphocytes from the blood circulation into the CNS, where a reactivation of specific lymphocytes by myelin-antigen-presenting dendritic cells, macrophages and residential microglia takes place, and the sequential development of perivascular and parenchymal inflammation is followed by demyelination and neuronal degeneration. Encephalitogenic T lymphocytes have CD4+ TH1 phenotype characterized by the production of interleukin (IL)-2 and interferon-γ. TH2 lymphocytes producing IL4, IL5, IL6 and IL10 cytokines are involved in the counter-regulation of TH1 effects, and promote clinical recovery. The TH1 / TH2 polarization is regulated by cytokines and chemokines. The transmigration of immune competent cells via the blood-brain barrier is aided by a temporal and spatial regulation of adhesion molecules on T lymphocytes and their counterparts on endothelial cells, and of chemokine ligands and their receptors in the residential CNS and hematogenous mononuclear cells. One of the most comprehensively studied CC chemokines in inflammatory demyelination is MCP-1 (CCL2). MCP-1 (CCL2) influences both innate immunity through its chemoattractant effect on monocytes / macrophages, and adaptive immunity through its effect on T cell polarization towards the TH2 subtype [ 11 ]. CCL2 primarily acts via the CCR2 receptor [ 1 ]. MIP proteins have both chemotactic and proinflammatory effects, but also promote homeostasis [ 6 ]. The MIP-1 family includes MIP-1α (CCL3), MIP-1β (CCL4), MIP-1δ (CCL9/10) and MIP-1γ (CCL15) that are produced by macrophages, microglia, astrocytes, dendritic cells and lymphocytes. These MIP-1 molecules act via CCR1, CCR3 and CCR5 expressed by lymphocytes and monocytes. MIP-1 proteins also regulate immune response by modulating T cell differentiation. The CCL3 and CCR5 interaction promotes polarization towards the TH1 subtype. Our understanding concerning the role of these CC chemokines and their receptors in inflammatory demyelination was greatly advanced by studies in the EAE model. In mice, the increased expression of MCP-1 (CCL2) by CNS immune cells is closely associated with the clinical activity of EAE [ 12 - 14 ]. Some studies, however, suggest that the presence of leukocytes is necessary for the production of CCL2 by astrocytes, as the expression of CCL2 prior to the accumulation of inflammatory mononuclear cells has not been observed in the CNS. Substantial MCP-1 (CCL2) expression may only occur in the late phase of acute disease and in the relapsing phases of CR-EAE. It was therefore postulated, that CCL2 is involved in the amplification rather than in the initiation of EAE [ 4 ]. In contrast, the MIP-1α (CCL3) expression correlates with the severity of acute disease and also is elevated during relapses in CR-EAE. RANTES (CCL5) is expressed in the CNS throughout the course, but does not correlate with the severity of acute or CR-EAE [ 13 ]. Jee et al [ 15 ] compared the histological features and MCP-1 (CCL2) and CCR2 expression levels in the lesions of Lewis rats during the acute attack of monophasic EAE and during the first two clinical events of CR-EAE. In concert with the mouse data [ 4 , 13 ], not only higher numbers of macrophages infiltrated the spinal cord during the first and second attacks of CR-EAE as compared to those at the peak of acute EAE in these rats, but the expression of MCP-1 (CCL2) was also significantly higher in the lesion of CR-EAE as compared to that of acute EAE. Similarly, CCR2, the main receptor for CCL2, was expressed by astrocytes, macrophages and T cells in higher amounts during CR-EAE than at the peak of acute EAE. This observation confirmed the role of CCL2 – CCR2 interaction in the development of relapses. Youseff et al [ 16 ] observed an increased mRNA transcription not only for MCP-1 (CCL2), but also for MIP-1α (CCL3) and MIP-1β (CCL4) at the onset of EAE in rat brains. MIP-1α (CCL3) and MCP-1 (CCL2) declined in two days even though the clinical disease further progressed. MIP-1β (CCL4) mRNA declined in correlation with the clinical recovery. RANTES (CCL5) mRNA, in contrast, increased in the brains only after recovery. The full length, reverse transcribed and PCR amplified DNA product for each of these four CCL molecules was transferred into a plasmid vector and injected as naked DNA vaccine into rats. Both the transcription of a relevant chemokine and the induced antibody response against it was monitored. The in vivo immune response to these CCL molecules differentially influenced the evolution of EAE. MIP-1α (CCL3) and MCP-1 (CCL2) DNA vaccines prevented EAE, while MIP-1β (CCL4) aggravated the disease and RANTES (CCL5) did not have an effect on the course of EAE. This study emphasizes the importance of CCL2 and CCL3 in the development of active EAE in rats. CCL1 also attracted attention in EAE. Teutscher et al [ 9 ] identified eae7 encoding CCL1 and other chemokines as a susceptibility locus and QTL in murine EAE. SNPs in CCL1 differentially segregated in mouse strains susceptible or resistant to EAE. mRNA molecules for both CCL1 and its receptor CCR8 were detected in spinal cord lesions of EAE, in correlation with the expression of tumor necrosis factor (TNF)-α by inflammatory leukocytes [ 17 - 19 ]. As both CCL1 and CCR8 were detected in microglia, an autocrine signaling mechanism was postulated. CCR8 (-/-) mice showed marked delay in the onset and reduced severity of EAE as compared to controls. Leukocyte infiltration in the spinal cord was not diminished in the CCR8 (-/-) mice, suggesting that that a defective microglial activation might have altered the clinical phenotype. Recent studies addressed the role of chemokines at the blood-brain barrier. Using intravital fluorescence videomicroscopy, Vajkoczy et al [ 20 ] demonstrated that the interaction between encephalitogenic T cells and endothelial cells of the BBB involves α4-integrin (VLA-4) which mediates a G-protein-independent capture (arrest) followed by G-protein-dependent adhesion strengthening of circulating T cells to VCAM-1 on endothelial cells. Postulating the involvement of chemokines in the integrin-mediated arrest of autoreactive T cells at the BBB, the investigators [ 3 ] subsequently aimed to identify the specific chemokines by performing in situ hybridization and immunohistochemistry on brain and spinal cord sections of mice with EAE. Constitutive expression of the lymphoid chemokine called EBV-induced molecule 1 ligand chemokine (ELC) / CCL19 in a subpopulation of CNS venules and induced expression of the secondary lymphoid chemokine (SLC) / CCL21 in inflamed CNS venules was detected. CCR7, the common receptor for these two chemokines was expressed on a subpopulation of cells in the perivascular cuffs. Encephalitogenic T cells in vitro showed expression of CCR7 and CXCR3, the alternative receptor for CCL21, and chemotaxed towards both CCL19 and CCL21 in a concentration-dependent and a Pertussis toxin-sensitive manner similar to naïve T cells. Functional deletion of CCR7 and CXCR3 or immune blockade of CCL19 and CCL21 reduced the binding of encephalitogenic T cells to inflamed venules in frozen brain sections. Altogether, these data suggest that CCL19 and CCL21 are expressed in cerebral endothelial cells and are involved in α4-integrin mediated adhesion strengthening of autoreactive T cells and subsequently of other inflammatory cells to the endothelial layer of the BBB. These molecular interactions may lead to permanent inflammatory cell immigration into the CNS in chronic autoimmune disease. CCL20 or MIP-3α (exodus-3 / LARC) is a chemokine active on dendritic cells and lymphocytes that express CCR6 [ 1 ]. Serafini et al [ 21 ]demonstrated the occurrence of dendritic cells in the spinal cord of mice immunized with the PLP139–151 peptide. Although dendritic cells were present during early acute, chronic and relapsing EAE, most prominent infiltration of spinal cord by mature dendritic cells was noted in relapsing disease. In all stages of EAE, CCL20 and CCR6 were upregulated in the CNS. This study emphasizes the importance of dendritic cells in antigen presentation and T cell restimulation, and links the immigration of dendritic cells to the expression of CCL20 in the CNS during EAE. CCL22 or macrophage-derived chemokine (MDC) is chemoattractant for monocytes, dendritic and NK cells, and T lymphocytes of the TH2 subtype. MDC / CCL22 acts via CCR4 which is preferentially detected on TH2 type, memory and regulatory T cells [ 22 ]. While MDC / CCL22 is considered to be predominantly involved in TH2 mediated immunity, Columba-Cabezas et al [ 22 ] demonstrated mRNA expression for MDC / CCL22 in the CNS of mice with relapsing-remitting and chronic-relapsing EAE induced by PLP139–151 or whole spinal cord homogenate. Immunohistochemistry demonstrated that MDC / CCL22 is produced by infiltrating leukocytes and residential microglia, while CCR4 is expressed by infiltrating leukocytes. In vitro activation of microglia resulted in secretion of bioactive MDC / CCL22 that induced chemotaxis of TH2 lymphocytes. This study concludes that MDC / CCL22 produced by microglia may play a role in a TH1 mediated CNS inflammation by inducing the homing of TH2 regulatory cells into the lesion site. To further clarify the role of chemokine receptors involved in EAE, Fife et al [ 23 ] examined CCR expression in normal (unprimed), PLP139–151 primed non-activated, PLP139–151 primed and reactivated lymph node derived T cells, and CNS-isolated CD4+ T cells from SJL mice receiving PLP139–151 specific, in vitro reactivated T cells. Normal resting CD4+ T cells and primed non-activated T cells expressed mRNA for CCR1, CCR2, CCR3, CCR5, CCR6, CCR7 and CCR8. In vitro activated T cells expressed in higher amounts most of the CCRs found in normal T cells as well as CCR4. After passive transfer of encephalitogenic activated T cells in naïve recipients, the donor derived encephalitogenic cells and the host-derived CD4+ T cells isolated only from the CNS lesions but not from spleen expressed mRNA for CCR1. This latter observation was confirmed at protein level, and appeared to be specific for acute EAE. Neutralization of the CCR1 ligand CCL3 (MIP-1α) diminished the inflammatory infiltrate in the CNS. The effects of anti-chemokine treatments in the mouse EAE is summarized by Elhofy et al [ 24 ] and Karpus et al [ 25 ] and is in consensus with data in the rat model. Although various strains and protocols were used, overall anti-RANTES (CCL5) had no effect in these models, anti-MIP-1α (CCL3) decreased the severity of acute EAE and anti-MCP-1 (CCL2) reduced the severity of both acute EAE and the relapses in CR-EAE. However, it is important noting that in some respect, these observations are model specific. While the impact of anti-CCL5 immune treatment was unremarkable in the autoantigen-induced forms of EAE, antibody treatment targeting CCL5 in a mouse hepatitis virus-induced inflammatory demyelination model resulted in diminished leukocyte infiltration and reduced neurological disability [ 26 ]. Genetic manipulations of the murine model provide further insights in the characterization of CCL / CCR molecules in EAE. In mice with the CCL2 transgene under the control of the lck (which directs the expression of transgene to cortical thymocytes) or MBP promoters (which directs the expression of transgene to the CNS), a spontaneous infiltration of monocytes / macrophages in the thymus and CNS was observed, respectively [ 27 ]. LPS injection induced higher CCL2 expression in the brain and markedly enhanced the mononuclear cell (MNC) infiltrate. The relationship between LPS treatment, CCL2 expression and MNC recruitment into the CNS remains partially understood, and seems to involve a complex immune regulatory mechanism rather than just a selective effect mediated by the upregulation of the CCL2 transgene. Nevertheless, these transgenic mice were clinically normal both before and after LPS injection. More recently, Elhofy et al [ 28 ] examined TH1 lymphocytes in a PLP-induced EAE model using a transgenic mouse strain that constitutively expressed low CCL2 levels in the CNS under the control of the astrocyte-specific glial fibrillary acidic protein promoter. CCL2 transgenic mice developed milder EAE than the littermate controls, despite similar numbers of CD4 and CD8 T cells in the CNS infiltrates and an increased number of monocytes in the CNS of the CCL2 transgenic animals. Functional studies revealed that encephalitogenic T cells from the CCL2 transgenic mice produced significantly less interferon-γ and proliferated less in the presence of PLP peptides than those of the non-transgenic controls. Increased CCL2 expression in the CNS also resulted in a decreased IL-12 receptor expression by PLP-specific T cells. Thus in this model, the overexpression of CCL2 in the CNS resulted in a suppression of the TH1 response and a milder clinical phenotype of EAE, despite the enhanced effect on monocytes. The CCL2 knock out (-/-) mice showed resistance to EAE and significantly decreased macrophage infiltration in the CNS following active immunization with MOG35–55 peptide. While T cells from CCL2 (-/-) mice transferred EAE to wild type mice, wild type T cells did not induce EAE in CCL2 (-/-) recipient mice. These observations suggest a key role for CCL2 in the recruitment of macrophages into the CNS and thus, in the pathogenesis of EAE [ 29 , 30 , 4 ]. The array of ligands for CCR2 includes MCP-1 (CCL2), MCP-2 (CCL8), MCP-3 (CCL7) or MCP-5 (CCL12). As CCL2 (-/-) mice did not show a compensatory upregulation of MCP-2 (CCL8), MCP-3 (CCL7) or MCP-5 (CCL12) mRNA molecules, MCP-1 (CCL2) is likely to be the main ligand for CCR2 in mice with EAE. The clinical phenotype of CCR2 (-/-) genotype was similar to that of the CCL2 (-/-) genotype, characterized by a reduced macrophage infiltration in the spinal cord and a decreased susceptibility to actively (MOG35–55) induced acute EAE in the studies by Fife et al [ 31 ] and Izikson et al [ 32 ]. T cells from CCR2 (-/-) immunized mice produced similar levels of interferon-γ and IL2 as those from controls, and were capable of transferring EAE in a naïve recipient. In contrast, T cells from wild type mice did not cause EAE in a CCR2 (-/-) recipient [ 31 ]. However, these observations again appeared to be model specific. Gaupp et al [ 33 ] reported that, even though the disease was milder or delayed, three CCR2 (-/-) mouse strains retained susceptibility to EAE in their experiments. Histological analyses revealed an abundance of neutrophils in lesions of the CCR2 (-/-) mice in contrast to the monocyte abundance in EAE lesions of wild-type mice. The development of compensatory immune mechanisms for the lack of CCR2 was evidenced by the increased mRNA expression for other CCL and CCR molecules (most notably IL8 and its receptor involved in neutrophil recruitment). This study emphasizes that promiscuity of chemokines and their receptors may overcome the deletion of a single CCR receptor with a resultant mild modification of the clinical and more profound modification of the histological phenotype. Further studies demonstrated an approximately 50% reduction of clinical EAE activity in the CCR1 (-/-) mice, likely involving the altered migration of monocytes and lymphocytes [ 34 ]. In contrast to the observed EAE suppression in the CCR1 (-/-) and CCR2 (-/-) models, the CCR5 knockout mice had the same disease severity as the wild-type controls [ 35 ]. These studies underscore the importance of CCR1 and CCR2 in the development of inflammatory demyelination and give support to novel alternative strategies targeting these CCR molecules. Such strategies include the development of small functional CCR antagonists, amongst which the most significant progress has been made with CCR1 antagonists [ 36 , 37 ]. CCR1 antagonist compounds were shown to inhibit CCL3 and CCL5 induced migration of MNCs in a dose dependent manner, and to reduce clinical EAE in rat [ 36 - 38 ]. In sum, CCL and CCR data from rodent EAE models using inbred, transgenic and knockout strains along with data from chemokine-specific antibody treatments or CCL DNA immunization in EAE suggest that concentration gradients of CCL2 and CCL3 decreasing from the CNS to the peripheral circulation are involved in the spatially and temporally regulated recruitment of mononuclear cells into the CNS which correlates with the course of clinical disease. CCL2 may play a more significant role during relapses than during the induction phase of the disease. CCL5 is expressed by mononuclear cells in the perivascular space during the recovery phase of an acute event, and may therefore be involved in the regulation of recovery rather than in the initiation of the disease. In addition, CCL19 and CCL21 are expressed by endothelial cells of the BBB, and are involved in the strengthening of leukocyte adhesion to inflamed venules followed by homing of encephalitogenic T lymphocytes to the CNS. CCL20 can control the recruitment of dendritic cells into lesions, whereas CCL22 may be involved in a TH2 mediated regulatory process during EAE. CCL1 is likely playing an important role in the autocrine regulation of activation of macrophages and microglia in EAE lesions. Thus, the functional involvement of CCL chemokines during EAE is not only restricted to a well orchestrated recruitment of dendritic cells, monocytes, macrophages, T effector and regulatory cells into the CNS, but also includes a temporal and spatial regulation of TH1 (CCL3, CCL5) or TH2 (CCL2, CCL22) polarization, and monocyte, macrophage and microglial activation (CCL1, CCL2, CCL7, CCL8). Their receptors, the CCRs play equally important roles in these processes. Experimental evidence now suggests that CCR1, CCR2, CCR3, CCR4, CCR5, CCR6, CCR7, CCR8 and CXCR3 on hematogenous mononuclear cells recognize these chemoattractant and regulatory molecules to induce cell differentiation, adhesion or migration of distinct inflammatory cells in peripheral lymphoid organs, at the BBB and in the CNS during the course of EAE. Even taking into consideration the complex and promiscuous nature of the CCL – CCR network, certain pathways may be associated with distinct biological function amenable to intervention. Targeting CCR molecules either by monoclonal antibodies or by small functional antagonists has become a novel and realistic strategy in the treatment and prevention of autoimmune diseases. Multiple sclerosis The complexity of disease pathogenesis, difficulties accessing the site of pathology and the descriptive nature of studies explain why the available CCL / CCR data are less comprehensive in MS as compared to those in EAE. Nevertheless, new observations support the generally accepted views that MS is a predominantly TH1 lymphocyte mediated disease, and CCL – CCR molecules play a significant part in the regulation of intercellular interactions in the peripheral lymphoid organs, at the BBB and in the CNS. In addition to defining chemotaxis, CCL-CCR interactions are involved in TH1 / TH2 polarization and regulation in MS. Recent studies also raise the possibility that distinct molecular mechanisms with characteristic CCL-CCR kinetics correlate with the development of histological subtypes of the disease. CCRs in the multiple sclerosis brain A recent review of chemokines and their receptors [ 39 ] suggests that every CC chemokine receptor (CCR1-CCR5) interact with multiple CCLs and vice versa. Five CCRs (CCR1, CCR2, CCR3, CCR5 and CXCR3) were detected on infiltrating monocytes, macrophages and lymphocytes in MS lesions. In contrast, several members of the CCL family [CCL2 = MCP-1, CCL3 = MIP-1α, CCL4 = MIP-1β, CCL5 = RANTES, CCL7 = MCP-3, CCL8 = MCP-2] were expressed in astrocytes, microglia and other inflammatory cells within MS lesions. While control brain specimens had only scarce appearance of CCR positive (microglial) cells throughout the CNS, foamy macrophages, microglia, perivascular lymphocytes and occasionally, astrocytes were positive for CCR2, CCR3 and CCR5 in chronic active plaques [ 39 , 40 ]. In other studies, CCR1, CCR2, CCR3 and CCR5 were detected on mononuclear cells and macrophages in demyelinating plaques [ 41 , 42 ]. Trebst et al [ 43 ] investigated the kinetics of CCR expression. In early demyelinating lesions, CCR1+/CCR5+ hematogenous monocytes and CCR1-/CCR5- microglial cells were detected. In later stages, macrophages became CCR1-/CCR5+, while microglia upregulated CCR5. This observation suggest that CCR1+/CCR5+ hematogenous monocytes enter into the CNS and stay there in the presence of appropriate ligands. During evolution of lesions, these cells down-regulate CCR1 while retain the CCR5 expression. A more recent study [ 44 ] reveals that this distinct temporal pattern, namely the decrease in CCR1+ and increase in CCR5+ cells, may be restricted to the histological type II demyelinating lesions characterized by mononuclear cell infiltration and immunoglobulin plus complement deposition, and is not seen in type III lesions characterized by oligodendrocytopathy and apoptosis [ 45 ]. CCR2 may also play a key role in the lesion development based on more indirect information. CCR2 is the main, but not exclusive, functional receptor for CCL2 [ 4 ], and as discussed above, the CCL2 – CCR2 interaction appears to play a key role in the development of EAE lesions. Microglia, macrophages and perivascular mononuclear cells show some degrees of immune reactivity for CCR2 in chronic active plaques in several studies, but the expression of CCR2 is generally low in MS lesions. Nevertheless, the data from EAE and observations in MS suggest that the CCR2 – CCL2 interaction is important in the development of plaques. This view was recently proposed and will be discussed below. CCR8, the receptor for CCL1, has been detected in vitro on TH2 and regulatory lymphocytes, macrophages and microglia. Using immunohistochemistry, Trebst et al [ 19 ] detected CCR8 on phagocytic macrophages and activated microglia in type II and type III demyelinating MS lesions. CCR8 expression correlated with the demyelinating activity, but was not restricted to the MS pathology. Phagocytic macrophages and activated microglia in stroke and progressive multifocal leukoencephalopathy also expressed CCR8. Thus, CCR8 seems to identify a subset of activated microglia in different CNS pathologies. CCLs in the multiple sclerosis brain Using methods of immunohistochemistry and in situ hybridization, McManus et al [ 46 ] investigated the expression of three monocyte chemoattractant proteins, MCP-1 (CCL2), MCP-2 (CCL8) and MCP-3 (CCL7) in correlation with the temporal evolution of plaques. All three proteins were detected in high amounts in the center, but sharply decreased at the edges of acute and chronic active lesions. Hypertrophic astrocytes showed the strongest expresion, while infiltrating mononuclear cells showed variable reactivity in plaques. MCP-3 (CCL7) was also detected in the extracellular matrix. Reactivity for these CC chemokine ligands outside of plaques was otherwise restricted to hypertrophic astrocytes. In situ hybridization confirmed the observation for CCL2 at mRNA level. There seemed to be an inverse correlation between the age of plaques and expression of these three CCL molecules, with only a scanty appearance of immunoreactive astrocytes in chronic silent lesions. These methods did not detect MCP chemokines in the brains of normal controls. Additional studies demonstrated the expression of CCL3 and CCL4 in macrophages and microglia, and CCL3 also in astrocytes [ 47 - 49 ]. CCL5 was primarily detected in perivascular inflammatory cells and astrocytes [ 48 - 50 ]. While most of the above studies used the method of immunohistochemistry, we recently assessed the mRNA expression levels for CCL2, CCL3, CCL5, CCL7, CCL8, CCL13 and CCL15 relative to β-actin in corresponding normal appearing white matter (NAWM), normal appearing gray matter (NAGM) and chronic active plaque containing specimens from ten post mortem MS brains. These specimens were characterized by hematoxyllin & eosin, Luxol Fast Blue and immune staining specific for CD68 and β 2 -microglobulin [ 51 ]. In addition, the expression distribution for pro- and anti-apoptotic molecules in these specimens was also assessed by real-time PCR [ 51 ]. The selection of the above listed CC chemokines was based on two considerations. First, we detected MS associated SNP haplotypes in the genes of CCL2, CCL11-CCL8-CCL13, CCL15 and CCL3 [ 8 ]. Second, previous studies suggested the involvement of CCL2, CCL7, CCL8, CCL5 and CCL3 molecules in the development of plaques [ 39 , 46 ]. While neither our genetic nor our mRNA studies revealed positive findings for CCL5, the three MCP chemokines CCL2 (MCP-1), CCL7 (MCP-3) and CCL8 (MCP-2) showed altered regional expressions in MS brains. We detected an increased expression of CCL2 in plaques as compared to NAWMs, and an increased expression of CCL7 in both plaques and NAWMs as compared to NAGMs. In contrast, the expression level of CCL8 was decreased in plaques as compared to NAWM or NAGM specimens (Banisor and Kalman, unpublished observation). This analysis of CCL mRNA molecules in various regions of MS brain complements the data from previous immunohistochemical studies, and further confirms the involvement of CCL2 and CCL7 (and possibly of CCL8) in the development of pathology. In consensus with others, however, we also note an increased CCL7, CCL8 and CCL13 expression in the white matter as compared to the gray matter in 5 other neurological disease controls (1 viral and 1 post-infectious encephalitis, 2 Alzheimer disease and 1 Parkinson disease). No differences were observed for any of these molecules in the white and gray matters of normal controls. We postulate that the expression of CCL molecules may be detected in various inflammatory conditions of the CNS, however, the temporal and cell specific upregulation of certain CCL and CCR molecules is pathology specific. Therefore, further exploration of the expression kinetics of these molecules may facilitate a better understanding of MS pathogenesis. CCL and CCR detected in blood and CSF Relatively limited numbers of studies are available regarding chemokines and their receptors in the blood circulation and in the cerebrospinal fluid (CSF) in MS patients. The expression of CCR5 was found to be higher on circulating T lymphocytes from MS patients than on those from normal controls. These T cells showed an increased migration towards CCL3 and CCL5, suggesting a functional significance of the altered receptor expression [ 42 , 52 ]. The migratory population represented predominantly TH1 / TH0 cells, while the non-migratory population was enriched for TH2 cells. The aberrant migration of T cells towards CCL3 and CCL5 was related to the increased expression of the CCR5 receptor, and could be blocked by anti-CCR5 antibodies. A fluctuation of CCR5 expression by T cells was also suggested in correlation with relapses and remissions in a small group of patients [ 53 ]. Sorensen and Sellebjerg [ 54 ] assessed the CCR expression profile on peripheral T cells of patients with relapse, remission or secondary progressive disease, and detected a higher percentage of CCR2-expressing T cells in secondary progressive MS (SPMS) than in other patient groups. CCR2-positive T cells displayed TH2 profile producing IL5 and tumor necrosis factor-α. The CCR5 expression associated with TH1 profile was significantly lower in SPMS than in patients with relapsing-remitting MS (RRMS) during relapse. Thus, the authors conclude that patients with SPMS have a high expression of CCR2, a chemokine receptor associated with TH2 profile, whereas patients with RRMS preferentially display T cells with CCR5 expression and TH1 profile. More CCR5 positive T cells produced tumor necrosis factor-α in patients with RRMS than those in patients with SPMS. CCR2 is known to be predominantly expressed on monocytes. However, when expressed on T cells, CCR2 is associated with the TH2 subtype as CCL2 induces differentiation of T cells into TH2 phenotype [ 11 ]. While Sorensen and Sellebjerg [ 54 ] detected significant differences in the CCR5 and CCR2 expression profile between RRMS and SPMS, they noted no differences in CCR expression between RRMS and controls. The observation regarding the association of CCR5 with RRMS is consistent with a previous study revealing that patients with the defective CCR5 receptor (CCR5 Δ32 deletion) have prolonged relapse free periods, but the long term prognosis of MS did not seem to correlate with the CCR5 Δ32 genotype [ 55 ]. Besides establishing the CCR characteristics in RRMS and SPMS, this study also suggests that targeting the CCL2-CCR2 axis with specific CCR2 antagonist or a combination of CCR2 and CCR5 antagonists might be an option in SPMS, whereas CCR5 antagonists alone may be considered in RRMS. However, CCR5 on peripheral MNCs was not uniformly found to be differentially expressed in MS subtypes [ 56 ]. MNCs from blood constitutively expressed CCL4 and CCL5, the ligands for CCR5, in all patient groups and controls. This study also failed to detect CCL2 and CCL3 by ribonuclease protection assay in peripheral blood MNCs. Further, the complexity of information regarding CCR5 is reflected by a recent study suggesting the association of the CCR5 Δ32 genotype with early death in MS [ 57 ]. There is relatively limited information available regarding CCR and CCL expression levels in the CSF. Some investigators showed that CCR5+ mononuclear cells of MS patients were enriched in the CSF, representing a significant proportion of monocytes and only a minority of T cells. However, neither cell population differed quantitatively from those of controls, suggesting that CSF leukocytes may not be fully reflective of CNS inflammation [ 39 , 55 ]. Giunti et al [ 58 ] detected CCR5, CCR7 and CXCR3 positive T cells in the CSF of patients with MS and other inflammatory neurological disease (IND) (meningitis, encephalitis, CIDP, neuroborreliosis). Coexpression of these receptors was noted on a subset of memory cells. The increased ratio of CXCR3 / CCR4 was suggested as a molecular correlate of disease activity by Nakajima et al.[ 59 ] TH1 clones established from the CSF of patients with IND and of controls similarly migrated in vitro towards CXCL10, CXCL12 and CCL5. CXCL10, CXCL12 and CCL19 were increased in the CSF of these patients [ 58 ]. Amongst CC chemokines, CCL3 and CCL5 were most consistently found to be elevated in the CSF of MS patients during relapses as compared to normal controls [ 59 - 61 ]. In contrast, decreased CCL2 was found in the CSF in all clinical forms of MS by Scarpini et al.[ 62 ] More consistently, however, low CCL2 levels were detected only during relapses by others [ 41 , 59 , 61 , 63 , 64 ]. The drop of CCL2 in the CSF was not found during relapses of neuromyelitis optica [ 65 ]. Mahad et al [ 64 ] also found that CCL2 in the CSF was decreased not only in patients with MS but also in patients with IND when compared to those of non-inflammatory CNS disease controls. In contrast, Bartosik-Psujek and Stelmasiak [ 61 ] observed an increase in both CCL2 and CCL5 in the CSF of patients with IND, and suggested that the drop of CCL2 during relapses is characteristic only of MS. Further, CCL2 concentration increased as time from the last relapse increased and following corticosteroid therapy [ 63 , 64 ]. With the exception of well defined changes in the CCL2, CCL3 and CCL5 levels in the CSF during relapses, most investigators observed no differences in various clinical forms of the disease [ 56 , 61 , 64 , 66 ]. Pashenkov et al [ 67 ] studied two secondary lymphoid organ chemokines, CCL19 (exodus-3, MIP-3β) and CCL21 (exodus-2, SCL) in CSF and sera of patients with MS, clinically isolated syndrome (CIS) presenting as optic neuritis (ON), isolated ON, IND and non-inflammatory neurological disease controls (NINC). CSF of the NINC group contained CCL19 but not CCL21, while both chemokines were elevated in the CSF of patients with MS, CIS-ON and IND. The authors postulate that CCL19 and CCL21 may control the retention of dendritic cells and the recruitment of naïve T cells and activated B cells, or a de novo formation of lymphoid structures in plaques. These cells are known to express CCR7, the receptor for CCL19 and CCL21. EAE studies also support the notion that CCL19 and CCL21 play important roles both at the BBB and in the CNS [ 3 ]. To correlate previous data on CCL concentrations in the CSF of MS patients, Kivisakk et al [ 68 ] measured mRNA for CCL2 / MCP-1 and CCL5 / RANTES in MNCs in the CSF and blood of patients with MS, acute meningitis and normal controls. While high numbers of MNCs expressing CCL2 and CCL5 were found in some patients, overall no differences were observed between MS and acute meningitis. This study would argue that there is no systemic dysregulation of CC chemokines contributing to MS pathogenesis. In sum, the above data suggest that CCL1, CCL2, CCL3, CCL4, CCL5, CCL7 and CCL8 are expressed by residential glia and perivascular leukocytes in plaques. Expression of the corresponding CCR1, CCR2, CCR3, CCR5 and CCR8 receptors has been demonstrated on infiltrating leukocytes, but also on microglia, dendritic cells and astrocytes. While the expression kinetics of CCR1 and CCR5 may discriminate between histological type II and type III lesions of MS, CCR8 is similarly expressed in both lesions types (Table 1 ). Table 1 CCR and CCL molecules in plaques, blood and CSF of MS patients. This Table summarizes in a cross-sectional manner major findings regarding CCR and CCL expression in brain, blood and CSF of multiple sclerosis patients. The dynamic nature of changes is detailed in the text. The interactions of CCL-CCR molecules on specific cell types are depicted in Figure 1. Type II and III lesions refer to the histological classification proposed by Lucchinetti et al [45]. References are indicated in brackets. In chronic active plaque expressed on In blood expressed on In CSF expressed on CCR1 Monocyte, macrophage, lymphocyte [39,41,42] CCR2 Monocyte, macrophage, lymphocyte [39-42] TH2 in SPMS [54] CCR3 Monocyte, macrophage, lymphocyte [39-42] CCR4 Monocyte, macrophage, lymphocyte [39] CCR5 Monocyte, macrophage, lymphocyte [39-42] TH1/TH0 in RRMS [42, 52-54] MNC [39,55,58] CCR7 T, dendritic [58] CCR8 Macrophage, microglia in type II and III lesions [19] In early -> late stage type II lesion CCR1+/CCR5+ -> CCR1-/CCR5+ Monocyte, macrophage [43] CCR1-/CCR5- -> CCR1-/CCR5+ Microglia [43] In acute, and to lesser degrees, in chronic active plaques expressed by In blood expressed by In CSF CCL2 Astrocyte, microglia, MNC [46] low in relapse [59,61,63,64] CCL3 Astrocyte, microglia, macrophage [47-49] increased in relapse [59-61] CCL4 Microglia, macrophage [47-49] MNC [56] CCL5 MNC, astrocyte [48-50] MNC [56] increased in relapse [59-61] CCL7 Astrocyte, microglia, MNC [46] CCL8 Astrocyte, microglia, MNC [46] CCL19 present in NIND, increased in MS, CIS-ON, IND [67] CCL21 increased in MS, CIS-ON, IND [67] The increase of CCL3 and CCL5 in the CSF during a relapse correlates with the increase in the expression of their receptor, CCR5 on TH1 lymphocytes, which results in an enhanced migratory activity of these cells towards CCL3 and CCL5. The consistently observed decrease in CCL2 levels in the CSF during or even prior to a relapse generated alternative considerations. The first consideration suggests, that the decreased CCL2 level likely relate to a decreased TH2 lymphocyte activity, as CCL2 induces TH2 polarization. Vice versa, CCL2 expression is controlled by TH2 cytokines such as IL4. The concept of CCL2 – TH2 coregulation is supported by the observation that clinical improvement and normalization of the inflammatory CSF profile after corticosteroid treatment correlate with the normalization of CCL2 in the CSF. Thus, measurements of CCL2 in the CSF may also reflect the fluctuation of TH2 activity during the course of MS. The second consideration was proposed by Dr. Ransohoff (oral presentation at the ECTRIMS meeting 2004) [ 4 , 69 ]. This interpretation reconciles the complex observations from the EAE model suggesting a key role for CCR2 – CCL2 in the development of inflammatory lesions, and from MS suggesting a low expression of CCR2 and increased expression of CCL2 in active plaques, but a decreased CCL2 level in the CSF. Based on this model: 1) CCL2 – CCR2 play an important role in the development of inflammatory demyelinating lesions both in EAE and MS; 2) CCL2 expressed in the CNS attracts CCR2+ monocytes and T cells into the developing plaque; 3) while CCR2 binds and internalizes CCL2 molecules in large amounts, CCL2 will be consumed resulting in a reduced CCL2 level in the intercellular fluids and the CSF; 4) when CCR2 encounters its ligand, the CCR2 / CCL2 complex will be internalized and CCR2 will be downregulated on the surface of inflammatory cells in the lesion. Conclusion Studies on EAE and MS suggest that CC chemokine ligands (most prominently CCL2, CCL3, CCL5, CCL7, CCL8, but also CCL1, CCL4, CCL19 and CCL21) expressed by residential immune cells in the CNS or by endothelial cells at the BBB are major chemoattractants for hematogenic immune cells (primarily monocytes / macrophages [CCL2, CCL7, CCL8, CCL22] dendritic cells [CCL19, CCL20, CCL21, CCL22] and T lymphocytes [CCL1, CCL2, CCL3, CCL4, CCL5, CCL19, CCL21, CCL22]) via interactions with their G-protein-coupled receptors (CCR1-CCR10). These CCL – CCR interactions play a key role in the recruitment, activation and retention of immune competent cells in the CNS, with the CCL1 – CCR8, CCL2 – CCR2, CCL3 – CCR1 / CCR5, CCL5 – CCR1 / CCR5, CCL7 – CCR1 / CCR2 / CCR3, CCL8 – CCR3, CCL20 – CCR6, CCL19 / CCL21 – CCR7, CCL22 – CCR4 interactions being the best characterized among them (Figure 1 ). The EAE model suggests that CCL19 and CCL21 produced by endothelial cells induce G-protein-mediated signaling via their receptor CCR7. This signaling leads to an enhanced adhesion of the leukocyte α4-integrin (VLA-4) to the endothelial VCAM-1 and results in a facilitated transmigration of leukocytes via the BBB. CCL-CCR interactions also define the differentiation and chemotaxis of T cell subpopulations, and thus may control the dynamic changes in the local balance of TH1 (CCL3 – CCR1 / CCR5, CCL5 – CCR1 / CCR5) and TH2 (CCL1 – CCR8, CCL2 – CCR2, CCL22 – CCR4) cell populations in lesion. Different CCL – CCR expression kinetics may characterize the different (initial, height, self-limiting) phases and histological subtypes (type II or type III) of inflammatory demyelination. This differential involvement of chemokines and their receptors in various stages and forms of MS, and the arising information concerning the involvement of genetic variants of CCLs suggest that small CCR antagonists may represent a realistic strategy in controlling the inflammatory activity that may have to be adjusted to individual disease characteristics. Figure 1 Interaction between CCL and CCR molecules at the blood-brain barrier. This figure depicts CCL-CCR interactions at the BBB (endothelial cells and astrocytic processes) interfacing a venule and the CNS. CCL molecules (most prominently CCL2, CCL3, CCL7 and CCL8, but also CCL1, CCL4, CCL19 and CCL21) are produced by residential microglia, astrocytes and endothelial cells throughout the course of lesion development, and by infiltrating MNCs (CCL5) during late phases of plaque formation, and attract functionally different subsets of monocytes / macrophages, dendritic cells and T lymphocytes from the circulation via the BBB into the CNS. The temporal and spatial regulation of molecular events, the association of distinct CCR molecules with different histological subtypes of demyelination and the involvement of different CCL-CCR interactions in T cell polarization are detailed in the text. Here we illustrate in a simplified and cross-sectional manner the main groups of interacting receptors on various hematogenous cells and ligands released by residential immune cells of the CNS or by components of the BBB. Group A of receptors and ligands expressed by and acting on monocytes / macrophages, respectively: CCR1 / CCR2 / CCR3-CCL7, CCR2-CCL2, CCR3-CCL8, CCR4-CCL22; Group B of receptors and ligands expressed by and acting on dendritic cells, respectively: CCR4-CCL22, CCR6-CCL20, CCR7-CCL19 / CCL21; Group C of receptors and ligands expressed by and acting on T lymphocyes, respectively: CCR1-CCL3 / CCL5, CCR2-CCL2, CCR4-CCL22, CCR5-CCL3 / CCL4 / CCL5, CCR7-CCL19 / CCL21, CCR8-CCL1. List of abbreviations BBB – blood brain barrier CCL – CC chemokine ligand CCR – CC chemokine receptor CIS – clinically isolated syndrome CNS – central nervous system CR-EAE – chronic-relapsing EAE CSF – cerebrospinal fluid EAE – experimental allergic encephalomyelitis EBV – Epstein-Barr virus IL – interleukin IND – inflammatory neurological disease LPS – Lipopolysaccharide MBP – myelin basic protein MCP – monocyte chemotactic protein MIP – macrophage inflammatory protein MNC – mononuclear cells MOG – myelin-oligodendrocyte glycoprotein MS – multiple sclerosis NAGM – normal appearing gray matter NAWM – normal appearing white matter NF1 – neurofibromatosis, type I NINC – non-inflammatory neurological disease controls NK cells – natural killer cells NOS2A – nitric oxide synthase 2A NPL – nonparametric linkage OMG – oligodendrocyte-myelin glycoprotein ON – optic neuritis PCR – polymerase chain reaction PLP – proteolipid lipoprotein QTL – quantitative trait loci RANTES – regulated upon activation normally T expressed and secreted cytokine RRMS – relapsing-remitting MS SNP – single nuclear polymorphism SPMS – secondary-progressive MS TH1 – T-helper 1 TH2 – T-helper 2 TNF – tumor necrosis factor VCAM-1 – vascular cell adhesion molecule-1 VLA-4 – very late antigen-4 Competing interests The author(s) declare that they have no competing interests. Authors' contributions Ileana Banisor, research assistant, was involved in the acquisition and analyses of our research data mentioned in the paper. She prepared the figure. Thomas P. Leist, collaborator, critically reviewed and edited the manuscript. Bernadette Kalman, P.I., designed the research studies mentioned from her lab, supervised the work processes, interpreted the data and drafted this manuscript. She also generated funding supports.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554759.xml
509304
p19 Arf Suppresses Growth, Progression, and Metastasis of Hras-Driven Carcinomas through p53-Dependent and -Independent Pathways
Ectopic expression of oncogenes such as Ras induces expression of p19 Arf , which, in turn, activates p53 and growth arrest. Here, we used a multistage model of squamous cell carcinoma development to investigate the functional interactions between Ras, p19 Arf , and p53 during tumor progression in the mouse. Skin tumors were induced in wild-type, p19 Arf -deficient, and p53 -deficient mice using the DMBA/TPA two-step protocol. Activating mutations in Hras were detected in all papillomas and carcinomas examined, regardless of genotype. Relative to wild-type mice, the growth rate of papillomas was greater in p19 Arf - deficient mice, and reduced in p53 -deficient mice. Malignant conversion of papillomas to squamous cell carcinomas, as well as metastasis to lymph nodes and lungs, was markedly accelerated in both p19 Arf - and p53 -deficient mice. Thus, p19 Arf inhibits the growth rate of tumors in a p53-independent manner. Through its regulation of p53, p19 Arf also suppresses malignant conversion and metastasis. p53 expression was upregulated in papillomas from wild-type but not p19 Arf -null mice, and p53 mutations were more frequently seen in wild-type than in p19 Arf -null carcinomas. This indicates that selection for p53 mutations is a direct result of signaling from the initiating oncogenic lesion, Hras, acting through p19 Arf .
Introduction Tumor development and metastasis is a multistep process of somatic cell evolution that includes uncontrolled proliferation, impaired apoptosis, loss of differentiation, immortalization, neovascularization, invasion, and metastatic spread ( Hanahan and Weinberg 2000 ). This evolutionary transformation can be operationally divided into distinct stages, including initiation, promotion, progression, and metastasis ( DiGiovanni 1992 ). Mutations in both oncogenes and tumor suppressor genes are found in end-stage tumors, implying their causal role in tumor development. However, the association of mutations in specific genes with specific phenotypic states during tumor progression is poorly characterized for most human solid tumors. It is also largely unknown whether each mutation is an independent event or whether there is a preferred sequence or combination of mutations that is favored. The purpose of this study is to investigate the functional interactions between the mutational activation of the oncogene Ras, and two tumor suppressors, p19 Arf , and p53, using a multistage epithelial tumor model. Ras is among the most frequently mutated oncogenes in human cancer, with approximately 30% of tumors carrying an activating mutation in one of three family members, Hras, Kras, or Nras ( Bos 1989 ). Cancer-associated mutations in Ras result in constitutively active Ras protein. Ras is a nodal signaling molecule that regulates multiple signaling pathways, leading to profound changes in cellular proliferation, apoptosis, differentiation, senescence, cytoskeletal organization, adhesion, and migration ( Campbell et al. 1998 ). Ras has also been shown to induce invasiveness and metastasis of cancer cells ( Pozzatti et al. 1986 ; Webb et al. 1998 ; Varghese et al. 2002 ). These pleiotropic effects suggest Ras may influence multiple steps in tumor progression. p53 is the most frequently mutated tumor suppressor gene in human cancer, with more than 50% of tumors showing mutations ( Hollstein et al. 1994 ). p53 is a nodal signaling protein that coordinates the cellular response to different types of stress, including oncogene activation, DNA damage, abnormal cell adhesion, altered ribonucleotide pools, hypoxia, and redox stress ( Ko and Prives 1996 ; Giaccia and Kastan 1998 ). These stress stimuli are thought to activate p53 by inducing posttranslational modifications that stabilize p53 and enhance its ability to act as a transcription factor ( Siognov and Haupt 1999 ; Vousden and Lu 2002 ). Loss of p53 function leads to loss of cell cycle checkpoints, impaired apoptosis, genomic instability, and tumor progression. However, a major unresolved issue is, of the many signals that have been shown to activate p53 using a variety of model systems, which one regulates p53 during autochthonous tumor progression. A mechanistic connection between Ras signaling and activation of p53 that involves the tumor suppressor p19 Arf was recently established. p19 Arf (p14 Arf in humans; Stott et al. 1998 ) is encoded by the p16 Ink4a /p19 Arf locus, but because it is transcribed in an alternative reading frame, the protein product is unrelated to the p16 Ink4a protein ( Quelle et al. 1995 ; Kamijo et al. 1997 ). Deletions or mutations at the p16 Ink4a /p19 Arf locus are frequently (more than 50% of cases) seen in human tumors ( Ruas and Peters 1998 ). p19 Arf was established as a bona fide tumor suppressor in studies showing that mice lacking p19 Arf are highly susceptible to spontaneous tumorigenesis ( Kamijo et al. 1997 ). In vitro studies had shown that enforced expression of oncogenes such as Ras, c-Myc, and E1A activated p53 and induced growth arrest, senescence, or apoptosis depending on the cell type or oncogene used ( Lowe and Ruley 1993 ; Hermeking and Eick 1994 ; Serrano et al. 1997 ). These cellular responses were impaired in cells lacking p53, indicating that functional p53 was required. The involvement of p19 Arf was first suggested by experiments showing that enforced expression of Ras, Myc, and E1A in cells induced p19 Arf , leading to G1 and G2 cell cycle arrest that was p53-dependent ( Kamijo et al. 1997 ; Stott et al. 1998 ). Cells lacking p19 Arf showed impaired p53 induction in response to these oncogenes, and, like p53-deficient cells, escaped growth arrest and were immortalized ( Zindy et al. 1998 ; de Stanchina et al. 1998 ; Palmero et al. 1998 ; Lin and Lowe 2001 ; Ferbeyre et al. 2002 ). In vivo evidence linking oncogene signaling to p19 Arf and p53 was obtained in a lymphoma model. B-cell lymphomas from transgenic Eμ-myc mice also show a dependence on p19 Arf to activate p53, and Eμ-myc mice lacking either p19 Arf or p53 developed lymphomas much faster ( Eischen et al. 1999 ; Schmitt et al. 1999 ). p19 Arf regulates p53 through mutual binding to the p53 regulator Mdm2. The levels of p53 in cells are normally kept low because of feedback regulation by the Mdm2 protein ( Haupt et al. 1997 ; Kubbutat et al. 1997 ). Mdm2 binds to p53 and targets it for degradation by nuclear to cytoplasmic shuttling and through the E3 ubiquitin ligase activity of Mdm2 ( Roth et al. 1998 ; Honda and Yasuda 1999 ). p19 Arf sequesters Mdm2 from p53 and inhibits the ubiquitin ligase activity of Mdm2, resulting in increased stability and accumulation of p53 ( Pomerantz et al. 1998 ; Tao and Levine 1999 ; Weber et al. 1999 ; Zhang and Xiong 1999 ). The importance of the Ras-p19 Arf -p53 pathway in growth arrest was established in culture systems involving the ectopic overexpression of mutant Ras in both murine embryonic fibroblasts ( Ferbeyre et al. 2002 ; Palmero et al. 1998 ) and primary epidermal keratinocytes ( Lin and Lowe 2001 ). Both in vitro and in vivo models have established that the gene dosage of mutant Ras is critical for its oncogenic function. A single copy of mutant Ras is insufficient to transform cells; at least two mutant alleles are required ( Finney and Bishop 1993 ). Duplication or even amplification of mutant Ras alleles is frequently observed in tumors ( Quintanilla et al. 1986 ; Bremner and Balmain 1990 ). Ras activates multiple signaling pathways, and quantitative differences in Ras activity can lead to activation of different signals and qualitatively different cellular phenotypes ( Shields et al. 2000 ). Cell culture conditions can add additional stress signals that are known to impinge on p19 Arf and p53, leading to cell cycle arrest ( Sherr and DePinho 2000 ; Lowe and Sherr 2003 ), and cannot recapitulate the complex cellular ecology of tumor progression. Thus, to understand the interactions between Ras, p19 Arf , and p53 that drive tumor progression, an autochthonous tumor model is required. Since more than 90% of human cancers are epithelial in origin, a carcinoma model system is favored. Mouse skin carcinogenesis is perhaps the best-characterized in vivo model of epithelial neoplasia and was instrumental in establishing the concepts of initiation, promotion, and progression ( DiGiovanni 1992 ). The two-stage chemical protocol involves treatment of mice with a carcinogen, DMBA, followed by multiple applications of TPA. This treatment induces benign squamous cell papillomas, nearly 100% of which have sustained an AT mutation in codon 61 of Hras ( Quintanilla et al. 1986 ). As this mutation results in constitutively active Ras protein, this protocol is ideal to study the biological consequence of Ras activation during the entire natural history of tumor progression. Papillomas consist of a series of folded epidermal or follicular hyperplasias that protrude from the skin surface. Papillomas have dysplastic characteristics including disturbed cell polarity, basal cell hyperplasia, disturbed maturational sequence, increased mitotic activity, and increased nuclear to cytoplasmic ratio ( Yuspa 1994 ). In most strains of mice, progression of these benign papillomas to malignant squamous cell carcinomas (SCCs) is a rare and late event. SCCs are usually endophytic tumors that present as plaques with an ulcerated surface. These tumors break through the basement membrane and progressively invade the underlying dermis and subcutaneous tissues, and rarely, can metastasize to regional and distant sites. SCCs are characterized by a disorderly proliferation of epithelial cells with increased cellular atypia and abnormal mitotic figures, and are classified into four grades: well-differentiated, moderately differentiated, poorly differentiated, and spindle cell carcinoma. The unique advantage of this skin tumor model is the ability to directly observe and quantify these evolutionary stages. Loss of p53 function is strongly associated with the benign to malignant transition of chemically induced SCCs. Mutations in p53 are seen more frequently in carcinomas than in papillomas ( Burns et al. 1991 ; Ruggeri et al. 1991 ). p53 knockout mice show accelerated malignant progression of SCCs ( Kemp et al. 1993 ). The strongest association of p53 with malignant progression was revealed in p53 +/− mice, in which loss of the remaining wild-type allele of p53 was seen in carcinomas but not papillomas, indicating a strong selective pressure to completely inactivate p53 during this transition. Accelerated malignant progression seen in the absence of p53 was accompanied by extensive loss of differentiation and lymph node metastasis, indicating that p53 inhibits multiple steps involved in malignant tumor progression. Here, we used the mouse skin tumor model to examine the role of p19 Arf in regulating the levels and tumor suppressor activity of p53. In addition, we addressed the biological and functional significance of alterations in p19 Arf , p53, or both during tumor initiation, promotion, progression, and metastasis. Similar to p53 -null mice, loss of p19 Arf resulted in increased malignant conversion, more aggressive tumors, and frequent and rapid metastasis. However, in contrast to p53 -null mice, p19 Arf -null mice had greater tumor numbers and tumor growth rates, indicating additional, p53-independent tumor suppressor functions for p19 Arf . Results Increased Papilloma Number and Size in p19 Arf -Deficient Mice Both in vitro and in vivo studies have demonstrated that p19 Arf is a tumor suppressor ( Eischen et al. 1999 ; Kamijo et al. 1999 ; Schmitt et al. 1999 ; Lin and Lowe 2001 ). However, other than regulation of p53, little else is known about the role of p19 Arf in tumor suppression. To address this, groups of p19 Arf+/+ , p19 Arf +/− , and p19 Arf− /− littermates were treated with a single dose of DMBA followed by twice weekly application of TPA for 15 wk (see Materials and Methods ). Papillomas began to appear in all three genotypes after 9 wk of promotion. By 20 wk, p19 Arf+/− and p19 Arf−/− mice showed a significant increase in papilloma number and size compared to wild-type mice ( Figure 1 A). Relative to wild-type littermates, p19 Arf−/− mice had an average of 2.97 more papillomas (95% CI (0.70, 5.24); p = 0.010) and p19 Arf+/− mice had an average of 2.60 more papillomas (95% CI (0.05, 5.14); p = 0.045) in weeks 18–30 after DMBA administration. Average papilloma size was also greater in both p19 Arf−/− and p19 Arf+/− mice compared to wild-type mice ( Figure 1 B). This effect was seen as early as 12 wk and increased through time so that by 28 wk, 33% (47/141) of papillomas from p19 Arf−/− mice were greater than 8 mm in diameter versus 14% (38/267) from wild-type mice ( p < 0.0001). Papillomas from p19 Arf -deficient mice measured up to 16 mm in diameter while very few papillomas on wild-type mice measured more than 9 mm. Thus, p19 Arf deficiency resulted in faster growing papillomas, indicating a role for p19 Arf in regulating the early stages of benign tumor growth. Figure 1 Skin Tumor Multiplicity, Size, and Progression in p19 Arf -Deficient Mice (A) Average number of papillomas (more than 2 mm in diameter) per mouse is plotted versus the number of weeks postinitiation. Both p19 Arf (Arf) +/− and p19 Arf−/− mice show greater numbers of tumors than p19 Arf+/+ mice. (B) Comparison of papilloma size (in mm) between p19 Arf+/+ , p19 Arf+/− , and p19 Arf−/− mice through 28 wk postinitiation. An increase in the largest size class of tumors is seen in p19 Arf+/− and p19 Arf−/− mice but not p19 Arf+/+ mice . (C) Percentage of mice bearing at least one carcinoma is plotted versus the number of weeks postinitiation . p19 Arf−/− mice show the shortest latency and greatest incidence of carcinoma conversion, with p19 Arf+/− mice showing an incidence between the p19 Arf−/− and p19 Arf+/+ mice. Time of appearance of lymph node metastasis is noted above the graph as a vertical line for each mouse analyzed. Metastasis to lymph node occurred frequently and sooner in p19 Arf -deficient mice than in wild-type mice. Mutations in Hras are found in more than 95% of DMBA/TPA-induced skin tumors ( Quintanilla et al. 1986 ; Kemp et al. 1993 ) . In vitro studies showed that mutant Ras induces p19 Arf , which, in turn, inhibits Ras-induced proliferation ( Sherr and Weber 2000 ). Thus, loss of p19 Arf might reduce or eliminate the need to mutate Ras . All papillomas from p19 Arf+/+ (5/5), p19 Arf+/− (4/4), and p19 Arf−/− (5/5) mice contained the identical A→T transversion at codon 61 of Hras, resulting in an amino acid change from glutamine to leucine and a constitutively activated Ras protein ( Quintanilla et al. 1986 ). Thus, mutation of Ras is very strongly selected for during epithelial carcinogenesis, with or without the presence of p19 Arf , and loss of p19 Arf cooperates with activated Ras to accelerate tumor growth. Increased Malignant Progression and Metastasis in p19 Arf -Deficient Mice The rate of malignant conversion of papillomas to carcinomas is greatly increased in the absence of p53 function ( Kemp et al. 1993 ). To determine if loss of p19 Arf had a similar effect, progression was quantified by visual inspection and confirmed by histologic analysis. The rate of conversion from papillomas to carcinomas was dramatically accelerated in p19 Arf -deficient mice. Carcinomas developed in p19 Arf−/− mice as early as 14 wk after initiation, whereas papillomas from p19 Arf+/+ mice began to convert much later, after 22 wk ( Figure 1 C). By 28 wk, 100% of the p19 Arf−/− mice had at least one carcinoma, compared to only 25% of the wild-type mice. The p19 Arf+/− mice showed an intermediate conversion rate, with 60% of the mice bearing at least one carcinoma, indicating an p19 Arf gene dosage effect on malignant progression. In addition to reducing the latency, p19 Arf deficiency increased the frequency of malignant conversion. The odds of developing a carcinoma within 30 wk after DMBA administration was 8.10 times higher for p19 Arf−/− mice (95% CI (1.90, 34.56); p = 0.005) and 3.11 times higher for p19 Arf+/− mice (95% CI (0.90, 10.77); p = 0.073) compared to wild-type mice. The reduced carcinoma latency and increased conversion frequency in the p19 Arf -null mice implicate loss of p19 Arf as a critical rate-limiting step in malignant SCC progression. Histologic analysis revealed that the carcinomas from control mice ranged in grade from well-differentiated to poorly differentiated SCCs. Carcinomas from p19 Arf+/− and p19 Arf−/− mice also showed a range of grades but a significant number (9/12) were characterized as spindle cell carcinomas. These were characterized by packed and spindle-shaped cells with elongated pleiomorphic nuclei and abundant abnormal mitotic figures. These cells grew in a homogenous pattern with very little evidence of the cellular organization typical of low-grade tumors. These tumors showed focal areas of squamous differentiation, indicating that they were derived from squamous epithelium. p19 Arf deficiency also increased dissemination and establishment of metastatic SCCs. Carcinoma-bearing p19 Arf -deficient mice frequently presented with enlarged lymph nodes, and in several cases tumors were noted on the lungs. ( Table 1 ; Figure 2 ). Histologic analysis revealed that these lymph nodes and lung tumors contained cells with features similar to the primary SCCs, including squamous differentiation, keratin pearls, high mitotic index, nuclear pleomorphism, and disturbed cell polarity ( Figure 2 D and 2 E). Immunostaining with a keratin-specific antibody showed that these cells were epithelial in origin, confirming that they were metastatic SCCs ( Figure 2 F and 2 G). 60% of local enlarged lymph nodes from carcinoma-bearing p19 Arf−/− mice contained such squamous carcinoma deposits, compared to 10% of those from wild-type mice ( Table 1 ). Metastatic lesions from p19 Arf -deficient mice were seen as early as 16 wk after initiation and must have occurred very soon after or simultaneously with papilloma to carcinoma conversion (see top of Figure 1 C). In contrast, only one metastatic lesion was seen in one p19 Arf+/+ mice through 36 wk of observation. p19 Arf+ /− mice displayed an intermediate frequency of metastasis. Newly formed blood vessels, some measuring up to 2 mm in diameter, were seen on the underside of each tumor and appeared to lead directly to the inguinal or brachial lymph node ( Figure 2 A and 2 B). Several primary p19 Arf -deficient carcinomas showed clear evidence of penetration of tumor cells through blood vessel walls, with intravascular rafts of tumor cells seen ( Figure 2 C), indicating a route by which tumor cells could migrate to distant organs through the circulation. Thus, in addition to increasing benign tumor growth, loss of p19 Arf accelerated both benign to malignant conversion and metastatic spread of epithelial tumors. Tumors lacking p19 Arf have a higher potential for metastatic spread. Figure 2 Metastasis of Primary SCC to Lymph Nodes and Lungs in p19 Arf -Deficient Mice (A) Underside of skin from tumor-bearing mouse shows newly formed blood vessels surrounding tumor site (arrow) and leading to inguinal lymph node (arrowhead). (B) Enlarged inquinal lymph node (left) containing metastatic SCC and blood vessel formation (arrow) compared to normal lymph node (right). (C) H&E stain of carcinoma section with prominent blood vessel (bv). Carcinoma cells (ca) have penetrated blood vessel wall (arrow). (D) H&E stain of lymph node bearing infiltrating SCC cells (arrow) among normal lymphocytes (arrowhead). (E) H&E stain of lymph node bearing metastatic differentiated SCC. (F) Immunostain with pan-keratin antibody of papilloma. (G) Immunostain with pan-keratin antibody of lymph node with metastatic SCC. (H and I) H&E stain of normal lung (arrowhead) with large metastatic SCC deposit (arrow). (J) H&E stain of lung metastasis with secondary site of infiltration (arrow). (D–G, J): 20× magnification. Inserts in (E–G): 40× magnification. Table 1 Metastatic Frequency of SCC Denominator indicates number of tissues examined. Numerator indicates number of tissues bearing SCC Reduced p53 Expression in Papillomas from p19 Arf -Deficient Mice The p53 expression in DMBA/TPA-induced papillomas is increased relative to adjacent normal skin ( Kemp et al. 2001 ). As multiple signals can lead to the accumulation of p53, including activated oncogenes, DNA damage, and hypoxia, it was not clear which was operative in this setting. As nearly all carcinogen-induced papillomas carry mutations in Hras, we questioned whether increased p53 expression was due to signaling from Ras through p19 Arf . Western blot analysis of nuclear lysates showed increased levels of both p19 Arf and p53 in wild-type papillomas compared to normal skin ( Figures 3 and 4 ). In contrast, p53 expression was not detectable in papillomas from p19 Arf− /− mice and was intermediate and variable in papillomas from p19 Arf+/− mice. Immunostaining of paraffin-embedded sections confirmed the Western analysis, with nuclear staining of p53 detected in the epidermal cells of papillomas from wild-type mice, reduced numbers of p53-positive cells in the p19 Arf heterozygous papillomas, and undetectable p53 staining in p19 Arf -null papillomas ( Figure 3 B). To determine if p53 could still be induced in the absence of p19 Arf by an alternative pathway, tumor-bearing mice were irradiated with 4Gy ionizing radiation and sacrificed 4 h later, and their tissues were examined for p53 expression. Both Western blot analysis and immunostaining revealed prominent induction of p53 in basal cells of normal skin and papillomas from both wild-type and p19 Arf -null mice ( Figure 3 A and 3 B). Thus, the induction of p53 seen in mutant Ras-containing tumors is due to signaling through p19 Arf . These results provide in vivo confirmation of the model, largely derived from in vitro studies, that posits that signaling from mutant Ras acts through p19 Arf to induce p53. Other pathways to activate p53, such as those initiated by DNA damage, remain functional in the absence of p19 Arf . Figure 3 Reduced p53 Expression in Skin Tumors from p19 Arf -Deficient Mice (A) Western blot analysis of nuclear lysates from skin tumors from p19 Arf (Arf) +/+ , p19 Arf+/− , and p19 Arf−/− mice using p53-specific antibody. PA, papilloma; skin IR, irradiated normal skin (B) p53 immunostain of paraffin-embedded skin tumor sections from p19 Arf+/+ , p19 Arf+/− , and p19 Arf−/− mice (arrows indicate positive stained cells) (top). p53 immunostain of irradiated papillomas (IR) from p19 Arf+/+ and p19 Arf−/− mice (bottom). p53 is not detected in normal skin or tumors from p19 Arf−/− mice, but is induced by irradiation in both normal and tumor cells from p19 Arf−/− mice. Figure 4 LOH of Wild-Type p19 Arf Allele in p19 Arf+/− Tumors (A) LOH analysis by semiquantitative PCR of the wild-type p19 Arf allele in p19 Arf+/− papillomas and carcinomas. Gradient made from kidney DNA used for quantitation of wt/mu ratio (top row). wt, wild-type allele; mu, knockout allele; asterisk, loss or reduction of p19 Arf wild-type band. (B) Western blot analysis of nuclear lysates from papillomas (PA) and carcinomas (CA) from p19 Arf+/+ , p19 Arf+/− , and p19 Arf−/− mice. Loss of the Wild-Type p19 Arf Allele in Tumors from p19 Arf+/− Mice Occurs During Benign to Malignant Conversion p19 Arf+/− mice displayed an intermediate rate of papilloma to carcinoma conversion (see Figure 1 C). Two genetic models could explain this heterozygous phenotype. p19 Arf could be haploinsufficient, in which case no mutation or loss of heterozygosity (LOH) should be seen in the remaining wild-type p19 Arf allele in carcinomas. Alternatively, p19 Arf could be recessive, in which case LOH or reduction to a homozygous null state would be expected. The fate of the wild-type allele of p19 Arf in tumors from heterozygous mice was assessed by semiquantitative PCR analysis of genomic DNA. Three of 15 (20%) papillomas examined showed evidence of loss of the wild-type p19 Arf allele, compared to ten of 15 (67%) carcinomas ( p = 0.0027) ( Figure 4 A), indicating LOH occurs primarily during malignant conversion. We next examined p19 Arf expression in tumor lysates by Western blot analysis with a p19 Arf -specific antibody. In wild-type mice, p19 Arf protein was elevated in all papillomas and three of six carcinomas compared to normal skin ( Figure 4 B). Increased expression of p19 Arf is consistent with activation of Ras in these tumors. p19 Arf expression was also increased in papillomas from p19 Arf+/− mice but not to the levels seen in wild-type mice, indicating that p19 Arf protein levels in tumors reflect p19 Arf gene dosage. p19 Arf protein was reduced or undetectable in four of six carcinomas from p19 Arf+/− mice, consistent with the LOH data. Collectively, these data indicate that p19 Arf expression is induced in tumors. Germline deletion of one p19 Arf allele provides a selective advantage during early tumor growth, and loss of the second allele confers an additional phenotype, destabilization of p53, and enhanced malignant progression. Independent Contributions of p19 Arf and p53 to Tumorigenesis The observations that p19 Arf and p53 were upregulated in papillomas, that p53 expression was reduced in p19 Ar f -null papillomas, and that loss of p19 Arf had a similar effect on tumor progression as that of loss of p53, provide strong in vivo support of the model whereby p19 Arf regulates p53 in response to mutational activation of Hras . However, enhanced tumor growth in p19 Arf -null mice, in contrast to reduced tumor growth in p53 -null mice ( Kemp et al. 1993 ), suggests additional tumor suppressor functions of p19 Arf , independent of p53. To examine the effect of the combined loss of p53 and p19 Arf tumor suppressors, skin tumors were induced in p19 Arf and p53 single and compound mutant littermates. Relative to wild-type mice, p53 -null mice developed fewer tumors, averaging 4.05 fewer papillomas (95% CI (−6.10, −2.00); p = 0.0001) 10–16 wk after the DMBA administration, while p19 Arf -null mice averaged 2.68 more papillomas (95% CI (0.52, 4.84); p = 0.015) 18–40 wk after the DMBA administration ( Figure 5 A). The p19 Arf /p53 double-null mice showed a papilloma multiplicity similar to wild-type mice. p53 −/− tumors were also smaller, while both p19 Arf−/− and p19 Arf−/− p53 −/− tumors were larger compared to wild-type tumors. p19 Arf and p53 also affected tumor size and morphology. Wild-type papillomas were highly exophytic, while tumors from both p19 Arf - and p53- deficient mice grew in a flatter, endophytic pattern ( Figure 5 B). Thus, loss of p19 Arf increased the number and size of both wild-type p53 and p53 -null tumors, demonstrating that p53 and p19 Arf contribute independently to the early stages of tumor development. Figure 5 Tumor Multiplicity and Proliferative Index in p19 Arf /p53 Compound Mutant Mice (A) Average number of papillomas (more than 2 mm in diameter) per mouse is plotted against the number of weeks post-initiation. (B) Image of wild-type, p19 Arf (Arf) −/− , p53 −/− , and p19 Arf−/− p53 −/− mice with skin tumors at time of sacrifice. Wild-type mice show large exophytic tumors, while both p19 Arf - and p53-deficient mice have endophytic tumors. Note larger tumors in p19 Arf /p53 compound mutant mice relative to p53 single mutants. (C) BrdU-positive cells in papillomas from wild-type , p53 −/− , p19 Arf−/− , and p19 Arf−/− p53 −/− mice at 10 wk postinitiation. (Bars represent average counts ± standard deviation from ten fields and five mice). p53 −/− tumors show significantly fewer BrdU-positive cells than either p19 Arf−/− or wild-type tumors ( p < 0.05, Wilcoxon one-sided t-test). To determine whether tumor growth in mice lacking p19 Arf or p53 is due to altered proliferation, additional cohorts of mice were treated with DMBA/TPA. Tumor-bearing mice 8–10 wk post DMBA treatment were injected with BrdU and sacrificed 1 h later. p53 -null papillomas showed a reduced BrdU labeling index compared to wild-type mice, while the p19 Arf -null tumors showed a higher BrdU labeling index ( Figure 5 C). The p19 Arf /p53 double-null papillomas showed a proliferative index similar to that of wild-type tumors. Thus, decreased proliferation contributes to the reduced tumor growth seen in the p53 -null mice. Apoptotic cells were very rare in papillomas regardless of p53 genotype or radiation exposure (apoptotic index over 40-fold less than proliferation index) (unpublished data). Thus, p53-regulated apoptosis does not appear to play a major role in SCC development, at least at the papilloma stage. Tumor Progression in p19 Arf /p53 Compound Mutant Mice To determine if p19 Arf and p53 cooperate during tumor progression, papilloma to carcinoma conversion was evaluated in p19 Arf−/− , p53 −/− , and p19 Arf−/− p53 −/− littermates. Tumor progression was accelerated in all p19 Arf - and p53 -deficient genotypes compared to wild-type littermates ( Figure 6 A). Carcinoma latency and multiplicity was almost identical for p19 Arf−/− mice regardless of p53 genotype (p53 +/+ , p53 +/− , or p53 −/− ) ( Figure 6 A), indicating no cooperation between p19 Arf and p53 for malignant conversion per se. However, the size of carcinomas in both p19 Arf−/− and p19 Arf−/− p53 −/− mice was considerably greater than that seen in p53 −/− mice at comparable time points (see Figure 5 B). This confirms a significant impact of p19 Arf on suppressing tumor growth that does not require p53. Figure 6 Tumor Progression and p53 LOH in p19 Arf /p53 Compound Mutant Mice (A) Average number of carcinomas per mouse is plotted against the number of weeks postinitiation. Tumor progression was accelerated in all p19 Arf (Arf)- and p53 -deficient genotypes compared to wild-type littermates . Carcinoma latency and multiplicity was almost identical for p19 Arf−/− mice regardless of p53 genotype ( p53 +/+ , p53 +/− , or p53 −/− ). (B) LOH of the wild-type p53 allele by semiquantitative PCR in p19 Arf /p53 compound tumors. Gradient made from kidney DNA used for quantitation of wt/mu ratio (top row). wt, wild-type allele; mu, knockout allele; asterisk, tumors with reduction of wild-type p53. In p53 +/− mice there is strong selective pressure to lose the wild-type allele during conversion to malignancy ( Kemp et al. 1993 ). As p19 Arf regulates p53, we next wished to determine if the selective pressure to lose p53 was reduced in the absence of p19 Arf . Tumors from p53 +/− mice of all three p19 Arf genotypes ( p19 Arf+/+ , p19 Arf+/− , and p19 Arf−/− ) were assessed for LOH of p53 by semiquantitative PCR analysis of genomic DNA. Two out of seven papillomas from p19 Arf+/+ p53 +/− mice show loss of the remaining p53 allele, while all papillomas examined from p19 Arf+/− p53 +/− ( n = 8) and p19 Arf−/− p53 +/− ( n = 8) mice show retention of wild-type p53 ( Figure 6 B). Seven of 14 (50%) carcinomas from p19 Arf+/+ p53 +/− , and three of 14 (21%) from p19 Arf+/− p53 +/− , but only one of 12 (8%) from p19 Arf−/− p53 +/− ( p = 0.036) showed loss of p53 . Thus, deletion of p19 Arf disrupts the activation of p53 and thereby reduces selection for mutations in p53 during malignant progression. Discussion Using a multistage model of tumor progression, we have examined the functional interactions between the oncogene Hras and the tumor suppressors p19 Arf and p53 . Somatic mutation of Ras is an early and frequent event in this model of tumor development. Against this backdrop, p19 Arf has at least two distinct tumor suppressor properties, which act at different stages of tumor development and which show a range of gene dosage effects. Loss of one or both p19 Arf alleles leads to accelerated growth of benign tumors, indicating p19 Arf is partially haploinsufficient for suppression of this early growth phenotype. Although p19 Arf regulates p53 at this stage, suppression of tumor growth per se by p19 Arf does not appear to be mediated through p53. p19 Arf also inhibits the benign to malignant transition and subsequent tumor cell dissemination and metastasis, and this effect of p19 Arf is, in contrast, mediated through p53. LOH of p19 Arf occurs preferentially in malignant tumors, indicating complete loss of p19 Arf is favored during progression. Thus, p19 Arf inhibits several stages in Ras-driven tumor progression. Furthermore, Ras is connected to both p19 Arf and p53 through a signaling pathway, indicating that selection for mutations in the p19 Arf / p53 pathway are a direct consequence of the initial Ras mutation. Ras, p19 Arf , p53, and Early Tumor Growth The observation that Hras mutations are found at high frequency in papillomas from wild-type and both p19 Arf - and p53-deficient mice indicates that squamous epithelial cells harboring Hras mutations have a strong selective advantage, with or without the presence of p19 Arf or p53. This permits analysis of the effects of Ras on defined genetic backgrounds in the natural setting of tumor cell evolution. The expression levels of both p19 Arf and p53 were increased in wild-type papillomas but not in p19 Arf−/− papillomas, indicating that p19 Arf regulates p53 in response to activated Ras in vivo. However, other signals to induce p53, such as those stemming from DNA damage, remain intact in the absence of p19 Arf , as shown by the rapid increase in p53 in irradiated p19 Arf -null tumors. Thus, of the many stimuli that have been shown to activate p53 using a variety of experimental systems ( Ko and Prives 1996 ; Giaccia and Kastan 1998 ), the Ras-p19 Arf pathway appears to be the major signal that operates during SCC development. This indicates that p53 activation is an intrinsic consequence of the oncogenic pathway that drives tumor growth, and is not due to other microenvironmental factors (e.g., those induced by hypoxia or due to lack of survival factors) or exogenous stimuli (e.g., DNA damage inducers). However, it remains possible that these other modes of p53 activation might predominate in other tumor types or in other circumstances. Despite the fact that both p19 Arf and p53 were induced in papillomas, loss of p19 Arf or p53 had opposite effects on early tumor growth. p19 Arf deficiency resulted in increased tumor cell proliferation and tumor growth while p53-deficient mice had reduced tumor cell proliferation, tumor numbers, and tumor size. The observation that tumors harboring mutant Ras grew faster in the presence of p53 than in the absence of p53 differs from in vitro studies, in which ectopically expressed Ras induces p53-dependent growth arrest or senescence ( Palmero et al. 1998 ; Zindy et al. 1998 ; Lin and Lowe 2001 ). These different outcomes are likely due to different experimental conditions. Ex vivo culture per se induces stress, which can further induce p53 and accelerate senescence ( Lowe and Ruley 1993 ; Serrano et al. 1997 ; Sherr and DePinho 2000 ). Moreover, the levels of active Ras protein likely differ; our autochthonous tumor model begins with a single cell that has undergone a mutation at the endogenous Hras locus, and subsequent tumor growth occurs in the context of surrounding normal cells. Locally produced growth factors or TPA treatment may attenuate the effect of Ras on p53, effectively dampening the response. Thus, the absolute levels of both Ras and p53, as well as cell type involved and the local cellular ecology, may dictate the outcome. Greenhalgh al. (1996) also reported reduction in Ras -transgene-induced skin tumors in a p53 -null background. We suggest that the acute effect of a constitutively active oncogene driving cellular proliferation, combined with lack of cell cycle checkpoints due to p53 deficiency, may generate excessive genetic instability, which may initially impair overall cellular fitness. This may be especially true at the early post-initiation stage where a small number of incipient tumor cells are competing with surrounding normal cells. This idea is consistent with the longstanding observation that p53-deficient mice ( Donehower et al. 1992 ), and Li-Fraumeni patients who carry a germline p53 mutation ( Vogelstein 1990 ), rarely develop multiple tumors, which would be expected if loss of p53 provided an early selective advantage. Also, in many human malignancies, mutations in p53 are infrequent in early premalignant lesions and much more common in late-stage disease, indicating a long latency between oncogene activation and loss of p53. Further genetic or epigenetic changes may be required for a cell to adapt to the combined effect of a dominant oncogene and loss of p53 to gain a fitness advantage. Loss of p19 Arf appears to be one such change, as reduced tumor growth due to the absence of p53 was rescued by loss of p19 Arf . Although p19 Arf regulates p53, suppression of tumor growth per se by p19 Arf does not appear to be mediated through p53. Other reports have suggested that p19 Arf has tumor suppressor functions that are independent of p53. p19 Arf -null mice show a broader spectrum of spontaneous tumors compared to p53 -null mice ( Donehower et al. 1992 ; Kamijo et al. 1999 ). Mice lacking both p19 Arf and p53 showed a wider range of tumor types than animals lacking either gene alone, and many developed multiple primary tumors ( Weber et al. 2000 ; Moore et al. 2003 ). Premalignant B-cells expressing oncogenic Eμ-myc and lacking both p19 Arf and p53 proliferated at a faster rate than cells lacking either p19 Arf or p53 alone ( Weber et al. 2000 ). Indeed, both p19 Arf and p53 are lost in a wide spectrum of human cancers, both familial and sporadic, at very high frequency ( Ruas and Peters 1998 ; Vonlanthen et al. 1998 ). Microarray and GeneChip analysis of genes induced by a conditionally regulated p19 Arf has identified members of the B-cell translocation gene family whose induction is independent of p53 ( Kuo et al. 2003 ). Expression of these genes inhibits cell proliferation and induces cell cycle arrest. p19 Arf can colocalize with the human replication protein A, suggesting a direct role for p19 Arf in DNA synthesis ( Yarbrough et al. 2002 ). p19 Arf has also been shown to inhibit ribosomal RNA processing ( Sugimoto et al. 2003 ) and to repress NF-κB transactivation ( Rocha et al. 2003 ). Finally, p19 Arf regulates vascular regression independent of p53, suggesting a role for p19 Arf in angiogenesis. ( McKeller et al. 2002 ). The functional relevance of these phenotypes for tumor suppression by p19 Arf remains to be elucidated. Ras, p19 Arf , p53, and Malignant Progression In the DMBA/TPA model, conversion of papillomas to carcinomas is a relatively rare event and can take up to 6–12 mo. Phenotypes associated with conversion include loss of basement membrane integrity, invasion of epithelial tumor cells into the dermis, loss of differentiation, and increased cellular atypia. Even more rarely, these carcinoma cells can metastasize, which involves additional phenotypic changes including, extravasation, migration, attachment, and establishment of tumor growth in an ectopic tissue. Although benign tumor growth differed between p19 Arf - and p53-deficient mice, both mutant mice showed dramatically accelerated progression to malignancy and rapid metastasis. Thus, benign tumors lacking either p19 Arf or p53 are at high risk for metastasis. In p19 Arf -deficient mice, progression was similar with or without the presence of p53 and did not involve p53 LOH, indicating that loss of p19 Arf decreased selection for p53 mutations during progression and that p19 Arf acts through p53 at this stage. Schmitt et al. (1999) also reported increased lymphoma dissemination in Eμ-myc p53 −/− mice relative to Eμ-myc mice alone. From a clinical perspective, then, the most relevant effect of the p19 Arf -p53 pathway may be to inhibit malignant conversion and metastasis. Loss of p19 Arf and/or p53 can increase progression and metastasis by several mechanisms. Deficiency in p19 Arf or p53 could indirectly affect progression via increased genetic instability, increased generation of mutants, and accelerated tumor evolution. This view postulates the existence of a distinct class of genes whose dysfunction increases progression and metastasis. It also requires a series of clonal evolutionary steps to select for cells carrying mutations in these genes. Alternatively, loss of p19 Arf or p53 can directly affect cellular phenotypes associated with progression through transcriptional regulation, or by relieving inhibition of Ras signaling. Both p19 Arf - and p53-deficient papillomas displayed several characteristics consistent with early malignancy. Conversion of papillomas to carcinomas is first characterized by a flattening, endophytic transition. p19 Arf - and p53-deficient papillomas had this morphology from the outset, suggesting an early propensity for malignant conversion. More detailed histological and immunochemical characterization showed that p53-deficient papillomas were more dysplastic and had aberrant expression of differentiation markers such as E-cadherin, P-cadherin, and Keratin-13 ( Cano et al. 1996 ). E-cadherin is a critical component of cell–cell adhesion and its down regulation is strongly associated with malignant progression ( Birchmeier et al. 1993 ; Perl et al. 1998 ). That these “high risk” p53 -null papillomas exhibited these early malignant features argues for a more direct effect of p53 on cellular phenotypes associated with progression. In addition, there was a lack of correlation between papilloma size and propensity to progress. p53 -null papillomas were fewer and smaller, yet these showed the most rapid progression. Finally, metastasis of skin tumors in both p19 Arf - and p53-deficient mice was observed within a matter of days after papilloma to carcinoma conversion, again, irrespective of precursor tumor size. Collectively, these data are inconsistent with a model in which loss of p19 Arf or p53 indirectly accelerates tumor progression by accelerating a series of independent genetic events, each followed by clonal selection, and instead favor a more direct model. We suggest that Ras may be a major driving force for multiple steps in tumor progression, with loss of p19 Arf and p53 playing a facilitating role. In addition to the well-known effects of Ras on proliferation, consistent with an early role in neoplasia, Ras also contributes to a number of phenotypes that are involved in malignant progression, including metastasis. Ras induces cell motility, invasiveness ( Lazarov et al. 2002 ; Dajee et al. 2003 ; Kim et al. 2003 ), epithelial to mesenchymal transition ( Oft et al. 1996 ; Zondag et al. 2000 ), angiogenesis ( Arbiser et al. 1997 ; Casanova et al. 2002 ), and metastasis ( Pozzatti et al. 1986 ; Webb et al. 1998 ; Oft et al. 2002 ; Varghese et al. 2002 ). Ras activates a number of signaling cascades that drive these processes ( Campbell et al. 1998 ). For example, Ras activation of the Raf-MAPK signaling cascade regulates the activities of nuclear transcription factors, including AP-1 ( Campbell et al. 1998 ). In addition to regulating proliferation, AP-1 induces a motility/invasion program ( Ozanne et al. 2000 ; Jochum et al. 2001 ; Young et al. 2003 ). Ras transgenic mice that lack Fos, a component of AP-1, develop benign skin tumors, but these fail to convert to carcinomas ( Saez et al. 1995 ). MAPK activation also determines the ability of Ras-transformed fibroblasts to metastasize to the lung ( Webb et al. 1998 ). Oncogenic Ras also works in concert with the Rho family of GTPases to regulate the intracellular actin cytoskeleton and promote cell motility and invasion leading to metastasis ( Zohn et al. 1998 ). Sustained signaling by oncogenic Ras can permanently downregulate Rac activity and lead to an epithelial to mesenchymal transition. This transition is associated with changes in gene expression, loss of E-cadherin-mediated cell–cell adhesions, and increased invasiveness of tumor cells ( Oft et al. 1996 , 2002 ; Zondag et al. 2000 ). This allows the cell to become mobile, invade surrounding tissue, and establish metastatic sites. An additional link between Ras and tumor progression was demonstrated in a genome-wide survey of Ras transformation targets that identified a significant increase in expression of genes triggering invasion and metastasis, such as the laminin receptor, collagenase (Mmp-1), stromolysin (Mmp-3), and the Cd44 glycoprotein ( Zuber et al. 2000 ). Ras also repressed genes involved in anti-invasive or anti-angiogenic activity, such as syndecan-2, tissue inhibitor of metalloproteases-2 (Timp-2), and thrombospondin-1. Further support for a continual involvement of Ras in tumor progression is indicated by the increase in copy number of mutant Ras alleles that is observed during tumor progression. In the DMBA/TPA skin tumor model, the mutant Ras allele is frequently duplicated in papillomas and amplified to multiple copies in carcinomas, especially in the most aggressive spindle cell tumors ( Bremner and Balmain 1990 ; Buchmann et al. 1991 ). Increased expression of mutant Ras genes by gene amplification or other mechanisms is found in other tumor types ( Schwab et al. 1983 ; Tanaka et al. 1986 ; Winter and Perucho 1986 ; Yokota et al. 1986 ) and has been shown to promote the growth of head and neck SCC and carcinoma of the cervix ( Hoa et al. 2002 ; Soh et al. 2002 ). This indicates that there is a selective advantage to progressively increasing levels of Ras throughout tumor progression. Quantitative differences in Ras activity are known to differentially activate signaling pathways leading to different cellular outcomes ( Shields et al. 2000 ). Loss of p53 may facilitate the increase in Ras levels, in that cells with inactivated p53 show greatly increased frequency of gene amplification ( Yin et al. 1992 ). Thus, mutation in Ras is much more than the initiating event: it directly contributes to many of the phenotypes associated with malignant progression ( Figure 7 ). As Ras induces both p19 Arf and p53, and both are antagonistic to Ras, we suggest that an important consequence of p19 Arf and p53 loss is that it permits increased Ras levels and signaling, fueling further tumor progression. In addition to counteracting Ras, p19 Arf and p53 likely contribute to tumor suppression through additional pathways. For example, loss of p19 Arf increases tumor growth, and loss of p53 confers resistance to apoptosis and loss of cell cycle checkpoints, leading to genetic instability. The observation that oncogene mutations are linked to tumor suppressor gene activation through mechanistic signaling pathways indicates that selective pressure in favor of tumor suppressor gene mutations originates from the initial oncogenic lesion and is thus intrinsic to the tumor. Figure 7 An Integrated Model of SCC Progression At the genetic level, treatment of mouse skin with DMBA induces mutations in Hras resulting in initiated cells that express constitutively active Ras protein (grey rectangles). TPA treatment is required for clonal expansion of these Hras mutant cells to form papillomas. Frequent duplication of the mutant Ras allele in papillomas indicates increased Ras signaling is favored. Mutation of p53, as well as additional Ras gene amplification, is seen in carcinomas, particularly in the most aggressive tumors (black rectangles). At the signaling level, mutant Ras upregulates p19 Arf (Arf), which in turn activates p53. p19 Arf , in turn, inhibits growth of Hras-driven tumors in a p53-independent manner. p19 Arf , acting through p53, also inhibits malignant progression and metastasis. As Ras signals through p19 Arf and p53, selection for subsequent mutations in p19 Arf or p53 is a direct consequence of the initial Ras mutation. In this model, Ras drives tumor progression through direct signaling effects and by dictating the evolution pathway of the tumor. Materials and Methods Mice p19 Arf -deficient mice (C57BL/6 × 129SvJ) were provided by Martine Roussel and Charles Sherr ( Kamijo et al. 1997 ). To increase their susceptibility to skin tumor development, p19 Arf−/− mice were backcrossed to the susceptible NIH/Ola strain (Harlan Olac, Oxfordshire, United Kingdom), and carcinogenesis studies were performed on the F 3 littermates of this cross. 20 mice of each genotype, p19 Arf−/− , p19 Arf+/− , and p19 Arf+/+ were treated. The backs of 8-wk-old male and female mice were shaved and treated with a single application of DMBA (25 μg in 200 μl of acetone; Sigma, St. Louis, Missouri, United States) followed a week later by twice weekly applications of TPA (200 μl of 10 −4 M solution in acetone) for 15 wk. The number and size of papillomas on each mouse were recorded every 2 wk. Mice were sacrificed if moribund or following detection of carcinomas. Tumors were frozen for DNA extraction and/or fixed in formalin to be processed and stained with hematoxylin/eosin (H&E) for histological examination. Mice deficient for p53 ( Donehower et al. 1992 ) (F 7 backcross to NIH) were crossed to p19 Arf -deficient (F 4 backcross to NIH) mice to generate p19 Arf+/− p53 +/− mice. These mice were intercrossed to generate all nine possible p19 Arf /p53 genotypes . Some 20–30 mice of each genotype were subjected to the same DMBA/TPA protocol and monitored as described above. Immunoblotting In order to remove abundant keratin present in papillomas and carcinomas, nuclear extracts were prepared as described ( Schreiber et al. 1989 ) with modifications. Pieces of skin or tumor were ground in liquid nitrogen with a mortar and pestle, and the resulting powder was dissolved in buffer A and further homogenized for 1 min on ice (PowerGen 125, Fischer Scientific, Pittsburgh, Pennsylvania, United States). Buffers A and C both contained 1 mM DTT, 0.4 mg/ml Pefablock, 25 mg/ml Aprotinin, 10 mg/ml Pepstatin, and 10 mg/ml leupeptin (Roche, Alameda, California, Unites States) to inhibit proteases. Protein concentrations were standardized using the Bradford assay (Bio-Rad, Hercules, California, United States) and equal loading (50 μg/lane) was confirmed by Ponceau S staining of the nylon membrane after blotting. Western blot analysis was performed using specific antibodies against p19 Arf (Novus Biologicals, Littleton, Colorado, United States) and p53 (Novocastra Laboratories, Newcastle-upon-Tyne, United Kingdom). Hras sequencing Genomic DNA was prepared from tumor and normal tissue by QIAamp DNA Mini Kit (Qiagen, Valencia, California, United States). A 400-bp PCR fragment containing exon 2 of Hras was amplified with standard PCR (3.0 mM MgCl 2 andannealing for 2 min at 37 °C with 40 cycles), with 5′-GACTCCTACCGGAAACAGGT-3′ and 5′-CTGTACTGATGGATGTCCTC-3′ primers. We used the internal primer 5′-TGGTCATTGATGGGGAGACA-3′ to sequence exon 2, using PE Biosystems (Applied Biosystems, Foster City, Calfornia, United States) Dye-Terminator and Big-Dye cycle sequencing. Histological analysis Sections of normal skin, papillomas, carcinomas, and other organs were removed and fixed in 10% normal buffered formalin for 4 h. After fixation, tissues were processed and then embedded in paraffin. From the tissues, 4-μm sections were cut and stained for either p53 (CM5, Novocastra) or pan-keratin (AE1/AE3, Novocastra) using high-temperature antigen retrieval in 10 mM citrate buffer (pH 6), or for BrdU (Dako, Glostrup, Denmark) after treating with 2N HCl followed by 0.1% trypsin. After staining with the primary antibody, the sections were stained with a biotin-conjugated secondary (Vector Laboratories, Burlingame, California, United States) followed by StreptABComplex/HRP (Dako). Slides were developed with DAB/NiCl and counterstained with methyl green. Control sections with no primary antibody were run concurrently. Other sections were cut and stained using a standard H&E method. Proliferation index was determined by counting the number of BrdU-stained cells per 40× field. The apoptotic index was determined by counting the H&E slides for the number of apoptotic figures per 40× field. All counts were done on a Nikon (Tokyo, Japan) Labophot-2 microscope. Semiquantitative PCR analysis Genomic DNA was prepared from tumor tissue or normal kidney by QIAamp DNA Mini Kit (Qiagen). To measure LOH of the p19 Arf wild-type allele in p19 Arf+ /− tumors, wild-type and knockout alleles were amplified by PCR separately then combined for electrophoresis. Primers 5′-AGTACAGCAGCGGGAGCATGG-3′ (Arf1), 5′-TTGAGGAGGACCGTGAAGCCG-3′ (Arf2), and 5′-ACCACACTGCTCGACATTGGG-3′ (ArfN) were used to amplify wild-type (Arf1 and Arf2) and knockout alleles (Arf2 and ArfN) from 100 ng of genomic DNA using 68 °C for annealing and extension at 90 s for 30 cycles. Equal amounts of each PCR product were then combined for electrophoresis on a 2% TAE agarose gel. Wild-type and knockout alleles of p53 +/− tumors were amplified in a separate reaction as described ( Timme and Thompson 1994 ) for 30 cycles. PCR products were electrophoresed on a 2% TAE agarose gel. Comparison gradients for p19 Arf and p53 were established by combining wild-type and knockout genomic DNA in quantified ratios, then amplifying as described above. Statistical methods In order to assess the impact of p19 Arf (or p19 Arf /p53 ) genotypes on the development of papillomas, longitudinal profiles of papilloma counts were analyzed using the generalized estimating equation (GEE) approach ( Zeger and Liang 1986 ). GEE is an established statistical approach to the regression analysis of longitudinal data. Our analysis used papilloma counts of each mouse measured every 2 wk as the outcome variable and incorporated within-mouse correlations of papilloma counts over time in making statistical inference. Using GEE, we estimated average differences of papilloma counts across genotypes after DMBA administration. Since the development of papillomas depends on the time since the DMBA administration and may differ by the sex of the mice, the effects of the weeks since the DMBA administration and sex were controlled for in the GEE analysis as covariates. A working correlation structure of the GEE analysis was specified as the first-order autoregressive structure over the time since the DMBA administration; however, GEE is robust against the misspecification of the correlation structure. In contrast to papillomas, the overall number of carcinomas developed during the experiment was relatively small. Thus, we analyzed differences by genotype in the probability of developing a carcinoma after DMBA administration. A logistic regression model ( Clayton and Hills 2003 ) was used to assess the odds of developing a carcinoma during the experimental period and to compare it across p19 Arf (or p19 Arf /p53 ) genotypes. Estimates of relative odds were adjusted for sex effects. Fisher's exact test was used for comparing two proportions such as comparing LOH proportions between papillomas and carcinomas. All statistical tests were two-sided.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509304.xml
553977
Low density DNA microarray for detection of most frequent TP53 missense point mutations
Background We have developed an oligonucleotide microarray (genosensor) utilizing a double tandem hybridization technique to search for 9 point mutations located in the most frequently altered codons of the TP53 gene. Isolated and multiplexed PCR products, 108 and 92 bp long, from exons 7 and 8, respectively, were obtained from 24 different samples. Single-stranded target DNA was then prepared from isolated or multiplexed PCR products, through cyclic DNA synthesis. Independent ssDNA's were annealed with the corresponding pairs of labeled stacking oligonucleotides to create partially duplex DNA having a 7-nt gap, which contains the sequence that will be interrogated by the capture probes forming double tandem hybridization. In the case of multiplexed ssPCR products, only two stacking oligonucleotides were added per target, therefore the gap for the PCR products having two consecutive codons to be interrogated in exon 7 was 12 nt long, so only single tandem hybridization was produced with these respective probes. Results 18 codon substitutions were found by DNA sequencing. In 13 of them a perfect correlation with the pattern of hybridization was seen (In 5 no signal was seen with the wt probe while a new signal was seen with the appropriate mutant probe, and in 8 more, as expected, no signal was seen with any probe due to the absence of the corresponding probe in the array). In 3 other cases a mutation was falsely suggested by the combination of the absence of the wild type signal along with a false signal in the other probe. In the other 2 cases the presence of the mutation was not detected due to the production of a false hybridization signal with the wild type probe. In both cases (false mutation or no mutation detected) relatively stable mismatched target/probe duplexes should be formed. These problems could be avoided by the addition of probes to improve the performance of the array. Conclusion Our results demonstrate that a simple TP53 microarray employing short (7-mer) probes, used in combination with single or double tandem hybridization approach and a simple or multiplex target preparation method, can identify common TP53 missense mutations from a variety of DNA sources with good specificity.
Background Cancer is a group of diseases characterized by uncontrolled cell growth. According to The World Cancer Report, cancer rates could further increase by 50% to 15 million new cases in the year 2020 [ 1 ]. The chances of surviving the onset of some common cancers depend largely on how early they are detected in addition to how well they are treated. The progression of mammalian cells towards malignancy is an evolutionary process that involves an accumulation of mutations at the molecular and chromosomal level. A candidate for involvement in this process is the tumor suppressor gene TP53 (MIM # 191170) which encodes a transcription factor (p53 protein) with cancer inhibiting properties. This gene is the most frequently mutated gene in human cancer. Between 30% and 70% of cancers of almost every organ and histological subtype have a point mutation in one of the two TP53 gene copies together with loss of the other copy [ 2 - 4 ]. In many cancers the distribution of mutations along the TP53 gene is tumor-specific, clustered between exons 5, 7 and 8. This region is highly conserved throughout evolution [ 3 ]. In this domain, six mutational "hotspots" have been identified at codons 175, 245, 248, 249, 273, and 282. The mutational profile of the TP53 gene is different between cancers, for example, codons 157, 158, 248, 249, and 273 have been designated as TP53 mutational hotspots in lung cancer [ 5 , 6 ], while in breast cancer the mutational hotspots are in codons, 175, 245, 248, 249 and 273 [ 7 ]. On the contrary, in cervical cancer the frequency of the p53 mutations is very low. [ 8 ]. Because mutations in the TP53 gene are so common in human tumors and have been widely reported in the literature, extensive mutational databases exist, such as that maintained by IARC [ 9 ]. Walker et al. [ 10 ], have utilized the IARC TP53 mutation database to define 73 "hotspots" for mutation in TP53 related to changes in protein structure and function. Martin et al. [ 11 ] performed a systematic automated analysis of the effects of TP53 mutations on the structure of the core domain of the p53 protein to functionally classify the various types of p53 mutants according to predicted effects on protein folding or DNA-protein interactions. Most mutations are currently identified by conventional methods of polymorphism detection and DNA sequencing [ 12 ]. A growing number of nucleic acid hybridization techniques have recently been applied to biomedical problems, including development of DNA probe arrays for detection of TP53 mutations [ 13 - 15 ]. In many applications, DNA chips containing surface-bound oligonucleotides (probes) are used to interrogate sample (target) sequence information through complementary recognition (hybridization) in a highly parallel fashion [ 16 ]. Major applications of DNA microarrays include gene expression profiling [ 17 , 18 ] and gene mutation analysis [ 19 , 20 ]. These techniques allow parallel analysis of multiple DNA samples. Mutation studies with DNA microarrays are still at an early stage and are in continuous development [ 21 , 22 ]. Although the field of microarray assays is still in a relatively early stage, it is generally anticipated that microarray assays will offer decreased cost and faster results, compared with the traditional, more labor intensive dideoxy sequencing approach. We recently reported a novel tandem hybridization strategy and its application for identification of mutations. Reliable discrimination of point mutations has been achieved by double tandem hybridization with a set of seven-mer probes [ 23 , 24 ]. In the present work we designed a novel small format DNA chip (genosensor array) to search for nine point mutations in codons 248, 249 and 273 of the TP53 gene. Results PCR products and reference hybridization patterns PCR products from exon 7 (codons 248 and 249) and exon 8 (codon 273) were obtained from each type of DNA sources listed in Materials and Methods. Figure 2A shows gel electrophoretic analysis of the fragments of 108 and 92 bp corresponding to exons 7 and 8, respectively. To prepare ssDNA targets PCR products were used as templates for cycle synthesis of a target strand, using the reverse PCR primer in single or multiplexed reactions cycled 25 times. The production of single-stranded target DNA's was assessed by its change in electrophoretic mobility in 4% agarose gels (Figure 2B ). The reference hybridization patterns produced using synthetic DNA targets are shown in Figures. 3A and 3B . As expected, all the synthetic targets, including those of the negative controls, yielded hybridization signals with their respective ("perfect-match") probes. Pairs of wild-type with each mutant synthetic target sequences (1:1 mixtures) were also hybridized to reproduce heterozygous conditions. All gave the expected signals, (Figure 3B ). Hybridization of DNA targets Hybridization of individual or multiplexed ssDNA targets prepared from the 24 DNA samples (that correspond to 72 codons in 24 samples) was performed with similar results. Base substitutions were seen in 11 out of the 24 samples tested (Table 3 ). One, two or three codons were altered in 5 (P6, L1, L2, B3 and B6), 5 (P1, L4, L7, L9 and L10) and 1 (L8) samples, respectively. In the cases of the samples having two or three codons affected, at least one of the base substitutions was silent. In the other 13 samples only WT sequences were identified (data not shown). Representative hybridization patterns are shown in Figure. 4A and the overall results are summarized in Table 3 . The DNA samples hybridized with one or two probes from each codon, corresponding to the homozygous or heterozygous condition, respectively. Among the 24 different samples multiplexed or individually tested in the microarray, six single base susbstitutions in codon 248 were found by sequencing the PCR product. All of them corresponded to base substitutions keeping the original aminoacid (Arginine). The same number of point mutations were also seen in codon 249 by DNA sequencing. However, no probes to search for these base substitutions were placed in the DNA microarray. Six point mutations in codon 273 were obtained. All of them, except the CGT→CCT change observed in plasmid P6, were searched with the set of probes contained in the DNA microarray. In this codon all the base substitutions are producers of aminoacid substitution. Discussion The TP53 tumor suppressor gene has proven to be one of the cellular genes most often mutated in human neoplasias. Analysis of the mutational events that target the TP53 gene has revealed evidence for both exogenous and endogenous mutational mechanisms. This gene mutational spectrum suggests evidence for a direct causal effect of tobacco smoke in lung cancer [ 26 , 27 ], aflatoxin B1 in liver cancer [ 28 ]. Therefore, identification of mutations in the TP53 gene may play an important role in the diagnosis, staging, and management of the cancer patients. More than 70% of the molecular studies have focused on the central region of the TP53 gene, more specifically on exons 5 through 8 which encode the DNA binding domain [ 3 ]. Thus, missense mutations in this region can cause a modification that may alter DNA binding. It is therefore essential that clinical trials aim to accurately establish the effect of TP53 mutations on all clinical parameters. One approach is the p53 GeneChip of Affymetrix used to identify mutations in this gene [ 13 , 14 ]. However, this technology is expensive and poorly available to many clinicians, which makes it necessary to develop new mutation detection methods [ 29 , 30 ]. The updated IARC TP53 mutation database [ 38 ] shows that more than 1700 different point mutations, which are associated to cancers, are distributed along the TP53 gene. TP53 somatic mutations are found in most types of sporadic human cancers at various frequencies (from 20% to 60%) [ 13 , 14 ]. The database also shows that approximately 20% of all mutations are located in the five codons (hotspots) in the core domain: 175, 245, 248, 249 and 273. Therefore a primary screen of these codons would be attractive. Most SNP diagnostic arrays use probes varying from 15 to 20 nt in length with similar Tm values [ 31 ]. However, under this circumstance some of the mutant sequences are difficult to discriminate, due to the relatively small destabilizing effect of certain single mismatches in the probe/target duplex [ 32 ]. The tandem hybridization method described here offers several advantages over the traditional oligonucleotide array configuration in which each interrogated target sequence is represented by a single surface-tethered probe. Because the long stacking probe targets the analysis to a single unique site within a nucleic acid, direct hybridization analysis of nucleic acid samples of high genetic complexity using short capture probes may be enabled [ 24 ]. Recently, we successfully used a double tandem hybridization strategy in the genesensor chip system to discriminate point mutations in codon 634 of RET oncogene in individuals affected by medullary thyroid carcinoma [ 23 ]. This DNA chip format involves a target sequence which is annealed with a pair of labeled oligonucleotides (stabilizing oligonucleotides) prior to hybridization to the array of capture probes, forming a single-stranded gap between the stabilizing oligonucleotides and corresponding to the length of the capture probes. The partially duplex product contains the sequences interrogated by the capture probes, and when the sequences between the target DNA and the capture probe (attached to slide) are perfectly complementary, the stability of the hybridization is enhanced due to coaxial stacking of bases between the contiguous ends of the probe and the stabilizing oligonucleotides. The energy involved varies with the identity of the base pairs and the coaxial stacking [ 22 ]. Tandem hybridization uses shorter capture probes than traditional oligonucleotide array hybridization, which may contribute to increased discrimination of point mutations because the single base mismatch acts on a less stable duplex formed between the target DNA and the shorter probe. Two additional attractive features inherent to tandem hybridization are that, i) the probability that formation of secondary structure within the single stranded DNA [ 25 ] will block accessibility of the target sequence is reduced, and ii) the label needed to reveal the hybridization can be introduced into the two stabilizing oligonucleotides, avoiding the extra step of labeling individual target strands [ 33 , 34 ]. The goal of the present work was to develop a simple format array able to perform a primary screening of the mutations occurring in the most frequently affected codons of the TP53 gene. It has been recently reported that detection of mutations yields variable results depending on the type of DNA sample [ 35 , 36 ]. Is clear that, the sensitivity to detect mutations in TP53, may also differ depending on the method used. In the current study the double tandem hybridization assay was applied to genomic DNA samples extracted from a variety of sources, including blood cells, cell lines and plasmids. There were not any differences in the ability of our system to detect mutations with respect to the different DNA sources. In the experiments employing synthetic DNA targets hybridized with arrays of mutant and WT probes the expected hybridization signals were seen both in homozygous and heterozygous conditions (Figure. 3 ). This result indicates that the probes are working properly. The hybridization of natural targets was done both with isolated and multiplexed PCR products with similar results. Multiplexing was applied in all (PCR reaction ssDNA target preparation, annealing of target with stacking oligonucleotides and hybridization) the steps. So these results are suggesting that multiplexing can be confidently employed for this analysis. Base changes were found in 11 DNA samples. Two were from plasmid, seven from cell lines and two from blood samples. In 5 of these samples a single codon was affected, in other 5 samples two codons were altered, and in the other sample three codons were affected. Therefore a total of 18 codon alterations were detected. In five of the 18 codon alterations (Samples L2 codon 273, L7 codon 273, L8 codon 273, B3 codon 273 and B6 codon 273) both the absence of hybridization signal with the wild type probe and the production of signal with the appropriated probe was seen. In eight cases (Samples P1 codon 248, P6 codon 273, L8 codons 248 and 249, L9 L8 codons 248 and 249 and L8 codons 248 and 249 and, L10 codons 248 and 249) no hybridization signal was seen even with the wild type sequence. This was due to the absence of the corresponding mutant probe in the array. In three cases (P1 codon 249, L4 codon 249 and L7 codon 249) having either homozygous (sample P1) or heterozygous (samples L4 and L7) genotypes (wtAGG→mutAGA) no hybridization signal with the wild type probe and, an additional false hybridization signal with the probe corresponding to the substitution of guanine for thymine in codon 249 (p-249 G 2 →T) was seen, which could be relatively stable under the conditions tested. Therefore an additional probe, searching by the AGA sequence, should be added in the future array to discriminate between the two alternative silent (AGA) and missense (AGT) genotypes. In the other two codons affected (L1 codon 248 and L4 codon 248) the presence of signal with the wild type probe was seen even when the sequencing was showing a base change (wtCGG→mutCGT in sample L1 or 248 wtCGG→mutCGK in sample L4). In both cases a G:N mismatch is formed which, according to SantaLucia [ 37 ], is relatively stable. So, as in the previous two cases the incorporation of additional probes in the array should help to improve performance of the DNA microarray to confidently identify the genotypes. The base substitution observed in B3 and B6 blood samples is a mutation that substitutes Arginine by Leucine. The presence of this mutation in blood suggests that it could be from germinal origin. In all the mutations (silent and missense) observed an Arginine was involved, which can be expected because it is coded by 6 codons and it frequently produces non conservative aminoacid substitutions. The array is able to confidently detect most of the mutations searched, however, additional probes should be added to improve its performance. Conclusions Our results demonstrate that a simple TP53 microarray using short (7-mer) probes, used in combination with a double tandem hybridization approach and a simple multiplex fragment preparation method, can identify common TP53 point mutations from different DNA sources. All the previous data confirms that the tandem hybridization technology, with the appropriated set of probes, is highly specific and can be confidently employed for diagnostic purposes. This system could provide rapid, economical and accurate mutational screening in the TP53 gene. We are presently improving our system by adding probes for these and other TP53 mutational hot spots. Methods IARC TP53 gene database search The distribution of nucleotidic changes in the TP53 gene associated with all cancers war retrieved from the IARC TP53 database [ 38 ]. Then, the single base substitutions occurring in the 3 exons of TP53 most frequently altered in cancers were obtained. Both wild type (accession # U94788) and mutated TP53 sequences were used to design the probes, stacking oligonucleotides, PCR primers and synthetic targets. Oligonucleotides Four new PCR primers of 20–25 nt length (Table 1 ) were used to amplify two DNA fragments of 108 bp (exon 7: codons 248 and 249) and 92 bp (exon 8: codon 273) within the TP53 gene. Six stacking oligonucleotides of 17–21 nt length were annealed with their corresponding target DNA's, in three pairs, to produce partially duplex DNA's having a 7-nt gap which is the sequence interrogated by the capture probes (Figure 1A ). Fourteen capture probes (seven-mer length; Table 1 ) were designed, two of them to search the wild type sequences for codons 157 and 158, which were used as negative controls, three more for the wild type 248, 249 and 273 codon sequences and nine to look for mutant sequences. According to our previous experience, 7-mer oligonucleotides with base changes located at central positions were selected as probes in order to yield good discrimination of point mutations by tandem hybridization at room temperature. The probes, which had a 3'-aminolink (3'-aminopropanol) modification for covalent binding to the glass slide surface [ 24 ], were purchased from Integrated DNA Technologies Inc. (Coralville, IA, USA). Biological samples (DNA targets) Twenty-four DNA samples from different sources were probed, some having known wild type or mutant TP53 sequences and others with unknown TP53 genotype. These included eight blood samples (B), two obtained from healthy subjects and six obtained from lung cancer affected patients, and ten established cancer cell lines or recombinant DNA's acquired from the ATCC (L). Two of the cell lines were a generous gift from Dr Mary A Salazar (IIB, UNAM) and one was a gift from Dr José Sullivan (INER, SS, Mexico). The other eight cell lines were provided by the DNA bank from Genomic Oncology Laboratory (CMN S XXI). Finally, six recombinant plasmids (P) harboring TP53 (wild type or different point mutations), provided by MSc. Irene Correa (CMN S XXI) were employed to standardize the hybridization arrays. The names, identification numbers, source and known genotypes in the 24 samples studied are described in Table 2 Target preparation and PCR assay Genomic DNA was isolated and purified using the Genomic DNA Extraction Kit (Life Technologies Inc., Gaithersburg, MD). DNA concentration was determined by absorbance at 260 nm in a spectrophotometer MBA 2000 (Perkin-Elmer) and by the PicoGreen ® dsDNA Quantitation Kit (Molecular Probes). Plasmid DNA was prepared from bacterial cultures using the alkaline lysis/SDS minipreparation according to the manufacturer's instructions (Life Technologies). The quality of the DNA preparations was assessed by gel electrophoresis in 1% agarose. All the reagents for PCR reactions were purchased from Promega (Madison WI. USA). The PCR primers are described in Table 1 . Oligonucleotides sequences and fragment sizes for each exon were as follows: PCREX7-F and PCREX7-R for a 108 bp product in exon 7, and PCREX8-F and PCREX8-R for a 92 bp fragment in exon 8. The final reaction mixture contained 0.2 mM each dNTP, 50 mM KCl, 10 mM Tris-HCl (pH 8.4), 1.5 mM MgCl 2 , 0.5 μM each primer, 2.5 U of Taq DNA pol (Promega) and purified DNA (50–100 ng) in a final volume of 100 μL. The PCR profile consisted of an initial heating at 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 30 s, with a final extension step at 72°C for 7 min. The PCR product was assessed by electrophoresis in 1 % agarose gel stained with ethidium bromide. Single-stranded target DNA was prepared by cycle synthesis as follows. A 30-μL aliquot of PCR product was processed using Ultrafree (Millipore, Bedford, MA) spin filters (30,000 M r cutoff), and resuspended in the same volume of HPLC-grade water (OMNI SOLV ® , EM Science). A 5-μL aliquot of this solution was added to a 100-μL PCR reaction, along with the primer corresponding to the target strand, then incubated for 27 temperature cycles using the same temperature profiles described above. The formation of each amplicon was assessed by mobility shift in agarose gel electrophoresis. Each one of the six stacking oligonucleotides was 5'-labeled by reaction with T4 polynucleotide kinase (Invitrogen, San Diego, CA) as follows. 5 pmol of each dephosphorylated stacking oligonucleotide; 5X Forward Reaction Buffer (5 μl); T4 Polynucleotide Kinase (10 unit); [γ 32 P]ATP (10 μCi/μl, 3000 Ci/mmol) 2.5 μl (NEN, Boston MA), specific activity 7000 Ci/mmol, diluted in sterile water to 7 μCi/μl, and autoclaved water HPLC-grade to 25 μl. The mix was incubated 10 min at 37°C and the reaction was stopped by adding 5 mM EDTA (2.5 μl). To prepare labeled partially duplex (gapped) target DNA's, each pair of γ 32 P-labeled stacking oligonucleotides was preannealed with the respective synthetic DNA target, single-stranded PCR fragments or multiplexed, single-stranded PCR products to form a 7-nt gap in each target. The annealing mixture contained 50 μL 20X SSC, 10 μL 1 M Tris-HCl (pH 8.0), 3 μL 0.5 M EDTA, 0.2 pmol of each labeled stacking oligonucleotide, 10 μL of target DNA, and HPLC-grade H 2 O to a final volume of 90 μL. The mixture was incubated at 95°C for 5 min, 45°C for 5 min, then 6°C for 5 min. Excess [γ 32 P]ATP was removed by filtration through an Ultrafree spin filter (3,000 M r cutoff) (Millipore. Bedford, MA), and the retained DNA was dissolved in 20 μL of 1X SSC. In the case of multiplexed targets, only 4 stacking oligonucleotides (SO-248-R, SO-249-L, SO-273R and SO-273-L) were added, one pair for each exon in the study, forming gap sizes of 12 and 7 nt for exons 7 and 8, respectively. Thus, with exon 7 a single tandem hybridization was conducted with the glass-tethered 7-mer capture probes, whereas with exon 8 double tandem hybridization was conducted. DNA probe arrays Glass microscope slides were prepared by soaking in hexane 20 min, then 20 min in absolute ethanol followed by drying 5 h at 90°C. Oligonucleotide probes containing 3'-terminal aminopropanol modification were dissolved in sterile water to a final concentration of 20 μM, and 200 μl droplets of each probe were applied to the glass slides using a model 417 Affymetrix (Santa Clara, CA) arrayer to place submicroliter droplets onto glass slides. Each array contained 14 probes: three wild type sequences and nine mutant sequences corresponding to the TP53 targets under study, two negative controls consisting of 7-mers targeted to the wild type sequences from codons 157 and 158 of TP53 gene, plus one blank spot without probe. As depicted in Fig. 1B , each TP53 capture probe was arrayed in triplicate (to assess the reproducibility of the results) in an upward direction starting from position 1 and the printing was divided into three groups. Positions 1–14 were printed first, then the pattern was repeated twice more across the slide. Position 15 was left empty (nothing at all was applied) and served as a negative glass control. After spotting, each slide was washed with water to remove unbound oligonucleotides, air-dried, then stored in a dessicator at room temperature. This attachment procedure results in a surface density of 10 10 -10 11 molecules mm 2 , which corresponds to intermolecular spacing of about 30–100 Å across the surface [ 25 ]. Hybridization experiments Just prior to hybridization the slides were soaked for 1 h at room temperature with blocking agent (10 mM tripolyphosphate) and then rinsed twice with water and air-dried. The hybridization cocktail contained: 118 μL 5 M tetramethyl-ammonium chloride (TMAC) (Life Technologies), 9 μL 1 M Tris-HCl (pH 8.0) (Life Technologies), 0.72 μL 0.5 M EDTA, 1.8 μL 10% (w/v) sodium dodecyl sulfate, 14.4 μL 40% (w/v) polyethylene glycol, 20 μL labeled target DNA in 1X SSC, and 35.3 μL HPLC-grade water. A 40-μL aliquot of this cocktail was placed onto each array and a cover slip was applied. Slides were incubated in a humid chamber for 3 h at 25°C. After hybridization, the slides were washed by dipping several times in the same hybridization solution without polyethylene glycol and DNA, air-dried, and then wrapped in plastic film and placed against X-ray film (Kodak BioMax) for autoradiography. Detection and quantitation of 32 P-labeled target molecules bound across the hybridization array was performed with a scanner (HP Scanjet 4400 c), and analysis of densitometric images was performed using ScanAnalyze 2 software (Stanford University). To provide confidence in the interpretation of the signals produced by human samples, reference hybridization patterns were prepared using a full set of synthetic targets. For this purpose, thirteen synthetic targets (44–50 nt), representing the wild type and mutant sequences, were used independently and in combination, to represent homozygous and heterozygous conditions. Sequencing of PCR products The PCR products obtained were purified using the QIAEX II kit (QIAGEN Inc. USA). For each PCR product the presence or absence of a mutation in exons 7 and 8 was demonstrated via sequencing using the Big Dye Terminator kit (Perkin-Elmer) and a model 373 automated DNA sequencer (Instruments Core Center of Health Research Council- IMSS). Authors' contributions ARL performed all molecular assays, data collection and analysis, and drafted the manuscript. RMR conceived and designed the study, coordinated and managed the study, performed data analysis and participated in drafting the manuscript. MSV conceived and designed the study, coordinated and managed the study, performed data analysis and participated in drafting the manuscript. JMEL participated in the design and coordination of the study. AMT designed the oligonucleotide probes, performed bioinformatic analysis of targets, and participated in drafting the manuscript. KLB assisted in coordination of the study, data analysis, and drafting the manuscript. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553977.xml
516581
WHO systematic review of maternal morbidity and mortality: the prevalence of severe acute maternal morbidity (near miss)
Aim To determine the prevalence of severe acute maternal morbidity (SAMM) worldwide (near miss). Method Systematic review of all available data. The methodology followed a pre-defined protocol, an extensive search strategy of 10 electronic databases as well as other sources. Articles were evaluated according to specified inclusion criteria. Data were extracted using data extraction instrument which collects additional information on the quality of reporting including definitions and identification of cases. Data were entered into a specially constructed database and tabulated using SAS statistical management and analysis software. Results A total of 30 studies are included in the systematic review. Designs are mainly cross-sectional and 24 were conducted in hospital settings, mostly teaching hospitals. Fourteen studies report on a defined SAMM condition while the remainder use a response to an event such as admission to intensive care unit as a proxy for SAMM. Criteria for identification of cases vary widely across studies. Prevalences vary between 0.80% – 8.23% in studies that use disease-specific criteria while the range is 0.38% – 1.09% in the group that use organ-system based criteria and included unselected group of women. Rates are within the range of 0.01% and 2.99% in studies using management-based criteria. It is not possible to pool data together to provide summary estimates or comparisons between different settings due to variations in case-identification criteria. Nevertheless, there seems to be an inverse trend in prevalence with development status of a country. Conclusion There is a clear need to set uniform criteria to classify patients as SAMM. This standardisation could be made for similar settings separately. An organ-system dysfunction/failure approach is the most epidemiologically sound as it is least open to bias, and thus could permit developing summary estimates.
Background Severe acute maternal morbidity (SAMM), also known as "near miss", is defined as "A very ill pregnant or recently delivered woman who would have died had it not been that luck and good care was on her side" [ 1 , 2 ]. This concept is relatively new in maternal care, but is increasingly becoming important in areas with low maternal mortality ratios or where the geographic area is small [ 3 , 4 ]. The use of data collected on SAMM has been shown to be a mechanism for identifying health system failures or priorities in maternal health care more rapidly than maternal deaths [ 5 ]. It has the advantage of events still being rare enough not to overload clinicians and data capturing personnel within a facility. Its routine use as an indicator, however, is limited due to the lack of uniform criteria of identification of the cases. This study was undertaken to systematically review all available studies on SAMM with a view to establishing the global prevalence and examining the usefulness as a maternal health indicator. Methods This study is a part of a bigger systematic review undertaken by the UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), Department of Reproductive Health and Research at the World Health Organization (WHO) to obtain prevalence/incidence data on maternal mortality and a range of morbidities including SAMM. The methodology of the systematic review followed an a priori protocol and involved an extensive search of all relevant published/unpublished data from 1997 to 2002. The methodology of the systematic review and the search strategy have been described elsewhere [ 6 ]. In brief, we searched 10 electronic databases, WHO regional databases, internet and reference lists, contacted experts in the field, and hand-searched relevant articles in the WHO Library. Criteria for inclusion of studies in the review were: inclusion of data relevant to pre-defined conditions, specified dates for data collection period, including data from 1990 onwards, sample size >200 and a clear description of methodology. A data extraction instrument was used to extract data from included studies. This instrument includes 48 items distributed in five modules three of which were relevant to this analysis. Modules were designed to collect information on (i) the general study level characteristics such as design, population, setting, (ii) prevalence/incidence of maternal morbid conditions, and (iii) quality assessment of morbidity reports. Reporting of definitions and of the procedures used for identification of cases for morbidities were part of quality assessment. We did not assign quality scores to articles, but preferred to present available information on variables regarded as quality components (including reporting of definitions, case-identification criteria, characteristics of setting and participants). Nearly 65 000 reports were screened initially by titles and/or abstracts of which more than 4500 were retrieved for full-text evaluation. More than 2500 of these were included in the review. Data extracted were entered into a specifically constructed database and tabulated using SAS software. A small number of the articles in the review report on SAMM, near miss or a similar definition such as severe morbidity, critically ill obstetric patient. Although we included other articles reporting on individual severe morbidities within their own category of conditions (e.g. severe hypertension within hypertensive disorders of pregnancy), this particular article is concerned with those papers which define a separate entity of SAMM or similar, or those which report on the most recognised end-points for SAMM (i.e. emergency hysterectomy and obstetric admissions to intensive care units). We later conducted an updated search of MEDLINE and Popline using the keywords 'near miss morbidity' and 'severe maternal morbidity' to find out about any recent publications that could be of importance, but were not included in the database due to the end date of the original search (2002) for the bigger systematic review. The date of this complementary search was June 2004. We describe below the included studies with an emphasis on the different definitions used and criteria for identification of the cases. Results A total of 30 reports of SAMM are included in the systematic review (twenty-seven identified by the original search and three by the complementary search). Study designs are mainly cross-sectional and most of them are conducted in tertiary care hospitals (see Additional file 1 ). Most of the studies describe the characteristics of the setting and participants as well as reporting definitions and procedures for identification of the cases. There are essentially two types of definition of severe acute maternal morbidity; one describing what the authors meant by a near miss; and the other describing a response to an event such as hysterectomy or admission to intensive care unit (ICU). Fourteen studies define a specific SAMM or near-miss 'condition' [ 1 , 2 , 7 - 18 ] while nine consider admissions to ICU as near-miss cases [ 19 - 27 ] and the remainder report on emergency hysterectomies [ 28 - 34 ]. In the majority of articles there is an intuitive agreement on what a near miss means – a woman who almost died but survived. Identification of cases, however, is complex and varies widely across studies. We listed the articles in Additional file 1 in three categories according to criteria used to classify patients as being near miss; disease-specific (specified criteria for common conditions, e.g. pre-eclampsia, haemorrhage); management-specific (specified criteria related to response to disease, e.g. hysterectomy or admission to ICU); and organ-system dysfunction/failure based (specified criteria for dysfunction or failure related to each organ system). One study reports the proportion of the admissions to ICU separately as well as the total number of SAMM cases [ 8 ]. We included this report under the section of organ-system based identification criteria. Individual prevalence rates for SAMM vary between 0.80% – 8.23% in the first category of studies that use disease-specific criteria. Case-finding criteria differ significantly within this category as well. In addition, some of the studies with similar criteria use mainly clinician's evaluation for identification [ 15 ], while others have established threshold levels for the degree of severity of the conditions of concern [ 7 , 18 ]. This is probably due to contextual factors such as the availability of facilities with sufficient diagnostic tools. The range is 0.38% – 1.09% in the group of reports that use organ-system based criteria and include unselected women. Rates are lower (0.01% – 2.99%) and variation is lesser in the category of studies using management-based criteria. An expected finding is the difference between resource-poor and more advantageous settings in the prevalence of SAMM. In resource-poor settings, 4–8% of pregnant women who deliver in the hospitals will experience SAMM when case-identification criteria are based on specific diseases. This rate is around 1% when the organ failure is considered. In more developed country settings, the rates are around 1% with disease-specific and 0.4% with organ-system based criteria, respectively. The results also suggest that the use of organ-system based criteria is more specific in identifying the real SAMM cases. Discussion Due to the wide variation in identification of the cases, it is not possible to pool data and make a summary estimate for SAMM. Because of the variation in case identifications in the three categories of identification criteria as well as variation within each category (e.g. for disease-specific criteria – the use of physician's evaluation versus technology requiring tests) it is difficult to make comparisons as well. Nevertheless, it is evident that the prevalence of SAMM is higher in studies conducted in less developed country settings. Rates seem to be higher also in studies that use disease-specific criteria as compared to those using organ-system based criteria for similar settings. This finding suggests less specificity of disease-specific criteria in identifying real SAMM cases. Although less specific, the use of disease-specific criteria has some advantages; it is easy to interpret, cases can be identified retrospectively, and the quality of care for that particular disease can be assessed [ 18 , 35 ]. However, the approach concentrates on certain diseases, and thus, other problems such as pulmonary embolus, which is an important cause of maternal death in developed countries could be ignored [ 36 ]. In addition, definition of conditions may not always be straightforward. For example, the same threshold for severe haemorrhage could have different consequences in women with normal haemoglobin levels or those with severe anaemia. Furthermore, although detailed objective criteria are established for case identification in developed country settings [ 18 ], the limited availability of resources in less developed settings may not permit this level of detail. Therefore, identification of cases is likely to be less accurate when the diagnosis depends on clinical estimates [ 2 , 15 ]. Use of management-specific criteria is advantageous in that it is simple to identify the cases, but it depends on many other variables such as the availability of ICU beds, the facilities in an ICU, or differing views about and indications for hysterectomy. Also, the approach does not include all SAMM cases. One study reporting SAMM according to organ-system based criteria and admissions to ICU separately shows that admissions to ICU represent only one third of all SAMM cases [ 8 ]. Use of organ-system based criteria allows for identifying all severe morbidities and then investigating the primary cause, thus does not discard any particular condition. It is the most reproducible across similar areas and criteria can be defined according to resources available. High technology requiring laboratory and haemodynamic investigations can be avoided. However, it is the most labour-intensive for identifying cases, hence criteria for inclusion as near miss must be strict. Bias can be introduced if data collection is incomplete and prevalence can be underestimated. Two approaches are used as potential methods of assessing the care SAMM cases receive. "Mortality Index-MI" is defined as the ratio of maternal deaths among the SAMM cases to the sum of maternal deaths and SAMM cases [ 35 , 37 ]. It represents the proportion of women who presents with a SAMM and subsequently dies [ 37 ]. Another approach is to calculate the ratio of SAMM to mortality [ 8 , 18 ]. We attempted to calculate the ratio of SAMM to mortality for all studies included under the categories of disease-specific or organ-system based case identification criteria (see Additional file 1 ). We did not calculate MI because it was not clear from some reports whether reported maternal deaths were identified as SAMM or not prior to death. It is clearly illustrated in the studies that more SAMM cases are likely to die in resource-poor settings than in more developed country settings. For example, the studies conducted in Niger, Benin and Malaysia give the morbidity to mortality ratio as 11–12 [ 10 , 15 , 17 ] while this is 117–223 in studies conducted in Europe [ 11 , 18 ] in the category where disease-specific criteria are used. The same applies to the category of organ-system based criteria; morbidity: mortality ratio is 5–8 in South Africa [ 1 , 14 , 15 ] and 49 in Scotland [ 8 ]. These findings suggest that an indicator that relates SAMM to maternal deaths could be a useful method in assessing the care SAMM cases receive. However, the definitions and identification of cases should be standardised at least for similar settings and the indicator needs to be clearly defined. Conclusion Considering all complexities in definition and case-identification of SAMM, it is necessary that studies clearly describe their identification criteria for the cases. There is a clear need to set criteria to identify SAMM cases. Use of organ-system based criteria seems to be a more useful approach in identifying cases as variation in defining criteria can be avoided, particularly for similar settings. It would then be easier to establish summary estimates for SAMM prevalence which could serve as a measure of maternal health and quality of care indicator. Competing interests None declared. Authors' contributions AMG and LS coordinated the conduct of the systematic review of maternal morbidity and mortality. RCP and LS outlined the manuscript. LS reviewed the studies and wrote the initial draft. AMG and RCP substantially improved the manuscript. Supplementary Material Additional file 1 Table describing important variables of all studies included in the systematic review Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516581.xml
532392
Drosophila Spastin Regulates Synaptic Microtubule Networks and Is Required for Normal Motor Function
The most common form of human autosomal dominant hereditary spastic paraplegia (AD-HSP) is caused by mutations in the SPG4 (spastin) gene, which encodes an AAA ATPase closely related in sequence to the microtubule-severing protein Katanin. Patients with AD-HSP exhibit degeneration of the distal regions of the longest axons in the spinal cord. Loss-of-function mutations in the Drosophila spastin gene produce larval neuromuscular junction (NMJ) phenotypes. NMJ synaptic boutons in spastin mutants are more numerous and more clustered than in wild-type, and transmitter release is impaired. spastin -null adult flies have severe movement defects. They do not fly or jump, they climb poorly, and they have short lifespans. spastin hypomorphs have weaker behavioral phenotypes. Overexpression of Spastin erases the muscle microtubule network. This gain-of-function phenotype is consistent with the hypothesis that Spastin has microtubule-severing activity, and implies that spastin loss-of-function mutants should have an increased number of microtubules. Surprisingly, however, we observed the opposite phenotype: in spastin -null mutants, there are fewer microtubule bundles within the NMJ, especially in its distal boutons. The Drosophila NMJ is a glutamatergic synapse that resembles excitatory synapses in the mammalian spinal cord, so the reduction of organized presynaptic microtubules that we observe in spastin mutants may be relevant to an understanding of human Spastin's role in maintenance of axon terminals in the spinal cord.
Introduction “Pure” autosomal dominant hereditary spastic paraplegia (AD-HSP) is an inherited disease characterized by bilateral spasticity in the absence of other phenotypes (reviewed in Fink 2003 ; Reid 2003 ). Afflicted patients experience difficulty in walking and have a distinctive gait. Degeneration of the lateral corticospinal tracts, which contain the axons of cortical neurons that innervate primary limb motoneurons, is observed in the lumbar regions of the spinal cord in patients with AD-HSP. The distal segments of long dorsal root ganglion axons also display degeneration. No evidence is seen for cell death or for primary myelination defects, and the axons of primary motor neurons do not degenerate ( Maia and Behan 1974 ; Wharton et al. 2003 ). AD-HSP thus appears to selectively affect the distal regions of the longest axons within the spinal cord. Because pathology is usually confined to long spinal cord axons, it has been suggested that the primary defect in pure AD-HSP is in axonal transport or some other process required for maintenance of axon terminals. Perturbation of anterograde or retrograde axonal transport might selectively affect the longest axons, because they would be most vulnerable to a reduction in efficiency of transport of material to or from their terminals. About 40% of cases of pure AD-HSP are caused by mutations in the SPG4 gene, which encodes an AAA ATPase called Spastin ( Hazan et al. 1999 ). AAA ATPases are a large and diverse set of proteins that include an approximately 250–amino acid (aa) conserved domain containing Walker A and B ATP-binding motif sequences (reviewed in Confalonieri and Duguet 1995 ; Patel and Latterich 1998 ; Neuwald et al. 1999 ). They use energy obtained from ATP hydrolysis to catalyze assembly or disassembly of a variety of protein complexes. AAA proteins are involved in many cellular processes, including vesicle trafficking, protein degradation, and microtubule dynamics. Many AAA ATPases form hexameric rings, and it is thought that the ring structures are required for catalytic activity ( Vale 2000 ). Spastin is a member of the “meiotic” subgroup of AAA ATPases ( Frohlich 2001 ; Frickey and Lupas 2004 ), which contains proteins involved in vesicle trafficking and microtubule dynamics. The only member of the subgroup whose activities have been biochemically characterized is Katanin-60, which is the catalytic subunit of a microtubule-severing protein ( McNally and Vale 1993 ; Hartman et al. 1998 ). Katanin-60 and Spastin are homologous only within their AAA domains. However, cell culture studies have provided evidence that Spastin is also involved in microtubule dynamics. Expression of wild-type human Spastin in transfected cell lines and cortical neurons caused disassembly of the microtubule cytoskeleton, while a mutant Spastin lacking catalytic activity colocalized with tubulin ( Errico et al. 2002 ; McDermott et al. 2003 ). The mechanisms by which spastin mutations produce dominant spasticity phenotypes in humans are controversial. A wide variety of nonsense and missense mutations, but no complete gene deletions, have been found in families with AD-HSP. It has been suggested that dominance arises from haploinsufficiency ( Charvin et al. 2003 ). This model, however, would require that the processes in which Spastin participates are vulnerable to a 50% decrease in its enzymatic activity. Another possibility is that truncated or missense mutant Spastins function as dominant negatives. Hexameric AAA ATPase ring complexes might be especially vulnerable to the presence of nonfunctional (“poison”) subunits that assemble into rings but lack catalytic activity. Consistent with this idea, expression of a mutant Spastin that associates with microtubules but cannot catalyze severing altered organelle distribution in transfected cells ( McDermott et al. 2003 ). In the dominant negative model, AD-HSP might only occur when Spastin activity is eliminated or greatly reduced. In this study, we describe the phenotypes arising from mutation of the Drosophila ortholog of human spastin . We initially identified this gene in a gain-of-function screen, in which we found that its overexpression in neurons causes axons in the embryonic central nervous system (CNS) to converge onto the midline (Sun 2000). Overexpression of Spastin in muscles erases their microtubule networks, consistent with the idea that Spastin is a microtubule-severing protein. We made loss-of-function (LOF) spastin mutations, and found that they produce recessive phenotypes affecting the larval neuromuscular system. The Drosophila neuromuscular junction (NMJ) uses glutamate as its neurotransmitter and employs ionotropic glutamate receptors homologous to vertebrate AMPA receptors ( Schuster et al. 1991 ; Petersen et al. 1997 ; Marrus et al. 2004 ). It is organized into presynaptic boutons that are surrounded by a postsynaptic scaffold, and its synapses exhibit plastic behavior during development. These properties make the fly NMJ a useful genetic model system for the study of glutamatergic synapses in the mammalian brain and spinal cord ( Keshishian et al. 1996 ; Koh et al. 2000 ). During the period from larval hatching through the third instar stage, the number of boutons at each NMJ increases by up to 10-fold in order to keep pace with the growth of its muscle target. New boutons are added by a process of budding ( Zito et al. 1999 ). As these boutons mature, their microtubule cytoskeleton is thought to progress through a regulated series of alterations ( Roos et al. 2000 ; Pennetta et al. 2002 ). In this paper, we show that synaptic growth and function are altered in spastin mutant larval NMJs. Boutons are more numerous than in wild-type larvae, and synaptic transmission is impaired. These changes could result from alterations in synaptic microtubule dynamics, because we find that microtubule bundles are depleted from the distal boutons of NMJs in spastin -null mutants. This is surprising, because the fact that Spastin overexpression destroys microtubule networks might lead one to expect that its removal would increase the number of microtubules. Morphological and microtubule phenotypes are seen only for a total gene deletion, indicating that complete loss of Spastin function is required to alter synaptic microtubules in the fly system. The phenotypes we see are quite different from those described in a recently published study of perturbation of Drosophila spastin using RNAi methods ( Trotta et al. 2004 ). In particular, the changes in synaptic microtubules that occur in spastin LOF mutants are opposite to those reported in the RNAi perturbation paper. spastin is not an essential gene, but mutant adults have severely compromised motor behavior. Null mutants cannot fly or jump, they climb slowly, and they often drag their hind legs. While it is intriguing that spastin mutant flies display such movement phenotypes, further work will be required to determine whether Drosophila can provide a useful organismal model system for human AD-HSP. Nevertheless, insights into the cellular functions of Drosophila Spastin obtained from our work should be relevant to an understanding of Spastin's functions in human neurons. Results The Drosophila spastin Gene We identified spastin in an “EP” screen for genes involved in embryonic CNS development. EPs are P element derivatives with a block of “UAS” sites recognized by the yeast transcription factor GAL4 near one end ( Rorth 1996 ). An EP element inserted in the proper orientation upstream of a gene will drive its expression in a cell-specific manner when the insertion line is crossed to the appropriate promoter- GAL4 “driver” line ( Brand and Perrimon 1993 ). We generated approximately 6,000 new EP insertion lines and screened them by crossing to pan-neuronal ( Elav C155 ) and pan-muscle ( 24B ) GAL4 driver lines ( Lin and Goodman 1994 ; Luo et al. 1994 ). Those lines for which crosses to either driver generated reduced numbers (<20% of expected) of viable adult progeny containing both the EP element and the driver were saved. About 2% of lines (131) exhibited lethality or reduced viability with one of the drivers, and 62 of these were lethal or semilethal with both drivers. The T32 insertion on the third chromosome conferred complete lethality when crossed to either driver, and produced a neuronal-driver-dependent axonal phenotype (see below). To identify the gene driven by the T32 element, we cloned a genomic DNA fragment adjacent to the insertion site and used it to identify a full-length cDNA encoding a 758-aa protein that is a member of the AAA ATPase family ( Figure 1 A). The T32 EP element is inserted into the 5′ UTR, 222 nucleotides upstream of the predicted ATG start codon ( Figure 1 B; Sun 2000 ). Figure 1 Drosophila Spastin: Sequence Alignment and Gene Map (A) Clustal alignment of complete D. melanogaster and H. sapiens Spastin amino acid sequences and the AAA region of Drosophila Katanin-60. Identical and similar residues are highlighted in dark and light gray, respectively. (B) Map of the Drosophila spastin gene, including exons (black boxes) and introns, the position of the T32 EP insertion (nucleotide 58 of the 5′ UTR), and the regions deleted by imprecise excision in lines 10-12, 17-7, and 5.75. The 3′ end of the adjacent Rox8 gene is also shown. Arrows indicate direction of transcription. (C) Unrooted phylogenetic tree generated by the “neighbor” algorithm, showing relationships between the AAA domains of Spastins and their close relatives in human and fly. Dm CG3326 is the counterpart of the human fidgetin/fidlik gene pair, while CG1193 probably encodes a second fly ortholog of human Katanin-60. In the mouse, fidgetin mutations produce inner ear defects that cause head-shaking and circling behaviors ( Cox et al. 2000 ). The gene driven by T32 is orthologous to the human SPG4 (spastin) gene that is mutated in the most common form of AD-HSP ( Hazan et al. 1999 ). Mammalian Spastins are the only proteins that are homologous to both the C-terminal AAA domain and the N-terminal region of fly Spastin ( Figure 1 A). Spastin exhibits homology to all other AAA proteins only within its AAA domain (approximately aa 460–754 of fly Spastin). There are about 30 AAA proteins encoded in the Drosophila genome. The Drosophila Spastin sequence from aa 233 to the C terminus is 49% identical to that of human Spastin (616 aa). The AAA domains of the two proteins are 67% identical. The other region that is conserved between the Spastins (34% identity) corresponds to aa 233–404 of the fly sequence. The same region is also weakly related (26% identity) to human Spartin, the product of the SPG20 gene mutated in Troyer syndrome, a form of “complicated” HSP ( Patel et al. 2002 ; Ciccarelli et al. 2003 ). Spartin is not an AAA ATPase. The AAA protein with a known biochemical function that is most closely related to Spastin (41% identity in the AAA domain) is Katanin-60 ( Figure 1 A and 1 C; McNally and Vale 1993 ; Hartman et al. 1998 ). Drosophila Spastin Localizes to the Cytoplasm Human Spastin is thought to be a cytoplasmic protein expressed in many cell types, based on localization of epitope-tagged proteins expressed in transfected cells, and antibody staining of human tissue and neuronal cell lines ( Errico et al. 2002 , 2004 ; McDermott et al. 2003 ; Wharton et al. 2003 ). However, antibodies generated against Spastin were also reported to stain nuclei in several cell lines and mouse spinal cord neurons ( Charvin et al. 2003 ; Errico et al. 2004 ). This dual subcellular localization has been proposed to reflect a role for human Spastin in processes involving highly dynamic microtubule states, such as during cell division (reflected in Spastin's nuclear localization) and in axon outgrowth, suggested by the distal cytoplasmic staining of growing axons in culture ( Errico et al. 2004 ). To investigate the subcellular localization of Drosophila Spastin, we generated a variety of antibodies against different regions of the protein (see Materials and Methods ). We evaluated these antibodies by staining stage 16 embryos overexpressing Spastin in the striped Engrailed pattern from an engrailed - GAL4 driver. Two different antibodies revealed Spastin expression in the expected striped pattern, which includes a subset of CNS neurons ( Figure 2 A). Anti-Spastin antibodies stained both the cell bodies and the axons of these neurons ( Figure 2 B). In the epithelial portions of the Engrailed stripes, where the cells are flat and spread out, we observed that Spastin is expressed uniformly in the cytoplasm, but did not detect any nuclear staining ( Figure 2 C; see also Figure 7 E, showing cytoplasmic expression in muscles). Figure 2 Spastin Protein Localizes to the Cytoplasm (A) In embryo “fillets” in which Spastin overexpression is driven by the engrailed-GAL4 driver, a polyclonal antibody, pAb1239, generated against the C-terminal half of Spastin (aa 380–758) recognizes the characteristic striped pattern of Engrailed cells. Anterior is up; the CNS is the structure in the center, and the lateral epithelial stripes extend to either side. (B) An enlarged view of the CNS shows Spastin protein in these embryos localizing to neuronal cell bodies (arrow indicates the ventral unpaired midline [VUM] neurons), as well as in commissural and longitudinal axons (arrowheads). (C) A high-magnification view of the Spastin-positive epithelial cells shows that the protein fills the cytoplasm (arrow), and is excluded from the nucleus (arrowhead). Scale bar: (A) 25, (B) 10, and (C) 6.5 μm. Figure 7 Spastin Overexpression in Muscles Erases the Microtubule Network (A) An antibody against β3-tubulin stains body wall muscles and chordotonal cap cells in stage 16 wild-type embryos. Two abdominal hemisegments are shown; muscle fiber numbers are labeled in one. The cap cells (brackets) are difficult to distinguish in this panel because of high levels of muscle tubulin staining. They extend diagonally from about the middle of muscle 18 to muscle 22. Anterior is to the left, and dorsal is up. (B) When Spastin is overexpressed in muscles (genotype: G14-GAL4/+; T32/+ ), β3-tubulin staining is very weak and has a disorganized pattern in most muscle fibers, but an intact microtubule network is still present in the cap cells, which do not express this driver (brackets). The muscle fibers are misshapen and partially (arrowhead) or completely (arrow) detached from their insertion sites. (C–H) Similarly, the microtubule network (recognized by antibodies to α-tubulin) is almost eliminated by high-level Spastin expression in third instar larval muscles. Larvae of genotype UAS-spastin / MHC-GS-GAL4 213-3 ; spastin 5.75 / TM3Ser-ActGFP overexpress Spastin protein specifically in muscles to varying degrees. Wild-type larval muscles had undetectable levels of Spastin using pAb 1239 (C) and displayed a dense network of microtubule bundles in the muscle (D and E), as well as in trachea (D, arrow) and neurons (D, arrowhead denotes a terminal arbor). In contrast, larval muscles expressing high levels of Spastin (F) show only faint muscle microtubule staining (G and H), while tracheal (G, arrow) and neuronal (G, arrowhead) staining remain robust. spastin mRNA is expressed at low levels within the embryonic ventral nerve cord (VNC) in wild-type embryos ( Sun 2000 ; Kammermeier et al. 2003 ). Endogenous Spastin appears to be a very rare protein, and we have not been able to define a staining pattern in wild-type embryos or larvae that disappears in null mutants. An independently generated anti- Drosophila Spastin antibody was reported to stain the cytoplasm of both neurons and muscles in wild-type larvae, and staining was also detected at NMJ boutons ( Trotta et al. 2004 ). Taken together, these results suggest that fly Spastin, like human Spastin, is likely to be a widely expressed protein that is primarily localized to the cytoplasm. Spastin Overexpression in Neurons Causes Collapse of the Embryonic CNS Crossing the T32 insertion line to scabrous (sca) - GAL4, which is expressed in neuronal precursors and neurons ( Klaes et al. 1994 ), produced very strong CNS phenotypes. In Figure 3 , these are visualized by staining with monoclonal antibody (mAb) 1D4 ( Van Vactor et al. 1993 ), which labels a set of three longitudinal axon bundles. At 23 °C, embryos displayed abnormal midline crossing of the inner 1D4 bundle, and the entire VNC was narrowed at these crossing sites ( Figure 3 B). When crosses were performed at 29 °C, a temperature at which GAL4 transactivation is stronger, the VNC collapsed onto the midline and discrete longitudinal bundles were no longer apparent ( Figure 3 C). We also made transgenic lines bearing a full-length spastin cDNA driven by a UAS-containing promoter. After crossing to sca - GAL4, such lines produced even stronger phenotypes, in which the VNC collapsed at 23 °C ( Figure 3 D). Consistent with this observation, we found that more Spastin protein was made in driver × UAS-spastin embryos than in driver × T32 insertion embryos. Figure 3 Neuronal Overexpression of Spastin Causes Midline Convergence of Embryonic CNS Axons (A) Anti-Fasciclin II (mAb 1D4) staining of filleted late stage 16 control embryos reveals three longitudinal axon bundles (arrow) on each side of the midline. Anterior is up. (B) In sca-GAL4/+; T32/+ embryos raised at room temperature, overexpression of Spastin in neurons causes the ladder to constrict toward the midline (e.g., arrow). (C) Increased Spastin expression at 29 °C causes collapse of the CNS onto the midline (arrow). Longitudinal axon tracts are thin or absent (arrowhead). (D) A phenotype similar to that in (C) is produced by sca-GAL4 -driven expression of the UAS-spastin cDNA insertion at 23 °C. Arrow and arrowhead indicate same as in (C). spastin LOF Mutations Produce Larval NMJ Phenotypes To evaluate Spastin's functions during development, we generated several deletion mutations from the T32 insertion by imprecise excision. We mapped their breakpoints by sequencing, and these data are displayed in Figure 1 B. In line 10-12, about half of the first exon is deleted. The 17-7 deletion ends within the second intron and thus removes the entire first exon (encoding sequence up to aa 251). In both of these lines, DNA encoding the protein region conserved between human and Drosophila Spastin is still present. Deletion 5.75 removes the entire spastin gene, as well as the intergenic region and 129 bp at the 3′ end of the sequence of the adjacent predicted gene Rox8 . Rox8 contains RRM RNA-binding domains. The 5.75 deletion removes the C-terminal 43 aa of the Rox8 protein, but does not delete into the RRM domains. The function of Rox8 is unknown, and there are no existing Rox8 mutations ( Brand and Bourbon 1993 ). Because the only null spastin mutation also affects Rox8, we relied on rescue experiments (see below) to demonstrate that the phenotypes we describe for the null mutant are due to loss of Spastin. Flies homozygous for spastin 10-12 and spastin 17-7 have behavioral phenotypes, but they eclose at normal frequencies and are fertile (see below). In contrast, most homozygous spastin 5.75 pupae do not eclose. spastin 5.75 adults have very severe behavioral phenotypes, and both sexes are sterile. These results suggest that the 10-12 and 17-7 alleles are hypomorphic, and that the spastin 5.75 phenotype represents the null condition. RT-PCR analysis of cDNA from spastin 10-12 and spastin 17.7 animals indicated that low levels of truncated spastin transcripts are still produced (data not shown). These may direct synthesis of proteins initiated from internal ATGs that could retain partial function, since they include the entire conserved AAA domain. We could not detect anatomical phenotypes in embryos homozygous for any of the spastin mutations. However, we saw striking morphological changes in the NMJs of spastin 5.75 third instar larvae. In Figure 4 , the two predominant types of glutamatergic boutons at the NMJs, Ib (big) and Is (small), are visualized by double-staining larval fillets with antibodies against Synaptotagmin (Syt, magenta; Menon and Zinn 1998 ) and Discs-large (Dlg, green; Woods and Bryant 1991 ). Syt is a presynaptic protein involved in neurotransmitter release that is localized to boutons, while Dlg is a primarily postsynaptic scaffold protein localized to the subsynaptic reticulum that surrounds each bouton ( Littleton et al. 1993 ; Lahey et al. 1994 ). Figure 4 A and 4 B show that NMJ boutons are smaller and more numerous at the muscle 6/7 NMJ of spastin 5.75 larvae than in Canton S w− ( WCS ) control larvae. ( WCS was chosen as a control because, like the lines used to generate our EP insertion mutants, it is derived from a Canton S wild-type background, but it is also w−, like the T32 excision derivatives. WCS is also commonly used for behavioral experiments.) Other NMJs are affected in a similar manner (e.g., Figure 4 D– 4 F, showing muscle 4 synapses). The sizes of muscle fibers are normal in spastin mutants. Figure 4 Synaptic Boutons Are Smaller, More Numerous, and Clustered in spastin LOF Mutants (A–F) Representative A3 NMJs on muscles 6/7 (A–C) or muscle 4 (D–F) stained with antibodies against Dlg (green) and Syt (magenta) are shown for control larvae ( WCS; A and D), spastin 5.75 larvae (B and E), and larvae expressing Spastin from the spin-GAL4 driver in a spastin 5.75 mutant background (Rescue; C and F). Boutons are arranged in a linear pattern in WCS larvae, whereas in spastin 5.75 larvae their distribution is more clustered and individual boutons are smaller. These phenotypes are rescued by Spastin expression via the spin-GAL4 driver. Scale bars, 10 μm. (G) Quantitation of bouton numbers in spastin mutants relative to wild-type and rescued larvae demonstrates complete rescue of the null phenotype by spin - or Elav - GS - GAL4 -driven expression of Spastin. spastin -null mutants have on average 1.6-fold more type Ib boutons on muscle 4 compared to WCS control larvae. Similarly, spastin -null mutants (of genotype spin-GAL4/CyOKr-GFP; spastin 5.75 ) have 1.7-fold more boutons compared to their sibling rescued larvae (genotype spin-GAL4/UAS-spastin; spastin 5.75 ). Boutons are also 1.6-fold more numerous in spastin -null larvae from a neuronal rescue cross (genotype +/ CyOKr-GFP; Elav - GS - GAL4 , spastin 5.75 / spastin 5.75 ) compared to their siblings in which UAS-spastin is expressed in neurons postembryonically (genotype UAS-spastin / CyOKr-GFP; Elav-GS-GAL4, spastin 5.75 / spastin 5.75 ). To quantify the NMJ phenotype, we counted the numbers of boutons at the muscle 4 NMJs of segments A2 and A3, where boutons typically form on the internal surface of the muscle and are thus easily imaged. Dlg is expressed at much higher levels at Ib compared to Is boutons, allowing the two types of boutons to be distinguished and counted. Because of the greater variability in Is bouton number between NMJs, we focused our quantitative analysis on the type Ib boutons. However, the numbers of both bouton types were similarly affected in spastin 5.75 larvae. The number of Ib boutons per muscle 4 NMJ was increased by 1.6-fold relative to WCS in spastin mutants at room temperature (approximately 23 °C) ( Figure 4 G), and the boutons often formed dense clusters, particularly at the ends of NMJ branches ( Figure 4 E). This morphology was rarely observed in wild-type muscle 4 NMJs, where boutons were arranged more linearly ( Figure 4 D). The clustered boutons resemble the “satellite” boutons described by other investigators ( Torroja et al. 1999 ; Franco et al. 2004 ; Koh et al. 2004 ; Marie et al. 2004 ). Hypomorphic spastin 10-12 and spastin 17-7 mutants had bouton numbers that did not differ significantly from controls. To confirm that loss of Spastin produced the observed NMJ alterations, and to determine whether Spastin is required presynaptically or postsynaptically, we needed to evaluate rescue of the phenotype by expression of Spastin from a UAS-spastin cDNA insertion. This was difficult because of the early lethality produced by expression of Spastin from most drivers. UAS-spastin animals bearing pan-neuronal (Elav-GAL4), motoneuronal (OK6-GAL4), or pan-muscle (24B-GAL4 or G14 - GAL4) drivers did not survive to larval stages at 23 °C, and few larvae appeared even at 18 °C. However, third instar larvae in which Spastin expression from the cDNA was conferred by spinster (spin) - GAL4, a weak driver that functions in both neurons and muscles ( Sweeney and Davis 2002 ), did survive at 23 °C. We were also able to obtain larvae in which Spastin expression was induced in neurons postembryonically. This was done by crossing UAS - spastin to Elav-GeneSwitch (GS)-GAL4, a driver line bearing a neuronally expressed GAL4 derivative that is only active in the presence of the progesterone analog RU486 ( Osterwalder et al. 2001 ; McGuire et al. 2004 ). Newly hatched larvae from this cross were maintained on RU486-containing food until the third instar stage. To assay rescue, we combined the spin-GAL4 and Elav-GS-GAL4 drivers and UAS-spastin insertions separately with spastin 5.75 , crossed the driver and UAS-spastin lines together, and assayed NMJ phenotypes in the F1 driver-GAL4/UAS-spastin; spastin 5.75 larvae. In each cross, we compared the rescued larvae to their unrescued spastin mutant siblings (driver-GAL4; spastin 5.75 ), because the presence of the driver chromosome had effects on the absolute number of Ib boutons (see Materials and Methods ). We observed that the ratio of muscle 4 Ib bouton numbers in spastin mutant controls versus rescued larvae was 1.7 for spin-GAL4, and 1.6 for Elav-GS-GAL4 ( Figure 4 G). Since the ratio of bouton numbers for spastin 5.75 versus WCS was 1.6, this indicated that rescue was essentially complete in both cases. We also observed that the abnormal bouton clustering was eliminated in rescued larvae using either driver ( Figures 4 E and S1 ). These results demonstrate that loss of Spastin from neurons during larval development causes the NMJ bouton phenotypes seen in spastin 5.75 mutants. To examine the consequences of driver-dependent postembryonic neuronal expression of Spastin in a wild-type background, we also counted boutons in UAS-spastin; Elav-GS-GAL4 larvae grown on RU486 food ( see Materials and Methods ). We observed that these larvae had fewer boutons than their siblings (0.83 ± 0.07 fold change), and some of their boutons appeared larger ( Figure S1 C). This phenotype is very mild, but it does suggest that loss and increased expression of Spastin can produce opposite effects on the NMJ. Neurotransmitter Release Is Impaired in s pastin Mutants To evaluate whether spastin mutations cause alterations in the electrophysiological properties of the NMJ, we evaluated synaptic transmission at the muscle 6 NMJ in WCS, mutant, and rescued larvae raised at 18 °C. In spastin 5.75 larvae, there was a reduction in the amplitudes of evoked responses (excitatory junction potentials [EJPs]) to depolarization of the innervating nerves. Average EJP amplitudes in the null were reduced to 78% of the levels in control (WCS) larvae ( Figure 5 A and 5 B; p < 0.003). We also examined the average amplitude and frequency of responses to single vesicles of spontaneously released neurotransmitter (“mini” EJPs [mEJPs]). mEJP amplitude was increased slightly, to 117% of WCS levels ( Figure 5 A and 5 C; p < 0.03). There was no significant change in mEJP frequency ( Figure 5 D; p = 0.3). Figure 5 NMJs in spastin Mutant Larvae Display Reduced QC (A) Representative EJP (upper) and mEJP (lower) traces are shown for control (WCS), spastin -null mutant (spastin 5.75 ), and spin-GAL4/UAS-spastin; spastin 5.75 (Rescue) larvae. All recordings were from the A3 or A4 muscle 6 NMJ. (B) The average EJP amplitude is decreased by about 20% in spastin -null mutants (37 ± 2.0 mV, n = 26) relative to control (48 ± 2.4 mV, n = 14) and Rescue (42.4 ± 1.2 mV, n = 28) larvae, and is intermediate between control and null levels in hypomorphic spastin 5.75/17-7 transheterozygotes (41.9 ± 1.5 mV, n = 22). (C) The average amplitude of spontaneous events (mEJPs) is increased slightly in spastin nulls relative to control and Rescue larvae. (D) The average frequency of spontaneous events is not affected in spastin mutants compared to control. Rescue larvae had a slightly higher mEJP frequency. (E) Average QC, a measure of the amount of neurotransmitter released per action potential, is significantly lower in transheterozygotes (28 ± 1.3) versus control (35 ± 1.7), and reduced even further in spastin nulls (23 ± 1.2). This decrease is completely rescued by spin-GAL4 -driven rescue (30 ± 1.9, p = 0.1 compared to WCS ). (F) Average QC is temperature dependent in spastin 5.75 /spastin 17-7 transheterozygous larvae, but not in homozygous spastin- null or control larvae. QC measured in transheterozygotes raised at 18 °C (light gray bars) is intermediate between that of control and nulls. At room temperature (dark gray) and 29 °C (black bars), similar QC values are measured in transheterozygotes and null mutants. *, p < 0.05; **, p < 0.005. Quantal content (QC), a measure of the number of vesicles released per evoked event, was calculated by dividing the EJP amplitude by the average mEJP amplitude. Because the evoked EJP was decreased and the mEJP increased in spastin 5.75 mutants, QC was reduced to 67% of WCS levels in these larvae ( Figure 5 E; p < 3 × 10 −6 ). spastin 17-7 /spastin 5.75 larvae had EJP amplitude, mEJP amplitude, and QC values intermediate between those of spastin -null and control larvae. QC in these transheterozygotes was also decreased significantly, to 78% of WCS levels ( Figure 5 E; p < 0.002). The changes in EJP amplitude and QC observed in spastin 5.75 mutants were completely rescued by spin-GAL4 -driven Spastin expression. Average QC in rescued larvae was not significantly different from wild-type ( p > 0.1), and was 30% greater than in spastin 5.75 mutants ( p < 0.005; Figure 5 E). We also observed that the electrophysiological phenotypes of the hypomorphic spastin 17-7 /spastin 5.75 larvae were temperature sensitive. While the average QC in transheterozygotes was reduced to 78% of controls at 18 °C, this effect was exacerbated at higher temperatures. At 29 °C, QC was 54% of wild-type ( Figure 5 F). However, QC in control or in spastin 5.75 null larvae was unaffected by temperature. These results suggest that the N-terminally truncated Spastin protein that is probably made from the spastin 17-7 allele is temperature sensitive. At 29 °C, this hypomorphic allele behaves as a null with respect to NMJ electrophysiology. Finally, we examined escaper larvae ( Elav-GAL4/+; T32/+ ) overexpressing Spastin in neurons, and found that synaptic transmission in these larvae was not significantly different from wild-type (data not shown). spastin Mutant Adults Have Severe Behavioral Phenotypes We observed that only approximately 20% of homozygous spastin 5.75 pupae were able to eclose at room temperature compared to 94% of heterozygotes, and the adults that emerged had severe movement defects (see Video S1 ). They could not fly at all, and did not even appear to move their wings, although the wings inflated and straightened in a normal manner immediately after eclosion. They also did not jump spontaneously, but would jump if persistently prodded in the abdomen. Their legs were weak: when standing, the metathoracic legs often slipped out from underneath them, and during walking they often dragged these legs (see Video S1 ). They also had difficulty holding on to surfaces when they were upside down. These phenotypes were temperature dependent. Null mutant flies that developed at 18 °C eclosed at much higher rates (56%) than at higher temperatures and moved more normally. Flies homozygous for the hypomorphic mutations, spastin 10-12 and spastin 17-7 , eclosed at normal frequencies at all temperatures. To evaluate these movement defects, we assayed flight and climbing ability in spastin -null and hypomorph flies ( Figure 6 ). The flight assay could only be used for hypomorphs since null mutants were flightless. In this assay, flies were released into the top of a vertical cylinder that had been coated on the inside with oil ( Benzer 1973 ; Atkinson et al. 2000 ). Poor fliers who took longer to fly fell to the bottom or collided with the lower walls of the cylinder, while good fliers who responded rapidly to being dropped collided with the upper walls. A histogram of the distribution of oil-trapped flies along the height of the cylinder showed that more than half of the spastin hypomorphs did not fly in time to avoid falling to the bottom of the cylinder ( Figure 6 A). In contrast, the majority (approximately 75%) of the controls, including w 1118 (another Canton S -derived w− control) flies and flies homozygous for T32 (in the absence of a GAL4 driver; these have orange eyes), flew well enough to distribute themselves along the sides of the column. Interestingly, for those hypomorphs that did fly out to the sides, their distribution paralleled that of the controls, suggesting that flight responses in the column were relatively normal in this subpopulation of the mutants ( Figure 6 B). Figure 6 spastin Mutant Flies Have Compromised Motor Behavior and Reduced Lifespans (A) In a flight test assay, over twice as many adult spastin hypomorphs (spastin 10-12 and spastin 17-7 ) fail to fly before falling to the bottom of a cylinder, in comparison to w 1118 and T32 homozygous controls. (B) Although fewer than half of the hypomorphs fly, compared to more than 70% of controls, the distribution of collision sites of the fliers along the height of the cylinder parallels that of the controls, suggesting that these spastin mutations affect flying ability in some animals but not in others. (C and D) spastin mutants are compromised in their climbing ability. (C) All control (WCS) and nearly all spastin hypomorphs climb to the top of a vial in 30 s, but only 40% of spastin nulls do so. (D) Climbing velocity (measured for those flies that reach the top in 30 s) is 3.8 ± 0.2 cm/s in WCS ( n = 45), but only 1.8 ± 0.2 and 1.4 ± 1.1 cm/s in spastin 10-12 ( n = 28) and spastin 17-7 ( n = 38) flies, respectively, and 0.3 ± 0.1 cm/s in spastin 5.75 null mutants ( n = 17; p < 1 × 10 −8 for all relative to WCS ). (E) Lifespan curves. The curve inflection point at which WCS and hypomorph flies begin to die off at a rapid rate occurs at 30–35 d after eclosion, and more than 70% of hypomorph flies and 95% of wild-type flies are still alive at 30 d. In contrast, approximately 45% of spastin 5.75 null mutant flies die prior to 4 d after eclosion. However, the majority of the remaining null flies survive more than 25 d, so that the curve inflection point for nulls occurs only a few days before that for controls and hypomorphs. (F) Mean lifespan is 46 ± 2.7 d in WCS controls ( n = 39) compared to 35 ± 3.2 and 35 ± 3.4 d, respectively, in spastin 10-12 ( n = 24, p < 0.02) and spastin 17-7 ( n = 32, p < 0.006), and 7.6 ± 0.6 d in nulls ( n = 62, p < 10 −31 ). (G) Only about 10% of spastin 5.75 flies eclosing at room temperature are males, while 40%–45% are males for controls and hypomorphs. In the climbing assay, flies were tested for their ability to climb up the side of a vial within a limited time period. All WCS and almost all homozygous spastin 10-12 and spastin 17-7 flies, but only about 40% of spastin 5.75 flies, climbed to the top of the vial within 30 s ( Figure 6 C). This difference did not reflect a loss of geotactic behavior, since spastin 5.75 flies were typically found at the tops of their vials after several minutes. Overall, mean climbing velocity was approximately 9-fold slower for spastin -null mutants than for wild-type flies, while the hypomorphs were about 2-fold slower than controls ( Figure 6 D). We also measured the lifespan of the flies. Under our conditions, WCS flies lived an average of 46 d at 25 °C . spastin 10-12 and spastin 17-7 flies had somewhat shorter lifetimes, surviving an average of 35 d. Lifespan was dramatically reduced in spastin 5.75 flies, which lived an average of only 8 d ( Figure 6 E and 6 F). Examination of mortality curves ( Figure 6 E), however, revealed that, as in the case of flight ability in the hypomorphs, these flies had a bimodal lifespan distribution. Only 55% of flies were still alive at 4 d post-eclosion, but most of these (37% of the total) then remained alive until about 25 d post-eclosion. After this time, they rapidly died off, and no flies remained alive more than 33 d. Another spastin phenotype observed in adults was male-specific lethality ( Figure 6 G). For WCS and spastin hypomorphs, more than 40% of eclosed adults were males. However, only approximately 10% of eclosed spastin 5.75 flies were male. We do not understand the origins of this phenotype. In summary, spastin -null adult flies had severely compromised movement behavior and were short-lived, while spastin hypomorphs displayed weaker movement and lifespan phenotypes. We also examined rescue for these behavioral phenotypes. When compared to their non-rescued siblings (spin-GAL4/CyOKr-GFP; spastin 5.75 ) from the same cross, spin-GAL4/UAS-spastin; spastin 5.75 flies climbed better, were more coordinated, and lived longer ( Figure S2 ; Video S1 ), indicating that partial rescue was achieved. These flies were still very slow, and it is clear that spin-GAL4 -driven Spastin expression did not restore behavior to the levels characteristic of control flies such as WCS . However, genetic background effects made the precise efficacy of rescue achieved in this experiment difficult to determine (the non-rescued spastin 5.75 sibling flies bearing the driver and balancer chromosomes used in the rescue cross were much more unhealthy and slow-moving than spastin 5.75 flies without these chromosomes). Spastin Overexpression Erases Microtubule Networks In Vivo To investigate whether Drosophila Spastin affects microtubule networks, we overexpressed it in embryonic muscles using the G14 - GAL4 or 24B-GAL4 drivers, and then visualized muscle microtubules in late stage 16 embryos with an anti-β3-tubulin antibody that preferentially stains polymerized tubulin ( Buttgereit et al. 1996 ). In wild-type embryos, a complex network of microtubules aligned along the muscle axes was observed ( Figure 7 A; two segments are shown). This was clearly seen in both vertically oriented (18 and 21–24) and diagonally oriented (5,11, and 19) muscles. When Spastin was overexpressed in muscles, the muscle microtubule network completely disappeared ( Figure 7 B). The muscles themselves appeared rounded, and were partially or totally detached from their insertion sites. This detachment may have been a consequence of the dissolution of the microtubule network. Oriented microtubule networks could still be seen within the cap cells of the chordotonal organs in each segment ( Figure 7 B, brackets); these cells did not express the GAL4 driver and therefore did not overexpress Spastin. Dissolution of the microtubule network was therefore specific to cells in which Spastin was overexpressed. We also overexpressed Spastin in larval muscles by crossing the muscle-specific MHC-GS-GAL4 driver line to UAS-spastin flies. MHC-GS-GAL4 is RU486-inducible, but we could not feed the larvae with RU486 as this was lethal. However, MHC-GS-GAL4 also confers late RU486-independent expression in third instar larvae, and we were able to obtain escaper larvae from the cross and double-stain these for Spastin and α-tubulin. As shown in Figure 7 C– 7 E, larvae that lacked detectable Spastin expression had a dense network of muscle microtubules. In contrast, MHC-GS-GAL4/UAS-spastin larvae that had high levels of muscle Spastin displayed a dramatic reduction in microtubules, so that only faint and sparse microtubules were observed in the muscle fibers ( Figure 7 F– 7 H). The strongly staining microtubules still visible in these larvae are those of the neurons and tracheae, which do not overexpress Spastin ( Figure 7 G– 7 H). These results also show that this α-tubulin antibody preferentially recognizes polymerized tubulin, since the total amount of tubulin dimers would be the same in both sets of muscles (tubulin dimers are very stable proteins and are unlikely to be proteolyzed after severing). Interestingly, when we overexpressed Spastin in the embryonic or larval CNS, we did not observe an obvious alteration of the axonal microtubule architecture (see also Figure 8 ). This suggests that Spastin may be unable to disassemble stable axonal microtubule bundles. Nevertheless, the dramatic effects of Spastin overexpression on muscle microtubules suggest that the embryonic CNS collapse phenotype conferred by neuronal overexpression (see Figure 3 ) may also arise from breakdown of key neuronal microtubules during the axonal growth phase. Figure 8 The Distribution of Stable NMJ Microtubule Bundles Marked by the MAP1B-like Protein Futsch Is Altered in spastin Mutant Larvae (A–C) Anti-Futsch labels stable microtubule bundles in axons, NMJ boutons, and interbouton regions. The muscle 4 NMJs in segment A3 of third instar wild-type (A), spastin 5.75 (B), and spin-GAL4/UAS-spastin; spastin 5.75 (Rescue) (C) larvae were immunostained with anti-HRP (A and B) or anti-Syt antibodies (C) to label presynaptic boutons (magenta), and mAb 22C10 to label Futsch protein (A–C, green). Arrows and arrowheads mark the terminal boutons (those at the ends of synaptic branches); boutons marked by arrows in (A–C) are enlarged in insets in the middle and right panels to show examples of Futsch patterns. (A) In control ( WCS ) larvae, terminal boutons have both looped (arrowheads) and diffuse, punctate (right panel, arrows and inset) patterns of Futsch staining. (B) In spastin mutants, Futsch staining appears similarly strong in axon bundles (not shown) and along the main branches of the bouton arbor. More distal and terminal boutons, however, have diffuse or no Futsch staining (arrows and arrowheads). Note the absence of green staining in insets. (C) The distribution of Futsch staining is restored to the control pattern by spin-GAL4 -driven expression of Spastin in the mutant background (arrows and arrowheads indicate loops). Scale bar, 5 μm. (D and E) Quantitative assessment of Futsch staining data. Futsch staining at A2 and A3 muscle 4 NMJs was classified as continuous (bundles or splayed bundles), looped, or diffuse or undetectable (none) for each bouton. (D) The percentage of boutons exhibiting continuous or looped Futsch staining (relative to the total number of boutons for each NMJ) is decreased in spastin mutants relative to controls, while the percentage of boutons having diffuse or no staining is increased. In total, 58% ± 4.2% of boutons in controls have a continuous pattern of Futsch staining, while only 42% ± 1.5% do in mutants. Boutons in this class are predominantly along the major (more proximal) branches of the axon arbor. Similarly, 11% ± 1.7% of wild-type boutons have Futsch loops, but only 6.0% ± 1.1% do in mutants. Most mutant boutons show only diffuse or no Futsch staining (52% ± 2.2%, versus 32% ± 5.2% in controls). Futsch distribution is restored to the control pattern by spin-GAL4 - or Elav-GS-GAL4- driven expression of Spastin. (E) The difference in Futsch distribution is most pronounced at terminal boutons. There is no detectable Futsch staining in the majority of terminal boutons (65% ± 4.5%) in spastin mutants, compared to only 7.8% ± 5.8% of terminal boutons in wild-type larvae ( p < 2 × 10 −6 ). Futsch staining is restored in most terminal boutons of spin-GAL4 - or Elav-GS-GAL4- rescued larvae, with only 20% ± 3.7% and 19% ± 5.9% of boutons, respectively, showing no staining ( p = 0.09 compared to WCS ). Terminal bouton staining in larvae overexpressing Spastin in neurons was unaffected relative to controls ( p = 0.12). **, p < 0.005; *, p < 0.03 relative to WCS; n > 8 NMJs scored in all cases. Microtubule Bundles Are Depleted in Distal NMJ Boutons of spastin LOF Mutants The finding that Spastin overexpression erases the microtubule network in muscles suggested that the spastin LOF NMJ phenotypes could arise from alterations in microtubule networks. To investigate this, we first examined the distribution of Futsch, a microtubule-associated protein related to vertebrate MAP1B ( Hummel et al. 2000 ; Roos et al. 2000 ). Futsch staining is restricted to stable neuronal microtubule bundles. Because Futsch is not expressed in the underlying muscle, Futsch antibody staining provides the optimal method for quantitatively evaluating stable microtubules within NMJ boutons. At muscle 4 NMJs in wild-type larvae ( Figure 8 A) we observed continuous microtubule bundles stained by anti-Futsch (mAb 22C10; green) within axons and along the axis of each branch of the NMJ (delineated by anti-HRP, which labels neuronal membranes; magenta). The intensity of Futsch staining weakens in the distal portions of the branches. Consistent with earlier findings, we also observed distinctive “loops” of Futsch staining within some boutons ( Roos et al. 2000 ; Packard et al. 2002 ; Pennetta et al. 2002 ). Loops were typically observed in terminal boutons at the ends of branches. In some terminal boutons, however, we detected only punctate staining or no staining at all. This last case may reflect the limits of detection rather than the complete absence of Futsch protein in a bouton. We quantified Futsch distribution by dividing the patterns of Futsch staining in boutons into three classes: continuous (bundles or splayed bundles), looped, and diffuse or undetectable. In spastin 5.75 larvae ( Figure 8 B), there was a shift in the Futsch pattern toward less organized morphologies (i.e., diffuse/undetectable). At the muscle 4 NMJ of spastin mutant larvae, 74% and 54% as many boutons contained continuous and looped Futsch, respectively, as compared to WCS . In contrast, 63% more boutons in spastin mutants displayed only diffuse or no staining ( Figure 8 D). These differences were most pronounced at the distal ends of the synaptic branches ( Figure 8 B, arrows and arrowheads). 65% of terminal boutons in mutants had no detectable Futsch staining, as compared to 8% in WCS ( Figure 8 E; p < 2 × 10 −6 ). The Futsch distribution phenotypes were rescued by expression of Spastin from the spin-GAL4 or the RU486-induced, neural-specific Elav-GS-GAL4 drivers. Rescued larvae had Futsch staining patterns very similar to those seen in WCS ( Figure 8 C– 8 E). These results show that the reduction in stable synaptic microtubules seen in spastin LOF mutants is due to loss of Spastin from neurons during larval development. We also examined Futsch staining in Elav-GS-GAL4/UAS-spastin larvae grown on RU486 food, but saw no difference from the pattern in wild-type controls. Thus, stable microtubules at the NMJ do not break down when Spastin is overexpressed at the levels induced by this driver. Having demonstrated statistically significant differences in Futsch localization between control and spastin mutant NMJs, we then directly examined tubulin in muscle 4 NMJ boutons using fixation conditions that reduce muscle microtubule staining (see Materials and Methods ). The pattern of NMJ microtubules is complex, and is difficult to quantitatively analyze because of residual signal from microtubules in the underlying muscle. However, using the α-tubulin antibody described above, we were able to clearly visualize looped microtubule structures ( Figure 9 A, green) within anti-HRP-labeled boutons (magenta). These loops were present both along the branches and in the terminal boutons (inset). Figure 9 The Microtubule Network in NMJ Boutons Is Altered or Absent in spastin Mutant Larvae (A) In wild-type ( WCS ) larvae, an antibody against α-tubulin (green) reveals the distribution of the network of microtubule bundles within the A3 muscle 4 NMJ bouton arbor. Presynaptic bouton membranes are labeled by anti-HRP antibody (magenta). The microtubule network has a complex structure and extends into the terminal boutons (arrowheads and inset). Many proximal boutons have loops (arrows). Microtubules are also observed outside of the boundaries of the NMJ; these are within the muscle fiber, which also expresses α-tubulin. Staining of these muscle microtubules is minimized by the use of Bouin's fix. (B) In spastin 5.75 mutants, the microtubule network is much sparser than in controls, particularly in the distal boutons at the edges of the bouton clumps that are characteristic of spastin mutant NMJs (arrowheads and inset). Many of these distal boutons have little or no detectable α-tubulin staining. More proximal boutons still have tubulin loops, however (arrows). Scale bar, 5 μm. spastin 5.75 NMJs exhibited weaker α-tubulin staining than wild-type controls, particularly in terminal boutons ( Figure 9 B). Looped microtubule structures could be seen in some boutons in mutants. However, boutons at the ends of NMJ branches or at the outer edges of the bouton clumps that are characteristic of spastin NMJs often lacked any tubulin staining (inset). Thus, our results indicate that microtubule bundles are selectively depleted from the distal boutons of NMJs in larvae lacking Spastin protein. Discussion Mutations in the human spastin gene, which encodes an AAA ATPase, are the most common cause of pure AD-HSP. We identified the Drosophila Spastin ortholog (see Figure 1 ) in a gain-of-function screen (see Figure 3 ). Drosophila Spastin is a cytoplasmic protein that can also localize to axons (see Figure 2 ). spastin -null larvae have altered NMJs in which presynaptic boutons are more numerous and smaller than in wild-type, and are organized in dense clusters (see Figure 4 ). These changes in bouton number and organization are rescued by expression of Spastin in neurons (see Figures 4 and S1 ). QC, a measure of the number of vesicles of neurotransmitter released in response to an action potential, is reduced at NMJs in both null and hypomorphic spastin mutants (see Figure 5 ). spastin -null flies have severe movement defects. They cannot fly at all, and do not jump. They climb and walk very slowly, often drag their hind legs when walking (see Video S1 ), and have greatly reduced lifespans. spastin hypomorphs have milder phenotypes, displaying flying defects and a decrease in climbing speed (see Figure 6 ). Regulation of Synaptic Microtubule Networks by Spastin The AAA domain of Spastin is quite similar to that of Katanin-60, which is a microtubule-severing protein. To determine whether Spastin might also sever or otherwise alter microtubules in vivo, we overexpressed the protein in embryonic and larval muscles. Strikingly, this overexpression erases or greatly reduces the microtubule network (see Figure 7 ). These data are consistent with the finding that overexpression of human Spastin in transfected mammalian cells causes microtubule disassembly ( Errico et al. 2002 ; McDermott et al. 2003 ). Having demonstrated that Spastin can cause disassembly of microtubules in vivo, we then examined how its absence affects the synaptic microtubule cytoskeleton. Based on the overexpression phenotype, one might have expected that microtubules would be more stable or more numerous in spastin LOF mutants. However, our observations indicate the opposite: microtubule bundles are depleted in NMJ boutons when Spastin is absent. At the wild-type muscle 4 NMJ, boutons are arranged along linear axes. Continuous microtubule bundles run along the axes and connect to larger bundles within the innervating axon. Microtubules within boutons are typically arranged in loops and swirls. In spastin -null mutants, boutons are arranged in clumps, and the distal boutons of these clumps often lack any detectable tubulin staining (see Figure 9 ). Looped microtubule structures are present within some proximal boutons, however, and the bundles connecting the NMJ to the axon are still present. These results suggest that the absence of Spastin selectively affects the construction of the presynaptic microtubule cytoskeleton, and that the severity of the microtubule defects in a bouton are correlated with its distance from the NMJ's axonal branchpoint. We quantitated these defects using an antibody against the microtubule-associated Futsch protein, which defines a subpopulation of stable neuronal microtubule bundles. In wild-type larvae, Futsch staining forms continuous lines along the main branches of the NMJ. Some individual boutons have Futsch loops, while others display only diffuse staining. A comparison of wild-type and spastin -null larvae shows that the distribution of Futsch within boutons shifts from organized structures (bundles and loops) toward diffuse patterns or the absence of detectable staining. This effect is most pronounced at terminal boutons, and is rescued by neuronal expression of Spastin (see Figure 8 ). If Spastin's function in vivo is to disassemble microtubules, as suggested by our overexpression experiments (see Figure 7 ), why does its absence produce a paradoxical reduction in microtubules within the NMJ (see Figures 8 and 9 )? One possibility is that microtubule severing is required for movement of microtubules into or within the presynaptic region. Some evidence for this idea has been published. In one study, injection of function-blocking anti-Katanin-60 antibody into cultured sympathetic neurons reduced process outgrowth, and microtubules were 4- to 5-fold longer in antibody-injected neurons than in control cells ( Ahmad et al. 1999 ). More recent work demonstrated that expression of dominant-negative Katanin-60 reduces axonal outgrowth ( Karabay et al. 2004 ). These results were interpreted as indicating that Katanin is required for severing microtubules to a length that allows their transport along the axon to its growing tip. When Katanin is inhibited, microtubule segments may be too long to be efficiently transported, and this results in a reduction in axon outgrowth. Based on these findings, we suggest that the depletion of microtubules in the distal boutons of spastin mutant NMJs arises because severing of axonal microtubules by Spastin is necessary to generate microtubule polymers that are short enough to be efficiently moved into and through the presynaptic terminals. Perhaps Spastin normally excises sections of microtubules at branchpoints where NMJ branches leave the axon trunk, and these severed microtubule segments (or individual tubulin dimers) are then moved distally into the boutons of the NMJ as it grows. Is Spastin also involved in axon outgrowth or guidance, as suggested by its embryonic gain-of-function phenotype (see Figure 3 )? Clearly loss of Spastin activity in Drosophila does not strongly affect outgrowth, since the embryonic CNS axon ladder develops in a normal manner and motor axons reach their appropriate targets in spastin mutants. Furthermore, axonal and muscle microtubules are not detectably altered in spastin -null embryos. Severing of microtubules in vivo, however, may usually involve the actions of multiple severing proteins. In addition to Spastin, the Drosophila genome encodes three AAA ATPases whose AAA domains are closely related to that of vertebrate Katanin-60. These are Katanin-60, CG1193, and an ortholog of mammalian Fidgetins, CG3326 (see Figure 1 C). None of these proteins have been genetically characterized. Of these four proteins, Spastin is most distant from vertebrate Katanin-60, yet we have shown that Spastin overexpression causes microtubule disassembly in vivo (see Figure 7 ). Thus, our results suggest that all four fly proteins are microtubule-severing enzymes or proteins that otherwise facilitate disassembly of microtubule networks. Perhaps each is dedicated to severing microtubules in particular cellular and subcellular contexts, and their functions may be partially redundant. If so, generation of severe phenotypes in which microtubule networks are disrupted might require loss of two or more of these AAA ATPases. In mammals, Katanin and Spastin are both expressed in CNS neurons ( Wharton et al. 2003 ; Karabay et al. 2004 ), consistent with the idea that they could have overlapping functions. After this manuscript was submitted for initial review, a paper appeared on perturbation of Drosophila spastin using transgenic RNAi techniques ( Trotta et al. 2004 ). In direct contrast to our results, this paper concluded that (1) spastin is an essential gene (since crossing spastin RNAi flies to a ubiquitous GAL4 driver line was reported to produce lethality), (2) spastin RNAi larvae have reduced NMJs and an increase in synaptic transmission, and (3) loss of Spastin from neurons produces an increase rather than a decrease in stable microtubules in the NMJ. The conclusions in our paper are based on phenotypic analysis of spastin mutations that delete part or all of the coding region and on rescue of null mutant phenotypes by neuronal expression from a transgene. Our results show that spastin is not an essential gene: even spastin -null flies can eclose and live for several days, and spastin hypomorphs, which would be expected to more closely resemble most RNAi-perturbed flies, eclose at normal rates and have lifespans and behavior that do not greatly differ from wild-type (see Figure 6 ). We also determined that the 17-7 mutation, which removes more than one-third of the coding region, produces no detectable alterations in bouton number or NMJ microtubules and slightly decreases synaptic transmission, while spastin -null mutants have more boutons than wild-type larvae, a reduction in NMJ microtubule bundles, and more severely reduced transmission (see Figures 4 , 5 , 8 , and 9 ). In most transgenic RNAi work in Drosophila, different transgenic lines yield phenotypes that range from hypomorphic to near-null, and transgenic RNAi does not completely eliminate expression of the target protein (e.g., Billuart et al. 2001 ; Kalidas and Smith 2002 ). In the Trotta et al. paper, it is unclear whether more than one transgenic RNAi line was analyzed, but RNAi is described as reducing the level of Spastin protein expression by less than 4-fold. Our findings on Spastin hypomorphic phenotypes imply that such RNAi larvae would not have morphological or microtubule bouton phenotypes and that adult flies would be relatively healthy. We do not understand the origin of the discrepancies between the two sets of results. Implications of Studies of Drosophila Spastin for the Understanding of Human AD-HSP spastin -null adult flies have severe movement defects. Their hind legs are particularly weak (see Video S1 ); this is interesting in light of the restriction of symptoms to the legs in most AD-HSP patients. Other aspects of the spastin mutant adult phenotypes also resemble observations made in human AD-HSP patients. The penetrance of the human spasticity phenotype is highly variable, so that some individuals carrying a spastin mutation appear unaffected, while others with the same mutation are confined to wheelchairs. In our experiments, we observed that some spastin hypomorphs exhibit normal flying behavior in a cylinder assay, while most fail to fly and crash into the cylinder base (see Figure 6 A and 6 B). Half of the spastin -null adults die within 4 d, but most of the survivors then live more than 25 d (see Figure 6 E). The selective male lethality we observed (see Figure 6 G) is also interesting in light of the discovery of a large SPG4 pedigree in which only males exhibit AD-HSP phenotypes ( Starling et al. 2002 ). Despite these apparent parallels, there is no evidence at present that the fly behavioral phenotypes arise through mechanisms related to those that cause human AD-HSP. The anatomy of the Drosophila nervous system is quite different from that of the mammalian spinal cord. Furthermore, AD-HSP is thought to be a neurodegenerative disease that progresses over a period of years, and it is unclear whether neurodegeneration as a result of spastin mutations could occur during the short lifespan of Drosophila . Further work will be required to determine to what extent the Drosophila system can provide an organismal model for AD-HSP pathology. The clearest implications of our work for AD-HSP emerge from the analysis of the cellular phenotypes arising from loss of Spastin. We show that the absence of Spastin alters the microtubule network at nerve terminals. Microtubule bundles are depleted in the distal boutons of the NMJ, which is a glutamatergic synapse that resembles excitatory synapses within the mammalian spinal cord. These results suggest that microtubules within the terminals of neurons in the human spinal cord could also be disordered or absent in patients with AD-HSP. The terminals of these neurons might eventually degenerate as a consequence of these microtubule defects, leading to a selective loss of distal axon segments within the spinal cord. Materials and Methods Genetics and molecular biology For the overexpression screen, approximately 6,000 new EP insertion lines were generated by crossing an X chromosome EP line, EP55, to a SbΔ2-3 transposase line. EP insertions on Chromosome II or III were crossed to the pan-neuronal driver line Elav C155 -GAL4 and the muscle driver line 24B-GAL4. Embryos from EP lines that exhibited less than 20% viability in combination with either driver were immunostained with mAb 1D4. Eighteen lines had embryonic CNS and/or motor axon defects when crossed to Elav C155 -GAL4, including line T32 . The flanking genomic region of T32 was cloned and sequenced by plasmid rescue. This sequence matched three overlapping EST's from BDGP, one of which, GH11184, contained the complete ORF of the gene downstream of the T32 insertion. This cDNA was sequenced in its entirety. Unrooted trees for Spastin and its closest relatives were constructed using six different algorithms (fitch, kitsch, neighbor, upgma, protein maximum likelihood, and parsimony) from the Phylip package, based on alignment to the PFAM AAA consensus. For the spastin excision lines, alleles 10-12, 17-7, and 5.75 were generated via imprecise excision of EP T32 using SbΔ2-3 . All alleles were homozygous viable, and their deletions were mapped by PCR and sequencing of larval or adult genomic DNA. Allele 5.75 causes sterility in both sexes. For the spastin rescue construct, the UAS-spastin cDNA construct was made by subcloning a 2.9-kb BglII fragment from GH11184 into the BglII site of pUAST ( Brand and Perrimon 1993 ). This fragment contains the spastin cDNA up to 350 bp after the stop codon (excluding 681 bp of the 3′ UTR) and including 28 bp of polylinker sequence from the pOT2 plasmid at the 5′ end. The construct was injected at approximately 300 ng/μl into KiΔ2-3 embryos and several transgenic lines recovered; experiments described here used the Chromosome II insertion line 8-3-5. Rescue of spastin -null phenotypes by spin-GAL4 -driven expression was assayed by crossing UAS-spastin/CyOKr-GFP; spastin 5.75 /TM3SerAct-GFP to spin-GAL4/CyOKr-GFP; spastin 5.75 /TM3SerAct-GFP . The numbers of Ib boutons in rescued larvae (UAS-spastin/spin-GAL4; spastin 5.75 /spastin 5.75 ) were compared to those in unrescued sibling mutants (spin-GAL4/CyOKr-GFP; spastin 5.75 /spastin 5.75 ) to calculate the ratios used to determine the efficacy of rescue (see Figure 4 G). This was done because we observed that the presence of the driver chromosome in the background increased the absolute number of Ib boutons, so that the appropriate control was to the sibling mutants also bearing this chromosome. Rescue of spastin -null phenotypes by postembryonic Elav-GS-GAL4 -driven expression was assayed by crossing UAS-spastin/CyOKr-GFP; spastin 5.75 /TM3SerAct-GFP to Elav-GS-GAL4, spastin 5.75 /TM6B and raising the larvae on RU486-containing food as described ( Osterwalder et al. 2001 ; McGuire et al. 2004 ). The numbers of Ib boutons in rescued larvae (UAS-spastin/+; Elav-GS-GAL4, spastin 5.75 /spastin 5.75 ) were compared to those in unrescued sibling mutants (CyOKr-GFP/+; Elav-GS-GAL4, spastin 5.75 /spastin 5.75 ) as described above for spin-GAL4 rescue. The effect of postembryonic neuronal overexpression of Spastin in a wild-type background was assayed by counting Ib bouton numbers in UAS-spastin/+; Elav-GS-GAL4, spastin 5.75 /TM3SerAct-GFP larvae from this same cross and comparing them to the numbers for UAS-spastin/+; Elav-GS-GAL4, spastin 5.75 /spastin 5.75 . The wild-type control lines used were Oregon R, w 1118 , and WCS , a Canton S line backcrossed ten times to white (gift of Anne Simon). For larval overexpression experiments, the MHC-GS-GAL4 213-3 driver was used to induce muscle-specific expression of UAS-spastin at 23 °C in the absence of RU486. Expression induced by this driver was confirmed in a separate cross using UAS-EGFP . Generation of Spastin antibodies Regions of the spastin cDNA encoding aa 136–416 (pGEX-T32PvuII), 1–167 (pAcG2T-T32BamRIa), and 380–758 (pAcG2T-T32BBA) were subcloned from GH11184 into bacterial (pGEX) or baculovirus (pAcG2T) expression vectors. Expressed protein (in the form of inclusion bodies for T32PvuII) was injected into guinea pigs (Covance, Princeton, New Jersey, United States) and the antiserum tested on en-GAL4/+2/+ embryos. Antisera against T32PvuII (86EX) and T32BBA (1239) both showed strong staining in the Engrailed pattern; thus, both recognized overexpressed Spastin. 86EX was purified by incubation with membrane-bound pGEX-T32PvuII protein and subsequent elution with 100 mM glycine (pH 2.5) followed by neutralization with 3M Tris (pH 8.8). pAb1239 was affinity purified using the immunogen bound to Affi-gel10 beads (Bio-Rad Laboratories, Hercules, California, United States), followed by preabsorption with spastin 5.75 larval fillets. Embryonic and larval immunocytochemistry Stage 16 embryos were fixed and stained using standard methods ( Patel 1994 ) with mAb 1D4 (1:5) or anti-β3-tubulin (1:500; a gift of R. Renkawitz-Pohl and D. Buttgereit). Third instar larvae were live-dissected in room temperature HL3 (see below) or PBS and fixed in 4% paraformaldehyde or Bouin's fix (for staining of NMJs with anti-tubulin) for 25 min. Primary antibodies used on larvae were mouse anti-Syt at 1:400 ( Menon and Zinn 1998 ), rabbit anti-Dlg (1:2,000; a gift of D. Woods and P. Bryant), mAb 22C10 (anti-Futsch, 1:50; Developmental Studies Hybridoma Bank, Iowa City, Iowa, United States), rabbit anti-HRP (1:250, Cappel, MP Biomedicals, Irvine, California, United States), mouse anti-α-tubulin DM1A (1:500; Sigma, St. Louis, Missouri, United States). Anti-Spastin pAb 1239 was used at 1:300. Staining was visualized with HRP-conjugated goat anti-mouse secondary (1:200; Jackson Laboratory, Bar Harbor, Maine, United States) or Alexa-Fluor 488 and 568 anti-mouse, -rabbit, or -guinea-pig secondaries (1:200; Molecular Probes, Eugene, Oregon, United States). All fluorescently labeled samples were imaged by acquiring z -series projections with a Zeiss (Oberkochen, Germany) 510 inverted confocal microscope and 63×/1.4 n.a. or 100×/1.2 n.a. PlanApo objectives. Only larval segments A2 and A3 were analyzed. Individual boutons were defined as a Syt-positive area encircled by Dlg-positive staining (see Figure 4 ), or a Syt- or HRP-positive varicosity in the synaptic arbor (see Figure 8 ). All type Ib boutons were scored for each muscle 4 NMJ. Average bouton numbers per muscle 4 (see Figure 4 ) were as follows (mean ± s.e.): WCS , 44 ± 2.6 ( n = 26 NMJs); spastin 5.75 mutant, 68 ± 2.4 (38); spin rescue, 50 ± 4.9 (26); mutant sibling with spin driver chromosome, 86 ± 4.6 (8); Elav-GS rescue, 70 ± 3.0 (19); and mutant sibling with Elav-GS driver chromosome, 110 ± 9.3 (21). p < 0.02 by one-way ANOVA for all paired comparisons. Electrophysiology Intracellular recordings were obtained at 18 °C, using sharp microelectrodes (boroscilicate glass, 1.0 mm OD; 18–35 MΩ resistance; World Precision Instruments Sarasota, Florida, United States) filled with 3M KCl, from body wall muscle 6 (segments A3 or A4) of filleted third instar larvae, following standard methods ( Jan and Jan 1976 ). Larvae were bathed in HL3 solution ( Stewart et al. 1994 ), in mM: NaCl, 70 (EM Science, Gibbstown, New Jersey, United States); KCl, 5; MgCl 2 , 20; NaHCO 3 , 10; HEPES, 5; Sucrose, 115; Trehalose, 5; and CaCl 2 , 1 (Sigma). Larvae were visualized with a 5×/0.10 n.a. Olympus (Tokyo, Japan) objective on an Olympus BX50WI microscope. EJPs were evoked by pulling the cut end of the innervating segmental nerve into a heat-polished suction electrode and passing a depolarizing pulse sufficient to depolarize both motoneurons (Grass SD9 stimulator). For each experiment, 10–15 single EJPs evoked at 0.2 Hz were recorded, and then spontaneous mEJPs recorded for 1 min afterwards. Only recordings with resting membrane potential below −60 mV were acquired. The average resting membrane potential for control (WCS) larvae was −72.2 mV, and did not differ significantly from any of the experimental groups. Average muscle input resistance in control larvae was 8.9 MΩ, and differed significantly only from the input resistance determined for spastin 5.75 / spastin 17-7 transheterozygotes (7.5 MΩ p < 0.04). Recordings were performed using an Axon Instruments (Foster City, California, United States) Axopatch 200B amplifier with CV203BU headstage operating in current clamp mode. The signal was low-pass filtered at 5 kHz, digitized through an Axon Instruments Digidata 1322A 16-bit acquisition system, and recorded using Axon Instruments Clampex 8.2 software. Mean EJP amplitude was determined by averaging all single EJPs with Axon Instruments Clampfit 8.2, and corrected for nonlinear summation according to McLachlan and Martin (1981) and Feeney et al. (1998) . mEJPs were measured using Mini Analysis Program (Synaptosoft, Decatur, Georgia, United States). mEJPs with a slow time course arising from neighboring electrically coupled muscle cells were excluded from analysis ( Zhang et al. 1998 ). QC for a given NMJ was estimated by dividing the average EJP amplitude by the average mEJP amplitude. Statistics were calculated using one-way ANOVA. Adult phenotypes Eclosion rates were determined by counting numbers of empty versus full (dead) pupae on the sides of bottles in which flies had been allowed to lay for comparable time periods. The flight test assay was performed at room temperature using an opaque cylinder (a 52-cm-tall pipette washer, 18 cm in diameter) coated on the inside with fresh mineral oil. Flies of a given genotype were dumped through a hole in the center of a lid at the top. The cylinder was divided into bins along its height, and the number of flies per bin counted. Flies of different genotypes were age-matched; more than 200 flies were counted for each. The climbing assay was performed on 4–5 d old flies maintained individually in vials. Climbing velocity for each fly was measured by transferring it to an empty vial, banging it to the bottom, and then measuring either the time required to reach the top of the vial or the maximum distance it climbed in 30 s, whichever came first. Three trials were performed, and the best speed was used. For lifespan tests, flies were maintained at 25 °C, transferred every 3 d to fresh food vials, and their lifespan noted. Supporting Information Figure S1 Spastin Expression in Neurons Rescues the spastin Mutant Morphology Representative muscle 4 NMJs stained with antibodies against Dlg (green) and Syt (magenta) are shown for (A) spastin 5.75 mutant (genotype +/ CyOKr-GFP; Elav - GS - GAL4 , spastin 5.75 / spastin 5.75 ), (B) neuronally rescued (UAS-spastin / CyOKr-GFP; Elav-GS-GAL4 , spastin 5.75 / spastin 5.75 ), and (C) neuronally overexpressing (UAS-spastin /+ ; Elav-GS-GAL4 , spastin 5.75 / TM3Ser-ActGFP) larvae. The clustered, smaller, and more numerous boutons observed in mutant NMJs (A, arrowhead) are absent in neuronally rescued larvae, which resemble controls ( WCS; see Figure 4 D). Spastin overexpression in neurons produces an opposite morphological phenotype compared to the loss of function: boutons appear slightly larger than in wild-type, and bouton counts show that they are reduced in number (83% of control; see text). Scale bar, 5 μm. (1.4 MB TIF). Click here for additional data file. Figure S2 Adult Behavior Is Partially Rescued by spin-GAL4 -Driven Expression of Spastin in the spastin -Null Background Behavioral tests were performed on flies from the four genotypes arising from the spin-GAL4 rescue crosses, raised at 18 o C. These genotypes were (1) spin-GAL4/UAS-spastin; spastin 5.75 (spin Rescue), (2) spin-GAL4 / CyOKr-GFP; spastin 5.75 (non-rescued spastin mutant, denoted 5.75[R]), (3) spin-GAL4/UAS-spastin; spastin 5.75 / TM3SerAct-GFP (Cy + Ctrl; heterozygous for the spastin mutation), and spin-GAL4 / CyOKr-GFP; spastin 5.75 /TM3SerAct-GFP (Cy Ctrl; heterozygous for the spastin mutation). (A) Climbing behavior. None of the spastin mutants (0%) from these crosses (5.75[R]; n = 21) reached the top of the vial in the prescribed 30 s time limit, compared to 8% for Rescue flies ( n = 75), and 100% for both spastin/+ controls ( n = 39 and 21). Twenty-seven percent of mutants (5.75[R]) did not climb at all, compared to only 4% of the Rescue flies and 0% of the spastin/+ controls. Thus, although both genotypes in the mutant background (homozygous for spastin 5.75 ) were much weaker than either spastin 5.75 heterozygous control, Rescue flies showed improved climbing ability compared to the mutants. (B) Similar to the results in (A), mean lifespan in spastin mutants (10 ± 1.3 d, n = 32) was significantly rescued by spin -driven expression of spastin (16 ± 1, n = 95, p < 0.004), although lifespans were much shorter in spastin 5.75 homozygotes than in heterozygous spastin/+ controls (43 ± 3.1 and 44 ± 2.7; n = 20 each). (218 KB PDF). Click here for additional data file. Video S1 Motor Behavior in Control, spastin 5.75 Mutant, and spin-GAL4 -Rescued Flies Flies are shown moving in a vial. Segment 1: Wild-type. One female and one male WCS fly are shown. Note the rapid rate at which they walk, as well as exhibiting climbing, jumping and flying behaviors. When still, their legs are controlled, and they are able to walk upside-down (out-of-focus fly near end of segment) for prolonged periods without falling. Segment 2: Mutant. One spastin 5.75 female is shown. Leg weakness is obvious, particularly for the mesothoracic and metathoracic legs, both when walking and standing still. She climbs poorly, and when rotated so that she is upside-down, is unable to maintain a hanging position. No wing movement or jumping is observed. Segment 3: Rescue. In spin-GAL4/UAS-spastin; spastin 5.75 flies, Spastin expression via the spin-GAL4 driver partially rescues the movement defects seen in spastin 5.75 mutants. Two males are shown, followed by one female. Note their improved leg steadiness, velocity, and hanging ability. These flies can also jump spontaneously. The female appears to be less fully rescued; however, she is able to walk upside-down for prolonged periods, and exhibits wing movement. (7.4 MB MOV). Click here for additional data file.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC532392.xml
449907
Great Ape Genomes Offer Insight into Human Evolution
null
Some primatologists have argued that to understand human nature we must understand the behavior of apes. In the social interactions and organization of modern primates, the theory goes, we can see the evolutionary roots of our own social relationships. In the genomic era, the age-old question, What makes us human? has become, Why are we not apes? As scientists become more adept at extracting biological meaning from an ever expanding repository of sequenced genomes, it is likely that our next of kin will again hold promising clues to our own identity. Lineage-specific gene gains and losses in humans and great apes Many comparative genomics studies have looked to our more distant evolutionary relatives, such as the mouse and even yeast, to help interpret the human genome. Because the genomes of mice, yeast, and humans have diverged significantly since their last common ancestor—about 75 million years ago for mouse and human, and about 1 billion years ago for yeast and human—there are enough differences between the functional and nonfunctional regions to home in on biologically significant sequences, based on their similarity. Sequences that are similar, or conserved, in such divergent species are assumed to encode important biological functions. These comparative studies have successfully identified and characterized many human genes. And a similar approach comparing primate genomes can help scientists understand the genetic basis of the physical and biochemical traits that distinguish primate species. In this approach, however, rather than looking for genes that are shared across many species, scientists look for those that are unique to a species. One of the primary agents of genome evolution is gene duplication. Duplicated genes provide the raw material for the generation of novel genes and biological functions, which in turn allow the evolution of organismal complexity and new species. (For more on duplicated genes, see the primer by Hurles in this issue.) James Sikela and colleagues set out to compare gene duplications between humans and four of our closest primate relatives to find the genetic roots of our evolutionary split from the other great apes. Collecting the DNA of humans, chimpanzees, bonobos, gorillas, and orangutans from blood and experimental cell lines, the researchers used microarray analysis to identify variations in the number of copies of individual genes among the different species. They analyzed nearly 30,000 human genes and compared their copy numbers in the genomes of humans and the four great apes. Overall, Sikela and colleagues found more than 1,000 genes with lineage-specific changes in copy number, representing 3.4% of the genes tested. All the great ape species showed more increases than decreases in gene copy numbers, but relative to the evolutionary age of each lineage, humans showed the highest number of genes with increased copy numbers, at 134. Many of these duplicated human genes are implicated in brain structure and function. The gene changes identified in the study, the authors conclude, likely represent most of the major lineage-specific gene expansions (or losses) that have taken place since orangutans split from the other great apes, some 15 million years ago. (Humans diverged from their closest cousins, the chimp and bonobo, roughly 5 million to 7 million years ago.) And because some of these gene changes were unique to each of the species examined, they will likely account for some of the physiological and morphological characteristics that are unique to each species. One cluster of genes that amplified only in humans was mapped to a genomic area that appears prone to instability in human, chimp, bonobo, and gorilla. This region, which corresponds to an ancestral region in the orangutan genome, has undergone modifications in each of the other descendent primate species, suggesting an evolutionary role. In humans, gene mutations in this region are also associated with the inherited disorder spinal muscular atrophy. This fact, along with the observation that there are human-specific gene duplications in this region, suggests a link between genome instability, disease processes, and evolutionary adaptation. In their genome-wide hunt for gene duplications and losses in humans and great apes, Sikela and colleagues have highlighted genomic regions likely to have influenced primate evolution. With the impending release of the chimp genome and more primate sequences to follow, scientists can take advantage of both sequence-based and microarray-based genome information to wrest additional insights from our primate cousins and flesh out the details of the human story.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC449907.xml
521488
UVA/UVA1 phototherapy and PUVA photochemotherapy in connective tissue diseases and related disorders: a research based review
Background Broad-band UVA, long-wave UVA1 and PUVA treatment have been described as an alternative/adjunct therapeutic option in a number of inflammatory and malignant skin diseases. Nevertheless, controlled studies investigating the efficacy of UVA irradiation in connective tissue diseases and related disorders are rare. Methods Searching the PubMed database the current article systematically reviews established and innovative therapeutic approaches of broad-band UVA irradiation, UVA1 phototherapy and PUVA photochemotherapy in a variety of different connective tissue disorders. Results Potential pathways include immunomodulation of inflammation, induction of collagenases and initiation of apoptosis. Even though holding the risk of carcinogenesis, photoaging or UV-induced exacerbation, UVA phototherapy seems to exhibit a tolerable risk/benefit ratio at least in systemic sclerosis, localized scleroderma, extragenital lichen sclerosus et atrophicus, sclerodermoid graft-versus-host disease, lupus erythematosus and a number of sclerotic rarities. Conclusions Based on the data retrieved from the literature, therapeutic UVA exposure seems to be effective in connective tissue diseases and related disorders. However, more controlled investigations are needed in order to establish a clear-cut catalogue of indications.
Background Unlike UVB radiation that can penetrate at the most into the papillary dermis, longer wavelengths in the UVA region have the capacity to reach the subcutis as well. Accordingly, as well as due to its lesser antiproliferative activity, UVB irradiation has not been established in the treatment of sclerotic disorders except for occasional cases of graft-versus-host disease (GvHD) [ 1 , 2 ]. Hence, this review examines different modalities of UVA phototherapy in the treatment of connective tissue diseases and related disorders. The term irradiance (e.g., in mW/cm 2 ), which is the most commonly used term in photobiology, relates to the subject (e.g., patient) struck by the irradiation. In photobiology, the time integral of the irradiance is commonly expressed as fluence (e.g., in J/cm 2 ), or even more loosely as dose [ 3 , 4 ]. Even though mostly combined with 8-methoxypsoralene or other photochemotherapeutic agents, broad-band UVA irradiation (315–400 nm), containing both UVA1 (340–400 nm) and UVA2 (315–340 nm), was used as monotherapy e.g. in the treatment of atopic dermatitis [ 5 ]. However, this phototherapeutic option was being replaced more frequently by the usage of irradiation devices which allow a more effective treatment by the administration of selected spectra. By eliminating shorter wavelengths in the UVA2 region adverse effects such as erythema are minimized and therapeutically effective higher UV doses can be administered. Thus, today broad-band UVA may play a subordinated role in modern phototherapy anymore, except for the combined application with psoralenes, even though it is still used for a large number of patients due to its wide availability and although it has not yet been directly compared with UVA1 for many sclerosing disorders. Conventional UVA1 treatment emitts wavelengths mainly between 340 and 400 nm, but may also produce scattered radiation >530 nm including infrared irradiation (780–3000 nm). Three different pattern of UVA1 dosage have been described: high-dose UVA1 phototherapy applying UVA1 doses ranging from 90–130 J/cm 2 single (975–1840 J/cm 2 cumulative) irradiation, medium-dose UVA1 phototherapy admitting doses between 20–90 J/cm 2 single (300–975 J/cm 2 cumulative) UVA1 and last but not least low-dose UVA1 phototherapy administering single UVA1 doses of ≤ 20 J/cm 2 or a cumulative doage ≤ 300 J/cm 2 , respectively [ 6 - 9 ]. Conventional UVA1 phototherapy may be accompanied by extensive heat load predominantly generated by infrared irradiation (780–3000 nm) and/or insufficient cooling systems of the phototherapy devices. For this reason, within the last years, lavish new UVA1 equipment was developed containing a special filtering and cooling system, in which a considerable amount of wavelengths >530 nm are eliminated and consequently the heat load due to heat-producing infrared radiation is strongly diminished [ 10 ]. Since more than two decades, the combination of oral 5- or 8-methoxypsoralen followed by broad-band UVA exposure is an effective treatment option in a widespread number of indications such as psoriasis and cutaneous T cell lymphoma [ 11 - 13 ]. Psoralens specifically belong to the best characterized agents of photosensitizing chemicals. Following its administration low-dose UVA irradiation is applied including an individual progression depending on the formation of the so-called PUVA erythema. In general, the overall mean cumulative dose has been found to comprise ≤ 400 J/cm 2 UVA [ 14 ]. In contrast to systemic PUVA therapy topical PUVA, for example applied as PUVA bath, PUVA shower or PUVA cream, is mainly characterized by absent/reduced systemic side effects and the restriction to a selective limited lesional area [ 15 ]. As a consequence, systemic PUVA has mostly been replaced by topical PUVA treatment representing an efficient well-tolerable alternative to oral methoxypsoralen administration. Generally, overall mean cumulative dosage has been found to comprise ≤ 200 J/cm 2 UVA [ 14 ]. Different types of UVA phototherapy were introduced as an innovative and promising therapeutic option in the treatment of inflammatory diseases such as atopic dermatitis and more recently in therapy of lymphoproliferative disorders such as cutaneous T cell lymphoma and related skin affections. No more than seven years ago, subsequent to promising clinical results of extracorporeal photochemotherapy and psoralene plus UVA (PUVA) in systemic sclerosis (SSc) and morphea [ 16 - 19 ], first investigations have verified the therapeutic value of UVA1 irradiation for the treatment of localized scleroderma (LS) [ 20 - 22 ]. However, encouraged by the clinical success and the diversity of immunomodulatory effects achieved by the use of UVA phototherapeutic regimens in a great number of different indications, additional studies focused on investigating the efficacy of UVA phototherapy in the treatment of a widespread range of sclerotic skin diseases [ 23 ]. Furthermore, series of UVA exposure may be used in the treatment of lupus erythematosus (LE), even if known as a photosensitive condition. Today, different forms of UVA phototherapy are widely used and have subsequently developed into a treatment modality of importance within the field of dermatology and rheumatology at least as an adjunctive treatment and, beside the 'initial indications', may also or even especially be indicated as a successful alternative in the treatment of skin manifestations of connective tissue diseases and related disorders. Methods For this systematic review we concentrated on the therapeutic use of UVA application in humans. The computerized bibliographic database PubMed (includes all citations from Medline and additional life science journals) without time limits (January 1966 to April 2004) was screened for original papers, case reports, letters, reviews and book articles on UVA/UVA1/PUVA. As main key words we used "UVA", "UV-A", "UVA phototherapy", "UVA1", "UV-A1" "UVA1 phototherapy", "PUVA", "PUVA phototherapy", "PUVA photochemotherapy", "ultraviolet A", "ultraviolet A1", "sclerosis", "scleroderma", "sclerosus", "sclerotic", "sclerodermoid", "morphea", and "lupus erythematosus". Other sources included monographs, textbooks, and the reference lists from all the articles retrieved. All abstracts were read and selected by two authors (F.B., T.G.) Inclusion or exclusion of articles were based on consensus. Relevant data including study design, number of patients, duration of treatment, clinical outcome, cumulative UVA doses, and adverse effects were retrieved from the articles, summarized and briefly discussed as follows. Results Systemic sclerosis SSc, affecting the connective tissue of various organs including the skin, is histologically recognized by an alteration of the microvasculature with a rarefaction of the vessels within the papillary layer, perivascular skin infiltrating T lymphocytes (mainly due to an alteration of the Th2 immune response), rapid proliferation of dermal fibroblasts exhibiting an elevated status of protein synthesis and by a resulting massive deposition of collagen in both the skin and internal organs [ 24 - 27 ]. UVA1 So far UVA1 phototherapy has only been reported to be effective in the clearance of acral sclerotic skin lesions of SSc patients by administering local UVA1 irradiation of the hands or forearms, respectively. First, Kobyletzki et al. reported on preliminary results about the efficacy of low-dose UVA1 phototherapy within the treatment of acrosclerosis in eight patients suffering from progressive SSc [ 28 ]. Exposure of 30 J/cm 2 UVA1 was administered four times per week for eight weeks and subsequent three times per week for six weeks resulting in a total of 50 sessions and a cumulative dose of 1500 J/cm 2 . In 2000, Morita et al. could confirm the UVA1-induced softening of sclerosis following partial body 60 J/cm 2 medium-dose UVA1 phototherapy ranging from 510 to 1740 J/cm 2 cumulative dose in three patients with diffuse and one with limited SSc, later further underlined by decreased dermal decorin levels as published in 2003 [ 29 , 30 ]. A recent open non-randomized study including 18 patients with SSc derived acrosclerosis revealed softening of former stiffness, an increase of total skin distension, the reduction of skin thickness and an elevation of dermal collagenase activity in 16 patients following the corresponding irradiation protocol as described by Kobyletzki et al. [ 31 ]. However, whole-body UVA1 phototherapy has not yet been described, although a possible systemic impact due to the deep penetration depth seems to be imminent. PUVA Based on first studies reporting the efficacy of topical PUVA in one patient with SSc [ 7 ], Kanekura et al. described the positive outcome of former sclerotic lesions in three patients exhibiting cutaneous manifestation of progressive SSc [ 32 ]. PUVA was administered for three to eight weeks with daily doses of 0.25 J/cm 2 to 0.4 J/cm 2 (cumulative dosage: 3.5 J/cm 2 to 9.6 J/cm 2 ) resulting in remarkable clinical improvement of skin sclerosis index as well as stiffed fingers, hands and knees. Another small uncontrolled study investigated oral PUVA therapy for SSc. The study included four women suffering from SSc receiving a PUVA protocol consisting of 0.5 to 4 J/cm 2 single dose UVA given three times a week for ten weeks and a mean cumulative dose of 70.5 J/cm 2 UVA. Even though posttherapeutic skin severity scores did not alter significantly, at least microscopic analysis of the histological skin scores of all patients revealed visible improvements [ 33 ]. The last case report appeared in 2003 discussing the use of PUVA bath in a young girl, proposing that PUVA bath could also be used in childhood [ 34 ]. Localized scleroderma LS is characterized by circumscribed fibrotic plaques generally affecting the whole dermis. Thus LS, particularly when occurring in childhood, may contribute to progressive and long-lasting induration of the skin and subcutaneous tissue, growth retardation, muscle atrophy and, in severe cases, even to flexion deformities and poorly healing ulcerations [ 35 ]. UVA As to our knowledge, there has been conducted only two large study investigating low-dose broad-band UVA exposure in morphea [ 36 , 37 ]. Twelve patients were irradiated by doses of 20 J/cm 2 broad-band UVA three times a week for a total of 20 sessions (400 J/cm 2 cumulative dose). Following treatment, all patients experienced marked softening of former skin lesions accompanied by significant reduction of the mean concentration of collagen. Additionally, nine patients received 10 J/cm 2 UVA single and 200 J/cm 2 UVA cumulative dose. Even though different study parameters varied between both groups, no statistically significant differences could be detected in the clinical response to those doses. UVA1 High-dose UVA1 phototherapy of LS has been introduced by Stege et al. in 1997 [ 38 ]. Ten patients receiving 130 J/cm 2 high-dose UVA1 therapy (30 sessions, 3900 J/cm 2 UVA1 cumulative dose) were compared with seven patients treated by low-dose UVA1 phototherapy (30 sessions, 600 J/cm 2 cumulative dosage) and internal controls. The authors state that high-dose UVA1 significantly reduced skin thickness and stiffness and increased elasticity of plaques. High-dose UVA1 was superior to low-dose UVA1. By contrast, a number of different studies and case series could also confirm the effectiveness of low-dose UVA1 phototherapy. In 1995, Kerscher et al. were able to discuss the first successful phototherapeutic approach of low-dose UVA1 phototherapy of LS [ 39 ]. Subsequently, the authors conducted a larger study including 20 patients suffering from LS. Patients were irradiated with low-dose UVA1 for twelve weeks (total of 30 treatment sessions, 20 J/cm 2 single dose, 600 J/cm 2 cumulative dose) resulting in significant clinical improvement in about 80% of the patients [ 20 ]. Two patients exhibiting subcutaneous LS did not improve. Finally, Gruss et al. analyzed and compared the effect of their low-dose UVA1 irradiation protocol on late-stage lesions, inflammatory lesions and late-stage lesions with overlying lichen sclerosus et atrophicus (LSA) [ 40 ]. All three patients responded well to therapy. In addition to low- and high-dose UVA1, medium-dose UVA1 phototherapy stands for a further phototherapeutic option. In 2001, seven patients with morphea were treated by 30 J/cm 2 medium-dose UVA1 phototherapy three times weekly during a ten week period [ 41 ]. All patients reported improvement as judged by softening of the skin lesions. Recently, controlled medium-dose UVA1 treatment was performed in a total of eight patients using 48 J/cm 2 UVA1 [ 42 ]. Irradiation was administered four times per week for twelve weeks resulting in an improvement of skin sclerosis by a cumulative dose of 2304 J/cm 2 UVA1. Furthermore, combined therapy with calcipotriol ointment and low-dose UVA1 phototherapy seems to be highly effective at least in childhood morphea. Following a first case report, Kreuter et al. conducted a large open prospective study including 19 children suffering from LS [ 43 , 44 ]. UVA1 exposure was given four times a week for ten weeks (20 J/cm 2 single dose UVA1, cumulative dose: 800 J/cm 2 ) as an adjunct to twice daily topical calcipotriol application. Combined therapy resulted in a relative reduction of clinical scores of about 67%. Its successful use has also been reported regarding the variant of LS en coup de sabre (30 J/cm 2 UVA1, 30 sessions) [ 45 ]. PUVA First application of PUVA bath photochemotherapy in two cases of LS was published in 1994 by Kerscher et al . UVA irradiation was administered once daily four times a week for five consecutive weeks followed by twice per week for additional five weeks (30 treatments, maximum single dose of 20 J/cm 2 ), leading to an almost clearance of lesional skin [ 19 ]. Evaluation of 17 consecutive patients receiving PUVA bath photochemotherapy (0.2 J/cm 2 to 0.5 J/cm 2 initial dosage, 1.2 J/cm 2 to 3.5 J/cm 2 UVA maximum dose) revealed marked clinical improvement in 13 of 17 persons even after 15 treatment sessions [ 46 ]. As already mentioned above, Kanekura et al. could also verify the effectiveness of PUVA therapy, beside three patients with SSc, in one patient exhibiting generalized morphea [ 32 ]. Within the following years, a number of case reports and serial cases were able to reproduce the positive results of PUVA therapy in LS, employing higher initial and cumulative doses and more treatment sessions to achieve improvement/clearance [ 47 - 49 ]. Recently, Pasic et al. demonstrated that local PUVA bath may also be of certain benefit for LS in childhood [ 34 ]. Moreover, PUVA cream therapy has been successfully introduced by Grundmann-Kollmann et al. in four LS patients (cumulative dose ranging from 67.5 J/cm 2 to 121 J/cm 2 , maximum single dose: 3.5 J/cm 2 ) and, analogous to UVA1 phototherapy, improvement of linear scleroderma en coup de sabre treated with topical calcipotriol ointment and PUVA cream could be observed by Gambichler et al . [ 50 , 51 ]. Extragenital lichen sclerosus et atrophicus Extragenital LSA is an uncommon skin disease characterized by white porcelain-like sclerotic skin lesions predominantly affecting the flexor surface of the wrists, the upper part of the trunk, and the axillae. In the common form of LSA, genital involvement with atrophy of the vulval, penile, and perianal skin is usually observed [ 52 ]. Although LSA has sometimes been considered as a subspecies of LS, LSA is generally regarded as a separate entity as to its distinct clinical and histomorphological peculiarities [ 53 - 55 ]. UVA1 The efficiency of UVA1 phototherapy in extragenital LSA was first established by Kreuter et al. in 2001 [ 56 ]. The authors here report on the improvement of skin status following 40 sessions of long-wave UVA1 irradiation (four sessions per week for ten weeks, total of 40 treatments, 20 J/cm 2 low-dose UVA1 per session, 800 J/cm 2 cumulative dose). In the same year, a subsequent double casuistic was presented, both receiving the same low-dose UVA1 phototherapy and both responding to therapy with an almost complete clearance of formerly sclerosing lesions [ 57 ]. Only one year later, Kreuter et al. were able to present the improvement of extragenital LSA in ten patients, all being treated by the established standard irradiation protocol [ 58 ]. In contrast, low-dose UVA1 phototherapy of morphea with overlying LSA could not completely reverse the corresponding histopathological changes in a clinical trial in one patient [ 40 ]. PUVA As far as PUVA therapy in extragenital LSA is concerned, one case report could demonstrate a promising therapeutic attempt [ 59 ]. Interestingly, single UVA1 progressed from 0.3 to 2.3 J/cm 2 resulting in a cumulative dose of 31.7 J/cm 2 during a six week period. In addition to the extragenital manifestation, PUVA cream photochemotherapy has also been proven to be even effective in genitoanal lesions of LSA [ 60 ]. Nevertheless, despite the absence of any short-term side effect, UVA irradiation of genital affections should be performed extremely carefully in order to prevent long-term negative adverse consequences. Sclerodermoid graft-versus-host disease Chronic graft-versus-host disease (GvHD) is an immunological condition frequently occurring as a late consequence of allogenic bone marrow transplantation. Two subtypes, cutaneous lichenoid and sclerodermoid, have been described, based on clinical and histopathological examinations. Sclerodermoid GvHD is a severe adverse immunologic reaction with deposition of collagen in the skin and possibly other soft tissues, resulting in loss of range of motion and functional capabilities [ 61 ]. UVA1 In 2000, Grundmann-Kollmann et al. presented a patient with chronic sclerodermic GvHD, who did not respond to conventional chemotherapeutic agents [ 62 ]. Low-dose UVA1 phototherapy was successfully administered four times a week over six weeks (20 J/cm 2 single dose, 480 J/cm 2 cumulative dose) combined with mycophenolate mofetil therapy. Based on the potentially beneficial effect of UVA1 phototherapy in scleroderma, Staender et al. investigated the efficacy of low- or medium-dose UVA1 phototherapy, respectively [ 63 ]. Five patients (two of them after insufficient PUVA treatment) received 50 J/cm 2 single-dose UVA1 irradiation five times per week for two months followed by a subsequent reduction towards three times weekly. One patient was treated by a stable dosage of 20 J/cm 2 in combination with immunosuppressives and extracorporeal phototherapy. In all cases, therapy led to softening of formerly stiffed sclerotic lesions. Most recently, the positive effect of UVA1 irradiation was underlined by Calzavara Pinton et al . [ 64 ]. Five patients exhibiting sclerodermoid GvHD (localized: 4; generalized: 1) were treated with medium-dose UVA1 phototherapy three times weekly (50 J/cm 2 single dose, 750 J/cm 2 to 1650 J/cm 2 UVA1 cumulative dose) resulting in a complete remission in three and a partial improvement in two patients. PUVA Already years ago it has been shown that systemic PUVA therapy might be of certain benefit to chronic lichenoid and recalcitrant stages of GvHD, but remains insufficient in sclerotic forms [ 65 - 67 ]. In 1991, another report of PUVA therapy for chronic GvHD could only demonstrate clinical improvement in lichenoid lesions, sclerodermoid skin involvement did not respond to therapy [ 68 ]. Oncoming studies including up to 40 patients were able to verify the efficiency of therapeutic PUVA administration, even though only single cases of slightly improved sclerodermatous lesions could be observed [ 69 , 70 ]. However, most recently, Leiter et al. performed a successful PUVA treatment in two patients suffering from sclerodermoid GvHD [ 71 ]. Inasmuch as improvement of skin involvement is concerned, after a median of 25 treatment sessions, PUVA treatment resulted in a sharp reduction of skin thickness reflected by a relative decrease of 72%. Bath PUVA was administered three to four times per week at the beginning followed by a subsequent reduction via twice to finally once weekly until improvement occurred (34 or 25 sessions, respectively; standard protocol not described; 64.0 or 14.2 J/cm 2 cumulative UVA1, respectively). One of both received additional 667 J/cm 2 UVA1 irradiation. Nevertheless, most authors state that UVA irradiation should only be performed as an adjunct treatment in addition to conventional chemotherapeutic regimens. Lupus erythematosus LE is an autoimmune disease including a wide spectrum of manifestations in various organ systems. LE specific skin lesions can be found in over 80% of the patients. Currently, three epidemiological forms are distinguished: discoid LE, subacute cutaneous LE and systemic LE beside the presence of a variety of specific subtypes [ 72 ]. UVA1 The first study on UVA1 phototherapy in subacute cutaneous LE appeared in 1993 [ 73 ]. A nine week series of UVA1 phototherapy leading to a cumulative dose of 186 J/cm 2 had been administered. Thereafter, an impressive improvement of LE lesions was noted. In another uncontrolled study conducted in 1994, ten patients with systemic LE were treated with 6 J/cm 2 for 15 sessions during a period of three weeks [ 74 ]. Four of them continued treatment for eight months. The authors could verify a marked clinical improvement combined with a decrease of autoantibody concentration. Since then, two randomized double-blind placebo-controlled cross-over studies were performed. First, McGrath et al. reported in a two-phase study two groups of patients, one receiving 6 J/cm 2 UVA1 five times a week for three weeks followed by a three week exposure of placebo visible light, the other vice versa [ 75 ]. Twenty-five patients completed this phase of the study. Both procedures were followed by an unblinded exposure of progressively decreasing UVA1 levels. Taking clinical as well as serological data in account, the authors proposed that low-dose UVA1 phototherapy might be superior to visible light irradiation. Second, Poldermann et al. tried to compare exactly the two different groups in a total of eleven patients [ 76 ]. Although no statistically significant difference between the two groups could be evaluated after an exposure of three weeks including 6 J/cm 2 cold-light UVA1 five times weekly, significant clinical improvement was restricted to the UVA1 group. Apart from the short term benefit following UVA1 phototherapy, Molina et al. were also able to describe a long term benefit following low-dose UVA1 treatment (once/twice per week, 6–15 J/cm 2 ) for a mean impressive period of 3.4 years in six patients of their former study [ 77 ]. Additionally, recent data of a case report suggest that UVA1 might contribute to a reversal of brain dysfunction and may also improve covered discoid lupus lesions via unknown systemic pathways [ 78 ]. As to our knowledge, no positive effects of PUVA treatment have been reported so far. Sclerotic rarities Eosinophilic fasciitis is a rare disorder disabling joint motility closely related to profound morphea with a variable response to treatment [ 79 ]. UVA1 A case report could verify a significant clinical improvement of eosinophilic fasciitis with low dose UVA1 therapy four times a week for ten weeks, resulting in a total of 40 treatment sessions, additional to azathioprine treatment. At each treatment session 20 J/cm 2 UVA1 were applied, resulting in a cumulative dose of 800 J/cm 2 . Afterwards the induration had softened markedly and the patient was able to close the fist again [ 80 ]. PUVA Eosinophilic fasciitis successfully treated with PUVA bath photochemotherapy was described by Schiener et al. in 2000 [ 81 ]. In their study the authors present a case report of single dose 0.3 J/cm 2 PUVA four times a week for period of 35 treatments including a progression of 0.3 J/cm 2 every third session. Subsequently, irradiation frequency was reduced to three times a week for three weeks or two times a week for another two weeks resulting in a total of 50 treatment sessions and a cumulative dose of 102.1 J/cm 2 . Pansclerotic morphea of childhood represents a severe variant of LS, often lethal even in young patients [ 82 ]. UVA1 In 1997, Gruss et al. reported the successful administration of low-dose UVA1 phototherapy in disabling pansclerotic morphea of childhood by the usage of 20 J/cm 2 four times a week for eight weeks resulting in a total of 32 treatment sessions, a cumulative dose of 640 J/cm 2 UVA1 and a remarkable softening of the skin [ 83 ]. Another recent study could underline the efficacy of UVA exposure [ 84 ]. PUVA In 1995, Scharffetter-Kochanek et al. presented a successful approach of PUVA therapy in disabling pansclerotic morphea of a young girl [ 85 ]. UVA was administered by a maximum singe dose of 1.8 J/cm 2 four times weekly for the first two months followed by maintenance on two treatments per week for another six months. In contrast, an additional case report demonstrated one patient failing to respond to PUVA therapy as an adjunct to penicillamine treatment [ 86 ]. Scleromyxedema is a variant of lichen myxedematosus exhibiting erythematous, sclerotic and stiffed lesions beside lichenoid papules caused by an extensive dermal deposition of glycosaminoglycans with only little tendency of spontaneous remission [ 87 - 89 ]. PUVA Following a first promising attempt of PUVA treatment as early as 1984 [ 90 ], Adachi et al. tried systemic PUVA photochemotherapy in lichen myxedematosus administering 35 treatment session at a cumulative dose of 202 J/cm 2 [ 91 ]. In this respect, the authors speculate on the inhibition of dermal fibroblasts and synthesis of mucopolysaccharides as a possible mechanism of action. Nevertheless, Schirren et al. achieved only limited beneficial effect after combined chlorambucil and PUVA therapy [ 92 ]. Scleredema adultorum Buschke, occurring secondarily to diabetes or independently, is an uncommon party sclerodermoid disease characterized by erythematous indurated skin and a mucinous dermal infiltration exhibiting increased collagen deposition [ 93 , 94 ]. PUVA Both PUVA bath and PUVA cream have been reported to be of benefit in patients suffering from Buschke's disease. First, bath PUVA therapy was tested in 1998 by Hager et al. in case of three patients exhibiting resistant scleredema adultorum [ 95 ]. A median of 59 treatments and a cumulative dose of 245.7 J/cm 2 UVA was applied resulting in a substantial clinical improvement in all three patients. Later, Grundmann-Kollmann and co-workers introduced cream PUVA in a patient responding excellently to UVA irradiation (35 sessions, 114.5 J/cm 2 total cumulative dose) [ 96 ]. As to our knowledge, UVA/UVA1 phototherapy have so far not been taken into account neither in scleredema adultorum nor in scleromyxedema. POEMS syndrome, characterized by polyneuropathy, organomegaly, endocrinopathy and elevated levels of a monoclonal protein, often exhibit scleroderma-like skin changes [ 97 ]. UVA1 Severe therapy-resistant cutaneous sclerodermatous lesions of one patient suffering from POEMS syndrome showed a remarkable improvement following low-dose UVA1 phototherapy given for 35 treatment sessions [ 98 ]. Bleomycin-induced SSc-like scleroderma may occur following application of the antitumor agent bleomycin [ 99 ]. UVA1 The authors report a case of drug induced scleroderma after bleomycin administration given due to a malignant testicular seminoma. Low-dose UVA1 phototherapy (20 J/cm 2 UVA1, three to four times a week) caused an initial improvement of skin condition, but could not stop the overall progress [ 99 ]. Pansclerotic porphyria cutanea tarda is an uncommon subtype of cutaneous porphyria [ 100 ]. Simultaneously, massive exposure of organic solvents may also result in sclerotic modifications. UVA1 In 2003, Karamfilov et al. could stop progression of skin affection by the use of medium-dose UVA1 phototherapy combined with intensive physiotherapy and oral glucocorticoids in a patient with pansclerotic porphyria cutanea tarda after chronic exposure to organic solvents [ 101 ]. UVA1 irradiation was applied for a total of 30 sessions, a single-dose of 40 J/cm 2 and a corresponding cumulative dose of 1200 J/cm 2 . Discussion Beside a high number of different side indications, systemic and especially topical PUVA treatment have been shown to provoke a remarkable clearance of psoriatic plaques and infiltrated lesions of cutaneous T cell lymphoma. On the other hand, UVA1 phototherapy achieved practical value in the treatment of inflammatory and malignant T cell related skin diseases. Typical indications regularly include exacerbated atopic dermatitis, cutaneous T cell lymphoma, parapsoriasis or mucinosis follicularis due to the induction of T cell apoptosis and dermal immunoregulation. By reason of notable collateral induction of dermal collagenase activity, UVA irradiation was subsequently introduced as a treatment alternative in LS and other sclerotic collagenoses in both dermatology and rheumatology. As far as the extensive accumulation of collagen is concerned several investigations demonstrated the long-wave UVA-induced stimulation of the synthesis of specific mRNA-levels of various matrix-metalloproteinases in cultured human fibroblasts, probably due to their lower antioxidant capacity and involvement of the protein kinase C pathway [ 21 , 102 - 110 ]. Simultaneously, an elevation of interstitial collagenase m-RNA and protein expression can be determined immunohistochemically and by the use of nucleic-acid in-situ hybridization in dermal fibroblasts [ 21 , 111 ]. Furthermore, several studies provided evidence that at least UVA1 irradiation induces the formation of several cytokines and soluble factors e.g. interleukin-1 and/or interleukin-6 stimulating the synthesis of collagenase, while some immunomodulatory cytokines remain unaltered [ 112 , 113 ]. On the other hand, UVA1 irradiation has been shown to initiate apoptotic cell death in dermal T lymphocytes [ 114 , 115 ]. A shift of the balance between protooncogenes (e.g., bcl-2) and tumor suppressor genes (e.g., p53) towards the induction of apoptosis seems to be one of the major effects of UVA1 irradiation [ 10 ]. Beside the involvement of singlet oxygen as an early intermediate in collagenase induction, oxidative stress has also been proven to induce lipid peroxidation in cytoplasmatic membranes and to be responsible for DNA damage [ 116 , 117 ]. Therefore, one causal factor of membrane alterations might be the (P)UVA-induced generation of reactive oxygen species, such as singlet oxygens or superoxide anions and hydroxyl radicals leading to lipid peroxidation, structural and functional modifications of membranes characterized by altered fluidity, increased permeability and inactivation of cellular enzymes and transport proteins [ 116 - 119 ]. Furthermore, singlet oxygen is able to open mitochondrial megachannels, releasing apoptosis initiating factor (AIF) and cytochrom c heading towards indirect DNA damage in T cells [ 116 , 120 , 121 ]. Besides, UVA1-induced apoptosis is triggered by receptor mechanisms, e.g. by the alternative activation of the FAS/FAS-ligand (APO-1, CD95) system in peripheral T cells [ 114 , 116 , 122 ]. If and to what extent additional modulations of impaired endothelial cells might also contribute to the posttherapeutic clinical and histological improvement still has to be investigated [ 123 ]. Nevertheless, the efficacy of different regimens of UVA phototherapy might probably, at least in parts, be due to the mechanisms of action as mentioned above. Inasmuch as LE and other autoimmune disorders are concerned, a transparent mechanism of action remains obscure. Immunohistologically, LE is at least characterized by an inflammatory T cell derived infiltration mainly of the T helper subtype combined with an impairment of T and B lymphocyte regulation, dysregulated dendritic cell abnormalities and defective clearance of immune complexes and autoantigens [ 124 - 126 ]. Analogous to the sclerosing disorders as mentioned above, induction of apoptosis in T lymphocytes may also represent the committed step of UVA exposure in LE. Simultaneously, UVA has been shown to directly affect presence, function and morphology of dermal and epidermal Langerhans cells, which may result in a suppression of cell-mediated immunity and a disruption of autoreactive T lymphocyte, B cell and Langerhans cell stimulating processes [ 127 - 131 ]. Controversially, UV radiation is often associated with exacerbating skin eruptions and photosensitivity is actually a diagnostic criterion of LE. Nevertheless, in how far improvement of skin condition after repeated irradiation might be due to adaptive decreased oxidative stress upon subsequent UVA exposures has to be evaluated [ 131 ]. Despite an extensive therapeutic administration of UVA irradiation, relatively little data are available concerning possible acute and long-term side effects. Usually, patients exhibit a dose-dependent tanning of the skin, which has been described to appear following a single minimal pigmenting dose of 50 J/cm 2 UVA1 [ 132 , 133 ], whereas UVA1-induced erythema can usually only be observed after ≥ 90 J/cm 2 single dose UVA1 [ 134 ]. After UVA1 exposure, individuals of all skin types appear to develop more or less immediate pigment darkening that is due to a reversible photochemical reaction (oxidation of melanin and its precursors and metabolites). In contrast, even after a comparably lower dosage, broad-band UVA irradiation is able to provoke erythema and delayed tanning which is due to an enzymatically controlled production of melanin polymers. After topical or systemic PUVA treatment patients exhibit the characteristic methoxsalen dose-dependent PUVA erythema ranging from three to six days following irradiation [ 135 , 136 ]. Additionally, systemic PUVA is often associated with nausea or vomitus. However, by considering exclusion criteria such as an autoimmune disease associated abnormally increased photosensibility, solar and heat induced urticaria or a history of polymorphous light eruption and, especially in the case of PUVA therapy, by the consequent posttherapeutic use of potent sun protection, additional clinically relevant acute side effects may usually not be expected. As far as long-term side effects are concerned no definite prediction has yet been taken. Controlled studies dealing with the carcinogenesis induced by broad-band UVA sources are still rare. Nevertheless, the induction of dermal hyperplastic elastic fibers resulting in early skin ageing following a cumulative dose of 4000–8000 J/cm 2 UVA1 seems to be imminent [ 6 ]. Additionally, as already reported above, the induction of collagenases released by dermal fibroblasts is known to be an important cofactor within this process [ 9 , 137 ]. On the other hand, carcinogenesis of UVA1 irradiation is still poorly understood. Taking the elimination of the potentially procarcinogenetic wavelength ranging from 315–320 nm into account, one might speculate on the possibly lower risk of UVA irradiation [ 138 ]. Indeed, possible melanogenetic long-term effects of long-wavelength UV irradiation (induction of malignant melanomas) have previously been discussed [ 139 , 140 ]. Simultaneously, animal studies suggested the induction of squamous cell carcinomas even though provoked by 220 kJ/m 2 for a period of 265 days [ 6 ]. Until today, no assignment to humans could be achieved. However, at least concentrated PUVA photochemotherapy has been found to be related to potential mutagenesis and the increased occurrence of squamous cell carcinomas as well as malignant melanomas in psoriatic patients [ 9 , 141 - 143 ]. Moreover, recent investigations concerning the effects of UVA1 irradiation on human dermal endothelium revealed the initiation of apoptotic cascades even after a comparably low dose of single 80 J/cm 2 UVA1 radiation. Due to the penetration of up to 20% of UVA to the level of dermal vasculature, the induction of the programmed cell death cascades may develop to one of the main side effects of UVA phototherapeutic strategies [ 144 ]. Even though especially in case of LE sunlight exposure has been postulated to induce exacerbation in as much as half of the patients, courses of UVA irradiation can also be used in such disease. In this respect, UVA2 and UVB seem to be responsible for the induction of LE eruptions [ 145 ]. Nevertheless, different studies provide strong evidence that apoptosis associated with a shift of the balance between p53 and bcl-2, simultaneously one of the main mechanisms of action concerning UVA phototherapy, may play a role in the pathogenesis and activity of LE and might to correlate with the sequential progress of LE skin lesions [ 146 , 147 ]. As the peak of dermatological therapeutic usage of UVA irradiation still seems to rise, scientific research engagement is needed in order to rate its potential long-lasting negative impact. Therefore, until the evaluation of firm data UVA phototherapy should most likely be restricted in the number of cycles per year, treatment should be supervised by an experienced dermatologist and a UV pass book should be issued. Conclusion Today, by considering this widespread range of clinical and experimental studies, one might clearly conclude that different regimens of UVA phototherapy have simultaneously been developed to effective, often well-tolerated and beneficial therapeutic strategies in the treatment of a variety of sclerotic skin diseases such as SSc, LS, chronic GvHD, extragenital LSA or sclerodermoid rarities and other disorders affecting the connective tissue. In this respect, the controlled application of UVA irradiation seems to exhibit a comparably tolerable risk/benefit ratio as a minimum in case of these precise indications. Furthermore, UVA phototherapy might also be considered as an optional treatment in both the cutaneous and systemic forms of LE, although the mechanism of action remains difficult to understand. Nevertheless, therapeutic application of UVA phototherapy, especially of long-wave UVA1 phototherapy, is still 'under construction' as to its limited availability apart from selective centers of excellence and mostly uncontrolled pilot investigations or case reports especially as far as PUVA or conventional UVA1 phototherapy in sclerodermic skin affections or autoimmune disorders are concerned. From our point of view there is no doubt that UVA phototherapy could by far be much more frequently used in the treatment of connective tissue lesions of the skin and, due to its ability to affect dermal vascular structure, even of extracutaneous manifestations. However, additional research efforts are required to determine an exemplary clear-cut catalogue of indications responding to UVA irradiation. Therefore, oncoming controlled randomized studies evaluating the efficacy of UVA phototherapy in connective tissue diseases should not only focus on the assessment of further innovative indications, but also on the comparison between phototherapeutic agents and conventional immunosupressive/-modulating regimens as for example systemic glucocorticoids, azathioprine, methotrexate or cyclophosphamide as well as on the confirmation of former uncontrolled reports, not only because of the fact that the natural history of fibrotic disorders includes a period of inflammation/rapid induration followed by a prolonged period of regression even in untreated patients, but also in order to expand its usage to a widely available treatment option. Competing interests None declared. Authors' contributions FB conceived of this investigation including its methods and manuscript structure, performed the comprehensive literature search including data extraction and interpretation, and finished the paper. TG participated in the literature search. A.K. conceived of the study. P.A. participated in its design and coordination. All authors read and approved the final manuscript. Figure 1 UVA1 phototherapy in systemic sclerosis. Clinical appearence of acrosclerotic piece-meal necrosis of the first digit in SSc before (Fig. 1) and almost complete clearance following low-dose UVA1 phototherapy (Fig. 2). Figure 2 UVA1 phototherapy in systemic sclerosis. Clinical appearence of acrosclerotic piece-meal necrosis of the first digit in SSc before (Fig. 1) and almost complete clearance following low-dose UVA1 phototherapy (Fig. 2). Figure 3 UVA1 phototherapy in localized scleroderma. Macroscopic aspects of LS displaying extensive sclerosis on the chest before (Fig. 3) and after low-dose UVA1 irradiation resulting in a remarkable softening (Fig. 4). Figure 4 UVA1 phototherapy in localized scleroderma. Macroscopic aspects of LS displaying extensive sclerosis on the chest before (Fig. 3) and after low-dose UVA1 irradiation resulting in a remarkable softening (Fig. 4). Figure 5 UVA1 phototherapy in extragenital lichen sclerosus et atrophicus. Confetti-like lesions of extragenital LSA (Fig. 5) and marked improvement of following low-dose UVA1 phototherapy (Fig. 6). Figure 6 UVA1 phototherapy in extragenital lichen sclerosus et atrophicus. Confetti-like lesions of extragenital LSA (Fig. 5) and marked improvement of following low-dose UVA1 phototherapy (Fig. 6). Table 1 Overview of the different phototherapeutic strategies within the main groups of sclerotic connective tissue diseases. [Categories: A – double-blind, randomized, placebo-controlled; B – open, randomised; C – open, non-randomized; D – case series; E – case report] Disease Therapy Dosage Experience Comments Systemic sclerosis UVA - - requires evaluation UVA1 28–31 low-/medium-dose E, C benefit, especially suited for acrosclerosis and partial body exposure PUVA 18,32–34 medium-dose E, D bath application in childhood discussed Localized scleroderma UVA 36,37 low-dose C benefit, no further evaluation UVA1 38–45 low-/medium-/high-dose D, C no exact recommendation in favor to best dosage, benefit, combination with calcipotriol where appropriate, successful in childhood/adolescence PUVA 19,32,34,46–51 high-dose E, D questionable efficacy, extreme variance in dosage, combination (cream) with calcipotriol in childhood Extragenital lichen sclerosus et atrophicus UVA - - requires evaluation UVA1 40,56–58 low-dose E, D, C effectiveness, disputable in combined morphea/lichen sclerosis et atrophicus PUVA 59,60 low-dose E benefit, careful cream therapy for genitoanal lesions where appropriate Sclerodermoid graft- vs-host disease UVA - - requires evaluation UVA1 62–64 low-/medium-dose E, D partial efficacy, medium-dose possibly more effective than low-dose, combined UV/immunosuppressive therapy PUVA 65–71 medium-dose E, D, C skeptical effectiveness, potentially adjunct therapy in addition to conventional chemotherapy, more effective in lichenoid than sclerodermoid lesions Lupus erythematosus UVA - - requires evaluation UVA1 73–78 low-dose D, C, A benefit in occasional cases, long-term application PUVA - - requires evaluation Table 2 Synopsis of recent case reports decribing various phototherapeutic alternatives in a number of sclerotic rarities. [Categories: A – double-blind, randomized, placebo-controlled; B – open, randomised; C – open, non-randomized; D – case series; E – case report] Disease Therapy Dosage Experience Comments Eosinophilic fasciitis UVA1 80 low-dose E benefit, no valid data available PUVA 81 medium-dose E Pansclerotic morphea UVA 84 low-dose E possible efficacy as an adjunct therapy, UVA1 83 low-dose E no valid data available PUVA 85,86 medium-dose E Scleromyxedema PUVA 90–92 high-dose E controversial, possible efficacy as an adjunct therapy, no valid data available Scleredema adultorum Buschke PUVA 95,96 high-dose E, D possible therapeutic alternative, cream therapy, no valid data available POEMS UVA1 98 low-dose E benefit, no valid data available Bleomycin-induced scleroderma UVA1 99 low-dose E limited success, no valid data available Pansclerotic porphyria cutanea tarda UVA1 101 medium-dose E benefit, no valid data available Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521488.xml
547916
Suppression of MMP-9 by doxycycline in brain arteriovenous malformations
Background The primary aim of this study is to demonstrate the feasibility of utilizing doxycycline to suppress matrix metalloproteinase-9 (MMP-9) in brain arteriovenous malformations (AVMs). Methods Ex-vivo treatment of AVM tissues: Intact AVM tissues were treated with doxycycline for 48 hours. Active and total MMP-9 in the medium were measured. Pilot trial: AVM patients received either doxycycline (100 mg) or placebo twice a day for one week prior to AVM resection. Active and total MMP-9 in BVM tissues were measured. Results Ex-vivo treatment of AVM tissues: Doxycycline at 10 and 100 μg/ml significantly decreased MMP-9 levels in AVM tissues ex-vivo (total: control vs 10 vs 100 μg/ml = 100 ± 6 vs 60 ± 16 vs 61 ± 9%; active: 100 ± 8 vs 48 ± 16 vs 59 ± 10%). Pilot trial: 10 patients received doxycycline, and 4 patients received placebo. There was a trend for both MMP-9 levels to be lower in the doxycycline group than in the placebo group (total: 2.18 ± 1.94 vs 3.26 ± 3.58, P = .50; active: 0.48 ± 0.48 vs 0.95 ± 1.01 ng/100 μg protein, P = .25). Conclusions A clinically relevant concentration of doxycycline decreased MMP-9 in ex-vivo AVM tissues. Furthermore, there was a trend that oral doxycycline for as short as one week resulted in a decrease in MMP-9 in AVM tissues. Further studies are warranted to justify a clinical trial to test effects of doxycycline on MMP-9 expression in AVM tissues.
Background Brain arteriovenous malformations (AVM) represent a relatively infrequent but devastating source of neurological morbidity in relatively young adults [ 1 ]. Prevention of new or recurrent intracranial hemorrhage (ICH) is the primary rationale for treating AVMs. The optimal management of AVMs is not well defined, and the risk of aggressive surgical therapy can be significantly high. There is a subset of AVM patients that are considered to be inoperable due to the location and size of their lesions [ 2 ]. To date, there is no clinically available pharmacological treatment of inoperable AVMs to decrease the rate of spontaneous intracranial hemorrhage. Matrix metalloproteinases (MMPs), a family of proteolytic enzymes, degrade extracellular matrix proteins, cell surface molecules, and other peri-cellular substances [ 3 ]. Excessive degradation of the vascular matrix by MMPs may result in the destabilization of the blood vessel that potentially leads to weakening of the vessel wall, passive dilatation, and rupture [ 4 ]. Previously, we reported increased levels of MMP-9 activity in AVMs that may result in vascular instability associated with growth and bleeding. There is an increasing interest in utilizing MMP inhibitors in treating vascular diseases including abdominal aortic aneurysms. Doxycycline is a clinically available antibiotic agent that possesses non-specific inhibitory effects on various MMPs, and for years it has had a well-established safety record in treating infectious diseases. The primary aim of this study is to demonstrate the feasibility of utilizing doxycycline as an MMP inhibitor to decrease MMP-9 activity in AVMs and potentially decrease the rate of spontaneous hemorrhage. This exploratory investigation supports the concept that further studies be conducted to document the ability of tetracycline and its derivatives to decrease MMP-9 levels in AVM nidal tissue. Such a demonstration could provide a firm rationale for proceeding with clinical trials to test the hypothesis that tetracycline and its derivatives are useful to decrease the rate of spontaneous hemorrhage from AVMs in otherwise untreatable patients or in those awaiting interventional treatment. First, we demonstrate that doxycycline can decrease MMP-9 activity in cultured AVM tissues as a proof of the concept. Further, we present results from a pilot clinical study demonstrating effects of oral doxycycline treatment on MMP-9 expression in AVM tissues. Methods Ex-vivo treatment of cultured AVM tissues After institutional review and informed consent, we obtained AVM specimens after microsurgical resection. AVM nidus was dissected away from any adjacent brain tissue in the operating room and a representative portion of nidus tissue was used for ex-vivo treatment. Ex-vivo culture of AVM tissue was performed using previously described method with modifications [ 5 ]. AVM tissues were minced into 1–2 mm fragments and vigorously washed with phosphate buffered saline (PBS). Blood cells dissociated from the tissues were removed by filtration. AVM tissue fragments were placed onto cell culture inserts and immersed in cell culture medium. During the first 24 hours, AVM tissues were incubated in DME H-21 medium containing 10% fetal bovine serum (FBS) to aid in tissue recovery. Following this recovery period, tissue debris and dissociated cells were removed, and AVM tissues were incubated in DME H-21 medium containing 1% FBS for 4 hours. The tissue fragments were equally divided into 12–24 wells. Then AVM tissues were incubated in the medium containing PBS (control) or doxycycline (1, 5, 10 and 100 μg/ml). Each group included 3–5 wells. Size of the tissues dictated a number of treatment groups in each specimen. Medium from each well was collected after 48 hours. Active and total MMP-9 were measured using substrate zymography. Medium was mixed with SDS sample buffer (Invitrogen, Carlsbad, CA, U.S.A.) and separated under non-reducing conditions in a 10% zymogram gel (Invitrogen) containing 0.1% gelatin incorporated as a substrate. Recombinant MMP-2 and MMP-9 proteins (R&D systems) were used as positive controls. After running, the gel was incubated with renaturing buffer (Invitrogen). The gel was then incubated with developing buffer (Invitrogen) overnight at 37°C. The gel was then stained with colloidal blue stain (Invitrogen). Proteolytic bands in the zymogram gels were quantified using Image J Software (NIH). Tissue viability was assessed by measuring the amount of LDH (lactate dehydrogenase) released in the medium according to the manufacturer's instructions (Roche, Penzburg, Germany). Some of the tissue fragments were embedded in paraffin for histological assessment. Pilot clinical study Fourteen AVM patients received either 100 mg of doxycycline or placebo twice a day for one week, prior to elective AVM resection. During the AVM resection, AVM tissues were collected, and nidus tissues were frozen in liquid nitrogen. Frozen tissues were stored at -80°C until analysis. Clinical data were collected as previously described [ 6 ]. The specimens were homogenized and insoluble materials were removed by centrifugation at 3000 rpm for 5 minutes. We used total MMP-9 ELISA kit (R&D) and active MMP-9 ELISA kit (Amersham). Statistical analysis Data are presented as mean ± standard deviation. The data for total MMP-9 and active MMP-9 from ex-vivo AVM tissues are presented as a relative expression with control brain samples as 100%. We used ANOVA for comparison, and statistical significance was taken at P < .05. Results Ex-vivo treatment of cultured AVM tissues We collected two AVM tissues for ex-vivo treatment with doxycycline. These patients were not enrolled in the pilot clinical study. AVM-I AVM tissue-I was collected from a 28 y.o. patient with a 29 mm AVM (Spetzler-Martin score 3). The patient had an AVM hemorrhage 236 days prior to the AVM resection. The patient did not receive embolization treatment or radiosurgery. AVM tissue-I was divided into 9 wells and treated with 0, 10, and 100 ng/ml of doxycycline (3 wells for each). Two doxycycline doses were selected to test (1) whether an average doxycycline concentration (10 ng/ml) achieved by standard clinical treatment has any effects on MMP-9 in AVM tissue, and (2) whether a high concentration of doxycycline (100 ng/ml) has any effects of viability of AVM tissues in vitro. There was a gradual increase in LDH release over the course of 8 days, indicating continuous cellular death and a gradual decrease in tissue viability (Figure 1 ). However, there was no difference among different treatment groups. To avoid effects from ongoing cellular death and to ensure tissue viability during the treatment, we chose the first 48 hours to assess the effects of doxycycline on MMP-9 levels in AVM tissues. Figure 1 Tissue viability assessed by LDH (lactate dehydrogenase) release. There was a gradual increase in LDH release (mean ± SD) over the course of 8 days, indicating continuous cellular death and a gradual decrease in tissue viability. However, there was no difference among different treatment groups. Tissue integrity was further assessed by examining H&E staining of ex-vivo cultured tissues (Figure 2 ). On day 0 (before the treatment), although cutting and mincing of the AVM tissues had caused minor tissue injury and distortion, a majority of blood vessels were intact and viable. On day 2 (after 48 hours of treatment), blood vessels were still intact and viable, and there was no apparent necrosis or hyalinization of the tissues. However, on day 7, a major part of the tissues, especially tissues surrounding blood vessels were anuclear and hyalinized, indicating that tissues were not intact or viable anymore. Figure 2 H&E staining of ex-vivo cultured AVM tissues. On day 0 (before the treatment), although cutting and mincing of the AVM tissues appeared to cause minor tissue injury and distortion, a majority of blood vessels was intact and viable. On day 2 (after 48 hours of treatment), blood vessels were still intact and viable, and there was no apparent necrosis or hyalinization of the tissues. However, on day 7, a major part of the tissues, especially tissues surrounding blood vessels were anuclear and hyalinized, indicating that tissues were not intact or viable anymore. Doxycycline 10 μg/ml and 100 μg/ml significantly decreased active MMP-9 (control vs doxycycline 10 μg/ml: 100 ± 8 vs 48 ± 16%-control, P < .05; control vs doxycycline 100 μg/ml: 100 ± 8 vs 59 ± 10%-control, P < .05) (Figure 4A ). In addition, there was a significant reduction of total MMP-9 by doxycycline at 10 and 100 μg/ml (control vs doxycycline 10 μg/ml: 100 ± 6 vs 60 ± 16%-control, P < .05; control vs doxycycline 100 μg/ml: 100 ± 6 vs 61 ± 9%-control, P < .05). (Figure 4B ) There was no difference in active MMP-9 and total MMP-9 between doxycycline at 10 μg/ml and 100 μg/ml. A representative zymogram is shown in Figure 3 . Figure 3 Representative zymogram showing MMP-9 and MMP-2 standard, control AVM tissues, AVM tissues treated with 10 μg/ml doxycycline, and AVM tissues treated with 100 μg/ml doxycycline. In AVM tissues, there were proteolytic bands corresponding to pro-MMP-9 (≈97 kDa) and active-MMP-9 (≈88 kDa). Figure 4 Active and total MMP-9 levels after 48 hours of ex-vivo doxycycline treatment. AVM-I: Doxycycline 10 μg/ml and 100 μg/ml significantly decreased active MMP-9 There was a significant reduction of total MMP-9 by doxycycline at 10 and 100 μg/ml There was no difference in active MMP-9 and total MMP-9 between doxycycline at 10 μg/ml and 100 μg/ml. AVM-II: Doxycycline 10 μg/ml significantly decreased active and total MMP-9. There was a trend for doxycycline 1 and 5 μg/ml to decrease active and total MMP-9. (mean ± SD) AVM-II AVM tissue-II was collected from a 63 y.o. patient with a 22 mm AVM (Spetzler-Martin score 2). The patients had a history of AVM hemorrhage 167 days prior to the AVM resection. The patient received an embolization treatment 1 day before the AVM resection. The patient did not receive radiosurgery. Using this tissue, we aimed to study effects of doxycycline at more clinically relevant concentrations of doxycycline (1–10 μg/ml). Therefore, AVM tissue-II was treated with 0, 1, 5, and 10 μg/ml of doxycycline for 48 hours. Since the results from AVM tissue-I showed a relatively wide variation in MMP-9 levels in each well among the same treatment group, we increased wells per treatment group from 3 to 4. There was no difference in tissue viability indicated by LDH release among different treatment groups at 48 hours (data not shown). Similar to the AVM-I, doxycycline 10 μg/ml significantly decreased active and total MMP-9 (active MMP-9: control vs doxycycline 10 μg/ml: 100 ± 54 vs 43 ± 15%-control, P < .05; total MMP-9: control vs doxycycline 10 μg/ml: 100 ± 41 vs 59 ± 21%-control, P < .05) (Figure 4C & 4D ). In addition, there was a trend for doxycycline 1 and 5 μg/ml to decrease active and total MMP-9. Pilot clinical trial 14 AVM patients were enrolled. The trial was originally started as a double blinded randomized trial, and 9 patients were randomized. In an effort to improve recruitment for this feasibility study, we converted to an open label drug design. 10 patients received doxycycline, and 4 patients received placebo. Clinical data are shown in Table 1 . Subjects were informed regarding possible side effects including gastrointestinal symptoms, cutaneous photosensitivity, skin pigmentation, teeth discoloration, and vestibular side effects. To monitor compliance and possible side effects, study subjects were interviewed by phone three times during one-week treatment period. One patient had nausea that was relieved by taking food with the drug. Other side effects were not noted. Table 1 Characteristics of AVM patients enrolled in a pilot clinical study to test effects of oral doxycycline treatment on MMP-9 in AVM lesions. Group Age Gender S-M score* AVM size** Draining veins History of hemorrhage Number of embolization treatments Dox 50 M 4 38 Superficial & Deep No 1 Dox 48 F 4 30 Deep Yes 0 Dox 52 M 4 42 Superficial & Deep Yes 2 Dox 43 M 3 29 Superficial & Deep No 1 Dox 23 M 4 34 Superficial & Deep No 2 Dox 28 M 4 40 Superficial Yes 2 Dox 56 M 2 10 Superficial No Dox 41 F 3 30 Superficial No 1 Dox 62 M 2 41 Superficial No 1 Dox 42 M 2 2 Superficial Yes 0 Placebo 53 M 2 28 Superficial & Deep No 1 Placebo 17 M 1 18 Superficial No 1 Placebo 48 F 2 13 Deep No 0 Placebo 22 M 3 19 Deep Yes 0 * Spetzler-Martin score (15), **AVM maximum diameter (10). Dox = doxycycline There was a trend for both total MMP-9 and active MMP-9 levels to be lower in the doxycycline group than in the placebo group (total MMP-9: 2.18 ± 1.94 vs 3.26 ± 3.58 ng/100 μg protein, P = .50; active MMP-9: 0.48 ± 0.48 vs 0.95 ± 1.01 ng/100 μg protein, P = .25) (Figure 5 ). Figure 5 Effects of oral doxycycline treatment on MMP-9 in AVM tissues. There was a trend for both total MMP-9 and active MMP-9 levels to be lower in the doxycycline group than in the placebo group. (mean ± SD) Discussion In this study, we demonstrated the feasibility of doxycycline, a tetracycline derivative, to decrease MMP-9 activity in AVM tissues. First, we demonstrated that a clinically relevant concentration of doxycycline decreased MMP-9 without affecting tissue viability in ex-vivo AVM tissues. Second, there was a trend that oral administration of doxycycline for as short as one week in AVM patients resulted in a decrease in MMP-9 at the target site – in the AVM nidus. By decreasing MMP-9 activity in AVM tissues, doxycycline may be able to restore the structural stability of AVM blood vessels and modify the clinical course of AVMs. Our data will provide the basis to conduct a clinical study to assess effects of tetracycline derivative treatment on the prevention of hemorrhage from AVMs. MMP-9 is a major enzyme that degrades the vascular extracellular matrix and has been implicated in a number of vascular diseases that involve abnormal angiogenesis and vascular remodeling [ 7 ]. MMPs are reported to be increased in cerebral aneurysms, atherosclerotic carotid plaque, and abdominal aortic aneurysm [ 8 - 10 ]. MMPs are emerging as a potentially new therapeutic target to treat vascular diseases. It has been proposed that pharmacological inhibition of MMPs may stabilize the unstable blood vessels and prevent complications such as vessel rupture [ 7 ]. In patients with abdominal aortic aneurysm, doxycycline treatment for one week prior to the repair surgery resulted in decreased MMP-9 and MMP-2 in the wall of the aneurysms [ 10 ]. Similar results have been reported in patients with atherosclerotic carotid plaques who received doxycycline for 2–8 weeks [ 11 ]. Our data using ex-vivo AVM tissues showed that abnormally high levels of MMP-9 expression in AVM tissues could be reduced by a clinically relevant concentration of doxycycline (i.e. 1–10 μg/ml). In patients with abdominal aortic aneurysm, doxycycline treatment at a conventional dose, 200 mg/day, resulted in a mean plasma concentration of doxycycline at 4.6 μg/ml with a range of 1.3 to 14.4 μg/ml in humans, and there was a corresponding reduction in plasma MMP-9 levels [ 12 ]. Similar plasma levels of doxycycline in mice successfully inhibited growth of experimental abdominal aortic aneurysms [ 13 ]. We were able to show that doxycycline at 10 μg/ml for as short as 48 hours can reduce MMP-9 expression in AVM tissues. A longer duration of treatment often used in clinical settings may result in a more pronounced reduction in MMP-9 expression in AVM tissues. One of the limitations of the ex-vivo study was the relatively short duration of doxycycline treatment. We chose 48 hours treatment to ensure tissue viability and integrity. Although we chose the doxycycline concentrations based on plasma doxycycline levels achieved by a commonly used therapeutic regimen, it is possible that tissue levels of doxycycline might be significantly lower than those of plasma. Furthermore, functional consequences of the reduction of MMP-9 in AVMs need to be examined in animal models or clinical studies. However, at the present time, there is no animal model that approximates recurrent intracranial hemorrhage from AVMs. There are models that can mimic certain aspects of the AVM phenotype. Hyperstimulation of mouse brain with VEGF using adenoviral transduction causes an increase in capillary density, increased MMP-9 activity, and may, in the appropriate genetic background, result in small vascular malformations [ 14 ]. Doxycycline can reduce capillary density and MMP-9 activity in this model [ 15 ]. Doxycycline has been shown in other human vascular diseases to reduce MMP levels. For example, Axisa et al. treated patients undergoing carotid endarterectomy with doxycycline or placebo for 2–8 weeks. Although MMP-1, MMP-3, and MMP-9 were reported to be increased, only MMP-1 had a statistically significant reduction by doxycycline; there was a trend for MMP-9 to be decreased by 22%. Baxter et al. used a much longer treatment period, 6 months, in patients with abdominal aortic aneurysm. They reported a gradual reduction of plasma MMP-9 levels by oral doxycycline over 6 month treatment period (baseline vs 3 months vs 6 months: 119 ± 38 vs 84 ± 33 vs 66 ± 24 ng/ml) [ 12 ]. To further explore the feasibility of a clinical application of doxycycline in modifying the clinical course of AVMs, we conducted a pilot clinical trial. This trial was primarily designed to test the effects of a short-term treatment with oral doxycycline on the expression of MMP-9 in AVM tissues. Despite the small sample size and short duration of treatment used in this study, there was a clear trend for one week of oral doxycycline treatment prior to surgical resection to reduce expression of both active and total MMP-9 in AVM tissues. These results indicate the feasibility of oral treatment with tetracycline derivatives in reducing MMP-9 activity in AVM tissues. Despite a large difference in the mean values of MMP levels in our pilot data, variance of MMP levels was high. Accordingly, a sample size of 46 in each group would be needed to demonstrate a difference in active MMP-9, assuming an alpha of 0.05 and power of 80%. However, variance may be reduced by utilizing a longer duration of treatment, as suggested by studies in other diseases [ 11 , 12 ]. A clinical study with a modestly larger sample size and longer treatment duration should suffice to verify effects of oral doxycycline on reducing MMP-9 activity in AVMs. There are some further considerations that may account for our high variance that could be avoided in future studies. Surgical AVM tissues consist mainly from nidal tissues, but they may contain intervening astroglial or neuronal tissues that may not express angiogenic factors to the same extent as nidal tissue. Histological examination of each AVM tissue fragments to confirm a predominance of nidal tissues may yield more homogenous tissue and decrease variability in MMP-9 levels. Some of the homogenized samples underwent a several "thaw-freeze" processes during the preliminary phase of the experiments, which may have increased variability in MMP-9 levels. Embolization treatment may affect production and activation of MMPs. Our previous study, however, showed no association between MMP-9 and embolization treatment in 37 AVM patients [ 16 ]. Based on the experience from this pilot study, future studies can be performed in a more standardized fashion to decrease tissue-to-tissue variability. We used different methods to measure MMP-9 in the medium collected from ex-vivo AVM tissues and homogenized AVM tissues from the pilot clinical trial. We chose zymography, a gel based method detecting enzymatic activity of MMPs, for ex-vivo AVM tissues, because the medium volume that can be used for MMP-9 assay was extremely limited, and our preliminary experiments showed higher sensitivity of zymography compared to ELISA. One of the disadvantages of using zymography is that "gel-to-gel" variability in sensitivity makes it difficult to combine the data obtained from multiple gels. On the other hand, ELISA method allows researchers to assay approximately 40 samples at a time when duplicates are used. In addition, establishing a standard curve, ELISA method can express MMP-9 expression values by the absolute unit such as μg/ml or ng/100 μg protein. Therefore, for the clinical pilot trial, in order to quantitatively compare MMP-9 expression among a large number of samples originally planned, we used ELISA to MMP-9 expression. Conclusions In summary, results suggest that doxycycline may have a therapeutic potential to reduce MMP-9 activity in AVM tissues and stabilize blood vessels that are prone to rupture. Our findings may provide a basis for a larger clinical trial to study effects of tetracycline derivatives in preventing AVM hemorrhage. Competing interests The author(s) declare that they have no competing interests. Authors' contributions TH, a corresponding author, participated in all aspects of the study and manuscript preparation. MMM participated in study design, experiments, analysis, interpretation of data, and manuscript preparation. JFL participated in experiments, analysis and interpretation of data. MTL participated in study design, analysis, interpretation of data, and manuscript preparation. WLY participated in study design, analysis, interpretation of data, and manuscript preparation. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC547916.xml
544589
Daily oviposition patterns of the African malaria mosquito Anopheles gambiae Giles (Diptera: Culicidae) on different types of aqueous substrates
Background Anopheles gambiae Giles is the most important vector of human malaria in sub-Saharan Africa. Knowledge of the factors that influence its daily oviposition pattern is crucial if field interventions targeting gravid females are to be successful. This laboratory study investigated the effect of oviposition substrate and time of blood feeding on daily oviposition patterns of An. gambiae mosquitoes. Methods Greenhouse-reared gravid and hypergravid (delayed oviposition onset) An. gambiae sensu stricto and wild-caught An. gambiae sensu lato were exposed to three types of substrates in choice and no-choice cage bioassays: water from a predominantly anopheline colonised ground pool (anopheline habitat water), swamp water mainly colonised by culicine larvae (culicine habitat water) and distilled water. The daily oviposition pattern and the number of eggs oviposited on each substrate during the entire egg-laying period were determined. The results were subjected to analysis of variance using the General Linear Model (GLM) procedure. Results The main oviposition time for greenhouse-reared An. gambiae s.s. was between 19:00 and 20:00 hrs, approximately one hour after sunset. Wild-caught gravid An. gambiae s.l. displayed two distinct peak oviposition times between 19:00 and 20:00 hrs and between 22:00 and 23:00 hrs, respectively. During these times, both greenhouse-reared and wild-caught mosquitoes significantly ( P < 0.05) preferred anopheline habitat water to the culicine one. Peak oviposition activity was not delayed when the mosquitoes were exposed to the less preferred oviposition substrate (culicine habitat water). However, culicine water influenced negatively ( P < 0.05) not only the number of eggs oviposited by the mosquitoes during peak oviposition time but also the overall number of gravid mosquitoes that laid their eggs on it. The differences in mosquito feeding times did not affect the daily oviposition patterns displayed. Conclusion This study shows that the peak oviposition time of An. gambiae s.l. may be regulated by the light-dark cycle rather than oviposition habitat characteristics or feeding times. However, the number of eggs laid by the female mosquito during the peak oviposition time is affected by the suitability of the habitat.
Background Although An. gambiae s.l. mosquitoes are nocturnal in their feeding and oviposition activities, the probable time of oviposition is determined by many factors including ambient temperature and light conditions, and the time the mosquito obtains its blood meal [ 1 , 2 ]. In addition, we hypothesised that the availability of a suitable larval habitat would also affect the mosquito's predisposition to oviposit. Anopheles gambiae is discriminative in its oviposition behaviour [ 3 ]. Its preferred larval habitats are fresh water pools that are generally small, transient and sunlit, devoid of vegetation and often turbid [ 4 - 6 ]. Oviposition tendency might therefore be related to location and availability of such sites. In this study, we compared the daily oviposition patterns and the number of eggs laid by An. gambiae s.s. and wild-caught An. gambiae s.l. on aqueous collections from habitats colonised by anopheline or culicine larvae respectively, and distilled water. Methods Mosquitoes Anopheles gambiae s.s. (MBITA strain; colonised since February 2001) mosquitoes from Mbita Point, western Kenya, were reared in a greenhouse [ 7 ] in water obtained from a natural ground pool colonised by anopheline larvae. Average temperatures and relative humidities were 31°C, 52 % during the day and 24°C, 72% at night. The mosquitoes were exposed to the natural photoperiod, 00° 25' South of the equator. A data logger (HOBO™) was used to record ambient conditions. Larvae were fed on Tetramin ® fish food. Adult mosquitoes were kept in standard mosquito rearing cages (30 × 30 × 30 cm) made of a metal wire frame with a solid metal base and covered with white nylon mosquito netting. They were offered a 6% glucose solution soaked in white paper towel wicks. Three-to-four-day-old females were offered two blood meals, one each day at 18.00 hrs, from the forearm of a human volunteer. The unfed mosquitoes were removed from the cage after each blood meal. Fully engorged females were left in the cages until they were gravid or hypergravid. Gravid mosquitoes are those that were provided with oviposition substrates on the third evening after their first blood meal. Hypergravid mosquitoes were provided with oviposition substrates one day later. Wild, indoor-resting, blood fed anopheline mosquitoes were collected during early morning hours from houses in Lwanda village of Suba district, western Kenya, by means of aspirators. They were immediately transported to the greenhouse, sorted out to obtain An. gambiae s.l. females and provided with 6% glucose solution. They were used in periodicity experiments on the second evening after collection, as described below. Oviposition substrates Turbid water taken from a natural ground pool colonised by anopheline larvae (anopheline habitat water), yellow-brown water from a reed swamp colonised by culicine larvae (culicine habitat water), and distilled water were used as oviposition substrates. Presence of larvae was determined by making five random dips using a 350 ml standard dipper. Oviposition substrate preference The experiments were carried out under greenhouse conditions in 25 cm cubic Plexi ® -glass cages, each fitted with a white netting top and a side sleeve opening. To determine oviposition substrate preference, individual gravid An. gambiae s.s. mosquitoes were exposed to 20 ml of each of the above substrates in a three-choice bioassay (n = 55). The substrates were held in black plastic oviposition cups (2 cm depth, 4 cm diameter), placed at equal distances from one another. Individual mosquitoes were released into the cages at about 17.00 hours and left overnight. The following morning, eggs oviposited on each substrate were counted under a dissection microscope. In subsequent replications, oviposition cups containing substrates were rotated such that they occupied different positions every time in the oviposition cages. Daily oviposition patterns in a no-choice bioassay Daily oviposition patterns of An. gambiae female mosquitoes on test oviposition substrates, which were offered individually, were determined as follows. Groups of five greenhouse-reared gravid and hypergravid An. gambiae s.s. females were held in separate cages into which anopheline or culicine habitat water or distilled water were introduced. Each mosquito and substrate combination treatment was replicated four times on each experimental day and the experiment repeated on three different days. At the end of the experiment, the mosquitoes that had laid in each group were identified by dissecting each under a dissection microscope and examining their ovaries for the presence of either retained eggs, coiled or uncoiled tracheolar skeins [ 8 ]. Daily oviposition patterns in a choice bioassay Groups of five gravid and hypergravid An. gambiae s.s. (ten cages of each) were placed in separate cages and allowed to choose from the three types of oviposition substrates. Similarly, groups of five wild-caught An. gambiae s.l. mosquitoes were offered a choice of the three substrates and their daily oviposition patterns monitored. The experiment was replicated twice on each experimental day and repeated on five different days with new mosquito batches. Individual species within the wild-caught An. gambiae mosquitoes that had laid were identified using polymerase chain reaction (PCR) [ 9 ]. Effect of the time of blood feeding on daily oviposition patterns The effect of the time of blood feeding of An. gambiae s.s. on its daily oviposition pattern was determined as follows. Four groups of three-to-four-day-old females were given two blood meals, one each day at 06.00 hrs, 18.00 hrs, 22.00 hrs or at 00.00 hrs, respectively. Unfed females were removed from the cages after every blood meal. Gravid mosquitoes were then provided with oviposition cups on the third day at 06.00 hrs and their daily oviposition patterns monitored. In all experiments, the oviposition cups were removed from the cages after every hourly interval, for 24 hours, starting at 18.00 hrs and replaced with freshly prepared ones. The eggs laid on each substrate were counted under a dissection microscope. To minimise disturbance that might have been due to exposure to white light, red light was used at night while replacing the oviposition cups. Data analysis Since oviposition trends for gravid and hypergravid females were similar, data for the two were pooled for analysis. The differences in the number of eggs laid on different oviposition substrates were compared statistically by analysis of variance using the General Linear Model (GLM) procedure. The effect of oviposition substrate on the number of either gravid or hypergravid mosquitoes contributing to the total egg number was similarly compared. Means were separated by the least significant difference (LSD) procedure. Data were subjected to log 10 (n+1) transformation to normalise their distribution. All the analyses were carried out using the SPSS ® statistical package, version 11.0. Results Oviposition substrate preference The mean number ± standard error (39.4 ± 6.1) of eggs oviposited on anopheline habitat water was significantly higher than that on the culicine (16.1 ± 4.6; P = 0.01) or distilled water (23.7 ± 5.3; P = 0.02). Daily oviposition patterns Daily oviposition patterns of An. gambiae s.s. on different substrates, offered in either no-choice or choice assays, are presented in Figures 1 and 2 , respectively. In both cases, the main oviposition time was between 19:00 and 20:00 hrs, approximately one hour after sunset, followed by a steady reduction in the number of eggs laid as the night progressed. In the choice bioassays, the gravid mosquitoes showed significant preference for anopheline habitat water over distilled ( P = 0.004) or culicine habitat water ( P = 0.001) throughout the daily cycle. In the no-choice bioassay, although the total number of eggs laid throughout the cycle on the different substrates was different, this was not statistically significant ( P = 0.4). However, during the peak oviposition time, the eggs laid on anopheline habitat water were significantly more than those on the culicine one ( P = 0.01) but not significantly more than those on distilled water ( P = 0.07). Egg-laying by mosquitoes of different ovary development stages was influenced considerably by the type of oviposition substrate ( P = 0.02). The hypergravid/ anopheline habitat water combination had the highest average number of mosquitoes (4.4 ± 0.3) laying their eggs, whereas gravid/culicine combination yielded the lowest response (2.5 ± 0.4; Table 1 ). Figure 1 Daily oviposition patterns of Anopheles gambiae s.s. on different oviposition substrates in a no-choice bioassay . Mean percentage (± SE) of the total eggs laid on each of three different oviposition substrates during 1-h time intervals. n = 24 cages containing five females each. Mosquitoes in each cage were exposed to one type of substrate under a natural LD cycle (sunset at 18:00). Figure 2 Daily oviposition patterns of Anopheles gambiae s.s. on different oviposition substrates in a choice bioassay . Mean percentage (± SE) of the total eggs laid on each of the three different oviposition substrates during 1-h time intervals. n = 20 cages containing five females each. Mosquitoes could choose from different substrates placed in the same cage under a natural LD cycle (sunset at 18:00). Table 1 The number of mosquitoes (Mean ± SE 1 ) contributing to the total eggs laid in each mosquito/ substrate combination. Mosquito/ Substrate Mean ± SE 1 Gravid/ Distilled water 3.3 ± 0.4 bc Gravid/ Anopheline habitat water 3.5 ± 0.4 ab Gravid/ Culicine habitat water 2.5 ± 0.4 c Hypergravid/ Distilled water 3.8 ± 0.4 ab Hypergravid/ Anopheline habitat water 4.4 ± 0.3 a Hypergravid/ Culicine habitat water 3.6 ± 0.4 ab 1 SE: Standard Error. n = 12 cages each containing five mosquitoes. Any two means sharing a letter in common are not significantly different at 5% level (LSD test). Unlike the greenhouse-reared An. gambiae s.s. , the wild-caught An. gambiae s.l. , which consisted of 23.9% An. gambiae s.s. ,71.7% An. arabiensis and 4.4% unidentified gravid females (n = 46), displayed two main oviposition times early in the night, between 19:00 and 20:00 hrs and between 22:00 and 23:00 hrs, respectively (Figure 3 ). These mosquitoes also showed significant preference ( P = 0.01) for anopheline habitat water over distilled or culicine habitat water. Figure 3 Daily oviposition patterns of wild-caught Anopheles gambiae s.l. on different oviposition substrates in a choice bioassay . Mean percentages (± SE) of the total eggs oviposited on each of the three different oviposition substrate during 1-h time intervals. n = 10 cages containing five females each. Mosquitoes could choose from different substrates placed in the same cage under a natural LD cycle (sunset at 18:00). An. gambiae s.s. females that obtained their blood meals later in the night displayed a somewhat broader oviposition peak time interval, ranging from 19:00 hrs to 22:00 hrs (Figure 4 ), than those that had fed earlier on, whose peak oviposition time interval was narrower (19:00 hrs to 21:00 hrs). Figure 4 Daily oviposition patterns of Anopheles gambiae s.s. fed at different times . Mean number (± SE) of eggs oviposited during 1-h time intervals. n = 8 cages containing five females each. Mosquitoes were kept under a natural LD cycle (sunset at 18:00). Discussion In the present study, the daily oviposition patterns of greenhouse-reared An. gambiae s.s. were well defined with oviposition peak times between 19:00 and 20:00 hrs, regardless of the type of oviposition substrate used. Haddow and Ssenkubuge [ 10 ] obtained comparable results using An. gambiae s.s. (KISUMU strain, western Kenya): about half of the eggs were laid during the first three hours of the night (18:00 – 21:00 hrs). On the other hand, oviposition by wild-caught mosquitoes from the coast of Kenya used by McCrae [ 1 ], comprising mostly An. gambiae s.s. , peaked much later at night in the hour following midnight. This suggests differences in oviposition patterns between our strain and that of Haddow and Ssenkubuge representing Lake Victoria populations, on one hand, and that used by McCrae representing the Kenyan coastal population, on the other. Studies of oviposition patterns of populations from different parts of eastern Africa may help shed further light on the question. In the current study, wild-caught An. gambiae s.l. , which were shown to contain a mixture of An. gambiae s.s. and An. arabiensis gravid females, displayed two distinct oviposition peak times in the first half of the night. The two peaks may be attributed to the two sibling species and suggests that this may also be an important factor in the diversity of oviposition patterns in the field in different geographical locations. The differences in the mosquito feeding times did not affect the timing of peak oviposition, although females that obtained their blood meals later in the night displayed a somewhat broader oviposition peak interval. Peak oviposition consistently occurred approximately one hour after sunset; therefore, a fall in light intensity might be one of the important cues that trigger oviposition in female An. gambiae that are physiologically ready to oviposit. On the other hand, McCrae [ 1 ] observed that the time of oviposition was a function of the time of blood feeding and not a result of an endogenous rhythm. Given the uniform oviposition peak times of mosquitoes that were fed at different times, daily oviposition among An. gambiae s.l. may also be endogenously regulated. Detailed experiments to demonstrate a free-running oviposition periodicity would clarify this. There was no difference in oviposition patterns displayed by gravid and hypergravid mosquitoes. Since significantly more gravid females exposed to the preferred substrate oviposited their eggs than those exposed to the less preferred one, gravid females that fail to find a suitable oviposition site on the night they are due may retain their eggs and oviposit early the next night as hypergravids. Gravid mosquitoes are generally attracted to water; however, the decision to oviposit may depend on additional olfactory signals [ 11 ] and /or contact stimuli received when the insects land on the water surface [ 12 ]. In this study and others [ 13 ], the gravid mosquitoes showed marked preference for the water taken from a site naturally inhabited by anopheline larval populations. This suggests 'memory' of similar information gathered by contact with the oviposition water at emergence or during larval period as in the case of Culex quinquefasciatus [ 14 ]. In this regard, gravid females might associate specific site characteristics from conspecific and heterospecific immatures, soil microbial activity [ 11 ], colour and turbidity of the oviposition substrate [ 13 ] with their suitability for sustaining progeny development. Conclusions This study shows that the peak oviposition time of An. gambiae s.l. may be regulated by the light-dark cycle rather than oviposition habitat characteristics or feeding times. However, the number of eggs laid during the peak oviposition time is affected by the suitability of the habitat. This suggests that there is a relationship between the investment made by the female mosquito with respect to the number of eggs laid in a given habitat and the potential fitness of the progeny. Females may use a series of site characteristics, including olfactory cues, to locate and oviposit at such sites. Our results on oviposition patterns differ from those reported on a coastal population, and suggest that a lot more work needs to be done to elucidate differences in this regard between different populations. Competing interests The authors declare that they have no competing interests. Authors' contributions LAS and KO conducted all the experimental work. AH, ALD, BGJK, JCB and JG co-ordinated and/or supervised the work. All authors actively contributed to the interpretation of the findings and development of the final manuscript and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544589.xml
543580
A survey of statistics in three UK general practice journal
Background Many medical specialities have reviewed the statistical content of their journals. To our knowledge this has not been done in general practice. Given the main role of a general practitioner as a diagnostician we thought it would be of interest to see whether the statistical methods reported reflect the diagnostic process. Methods Hand search of three UK journals of general practice namely the British Medical Journal (general practice section) , British Journal of General Practice and Family Practice over a one-year period (1 January to 31 December 2000). Results A wide variety of statistical techniques were used. The most common methods included t-tests and Chi-squared tests. There were few articles reporting likelihood ratios and other useful diagnostic methods. There was evidence that the journals with the more thorough statistical review process reported a more complex and wider variety of statistical techniques. Conclusions The BMJ had a wider range and greater diversity of statistical methods than the other two journals. However, in all three journals there was a dearth of papers reflecting the diagnostic process. Across all three journals there were relatively few papers describing randomised controlled trials thus recognising the difficulty of implementing this design in general practice.
Background "Diagnosis is the keystone of good medical practice"[ 1 ] General practitioners (GPs) are primarily diagnosticians [ 2 ] yet it appears that diagnosis remains their Achilles heel[ 3 ]. The problem has its origins in a misunderstanding of the differences of the five Ps (patients, pathologies, presentations, prevalences and predictive values) in hospital practice compared to primary care[ 4 ]. Decisions made by GPs are different from those made by hospital clinicians. The precise diagnostic labels may be less important than deciding on an appropriate course of action. Hence, diagnoses are often framed in terms of binary decisions; treatment versus non-treatment, disease versus non-disease, referral versus non-referral, and serious versus non-serious for example[ 4 ]. From a statistical viewpoint the binary decision making process has a lot of appeal. For example, the use of the naïve Bayes' discriminant function (and from it the derivation of likelihood ratios) is appropriate. Proponents of Bayes' argue for its simplicity and ease of interpretation[ 5 , 6 ]. In contrast, opponents argue that data are not used efficiently if they are simply ploughed through the "black box" of Bayes'[ 7 , 8 ]. Whatever the rights and wrongs of Bayes' as a technique it is time for GPs to become more familiar with statistical methods aimed at diagnosis. In relation to haematuria (blood in the urine) and the diagnosis of urological malignancy two of the authors of this paper (NS and ASR) have used Bayesian techniques in order to seek to refine diagnostic discrimination by general practitioners [ 9 ]. The results from this work have been incorporated successfully into local primary care oriented referral guidance. Many medical journals, both generalist[ 10 , 11 ] and specialist [ 12 - 18 ], have been reviewed for their statistical content. Articles have been published in the fields of radiology, [ 12 - 14 ] otolaryngology, [ 15 , 16 ] rehabilitation medicine[ 17 ] and ophthalmology[ 18 ] to name but a few. However, general practice is under researched in this area[ 19 ]. The aim of this paper is to review three leading UK journals in general practice and to see what statistical methods are being used. It is not our intention to see if the methods are being used correctly but to look at the range of techniques reported. The outcome of this research should give pointers to the future education of GPs who wish to undertake research. Methods Three statisticians (MJC, ASR and GKA) (two of them holding Chartered status of the Royal Statistical Society) including one Professor, one Senior Lecturer and one Lecturer each reviewed one leading UK journal in general practice. The fourth author (NS) is a Primary Care Physician. The journals chosen were the British Medical Journal (BMJ) (general practice section) , British Journal of General Practice (BJGP) and Family Practice . These three journals were chosen because they reflected the main primary care journals in the UK. The journals were hand searched for original research articles over a one-year period (1 January to 31 December 2000). Articles were classified for both their statistical content and methods of design according to criteria laid down elsewhere[ 10 , 20 ]. Tables 1 and 2 list the classification criteria used for both study design and statistical methods. Letters were excluded on the grounds that they are typically responses to previously published material rather than original contributions in themselves. We are aware, of course, that not all primary care research is published in these three journals alone and we comment on this later. Table 1 Classification of design methods (after Wang and Zhang, 1988) [19] Design method Case report Cross-sectional survey Retrospective study Prospective study Clinical trial basic science study Table 2 Classification of statistical methods (after Emerson and Colditz, 1983) [10] Category Brief description No statistical methods or descriptive statistics No statistical content, or descriptive statistics only (e.g., percentages, means Standard deviations, standard errors, histograms Contingency tables Chi-square tests, Fisher's test, McNemar's test Multiway tables Mantel-Haenszel procedure, log-linear models Epidemiological studies Relative risk, odds ratio, log odds, measures of association, sensitivity, specificity t-tests One-sample, matched pair, and two sample t- tests Pearson correlation Classic product-moment correlation Simple linear regression Least-squares regression with one predictor and one response variable Multiple regression Includes polynomial regression and stepwise regression Analysis of variance Analysis of variance, analysis of covariance, and F-tests Multiple comparisons Procedures for handling multiple inferences on same data sets (e.g., Bonferroni techniques, Scheffe's contrasts, Duncan's multiple range procedures, Newmann-Keuls procedure) Non-parametric tests Sign test, Wilcoxon signed ranks test, Mann- Whitney test, Spearman's rho, Kendall's tau, test for trend Life table Actuarial life table, Kaplan-Meier estimates of survival Regression for survival Includes Cox regression and logistic regression Other survival analysis Breslow's Kruskal Wallis, log rank, Cox model for comparing survival Adjustment & standardisation Pertains to incidence rates and prevalence rates Sensitivity analysis Examines sensitivity of outcome to small changes in assumptions Power Loosely defined, includes use of the size of detectable (or useful) difference in determining sample size Transformation Use of data transformation (e.g., logs) often in regression Cost-benefit analysis The process of combining estimates of cost and health outcomes to compare policy alternatives Other Anything not fitting the above headings includes cluster analysis, discriminant analysis, and some mathematical modelling The main study was preceded by a pilot phase in which a random sample of 10 articles was classified both by statistical content and study design by the three statisticians. Where there were differences of opinion, consensus was reached by discussion. We met once to discuss our classification system, and to iron out differences of opinion. One problem lay in how we actually classified study design. For example, one of use used the phrase 'cross-sectional survey' while another used the phrase 'questionnaire survey' when both meant the same in terms of study design. Another problem was that we missed some of the statistical techniques (where there were many) and this required much more careful reading of the articles when we carried out the main survey. We did not carry out a formal reliability study of the pilot phase but instead relied on our experiences both as statisticians, and as journal reviewers. Similarly we chose not to carry out a formal reliability analysis in the main study. Results The total number of articles reviewed over a one year period was as follows: BMJ (general practice section) (n = 79), BJGP (n = 145) and Family Practice (n = 81). Study design The most common design was that of a cross-sectional survey being found in 24.1%, 39.3% and 35.1% of articles in the BMJ , BJGP and Family Practice respectively (Table 3 ). Although we classified articles by the term 'cross-sectional survey' this was not necessarily the choice term adopted by the journal. Sometimes the phrase 'questionnaire survey' was used and we assumed this was data collected cross-sectionally. We found a similar difference in nomenclature for our phrase 'cohort study' in which the phrase 'prospective survey' was also found. The highest proportion of qualitative studies was in Family Practice (21.0% compared to an average of 11.8%). Qualitative studies included those encompassing terms such as 'focus groups' and 'semi-structured interviews' for example. Figure 1 shows the proportion of papers ranked by a qualitative design. For all three journals, diagnostic studies were infrequently used. Examples of these include those based on screening (e.g., the usefulness of N-terminal brain natriuretic peptide level for screening of patients with heart failure), and calculating the sensitivity and specificity of diagnostic tests (e.g., Helicobacter pylori for the detection of peptic ulcer). Examples of more unusual study designs include those based on video recordings, literature reviews and quasi-experimental designs. Table 3 Design methods BMJ BJGP Family Practice Overall Designs n (%) n (%) n (%) n (%) Cross-sectional survey 19 (24.1) 57 (39.3) 31 (34.8) 107 (35.1) Qualitative study 3 (3.8) 16 (11.0) 17 (21.0) 36 (11.8) Cohort study 8 (10.1) 21 (14.5) 4 (4.9) 33 (10.8) RCT 14 (17.7) 7 (4.8) 8 (9.9) 29 (9.5) Reviews 4 (5.1) 8 (5.5) 2 (2.5) 14 (4.6) Reliability/diagnostic 2 (2.5) 8 (5.5) 1 (1.2) 11 (3.6) Case-control study 4 (5.1) 1 (0.7) 3 (3.7) 8 (2.6) Cluster RCT 4 (5.1) 1 (0.7) 2 (2.3) 7 (2.3) Other 21 (26.6) 26 (17.9) 13 (16.0) 60 (19.7) Total articles 79 145 81 305 Note RCT = randomised controlled trial. Figure 1 Proportion of papers ranked by a qualitative design Statistical methods The range of statistical methods reported can be seen in Table 4 . The number of methods exceeds the number of articles as some reported more than one technique. There are differences between the journals. The BMJ shows a greater range and breadth of articles than Family Practice . More sophisticated techniques are reported more often in the BMJ than either of the other two journals. In the BMJ , the two most common statistical methods used were logistic regression (n = 14, 17.7%) and the Chi-squared test (n = 13, 16.5%). The two least common were the Mantel-Haenszel statistic (n = 1, 1.3%) and Cronbach's alpha (n = 1, 1.3%). Relatively new innovations such as random effects models were seen in both the BMJ and the BJGP . The least sophisticated statistical methods appeared in Family Practice . Methods based on likelihood ratios were seldom found in either the BMJ or BJGP and not at all in Family Practice . Nonparametric tests were often unspecified but where they were included Mann-Whitney U test, Spearman's correlation coefficient and the Wilcoxon matched-pairs signed ranks test. Multiple comparisons included Bonferonni techniques and Scheffe's contrasts. Survival analysis included Kaplan-Meier curves and Cox regression. Table 4 Statistical methods BMJ BJGP Family Practice Overall Methods n (%) n (%) n (%) n (%) No statistics or simple summaries 23 (29.1) 47 (32.4) 33 (40.7) 103 (33.8) Chi-squared tests 13 (16.5) 40 (27.6) 19 (23.5) 72 (23.6) t-tests 7 (8.9) 22 (15.2) 17 (21.0) 46 (15.1) Logistic regression 14 (17.7) 19 (13.1) 11 (13.6) 44 (14.4) Nonparametric 11 (13.9) 24 (16.6) 4 (4.9) 39 (12.8) Odds ratios/relative risks 11 (13.9) 13 (9.0) 14 (17.3) 38 (12.5) Regression 9 (11.4) 10 (6.9) 11 (13.6) 30 (9.8) Sample size/power 6 (7.6) 17 (11.7) 3 (3.7) 26 (8.5) Summaries with CIs 9 (11.4) 3 (2.1) 6 (7.4) 18 (5.9) Kappa 2 (2.5) 9 (6.2) 4 (4.9) 15 (4.9) Sensitivity/specificity 4 (5.1) 10 (6.9) 1 (1.2) 15 (4.9) Pearson correlation 2 (2.5) 6 (4.1) 6 (7.4) 14 (4.6) Multiple comparisons 2 (2.5) 4 (2.8) 4 (4.9) 10 (3.3) ANOVA 5 (6.3) 4 (2.8) 9 (3.0) Mantel-Haenszel 1 (1.3) 5 (3.4) 2 (2.5) 8 (2.6) Random effects models 4 (5.1) 4 (2.8) 8 (2.6) Cronbach's alpha 1 (1.3) 5 (3.4) 1 (1.2) 7 (2.3) Fisher's exact test 7 (4.8) 7 (2.3) Likelihood ratio 3 (3.8) 3 (2.1) 6 (2.0) Survival analysis 6 (7.6) 6 (2.0) Other 4 (5.1) 37 (25.2) 10 (12.3) 51 (16.7) Total articles 79 145 81 305 Notes CIs = confidence intervals. ANOVA = analysis of variance. One-third of all articles reported no statistics or simple summaries (for example, mean, median, percentage, standard deviation, interquartile range). No journal article with a qualitative design had any statistical content. A large number of articles reported other statistical methods, in particular the BJGP . This was due to a wide range of statistical methods being reported only once. Examples include time series, multilevel modelling and factor analysis. In others, we could not decipher which statistical techniques had been used. Table 5 shows the rank order of the statistical methods by each journal. Differences between the journals can be seen more clearly. Table 5 Ranking of statistical techniques BMJ BJGP Family Practice Methods Rank Rank Rank Chi-squared tests 2 1 1 t-tests 7 3 2 Logistic regression 1 4 4.5 Nonparametric 3.5 2 9 Odds ratios/relative risks 3.5 6 3 Regression 5.5 7.5 4.5 Sample size/power 8.5 5 11 Summaries with CIs 5.5 17 6.5 Kappa 15 9 9 Sensitivity/specificity 11.5 7.5 13.5 Pearson correlation 15 11 6.5 Multiple comparisons 15 15 9 ANOVA 10 15 Mantel-Haenszel 17.5 12.5 12 Random effects models 11.5 15 Cronbach's alpha 17.5 12.5 13.5 Fisher's exact test 10 Likelihood ratio 13 17 Survival analysis 8.5 Notes Excluding other methods and no statistics/simple summaries. CI = confidence interval. ANOVA = analysis of variance. Discussion Two-thirds of all journal articles relied on some type of statistical analysis beyond descriptive statistics (Table 4 ). The Chi-squared test and t-tests were commonly used in the BJGP and Family Practice . Papers in the BMJ and the BJGP used more sophisticated statistical methods than Family Practice (Table 4 ). While both the BMJ and the BJGP used sophisticated methods, the BMJ used them more often. Why might this be so? The sophistication of methods used is influenced by three factors. First, issuing instructions to authors of a statistical nature. This requires a bank of statisticians available for review to which the BMJ has access. Second, general articles on statistical aspects of writing papers. Third, tutorial type articles explaining specific techniques. The BMJ continues to take a lead in the latter two areas and indeed published statistical guidelines for contributions to medical journals over 20 years ago[ 21 ]. Despite the lack of sophistication in Family Practice , there has been a trend of using more advanced statistics elsewhere,[ 14 , 15 , 17 , 20 ] and this has been linked to the increasing availability of computer packages[ 14 ]. The BJGP is currently struggling to find statistical reviewers (personal communication by Editor to ASR). It is perhaps too easy for us to lay blame at the Editors door for this lack of sophistication. Statisticians are relatively rare, and review, for the most part, is unpaid. Although these three journals publish a large proportion of the research in general practice within the UK, they by no means represent 100% of it. To look at this further we examined the year 2000 and undertook a MEDLINE search using the key indexing phrase 'General Practice'. We found over 800 articles in a diversity of journals. Articles were published in the fields of rheumatology, medical ethics, obstetrics, public health, clinical pharmacology, clinical neurology and telemedicine to name but a few. We chose to look at the year 2000. Would our results be different had we selected a different year? The published literature suggests otherwise. In a 20 year old study, Emerson and Colditz[ 10 ] found t-tests (44%) and Chi-squared tests (27%) were the most common statistical methods reported although now Chi-squared tests are more common than t-tests (Table 5 ). Given the emphasis on statistical computing today we might have expected less reliance on these two methods. What lies behind this lack of progress? Altman and Goodman[ 22 ] looked at the speed of the transfer of technology of new statistical methods into the medical literature. They concluded that many methodological innovations of the 1980s had still not made their way into the medical literature of the 1990s suggesting a typical lag-time of 4–6 years. Lag-time is likely to be related to quality statistical review and this may be longer in journals with less impact. It is also worth reporting that since we carried out this survey (year 2000) there will have been a modest increase in the use of newer, more sophisticated statistical techniques. Now let us turn to study design. The gold standard research design is considered to be the randomised controlled trial (RCT). It has been acknowledged that carrying out RCTs in general practice are difficult[ 23 , 24 ]. In our survey we found few RCTs (Table 3 ). There are particular problems of recruitment with respect to primary care. Many issues have been discussed. For example, most practices have no formal contractual arrangement to participate in research and may be unwilling to participate unless there is immediate benefit to their patients. It is known that motivating practices for long-term follow-up studies particularly is not easy[ 25 ]. Practices may feel uncomfortable about randomising their patients[ 26 ] but, delegation of this duty to another may lead to a breakdown of the special doctor/patient relationship. There are statistical and sample size concerns also. Randomisation by practice (so-called cluster randomisation) leads to larger sample sizes being required[ 27 , 28 ]. What are the issues here? Are they really that different from secondary care? A recent publication posed the question 'What do residents really need to know about statistics?'[ 29 ]. The authors surveyed six journals and catalogued them for their statistical and methodological content. The most popular statistical tests across the whole range of journals were the Chi-squared test followed by the t-test. The authors concluded that with knowledge of each of these two tests clinicians should be able to interpret up to 70% of the medical literature. Conclusions For all three journals there was a dearth of articles reflecting the diagnostic process. Why is this? It has already said that diagnosis is the Achilles Heel of GPs[ 3 ]. If it is not to remain this way we must start to educate doctors. The question is how. The latest "Tomorrow's Doctors"[ 30 ] states that students must have "Adequate knowledge of the sciences on which medicine is based and a good understanding of the scientific methods including principles of measuring biological functions, the evaluation of scientifically established facts and the analysis of data". Clearly, there is a role for teaching statistics in the education of doctors who wish to undertake research. The much greater prevalence of methods concerning binary data (Chi-squared test, logistic regression, odds ratios/relative risks) over methods concerned with continuous data should be reflected in our (statistical) teaching. Initial training in means, medians and modes should be replaced by relative risk, absolute risk and numbers needed to treat. Competing interests The author(s) declare that they have no competing interests. Authors' contributions Three authors (ASR, GKA and MJC) carried out the literature review while all four authors contributed to the writing. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543580.xml
520755
Determination of the differentially expressed genes in microarray experiments using local FDR
Background Thousands of genes in a genomewide data set are tested against some null hypothesis, for detecting differentially expressed genes in microarray experiments. The expected proportion of false positive genes in a set of genes, called the False Discovery Rate (FDR), has been proposed to measure the statistical significance of this set. Various procedures exist for controlling the FDR. However the threshold (generally 5%) is arbitrary and a specific measure associated with each gene would be worthwhile. Results Using process intensity estimation methods, we define and give estimates of the local FDR, which may be considered as the probability for a gene to be a false positive. After a global assessment rule controlling the false positive error, the local FDR is a valuable guideline for deciding wether a gene is differentially expressed. The interest of the method is illustrated on three well known data sets. A R routine for computing local FDR estimates from p -values is available at . Conclusions The local FDR associated with each gene measures the probability that it is a false positive. It gives the opportunity to compute the FDR of any given group of clones (of the same gene) or genes pertaining to the same regulation network or the same chromosomic region.
Background Microarrays are part of a new class of biotechnologies that allow the monitoring of the expression level of thousands of genes simultaneously. Among the applications of microarrays, an important task is the identification of differentially expressed genes, i.e genes whose expressions are associated with the status of the patient (treatment/control for example). The biological question of the identification of differentially expressed genes can be restated as a one (for paired data) or two-sample (for unpaired data) hypothesis testing procedure: is the gene differentially expressed between the two situations? However, when thousands of genes in a microarray data set are evaluated simultaneously by fold changes or significance tests approach, multiple testing problems immediately arise and lead to many false positive genes. In this 'one-by-one gene' approach the probability of detecting false positives rises sharply. The False Discovery Rate (FDR), is defined as the expected fraction of false rejections among those hypotheses rejected. In their seminal paper Benjamini & Hochberg [ 1 ] provided a distribution free procedure (BH) for choosing a threshold on p -values that guarantees that the FDR is less than a target level α. The same paper demonstrated that the BH procedure is more powerful than the Bonferroni method that controls the familywise error rate. The FDR gives an idea of the expected number of false positive hypotheses that a practitioner can expect if the experiment is done an infinite number of time. As usual with expectation, it gives very little information about the number of false discovery hypotheses in a given experiment. Motivation The value of 1, 5 or 10% for the FDR, which determines the threshold t , is arbitrary. Storey and Tibshirani [ 2 ] stressed the importance of assessing to each feature its own measure of significance. They proposed to use the q -value, where P i is the p -value of the ordered gene i , R i is the total number of rejected genes whose p -values are less than the threshold t = P i and is an estimate of the total number of non differentially expressed genes, m 0 . The q -value is appealing because it gives a measure of significance that can be attached to each gene, but it must be stressed that it is not an estimate of the probability for the gene to be a false positive. The q -value is generally lower than the latter because it is computed using all the genes that are more significant than gene i . Obviously a gene whose p -value is near to the threshold t does not have the same probability to be differentially expressed than a gene whose p -value is close to zero. Therefore the q -value gives a too optimistic view of the probability for the gene to be a false positive. Therefore it is interesting to obtain an estimate of the FDR attached to each gene, called local FDR, from an inferential point of view and without any assumption about the distribution of the p -values under H 1 . Results Let H 0 ( i ) = {gene i is not differentially expressed}. Let the local FDR be the probability that a given gene is not differentially expressed. More specifically, FDR ( i ) is the probability that a gene, whose p -value is P i , is not differentially expressed, taking into account the whole set of tests. A raw local FDR estimate is defined in a first step. In a second step the local FDR estimate is defined as a smoothed value based on the raw values. Let P 1 < … < P m denote the ordered p -values for testing H 0 ( i ). The raw local FDR estimate for gene i is: where where λ is a tuning parameter and W (λ) = #{ i , P i > λ}, see Storey [ 3 ]. Assume that the p -values for the non-differentially expressed genes are independent. The raw local FDR estimate has the following properties: • Under H 0 ( i ) and H 0 ( i - 1) and if E ( ) = m 0 , ( i , λ) is unbiased with mean 1. • Let ( i , m 0 ) = m 0 ( P i - P i -1 ). Under H 0 ( i ) and H 0 ( i - 1) and if m 0 is known, V ( ( i , m 0 )) = /[( m 0 + 1) 2 ( m 0 + 2)] ≈ 1, for m 0 large enough. This value is a lower bound for V ( ( i , λ)) when m 0 is unknown. • The variance of the raw local FDR under H 1 is generally much smaller than under H 0 . • where q j is the q -value of gene j . The q -value may thus be viewed as the mean of the local FDR of the genes with p -values lower than P j . ( i , λ) is generally a very variable estimator. Moreover the local FDR should increase with the p -value. This is not the case for the raw local FDR. Therefore it is necessary to use a smoothed estimate. The smoothed local FDR( i ) is where f i is a smoothing function of the ( j , λ) for j = 1, m , computed at position P i . ( i , λ) gives a very valuable guideline for the choice of a threshold. One may consider the curve of the local FDR versus the index of the gene ordered by their p -values: a good candidate for the threshold should be a point with a high second order derivative, which corresponds to an abrupt change in the slope of the curve (see the examples of the following section). The second order derivative of the smoothed local FDR can be computed numerically using finite differences. As an interesting application of the local FDR, it is possible to compute the FDR associated with a class of genes or clones by summing up the local FDR estimate of each clone or gene: one may consider for example clones corresponding to the same gene, genes known involved in a given regulatory network, or gene from the same chromosomic region, and associate a FDR with the whole class. These genes do not need to have consecutive p -values. The following sections demonstrate how the local FDR can be useful using the data of well known experiments. Local FDR on Golub data set Golub [ 4 ] were interested in identifying genes that are differentially expressed in patients with two types of leukemias (ALL, AML). Gene expression levels were measured using Affymetrix high-density chips containing 6817 human genes. The learning set comprises 27 ALL cases and 11 AML cases. Data are available in the R multtest package. We used the preprocessing proposed by the authors and the p -values based on random permutations of the ALL/AML labels on Welch t -statistics for each gene, Dudoit [ 5 ], on the 3051 remaining genes. m 0 is estimated with bootstrap method as suggested by Storey and Tibshirani and implemented in the library GeneTS of software R. Figure 1(a) presents the ( i ) for ordered genes and 1(b) presents the smooth curves obtained using lowess with a span of 0.2 and an adaptative moving average method. We can see that there is an abrupt change of the smoothed local FDR around gene number 500 which corresponds to a threshold t = 0.15 for the p -value. This may be an indication about the threshold. The Figure 1(c) presents a zoom of the Figure 1(b) for the first 600 p -values. We can see in Figure 1(c) that if we select the 438 (14%) top genes, we obtain a q -value equal to 0.0078 while the 438 th gene has a local FDR equal to 0.027. It must be noticed that there is a big difference between the two measures of FDR because the numerous regulated genes with very small p -values have a great influence on the q -value, which is not the case of the local FDR (see Figure 1(c) ). The p -values have been obtained using random permutations. Therefore the p -values are discrete with several genes possessing the same p -value. Therefore the values of ( i , λ) may be equal to 0 because the difference between two successive p -values is 0. The discrete structure of the p -values implies a departure from the theoretical continuous uniform distribution. This explains why the moving average smoothing creates discrete jumps which appear in Figure 1(c) . If the distribution of the statistics under H 0 is correct, the p -values are distributed as a uniform distribution over [0, 1]. The empirical distribution of the high observed p -values (say above 0.5) is far from the uniform distribution. There are several non-exclusive possibilities to explain this: more than 50% of the genes are differentially expressed, the gene results for non-differentially expressed are correlated or there is a technical problem in the random permutations of the Welch t -statistics. Local FDR on Breast Cancer data set Storey and Tibshirani [ 2 ], have analysed in detail data from Hedenfalk [ 6 ] on 15 microarrays on breast cancer. Using the same p -values, we have computed local FDR estimates. The three genes which have been analysed in detail by Storey and Tibshirani [ 2 ] are presented in Table 1 . One can see that the smooth local FDR estimate is generally greater than the q -value and gives a better idea of the probability that a gene is a false positive. For example, at the level of 5%, CTGF will be considered as differentially expressed on the basis of the q -value while it will be considered as non differentially expressed using the local FDR. Figure 2(a) presents the ( i ) for ordered genes and 2(b) presents the smooth curves obtained using lowess with a span of 0.2 and moving average methods. The two smoothing methods give similar results. Setting λ = 0.5, Storey and Tibshirani [ 2 ] estimate that 67% of the 3170 genes in the data are not differentially expressed. The asymptote near 1 of the smooth curve supports this estimation. Local FDR on ApoAi data The goal of the study is to identify genes with altered expression in the livers of two lines of mice with very low HDL cholesterol levels compared to inbred control mice. The mouse model is the apolipoprotein AI (ApoAI) knock-out mice. ApoAI is a gene known to play a pivotal role in HDL metabolism. The statistical analysis is described in Dudoit [ 7 ]. Height clones are expected to be differentially expressed between the control and the knock-out mices because they are clones of the ApoAI gene or of genes coregulated with ApoAI. The height clones are actually the 8 top clones detected by the statistical tests. However there are other following clones which seem statistically significant if we consider the q -value. We can see on the Figure 3(c) that the local FDR values are much higher than the q -values. Figure 3(a) presents the ( i ) for ordered clones and Figure 3(b) presents the smooth curves obtained using lowess with a span of 0.2 and moving average methods. The two smoothing methods give different results at the two ends of the [0, 1] interval. The moving average method which uses a special adaptative algorithm for the ends gives a better smoothing. This is particularly important for the clones with a small p -value for which it is crucial to obtain good estimates of the probability of being false positives. The lowess smoothing does not work well for the 50 first clones. In this particular case the default smoothing parameter f = 0.2 is not well suited and should be lower. However if it is chosen too low, the smoothing will not fit well the rest of the curve. There are two clones of the gene Apo-AI. If we want to estimate the FDR of these two clones taken in a whole, we compute the mean of the smoothed local FDR of the two clones (the first and the height top clones) and obtain a local FDR for the gene Apo-AI, which is equal to . This example shows that it is possible to estimate the local FDR of any group of clones. This opportunity provided by the local FDR is certainly one of its major advantage with many potential applications. Discussion The curve of the smoothed local FDR is an efficient tool to summarize the information about the number and the statistical significance of differentially expressed genes, and may also be used to give an indication about the validity of the statistical assumptions. Moreover it is a valuable tool to choose the threshold for separating the differentially expressed genes from the non-differentially expressed one: one can choose a value of t maximizing the second derivative. Alternatively one can use a cost function and choose the threshold that minimizes the mean cost for a given cost function: using cost of the experiment, cost of false positive gene validation and the profit of discovering a differentially expressed gene, it is direct to compute the optimal strategy for choosing the threshold. Note that a decision rule based on the local FDR would lead to a different set of selected genes than the usual one obtained by controlling the FDR. Consider the set of tests for which the local FDR is below 0.05, say. This set is not identical to the set identified by the standard criterion that FDR < 0.05. The local FDR is higher than the q -value. Therefore the first set is strictly included in the second one. The local FDR rule is therefore more conservative than the usual FDR one. Conclusions The p -value gives the probability that a non differentially expressed gene would be as or more extreme than the gene under concern. The q -value indicates the estimated proportion of genes as or more extreme than the gene under concern that are a false positive. The local FDR gives the estimated proportion of genes around the gene under concern which are false positive. The latter may be used as the probability that the gene under concern is a false positive, taking into account the multiplicity of the test. One of the major interest of the local FDR is that it gives the opportunity to compute the FDR of any given group of clones (of the same gene) or genes pertaining to the same regulatory network or the same chromosome. Methods Model Basically, the various procedures proposed in the literature aim to test the null hypothesis H 0 ( i ) = {gene i is not differentially expressed}. Let consider a particular experiment. We observed the differential expression of the genes and compute the associated ordered p -values P i . In the following we will use the classical property: the p -values corresponding to non differentially expressed genes are uniformly distributed over [0, 1]. Furthermore, we will assume, as often, that these p -values are independent. However, the independence of the p -values of differentially expressed genes is not required. Consider a multiple testing situation in which m tests are being performed. Let m 0 be the number of non differentially expressed genes. Let I ( t ) be the set of the genes having a p -value lower than t : I ( t ) = { i : P i ≤ t } and R ( t ) = # I ( t ), its cardinal. Let V ( t ) = #[ I ( t ) ∩ ( i ∈ H 0 )] and S ( t ) = #[ I ( t ) ∩ ( i ∈ H 1 )]. Using a threshold t , the m genes can be classified according to the following 2 × 2 table 2 : The Family Wise Error Rate (FWER) is defined to be FWER = P [ V ( t ) ≥ 1]. A classical way to control FWER is given by the Bonferroni inequality. This quantity corresponds to the most direct extension from a test hypothesis procedure but can be very restrictive in a multiple testing procedure. The status of the gene associated with the P i is an unobserved value. It is the same framework as point process (see for example [ 8 ]). In fact we observe R ( t ) = V ( t ) + S ( t ) the sum of two counting processes. The first one V ( t ) is a counting process associated with non differentially expressed gene. Since the p -values under H 0 are uniformly distributed, V ( t ) has a binomial distribution with parameter m 0 and t . The intensity of V ( t ) is constant and proportional to m 0 . S ( t ) is the counting process associated with gene under H 1 and very few can be said about its distribution. One may expect the intensity of S ( t ) to be decreasing with t . The false discovery rate is defined as: It corresponds to the expected proportion of rejections that are incorrect. The BH procedure works as follows. Let P 1 < … < P m denote the ordered p -values. Calculate k = max i { P i ≤ α i / m }. The procedure rejects all null hypotheses for which P i ≤ P k . If the tests are independent, this procedure ensures that Let FDR( t ) be the FDR when rejecting all null hypotheses with P i ≤ t . Because the p -values of non-differentially expressed genes are uniformly distributed over [0, 1], a natural estimate of FDR( t ) is Therefore the problem is to estimate m 0 . Storey [ 3 ], proposed to estimate m 0 with where λ is a tuning parameter. In particular the case λ = 0 leads to . This is the most conservative case and corresponds to the BH procedure. Since the practical implementation of Storey method gives reasonably good results, we used it in the examples. FDR is defined as the expectation of the ratio of two counting processes V ( t ) and R ( t ): FDR( t ) = E [ V ( t )/max( R ( t ), 1)]. The expectation of V ( t ) is m 0 t and R ( t ) is observed. Therefore, Storey [ 3 ] propose to use the following estimate: The ratio of the expectations differs from the expectation ratio but Storey [ 3 ] proved that E ( ( t , λ)) ≥ FDR( t ) using a convexity argument. Definition and Estimation of the Local FDR As stated before, V ( t ) and R ( t ) are counting (i.e. cumulative) processes. It would be very interesting to estimate the ratio of the local intensities of the two processes at point t . The intensity of process V ( t ) is equal to m 0 and thus is known, provided that we know m 0 . The intensity of process R ( t ) is unknown, but R ( t ) is observed. Therefore, using point process methods it is possible to estimate its intensity at each point t . We first define the cumulative processes from t 1 to t 2 : Let 0 ≤ t 1 < t 2 , I ( t 1 , t 2 ) = { i : t 1 < P i ≤ t 2 }, R ( t 1 , t 2 ) = # I ( t 1 , t 2 ), V ( t 1 , t 2 ) = #[ I ( t 1 , t 2 ) ∩ ( i ∈ H 0 )] and S ( t 1 , t 2 ) = #[ I ( t 1 , t 2 ) ∩ ( i ∈ H 1 )]. FDR ( t 1 , t 2 ) is defined as the expected ratio of V ( t 1 , t 2 ) and R ( t 1 , t 2 ): It is a generalization of the usual FDR: if t 1 = 0 and t 2 = t then FDR( t 1 , t 2 ) = FDR( t ). So, the natural estimate of FDR( t 1 , t 2 ) is: The substitution of 0 by t 1 does not change the proof, so using the same convexity argument as Storey [ 3 ], we obtain the following property: E ( ( t 1 , t 2 , λ)) ≥ FDR ( t 1 , t 2 ). The local FDR is the FDR( t 1 , t 2 ) for small intervals [ t 1 , t 2 ]. If we want to estimate the local FDR around the p -value of the gene i , the question can be restated as how to estimate the ratio of the intensities of two processes around a given point P i . The intensity of process R ( t ) has to be estimated at each value of t . It is possible to consider small windows of size h , or alternatively, to consider windows of different sizes corresponding to a fixed count for R ( t ). We have chosen the latter solution, for windows of variable size seem more appealing in the particular context. Let FDR( i ) be the local FDR around P i . To estimate FDR( i ) we need to define a neighborhood around P i . Let V i = V ( P i -1 , P i ). Remarking that R ( P i -1 , P i ) = 1, we have FDR( i ) = E ( V i ). Furthermore E ( V i ) = P ( V i = 1) since V i is a binary variable. Thus FDR( i ) provides an unbiased estimation of P ( V i = 1), the probability for gene i to be a false positive. The raw local FDR estimate for gene i is: Assume that H 0 ( i ) and H 0 ( i - 1) are true and E ( ) = m 0 . Therefore this estimate is unbiased with mean 1. Using definition (1), it is direct to obtain: which equals the q -value of gene j . The q -value may thus be viewed as the mean of the raw local FDR of the genes with p -values lower than P j . Under the hypothesis H 0 , it is known that the differences between successive ordered values of independent realizations of the uniform([0,1]) distribution have a Beta distribution with parameters 1 and m 0 (see Johnson [ 9 ] Chap. 26). Therefore the variance of the raw local FDR estimate for non-differentially expressed genes when m 0 is known is equal to /[( m 0 + 1) 2 ( m 0 + 2)] ≈ 1, for m 0 large enough. The variance of estimates (1) under H 1 is generally much smaller than under H 0 (see Figures 1(a) , 2(a) and 3(a) for an illustration). However, one may see on these Figures that ( i , λ) is a very variable estimator. This fact is well known in point process literature, [ 8 ]. Moreover, the interval ] P i -1 , P i [ is not symmetric. If we consider the neighborhood interval around P i defined by t 1 = ( P i -1 + P i )/2, t 2 = ( P i +1 + P i )/2 then we obtain another estimate of the local FDR: Note that (2) is a moving average of order 2 of (1). It is well known that estimates provided by moving average (or kernel estimators) are more stable, see [ 8 ]. This smoothing is generally not enough to obtain usable results and we can consider any kind of smoothing. We propose to estimate FDR( i ) by where f i is a smoothing function of the ( j , λ) for j = 1, m , computed at position P i . The smoothing method must be suited to the properties of the raw FDR: • its variance is low for low p -values corresponding to highly differentially expressed genes • its variance is very high for p -values corresponding to non differentially expressed genes Therefore the window of smoothing should be short for low p -values and large for p -values corresponding high p -values. The lowess smoothing method has a fixed number of neighbor points. Therefore its window size depends of the density of points around the p -value under concern. The density of points is higher for low p -values which in turn implies a shorter window size, which is a good property. However the adaptation of the window size is not sufficient in some cases such as in the Apo-AI example. Moreover the smoothed FDR should be an increasing function of the p -values, a property which is not satisfied by the lowess smoothing. Therefore we prefer to use an ad hoc moving average smoothing using the following algorithm for computing ( i , λ): let 0 < t 1 < t 2 < t 3 be three pre-definite thresholds and m 1 < m 2 < m 3 < m 4 four pre-definite integers. • if max j ≤ i ( j , λ) < t 1 use a moving average of order min (2 i - 1, m 1 ) • if t 1 < max j ≤ i ( j , λ) < t 2 use a moving average of order min (2 i - 1, m 2 ) • if t 2 < max j ≤ i ( j , λ) < t 3 use a moving average of order min (2 i - 1, m 3 ). • if max j ≤ i ( j , λ) > t 3 use a moving average of order min (2 i - 1, m 4 ). We have obtained good empirical results on many data sets with t 1 = 0.01, t 2 = 0.05, t 3 = 0.2, m 1 = 3, m 2 = 5, m 3 = 15 and with the constraint that ( i , λ) is not decreasing. This adaptative moving average method is quite empirical. This topic deserve some more work to build a well assessed smoothing method. This is one of our ongoing research project. Authors' contributions Avner Bar-Hen, Jean-Jacques Daudin and Stephane Robin equally contributed to the statistical work and the redaction task. Julie Aubert coded the R-program and analyzed the three data sets.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC520755.xml
549536
Proton NMR visible mobile lipid signals in sensitive and multidrug-resistant K562 cells are modulated by rafts
Background Most cancer cells are characterized by mobile lipids visible on proton NMR ( 1 H-NMR), these being comprised mainly of methyl and methylene signals from lipid acyl chains. Erythroleukemia K562 cells show narrow signals at 1.3 and 0.9 ppm, corresponding to mobile lipids (methylene and methyl, respectively), which are reduced when K562 cells are multidrug resistant (MDR). While the significance of the mobile lipids is unknown, their subcellular localization is still a matter of debate and may lie in the membrane or the cytoplasm. In this study, we investigate the role of cholesterol in the generation of mobile lipid signals. Results The proportion of esterified cholesterol was found to be higher in K562-sensitive cells than in resistant cells, while the total cholesterol content was identical in both cell lines. Cholesterol extraction in the K562 wild type (K562wt) cell line and its MDR counterpart (K562adr), using methyl-β-cyclodextrin, was accompanied by a rise of mobile lipids in K562wt cells only. The absence of caveolae was checked by searching for the caveolin-1 protein in K562wt and K562adr cells. However, cholesterol was enriched in another membrane microdomain designated as "detergent-insoluble glycosphingomyelin complexes" or rafts. These microdomains were studied after extraction with triton X-100, a mild non-ionic detergent, revealing mobile lipid signals preserved only in the K562wt spectra. Moreover, following perturbation/disruption of these microdomains using sphingomyelinase, mobile lipids increased only in K562wt cells. Conclusion These results suggest that cholesterol and sphingomyelin are involved in mobile lipid generation via microdomains of detergent-insoluble glycosphingomyelin complexes such as rafts. Increasing our knowledge of membrane microdomains in sensitive and resistant cell lines may open up new possibilities in resistance reversion.
Background When studied by NMR proton spectroscopy, most cancer cells are characterized by increased narrow signals at 0.9 and 1.3 ppm corresponding, respectively, to methyl and methylene resonances that belong to lipid acyl chains moving isotropically. This so-called "mobile lipid signal" (or ML signal) has been studied for several decades (for review see[ 1 ]). However, research has failed to elucidate the molecular origin of mobile lipids, their subcellular localization or their physiological significance. Regarding their molecular origin, acyl chains can form part of triglycerides or esterified cholesterol. Phospholipidic acyl chains may also be involved if they are not embedded in membrane lipid bilayers. As far as localization is concerned, two types of subcellular origin are still debated. As early as 1988 [ 2 ], Mountford et al proposed an origin in lipoprotein-like microdomains within the plasma membrane. More recently, some authors [ 3 ] have suggested the occurrence of cytosolic droplets, which appear concomitantly with the ML in stressed cells [ 4 ] or which are associated with necrosis [ 5 ] and apoptosis. However, it has been shown that ML intensity is not always correlated with the number of cytoplasmic droplets [ 6 ]. These studies (op cit.) tend to show that cytosolic lipid droplets generate mobile lipid signals in NMR spectra, without excluding the possibility that such signals could arise independently of the presence of cytosolic lipid bodies. 1 H-NMR spectroscopy shows that ML are decreased in erythroleukemia K562 cells when these become resistant (K562adr)[ 7 ], whereas the compositions of lipids extracted from K562wt (drug sensitive) and K562adr cells are the same, except for sphingomyelin content which is increased in resistant cells [ 8 ]. Thus, although we may explain this by a structural difference in lipid "organization", the significance of such a behaviour is not yet elucidated. In a previous study [ 9 ], we showed that the ML signal is not linked to the externalization of phosphatidylserine (a membrane phospholipid), this being a phenomena that precedes apoptosis. In model membranes, free cholesterol interacts with phospholipids and sphingolipids to influence membrane fluidity [ 10 ]. In vivo , cellular-free cholesterol is located in the plasma membrane [ 11 ], which exhibits increasing structural order as demonstrated in erythrocytes [ 12 ], LM and CHO cells [ 13 , 14 ]. Moreover, cholesterol in model membranes is able to promote microdomains towards an intermediate state called the liquid-ordered phase (Lo), with less fluidity than the gel phase state and more fluidity than the surrounding membrane in the liquid crystalline state [ 15 ]. The aim of this study was to investigate a possible contribution from the cholesterol-containing membrane domain to the ML detected in the NMR proton spectra of wild-type K562 and K562 adriamycin-resistant cells. In a first set of experiments, cell membrane cholesterol was extracted by methyl-β-cyclodextrin (MCD). MCD is an oligosaccharid able to pump cholesterol out of the cell [ 16 ]. Moreover, cholesterol extracted by MCD originates primarily from the plasma membrane [ 17 ]. Cell incubation with MCD was expected to change the membrane microdomains via cholesterol extraction, and consequently produce variations in ML. In a second set of experiments, we used triton-X100 to separate membrane microdomains and then investigated their role in ML generation. In this study, we use the term "Detergent Insoluble Glycosphingolipid Complexes" (DIGCs) or "rafts" to refer to the membrane microdomains isolated by treatment with this mild non-ionic detergent (for review see [ 18 ]). Finally, as these raft microdomains are also enriched in sphingolipids, we used sphingomyelinase to modify their "organization", with the aim of inducing changes in ML signal intensity. Results Cholesterol quantification K562wt and K562adr cells do not differ in their total cholesterol contents. The proportion of cholesterol-ester were found to be lower in the resistant cells than in the wild type cells (Table 1 ). This implies that K562wt cells contain more free cholesterol than K562adr cells. Total cholesterol content after MCD treatment (Table 1 ) shows a 45 % decrease in both K562wt and K562adr cells with respect to the control. Table 1 Quantification of cholesterol in K562 cells. Results are expressed as means ± SD. K562 wt K562adr Total cholesterol (μg/10 6 cells) 3.6 ± 0.5 (n = 5) 3.5 ± 0.3 (n = 4) Cholesterol ester (percentage of total cholesterol) 26.5 ± 1.3 (n = 6) 19.4 ± 2.4 (n = 5) Cholesterol variation with MCD treatment (% of control) 57 ± 13 (n = 5) 54 ± 15 (n = 4) Cell proliferation after incubation with methyl-β-cyclodextrin The cell proliferation results for K562wt and K562adr are presented in Figs. 1 and 2 . The two-hour incubation without FCS had no effect on proliferation in K562wt or K562adr when compared to cells incubated in RPMI+FCS. Treatment with MCD inhibited proliferation in K562wt. This was not the case with K562adr, which grew normally. Figure 1 K562wt cells (up) and K562adr cells (down) were incubated for 2 hours in medium alone (WT RPMI; full line and square, n = 6) supplemented in FCS 10% (dot line and lozenge, WT R10; n = 6), in methyl-β-cyclodextrin 5 mM (dot line and triangle, WT MCD; n = 6) and successively incubated in R10 for 72 hours. Data points are the percentages of the cellular concentration normalized to cellular concentration at T = 0 hours expressed as means, with vertical bars representing standard deviation (SD). Figure 2 K562adr resistance modulation after MCD treatment. K562adr cells were incubated for 2 hours in medium supplemented in SVF 10% (R10-DN; n = 4) or in methyl-β-cyclodextrin 5 mM (CD-DN; n = 4) and seeded in R10 containing daunorubicin for 72 hours. Data points are the percentages of the cellular concentration normalized to cellular concentration at T = 0 hours expressed as means, with vertical bars representing standard deviation (SD). When K562adr cells were resuspended in R10 containing daunorubicin (0.25 nM), they were able to proliferate at this daunorubicin concentration. However, after treatment with MCD, the K562adr proliferation rate was decreased as shown in figure 2 . This shows that MCD treatment reversed the resistance phenotype in K562adr; despite having no effect on cell growth, MCD affected the K562adr cell membrane. Caveolin-1 expression: Western-Blot Caveolin-1 expression was studied in K562 cells by SDS-page and western-blotting. Caveolin-1 is a protein of 21 kD [ 19 ]. As a positive control, we chose a protein extract of NIH-3T3 cells that are known to express caveolin-1. In this control, a band appeared between the 19 and 32 kD size markers. In the present study, however, neither K562wt nor K562adr expressed caveolin-1. 1 H-NMR spectra Incubation with methyl-β-cyclodextrin Figure 3 shows spectra obtained on K562wt and K562adr cells, and resonance ratios are presented in Table 2 . These results show that MCD increases the lipid signals only in K562wt cells. This increase can also be seen in Table 3 , which reports a significant increase of CH 3 /Ct and CH 2 /Ct ratios in MCD-treated K562wt cells compared with non-treated cells. Figure 3 NMR proton spectra obtained at D3 on K562wt (left side) and K562adr (right side) on control cells (top spectra) or treated with 5 mM MCD during 2 hours (bottom spectra). Peak assignment – 1: CH3 from fatty acid chains; 2: CH2 from fatty acid chains; 3: Alanine; 4 : glutamine and glutamate; 5: creatine; 6 N-trimethyl from choline, 7: Inositol; 8 : CH 2 in β position from ester function in fatty acid ester; 9: CH2 in a position of a double bond in fatty acids; 10 : CH 2 in α position from ester function in fatty acid ester. Table 2 NMR peak ratios measured on K562wt and K562adr spectra. Cells were incubated for 2 hours in RPMI, or in RPMI containing methyl-β-cyclodextrin 5 mM (MCD). Results are expressed as means of arbitrary units ± SD multiplied by tenStatistical analysis : *: P < 0.05 when compared with RMPI incubated cells. K562wt K562adr Group RPMI (n = 6) MCD (n = 5) RPMI (n = 4) MCD (n = 4) CH 3 /Ct 50 ± 5 101 ± 22* 34 ± 3 34 ± 2 CH 2 /Ct 247 ± 35 537 ± 13* 142 ± 22 130 ± 18 N + (CH 3 ) 3 /Ct 23 ± 3 22 ± 7 17 ± 3 21 ± 16 Table 3 NMR peak ratios measured on K562wt and K562adr spectra. PBS: incubation in PBS; PFA: incubation in PFA; Triton : incubation in PFA and triton X; Smase: incubation in PFA, triton X-100 and 0.5 units sphingomyelinase. Results are expressed as means of arbitrary units ± SD multiplied by ten. K562wt K562adr Group PBS (n = 4) PFA (n = 7) Triton (n = 7) Smase (n = 4) PBS (n = 4) PFA (n = 7) Triton (n = 8) Smase (n = 5) CH 3 /Ct 198 ± 21 231 ± 12 312 ± 42* 392 ± 105° 161 ± 19 192 ± 22 265 ± 62* 267 ± 42* CH 2 /Ct 541 ± 52 611 ± 78 1060 ± 164* 1696 ± 441° 308 ± 93 386 ± 56 695 ± 141* 565 ± 167* N(CH 3 ) 3 /Ct 132 ± 6 141 ± 18 53 ± 10* 79 ± 17° 114 ± 12 132 ± 36 60 ± 15* 94 ± 39* Statistical analysis : * : P < 0.05 when compared with PBS-incubated cells, °: P < 0.05 when compared with Triton-treated cells, Triton X-100 and sphingomyelinase treatments spectra obtained after treatments with PFA, triton and sphingomyelinase are presented in Figures 4 and 5 for K562wt and K562adr cells, respectively. The ratios of the peak intensities with respect to the creatine peak are reported in Table 3 . When compared to the control, PFA treatment did not modify the cell spectra. After microdomain separation by triton treatment, the CH 2 /Ct ratio was increased for K562wt cells and, to a lesser extent, for K562adr cells. N-trimethyl to creatine peak ratios were decreased in both cell lines. After treatment of these microdomains with sphingomyelinase, the spectra showed an increased CH 2 /Ct ratio compared with the previous spectrum obtained on "DIGCs" in K562wt cells. However, the spectra of K562adr cells treated with Smase remained unchanged compared with spectra obtained on DIGCs from K562adr cells. Figure 4 K562wt 1 H-NMR spectra: effects of Triton X-100 and sphingomyelinase treatments. A: non-treated cells; B: cells fixed with PFA; C: cells fixed with PFA 4% in triton-X100 1%; D: cells fixed with PFA 4% incubated in triton-X100 1% and with 0.5 units sphingomyelinase. For peak assignment, see figure 4. Peaks at 3.6 ppm after SMase treatment arise from enzyme working buffer. Figure 5 K562adr 1 H-NMR spectra: effects of Triton X-100 and sphingomyelinase treatments. A: non-treated cells; B: cells fixed with PFA; C: cells fixed with PFA 4% in triton-X100 1%; D: cells fixed with PFA 4% incubated in triton-X100 1% and with 0.5 units sphingomyelinase. H-NMR spectra: Triton X-100 and sphingomyelinase treatment. For peak assignment, see Fig. 4. Peaks at 3.6 ppm after SMase treatment arise from enzyme working buffer. Discussion We studied the modification of cell spectra using cholesterol and sphingomyelin because: i) cholesterol levels affect membrane fluidity; ii) membrane microdomains, rafts and caveolae are all enriched in these lipids. Western-blot did not detect the presence of caveolin-1 in either K562wt or K562adr. This confirms previous results showing the absence of caveolin-1 in K562wt [ 20 ]. Even if membrane microdomains exist in this cellular type, this implies they represent rafts and not caveolae. Rafts have been described as corresponding to membrane microdomains enriched in cholesterol and sphingomyelin lipids [ 18 ]. In our first set of experiments, we disrupted the rafts using MCD to extract cholesterol from the cells. Before MCD treatment, K562wt and K562adr cells displayed the same cholesterol concentrations, which were found to be in agreement with previous results obtained for other cancer cells [ 21 , 22 ]. In K562wt cells, this cholesterol was preferentially in the form of cholesterol-ester, implying that the absolute concentration of free cholesterol is lower in K562wt cells compared with K562adr cells. 90% of the free cholesterol is contained in plasma membrane [ 11 ]. On the other hand, cholesterol-ester synthesized in the endoplasmic reticulum is generally localized in the cytosol and is able to form cytoplasmic lipid bodies. This cholesterol is less available to MCD extraction. Thus, MCD treatment may affect the free-cholesterol plasma membrane pool [ 17 ] and not the intracellular pool that mainly contains cholesterol-ester in K562wt cells. Consequently, in our study, a higher proportion of membrane cholesterol (corresponding to free cholesterol) was extracted in K562wt cells compared against K562adr cells. This may account for the fact that the proliferation rate for K562wt was more affected than for K562adr. Moreover, the removal of a greater proportion of membrane cholesterol modified the NMR proton spectra of the cells, since the mobile lipid signals only increased in the K562wt cells. Nevertheless, we observe that MCD treatment affected the K562adr cell line. This is demonstrated by the reversion of daunorubicin resistance. In the plasma membrane, K562adr cells express the P-gp glycoprotein responsible for the extrusion of drugs. When K562adr cells were treated with MCD, the resistance to daunorubicin was decreased, showing that the P-gp is unable to extrude daunorubicin from the cells. P-gp has been widely reported to be sensitive to the lipid content of plasma membrane [ 23 ]. The present study shows that MDR is sensitive to the cellular cholesterol content. Indeed, it has been shown that drugs can accumulate in MDR cells grown in the presence of MCD [ 24 ]. This first set of experiments suggests that ML signals are generated during the disruption of membrane microdomains in K562wt cells, because MCD has been shown to extract cholesterol preferentially from membrane microdomains [ 17 ]. Overall, these results tend to show a link might exist between microdomains and the ML signal in K562 cells. In a second set of experiments, we isolated rafts using triton X-100 at 4°C, a mild detergent in which they are insoluble [ 25 ]. We applied this method to K562wt and K562adr cells. NMR spectra obtained on these isolated raft fractions show an increase of the ML signals for K562wt and K562adr cells. In this particular case, this implies that ML could be produced by raft domains. Although it has been shown that raft structures are tightly packed at 4°C, there is no report of their ?organization/structure at 25°C, which is the temperature chosen for the NMR experiments [ 26 ]. At this temperature, microdomain lipids could be in another physical state allowing them to move more isotropically and produce an NMR-visible signal: their solubility in triton at 25°C supports this hypothesis [ 27 ]. In the final set of experiments, we incubated detergent insoluble fractions with sphingomyelinase, which was supposed to metabolise the sphingomyelin included in the rafts. Only the K562wt spectra exhibited an increase in ML signal compared with the raft spectra. Again, this suggests that microdomains are differently organized between K562wt and K562adr. Several hypotheses may be proposed to explain the difference between K562wt and K562adr: i) sphingomyelin concentrations are higher in K562adr; ii) lipid microdomains in K562adr may be organized in such a way that sphingomyelin is inaccessible to sphingomyelinase and iii) there are more microdomains in K562adr, as suggested previously by some authors [ 28 ]. Conclusions These results suggest that, while isolated rafts generate ML signals within the membrane, the presence of rafts seems to maintain a relatively tight organization. When this organized structure is disrupted by MCD or sphingomyelinase, the ML signal may be detected in K562wt cells and the P-gp function can then be modulated in K562adr cells. Taken together, the results suggest that ML signals may be sensitive to modifications in the lipidic organization of the cells and membrane, in which microdomains behave as rafts enriched in cholesterol and sphingomyelin. A better knowledge of these microdomains, as well as the differences in behaviour between sensitive and resistant cells, could open up new therapeutic perspectives for reversing the drug-resistance phenotype. Materials and Methods Chemicals Methyl-β-cyclodextrin, sphingomyelinase, cholesterol, chloroform, isopropanol and paraformaldehyde, as well as deuterium oxide (D 2 O) and all chemicals for western-blot analysis were provided by Sigma-Aldrich (Saint-Quentin Fallavier, France). Triton X-100 was supplied by Merck (Schuchardt, Darmstadt, Germany), and methanol by Acros Organics (Geel, Belgium). Cell culture Human erythroleukemia K562 cells were grown in the culture medium RPMI 1640 (Bio-Whittaker Europe, Verviers, Belgium) supplemented with fœtal calf serum (FCS) (10%) and glutamine (2 mM) (R10 medium) at 37°C and 5 % CO 2 . The K562 resistant variant cells (K562adr) were a gift from F. Calvo, Université Paris 7, and were grown in the same medium with 0.5 μg/ml adriamycin. Cholesterol quantification Lipids were extracted as previously described [ 29 ], and then desiccated and resuspended in isopropanol. Total cholesterol and free cholesterol were assayed using a kit (Cholesterol RTU, Biomérieux Craponne, and Boeringher, France) based on cholesterol esterase. This was followed by cholesterol oxidase reaction and colour development measured on an absorption spectrometer at 500 nm. Concentration was determined using a standard solution of 0.5 mg/ml cholesterol. We first checked that the solvent (isopropanol or chloroform) did not have any effect on colour development. Cholesterol extraction using methyl-β-cyclodextrin Cells were incubated in RPMI, glutamin 2 mM, containing methyl-β-cyclodextrin 5 mM for 2 hours at 37°C, 5% CO 2 . Controls were incubated in RPMI, glutamin 2 mM, with or without FCS 10% for 2 hours. Proliferation After incubation with methyl-β-cyclodextrin, with RPMI+ FCS or with RPMI (the last being used as a control for RPMI+FCS, considering that methyl-β-cyclodextrin is not used in the medium with FCS), cells were collected, washed and resuspended at 150,000 cells/ml in RPMI, glutamin 2 mM, FCS 10%. Cells were seeded in triplicate on 96-well microplates and incubated at 37°C and 5% CO 2 . Proliferation was assessed by counting cells with a cell coulter (Beckman Coulter, ZI model) at 24, 48 and 72 hours after seeding. Caveolin-1 detection: western-blot Cells were collected in late log-phase and washed twice in PBS 150 mM and were resuspended in loading buffer (urea 4 M, lauryl sulphate 3.8%, glycerol 20%, tris base 75 mM pH 6; β-mercaptoethanol 5%; bromophenol blue 5%) for 5 mn at 95°C. They were centrifuged at 12500 g, 4°C for 10 minutes, and 50 μg of extracted protein were deposited per well on an 8% acrylamide gel. Migration was performed in an electrophoresis buffer (TRIS base 125 mM; glycine 250 mM; lauryl-sulphate 0.1%, pH 8.1) for 1 hour at 125 V with a size marker (Kaleidoscope prestained standards Bio-rad). Protein was transferred on a nitrocellulose membrane for 2 hours at 80 V and 4°C in a transfer buffer (tris-base 48 mM; glycine 39 mM, SDS 0.037%, methanol 20%). Non-specific sites were blocked with 1% milk (Regilait, skimmed milk, Saint-Martin-Belle-Roche, France) in TBS 1X (tris-base 20 mM, NaCl 500 mM, and pH 7.5). The nitrocellulose membrane was incubated for 1 hour at room temperature with 1 μg/ml anti-caveolin 1 rabbit antibody (Tebu, France) and washed three times for 10 minutes with TBS before incubation for 1 hour at room temperature with goat anti-rabbit and conjugate secondary antibody (1:3000). After three washes in TBS for 10 minutes, staining was developed in a phosphatase alkaline buffer (NaHCO 3 0.1 M; MgCl 2 , 6H 2 O 1 mM) with tetrazolium nitro-blue 3% in N,N-dimethylformamide 70% and bromo-4-chloro-3-indolylphosphate p-toluidine 1% in N,N-dimethylformamide 70% for 30 minutes at room temperature in the dark. Triton treatment Cells were collected at the end of the log-phase, 4 days after seeding, when ML were at a maximum, and washed twice with 1 ml PBS 150 mM at 4°C. They were resuspended in paraformaldehyde (PFA) 3% in PBS 150 mM for 30 minutes at 4°C and washed twice with 1 ml PBS 150 mM at 4°C to be resuspended in triton X-100 1% in PBS 150 mM at 4°C for 30 minutes. Cells or triton-insoluble fractions were collected and washed for NMR experiments at 25°C. Sphingomyelinase treatment After triton treatment, cells were collected, washed with PBS 150 mM and treated for 35 mn at 37°C with 0.5 units of sphingomyelinase. Cells were washed as described for NMR experiments at 25°C. NMR analysis Cells were washed twice in 1 ml PBS 150 mM, twice in PBS/D 2 O, centrifuged at 250 g and counted. Then, 10 7 cells were resuspended in 400 μl PBS/D 2 O before transfer to a 5-mm Shigemi NMR tube. Experiments were performed without rotation, and the ?analysed cell pellet was maintained in the coil volume in the Shigemi NMR tube. The NMR proton spectra of whole cells were obtained at 25°C on a Unity Inova spectrometer (Varian, France) working at 500 MHz. One-dimensional runs were performed by accumulating 128 transients of 90° pulse with 2 s relaxation time. The signal from the residual water was suppressed by the presaturation technique, by using 0.03 mW irradiation for 2 s. Acquisition time was 1.34 s on 16K data points, corresponding to a spectral width of 6 kHz. The Fourier transform was applied without zero-filling using an exponential window multiplication function corresponding to 1 Hz line broadening. The resonances were integrated after automatic baseline correction. Two-dimensional COSY runs were performed with 2K data points in the F2 direction and 256 data points in the F1 direction. The two-dimensional Fourier transformation was applied after zero filling to 512 data points in the F1 direction with a sine-bell function in both directions. Each run consisted of a one-dimensional acquisition and a two-dimensional COSY spectrum. Peak assignments were based on data from the literature. The peak areas were measured by manual integration with the Bruker WINNMR software using a manual tangential baseline correction for each peak, and the assigned peak areas were normalized to the creatine peak area. The following resonances were integrated: methyl group (CH 3 at 0.9 ppm), methylene group (CH 2 at 1.3 ppm), choline N-trimethyl group (N + (CH 3 ) 3 at 3.2 ppm) and creatine (CH 3 at 3.05 ppm). The values obtained for the different treatments of the two cells lines were compared by ANOVA analysis followed by a Student-Neumann-Keuls test for group-to-group comparison. P < 0.05 was considered as a significant value. NMR proton spectra of PFA or triton- or SMase-treated cells were acquired at 500 MHz and 25°C on a Bruker Advance DMX500 spectrometer (Bruker, Wissembourg, France). The runs consisted of a 1D acquisition with similar parameters as the previous 1D spectra, and the spectrum analysis was performed as previously with Bruker WINNMR software. The possible contribution of lactate to the 1.3 ppm signal was eliminated by analysing 2D Cosy spectra in whole cells, which resolved the resonances of lactate from fatty acid chains. In fixed cells, we calculated the ratio of the double-bond signal (CH = CH at 5.4 ppm) to CH 2 peak area. As both groups belong to fatty acyl chains, this ratio remained constant since lactate did not significantly contribute to the increase of the CH 2 signal [ 1 ]. List of abbreviations NMR: Nuclear Magnetic Resonance; MDR: multi-drug resistance; ML: mobile lipid; K562wt: K562 wild type; K562adr: K562 adriamycin-resistant; MCD: methyl-β-cyclodextrin; DIGC: detergent-insoluble glycosphingolipid complex; FCS: fœtal calf serum; R10: RPMI culture medium with 10% FC; PBS: Phosphate buffer saline; TRIS: Tris(hydroxymethyl)-1,3-propanediol; TBS: TRIS buffer Saline; PFA: paraformaldehyde; COSY: correlated-spectroscopy; CH 3 : methyl; CH 2 : methylene; N(CH 3 ) 3 : N-trimethyl; Ct: creatine Authors' contributions all authors contributed equally to this work
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549536.xml
516032
The contribution of demographic and morbidity factors to self-reported visit frequency of patients: a cross-sectional study of general practice patients in Australia
Background Understanding the factors that affect patients' utilisation of health services is important for health service provision and effective patient management. This study aimed to investigate the specific morbidity and demographic factors related to the frequency with which general practice patients visit a general practitioner/family physician (GP) in Australia. Methods A sub-study was undertaken as part of an ongoing national study of general practice activity in Australia. A cluster sample of 10,755 general practice patients were surveyed through a random sample of 379 general practitioners. The patient reported the number of times he/she had visited a general practitioner in the previous twelve months. The GP recorded all the patient's major health problems, including those managed at the current consultation. Results Patients reported an average of 8.8 visits to a general practitioner per year. After adjusting for other patient demographics and number of health problems, concession health care card holders made on average 2.6 more visits per year to a general practitioner than did non-card holders (p < .001). After adjustment, patients from remote/very remote locations made 2.3 fewer visits per year than patients from locations where services were highly accessible (p < .001). After adjustment for patient demographics, patients with diagnosed anxiety made on average 2.7 more visits per year (p = 0.003), those with diagnosed depression 2.2 more visits than average (p < .0001), and those with back problems 2.4 more visits (p = 0.009) than patients without the respective disorders. Conclusions Anxiety, back pain and depression are associated with greater patient demand for general practice services than other health problems. The effect of sociodemographic factors on patient utilisation of general practice services is complex. Equity of access to general practice services remains an issue for patients from remote areas, while concession health care card holders are attending general practice more frequently than other patients relative to their number of health problems.
Background The frequency of patient visits to general/family practice is affected by a range of factors including patient characteristics, physician/practice factors and broader issues such as access to services. A major contributor to the variance in visit rates to general/family practitioners is the complexity of health problems experienced by the patient. Those patients who attend general practice frequently report poorer health than those who attend less frequently [ 1 , 2 ]. Not all health problems however, have an equal effect on patient visit rates to general practice. Numerous studies have found that for a subset of patients, the presence of psychological health problems is associated with frequent attendance at medical services [ 3 - 6 ]. Higher depression scores and increased levels of health anxiety have been found among patients who attend general practice frequently [ 3 ], and practitioners are more often involved in managing psychosocial issues for frequent attenders than other patients [ 1 ]. Frequent attendance at general practice has also been found to be associated with an increased use of other health services such as out of hours services and "inappropriate attendance" at accident and emergency wards [ 6 - 8 ]. The general practitioner (GP) also influences patients' attendance rates through doctor-initiated visits and some practitioners attract a larger proportion of frequent attenders than others [ 9 ]. Remuneration may also affect GP behaviour. For example, GPs paid a flat rate capitation per patient may be more motivated to manage their own workload by reducing patient return visits [ 10 ]. Positive aspects to increasing general practice attendance includes recalling patients for chronic conditions as part of a program of structured care [ 11 , 12 ]. Patients who attend the GP frequently often receive improved continuity of care [ 9 ], while those who attend general practice infrequently may be receiving less than optimal care, especially where geographical barriers to access result in lower utilisation of general practice services [ 13 , 14 ]. The available evidence on visit frequency to general/family practice comes from a range of settings in several countries. There are differences in general practice between countries that may affect patterns of patient visits and some studies are based on a limited number of practices, which may affect the generalisability of the findings [ 3 , 15 ]. Although the association between frequent medical visits and increased psychological health problems has been well-researched, the effect of other specific chronic conditions on general practice visit rates is less clear. In Australia there are few studies that examine the relationship between patient morbidity, social and demographic factors and the frequency of visits to general practice. We therefore investigated patient sociodemographic and morbidity factors associated with patient self-reported visit frequency among general practice patients in Australia, as part of an ongoing national study of general practice activity. Methods The study reported here is part of the Bettering the Evaluation and Care of Health (BEACH) program, a national study of general practice activity in Australia. The BEACH method has been described in detail elsewhere [ 16 ]. In summary, BEACH is a continuous cross-sectional study, which commenced in April 1998. A random sample of approximately 1,000 GPs is recruited throughout each year in a rolling sample. The GPs are selected from a sampling frame of all Australian GPs who claimed 375 or more general practice (A1) items from the Medicare Benefits Schedule in the previous quarter. Each GP provides details for 100 consecutive patient encounters. Design This paper is based on a sub-sample of GPs who participated in BEACH between August 2001 and March 2002. For a subset of 30 out of the 100 encounters, the GP recorded, in addition to the problems managed at the encounter, any other major health problems of the patient that had not been managed at the current encounter. GPs were instructed to include chronic illnesses requiring ongoing care, past problems that affect present and future care and social problems that influence health. As many as 12 extra problems could be recorded per patient and when added to the maximum of four problems managed at the encounter this allowed up to 16 health problems to be recorded per patient. The GP recorded problems in free text, which were secondarily classified according to the International Classification of Primary Care (ICPC-2) using an extended vocabulary of terms [ 17 , 18 ]. ICPC-2 includes components for diagnoses and symptoms, and process codes (e.g. "Test results", "Prescription renewal"). Process codes were excluded from counts of health problems, unless they clearly identified the nature of the underlying condition. Synonymous or related problems were grouped according to standard ICPC-2 categories to ensure that problem categories included all patients with the disorder [ 16 ]. Morbidity variables Mental health (particularly depression), asthma, diabetes, osteoarthritis and cardiovascular disease are five National Health Priority Areas that are major contributors to mortality, morbidity and health service costs in Australia [ 19 ]. They are also chronic health problems commonly managed in general practice [ 20 ]. In addition hypertension, back complaint and oesophageal disease are among the most common problems managed in general practice [ 20 ]. We therefore focussed on these health areas of particular relevance to general practice when investigating the association between morbidity and frequency of patient visits to a GP. Number of annual visits to general practice The GP asked each patient in the sub-sample to recall how many times he/she had visited a GP in the previous 12 months. Demographic variables Commonwealth concession health care cards are available to people on limited incomes and entitle the holder to health services at greatly reduced cost. In 1998, around one third of Australians aged 15 years and over held a concession health care card [ 21 ]. Concession health care card status was included in the analysis as a marker of socioeconomic status [ 22 ]. The Remote Index of Australia (ARIA) was used to classify patients according to geographical remoteness and accessibility of services [ 23 ]. Patients were classified by patient residential postcode into three broad categories: those from areas that were "highly accessible" to services, those from areas that were "accessible/moderately accessible" and those from areas "remote/very remote" from services. These three categories were used to investigate the association between geographical barriers to access and the frequency of patient visits to a GP. Statistical analysis The patient sample was a single stage cluster design with the GP as the primary sampling unit. Observations recorded by the same GP were therefore not independent and the statistical analysis adjusted for the correlation between patients within each cluster. We used procedures in SAS software V8.2 that adjust the standard error for the intra-cluster correlation [ 24 ] and all reported p-values and confidence intervals in both the descriptive and multivariable analyses include this adjustment. Multiple regression Number of annual visits to a GP was the main outcome of interest. After checking the linear relationship between number of visits and other ordinal variables (age and number of health problems), multiple regression was performed in two stages to identify predictors of visit frequency. 1) Demographic predictors We performed multiple regression, with self-reported annual GP visits as an ordinal outcome to identify the independent demographic predictors of the frequency of visits to a GP after controlling for morbidity. We fitted the morbidity covariate as the total number of major health problems recorded for the patient. 2) Morbidity predictors We fitted separate regression models to estimate the effect of specific morbidity on visit frequency, after adjusting for patient age, sex, other significant demographic predictors and the number of other health problems of the patient. The morbidities of interest were chronic problems most commonly managed in general practice, specifically depression, anxiety, back problems, diabetes, hypertension, asthma, ischaemic heart disease, and osteoarthritis [ 16 ]. We interpreted the partial regression co-efficient of morbidity in each model as the mean difference in annual GP visits between patients who had the problem versus those who did not, when other factors were kept constant. A large number of models was fitted, therefore an alpha of 0.01 was used as the test of significance to reduce type I error. Results Three hundred and seventy nine GPs participated in the sub-study. Six hundred and eighteen patients (5.4%) did not answer the question on visit frequency, giving a final sample of 10,755 respondents. Patient respondents and non-respondents were not significantly different in mean age, sex distribution or mean number of health problems recorded. While all states of Australia were well represented in the GP sample, the sample had a smaller proportion of GPs aged less than 35 years and was somewhat more urban than the population of Australian GPs (Table 1 ). Table 1 Comparison of general practitioner (GP) characteristics for the sample and the population of Australian general practitioners in 2001. GP sample Australian population of GPs (a) N 379 17,534 Male 61.4% 67.5% Age group <35 years 8.2% 12.1% 35–44 29.3% 27.1% 45–54 36.7% 32.0% 55 + 25.9% 28.8% Urban/Metropolitan practice 76.0% 72.6% Solo practice 16.1% Not available 5 plus GPs in practice 45.2% Not available (a) defined as practitioners who claimed the equivalent of 1,500 general practice (A1) Medicare items of service in 2001 (Medicare Benefits Schedule unpublished data, Australian Department of Health and Ageing) Patients aged less than 25 years were under-represented in the sample relative to all patients who had visited a GP in Australia at least once in 2001, while patients aged 65 years and older were over-represented (Table 2 ). Table 2 Characteristics of the patient sample versus the general practice patient population in Australia in 2001. Sample n (%) Australian population of general practice patients (a) % Female patient 6,404 (59.5) 53.4 Age group <15 years 1,360 (12.7) 19.9 15–24 1,005 (9.4) 13.1 25–44 2,875 (26.9) 29.7 45–64 2,677 (25.0) 23.8 65–74 1,274 (11.9) 7.6 75+ 1,514 (14.1) 5.9 Address in "highly accessible" location 8,702 (83.3) Not available Speaks language other than English at home. 890 (8.3) Not available Holds concession health care card 4,448 (41.4) Not available (a) Defined as persons who claimed at least one general practice (A1) Medicare item in 2001 (Medicare Benefits Schedule unpublished data, Australian Department of Health and Ageing) The majority of patients in the sample (83.3%) were from locations where services were "highly accessible", and 8.3% spoke a language other than English at home. The mean patient self-reported visit rate for the sample was 8.8 (95% CI: 8.3–9.2) visits to a GP per year. On average the GPs recorded 2.9 morbidities per patient, and half (51.1%) of the recorded problems were being managed by the GP at the current consultation (results not tabled). There was no difference in self-reported visit frequency between the sexes (p = 0.11) (results not tabled). The mean visit rate for children aged less than 5 was 5.8 visits per year, which fell to 4.3 for children 10–14 years of age, then increased linearly after age 15 years (p < .0001). There was a simple linear increase in frequency of visits with increasing number of recorded health problems (p < .0001). Therefore age was fitted as a categorical variable and number of problems as a numeric variable in the multiple regression models. After adjusting for age and sex, there was no difference in visit frequency according to language spoken at home (Table 3 ). After adjusting for age, sex, remoteness and total number of health problems, concession health care card holders made on average 2.6 more visits per year to a GP than non-card holders (p < .001, Table 4 ). After adjusting for age, sex, concession health care card status and total number of health problems, patients from remote/very remote locations made on average 2.3 fewer visits per year to a GP than patients from locations where services were highly accessible (p < .001). Table 3 Mean number of recorded problems (crude rate) and differences in mean annual visits to a GP (crude rates and adjusted for age and sex) by patient demographics Annual visits Crude Adjusted Patient characteristics (n) Mean problems Mean visits Mean visits p-value Highly accessible (8,702) 2.9 9.1 9.0 Moderately accessible (1,599) 2.7 7.7 7.9 0.05 Remote/very remote (145) 2.5 5.8 6.5 <.0001 Concession health care card holder (4,448) 3.5 11.4 10.7 <.0001 Non health care card holder (6,307) 2.4 6.8 7.5 Speaks language other than English at home (890) 2.8 9.0 9.1 0.369 Speaks English at home (9,865) 2.9 8.7 8.8 Male (4,285) 2.7 8.5 N/A N/A Female (6,404) 2.9 8.9 N/A Table 4 Multiple regression model of sociodemographic predictors of visit rates and number of recorded health problems Partial coefficient (a) (95%CI) p-value Patient sex (ref: male) 0.0 (-0.6;0.5) 0.88 Patient age (ref: <1 year) 1–4 years 2.2 (1.5;2.9) <.001 5–14 -0.2 (-0.8;0.4) 0.46 15–24 0.5 (-0.1;1.2) 0.08 25–44 2.2 (1.5;3.0) <.001 45–64 1.9 (1.1;2.8) <.001 65–74 2.0 (0.9;3.1) <.001 75+ 2.7 (1.4;4.1) <.001 Accessibility (ref: Highly accessible) Moderately accessible -1.0 (-2.1;0.2) 0.12 Remote/very remote -2.3 (-3.1;-1.5) <.001 Concession health care card holder 2.6 (1.8;3.3) <.001 Each extra health problem 1.6 (1.3;1.8) <.001 (a) Interpreted as the change in mean number of annual visits after adjusting for all other variables in the model. Morbidity and visit rates The mean annual visit rates for patients with common chronic disorders were well above average (Table 5 ). After adjusting for demographic predictors and total number of health problems, patients with anxiety made on average 2.7 more visits to a GP per year (p = 0.003), those with depression 2.2 more visits (p < .0001) and those with back problems 2.4 more visits (p = 0.009). Ischaemic heart disease, diabetes, asthma, oesophageal disease, osteoarthritis or hypertension did not affect patient visit rate beyond that expected from the patients' demographics and overall number of health problems. Table 5 The effect of specific morbidity on annual visits to a GP (Crude rates, rates adjusted for age, sex, concession health care card, remoteness and number of other problems) Crude rates Effect on visit rate (adjusted) Health problem (a) (n) Mean annual visits No. other problems Morbidity coefficient (b) (95%CI) p-value Anxiety (381) 14.1 3.4 2.7 (0.9;4.5) 0.003 † Depression (1,100) 13.1 3.2 2.2 (1.4;3.1) <.0001 † Back complaint (578) 12.8 2.9 2.4 (0.6;4.2) 0.009 † Diabetes (706) 13.7 3.8 0.9 (-0.1;1.9) 0.08 Ischaemic heart disease (585) 14.3 4.3 0.6 (-0.7;1.9) 0.38 Asthma (876) 10.5 2.8 0.3 (-0.4;1.0) 0.40 Oesophageal disease (636) 12.9 4.1 -0.2 (-1.2;0.8) 0.65 Osteoarthritis (859) 13.0 4.1 -0.4 (-1.3;0.5) 0.39 Hypertension (2,094) 11.8 3.5 -0.4 (-1.2;0.4) 0.30 (a) Each line represents a separate regression model (adjusted for age, sex, concession health care card holder, remoteness and number of problems). (b) Interpreted as the mean difference in number of annual visits between patients with the specific problem versus those without after adjusting for age, sex, concession health care card holder, remoteness and overall number of health problems. † significant p < .01 Discussion As would be expected the number of health problems recorded for the patient was a major predictor of frequency of attendance at general practice. However this study demonstrated that specific chronic disorders had differential effects on visit frequency. Anxiety, back complaints, and depression had the greatest effects on increasing patient annual visits to a GP. The relationship between psychological health problems and increased medical visits has been demonstrated across many settings [ 1 - 3 , 5 ]. The current study further demonstrates the generalisability of this relationship. However the interpretation of this finding in the Australian context is less clear. Patients with depression may be visiting GPs more frequently because GPs fail to recognise and treat the root causes of the patient's poor mental health [ 5 ]. Alternatively, above average visits for depression and anxiety may be appropriate if the practitioner has initiated the return visit for ongoing management. GPs in Australia have been targeted to improve the detection and management of depression among general practice patients [ 25 ], so higher visit rates for some patients with depression could reflect a more active intervention strategy by their GPs. Other research has reported a link between unexplained back pain and frequent attendance at medical services [ 26 ]. In the current study back complaints included problems described as symptomatic back pain as well as diagnosed disc and nerve problems [ 18 , 20 ]. It is unclear the degree to which the frequent visits of patients with back complaints may be due to somatisation [ 27 ] and how much is explained by the need to manage acute/intractable pain from organic causes. Other research indicates that patients with diabetes make more general practice visits than average [ 28 ] and structured care programs might be expected to result in above average visit frequency for patients with diabetes [ 12 , 29 ]. The current analysis controlled for other health and demographic factors and found that the higher visit rates of patients with diabetes was more related to the high number of health problems reported by the patient rather than to extra visits related to diabetes per se. Regression models revealed similar explanations for the visit rates of patients with asthma, ischaemic heart disease, osteoarthritis, oesophageal disease and hypertension: visit rates were mostly explained by the overall level of health problems experienced by the patient rather than to any increased demand for health care related to the specific morbidity. As expected, patients from remote areas reported visiting general practice less frequently relative to their health needs, confirming that geographic access to services remains a significant barrier to meeting the health care needs of patients from remote locations [ 13 , 14 ]. Concession health care card holders have more chronic and psychological health problems managed in general practice relative to the rest of the Australian general practice population [ 22 ]. In this study concession health care card holders reported a higher than average number of visits even after accounting for their number of health problems. This relatively high visit rate coupled with high levels of ill-health places concession health care card holders among the most frequent attenders at general practice. Shorter consultation length is associated with lower socioeconomic status in Australia [ 30 ]. Perhaps concession health care card holders are partly compensating for less time spent with the GP at each visit by visiting more often. This study had the advantage of randomly sampling clusters of patients across a wide range of practices in all states and regions of Australia. A further advantage was having a qualified medical practitioner to record patient morbidity and the use of the medical record to help validate patient recall of GP visits. The methodology of this study was limited by using patient self-report of visit frequency and by sampling patients at the point of consultation. The average number of GP visits among Australians who visit a general practitioner at least once in a 12 month period is known to be six visits per year, increasing with increasing age (Australian Health Insurance Commission, unpublished data), while the self-reported visit rate in the current study was 8.8 visits per year. Because the patients were sampled at the consultation, older patients who attend general practice more frequently had a greater chance of being sampled and this was reflected in the older age distribution of the patient sample. Even allowing for this selection bias, the mean annual self-reported visits was somewhat higher than expected given the sample's age structure, indicating some overestimation by patients in their recall of the number of GP visits. One Australian study that validated patient self-report against actual Medicare benefit claims found that patients who overestimate recent use of medical services are "telescoping" real but less recent events into the nominated time period and are in fact higher users of health services over the long term [ 31 ]. In the current study the number of reported visits increased with age and number of morbidities as expected. Therefore even if somewhat overestimated, self-reported visit frequency had sufficient face validity to allow a comparative analysis of the effect of morbidity and demographic factors on the relative frequency of visits to a GP. Conclusion This study has demonstrated that number of visits to a GP by patients in Australia is largely explained by the number of health problems the patient is experiencing. However psychological health problems and sociodemographic factors have differential effects on visit rates. GPs need to be alert to the particular psychological and social needs of patients who are attending frequently with multiple health problems. Equity of access, in terms of equal utilisation of services, continues to be an issue for patients living in remote locations. Competing interests None declared. Authors contributions SK designed and performed the statistical analysis, researched and wrote the main draft. HB directed the study, participated in the design and assisted in writing and reviewing the final article. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516032.xml
539238
Walkability and self-rated health in primary care patients
Background The objective of this study was to investigate the relationship between perceived walkability and overall self-rated health among patients who use community-based clinics. Methods A cross-sectional survey was distributed to a convenience sample in three community clinics. Forms were completed by 793 clinic patients. Multiple logistic regression analysis was to control for the effects of demographic variables and lifestyles. Results Perceiving the availability of places to walk was related to better self-rated health. The most important places were work (OR = 3.2), community center (OR = 3.12), park (OR = 2.45) and day care (OR = 2.05). Respondents who said they had zero (OR = .27) or one (OR = .49) place to walk were significantly less healthy than persons who said they had five or more places to walk. Conclusion Persons who perceived that they had no place to walk were significantly less healthy than persons who thought they had at least one place to walk (OR = .39). Support for walkable neighborhoods and education of patients about options for walking may be in the best interests of community medicine patients.
Background Much recent interest and research has been directed at the relationship between one's health and where one lives [ 1 ]. A number of measures of health (including all-cause mortality [ 2 , 3 ], low infant birth weight [ 4 ], unintentional childhood injuries [ 5 ], hospitalization for asthma [ 6 ], and the risk for certain arboviral diseases [ 7 , 8 ] have been associated with neighborhood effects that are independent of demographic health markers such as age, gender, race, and poverty. The neighborhood environment appears to exert health effects independent of or in addition to the health behaviors of the neighborhoods inhabitants [ 9 ]. Neighborhoods where walking is convenient might encourage their inhabitants to exercise. Indeed, having convenient places to walk in the neighborhood has been related to the proportion of persons in the neighborhood who met current activity recommendations [ 10 ] and with a decreased prevalence of overweight [ 11 ]. Neighborhoods with convenient places to walk are characterized by good "walkability." Walking is now recommended for the prevention and treatment of many common diseases, such as hypertension, diabetes, coronary heart disease, osteoporosis, colon cancer, and obesity [ 12 ]. Therefore, one would postulate that neighborhoods characterized by good walkability would be inhabited by healthier residents. The hypothesis that perceived walkability is directly related to an individual's overall self-rated health has not previously been investigated. The purpose of this project was to test that hypothesis in patients attending community-based clinics. Self-rated health has been shown to accurately predict overall health as measured by other more traditional measures of health [ 13 ]. Methods A cross-sectional survey was used to test the hypothesis that patients with good self-rated health perceived the neighborhoods in which they live as having good walkability, and that this effect was independent of demographic characteristics and lifestyle habits that would predict good health. The sample was drawn from three clinics that primarily serve low-income populations. Clients, parents (if the patient was a child) and accompanying visitors were asked to complete surveys and drop them into a box. Participation was voluntary. A total of 1471 surveys were distributed with an overall return of 825 (56.1%) surveys of which 793 (54%) met eligibility parameters. Pregnant women and persons under age 18 were excluded. Assuming 80% power, p < .05, 20% poor health among the persons inhabiting neighborhoods with good perceived walkability and 30% poor health among persons inhabiting neighborhoods with poor perceived walkability, 626 cases were needed to test the hypothesis. Completed forms were received from a total of 793 persons. Return rates varied by clinic. Clinic 1 is a university-based family medicine clinic providing a full range of primary care services to cross-generational clients. It is staffed by family medicine physicians and residents. Census was approximately 85 clients daily, of which, less than 5 % were non-English speaking. Clinic personnel distributed 500 survey forms over an eight week period with an 80.8% return rate. Clinic 2 serves women and children, providing obstetrical, well care (including immunizations), and acute care services to a targeted high-risk, low socioeconomic sub-population. It is staffed by pediatric and OB-GYN physicians and residents. Approximately 30% of the clinic clients do not speak English. A total of 471 surveys were distributed over a period of 18 weeks with a return of 37.4%. Both the large number of obstetrical patients (ineligible for survey) and percentage of non-English speaking clients contributed to the low return rate. Clinic 3 provides primary care services to a population of indigent adults meeting residential and income screening requirements. It is staffed by internal medicine physicians and residents. A total of 500 surveys were distributed in this clinic over a ten week period with a return of 42.6%. The dependant variable for the study was self-rated health. Subjects were asked whether in general they would say their health was excellent, very good, good, fair or poor. Excellent, very good, and good responses were combined to form a category called 'good health' while fair and poor comprised 'poor health'. Leyden's scale of walkability was modified for a US population for this study[ 11 ] by dropping the terms newsagent, chemist, and crèche. The question reads: "A lot of people are very dependent on a car these days to get where they want to go. If you or another family member wanted to which of the following could you walk to without too much trouble. Circle all you could walk to without too much trouble." Possible answers were a local corner shop, a church, a park, a local school, a community center or recreation center, a day care center, a drug store, a bar or pub, the place that I work, or "none of the above. It is really hard to go anywhere without a car." Both demographic characteristics and lifestyle variables were used to adjust the associations between perceived walkability and self-rated health. Lifestyle variables were: numbers of fruits and vegetables eaten per day (zero, one, two, three, four, five or more), smoking status (not a smoker, smokes one-20 cigarettes per day, smokes more than 20 per day), days of physical activity per week that involve at least 20 minutes of exercise (zero, one, two, three, four, five or more), and obese (yes vs no). Obese was defined as body mass index (BMI) >30. BMI was computed from self-reported height and weight. Demographic variables were race/ethnicity (Hispanic, non-Hispanic white, non-Hispanic black, non-Hispanic Asian, non-Hispanic other), gender, age category, marital status (married vs. not), and highest level of education achieved. Chi square tests were performed to test for any unadjusted associations between self-rated health and each categorical independent variable. Multivariate logistic regression modeling was employed to determine if associations between perceived walkability and self-rated health remained significant after adjustment for demographic and lifestyle variables. Separate logistic regression models were run for each variable. Statistical analysis was performed using EpiInfo 3.2.2. Results The question about self-rated overall health was answered by 793 respondents. Of these, 67 percent were classified as healthy because they said their health was excellent, very good, or good. The remainder was classified as having poor health because they said their health was fair or poor. The typical respondent was non-Hispanic white, female, had at least a high school education. The sample was evenly spread across age groups. About half were married and about half were not married (Table 1 ). Table 1 Association between good self-rated health and demographic variables (chi-square) Variable Overall Percent Percent Healthy Pct Not Healthy P Age 0.0717 18–25 17.0 77.6 22.4 26–35 21.6 66.5 33.5 36–45 17.0 64.2 35.8 46–55 19.1 64.0 36.0 56–65 12.3 69.1 30.9 Over 65 12.8 60.4 39.6 Race/ethnicity 0.0303 Hispanic 15.5 58.5 41.5 NH* asian .5 75.0 25.0 NH* black 7.8 54.8 45.2 NH* white 74.9 69.9 30.1 NH* other 1.3 70.0 30.0 Gender 0.4852 Male 18.7 64.2 35.8 Female 81.3 67.6 32.4 Education 0.0011 Less than high school 9.0 54.9 45.2 High school 37.8 62.0 38.0 Some College 37.7 69.6 30.4 College degree 15.5 79.7 20.3 Married <0.0001 Yes 49.8 74.7 25.3 No 50.2 59.3 40.7 *NH = Non-Hispanic Unadjusted relationships between demographic variables and self-rated health are shown in Table 1 . Statistical significance was determined using chi-square tests. As expected, the percent healthy declined with age. Non-Hispanic white respondents were more likely to report good self-rated health than Hispanics or non-Hispanic blacks. The percent healthy increased with level of education and married persons were more likely to be healthy. No significant difference was seen between male and female subjects in self-rated health. The most common number of days of exercise per week was zero (34.9 percent). Most people said they ate at least one fruit or vegetable each day, though only less than half had three or more. The percent obese was 43.6. Over 70 percent were non-smokers (see Table 2 ). Table 2 Association between good self-rated health and lifestylevariables (chi-square) Variable Overall Percent Percent Healthy Pct Not Healthy p Days of exercise per week 0.0104 Zero 34.9 62.9 37.1 One 8.0 52.4 47.6 Two 16.6 67.9 32.1 Three 18.8 72.3 27.7 Four 5.8 80.4 19.6 Five or more 16.0 71.4 28.6 Fruit and vegetables 0.1522 Zero 15.2 62.5 37.5 One 19.6 70.8 29.2 Two 27.6 68.7 31.3 Three 20.1 66.5 33.5 Four 9.9 56.4 43.6 Five or more 7.6 75.0 25.0 Obese <0.0001 Yes 43.6 58.1 41.9 No 56.4 73.8 26.2 Smoking Status 0.1496 Non-smoker 72.0 68.9 31.1 One – 20 per day 26.1 61.8 38.2 Over 20 per day 1.9 60.0 40.0 Self-rated health was related to lifestyle variables (Table 2 ). Persons who exercised more times per week were more likely to report good self-rated health. Obese persons were less likely to be healthy. Smoking and consumption of fruit and vegetables were not significantly related to self-rated health, though more smokers reported poor health. The association between perceiving particular places to walk and self-rated overall health is shown in Table 3 . Respondents typically said they could not walk to the places included in the walkability scale. The highest percents were corner store (43.1), park (44.0), and school (40.9). When the association between particular places to walk and self-rated health was adjusted for the demographic and lifestyle variables, eight were significant: workplace (OR = 3.2, p = .0011), church (OR = 1.76, p = .0031), community center (OR = 3.12, P = .0019), corner store (OR = 1.71, p = .0032), day care (OR = 2.05, p = .0216), drug store (OR = 1.88, p = .0055), park (OR = 2.45, p < 0.0001), and school (OR = 1.86, p = .0008). Table 3 Association between Good Self-Rated Health and Places to Walk (N = 793) Variable Overall Percent Percent Healthy Pct Not Healthy p Ajusted Odds Ratio (Confidence Interval)* p 67.0 33.0 Pub 0.5043 Yes 10.1 72.5 27.5 1.38 (.77–2.48) 0.2750 No 83.9 66.2 33.8 reference Workplace 0.0146 Yes 8.7 82.6 17.4 3.20 (1.5955–6.4271) 0.0011 No 85.1 65.3 34.7 reference Church 0.0010 Yes 32.2 75.7 24.3 1.76 (1.21–2.57) 0.0031 No 61.8 62.2 37.8 reference Community Center 0.0030 Yes 9.6 84.2 15.8 3.12 (1.52–6.38) 0.0019 No 84.4 64.9 35.1 reference Corner Shop 0.0002 Yes 43.1 74.6 25.4 1.71 (1.20–2.44) 0.0032 No 50.8 60.3 39.7 reference Day Care 0.0176 Yes 11.3 80.0 20.0 2.05 (1.11–3.78) 0.0216 No 82.6 65.0 35.0 reference Drug Store 0.0037 Yes 21.7 77.3 22.7 1.88 (1.20–2.93) 0.0055 No 72.3 63.7 36.3 reference Park <0.0001 Yes 44.0 76.8 23.2 2.45 (1.69–3.55) <0.0001 No 49.9 58.1 41.9 reference School <0.0001 Yes 40.9 76.2 23.8 1.86 (1.29–2.67) 0.0008 No 53.1 59.6 40.4 reference *adjusted for marital status, age, gender, obesity, smoking status, days of exercise per week, educational level, race/ethnicity using multiple logistic regression (N = 775) Table 4 shows the association between the number of places to walk and self-rated health. Nearly 30 percent of respondents reported that they had no places to which they might walk. Using five or more places to walk as the reference category and adjusting for demographic variables and lifestyle, persons with no places to walk had lower odds of being in good health (OR = .27, p < 0.0001). Persons who had only one place to walk also were not as healthy (OR = .49, p = .0257). The other levels of walkability were not significantly different from five or more places. Table 4 Association between Good Self-Rated Health and Number of Places to Walk Variable Overall Percent Percent Healthy Pct Not Healthy p Ajusted Odds Ratio (Confidence Interval)* p 67.0 33.0 Walk category <0.0001 None 27.9 51.1 48.9 .27 (.16–.47) <0.0001 One 14.6 65.5 34.5 .49 (.26–.92) 0.0257 Two 11.6 75.0 25.0 .76 (.37–1.54) 0.4414 Three 12.5 72.7 27.3 .66 (.34–1.30) 0.2281 Four 10.2 72.8 27.2 .72 (.35–1.45) 0.3505 Five + 17.2 80.1 19.9 reference Missing 6.1 68.8 31.3 .59 (.26–1.30) 0.1898 No place to walk <0.0001 Yes 27.9 51.1 48.9 .39 (.27–.57) <0.0001 No 72.1 73.1 26.9 reference *adjusted for marital status, age, gender, obesity, smoking status, days of exercise per week, educational level, race/ethnicity using multiple logistic regression (N = 775) Also shown in Table 4 is the result of a multiple logistic regression analysis in which perceived walkability is scored simply as none versus more than none. In this model, the adjusted odds ratio for perceived walkability was .39 (p < 0.0001) Discussion Studies of physical activity in public health may be classified according to the intensity of physical activity on which they focus (the "meet recommended guidelines" group versus the "any activity will do" group). Meeting recommended physical activity guidelines is related to better self-rated health [ 14 ]. Certainly, more fitness is better, but arguments can be made against setting high standards for physical activity in the population. Foremost among these is the observation that effective interventions to promote high levels of fitness in the general population may not be available. Harlan reported that a brief intervention in primary care was not effective [ 15 ], as did Yeazel [ 16 ]. In contrast, Long's study supported primary care based promotion of physical activity [ 17 ] and Eakin also offered support [ 18 ]. Overall, most studies of the elderly do not show that exercise reduces disability [ 19 ]. If promotion of intensive physical activity is problematic, then more moderate activity might be a more reasonable goal. Simple walking, for example, reduces the cost of medical care for the elderly [ 20 ] and interventions have been developed to increase walking [ 21 ]. Besides, sedentary adults may not be able to accurately recall the intensity of physical activity, casting doubt on the accuracy with which it can be measured [ 22 ]. In addition to valuing leisure-time walking, researchers and clinicians should take into account the benefits of work related activity and housework [ 23 ]. So called "lifestyle" interventions are more cost-effective than supervised center-based exercise [ 24 ]. In short, encouraging moderate physical activity is important. Successfully doing so in the most sedentary and unfit portion of the population (the bottom 20 or 25 percent) would generate large gains in population health [ 25 ]. One approach to encouraging moderate physical activity is to increase walkability. The convenience of places to exercise is widely recognized to be important; adults generally support local policies that increase the availability of places to exercise [ 26 - 28 ]. Walkability in Georgia was investigated by Powell et al, using the state-wide Behavioral Risk Factor Surveillance System (BRFSS) [ 10 ]. Over 90 percent of Georgians reported that they knew of a place where they felt safe walking. The most common place was the respondent's neighborhood (32%). About half said that they could get to their walking place in less than ten minutes. A direct relationship was found between convenience of the walking place and the proportion of respondents meeting current activity recommendations. Better health should result from more walking, but this relationship was not tested in the Georgia study. Our purpose was to investigate the effect of perceived walkability on the health of patients attending community based clinics. The results should not be interpreted as being relevant to all persons in the community or even to all persons who live in neighborhoods they perceive to be unwalkable, but only to patients attending clinics that serve a disproportionate share of disadvantaged persons. This is important to public health because the success of strategies designed to improve in overall community health will have to focus on the disadvantaged people who bear the greatest burden of poor health. We believe that more studies of disadvantaged clinic populations such as this one are needed in the public health literature. Conclusions The study reported here of low income primary care patients confirms the relationship between perceptions of convenient walking locations and self-rated health. Health status is measured by just one question, a practice which has become increasingly common in the public health literature[ 29 , 30 ]. However, since the study employs a cross-sectional design the relationships between perceived walkability and good health might be due to an omitted third variable, such as a tendency to look for excuses for inactivity among persons of poor health. Or a tendency for negative persons to report no places to walk may have influenced the results of this study. In addition, since a large proportion of our respondents were female, the results might be less generalizable to male populations. Another limitation of the study is the subjective nature of the walkability measure. Perceptions of walkability may not be accurate and thus objective assessment of walkability might have led to a different conclusion. However, since our purpose was to investigate the relationship between perceptions of walkability and self-rated health, the results are valid for this sample. If community medicine patients are incorrect in their perceptions of neighborhood walkability, then public health education campaigns and personal health education in clinics can inform them about options for walking in their neighborhoods and how they might overcome any perceived barriers. Our results also may have been influenced by the context from which subjects were drawn. Amarillo is located in the Panhandle of Texas. Most of the year the climate is mild, but during the summer season temperatures can average over 90 degrees. In some neighborhoods, crime rates are above the national average and walking after dark might be worrisome. Furthermore, cities in West Texas are designed for automobile traffic, to the detriment of pedestrians and bicyclists. Furthermore, the vast open spaces in the region make it impossible to reach many locations by foot. Accordingly the, culture of the area has not evolved to be supportive of walking. Therefore, our results may not be generalizable to communities where walking has historical been a visible aspect of the culture. Despite these limitations, the results of this analysis are intriguing and warrant further investigation using a prospective study design. The evidence presented suggests that support for walkable neighborhoods and health education about options for walking in their neighborhoods may be in the best interests of community medicine patients. Competing interests The author(s) declare that they have no competing interests. Authors' contributions JER planned the project, analyzed the data and wrote the first draft of the manuscript. JRP contributed comments and revisions to the manuscript. AD organized data collection and wrote a portion of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539238.xml
516797
Adaptation and Immunity
The ongoing battle between hosts and pathogens has long been of interest to evolutionary biologists.
The ongoing battle between hosts and pathogens has long been of interest to evolutionary biologists. Because hosts and pathogens act as environments for each other, their intertwined struggle for existence is both continual and rapid. At the molecular level, this cycle of environmental change and evolutionary response means that mutations are continually being tried out by natural selection. It is therefore little wonder that the host and pathogen genes that control infection and immunity frequently show high levels of genetic diversity and present some of the best examples of positive selection (adaptive evolution) reported to date ( Yang and Bielawski 2000 ). In particular, rates of nonsynonymous substitution per site (resulting in an amino acid change; d N ) often greatly exceed those of synonymous substitution per site (silent change; d S ), as expected if most mutations are fixed because they increase fitness ( Figure 1 ). Figure 1 Measuring Selection Pressures by Comparing the Ratio of Nonsynonymous to Synonymous Substitutions Per Site (A) Classification of substitutions. Nonsynonymous substitutions (red) are those that change the amino acid sequence of the protein encoded by the gene, while the degeneracy of the genetic code ensures that synonymous substitutions (yellow) result in the same amino acid sequence. (B) Calculation of d N /d S . By assuming that synonymous mutations are neutral and fixed by random genetic drift, it is possible to determine the mode of selection acting on nonsynonymous mutations. If all nonsynonymous substitutions were neutral, then their rate of occurrence per site (d N ) would be the same as that of synonymous substitutions per site (d S ), so that d N /d S equals one. A lower ratio of nonsynonymous to synonymous substitutions per site (d N /d S < 1) means that some proportion of the nonsynonymous mutations are deleterious and removed by purifying selection. Conversely, positive selection fixes advantageous nonsynonymous mutations faster than genetic drift fixes synonymous mutations (d N /d S > 1), although this is usually restricted to a small proportion of amino acid sites within any gene. In the hypothetical example of five gene sequences shown here, with d S given above the diagonal and d N below the diagonal, there is no evidence for positive selection because mean d N /d S the (0.577) is less than one. At the host level, most studies of the selection pressures acting on immune system genes have concentrated on genes implicated in the adaptive immune response against microbial pathogens, particularly those producing antibodies ( Sitnikova and Nei 1998 ; Sumiyama et al. 2002 ), or genes encoding reconnaissance molecules known as the major histocompatibility complex (MHC), which control the action of T-cells ( Hughes and Nei 1988 ; Yeager and Hughes 1999 ) ( Box 1 ). As its name suggests, the role of the adaptive immune response is to stimulate and ‘memorise’ immunity to specific pathogens. As microbial pathogens such as viruses are both abundant and rapidly evolving, positive selection on components of the adaptive immune response is often very strong ( Yeager and Hughes 1999 ). Far less attention has been directed toward the less specific innate (‘nonadaptive’) immune response, even though this response requires a wide array of genes and acts as the front line of immune defence ( Box 1 ). Would we expect the same strength of positive selection on a generalized pathogen control system? This is a question of fundamental importance because the luxury of adaptive immunity is not available to most organisms, having probably evolved along with the vertebrates ( Bartl et al. 1994 ), whereas the more widespread innate immune system is often depicted as a primitive characteristic. Molecular Evolution of the Innate Immune System The genes involved in innate immunity have recently come under the molecular evolutionists' gaze. One important group are the defensins, a large class of short antimicrobial peptides that constitute an effective immune response team in organisms as diverse as plants and primates ( Boman 1995 ). Because defensins are cationic (positively charged), they are able to interact with negatively charged molecules on the surface of microbes and permeate their membranes. Sequence analyses of defensins and similar antimicrobial peptides have revealed the telltale signatures of positive selection, with d N greater than d S in many comparisons ( Hughes 1999 ; Duda et al. 2002 ; Maxwell et al. 2003 ). Other genes of the innate immune system also seem to be subject to powerful positive selection. One dramatic example described in this issue of PLoS Biology is the APOBEC3G gene of primates ( Sawyer et al. 2004 ). This case is especially striking because rather than killing pathogens through protein or cellular interactions, like most immune genes, APOBEC3G works by manipulating the genome sequence of the invading microbe. The genomes of primates contain a family of nine APOBEC genes that encode enzymes involved in the editing of RNA and/or DNA through the deamination of cytosine (C), so that this nucleotide mutates to uracil (U). This is essential for various aspects of cellular function. APOBEC1, for example, is involved in the C→U editing of apolipoprotein B mRNA (therein christening the family), while another family member, the activation-induced deaminase, has a vital role in adaptive immunity in that it assists in the diversification of antibodies. Two more enzymes, APOBEC3G and APOBEC3F, form part of the innate immune system; they function as antiviral agents and are being intensively studied in the context of infection with the human immunodeficiency virus (HIV), the cause of AIDS. In particular, APOBEC3G targets the reverse transcription step of the HIV life cycle, in which the viral genomic RNA is converted into proviral DNA, which is then integrated into the host genome ( Mangeat et al. 2003 ). APOBEC3G-induced deamination at this stage results in monotonous guanine-to-adenine (G→A) nucleotide changes, a phenomenon called G→A hypermutation that had long been noted by HIV researchers without a clear understanding of its cause. We now know that G→A hypermutation is part of the innate immune response to retroviral infections. Although there is still some debate over exactly how APOBEC3G leads to viral eradication, the most likely scenario is that G3A hypermutation results in the generation and incorporation of a multitude of deleterious mutations that fatally disrupt viral functions. This strategy is likely to work well for retroviruses like HIV because their genomes are so compact that individual sequence regions often perform multiple functions. Under these cramped conditions, most mutations are likely to severely disrupt some aspect of viral function and thereby reduce fitness ( Holmes 2003 ). Indeed, it has been estimated that the deleterious mutation rate in viruses that replicate using RNA polymerases (either reverse transcriptase in the case of retroviruses or RNA-dependent RNA polymerase for other RNA viruses) is on the order of one error per replication cycle, so that many of the viral progeny produced by replication are defective ( Elena and Moya 1999 ). HIV is normally able to overcome this burden of deleterious mutation because of its remarkable reproductive power; each day, on the order of 10 10 virions are produced in a single infected individual ( Perelson et al. 1996 ), so that enough fit and able recruits will make it through to the next generation. Treating RNA Virus Infections Through Lethal Mutagenesis The high mutation rates of RNA viruses mean that adaptively useful genetic variation is produced frequently. The rub, however, is that fitness-enhancing mutations are a small minority, and the preponderance of deleterious mutations means that RNA viruses live on the edge of survival ( Domingo 2000 ). By increasing the rate at which deleterious mutations appear, APOBEC3G pushes viruses over this edge, causing a form of ‘lethal mutagenesis’ that results in their destruction; the rate of mutation becomes so high that no genome can reproduce itself faithfully, and the population crashes. Intriguingly, researchers designing new antiviral drugs have also begun to realise that forcing viruses into this sort of ‘error catastrophe’ might be an effective way to treat them ( Figure 2 ). There are a growing number of studies in which mutagens, such as ribavirin and 5-fluorouracil, are applied to viral infections in vitro and in vivo, including HIV, in the hope that these will induce so many deleterious mutations that the virus suffers an error catastrophe and is cleared ( Loeb et al. 1999 ; Sierra et al. 2000 ; Crotty et al. 2001 ; Ruiz-Jarabo et al. 2003 ). The results produced to date are highly encouraging, particularly when these error-inducing drugs are combined with more conventional treatment strategies that aim to reduce the rate of viral replication ( Pariente et al. 2001 ). The discovery that a natural antiviral agent, APOBEC3G, probably works in much the same way should provide even more encouragement. Figure 2 Lethal Mutagenesis As a Means of Controlling RNA Virus Infections (A) In a viral population prior to the application of mutagens, the mean error rate (white) is on the order of one per genome per replication (mutations marked by asterisks). (B) If a mutagen such as ribavirin is then applied to an infected patient, the mean error rate of the virus (black) is increased so that the population crosses a threshold of ‘error catastrophe’; after this point fitness declines dramatically and the population crashes. This drug-induced lethal mutagenesis seems to work more efficiently when it is used in combination with drugs that reduce the rate of viral replication. Sadly, however, the pathogens have fought back. The anti-HIV properties of APOBEC3G were discovered because most viral strains escape its neutralising properties. Lentiviruses like HIV possess a gene that encodes an protein called Vif (‘viral infectivity factor’) that counters APOBEC3G ( Sheehy et al. 2002 ). Hence, it is probably only in naturally occurring Vif-defective mutants that APOBEC3G is effective against HIV. Furthermore, because positive selection on APOBEC3G has operated for at least 30 million years and lentiviruses in general, and HIV in particular, are likely to be more recently evolved than this, it is clear that a broad range of retroviral pathogens have been responsible for the adaptive evolution of this particular immune gene ( Sawyer et al. 2004 ). Given the frequency with which the remnants of past retroviral infections are found in the mammalian genome ( Smit 1999 ), in the form of usually defunct endogenous retroviruses ( Box 1 ), it is likely that our genomes are continually bombarded with retroviruses like HIV but that the majority are cleared by innate immune mechanisms like APOBEC3G. It is possible that the retroviruses that successfully infect us are those, like HIV, that have managed to evolve strategies to avoid the destructive capacities of APOBEC3G. The intense selective pressure on the defensins and APOBEC3G illustrates that although the innate immune response is generalist in its action, it is as highly and intricately evolved as its better-studied ally, the adaptive immune system. Rather than being an evolutionary remnant, it is a dynamic and continually adapting system. Less clear is whether other host proteins act in the same manner as APOBEC3G. In particular, the most common and destructive pathogens faced by humans and other mammals are RNA viruses, such as influenza A, yellow fever, and hepatitis C. In most cases, our ability to survive these viral infections is simply a combination of good luck and good breeding; with the right combination of MHC alleles, itself a function of population history and what we by chance inherit from our parents, some individuals may be more able to fight off viral infections than others. The ubiquity of RNA viruses hints that our genomes might also contain an innate, yet highly adapted, defence system that targets this abundant class of pathogens by manipulating their mutation rate. Although utilizing lethal mutagenesis might one day be an important way to design new drugs against a variety of viral pathogens, it would come as no surprise if nature got there first. Box 1. Glossary Adaptive immune system. The pathogen-specific part of the vertebrate immune system. It is comprised of two major arms, antibodies (the humoral response) and T-cells (the cellular response), both of which lay down an immunological memory for future defence. Innate immune system. The nonspecific part of the vertebrate immune system. It has a wide variety of components, ranging from lysozymes in saliva to cytokines, defensins, interferons, and natural killer cells in a variety of tissues. Endogenous retroviruses. The (usually) dead remnants of functional retroviruses that are now passed on through the germ line like normal genes. It has been estimated that approximately 5% of the human genome is composed of endogenous retroviruses.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516797.xml
340946
T-Cell Differentiation and Progression of HIV Infection
null
Of the 300 or so viruses that cause disease in humans, HIV may have the greatest adaptive advantage. Like most persistent viruses—including the herpesviruses Epstein–Barr and cytomegalovirus (CMV)—HIV employs various strategies to counteract its host's response to infection. But HIV possesses a unique ability to sustain a progressive attack on the immune system—infecting the very cells that coordinate the immune response—leaving the body susceptible even to normally harmless microorganisms. It is these so-called opportunistic infections, rather than the human immunodeficiency virus itself, that makes HIV so deadly. The specific mechanisms that engineer this ongoing systemic attack have been the subject of intense research. HIV targets white blood cells with protein surface receptors called CD4. These CD4, or helper, T-cells normally orchestrate the body's immune response by signaling killer T-cells (which are also called CD8 T-cells, after their CD8 surface receptors) and other immune cells to multiply and differentiate—that is, become specially equipped to recognize a particular pathogen, or antigen. At the onset of infection, the immune system appears to respond normally, with a strong attack led by HIV-specific CD8 T-cells that initially contain the virus. But as the infection progresses, CD4 counts drop and the body's ability to renew T-cells decreases while its proportion of “antigen-experienced” CD8 T-cells increases. While the biological effect of this hyperactivity is unclear, it is apparent that patients with elevated immune activity face a poor prognosis. Investigating the interaction among immune activation, CD8 T-cell differentiation, and HIV prognosis, Victor Appay and colleagues report that a close connection between elevated immune activation and elevated levels of highly differentiated T-cells may bring further insights into how HIV exhausts the immune system. To examine the effect of elevated immune activation on T-cells, the researchers analyzed T-cells from a group of HIV-infected individuals collected at two distinct points in time: at the onset of acute infection—which is characterized by vigorous HIV replication—and after treatment, when viral replication is suppressed. To explore the connection between T-cell differentiation and clinical status, the researchers analyzed the T-cells from a group of untreated infected individuals divided into three subsets based on stage of infection: acute infection, chronic infection without progression, and chronic infection with signs of progression. During acute HIV infection, the vast majority (80%–90%) of the CD8 T-cell population was activated—not just the HIV-specific CD8 T-cells. Surprisingly, CD8 T-cells specific to the Epstein–Barr and CMV viruses showed significant activation levels during acute infection, suggesting that HIV may indirectly promote the replication of these viruses. When the researchers investigated the effects of this activation on T-cell differentiation, they found a correlation between increasing antigen concentrations and increasing CD8 T-cell activation and proliferation. And when Laura Papagno et al. analyzed the differentiation state of CD8 T-cells in individuals at different stages of infection, they found a progression in the proportion of highly differentiated CD8 T-cells associated with HIV disease progression. These results, the researchers conclude, show that chronic overactivation of the immune system during HIV infection produces the large pool of highly differentiated T-cells observed in HIV infection. T-cells go through various stages toward late differentiation, and it may be that the early-differentiated CD8 T-cells, which maintain the ability to proliferate, offer protective immunity. But highly differentiated cells, they propose, exhibit characteristics associated with “replicative senescence”—they are in effect old, worn-out cells that can no longer proliferate. Though replicative senescence is a natural process for most cells, in the context of HIV—in which infected individuals also lose the ability to replenish T-cells—it creates an aging population of T-cells that are less effective at fighting infection. Two T-cells, one of which recognizes a target cell
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC340946.xml
553963
Contribution of gap junctional communication between tumor cells and astroglia to the invasion of the brain parenchyma by human glioblastomas
Background Gliomas are "intraparenchymally metastatic" tumors, invading the brain in a non-destructive way that suggests cooperation between glioma cells and their environment. Recent studies using an engineered rodent C6 tumor cell line have pointed to mechanisms of invasion that involved gap junctional communication (GJC), with connexin 43 as a substrate. We explored whether this concept may have clinical relevance by analyzing the participation of GJC in human glioblastoma invasion. Results Three complementary in vitro assays were used: (i) seeding on collagen IV, to analyze homocellular interactions between tumor cells (ii) co-cultures with astrocytes, to study glioblastoma/astrocytes relationships and (iii) implantation into organotypic brain slice cultures, that mimic the three-dimensional parenchymal environment. Carbenoxolone, a potent blocker of GJC, inhibited cell migration in the two latter models. It paradoxically increased it in the first one. These results showed that homocellular interaction between tumor cells supports intercellular adhesion, whereas heterocellular glioblastoma/astrocytes interactions through functional GJC conversely support tumor cell migration. As demonstrated for the rodent cell line, connexin 43 may be responsible for this heterocellular functional coupling. Its levels of expression, high in astrocytes, correlated positively with invasiveness in biopsied tumors. Conclusions our results underscore the potential clinical relevance of the concept put forward by other authors based on experiments with a rodent cell line, that glioblastoma cells use astrocytes as a substrate for their migration by subverting communication through connexin 43-dependent gap junctions.
Background Glioblastoma, the most aggressive primary brain tumor, is invariably associated with profuse and long-distance invasion of the brain parenchyma. Accordingly, this tumor has been defined as "intraparenchymally metastatic" [ 1 ]. Intra-parenchymal dissemination is largely responsible for systematic recurrence despite treatments [ 2 ], and poor prognosis with a median survival time not exceeding one year [ 3 ]. Preferential routes of long-distance dissemination have been identified, among which the basal lamina of blood vessels and fiber tracts [ 1 , 4 ]. Numerous studies have revealed mechanisms of invasion in those conditions, specifying the molecular bases of glioma cells interaction with specific ligands of the brain extracellular matrix, their proteolytic modification by glioma cells, and the corresponding cellular receptors [ 2 , 5 ]. Less is known about the mechanisms that underlie the migration of glioma cells among neural cells, in the brain parenchyma, despite its profuseness in the area that surrounds the tumor mass, and its major role in recurrence. We have undertaken to identify the mechanisms of this particular type of glioma invasion, on the basis of the hypothesis that cells in the brain parenchyma may play a supportive role in this process. One key for the elaboration of this working hypothesis has been the demonstration that this intra-parenchymal invasion occurs in a non-destructive way (see refs and discussion in Bernstein, 1996), as this suggests that invading human glioblastoma cells may establish relationships of cooperation with their environment and in particular with resident brain cells. We have focused our attention on gap junctions because of two complementary reasons. First, evidence from peripheral tumors such as melanoma has implicated gap junctions in tumor cell migration [ 6 ]. Second, gap-junctional communication via connexin 43 (Cx43) between astrocytes and glioma cells has been demonstrated [ 7 ], and linked to a phenotypic transformation of astrocyte which may render the brain parenchyma permissive to glioma invasion. In a recent study, Lin et al. have strongly supported this hypothesis by establishing that non-migratory rodent C6 tumor cells display a migratory behavior following transplantation into the rat brain, when engineered to synthesize Cx43 [ 8 ]. Although Cx43 was shown to exert additional cell adhesion effects, the authors demonstrated that promotion of migration was not associated with this role but with the establishment of functional gap junctional communication (GJC). Astrocytes -to which glioblastoma cells are phenotypically most closely related- are widely inter-connected through gap junctions, formed predominantly by the protein Cx43 [ 9 ] and glioma cells also express Cx43 [ 10 - 13 ]. Building up upon the concepts defined by Zhang et al. (1999) and Lin et al. (2002), we have explored their clinical relevance by looking at the ability of human glioblastoma cells to establish functional gap junctions with astrocytes, and at the role that such functional interaction may play in their migratory behavior. To observe and quantify the migration of human glioblastoma cells in a three-dimensional brain environment ex vivo , our analysis has taken advantage of ISIS , the "intra-slice implantation system" that we have recently developed [ 14 ], in addition to in vitro glioma cell cultures and glioma/astroglial co-cultures. The ISIS set-up allows the implantation not only of human tumor cells but also of fragments of biopsied glioblastoma into rodent brain slices maintained in culture [ 14 , 15 ]. It allows for long term analysis of glioma cell migration in a three-dimensional organotypic environment that mimics that of the brain parenchyma and application of pharmacological agents to disrupt GJC. Results Heterocellular coupling and migration of human glioblastoma cell lines Heterocellular coupling was assessed in glioma-astrocytes co-cultures, using the dual-label technique, in which glioma cells were pre-labeled donors cells and astrocytes potential recipient cells labeled through gap-junctional diffusion of the dye. Heterocellular coupling was more extensive for GL15 cells (heterocellular coupling astrocytes/glioma ratio: 4.05 ± 0.6 recipient astrocytes per one GL15 donor cell) than for 8-MG cells (0.6 ± 0.2; p < 0.001) (Fig. 1A ). Figure 1 Heterocellular coupling of glioma cell lines, and migration assay in brain slice cultures. A, heterocellular coupling between astrocytes and human glioma cell lines (using the preloading method, see also fig. 6A). In contrast to 8-MG cells, GL15 cells established extensive GJC with normal astrocytes. Histograms represent mean heterocellular coupling indices (number of recipient astrocytes per double labeled tumor cell) of a minimum of three independent experiments (range 4 – 6; n= 120 to 225 donor cells per group; ± SEM; ***, p < 0.001). B, 8-MG cells and GL15 cells migration in brain slice cultures. Histograms of the average migration indices showing that GL15 were more invasive than 8-MG cells (mean values ± SEM. ***, p < 0.001). Analysis of tumor cells invasion was assessed in brain slice cultures, in which both homocellular coupling (between glioma cells) and heterocellular coupling with astrocytes can occur. GL15 cells displayed a significantly higher invasive potential than 8-MG cells [(S - S 0 ) / S 0 = 7.82 ± 0.68 vs. 2.67 ± 0.67, p < 0.001] (Fig. 1B ). Addition of CBX induced a significant decrease in the heterocellular coupling index of GL15 cells with astrocytes (down to 1.89 ± 0.74; p < 0.001) in co-cultures (Fig. 2B ) whereas coupling was similar in control cultures either untreated, or treated with the functionally inactive analogue of CBX, GZA. Inhibition of GJC by CBX following implantation of GL15 into brain slices, resulted in a significant decrease in GL15 cells invasion into the brain parenchyma. The surface ratio (S-S 0 ) / S 0 of migration was reduced by inhibition of the gap junction function (5.59 ± 0.49 for CBX-treated vs. 7.82 ± 0.70 in untreated controls; p < 0.01) (Fig. 2C ), as well as the number of GL15 cells that left the tumor mass (105 ± 6.3 cells/mm of perimeter vs. 132 ± 7.2; p < 0.01) (Fig. 2D ). There was no difference between results obtained in the GZA and control groups. Those results were confirmed when migration of glioma cells out of spheroids was analyzed over an astrocyte monolayer culture in a 42 hours-long time lapse experiment (Fig. 2E ). Figure 2 Effects of inhibition of heterocellular coupling between GL15 cells and astrocytes . A, Photomicrographs in co-cultures (preloading method) showing heterocellular coupling without (upper panel) or with addition of the inhibitor CBX. GL15 donor cells pre-labeled with a non-diffusing membrane-bound dye (DiI, red ) and loaded with a fluorescent gap junction-permeable dye (calcein, green ) were seeded on a monolayer of unlabeled astrocytes. Functional heterocellular GJC is visualized as the transfer of calcein from DiI-labeled donor cells (small arrows) to surrounding recipients cells. Donors cells appear yellow because of the merge of red and green labeling. CBX reduced the number of calcein-containing ( green ) recipient astrocytes, (×200). B, Histograms representing mean values of a minimum of three independent runs as in A (range 4 – 6; n = 160 to 225 donor cells per group ± SEM). (***, p < 0.001). C and D, histograms of the migration indices of GL15 in brain slices without or with CBX, illustrating decreased migration in the treated group (mean values ± SEM. **, p < 0.01). E, Cumulative migration indices plotted against time in co-cultures where GL15 cell spheroids were plated upon an astrocyte monolayer, without or with CBX; (S - S 0 ) / S 0 values were measured every 12 h, for 48 h. Integrated areas under the two curves in arbitrary units were compared, p < 0.0001. In contrast, CBX treatment had no significant effect on 8-MG migration, either in glioma-astrocytes co-cultures or following implantation into brain slices (data not shown). Inhibition of GJC reduces invasion ability of cells either from human glioma xenografts maintained in nude mice or from fresh human glioma biopsies The composite cellular content of glioma fragments and the technical difficulty in maintaining cells alive after seeding them in culture as cell suspensions, precluded analysis of the functional coupling of tumors, either maintained in nude mice or freshly biopsied. We, therefore, relied upon the effects of CBX to assess whether, similarly to results obtained in established glioma cell lines, inhibition of GJC may affect the migration of tumor cells out of a fragment and into the neural parenchyma of a brain slice. Invasive capacities of the seven tumors maintained in nude mice were quite heterogeneous. Three were minimally invasive according to our criteria [total number of cells that detached from the tumor mass was less than 15 cells and the distance of migration less than 50 μm (T3, T6 and T8)]. These three tumors were precisely the ones in the series of 7 tumors to not exhibit any response to CBX treatment (Table 1 ). Among tumors characterized as "invasive" according to our criteria, CBX treatment significantly inhibited invasion in three (T2, T4 and T7), and had no statistically significant effect in one (T1). Inhibition concerned both the surface ratio of migration (35 to 50% decrease) and the number of migratory cells (25% to 40% decrease) (Table 1 ). Treatment with the inactive analogue GZA had no effect on migration as compared to controls. Table 1 Cx43 expression and invasive capacity of gliomas maintained in nude mice Controls CBX Tumor Tumor § Cx43 (S-S 0 )/S 0 Cells/mm perimeter (S-S 0 )/S 0 Cells/mm perimeter T1 TG-4-GEN +++ 1.15 ± 0.23 77 ± 9.4 1.05 ± 0.11 68 ± 5.6 T2 TG-1-HAM +++ 0.56 ± 0.05 81 ± 9.4 0.29 ± 0.08 *** 50 ± 7.5 * T3 TG-7-ROM + 0 a - 0 a - T4 TG-14-RAV +++ 1.11 ± 0.08 73 ± 6.7 0.71 ± 0.04 ** 54 ± 4.7 * T6 TG-5-RAI +/- 0 a - 0 a - T7 TG-14-CHA +++ 0.45 ± 0.08 48 ± 4.2 0.22 ± 0.01 *** 35 ± 4.4 * T8 TG-20-THO +/- 0 a - 0 a - a , cut off : a tumor which exhibited less than 15 migrating cells distance of migration, if any, less than 50 μm, was considered non invasive. § Xenograft tumor nomenclature as originally described by Leuraud et al (2003). n = 7–10 per group and condition. ***, p < 0.001. **, p < 0.01, *, p < 0.05 In order to test whether the results obtained with experimental gliomas could be extrapolated to native tumors, the migration potentials of biopsies from 4 human high grade gliomas (HG) were studied in brain slice culture, with or without addition of CBX or GZA. Overall, CBX treatment reduced migration indices of these four tumors by 30% (range 20 to 50%; p < 0.05 unpaired t test and Mann & Whitney – Table 2 ); this was the most pronounced for HG-23 (p < 0.02, Fig. 3A ). Table 2 Cx43 expression and invasive capacities of human biopsed gliomass Controls CBX Tumor Cx43-IR (S-S 0 )/S 0 (S-S 0 )/S 0 HG-18 + 4.89 ± 0.85 (n = 6) 3.63 ± 0.86 (n = 6) HG-19 + 8.12 ± 0.83 (n = 9) 6.69 ± 0.88 (n = 7) HG-22 + 3.64 ± 0.66 (n = 9) 2.92 ± 0.21 (n = 8) HG-23 +++ 5.48 ± 0.85 (n = 6) 2.92 ± 0.47 (n = 6) IR: immunoreactivity Figure 3 Inhibition of gap junctional communication in biopsied human gliomas. A. Vimentin immunostaining of the biopsied human glioblastoma HG-23 7 days after implantation into a cultured brain slice without (left panel) or with (right panel) CBX. Note the inhibition of cell migration when GJC is blocked. B, Connexin 43 immunocytochemistry (DAB, counterstained with hematoxylin) in two glioblastomas showing either moderate (HG-22) or abundant (HG-23) expression of the protein (×200). At the cellular level (inserts, ×1000), staining is dot-like or linear, predominantly at the membrane. Cx43 expression in glioma cells In order to investigate whether Cx43 was instrumental in the establishment of functional coupling between glioma cells and astrocytes, Cx43 expression was analyzed in glioma cells from GBM cell lines and fresh human biopsies used in previous experiments. We chose to focus on Cx43 because it is the major Cx in astrocytes and is expressed by glioma cells [ 10 - 13 ]. In agreement with previous studies [ 16 ], three isoforms of Cx43 could be separated on SDS-PAGE and visualized by immunoblotting as three distinct bands in control astroglial primary cultures: Cx43-NP, a non-phosphorylated isoform, Cx43-P1 and Cx43-P2, the two phosphorylated isoforms of the protein (Fig. 4 ). The Cx43-P1 isoform was prominent and Cx43-P2 and Cx43-NP isoforms were equally present (Fig. 4 ). As for invasive potential, Cx43 expression pattern was different in the two GBM cell lines with the GL15 cells exhibiting a three-fold higher concentration of the Cx43-P1 as compared to 8-MG cells (p < 0.001; Fig; 4A ). Figure 4 Connexin 43 expression in glioma cells . A, Western blot analysis of Cx43 expression in cell lines GL15 (lanes 1 and 2) and 8-MG (lanes 4 and 5), and in primary mouse astrocytes (lane 3). Three isoforms of Cx43 were detected in normal astrocytes, that corresponded to two phosphorylated (Cx43-P1 and Cx43-P2) and one non-phosphorylated (Cx43-NP) isoforms. GL15 cells exhibited higher level of Cx43 protein than 8 MG cells. B, Western blot analysis of Cx43 expression in seven human gliomas maintained in nude mice (T1 to T8). All tumors expressed one phosphorylated isoform. In addition, the non-phosphorylated isoform was detected in 4 tumor samples (T1, T2, T4 and T7). Tubulin was used as internal control The gels shown, are representative of a minimum of three independent experiments. Histological analysis of Cx43 immunoreactivity in brain slices implanted with GL15 cells, showed that the protein was overexpressed in the tumor margin where glioma cells and astrocytes processes were intimately apposed and intermingled (Fig. 5 ). Confocal microscopy of brain slices allowed us to identify regions of enhanced immunostaining for Cx43 at the interface between GL15 cells and astrocytes, suggesting the existence of heterocellular gap junctions (Fig. 5 ). Interestingly, these astrocytes displayed a bipolar morphology that extended in arrays a long process that paralleled the outward migrating cells. This was not observed when 8-MG glioma cells were implanted. Figure 5 Cx43 immunoreactivity in brain slice culture injected with GL15 glioma cells. A, 3D projection of 15 consecutive confocal sections (0.5 μm) (×630), showing cooperation between migrating GL15 glioma cells (vimentin imunostaining, blue) with host astrocytes of the brain slice (GFAP immunostaining, red). Cx43 protein (green), is overexpressed by both tumor cells and astrocytes at the tumor margin where their processes are entwined. B, a single confocal section and two orthogonal projections (right and lower margins) corresponding to the overlayed perdendicular line. Enlarged view, of area boxed in B shows punctuate Cx43 immunoreactivity corresponding to gap junction plaques, at the interface of glioma cells and astrocytes strongly suggesting formation of heterocellular gap junctions. Note astrocyte morphologic changes which extend bipolar processes parallel to the long axes of the tumor cells A general expression of the P1 isoform was observed in the seven tumors maintained in nude mice, and a variable expression of the Cx43-NP isoform. The four invasive tumors (T1, T2, T4 and T7) expressed abundantly both Cx43 isoforms Cx43-NP and Cx43-P1, whereas the three others (T3, T6 and T8) which were not invasive and did not exhibit any response to CBX treatment, expressed lower levels of Cx43-P1 isoform and did not present detectable levels of Cx43-NP. (Fig. 4B ). No expression of Cx43-P2 was recorded. The amount of tissue available for experimental analysis did not allow us to carry out western blotting on freshly biopsied tumor fragments. Thus Cx43 expression in the human tumor biopsies was evaluated by immunohistochemistry on paraffin embedded tissue section. Cx43 immunostaining was seen in all 4 specimens. Staining was patchy, and both the intensity of staining and the number of immunostained cells in patches differed from one sample to another. In the typical case, a positive microscopic field contained several strongly stained cells, and several unequivocally labeled cells of lower intensity (Fig. 3B ). Homogenous fields of particularly strong immunolabeling intensity was observed in the HG-23 specimen. Homocellular coupling and migration of human glioblastoma cell lines We have next evaluated homocellular coupling between glioma cells by the scrape loading method, in which homocellular coupling is demonstrated by the diffusion of the dye from mechanically lesioned cells to their neighbors in a monolayer culture. Assay revealed extensive coupling of GL15 (homocellular coupling index: 63.75 ± 4.4 %), comparable to that displayed by astrocytes in parallel control experiments (60 ± 5.2%), whereas 8-MG cells were only weakly coupled (12.7 ± 1.87 %; p < 0.001 vs. GL15), a result in keeping with their low level of Cx43 expression. Homocellular coupling of all cell types was strongly affected by addition of the gap junctional blocker CBX: down to 7.4 ± 1.3 % for astrocytes (p < 0.001), to 1.97 ± 0.47 for GL15 (p < 0.001) and to 2.2 ± 0.47 for 8-MG cells (p < 0.001) (Fig. 6 ), whereas GZA had no effect. Figure 6 Homocellular coupling in human glioma cell lines using the scrape loading method. A, Dye coupling in absence (upper row) or presence (lower row) of the gap junctional blocker CBX. After scraping, injured cells were loaded with Lucifer Yellow (LY) and Rhodamin Dextran, and GJC was assessed by the number of neighboring cells that secondarily acquired green fluorescence. Left panel , confluent monolayer of GL15 cells in phase contrast. Right panel , centrifugal (left to right) diffusion of the dye tracer LY ( green ) from initially loaded injured cells ( yellow , because of the merge of red and green ) in the same monolayer. (×100). B, histogram representing the homocellular coupling indices (mean values ± SEM) from 4 to 6 independent cultures for each condition and cell type (astrocytes, GL15 or 8-MG). Homocellular coupling index = % labeled cells out of a total of ~400–800 analyzed cells per cell culture. Addition of CBX resulted in a dramatic decrease of the homocoupling in all cases ***, p < 0.001. To ensure that CBX-induced decreased migration of GL15 cells was mediated by the effect of the compound on gap junction communication, rather than more generally on the phenotype of tumor cells, analysis of tumor cell migration was performed in experimental conditions in which only homocellular coupling could exist, i.e. after seeding of glioma spheroids on a collagen IV matrix. Surprisingly, when homocellular coupling between tumor cells was inhibited by CBX, both tumor cell lines displayed a significant increase in migration out of the spheroids [(S-S 0 ) / S 0 = 55 ± 4.1 and 25.4 ± 5.25 for GL15 and 8 MG, respectively, p < 0.001 vs. untreated in both cases] (Fig 7 ). Figure 7 Migration assay of glioma cell lines on collagen IV. A, vimentin-immunostained GL15 and 8-MG spheroids, after 4 days of culture on collagen IV without (upper row) and with (lower row) addition of CBX, (x50). B, (S - S 0 ) / S 0 ratios (mean values ± SEM) obtained from 10 to 30 spheroids per group and condition. Addition of CBX enhanced spheroid disintegration and centrifugal dispersion of migrating cells. ***, p < 0.001. Induction of glioma cell migration in vivo by overexpression of Cx43 In order to validate further the hypothesis of a role of GJC formed by Cx43 in tumor cell migration, we have used a technique adapted from that used by Lin et al. (2002) [ 8 ]. C6 tumor cells have been genetically engineered by transfection, using a plasmid containing a fusion gene encoding GFP attached to the carboxyl terminus of Cx43. The gap junctions formed by this chimeric protein were dye-permeable [ 17 ]. C6 cells are known to proliferate without migrating out of the tumor mass, following implantation into the rat brain, and do not express a significant amount of Cx43. By transplanting mixed population of native and GFP-Cx43-transfected cells, we explored the capacity for the latter to display specifically a migratory behavior. Indeed, GFP-Cx43 expressing C6 cells were preferentially located at the periphery of the tumor mass and some were even observed scattered in the surrounding brain parenchyma, strongly suggesting a migratory behavior (Fig. 8 ). This was not observed with C6-GFP cells which were, as classically described, only present in the tumor mass. Figure 8 In vivo parenchymal invasion of C6 glioma cells overexpressing the fusion protein GFP-Cx43. A, vimentin-immunolabeled C6 glioma cells ( red ) 8 days after striatal injection in the nude mice brain. C6 cells overexpressing the Cx43-GFP protein in green (arrow heads), are preferentially located at the tumor margin and display an invasive phenotype (×100). B, Microscopic field of the area boxed in A, at a higher magnificence (×200). C, D, confocal section (1 μm) of the border of another tumor showing invasive C6-Cx43-GFP cells at the tumor periphery migrating out. C6 cells which did not express Cx43-GFP protein, only labeled in red for vimentin remain in the tumor mass. E, is the overlay of the two channels. Migrating C6-Cx43-GFP cells appear yellow because of the merge of red and green . Dashed lines indicate the extremity of the tumor mass. Scale bar, 25 μm. tm, tumor mass. Discussion The main result of this study is the demonstration that functional GJC between tumor cells and host astrocytes is instrumental in the invasion of the brain parenchyma by human glioblastoma cells. Establishment of functional coupling between those two cell populations may be mediated by the gap junction protein Cx43 as ability of glioblastoma cells of various origins to migrate out of the tumor bulk into the brain parenchyma, appeared related to the level of Cx43 expression. In support of this hypothesis, overexpression of a fusion Cx43-GFP gene construct triggered a migratory behavior in otherwise non-infiltrative C6 glioma cells implanted into the brain. The similarities between those results and mechanisms of neural precursor cell migration during embryonic development suggest that glioma cells may subvert and take advantage of a physiological mechanism of cell migration into the brain. In contrast, homocellular interaction between glioma cells through gap junctions hampers their dissemination out of the tumor mass, pointing to an additional role for gap junctions, in cell-to-cell adhesion. Over the past decade, studies on GJC in gliomas have mainly addressed its role in uncontrolled cell proliferation [ 18 - 20 ], by analogy with epithelial cancers in which a low level of connexin expression is a sign of profoundly perturbed tissue homeostasis and malignant evolution [ 21 ]. The notion that, in gliomas, GJC may also favor migration draws scientific argumentation from embryology and from observations that other neurectodermal tumors such as melanoma need to establish GJC with host cells in order to migrate out [ 6 ]. This concept has been strongly supported recently by the demonstration that a C6 rodent glioma cell line that had been genetically engineered to express Cx43 displayed a migratory behavior following transplantation into the rat brain [ 8 ]. This phenomenon was the more striking that the value of the rodent C6 cell line as an experimental model for glioblastoma has always been questioned because it did not reproduce the infiltrative phenotype which is a major characteristic of the human brain tumor. The induction of C6 migration was related to the establishment of functional GJC between glioma cells and host astrocytes through homotypic Cx43-Cx43 gap junctions. Several lines of evidence drawn from our results converge to indicate that this concept is clinically relevant, as GJC with astrocytes also plays a role in the migration of brain tumor cells for human glioblastomas. Experimental results obtained here with native human glioblastoma cells of various origins, can indeed be interpreted as a demonstration of a similar role for GJC. First, pharmacological blockade of GJC by the specific inhibitor CBX [ 22 ] had negative effects on tumor cell migration, whereas its inactive analogue GZA [ 23 ] did not. Second, the gap junction protein Cx43, which is overexpressed by reactive astrocytes in the margin of glioblastomas [ 10 ], was also expressed by all studied human tumors, including biopsy specimens, providing a molecular basis for the potential formation of homotypic heterocellular gap junctions. Last, close membrane appositions at regions of reinforcement of Cx43 immunostaining, suggesting intercellular contacts between astrocytes and glioma cells migrating out of a tumor spheroid or the body mass, were observed in co-cultures of glioma cells on an astroglial monolayer, as well as following implantation of tumors into brain slices maintained in organotypic cultures. The pharmacological approach used here had both advantages and drawbacks, which should be indicated as they may limit the conclusions to be drawn. The main advantage of that technique is the fact that it allowed us to study GJC inhibition in fresh tumor samples, in which biological properties and diversity of glioblastoma cells are well preserved, in contrast to the clonal selection and genetic variation that are commonly observed with cells lines. Conversely, we probably introduced some biological variability in the experimental system. This may be relevant to the fact that GJC inhibition with CBX did not abolish but rather reduced the rate of human glioma cell migration in our study, whereas Lin et al. (2002) obtained an "all-or-nothing" effect following genetic engineering of the C6 line. The partial effect on migration observed here may be due to the partial pharmacological inhibition consistent with the results of other studies [ 24 , 25 ] that demonstrated that CBX only reduced GJC by 50%. Alternatively, glioblastoma cell migration into the brain parenchyma does not relate exclusively to cell-to-cell interaction with astrocytes. Interactions between tumor cell and matrix components (reviewed by Giese et al., 2003) [ 5 ] are also major actors in glioma cell migration, and they are most likely not effected by inhibition of GJC, at least not directly. The use of a pharmacological compound also leads to questions about specificity. Various compounds used to block GJC have additional non-specific and toxic effects, among which halotane, octanol or acidifying agents. CBX has been shown to be nontoxic and has no effect on cell proliferation, total protein synthesis or cell viability [ 23 , 26 , 27 ]. This is ascribed, by some authors, to its mode of action, through direct binding on connexon particles [ 22 ]. We have checked the specificity of the GJC effects via different experimental ways, including a systematic negative control with glycyrrhyzic acid (GZA), the CBX precursor which only differs from it by its inability to disrupt GJC. CBX and its inactive analog GZA are known to both interact with mineralocorticoïd and glucocorticoïd receptors (see discussion in [ 28 ]), but these receptors do not mediate CBX-decreased GJC [ 23 ]. We also tested CBX effect on cell apoptosis, cell proliferation or cell viability in our conditions. No significant effect was observed. Although it has been reported that Cx43 may induce cell growth [ 18 , 29 ], we did not observe any effect of CBX on cell proliferation. It is interesting to mention that CBX-decreased GJC is mainly due to alteration of gap junction channel permeability, whereas Cx overexpression is sufficient to regulate cell growth [ 20 , 30 ]. As an interesting positive control of the specificity of the effects recorded in co-cultures and following implantation of tumor cells into brain slices, we have observed an opposite effect of CBX on cell migration when only homocellular interaction between glioblastoma cells was concerned (on collagen IV). Altogether, these results support the conclusion that non-specific pharmacological effects, if any, did not alter results of our study in a consequential manner. Levels of Cx43 expression were positively related to glioma invasiveness. Analysis of Cx43 expression by western blot allowed detection of two distinct isoforms of the Cx43 protein, Cx43-P and Cx43-NP. Both isoforms were detected in the four highly invasive tumors (T1, T2, T4, T7), whereas the three minimally invasive tumors (T3, T6, T8) displayed no detectable levels of Cx43-P isoform. The presence of both Cx43 isoforms strongly correlated with the invasive capacities of the tumor. However, whether the migratory capacity may depend on the overall levels of expression, and/or upon the concomitant presence of both Cx43 isoforms is, at this stage, a matter of speculation. Published data on the association of either isoform with the functional state of gap junctions are conflicting. Phosphorylation of Cx43 either on serine or tyrosine residues has been reported to disrupt GJC (for review see [ 31 , 32 ]). This could in part explain why T3, T6 and T8 tumors, which only displayed the phosphorylated isoform -the inactive isoform of Cx43-, were non-invasive tumors. However, the reverse association between Cx43 dephosphorylation and GJC inactivation has been suggested in astrocytes [ 33 ]. Further biochemical analysis of Cx43 isoforms function in these cells may reveal whether the pattern of Cx43 isoform expression is involved in the invasive process and how this pattern can be altered by specific phosphorylation. On the other hand, it cannot be excluded that the abundance of both -P and -NP isoforms in migrating tumors simply relates to a dynamic turnover of connexons between the cytoplasm (Cx43-NP) and the cell membrane of a migrating cell, where Cx43-P may participate in short-lived contacts with host astrocytes [ 21 ]. Recent data have shown that CBX itself did not have any effect on the phosphorylation pattern of Cx43 [ 34 ]. In the biopsy specimens, levels of Cx43 expression were different from one tumor to another, and from one zone to another within a specimen, in agreement with results reported by other authors [ 10 - 13 ]. Interestingly, the highest expression of Cx43 was seen in HG-23, the tumor in which migration was the most substantially inhibited by CBX. In addition, transfection of C6 glioma cells with a Cx43 protein fused to a green fluorescent protein allowed us to trace C6-Cx43 cells after intrastriatal stereotactic implantation. Consistent with results of Lin et al (2002) and [ 35 ], only C6-Cx43 cells switched to an invasive phenotype providing a direct evidence of the positive correlation between glioma cells Cx43 expression and invasive potential. We hypothesize that Cx43 may favor glioma cells migration by facilitating glioma cells to form gap junction interactions with host astrocytes that will allow them, through currently unknown molecular mechanisms, to drag out from the tumor mass. A peculiar morphological finding in this study was the phenotypic alteration of astroglial processes around tumors. Changes in the astroglial phenotype induced by glioma cells through GJC have been observed by other authors [ 7 ]. The ISIS protocol used here allowed us to gain a comprehensive three-dimensional view of the astroglial meshwork, and to identify the radial orientation of astrocytic processes toward the body of the tumors. This organization is reminiscent of the scaffolding of radial glial cells observed in the developing neocortex in vivo [ 36 ]and in vitro [ 37 ]. Radial glial cells extending long processes from the ventricular zone to the cortical plate [ 36 ]communicate with migrating neurons through Cx43 gap junctions [ 38 - 40 ]. Although tentatively, we may therefore speculate that astroglial Cx43 provide similarly support and guidance to the migrating tumor cells. In the same way as immature neuronal cells do [ 38 , 39 , 41 ], glioma cells may trigger the phenotypic change of surrounding astrocytes into radial glia-like cells, in order to make them suitable as a substrate for migration. The three different migration assays allowed us to explore the importance of intercellular interactions between migrating cells and cellular environment occurring during invasion. We have shown that inhibition of homocellular GJC in GL15 spheroids seeded on collagen IV (i.e. in the absence of astrocytes), boosted glioma cell motility, resulting in a higher centrifugal dispersion of cells. This result is in keeping with that obtained by Lin et al (2002) in an experiment in which Cx43 was shown to mediate glioma cell adherence and aggregation. However, accelerated motility of GL15 glioma cells when CBX blocked homocellular glioma/glioma GJC was not observed when glioma cells additionally formed heterocellular GJC with astrocytes, i.e. on astrocyte monolayer or in brain slices. Quite the contrary, heterocellular glioma-astrocytes GJC seemed to be necessary for the invasive process to occur normally, as CBX significantly blocked it in those conditions. Conclusions The purpose of our study was to determine whether the concept that functional gap junctional communication with host astrocytes facilitates invasion of malignant glioma cells put forward by other authors on the basis of experiments using a rodent glioma model (Lin et al, 2002), indeed apply to human glioblastoma and may, as a consequence, bear clinical relevance. This has indeed been clearly confirmed by the inhibitory effect of the specific GJC blocker, carbenoxolone, on the migration of human glioblastoma cells. Results obtained ex vivo on different supports of migration point to a functional difference between homocellular (glioma-glioma) and heterocellular (glioma-astrocytes) GJC. This may eventually be of prime therapeutic interest, by revealing mechanisms by which glioma cells disengage themselves from their neighbors in the tumor bulk, and establish new contacts with host astrocytes in order to migrate away [ 2 , 7 ]. Methods The role of GJC in human glioma invasion was investigated in vitro using three different types of preparations, namely glioblastoma cell lines, xenografts in nude mice, and freshly biopsied tumors. Because of selection for genotypes, the original cell heterogeneity of gliomas is restricted in xenografts and absent from cell lines. These models, therefore, do not fully reproduce the cell content of in vivo gliomas but, reciprocally, they present the advantage of being homogenous, easy to handle, and to allow analysis of large numbers of cells. ISIS uniquely permits to study, in addition, the invasiveness of freshly biopsied tumors with preserved cell repertoire. In those conditions, however, investigation is limited by the amount of available tissue. Human glioma cells The two human glioblastoma (GB) cell lines used in the study, GL15 [ 42 ] and 8-MG [ 43 ] were grown as cell culture in glioma-cell medium: 50% minimum essential medium (MEM) and 50% Dubelco Modified Eagle's Medium (DMEM) complemented with 10% fetal calf serum (ATGC Biotechnological, France), glutamine 2 mM, D-glucose 3.3 mM, Penicillin 100 UI/mL, Streptomycin 100 μg/mL (all from Eurobio, France) at 37°C in a 5% CO2 humidified incubator. On confluence, cells were trypsinized, centrifuged at 300 G for 10 minutes and resuspended in 1 mL of glioma-cell medium. 5 × 10 5 cells were seeded in 1.5% agar-coated flasks (25 cm 3 , Falcon, France) for 7–10 days to obtain GL15 and 8-MG spheroids. Spheroids with a diameter ranging from 50–100 μm were selected for migration assays. Seven grade III-IV human glioma biopsies xenografted and maintained subcutaneously in Swiss Nu/Nu mice (Charles River, France), kindly provided to us by Dr M-F. Poupon and P. Leuraud [ 44 ], were used in this study. For the sake of simplicity these tumors are referred to in the text as T1, T2, T3, T4, T6, T7 and T8 (see table 1 ). Mice were sacrificed when the tumor bulk reached a diameter larger than 1 cm. The tumors were retrieved and immediately transferred in the glioma-cell medium. Areas of necrosis and hemorrhage were identified under a microscope and discarded. The remaining tissue was cut into three fragments that were used as follows: -one was immediately implanted into a healthy nude mouse, in order to maintain the model; -a second one was immediately frozen for Western Blot analyses; -the third one was cut into smaller pieces for implantation into rodent brain slices maintained in organotypic cultures (see below). Finally, four fresh human biopsy specimens were obtained from stereotactic or resection biopsies in the operating room. All samples were retrieved by the neurosurgeon from areas of enhanced contrast on magnetic resonance imaging, which correspond to a hypercellular zone on histological examination of the tumor. After retrieval, samples were immediately transferred in glioma-cell medium and prepared for implantation into brain slices. Pathological analysis of the same tumors was carried out in parallel on adjacent specimens; it systematically included immunohistochemistry for the connexin protein Cx43 in addition to usual stains. According to the WHO 2000 classification, these tumors were glioblastomas in three cases, an anaplastic grade III astrocytoma in one. For the sake of simplicity the generic term of glioblastoma will be used in the rest of the text. Functional assessment of gap junctional communication Functional assessment of gap junctional communication (GJC) was performed on glioblastoma cell lines. Homocellular GJC was determined by the scrape-loading/dye transfer technique [ 45 , 46 ]. Briefly, either GL15, 8-MG or astrocytic cell cultures were incubated for 10 minutes in HEPES buffer solution (140 mM NaCl, 5.5 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 10 mM glucose and 10 mM HEPES pH 7.3). Cells were then washed in a Ca 2+ -free HEPES solution to prevent uncoupling. After scraping with the razor blade cells were incubated in the same solution containing 1 mg/ml of Lucifer Yellow (LY) (Sigma, France) and 1 mg/mL of Rhodamin Dextran 10,000 MV (Molecular Probes) for 1 minute. Dye transfer through GJC was assessed 10 minutes after scraping, using an Axioplan-2 microscope and appropriate filters (Carl Zeiss, Inc; Oberkochen, Germany). Fluorescent and bright light images digitized with a CoolSNAP camera (Ropper Scientific; Tuscon, AZ) from every microscopic field were merged, and the homocellular coupling index was calculated as the ratio of the number of LY-labeled cells to the total number of cells. Heterocellular GJC was evaluated in glioma/astroglial cells co-cultures by a method adapted from Goldberg et al. (1995) [ 27 ]. Tumor cells were loaded for 20 min with a dual-label dye solution containing isotonic PBS-glucose (0.3M), 5 μM Calcein-AM and 10 μM DiI (both from Molecular Probes, France). They were then rinsed three times with isotonic PBS-glucose 0.3M, trypsinized and centrifuged. Five hundred Calcein/DiI-labeled cells (donor cells) were seeded on an unlabeled confluent astrocytic monolayer (recipient cells), and incubated at 37°C in a 5% CO2 humidified incubator for 20 h. Heterocellular coupling was calculated as the ratio of recipient cells (calcein-labeled astrocytes) to donor cells (Calcein/DiI-labeled glioblastoma cells). It is important to underline that this technique is different from the scrape loading method and uses another dye tracer to assess the gap junctional coupling. Hence, their results cannot be readily compared in quantitative terms. Effects of inhibition of GJC were evaluated by treating the cultures with carbenoxolone (CBX); controls received its inactive analog, glycyrrhyzic acid (GZA) [ 23 ] (both from Sigma, France). CBX is a widely used specific inhibitor of GJC likely exerting its action by changing the conformation of gap-junction channels [ 22 ], while GZA has no effect on GJC. CBX and GZA treatments were carried out at a concentration of 30 μM. Cytotoxicity of these molecules on the two glioblastoma cell lines and on astrocytes had been tested in preliminary experiments. There was no significant increase in the lactate dehydrogenase medium/cell ratio for up to 7 days at the concentrations used (Cytotoxicity Detection Kit, Roche Molecular Biochemicals, France). Cell migration assays Migration of glioblastoma cell lines was analyzed in three different experimental in vitro set ups: on collagen IV-coated cell culture dishes, after seeding on an astrocytic monolayer culture, and following implantation into organotypic brain slices ( ISIS , Intra-Slice Implantation System). Migration out of tumor fragments issued from xenografted gliomas or from freshly biopsied specimens were only studied using the ISIS method. Migration of GL15 and 8 MG cells out of spheroids deposited onto collagen-coated (0.5 μg/mL; Sigma, France) cell culture dishes was analyzed over 4 days. For astrocytic co-cultures, primary astrocyte cultures were obtained from cerebral hemispheres of neonatal C57/bl6 mice (Charles River, France) as previously described [ 47 ]. Cells were grown in astrocyte medium: MEM (GIBCO, France) complemented with 10% FCS (ATGC Biotechnological, France), glutamine 2 mM, D-glucose 33 mM, Penicillin 100 UI/mL, Streptomycin 100 μg/mL and amino acids 0.5X (all from Eurobio, France). Medium was changed every 3 days. On confluence, GL15 and 8-MG spheroids were seeded on top of the astroglial monolayer, and cultures were continued for 4 days. The ISIS migration assay was carried out as described previously [ 14 ], using brain slices obtained from postnatal day 6 C57/bl6 mice. Briefly, after decapitation the two hemispheres were separated and 400 μm-thick coronal slices were cut using a McIllwain tissue chopper (Mickle Laboratory, England). Slices were transferred into brain slices medium: MEM (GIBCO, France), 1 g/l D-glucose, 10% heat-inactivated fetal calf serum, 0.1 g/l transferrin, 16 μg/l putrescin, 40 μg/l N-selenium, 30 μg/l tri-iodothyronin, 5 mg/l insulin, and 60 μg/l progesterone (all from Sigma, France). Slices were separated and transferred onto Millicell-CM membranes (Millipore, France). The Millicell-CM membranes were kept in six-well plates, above 1 ml of brain slice medium. Slices were incubated at 37°C in a humid atmosphere with 5% CO 2 . The medium was changed three times a week. 24 hours after brain slice preparation, GL15 and 8-MG spheroids or small fragments of xenografted tumors and freshly biopsied specimens were deposited onto the slice surface and gently pushed with the tip of a needle, until the tumor tissue was well enveloped by the parenchyma of the slice. Cultures were maintained for 5 or 7 days for cell lines and tumor fragments, respectively. In all experimental conditions, effects of the blockade of GJC were assessed using CBX (30 μM for collagen-coated and astrocytic co-cultures; 60 μM for brain slice cultures, a stronger dose because of the thinness of the slice), with two negative controls: untreated and GZA supplemented to the medium instead of CBX. These compounds were added to the culture medium 24 h after glioma spheroid or tumor implantation. In vivo experiments For a control experiment, we transfected the C6 glioma cell line with a plasmid vector encoding either a green fluorescent protein (GFP) fused to the carboxyl terminus of Cx43 [ 17 ] kindly provided by Dr P. Martin, or else only GFP. C6-Cx43-GFP cells and C6-GFP control cells were implanted into the striatum of nude mice (5. 10 4 cells in 1 μL) (n = 5). Mice were sacrificed at day 8 and perfused with paraformaldehyde (4%) 8 days after implantation. Brains were sectioned serially on cryostat at 30 μm thickness as previously described [ 48 ]. Immunoblotting and immunohistochemistry Immunoblot analysis for Cx43 was performed on cells and tumor samples that were suspended in lysis buffer containing Tris base 50 mM, pH 7.4, NaCl 150 mM, Triton X-100 0.5%, EDTA 1 mM, protease inhibitor cocktail 0.5% (Sigma, St. Quentin, France) and triturated to homogeneity. Homogenates were centrifuged at 13,000 G for 10 or 30 minutes (for cell lines and tumor samples, respectively), and supernatants were aliquoted and stored at -80°C. Total protein content was determined by the Lowry method with bovine serum albumin as a standard. Total protein (30 μg) was boiled for 5 min after addition of 10% glycerol/5% mercaptoethanol and separated on SDS-PAGE 12.5% gels, transferred on PVDF membrane (Bio-RAD, France) under classical conditions [ 49 ] and immunoblotted overnight at 4°C, using a polyclonal anti-Cx43 antibody (1:500, Zymed, France). Bound primary antibody was detected with a horseradish peroxydase-coupled anti-rabbit antibody (1:5,000; Amersham Pharmacia Biotech, France). In preliminary positive control experiments, Cx43 was revealed as three bands of about 41, 43 and 45 kDa, representing three different isoforms in heart sample. The two higher molecular mass bands have been shown to correspond to phosphorylated forms of the protein, and the 41 kDa migrating species to the unphosphorylated form of the protein [ 16 ]. Immunohistochemistry was carried out on cell and slice cultures and brain slices from in vivo experiments following fixation using 4% paraformaldehyde in PBS at 4°C (30 min for mono- and co-cultures, 4 h for ISIS ). For fluorescent immunostaining, cultures were rinsed several times in PBS, then incubated for 1 hour in PBS containing 10% normal horse or goat serum and 0.6% Triton X-100 (Sigma). They were incubated overnight at 4°C with monoclonal antibodies against vimentin (1/400, Neo Markers, France), or polyclonal antibodies against glial fibrillary acidic protein-GFAP (1/100, Dako) or Green Fluorescent Protein-GFP (1/200,. Molecular Probes). After 3 washes in PBS, they were then incubated with a PBS solution containing anti-rabbit or anti-mouse immunoglobulin conjugated either to fluorescein isothiocyanate-FITC, AMCA (1/400; Vector, France), or to cyamidine-Cy3 (1/1000; Jackson Immuno Research, France). For human biopsies, immunocytochemical detection of Cx43 was carried out on formalin-fixed, paraffin-embedded 5 μm sections. Antigen-retrieval consisted in heating slides immersed in citrate buffer (pH 6) in a microwave oven (3 × 5 min at 750 W). To block non-specific binding sites slides were incubated in PBS containing 10% BSA. A primary polyclonal antibody (Zymed) directed against the intracellular portion of the Cx43 molecule was applied (1:200) to sections overnight. Bound primary antibodies were detected by a Vector Elite ABC kit (Vector Laboratories, Burlingame, California) following the manufacturer's instructions and using DAB as the chromogen. Slides were lightly counterstained in Harris's hematoxylin. Negative controls had the primary antibodies omitted. A specimen of normal human cortex from epilepsy surgery showed diffuse staining of astrocytic cells accentuated at perivascular and superficial locations, corresponding to astrocytic endfeet and the glia limitans, respectively. Quantification and statistical analysis Fluorescent microscopy on a Zeiss Axioplan2 microscope and image analysis using the KS.400 (3.0 version) software were performed to quantify invasion of glioma cells, according to techniques that have been detailed elsewhere [ 14 ]. Two invasion parameters were assessed, (i) the maximum area of invasion, given as the ratio (S-S 0 ) / S 0 , where S is the maximum area containing migrating glioma cells and S 0 is the initial tumoral mass area, this ratio compensating for the heterogeneous sizes of implanted spheroids and tumor fragments, and (ii) the number of invading cells (cells/mm of S 0 perimeter). For collagen-coated migration assay and astrocytic co-cultures, only the maximum area of invasion was assessed. The statistical significance was evaluated using Mann & Whitney test and Student unpaired t test. Authors'contribution ROconceived and designed the study, performed acquisition, analysis and interpretation of data and drafted the manuscript. CCconceived and designed the study, performed acquisition, analysis and interpretation of data and drafted the manuscript. JSGparticipated in the design of the study and carried out xenografts in nude mice. SdB participated in the design of the study. SP provided human biopsies tissue and diagnosis. LV participated in the design of the study. MT have been involved in revising the article critically for important intellectual content. MPconceived, designed and coordinated the study, drafted the manuscript and have given final approval of the version to be published. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553963.xml
518962
Evidence for positive selection on Mycobacterium tuberculosis within patients
Background While the pathogenesis and epidemiology of tuberculosis are well studied, relatively little is known about the evolution of the infectious agent Mycobacterium tuberculosis , especially at the within-host level. The insertion sequence IS 6110 is a genetic marker that is widely used to track the transmission of tuberculosis between individuals. This and other markers may also facilitate our understanding of the disease within patients. Results This article presents three lines of evidence supporting the action of positive selection on M. tuberculosis within patients. The arguments are based on a comparison between empirical findings from molecular epidemiology, and population genetic models of evolution. Under the hypothesis of neutrality of genotypes, 1) the mutation rate of the marker IS 6110 is unusually high, 2) the time it takes for substitutions to occur within patients is too short, and 3) the amount of polymorphism within patients is too low. Conclusions Empirical observations are explained by the action of positive selection during infection, or alternatively by very low effective population sizes. I discuss the possible roles of antibiotic treatment, the host immune system and extrapulmonary dissemination in creating opportunities for positive selection.
Background How actively do populations of Mycobacterium tuberculosis cells undergo adaptive evolution on the spatial and temporal scales of individual infections? On the one hand, the long generation time and limited sequence diversity of this organism might suggest a slow pace of adaptive evolution. On the other hand, the rapidity and ease with which antibiotic resistance is generated during infection suggests otherwise. The physiology and immunology of tuberculosis pathogenesis have been well studied. The infectious agent M. tuberculosis is known to invade and replicate within alveolar macrophages. There is a spectrum of responses by the immune system, corresponding to the relative involvement of Thl and Th2 immune cells, which respectively stimulate the cytotoxic response (more effective against infected cells), and the humoral/antibody response (more effective against extracellular pathogens) [ 1 , 2 ]. Some progress has been made in describing the population dynamics of mycobacterial infection quantitatively [ 3 - 5 ]. At the wider spatial and temporal scales of populations, the molecular epidemiology and the evolution of M. tuberculosis have been carefully studied. Genotypic data are rapidly accumulating in the molecular epidemiology of infectious diseases. These are usually compiled and summarised to make inferences about the state of an epidemic in a given geographic location, or at the global level. For example, epidemiologists seek to identify risk factors for infection, and to locate particular strains that are especially transmissible or pathogenic [ 6 - 8 ]. The evolutionary history of M. tuberculosis has also been characterised. For example, it has been argued that the limited variation at the nucleotide is due to a recent population bottleneck [ 9 ], and that the common ancestor of Mycobacterium bovis and M. tuberculosis may well have been a human rather than bovine pathogen [ 10 ]. Less understood is the evolution of M. tuberculosis at the cellular level inside bodies. There has been little integration of the genetic information from markers with the population genetics of the bacterial population within hosts. In this article, I examine data collected for the purposes of molecular epidemiology to present three lines of evidence supporting the action of positive selection on M. tuberculosis . The data come from the marker IS 6110 , which is currently the standard method of typing tuberculosis isolates. These genotypic data will be considered under assumptions of neutrality, and then under the assumption that positive selection is acting. The case for the action of selection is based on the following three arguments. • Under the assumption of neutrality, the observed mutation (or transposition) rate of the genetic marker IS 6110 is unusually high; the estimated mutation rate is lower if selection is acting. • The observed times associated with change are too low to be explained by neutrality; positive selection lowers the expected substitution time. • The observed level of polymorphism is too low to be explained by neutrality. Results: Models and observations In each of the following sections a comparison is made between the strictly neutral model and a generalised model including selection through a single parameter s (described in the Appendix). Although the analyses start with strict neutrality ( s = 0) in each argument, alleles for which s < 1/ N , where N is the effective population size, can be considered nearly neutral , in that the effects of drift outweigh the force of selection [ 11 ]. In each case, explaining observations in this range of selective coefficients requires very low effective population sizes. Transposition rates of IS6110 When genetic mutations are selectively neutral, the substitution rate is equal to the mutation rate [ 11 ]. In the present case, the within-host substitution process is of interest. Rosenberg et al. [ 12 ] determined the within-host substitution rate of the IS6110 marker to be around 0.00184 to 0.0390 events per copy per year, with the maximum likelihood estimate at 0.0287. Under neutrality, therefore, this rate corresponds to a per insertion mutation rate of μ i ~ 7.9 × 10 -5 events per site per generation, assuming a generation time of 1 day in active infections. This figure comes from a measured doubling time of close to 24 hours, based on clinical isolates grown in human monocyte cultures and in culture media [ 13 - 15 ]. Rates of point mutation (events per nucleotide per generation) are usually in the vicinity of 10 -9 . In mutator strains, that is, genomes in which the DNA repair machinery is damaged, leading to elevated mutation rates, the mutation rate rises orders of magnitude, up to ~10 -7 – 10 -6 [ 16 ]. The mutation rate of IS 6110 under neutrality therefore seems suspiciously high, although this is only "circumstantial evidence", since it is not inherently problematic. Indeed, mutation rates as high as 10 -4 per element per generation have been measured for IS 10 in vitro [ 17 ]. Nevertheless, if positive selection is allowed the estimated mutation rate decreases. Leaving aside the complicating influence of clonal interference [ 18 ], the rate of substitution is K = uN μ i (1) where u is the probability of fixation of a mutant, μ i is the mutation rate and N is the population size [ 11 ]. An estimate of the mutation rate when mutants have advantage s is = K /( uN ). The diffusion model of drift provides an expression for u as a function of the population size N and selective coefficient s (see the Appendix). Figure 1 plots over s for a few different values of N . In each curve the estimated mutation rate decreases as the selective coefficient rises. According to this analysis, lower mutation rates are possible when there is some selection and a large population size, or when selection is strong and the population size is small. Note that the estimated mutation rate remains high if mutations are nearly neutral. Figure 1 Estimate of mutation rate when positive selection is acting. The estimate is plotted on a logarithmic scale in base 10. Solid curve: N = 10; Dashed; N = 1000; Dotted: N = 10 5 . Fixation times Various studies have measured the stability of IS 6110 as a genetic marker by examining genotypes of serial isolates from patients with persistent infection. A small number of changes in the genotypes between serial isolates indicates a stable marker. Differences in genotypes due to exogeneous reinfection by unrelated strains are excluded from consideration. In the data of Niemann et al. [ 19 ] and Rosenberg et al. [ 12 ], the median time interval associated with changes in IS 6110 genotypes from serial samples of M. tuberculosis is 212 days, and the maximum is 683 days. Because the second sample is taken some time after fixation of the mutant, the actual substitution times are unknown, but they were clearly all under 683 days. I will now show that the expected substitution times under strict neutrality are well in excess of this value. Let us start with the assumption that the expected time for substitution to occur is the average time taken for the successful mutant to appear plus the time taken for that mutant to reach fixation conditional on its eventual fixation. (I will later drop the assumption about waiting for the mutant to appear). The average appearance time is 1/( μNu ) = 1/ μ since u = 1/ N under strict neutrality. The average time for a successful neutral mutant to reach fixation is 4 N generations. The mutation rate of interest in this context is the rate per genome per generation, since what is of concern is whether any of the elements in a given genome produce change. For simplicity, assume that the genomic mutation rate scales linearly with copy number. (At the resolution of this analysis, this is a reasonable approximation.) Considering a typical strain has 10 copies of the IS element, the relevant mutation rate here is μ = μ i × 10 = 7.9 × 10 -4 . Therefore, for N = 10, 10 3 , 10 5 , the expected substitution times are roughly 1300, 5300, 4 × 10 5 generations, respectively. With the generation time set to one day, the upper bound of observed substitution times was 683 generations, which is well below theoretical expectations. Now consider the possibility of positive selection under two alternative conservative assumptions. The earlier assumption that there are no successful mutants at the time of the first sample is favourable to the parental strain. A more conservative approach (favouring mutants) would be to say that the mutant destined to reach fixation appears exactly at the time of the first sample. We can then ask how long it takes on average for this mutant to reach fixation if it is positively selected. An even more conservative model would be that not only is the successor strain present at the time of the first sample, but is present at a frequency of 30%. Furthermore, let us say the subdominant strain only needs to be at 70% at the time of the second sample to be considered to have replaced the parental strain. A model of the sojourn times of alleles in populations conditional on fixation must now be specified. Again using the diffusion model of drift (see Appendix), the mean time spent by a mutant in the range of frequencies ( a, b ) (provided a is greater than the initial frequency), conditional on fixation, was found by Ewens [ 20 ] and by Maruyama [ 21 ] to be Figure 2 shows the two conservative models, corresponding to two different boundary values for ( a, b ). Even in the extremely conservative model shown in the right-hand plot, the effective population size must be below 400 in order to explain the observed substitution times under strict neutrality. The data are difficult to account for even in terms of nearly neutral mutations ( s < 1/ N ) and an effective population size of N = 1000. The alternative explanation is that the effective population size is larger, but positive selection is acting to make changes sweep through the population faster. Figure 2 Mean sojourn times as functions of selective coefficient s , for different values of N . Left: from a = 1/ N to b = 1 - 1/ N ; Right: from a = 0.3 to b = 0.7. Polymorphism Many analyses of pathogen genotypes assume isolated strains to be clonal, that is, to be monomorphic. This assumption has been scrutinised by De Boer et al. [ 22 ], who showed that, in fact, a large proportion (93%) of M. tuberculosis isolates are monomorphic using IS 6110 as the marker. They also show that the limits of detection of a second strain are around frequencies of 0.1 to 0.3. More sensitive instruments and refined genotyping procedures are likely to reveal greater polymorphism. The current information can be used, however, to study the population of the organism in hosts by using ranges of detectable polymorphism . In this section, two ranges will be considered in examining predictions from models: first, 0.1 to 0.9, and second, 0.3 to 0.7. The polymorphism argument rests on the assumption that the isolates reported in [ 22 ] can be viewed as a random sample from a set of populations in mutation-drift equilibrium. It should be noted that because the isolate represents a sample of cells from the patient, it presumably does not always reflect the diversity of cells in the greater within-host population. Thus the polymorphism or heterogeneity observed from isolates is an underestimate of the actual levels. Wright [ 23 ] found the stationary probability distribution of allele frequencies under the diffusion model with mutation and two alleles. Let f ( x ) be the probability density function of this distribution and F(x) be the cumulative probability function F(x) (see Appendix). The probability that a given population (patient) is between frequencies a and b (where a < b ) is This quantity can be alternatively interpreted as the proportion of populations observed to be polymorphic according to the detection limits set by ( a, b ). First consider the neutral case. When there is no selection ( s = 0), the distribution described by f(x) is a Beta distribution. Figure 3 shows the probability of an isolate being scored as a polymorphic population, using two alternative detectable polymorphism ranges ( a, b ) = (0.1, 0.9) and (0.3, 0.7), and a mutation rate of μ = 7.9 × 10 -4 per cell per generation. Figure 3 Probability of detecting polymorphism in the absence of selection, as a function of N . Two different ranges of detectable polymorphism were used. Dashed curve: (0.1, 0.9); dotted: (0.3, 0.7). We use μ = 7.9 × 10 -4 . The horizontal bar indicates the observed fraction of polymorphic populations (0.074) from de Boer et al . [22]. Next, consider the model that includes selection. For the two detectable polymorphism ranges, Figure 4 shows how selective coefficient s and effective population size N are related to the probability of observing polymorphism. As s increases, the predicted polymorphism decreases dramatically, particularly for large N . Again, an explanation of the observed level of polymorphism is only possibly by setting N to be extremely low. Figure 4 Probability of polymorphism as a function of s . Left: the detection threshold is set at 0.3; Right: the detection threshold is set at 0.1. The mutation rate is set to μ = 7.9 × 10 -4 Discussion The three lines of evidence presented in this article suggest positive selection on M. tuberculosis within hosts. There are, however, limitations to these analyses. In the first argument there is no inherent problem with finding transposition rates that are high. In the second argument, it is possible to lower the effective population size far enough to explain the speed of substitution. In the third argument, 1) the bacterial populations sampled in [ 22 ] might not be close to mutation-drift equilibrium, 2) the sampled cells might not reflect the true diversity of the bacterial population in a patient, and 3) the levels of polymorphism again may be explained by very low effective population sizes. Consistency with observations nevertheless requires N values of around 100 or lower, which seems grossly at odds with the usually large census population sizes of bacteria. In mouse models of TB infection, for instance, bacterial loads reach around 10 5 – 10 7 colony forming units per lung [ 2 , 24 ]. It has been noted, however, that effective population sizes of bacteria can be much lower than actual sizes [ 11 , 25 ]. I will also comment on why I have not attempted to statistically fit the model to data to estimate N and s . First, from the plots shown here, it is clear that different combinations of the two parameters can explain the observations. This would make it difficult to locate the best fit. Second, although the model can be used to assess the possibility of neutrality in the current context, it cannot adequately serve as a framework for estimation given the intricacies of host-pathogen interactions. Further, adding more parameters to the model would increase the complexity of the analysis beyond what can be sustained by the resolution of the currently available data. Taken together, the results suggest positive selection, although the evidence is not conclusive. A possible alternative is that the effective population sizes of M. tuberculosis within patients are very low due to population structure, background selection, or other factors. If there is indeed detectable adaptive evolution of tuberculosis within patients, what are the sources of selection? Two important candidates are antibiotic treatment and the host immune system. Studies using serial isolates have found no correlation between IS 6110 genotype instability and (a) drug resistance/susceptibility of the isolate [ 26 , 27 ], (b) change in drug resistance status [ 19 ] or (c) drug adherence by the patient [ 26 ]. It is still possible, however, that the collection of observed changes involve a variety of different genetic loci, with at least some conferring drug resistance, although such events may not be statistically detectable. Further, mutation in drug resistance loci will not necessarily be revealed by a marker. Genetic analysis of isolates of M. tuberculosis from the lung lesions of six patients has shown heterogeneity in resistance-associated alleles, but not with respect to IS 6110 [ 28 ]. Alternatively, fingerprint changes may reflect (evolutionary) escape from the immune system. The analysis here hints at low effective population sizes – perhaps the immune system induces a heavy decline in population sizes of M. tuberculosis within patients, i.e., bottlenecks – which is overcome by survivors with new genotypes. If the observed patterns are to be explained by severe bottlenecks, the surviving cells are not necessarily better adapted to residing in the host than the parental cells that were eliminated by the immune response. It is noteworthy that Yeh et al. [ 26 ] found no relationship between HIV status of the patient and genotype instability. This suggests that genetic changes in M. tuberculosis are not primarily driven by the immune system. However, the extraordinary ability of M. tuberculosis to manipulate the T cell response [ 2 ] suggests the role of adaptation to the immune system in the deeper evolutionary history of the organism. Interestingly, de Boer et al. [ 27 ] found an association between IS 6110 change and extrapulmonary disease or pulmonary+extrapulmonary disease and extrapulmonary origin of isolates. Dissemination is a major factor in the pathogenesis of tuberculosis. Since the lungs are the preferred environment of the organism, the new environments outside lungs may create opportunities for adaptive evolution. Adaptive evolution leading to specialisation to tissue types is to be expected. A recent article [ 29 ], for example, has found the occurrence of tissue-specific adaptations in Streptococcus pyogenes by examining ratios of non-synonymous to synonymous substitution rates ( d n / d s ). Is IS 6110 directly responsible for adaptive mutations? On one hand, the apparently strict asexuality of M. tuberculosis implies that all genes good, bad or neutral are tightly linked to each other. It is likely then that IS-induced changes hitchhike to fixation with other mutations that confer advantage to the genome. On the other hand, it has recently been demonstrated that IS 6110 carries a promoter that can modify the the expression of neighbouring genes, raising the possibility of a direct role for the element in adaptive evolution [ 30 ]. Note that changes caused by IS 6110 can be not only beneficial, but also neutral or deleterious [ 31 ]. At the within-patient level, the best studied pathogen is perhaps HIV. While M. tuberculosis shares with viruses the characteristic of replicating within cells, a major difference is that mutation rates in viruses are much higher, particularly in retroviruses, which depend on reverse transcriptase (a low-fidelity enzyme) to copy their genomes. Hence, the extent of nucleotide variation of M. tuberculosis is not expected to be the same as is commonly observed for example in HIV [ 32 ]. There is ongoing controversy among HIV researchers about the role of stochasticity due to low effective population sizes in the evolution of the virus [ 33 - 35 ]. In any case, investigating the ratio of non-synonymous to synonymous substitutions ( d n / d s ) has established the action of positive selection on HIV within patients [ 32 , 36 ]. In M. tuberculosis , the level of polymorphism at synonymous sites has been noted to be extremely low [ 9 ]. It would be of interest to measure the ratio of non-synonymous to synonymous polymorphisms in key genes, such as loci conferring resistance to drugs, or those implicated in interactions with the immune system. These d n / d s ratios may provide further insight into the nature of positive selection in M. tuberculosis . Appendix: Bacteria and the Wright-Fisher process The analyses here rely on the commonly used diffusion model of drift and selection in a population, based on the Wright-Fisher process [ 37 , 38 ]. It is also possible to use the Moran model, in which at each time step an individual is chosen randomly to reproduce, and then another individual is chosen to die. The individual to die may be the same as the individual that reproduced, but not the offspring. Selection can be incorporated by including differential probability of birth or death for different genotypes. As noted by Ewens [ 38 ], the Moran model closely resembles the Wright-Fisher model; the critical difference between the two models arises from differences in the distribution of offspring number. The theory is usually discussed in relation to a diploid population of size N e in which there are 2 N e copies of the (autosomal) gene in question. The diffusion model is used here with minor adjustments to describe bacterial populations. Let the number of bacterial cells in a population be N . Each mutant appears in the population at frequency 1/ N . Realistically, population sizes fluctuate and only a subset of cells actively divide. The number N should therefore be considered to be the effective rather than the actual population size, which may be much larger than N . Mean and variance of change It can be shown that the deterministic dynamics of selection in a haploid model are well approximated by a logistic model. The mean change in frequency x of an allele per generation is m ( x ) = sx (1- x ), which is identical to the diploid model with additive fitnesses (no dominance) if each copy of the advantageous allele adds s to the fitness (see [[ 37 ], p. 192]). That is, heterozygotes enjoy a fitness advantage s and homozygotes have advantage 2s . The variance component v(x) , the variance in change of allele frequency per generation, can also be taken from diploid theory. Replacing the diploid model of the random union of gametes with choosing cells randomly from each generation to the next, the effective population size is adjusted according to the distribution of offspring number under a given model of cellular division. Binary fission There are numerous ways to model drift in populations of organisms that reproduce by dividing to produce two daughters [ 39 ]. Here, cells are assumed to undergo fission synchronously and daughter cells are chosen randomly at each generation. In the absence of selective effects, the offspring distribution is p 0 = 1/4, p 1 = 1/2, p 2 = 1/4, where p i is the probability of producing i offspring. The variance in offspring number here is 1/2 and the variance-effective population size equals N / = 2 N . Thus in this case, the diploid theory can be directly used as far as v(x) is concerned (replacing 2 N e with 2 N ). Johnson and Gerrish [ 39 ] consider alternative models. These alternatives are associated with different rates at which drift proceeds in a population, and would not affect the qualitative conclusions drawn here. Fixation probability Diffusion models of genetic drift have shown [ 11 ] that the probability of fixation of an allele at frequency p in a randomly mating diploid population of size N e (with 2 N e copies of the gene in question) is Therefore, using p = 1/ N rather than the usual p = l/(2 N e ) the probability of fixation of a mutant bacterial cell with selective advantage s is Note that when 4 N s >> 1, u ~ 4 s . This agrees with a result of Gerrish and Lenski [ 18 ], using a branching process model (rather than the diffusion model) to find the fixation probability under this same model of binary fission. See [ 39 ] for discussion of u for alternative models. Steady state distribution Let the mutation rates from any genotype to any other be equal ( μ ). As stated above, selection is additive. As shown by Wright [ 23 ], the steady state distribution of allele frequency is then given by the density function See also [ 37 , 38 ] for further details.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC518962.xml
549522
Passive and active ventricular elastances of the left ventricle
Background Description of the heart as a pump has been dominated by models based on elastance and compliance. Here, we are presenting a somewhat new concept of time-varying passive and active elastance. The mathematical basis of time-varying elastance of the ventricle is presented. We have defined elastance in terms of the relationship between ventricular pressure and volume, as: dP = EdV + VdE , where E includes passive (E p ) and active (E a ) elastance. By incorporating this concept in left ventricular (LV) models to simulate filling and systolic phases, we have obtained the time-varying expression for E a and the LV-volume dependent expression for E p . Methods and Results Using the patient's catheterization-ventriculogram data, the values of passive and active elastance are computed. E a is expressed as: ; E p is represented as: . E a is deemed to represent a measure of LV contractility. Hence, Peak dP/dt and ejection fraction (EF) are computed from the monitored data and used as the traditional measures of LV contractility. When our computed peak active elastance (E a,max ) is compared against these traditional indices by linear regression, a high degree of correlation is obtained. As regards E p , it constitutes a volume-dependent stiffness property of the LV, and is deemed to represent resistance-to-filling. Conclusions Passive and active ventricular elastance formulae can be evaluated from a single-beat P-V data by means of a simple-to-apply LV model. The active elastance (E a ) can be used to characterize the ventricle's contractile state, while passive elastance (E p ) can represent a measure of resistance-to-filling.
Background The heart may be conceived as a pump that receives blood from a low-pressure system and raises it to a high-pressure system. Although the mechanism responsible for generation of wall stress (and hence left ventricular (LV) pressure) is contraction of the myocardial fibers, an analytical formulation linking the myocardial and LV dynamics is still lacking. In the absence of this formulation, a popular way of linking LV pressure and volume dynamics is by means of LV compliance (or elastance) [ 1 , 2 ]. Although this yields the cyclic values of elastance and compliance, this concept does not provide an intrinsic measure of elastance and compliance for the contractile state of the LV. The concept of compliance or elastance was first employed for blood vessels [ 3 ], by relating incremental cross-section area (or volume) and transmural pressure. Warner appears to be have been the first to adopt a compliance description for a dynamic heart [ 4 ]. In Warner's description, a mean value of compliance is adopted for diastolic phase and another mean value for systole, with abrupt transitions between the two states. Defares [ 5 ] avoided the stepwise transition between diastolic and systolic compliance, by making elastance a continuously varying function of time. Later, the concept of a continuously varying compliance or elastance was adopted by a number of investigators with diverse variations [ 2 , 6 , 7 ]. Nevertheless, they all share the concept of a simple and extrinsic compliance term as an adequate description of ventricular mechanics during the cardiac cycle, based on monitored values of LV pressure and volume. Classically, ventricular compliance is defined, at any point in time, as the change in ventricular volume concomitant with the change in ventricular pressure, such that C = dV / dP , E = dP / dV (1) If C is assumed constant, this equation becomes a linear relation, whose integration gives V = CP + V c (2) where V c is an integration constant. If the compliance varies with time, then all terms in equation (2) may vary with time, as: V ( t ) - V c ( t ) = C ( t ) P ( t )     (3) In this context, Suga [ 8 ] opted for the definition of ventricular elastance as to represent the elastance of the contracting LV, where V d represents the unstressed LV volume corresponding to zero LV-pressure, obtained by drawing a tangent to P-V curves at end-ejection, as illustrated in Figure 1 . This model gave rise to the development of the end-systolic pressure-volume relation (ESPVR) as a measure of contractility [ 9 - 14 ]. Figure 1 Schematic drawing of P-V loops and end-systolic P-V relation (ESPVR) . Schematic drawing of P-V loops and end-systolic P-V relationship (ESPVR) with a positive volume intercept V d . The slope of ESPVR line is deemed to be E max . However, the determination of the maximal slope E max and of the volume-axis intercept (V d ) of the tangent to the P-V curve at end-ejection (as a measure contractility of the cardiac muscle) is not only unreliable [ 11 , 15 ], but also requires generation of multiple P-V loops under variations loading conditions [ 11 , 15 ]. It is hence impractical to use clinically for a specific LV catheterization-ventriculography data. Above all, all of these variations in the concept of E obtained from LV pressure-volume data fail to explain the phenomena of LV suction and LV pressure drop during early filling as well as the generation of LV pressure increase during isovolumic contraction. We have hence come up with a new concept of dual passive and active elastances operating throughout the cardiac cycle. The passive elastance (E p ) represents the LV pressure response to LV volume change (to LV volume increase during LV filling phase and to LV volume decrease during LV ejection phase). However, simultaneously, we also have active elastance (E a ) representing the contraction of the left ventricle due to its sarcomeric activation (and the development of force between the actin-myosin units) and relaxation (due to disengagement of the actin-myosin units). LV E a develops after the start of isovolumic contraction, becomes maximum some time during late ejection and thereafter decreases and becomes zero during diastolic filling. On the other hand LV E p starts increasing after the initiation of LV filling as the LV volume increases. It reaches its maximum value at the end-of-filling phase, remains constant during isovolumic contraction, and thereafter decreases during ejection (as the LV volume decreases). While the generation of E a helps us to explain the development of the LV pressure increase during isovolumic contraction, the decrease of E a during diastole helps us to explain the decrease in LV pressure during early filling. The incorporation of both E p and E a helps us to explain the LV pressure changes during the filling and ejection phases. Methods Data acquisition The subjects in this study were studied in a resting recumbent (baseline) state, after premedication with 100–500 mg of sodium pentobarbital by retrograde aortic catheterization. Left ventricular chamber pressure was measured by a pigtail catheter and Statham P23Eb pressure transducer; the pressure was recorded during ventriculography. Angiography was performed by injecting 30–36 ml of 75% sodium diatrizoate into the LV at 10 to 12 ml/s. It has been found by using biplane angiocardiograms that calculated orthogonal chamber diameters are nearly identical [ 16 ]. These findings are used to justify the use of single-plane cine techniques, which allow for beat-to-beat analysis of the chamber dimensions. For our study, monoplane cineangiocardiograms were recorded in a RAO 30° projection from a 9 in image intensifier using 35 mm film at 50 frames/s using INTEGRIS Allura 9 system at the Nation Heart Centre (NHC), Singapore. Automated LV analysis was carried out to calculate LV volume and myocardial wall thickness. The LV data, derived from the cineangiographic films and depicted in Figure 2 consists of measured volume and myocardial thickness of the chamber as well as the corresponding pressure. All measurements are corrected for geometric distortion due to the respective recordings systems. Figure 2 A case study of measured LV pressure, volume and wall thickness during a cardiac cycle . An example of a patient's measured LV pressure, volume and wall thickness during a cardiac cycle; t = 0-0.08s is the isovolumic contraction phase, t = 0.08s-0.32s is the ejection phase, t = 0.32s-0.40s is the isovolumic relaxation phase, and t = 0.40s-0.72s is the filling phase. Note that even after 0.4 s, the LV pressure still continues to decrease from 17 mmHg (at 0.4s, at start of filling) to 8 mmHg at 0.44s. In Figure 2 , it is noted that during the early filling phase, LV pressure decreases even though LV volume increases. This phenomenon is defined as the 'LV suction effect', which will be explained later by using our concepts of active and passive elastances. This phenomenon is also depicted in Figure 3 and Table 1 . Figure 3 Relationship between LV volume and pressure for the data of figure 2 . Relationship between LV volume and pressure for the data of Figure 1. Points (21–36) constitute the filling phase, (1–5) constitute the isovolumic contraction phase, (5–17) constitute the ejection phase, and (17–21) constitute the isovolumic relaxation phase. Note that after point 21, the LV pressure decreases; this characterizes LV suction effect. Table 1 Computed values of E a and E p during the cardiac cycle, for subject HEL Point Phase Time Pressure Volume E a E p 1 Isovolumic contraction 0 18 136.7 0 0.968314 2 0.02 22 135.7 0.051287 0.930811 3 0.04 32 134.6 0.153167 0.891234 4 0.06 52 133.5 0.282331 0.853339 5 0.08 80 132.5 0.424912 0.820289 6 Ejection 0.1 94 129 0.570602 0.714374 7 0.12 103 124.5 0.711775 0.598039 8 0.14 110 119.3 0.843116 0.486996 9 0.16 113 114 0.961299 0.395008 10 0.18 116 107.3 1.06464 0.303159 11 0.2 118 101.8 1.15275 0.243961 12 0.22 120 97.5 1.22616 0.205852 13 0.24 121 94 1.28607 0.179272 14 0.26 122 91 1.334 0.159239 15 0.28 120 88 1.37114 0.141444 16 0.3 116 85.5 1.34002 0.128144 17 0.32 100 84.3 1.15107 0.122212 18 Isovolumic relaxation 0.34 74 85 0.846411 0.125638 19 0.36 50 85.5 0.523931 0.128144 20 0.38 30 86.4 0.269515 0.132782 21 0.4 17 90.6 0.113989 0.156743 22 Filling 0.42 10 105 0.0392726 0.276831 23 0.44 8 112 0.0109316 0.365003 24 0.46 8.4 117 0.00244008 0.444703 25 0.48 9 119 0.00043378 0.481259 26 0.5 9.6 120.2 6.10272e-005 0.50462 27 0.52 10.2 121 6.75442e-006 0.520821 28 0.54 10.5 121.4 5.84848e-007 0.529115 29 0.56 10.7 121.6 3.94098e-008 0.533312 30 0.58 10.8 121.8 2.0564e-009 0.537542 31 0.6 11 122 8.2698e-011 0.541805 32 0.62 11.8 124 2.55156e-012 0.586344 33 0.64 12.8 127 6.01397e-014 0.66011 34 0.66 14.5 130.7 1.07837e-015 0.763991 35 0.68 17 134 1.46518e-017 0.87036 36 0.7 20 136.6 1.50271e-019 0.964497 37 0.72 18 136.7 1.15909e-021 0.968314 Definition of passive elastance and active elastance of the LV At the start of diastolic-filling phase, the LV incremental pressure dP LV is the response to (i) LV E a continuing to decrease due to the sarcomere continuing to relax well into the filling phase, and (ii) to the rapid inflow of blood and the corresponding increase in LV volume, along with increase in LV E p . The associated governing differential equation, relating LV pressure and volume, can be put down (by referring to the Appendix for its derivation) as [ 17 ]: where t represents the time variable (s) from the start of filling phase; V represents the volume of LV (ml) during the filling phase; P LV represents pressure of the LV, in mmHg (hereafter symbolized by P ) (mmHg); M represents the inertia term = [LV wall-density ( ρ )/(LV surface-area/wall-thickness)] = ρh / 4 πR 2 , for a spherical LV model (in mmHg/(ml/s 2 )); E represents LV elastance (mmHg/ml). Likewise during ejection, the LV pressure variation (dP LV ) is caused by both E a variation as well as E p decrease. The instantaneous time-varying ventricular elastance (E) is the sum of (i) volume-dependent passive elastance (E p ) and (ii) active elastance (E a ) due to the activation of the LV sarcomere. Hence, E = E a + E p (6) We will now provide the expressions for E p and E a , and then their formulations. The passive (unactivated) myocardium exhibits properties of an elastic material, developing an increasing stress as strain increases, as occurs during ventricular filling. The passive stress-strain relation of myocardial muscle strip is nonlinear, and therefore cannot be described by Hooke's law. As an approximation, it follows an exponential relationship [ 18 - 20 ]. Therefore, the relation between LV passive pressure and volume has also been assumed to be exponential. Since E p (= dP/dV) is volume-dependent, we can express it as: where E p 0 is the passive elastance coefficient, z p is the passive elastance exponent, and V is the LV volume; its derivation is provided in a subsequent section. On the other hand, we will represent E a as an intrinsic property of the LV (derived later), as: where (i) t is measured from the start of isovolumic contraction, (ii) the parameter E a 0 is the active elastance coefficient, (iii) the time-coefficient ( τ C ) describes the rate of elastance rise during the contraction phase, while ( τ R ) describes the rate of elastance fall during the relaxation phase; (iv) the exponents " Z C " and " Z R " are introduced to smoothen the curvatures of the E a curve during isovolumic contraction and relaxation phases; (v) the parameter d is a time constant whose (to be determined) value is during the ejection phase, and (vi) u(t-d) is the unit step function, u(t-d) = 0 for t<d. The rationale for equation (8), as provided in the next section, is based on E a incorporating parameters reflecting the (i) generation of LV pressure during isovolumic contraction, (ii) decrease of LV pressure during isovolumic relaxation and early filling, and (iii) the LV pressure-volume relationship during filling and ejection phase. Our hypothesis Based on equations (5–8), our hypothesis is that both E a and E p contribute to the relationship of LV pressure and volume. While E p incorporates LV pressure change due to LV volume change, E a incorporates the effect of LV myocardial activation in the generation of LV pressure during the isovolumic phases (when the LV volume remains constant). Since E a is deemed to be the basis of LV pressure generation, its variation (as given by equation 8) corresponds to the LV pressure variation. Determination of E a and E p expressions a) Active elastance (during isovolumic contraction and relaxation) During isovolumic contraction (because dV = 0, and E p is constant), the governing equation (5) becomes VdE = dP LV , which can be detailed as: V i ( E i - E i -1 ) = V i [( E a,i + E p,i ) - ( E a,i -1 + E p,i -1 )] = V i ( E a,i + E ped - E a,i -1 - E ped ) = dP LV,i (9) where i is a time instant during the isovolumic contraction and relaxation, V i and P LV,i are the monitored LV volume and pressure at this instant, and E ped is the passive elastance at the end-diastolic phase. Also, during isovolumic relaxation (because dV = 0, and E p is constant), the governing equation (5) becomes VdE = dP LV , which can be represented as: V i ( E a,i + E p,i - E a,i -1 - E p,i -1 ) = V i ( E a,i + E pes - E a,i -1 - E pes ) = dP LV,i (10) where E pes is the passive elastance at the end-systolic phase Now, applying equations (9 & 10) to the case shown in the Figure 2 , we have (using the monitored LV pressure-volume data): 1. For isovolumic contraction E a ,1 = 0     (11-a) E a ,2 = ( P 2 - P 1 )/ V 2 + E a ,1 = 0.029477 mmHg/ml     (11-b) E a ,3 = ( P 3 - P 2 )/ V 3 + E a ,2 = 0.103771 mmHg/ml     (11-c) E a ,4 = ( P 4 - P 3 )/ V 4 + E a ,3 = 0.253584 mmHg/ml     (11-d) E a ,5 = ( P 5 - P 4 )/ V 5 + E a ,4 = 0.463599 mmHg/ml     (11-e) 2. For isovolumic relaxation E a ,18 = ( P 18 - P 17 )/ V 18 + E a ,17 = E a ,17 - 0.058954 mmHg/ml     (11-f) E a ,19 = ( P 19 - P 18 )/ V 19 + E a ,18 = E a ,17 - 0.177824 mmHg/ml     (11-g) E a ,20 = ( P 20 - P 19 )/ V 20 + E a ,19 = E a ,17 - 0.312656 mmHg/ml     (11-h) E a ,21 = ( P 21 - P 20 )/ V 21 + E a ,20 = E a ,17 - 0.463599 mmHg/ml     (11-i) Now in the above expressions 11(f-i), E a,17 at end-ejection is unknown. For different representative values of E a,17 , we can get different E a curves. We need to determine the optimal value of E a,17 , such that E a can be described by a smooth curve to fit both isovolumic contraction and ejection phases. In Figure 4 , we have determined E a,17 = 1.1 mmHg/ml, and the polynomial expression for E a (t) to fit its above calculated values during isovolumic contraction and relaxation. However, in order to (i) more suitably represent E a to correspond with its role during the cardiac phases (isovolumic, ejection and filling), and (ii) because of the sigmoidal shape of E a curve and its variation resembling the LV pressure variation during systole (as seen in Figure 4 ), we express E a (according to equation 8) as: Figure 4 E a vs. time for the data of figure 2. Points (1–5) are the computed values of E a during isovolumic contraction phase, based on equations 11(a-e). In the isovolumic relaxation phase, the computed values are represented by the symbol Δ, for E a ,17 = 1.1 mmHg/ml . The best fit for E a during isovolumic contraction and relaxation phase is given by the curve: E a = -12000 t 6 + 17000 t 5 - 7700 t 4 + 1100 t 3 + 19 t 2 + 0.59 t + 0.00056 So that its constituent parameters have physiological significance as indicated following equation (8). Hence, to compute E a (s) during isovolumic contraction, when u(t-d) = 0, we employ the expression and determine its parameters E a0 , τ C and Z C to fit the monitored pressure-volume data. Then, to compute E a (s) during isovolumic relaxation, we employ the total expression and determine its remaining parameters d, τ R and Z R to fit the measured pressure-volume data. For the sample data of Figure 2 , the variation of E a is depicted in Figure 5 , along with the values of its parameters. We now propose that E a can be employed as an index of contractility. Figure 5 The data of figure 2 is fitted with equation (8). When the patient data of Figure 2 is fitted with equation (8), the resulting parameters values are obtained as: E a 0 = 1.48 mmHg/ml, τ C = 0.1555 s, Z C = 1.631, d = 0.28 s, τ R = 0.07935 s, Z R = 2.267 s, E a ,17 = 1.1 mmHg/ml, RMS = 0.026 mmHg/ml. b) Passive elastance determination during diastolic filling During the diastolic filling phase, equation (5) becomes Now because E p is constant at a particular volume V i , equation (13) becomes where i is a time-instant during diastolic filling, V i and P LV,i are the monitored LV volume and pressure at this time, and M = ρh / 4 πR 2 . For the patient data (shown in Figure 2 ), we can get the mean value for M during diastolic filling, M = 8.03 × 10 -6 mmHg/(ml/s 2 ). Therefore, from equation (15), we can calculate the values of E p at various instants during filling phase. We then plot E p vs V, in Figure 6 . By fitting equation (7) to these calculated values of E p , we obtain the values of the parameters E p 0 and z p , as: Figure 6 Passive elastance E p vs LV volume for the data of figure 2 . Passive elastance E p vs LV volume V, for the sample case shown in Figure 2. z p = 0.0395 ml -1 , E p 0 = 4.375 × 10 -3 mmHg / ml (16) and the E p function (corresponding to its expression given by equation 7) as follows: E p = 4.375 × 10 -3 e 0.0395 V (17) We now propose to adopt E p as a measure of LV resistance-to-filling. Hence during ejection, both E a and E p are varying. During ejection and filling phases, E p can be calculated at any time using equation (17). Likewise, E a can be calculated during ejection and filling phases using equation (8), once its parameters have been determined by employing equation (12-b & 12-b) during isovolumic contraction and relaxation phases. Their values during the cardiac cycle are listed in Table 1 . Clinical application results The analyses, presented herewith, are now applied to clinically obtained data consisting of the subject's left ventricular (instant-to-instant) dimensions (obtained by cineangiocardiograph) and chamber pressure (obtained by cardiac catheterization). For each subject, passive and active elastances are determined from the left ventricular data. Table 2 provides the measured data and the model-derived parameters for three subjects (subject HEL, DDM, and ML). Subject HEL serves as a sample patient with myocardial infarct, subject DDM with double vessel disease (DVD) and hypertension, treated with PTCA; subject ML with idiopathic myocardial hypertrophy (IMH). Table 2 Clinical history, measured hemodynamic data and calculated passive and active elastance parameters (E p and E a ) for subjects (HEL, DDM and ML). Where LVP: left ventricle chamber pressure, AOP: aortic pressure, EDV: end-diastolic volume, ESV: end-systolic volume, EF: ejection fraction, MI: myocardial infarct, DVD: double vessel disease, HTN: hypertension, IMH: idiopathic myocardial hypertrophy, E a ,max : maximum active elastance Subject H.E.L D.D.M M.L Disease MI, DVD DVD, HTN IMH LVP (mmHg) 122/18 170/24 109/12 AOP (mmHg) 125/75 169/99 115/70 EDV/ESV (ml) 132.5/84.3 121.7/41.3 368/284 EF 0.36 0.66 0.23 E p 0 (mmHg/ml) 4.375 × 10 -3 6.74 × 10 -5 1.442 × 10 -8 z p (ml -1 ) 0.0395 0.07499 0.05024 E a 0 (mmHg/ml) 1.48 4.4 0.595 τ C (s) 0.1555 0.207 0.1082 Z C 1.631 1.536 1.977 d (s) 0.28 0.26 0.18 τ R (s) 0.07935 0.1536 0.1377 Z R 2.267 2.943 1.873 E a ,max (mmHg/ml) 1.37 3.58 0.57 dP / dt max (mmHg/s) 1200 1475 1125 The variations of model-derived nonlinear passive and active elastances for the subject HEL are shown in Figure 7 . For this particular subject (HEL), the maximum active elastance is 1.37 mmHg/ml. In Figure 8 , we have plotted E a vs incremental pressure (P-P ed ) for this patient HEL. Note that the elastance is much higher at late-ejection than early ejection. This is because of a continuing sarcomere stress development and shortening. The active elastance reaches its maximum value at late-ejection (point 15), and thereafter decreases. However as shown in Figure 7 , even after the end of relaxation phase (point 21) the active elastance continue to decrease into the filling phase. This decrease can explain the suction effect during the rapid filling sub-phase, even after LV filling has commenced. Figure 7 Pressure, active elastance E a , and passive elastance E p and total E = E a + E p for the data of figure 2. Pressure, active elastance E a , passive elastance E p , and total E = ( E a + E p ) for the sample subject shown in Figure 2. In this figure, 1–5 represents the isovolumic contraction phase, 5–17 represents the ejection phase, and 17–21 represents the isovolumic relaxation phase, 21–37 represents the diastolic filling phase. Figure 8 Active elastance vs. incremental pressure. Active elastance vs. incremental pressure (P-P ed ) for the same subject as shown in Figure 2. The arrow direction indicates progression of time; 1–5: isovolumic contraction phase; 5–17: ejection phase. Note the rapid decrease in E a during the isovolumic relaxation that also extends into the filling phase, and causes suction of blood into the LV even before initiation of left-atrial contraction. Figures 9 and 10 provide representations of the nonlinear passive and active elastance for subject HEL (with hypertension), DDM (myocardial infarct), and ML (idiopathic myocardial hypertrophy). The E p vs LV volume plots, in Figure 9 , clearly reveal that E p increases exponentially with increase LV volume; the parameters ( E P 0 and z p ) characterize this relationship. The passive elastance curve is steeper for a stiffer myocardium, with a corresponding bigger value of the exponential coefficient Z p (subject ML). The E a vs incremental pressure (P-P ed ) plots, in Figure 10 , reveal the development and decrease of E a during systole, which in turn governs the generation of LV pressure. Figure 9 LV volume and corresponding volume-dependent passive elastance. Volume-dependent passive elastance (E p ) for subjects HEL, DDM, and ML. Figure 10 Active elastance vs. incremental pressure. Active elastance vs. incremental pressure for subjects HEL, DDM, and ML. Discussion Ea as a contractility index Yet another way to study E a variation is by means of the E a vs normalized time (t/t s ) plot, shown in Figure 11 . These two figures 10 and 11 make us realize that E a,max could be regarded as an index of LV contractility. Hence, we decided to plot E a,max vs the traditional contractility indices of EF and (dP/dt) max . These plots are displayed in Figures 12 and 13 . It is noted that E a,max has a high degree of correlation with both EF and (dP/dt) max . It is interesting to compare our correlation-coefficient value (0.8972) with the value of 0.89 obtained by Mehmel et al [ 12 ], although this paper computes elastance as an extrinsic property = [P/(V-V d )] es . Figure 11 Active elastance vs. normalized time. Active elastance vs normalized time (t/t s ) for subjects HEL, DDM, and ML. Herein, t s is the duration from start-of-isovolumic contraction phase to end-of-isovolumic relaxation. Figure 12 E a,max vs EF. Relating our contractility index E a,max to EF factor, with r being the correlation coefficient. Figure 13 E a,max vs dP/dt max . Relating our contractility index E a,max to the traditional contractility index dP/dt max , with r being the correlation coefficient. Demonstrating LV suction phenomenon The active elastance curve can explain some critical LV physiological phenomena, namely LV pressure generation during isovolumic contraction and LV suction during early filling. The rapid decrease in elastance during isovolumic relaxation extends into the filling phase, and can explain the decrease in LV pressure (in Figure 7 ) even after LV filling has commenced. Let us show how this happens, by rewriting equation (15) as follows: P i - P i-1 = (E p,i + E a,i )(V i - V i-1 ) + V i (E a,i - E a,i-1 )     (18) by neglecting the term, based on the calculation of its value being of the order of 10 -2 compared to (i) ( E p,i + E a,i )( V i - V i -1 ), which is of the order of 10 0 and (ii) V i ( E a,i - E a,i -1 ) which is the order of 10 1 . In equation (18), it is seen that P i can be less than P i-1 (or that P i - P i-1 < 0) only if (E a,i - E a,i-1 ) is negative, i.e., active elastance is decreasing. Now for subject H.E.L (Figure 1 ), the computed values of E a and E p at these 2 instants (based on equations 8 & 17) are: E p ,22 = 0.2768 mmHg/ml , E a ,22 = 0.0393 mmHg/ml , E a ,21 = 0.1140 mmHg/ml , V 22 = 105 ml , V 21 = 90.6 ml (19) Substituting these values into equation (18) gives ( P 22 - P 21 ) = -3.22 mmHg , confirming decrease of pressure during early filling. Hence, our novel concept of "decreasing E a during the early phase of filling" enables us to explain the phenomenon of decreasing LV pressure during the early stage of filling. In other words, it is suggested that the sarcomere actin-myosin activity continues into the filling phase. The decreasing E a during the filling phase seems to reflect decreasing sarcomeric activity during filling. Likewise, the increase in E a during isovolumic contraction is responsible for increase in LV pressure at constant volume, as demonstrated by means of equation (11). Demonstrating variation of LV pressure during ejection in terms of E a and E p Similarly, both the active and passive elastances can explain LV pressure variation during ejection, using equation (18). Let us show how this happens by just taking two instants (t 10 & t 9 ), as follows: P 10 - P 9 = (E p,10 + E a,10 )(V 10 - V 9 ) + V 10 (E a,10 - E a,9 )     (20) Substituting these computed values (listed in Table 1 ) into equation (20) gives P 10 - P 9 = 1.92 mmHg , which is approximately equal to the actual value of 2 mmHg. Concluding comments Previous works have described elastance as a derived parameter from LV pressure-volume data, based on the definition of P / ( V - V d ) [ 9 , 10 , 12 - 14 ]. Also, E max (given by ESPVR) as illustrated on Figure 1 varies with different arterial loading states and is therefore load dependent [ 21 ], and not an intrinsic index independent of LV loading states. Our definitions of E p and E a enable us to explain the phenomena of (i) LV suction during early filing, (ii) LV pressure rise during isovolumic contraction (iii) LV pressure variation during the ejection phase, and (iv) LV pressure drop during the relaxation phase. Both the concepts of E a and E p are made possible by our redefining elastance and compliance as dP = d(EV) = d(V/C) = VdE + EdV (21) Our concept of E a enables us to explain (i) the generation of pressure during isovolumic contraction and decrease of pressure during isovolumic relaxation, when the volume is constant, and (ii) the decrease of pressure during the rapid filling phase. Our concept of decreasing E a along with increasing E p during filling, which enables us to explain the LV suction effect during early filling, is indirectly supported by the work of Shoucri [ 22 ]. He has indicated that the LV wall stress during the filling phase is made up of two components: (i) a passive increasing component due to LV pressure and (ii) an active decreasing component due to decreasing active fiber stress. However, the determination of this active fiber stress in his paper is empirical and not based on a well-defined concept and expression for E a , as provided by us. Finally, both E a and E p are invoked to explain the variations of pressure during the ejection and filling phases. From the viewpoint of intrinsic indices of LV assessment, E p can represent LV myocardial stiffness property and resistance to LV filling. On the other hand, E a has been shown to correspond to LV contractility, by means of Figures 12 & 13. Therein, we have shown a high degree of correlation between E a,max and (dP/dt) max as well as EF. In future, we can also couple our LV model (of E p and E a ) with an arterial lumped-parameter model (consisting of total peripheral resistance R, total arterial compliance C, aortic characteristic impedance Z 0 and inertial L) in order to simulate and explain the mechanisms of chronic hypertensive states [ 23 ], in terms of alteration in E a itself as a measure of LV adaptation to chronic hypertension induced in the circulation system. Appendix: Derivation of Equation (5) Dynamic equilibrium of the LV myocardial element gives (based on figure 14 ): Figure 14 Dynamic equilibrium of a myocardial element. Dynamic equilibrium of a myocardial element. Element mass m e = ρ (2 r 1 dθ 1 )(2 r 2 dθ 2 ) h = ( ρh )(4 r 1 r 2 dθ 1 θ 2 ) = ρhA e = m s A e ; dσ i = ( dP el )( r i ) / (2 h ), (i = 1, 2); dP el is the incremental elastic-recoil pressure. For dynamic equilibrium of the myocardial element, m e + 2( dσ 1 )(2 r 2 dθ 2 h ) + 2( dσ 2 )(2 r 1 dθ 1 h ) - A e dP LV = 0, or m e + A e dP el - A e dP LV = 0. For our LV spherical geometry model, r 1 = r 2 and dσ 1 = dσ 2 . where the myocardial element mass, m e = m s A e , m s (the myocardial surface-density or mass per unit surface area) = ρh , ρ is the myocardial density, and u is the radial displacement, dP el and dP LV are the incremental elastic-recoil and left ventricular pressures (as depicted in Figure A-1). Now since, By assuming r 1 = r 2 = R we can write: where ρ s is the surface density, M = ρ s h , and V is the LV volume. Now, referring to Figure A-1, and where R e is the resistance to LV filling (through the open mitral-valve), ρ f is the blood density and v is the blood velocity at site 2. Let us now define incremental elastic recoil pressure (in response to incremental LV pressure dP LV ) as: Hence, from equations (A-1, A-4, A-7 & A-8), we have Introducing the term elastance (= 1/C), we can put down Authors' contributions Liang Zhong carried out the elastance studies, participated in data acquisition, performed statistical analysis and drafted the manuscript. Dhanjoo N. Ghista conceived of the study, and participated in its design and coordination and helped to draft the manuscript. Eddie Y-K Ng participated in coordination and helped to draft the manuscript. Soo T Lim participated in its design and coordination. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549522.xml
506786
Cancer of the endometrium: current aspects of diagnostics and treatment
Background Endometrial cancer represents a tumor entity with a great variation in its incidence throughout the world (range 1 to 25). This suggests enormous possibilities of cancer prevention due to the fact that the incidence is very much endocrine-related, chiefly with obesity, and thus most frequent in the developed world. As far as treatment is concerned, it is generally accepted that surgery represents the first choice of treatment. However, several recommendations seem reasonable especially with lymphadenectomy, even though they are not based on evidence. All high-risk cases are generally recommended for radiotherapy. Methods A literature search of the Medline was carried out for all articles on endometrial carcinoma related to diagnosis and treatment. The articles were systematically reviewed and were categorized into incidence, etiology, precancerosis, early diagnosis, classification, staging, prevention, and treatment. The article is organized into several similar subheadings. Conclusions In spite of the overall good prognosis during the early stages of the disease, the survival is poor in advanced stages or recurrences. Diagnostic measures are very well able to detect asymptomatic recurrences. These only seem justified if patients' chances are likely to improve, otherwise such measures increases costs as well as decrease the patients' quality of life. To date neither current nor improved concepts of endocrine treatment or chemotherapy have been able to substantially increase patients' chances of survival. Therefore, newer concepts into the use of antibodies e.g. trastuzumab in HER2-overexpressing tumors and the newer endocrine compounds will need to be investigated. Furthermore, it would seem highly desirable if future studies were to identify valid criteria for an individualized management, thereby maximizing the benefits and minimizing the risks.
Incidence and mortality Nearly 170,000 new cases of endometrial carcinomas were estimated worldwide in 1997 [ 1 ]. However, incidences throughout different regions of the world vary considerably. Compared to Africa and Asia having the lowest rates of incidence, Western Europe, USA and Canada are shown to have the highest incidence worldwide (Figure 1 ). As shown in figure 2 , even within Europe the incidence rates are very heterogeneous. In some of these countries, e.g. Germany, endometrial carcinoma is the most common among genital carcinoma [ 2 ]. Figure 1 Estimated endometrial cancer incidences throughout different regions of the world [1]. Figure 2 Estimated endometrial cancer incidences throughout different countries in Europe [1]. Endometrial carcinomas occur in advanced age (postmenopausal). The age-related incidence for Germany is shown in Figure 3 . The overall increase in the incidence of this disease during the last decades is mainly related to higher life expectancy within the developed world. Figure 3 Age dependent incidence of endometrial cancer in Germany. Robert-Koch-Institute at: Etiology and risk factors The etiology of the endometrial carcinoma is not fully understood. Most cases appear sporadically whereas about 10% are hereditary. Chief among the latter is the autosomal dominantly inherited hereditary non-polyposis colorectal cancer (HNPCC) [ 3 - 5 ]. The risk of developing endometrial cancer is believed to be ten times higher for women carrying the gene compared to the general population [ 6 ]. The likelihood of a synchronous or metachronous development of endometrial carcinomas is, however, higher for patients with breast, ovarian, and non-hereditary colorectal cancer [ 7 ]. Within the current concept of multi-step progression of normal cells to malignancy, recent molecular work has identified several gene alterations important for tumor development. In summary, mutations and amplifications of oncogenes K-ras and HER2/neu, mutations or deletions of tumor suppressor genes p53, p21, p16, and pTEN/MMAC1 as well as impaired DNA repair functions through mutations of hMLH1, hMSH2, and hMSH6 have been connected with the development of endometrial carcinomas [ 8 ]. Environmental, dietary and hormonal factors as well as an aging female population have been attributed to an observed increase of endometrial carcinomas over the past few years. Epidemiologic studies have observed correlations between the incidence of endometrial cancer and the usage of estrogens, especially when applied to alleviate perimenopausal and postmenopausal symptoms. Therefore, it appears that estrogen plays a key role in the development and progression of endometrial carcinomas. There is also convincing evidence that high body mass increases the risk of endometrial carcinomas. Current estimations figure that about 40% of these relate to excess body weight [ 9 ]. A plausible biological explanation for obesity influencing the risk of endometrial cancer is found in increased aromatization of androstendione to estrone in adipose tissue [ 10 ]. Hyperadrenocorticism, which is more common in obese individuals, also disturbs the estrogen metabolism. The correlation between obesity and hyperadrenocorticism is possibly increased by hyperinsulinism. This also explains the higher frequency of endometrial carcinomas in combination with diabetes mellitus. A high intake of saturated fat may also increase the risk, whereas high consumption of vegetables and fruits may do the opposite [ 1 , 11 ]. Due to the fact that most factors relate to prolonged or intensive exposure to estrogen (hyperestrogenism), this may likely be identified as the underlying concept leading to the development of endometrial cancer. In this process estrogens are believed not to act as carcinogenic agents, but as promoters of carcinogenesis. A state of hyperestrogenism may be caused by exogenous or endogenous factors and may relate to reproductive factors, estrogen exposure or menopausal years [ 9 , 12 , 13 ]. Several endogenous risk factors are also associated with the increase of endometrial cancer risk e.g. early menarche, late menopause (2.4-fold), nulliparity (2-fold), the polycystic ovarian syndrome, diabetes mellitus (2.8-fold), high blood pressure (1.5-fold), obesity (up to 10-fold), other tumors with estrogen production, atypical endometrial hyperplasia (see Table 1 ) and the aforementioned inherited forms of colorectal cancer (10-fold). Basically two hypotheses endeavor to explain the protective effect of high parity: The first one assuming the mechanical removal of premalignant and malignant cells with each delivery, the second describing a protective effect of high progesterone values during pregnancy. It is most likely for both hypotheses to be correct to some extent, whereas a large population based study from Finland favors the second hypothesis [ 14 ]. An additional factor may be anovulatory ovarian insufficiency, especially in patients with polycystic ovaries (PCO syndrome). As these patients have a persistent progesterone deficiency due to non-formation of a corpus luteum, they lack this important protective mechanism. Endocrine risk factors also include hormone producing ovarian tumors. This goes for estrogen as well as androgen producing tumors since androgens may be converted to estrone in adipose tissue. Table 1 Type of endometrial hyperplasia and rate of progression to cancer [20]. Type of hyperplasia Rate of progression Simple (cystic without atypia) 1 Complex (adenomatous without atypia) 3 Atypical Simple (cystic with atypia) 8 Complex (adenomatous with atypia) 29 Previous irradiation of the pelvis, estrogen replacement therapy (HRT) (especially unopposed HRT) and tamoxifen therapy represent exogenous risk factors. Prolonged and exclusive intake of synthetic estrogens is associated with an up to 15-fold higher incidence of endometrial carcinomas [ 15 ]. Similarly, some selective estrogen receptor modulators (SERMs) and especially tamoxifen may have an estrogenic, proliferative effect on the endometrium. Thus tamoxifen therapy is frequently associated with polyps, hyperplasia and carcinomas of the endometrium [ 16 , 17 ]. Tamoxifen seems to induce a 6.4-fold increase in the risk for endometrial cancer [ 18 ]. On the other hand, cyclic application of estrogens combined with progestins does not increase endometrial cancer incidence, e.g. hormonal contraception (combined preparations) and may even reduces the risk of endometrial carcinomas by up to 50% (see prevention). In essence, hyperestrogenism will lead to stronger proliferation, thereby causing hyperplasia which may gradually acquire more and more cellular atypia (atypical hyperplasia) and later transform into an endometrioid adenocarcinoma. Recently a case-control study identified antipsychotic drugs as being a great risk factor for endometrial cancer [ 19 ]. These findings may also be explained to some extent by the side effects of the drugs (obesity, insulin resistance, amenorrhea, low gonadal steroids leading to hyperprolactinemia). Precancerosis In some cases endometrial cancer develops from atypical endometrial hyperplasia. The likelihood of this happening correlates with the degree of hyperplasia (Table 1 ) [ 20 ]. According to the World Health Organization (WHO) classification, endometrioid carcinomas are divided into simple and complex forms, each with and without atypia [ 21 ]. This classification has also been accepted by the International Society of Gynecological Pathologists. Endometrial hyperplasia is regarded as a preliminary stage of endometrioid carcinomas. Serous and clear cell carcinomas are on the other hand frequently associated with an atrophic endometrium [ 22 ]. A precursor of the serous carcinoma and possibly of some clear cell carcinomas is the endometrial intraepithelial carcinoma (EIC) [ 22 , 23 ]. Prevention To prevent an outbreak of the disease in menopausal and postmenopausal women long-term estrogen replacement therapy (treatment of menopausal symptoms, osteoporosis etc.) should be supplemented by intermittent application of gestagens. A recent study suggested the use of intrauterine devices (IUD) also possibly reducing the risk for endometrial cancer due to improved elimination or decrease of hyperplastic endometrial cells [ 24 ]. Numerous studies have shown that cigarette smoke reduces the risk of an endometrial carcinoma for women after the menopause, although it may increase the risk for premenopausal women [ 25 ]. The greatest reduction of risk was found in obese, multiparous women who did not receive hormone treatment [ 26 ]. Women in advanced stages of disease (stages II – IV), however, were found more likely to smoke than women in early stages (0 – I). This may reflect a smoking-related decrease in the incidence of early-stage tumors as well as an increase in tumor invasiveness and metastases [ 27 ]. The risk of endometrial carcinomas may be reduced significantly by prolonged progestin therapy every month (for 10 days) alone or in combination with estrogen [ 28 ]. Since progestins are known to act as cofactors of cancerization in breast and cervical cancer such concepts are better interpreted cautiously [ 29 ]. Therefore hormonal preventive concepts need to undergo a general assessment of benefits and risks. To sum up, apart from excess body fat reduction and omission of unnecessary estrogen therapy, there appears to be no reasonable way of preventing endometrial carcinomas. Early detection To date there is no procedure that seems appropriate as a screening method for early detection of endometrial carcinomas. Current guidelines of the American Cancer Society suggest informing patients of risks and symptoms involved with endometrial cancer and furthermore firmly emphasize the importance of reporting unexpected bleeding or spots to their physician [ 30 ]. The fact of most endometrial carcinomas (with the exception of the rare serous and clear cell types) showing these kinds of preliminary symptoms leads to the diagnosis of over 75% of cases while still in stage I [ 7 ]. A recent case-report on the usage of the Mirena ® intrauterine system maintains that irregular menstrual bleeding should not be treated simply with this system without prior diagnostic [ 31 ]. The necessity of finding a screening method is discussed controversially in this context. So far, most epidemiological studies have failed to show significant effects of screening on mortality. In some cases the Pap smear may lead to the diagnosis of endometrial cancer. However, in many cases cells from inside the uterus are not assessed by the sampling procedure. Positive cervical cytology was found to correlate with nodal spread in 91% of cases, whereas the risk of lymph node spread in patients with normal cervical cytology was estimated at around 2% [ 32 ]. It would seem too early to suggest that this could help to reach a decision on the necessity of lymphadenectomy. Maybe the ThinPrep Pap tests will be able to allow further conclusions in the future [ 33 ]. Transvaginal ultrasound has also been suggested as a potential means of early detection of endometrial carcinomas. A recent meta-analysis involving 9,031 patients and covering 57 separate studies on the diagnostic accuracy and positive predictive power of endometrial thickness measurement by pelvic ultrasound in patients with postmenopausal bleeding concluded that these measurements cannot solely be used to accurately rule out endometrial pathology. Measurement of both endometrial layers of ≤ 5 mm coincides with endometrial pathology in only 2.5% [ 34 ]. Other studies have used saline infusion sonography and color Doppler sonography to differentiate between endometrial cancer, endometrial hyperplasia, fibroids, endometriosis, myoma or tamoxifen induced endometrial thickness [ 35 - 42 ]. A recent comparison of saline infusion sonography and office hysteroscopy revealed similar ratings of patients' pelvic pain during the procedures. Sensitivity and specificity coefficients as well as negative and positive predictive values were higher for the office hysteroscopy [ 43 ]. It seems that biopsy remains the only accurate way of diagnosing endometrial cancer [ 44 ]. Under optimal circumstances the gynecologist will remove a tissue sample from the uterine lining under hysteroscopic control [ 44 - 47 ]. If hysteroscopic control is neglected, the false negative rates for dilatation and curettage (D&C) will range between two and six percent, thereby emphasizing the limitations of D&C alone [ 45 , 48 - 50 ]. Newer methods, like magnetic resonance imaging (MRI), positron emission tomography (PET), intraoperative ultrasound or three dimensional sonography are not likely to gain importance with respect to the diagnosis or early diagnosis of the disease [ 51 - 56 ]. It is, however, reasonable to believe that they are likely to deliver more information about the invasion depth of the myometrium or lymphatic metastases [ 52 , 55 , 57 ]. Classification of endometrial carcinoma The term endometrial carcinoma describes a variety of different tumor types originating from the inner ling of the uterus. Many authors differentiate between two basic types, which may be divided into estrogen-dependent and estrogen-independent types or tumors with favorable or unfavorable prognosis [ 58 ]. Although there are no cross-sectional studies comparing tumors from various ethnic groups and significant differences in tumor biology the frequency of tumors of a certain category in various geographical areas have been assumed. This may explain why mortality in the USA is higher in black women than in white women (5.8 deaths per 100,000 persons in black women vs. 3.1 deaths per 100,000 persons in white women) [ 59 ]. The different categories may be summarized as follows: 1. It comprises of estrogen related tumors occurring in younger, perimenopausal women. These tumors are often said to be highly differentiated, mainly adenocarcinomas with positive steroid hormone receptor status (ER, PR), the known risk factors (estrogen etc.) originating from atypical endometrial hyperplasia. The patients in this group have a longer history, lower grade tumors, less myometrial invasion and low potential for lymphatic spread. They may be associated with concomitant carcinomas of the ovary, breast and colon and respond to progestin therapy. The overall prognosis is generally favorable. 2. This category is comprised of tumors of patients with shorter history, higher grade tumors, deeper myometrial invasion and high risk of lymphatic spread. The tumors will not respond to progestin therapy and there are no associated tumors. Histologically these tumors are identified as serous carcinomas or clear cell adenocarcinomas. The prognosis of patients with this type of tumor is poor. 3. Findings of carcinomas in atrophic endometrium being associated with an intermediate prognosis lead to the suggestion of a third category. In these cases the endometrioid carcinomas are not likely to be estrogen-related [ 21 ]. 4. It was furthermore suggested to incorporate endometrial neoplasia originating from an inherited predisposition into a fourth category. These types of tumors tend to develop about 15 years earlier and are associated with a favorable prognosis. Most patients in this category have a hereditary non-polyposis colorectal carcinoma syndrome (HNPCC). Histological types according to the WHO classification The following section will briefly describe the most important characteristics of endometrial carcinomas. Detailed descriptions may be found in specialized pathological textbooks e.g. Anderson et al. [ 21 ]. • Endometrioid adenocarcinoma : The endometrioid adenocarcinoma, whose glands resemble those of the normal endometrium, is the most common type (60 – 80%). This type is considered part of the 1 st category • Endometrioid adenocarcinoma with squamous cell differentiation (adenoacanthoma, adenosquamous carcinoma): Approximately 25% of endometrioid adenocarcinomas display partial squamous differentiation. The squamous elements are interpreted as terminally differentiated indicating that the tumor is incapable of independent growth. Prognosis depends on the glandular components of the lesion. Highly differentiated tumors (adenoacanthomata) have a favorable prognosis, whereas poorly differentiated tumors (adenosquamous carcinomas) have an unfavorable course. • Serous adenocarcinoma : This tumor type is similar to the serous ovarian carcinoma. It is characterized by aggressive growth and poor prognosis. Lymphogenic and hematogenic metastases are usually already present at the time of diagnosis. Nearly all tumors are poorly differentiated. Serous adenocarcinomas belong to the 2 nd category of endometrial cancer. • Clear cell adenocarcinoma : This type comprises 3 to 6% of all endometrial carcinomas. Like the serous adenocarcinoma, clear cell adenocarcinomas tend to progress rapidly. They share the 2 nd category. • Mucinous adenocarcinoma : Diagnosis is based on the presence of mucus within the tumor cells. Purely mucinous carcinomas are rare, although a mucinous component within endometrioid carcinomas is more common. The tumors are usually highly differentiated and have a good prognosis. It is important to exclude a primary mucinous adenocarcinoma originating from the endocervix, spreading into the uterine body. Mucinous adenocarcinomas are part of the 1 st category. • Squamous cell carcinoma (2 nd category): A very rare entity associated with very poor survival. Primary squamous cell carcinomas of the uterine cervix should be ruled out. Adjuvant platinum-based radiochemotherapy may result in improved survival. • Mixed carcinoma : A carcinoma composed of two or more different non-squamocellular components with each component occupying at least 10% of the tumor. Prognosis varies, e.g. in case of a serous component, the prognosis is poor. • Undifferentiated carcinoma : A rare carcinoma without glandular, squamous or sarcomatous differentiation. Prognosis is unfavorable. • Rare types of endometrial carcinomas: small cell carcinomas, microglandular adenocarcinomas, signet-ring cell carcinomas, transitional cell carcinomas, glassy cell carcinomas, mucinous adenocarcinomas of the intestinal type, lymphoepithelioma-like carcinomas, and endometrial adenocarcinomas with trophoblastic differentiation have been reported. Grading The tumor grading is a highly significant prognostic parameter although it is subjective with a considerable inter-observer and intra-observer variability. It is determined by the percentage of non-squamocellular, solid portions as follows: G1: 5% and less, G2: 5–50%, G3: more than 50%. A significant nuclear atypia increases the grade of differentiation by one grade. Preoperative diagnostic procedures After histological confirmation of an endometrial carcinoma clinical palpation and vaginal sonography should be performed. Several additional examinations have been suggested: Rectoscopy, cystoscopy, computed tomography and/or MRI. They may be omitted in clearly diagnosed early cases but are strongly recommended in advanced stages. Recent studies indicate some value of tumor marker diagnostics. At a cut-off level of 40 U/ml elevated CA125 serum levels indicate nodal metastases with a sensitivity of 77.8% and a specificity of 81.1% [ 60 , 61 ]. Surgical therapy In all stages of endometrial carcinomas, surgery is the primary treatment of choice. Preoperative intracavitary radiation treatment, often recommended in earlier times, is not considered advisable any more since the information about the depth of myometrial invasion and thus information on an important prognostic factors is lost [ 62 ]. A study indicates that the Maylard-type incision is superior to transverse (Pfannenstiel-type) or longitudinal incision [ 63 ]. If possible, an abdominal hysterectomy with removal of the adnexa and a peritoneal lavage should be performed. After removal of the uterus, the depth of tumor invasion into the myometrium has to be determined to estimate the probability of extrauterine spread. If the myometrium is infiltrated to more than 50%, a pelvic and paraaortic lymphadenectomy should be performed. For appropriate staging, more than 20 lymph nodes should be dissected [ 64 ]. But also in cases of additional adverse prognostic factors (poor grading, lymphangiosis, see below), pelvic and paraaortic lymphadenectomy are recommended by many [ 65 ]. A decision tree on primary therapy is given in figure 4 . Unfortunately, studies found no correlation between depth of invasion, histological grade, cervical invasion, peritoneal cytology, menopausal status, preoperative serum CA125 level or primary tumor diameter. Only lymphvascular space involvement (P < 0.0001) was significantly correlated to pelvic lymph node metastases, which lead the authors to the conclusion that all patients should undergo extended surgical staging, except when clinical or operative factors increase patients' morbidity [ 66 ]. Lymphadenectomy may be omitted in cases of more favorable prognosis. Figure 4 Decision tree concerning the primary treatment of endometrial carcinoma.Legend: RT = radiotherapy; HT = endocrine treatment; CHT = chemotherapy; D&C = dilatation and curettage There is some controversy on the value of radical hysterectomy in stage II carcinomas. While several earlier studies advocated this procedure [ 67 - 70 ] a recent study reported no prognostic advantage [ 71 ]. Maybe only the patients with stage IIb tumors and clinically evident tumor infiltration profit from radical hysterectomy. This requires further investigation. Complications after pelvic lymphadenectomy may be reduced by omentoplasty and omentopexy [ 72 ]. In advanced stages of the disease complete removal of all tumor sites is warranted. In case of serous histology and peritoneal spread, some authors also advocate omentectomy. A recent study indicates that optimal cytoreduction results in a significant survival benefit for stage IVB endometrial cancer patients with a reasonable surgical morbidity rate [ 73 ]. Vaginal hysterectomy as primary treatment of endometrial cancer has also been investigated especially in medically compromised women [ 72 , 73 ]. Such approaches may be combined successfully with laparoscopically assisted radical vaginal hysterectomy [ 74 - 77 ]. The aforementioned basic goals of surgery (hysterectomy, removal of the adnexa and lymphadenectomy in stages Ic and higher) should be reached especially in medically fit patients, because treatment along the recognized guidelines has been found to be prognostically favorable. Some studies, however, have reported the standards mentioned in many cases as not having been realized [ 78 - 80 ]. As shown in Italy there seems to be certain reluctance towards bringing current topics discussed in literature into practice [ 81 ]. The problem mentioned may partly be due to the fact that the current guidelines are insufficiently supported by randomized surgical trials. Interestingly the COSA-NZ-UK Endometrial Cancer Study Group trial showed that lymphadenectomy showed no advantage for endometrial cancer if primary surgery was followed by adjuvant radiotherapy [ 82 ]. Therefore studies on all surgical aspects are warranted. This also includes newer surgical approaches which await further evaluation in prospective studies. Staging Endometrial carcinomas are staged surgically. Procedures previously used for determination of stages, such as fractional dilatation and curettage to differentiate between stage I and II, are no longer applicable unless the patient is to be treated by primary radiation therapy. The old (1971) and new (1988) staging system of the International Federation of Gynecology and Obstetrics (FIGO) are shown in table 2 . The prognostic importance of adequate surgical staging was recently demonstrated [ 83 ]. Table 2 Tumor classification of the international Federation of Obstetrics and Gynecology (FIGO). The surgical staging system is obligatory unless patients are to undergo primary radiotherapy when the older clinical staging system may be used. Stage Stage – Clinical Staging Stage – Surgical Staging I I Carcinoma confined to corpus Ia Tumor limited to endometrium Ia Length of uterine cavity ≤ 8 cm Ib Invasion ≤ 1/2 myometrium Ib Length of uterine cavity > 8 cm Ic Invasion > 1/2 myometrium II II Carcinoma involves corpus and cervix IIa Endocervical glandular involvement only IIb Cervical stromal invasion III III Carcinoma extends outside uterus but not outside the true pelvis IIIa Tumor invades serosa or adnexa or positive peritoneal cytology IIIb Vaginal metastasis IIIc Metastases to pelvic or para-aortic lymph nodes VI IV Carcinoma extents outside true pelvis or involves bladder or rectum IVa Tumor invades bladder, bowel mucosa, or both IVb Distant metastases, including intra-abdominal and/or inguinal lymph nodes Prognostic and predictive factors Tumor stage, patient age, histologic type and grade, hormone receptors and DNA ploidy represent the traditional prognostic factors. Respective of the response to progestin therapy steroid hormone receptors may also be regarded a predictive factors in recurrent and advanced disease [ 8 ]. The strong prognostic impact of tumor stage is underlined by the cumulative 5-year survival rates (surgical/pathological staging) which are 85% for stage I, 70% for stage II, 49% for stage III and 18% for stage IV disease. Obviously both major factors which make up the staging system of endometrial carcinomas, depth of invasion into the myometrium and lymph node status, are major prognostic factors [ 84 , 85 ]. A recently published analysis draws the attention to lymphvascular space involvement and suggests that its presence should indicate lymphadenectomy or adjuvant therapy [ 86 ]. Lymphvascular space involvement was also closely linked with advanced stage (unpublished observation). Furthermore, several additional prognostic factors have been suggested: Nulliparity, high tumor cell proliferation (KI-67), high tumor vessel density (angiogenesis), oncogene amplification or overexpression (HER2/neu, K-ras) and alterations of tumor suppressor genes (PTEN, p53, p21, p16) are believed to be associated with adverse prognoses. Especially Ki-67 could be of greater importance seeing that this parameter proved to be independent in multifactorial analyses in a prospective study [ 87 ]. There is a divergence of opinion concerning the value of a positive peritoneal cytology as an independent prognostic factor [ 88 ]. In stage I, depth of myometrial invasion, vascular invasion, mitosis count and progesterone receptor negativity are statistically significant prognostic factors [ 89 , 90 ]. Overexpression of p53 is observed in approximately 20% of all endometrial carcinomas and in up to 90% of serous carcinomas [ 91 ]. The plasminogen activator inhibitor type 2 has been discussed as a possible independent prognostic marker [ 92 ]. Overexpression of the oncogene Her-2/neu is significantly more common in advanced than in early stages. Finally, ploidy status is possibly an independent prognostic factor [ 93 ], with aneuploidy being mainly associated to prognostically unfavorable serous, clear cell and poorly differentiated carcinomas. There are, unfortunately, still too many controversies to draw final conclusions or even to make suggestions on factors to be determined routinely. Radiotherapy Primary radiotherapy In medically compromised women, primary irradiation may be suitable. Analyses from Germany show that approximately 20% of all patients undergo primary radiotherapy [ 7 , 64 ]. There are three possible approaches: The after loading technique alone, a combination of after loading and percutaneous techniques or by percutaneous radiotherapy alone. The literature also reports on considerable variation in the number of fractions and the doses of each fraction [ 64 ]. In case of brachytherapy only 5 × 8 Gy or 8 × 5 Gy may be applied. Combined therapy usually delivers 50 Gy percutaneously with partial blocking of the bladder and intestines after 24 and 30 Gy and 2 to 5 fractions via brachytherapy. Brachytherapy may also be applied before commencement of percutaneous therapy. The selected dose may be applied by a variety of techniques e.g. by Heyman capsules, double-rod-shaped applicators, indwelling applicators etc. [ 64 ]. An appropriate applicator should ensure adequate irradiation of the entire uterus [ 94 ]. Nowadays, computer controlled treatment planning allows optimal treatment planning and an individual adaptation of the dose distribution to the uterine cavity. This may vary between patients of course. Only few clinical trials on primary radiotherapy have been performed. The results suggest that dividing doses into smaller fractions allows better tumor control and has less side effects [ 95 ]. The old question on high-dose-rate or low-dose-rate after loading still remains unsolved. There is still limited data on the efficiency of primary radiotherapy in endometrial carcinomas. An analysis of 154 patients having undergone primary radiotherapy at our department showed that local recurrences are more common in this group compared to surgery and adjuvant radiotherapy. The rates of recurrences are stage dependent: 23.3% vs. 13.2% in stage I and 39.2% vs. 25.9% in stage II [ 7 ]. As shown in figure 5 there is a significant difference in overall survival between both groups. Thus, primary radiotherapy represents an effective but suboptimal measure for this group of patients being in generally poorer physical condition. Figure 5 Decision tree concerning the treatment of recurrent endometrial carcinoma.Legend: ENCA = endometrial carcinoma; QoL = Quality of Life; RT = radiotherapy; HT = endocrine treatment; CHT = chemotherapy; Adjuvant postoperative irradiation Most studies indicate that advance age, grade 3 histology or deep myometrial invasion relate to a higher risk of disease recurrence. Therefore adjuvant radiotherapy seems important in this subset. External beam radiation should also be considered in cases of multiple infiltrated lymph nodes (> 5). But many questions remain unanswered and the value of postoperative irradiation is still under debate. Survival rates of patients with early stage disease are excellent, no matter whether they underwent extended-surgical staging with more restricted use of the adjuvant therapy or simple hysterectomy bilateral salpingoophorectomy with more frequent use of adjuvant radiotherapy [ 96 ]. Prospective-randomized trials have so far only demonstrated improved local control yet no overall survival benefit, and have higher rates of treatment related complications [ 97 ]. This accords to larger retrospective analyses, most likely due to the fact that the majority of these recurrences can be salvaged through radiation therapy [ 98 ]. With respect to the importance of the problem, trials to evaluate the therapeutic benefit of adjuvant radiotherapy in the several subsets of patients at higher risk are warranted. The American Brachytherapy Society has now issued recommendations for brachytherapy for carcinomas of the endometrium [ 94 ]. According to these, the applicator selection should be based on patient and target volume geometry, the dose prescription point should be clearly specified and the treatment plan should be optimized. For intravaginal brachytherapy selection of the largest diameter applicator is to ensure close mucosal apposition. Finally, doses should be reported both at the vaginal surface and at 0.5-cm depth irrespective of the dose prescription point. Adjuvant medical treatment The data on adjuvant medical treatment is not conclusive. Most studies have their limitations and therefore there is still no final answer to the question, who should receive what type of adjuvant treatment. Merely in cases of uterine papillary serous carcinomas, which affect 1% to 10% of patients, there is consensus that patients should receive chemotherapy (with or without adjuvant radiotherapy) with a platinum-based regimen, combined with doxorubicin and cyclophosphamide. Newer regimens consider paclitaxel, with or without platinum [ 99 ]. Adjuvant hormonal treatment A multicenter, open, controlled, prospectively randomized trial on adjuvant endocrine treatment with medroxyprogesterone acetate (MPA) or tamoxifen in stage I and II endometrial carcinomas failed to detect differences in the disease-free and overall survival rates for a tamoxifen group compared with a control or a MPA group [ 100 ]. However, the total number of patients on trial (n = 388) seems too low in relation to the favorable prognosis of early stage disease and the low total response rate of tamoxifen which ranged around 10% in this situation [ 101 ]. In the aforementioned study tamoxifen demonstrated only modest activity which lead the authors to the conclusion that tamoxifen does not warrant further investigation as a single agent but perhaps a sequential use of tamoxifen and progestational agents [ 101 ]. Adjuvant chemotherapy The generally good prognosis of endometrial carcinomas does not justify a general recommendation of chemotherapy, especially in the early stages. Even so, patients at high risk (unfavorable histological type, deep myometrial invasion and lymph-vascular space involvement) seem to profit from adjuvant chemotherapy [ 102 , 103 ]. Studies on the subject unfortunately often lack a control arm so the effects of chemotherapy remain unclear [ 104 ]. Well designed studies in the group of high risk patients are warranted. Preservation of fertility The development of endometrial cancer in young patients is usually related to unopposed estrogen stimulation. In patients with continuing a desire to have children, approaches have been made to preserve fertility. Primary hormonal therapy with progestins was suggested as an alternative treatment for surgery hereby offering the preservation of fertility. These must be interpreted with caution because of low case numbers and a publication bias. Treatment with megestrol acetate at 160 mg/day for 3 months or medroxyprogesterone acetate (MPA) at 200–800 mg/day for 2–14 months resulted in disease regression in 60 to 75%, however, the percentage of patients who actually delivered healthy children was much lower, ranging from 20 to 25% [ 105 - 107 ]. Furthermore, in several patients, persistent or recurrent disease was observed at the time of a later hysterectomy. In all these cases of unsatisfactory progestin therapy and delayed definitive surgical treatment may have adversely affected patient prognosis. In essence, the problems regarding the optimal selection of patients for conservative progestin therapy are unsolved. Only cases with good prognostic factors are to be selected: Well differentiated tumors and tumors with known favorable prognoses (e. g. endometrioid type tumors) with no or minimal myometrial invasion and early stage disease. The greatest problem lies in the difficulty of appropriate staging, resulting in a potential underestimation of the problem. Maybe positron emission tomography, magnetic resonance imaging or a combination of both may be useful in this respect in due course. Patients have to be carefully informed that this fertility preserving concept is still experimental. Moreover, these women must realize the low overall pregnancy rate which may partially be related to the origin of the disease (obesity, irregular menses, polycystic ovarian disease (PCOD) with chronic anovulation and infertility) [ 108 ]. Nowadays when treatment of infertility is frequently offered to elderly women, such conservative treatments should, however be investigated more thoroughly. Further questions concern follow-up of these patients. In the aforementioned studies, many patients were treated with maintenance therapy (oral contraceptives or cyclic progestins) to prevent recurrence, which was excluded by routine combination of sonography and D&C every 3–6 months. Palliative treatment – treatment of local recurrences Generally the prognosis for patients with recurrent disease is poor, therefore a thorough staging procedure should be performed. Analyses from other tumor entities, e.g. ovarian carcinomas, have demonstrated that a second surgical intervention may be useful in improving patients' overall survival [ 109 , 110 ]. Thus, even in absence of clinical studies on the subject, patients with single site recurrence should be evaluated for their suitability to surgery at relapse. In patients with isolated central recurrences, pelvic exenteration may be a potential option for cure [ 111 ]. The choice with respect to therapy of cancer recurrence strongly depends on prior treatment. In case of prior radiotherapy, a second intervention may often not be possible. If radiotherapy is possible, only nonbulky vaginal recurrences (< 0.5-cm thick) should be treated by intracavitary brachytherapy. Patients with bulky (> 0.5-cm thick) recurrences should receive interstitial techniques [ 94 ]. A general decision tree regarding the procedures in case of a recurrence is depicted in figure 6 . Regarding the decision, whether or not to start with endocrine or cytotoxic treatment, the individual risk seems important. But equally important are aspects of patients' quality of life. A risk assessment scale, which was originally introduced by Possinger for breast cancer, seems helpful [ 108 ]. Similar to the adjuvant situation, treatment with tamoxifen, medroxyprogesterone acetate (MPA) alone or in combination may be used. However, there are only few, relatively old studies on the topic, all of which do not allow a final conclusion on the value of such therapies. MPA seems to be the best substance producing remission rates up to 80% in receptor positive tumors [ 113 ]. Remission rates of tamoxifen range around 25–30%. After failure of MPA, tamoxifen may be added to MPA producing remission rates in this combination of 50–60% [ 114 ]. Also aminogluthemide seems to be active in endometrial cancer [ 115 ]. Responses to all kinds of treatments unfortunately do not tend to last for long. Figure 6 Comparison of overall survival between patients undergoing surgery and radiotherapy or primary radiotherapy. Note that patients in the group undergoing primary radiotherapy were generally older Some studies also suggest a potential benefit for luteinizing hormone-releasing hormone analogues (GnRH), although studies in endometrial carcinomas have not shown any convincing activity [ 116 - 118 ]. To sum, up on the background of the few and old studies with often low case numbers, newer large studies on endocrine strategies are warranted. With respect to chemotherapy, several cyctostatic compounds have demonstrated efficiency. Among these are paclitaxel, carboplatinum, doxorubicin, cisplatin, etoposide, and 5 fluorouracil. Topotecan showed only limited activity [ 119 - 124 ]. All drugs may be used alone or in combination, perhaps even in combination with hormonal therapy [ 125 , 126 ]. Again, realizing that combination chemotherapy produces greater side effects, every option, endocrine treatment, single drug or combination chemotherapy should be considered. Follow-up Generally recommendation for clinical follow-up of patients advocate patients to be monitored at three-monthly intervals during the first 3 years, at 6-monthly intervals up to the 5th year and at yearly intervals thereafter. Apart from provision of general information on all aspects of the disease and its treatment including unconventional cancer therapies, the documentation of patients' history, a clinical gynecological examination including a pap smear, a vaginal sonography and even the determination of tumor markers (SCC, CA125) have been recommended. Up to 95% of all recurrences may be detected early this way [ 127 ]. Again there are no prospective studies on the subject to enforce such recommendations. On the contrary, several retrospective studies indicate that an intensive follow-up does not result in a survival advantage for patients with recurrent disease but merely increases costs [ 7 , 128 - 131 ]. Also in this area the most appropriate management of endometrial carcinomas remains to be determined. In any case an annual examination of the breast, including mammography, is recommended, due to the fact of the frequent coincidence of malignancies of the breast [ 7 ]. Estrogen replacement therapy in endometrial cancer patients Hormone replacement therapy (HRT) with estrogen with or without progestins is frequently used to alleviate menopausal symptoms but also to reduce the risk of osteoporosis and cognitive dysfunctions. In endometrial carcinomas HRT may be believed to be critical due to fear of initiating growth of occult residual tumor cells, resulting in disease recurrence and shortened survival. As summarized by the American College of Obstetricians and Gynecologists (ACOG) there is not enough data to draw any final conclusion. Any decision on the subject should thus be individualized based on potential benefit and risk to the patient [ 132 ]. Some studies which have addressed this subject covered only a small number of patients in regard to the overall excellent prognosis of endometrial cancer patients. Of all these studies, however, none produced evidence that patients should not receive estrogen [ 133 - 136 ]. On the contrary, some studies reported an even prolonged survival for patients (with low-risk factors for recurrence, namely, low tumor grade (grades 1 and 2), less than 1/2 myometrial invasion, and no metastases to lymph nodes or other organs,) who received estrogen [ 134 , 136 ]. Furthermore, the introduction of selective estrogen receptor modulators (SERM) has also offered new treatment options which will also have to be studied in the future. Final remarks Endometrial carcinomas represent a very frequent tumor entity in industrialized countries. It is hard to believe how little evidence-based data exists on even the important aspects of the disease. This may be partially due to the overall good prognosis even if surgery is reduced to hysterectomy and adnectomy. However, the high incidence in the developed world and consequently many women suffering relapses, necessitates the research of new approaches for cancer recurrence. Recent research suggests that therapy with trastuzumab (Herceptin ® ) could perhaps improve the outcome in HER-2/neu overexpressing tumors [ 137 ]. Further research will focus on molecules and pathways responsible for the initiation and growth of endometrial carcinomas, including tumor suppressor genes, DNA mismatch repair genes, oncogenes, molecules involved in adhesion and invasion and angiogenesis [ 138 ]. This research will hopefully allow the development of specific and selective inhibitors. Some advances may also be possible with – conventional treatment, especially radiotherapy. Recent findings of a retrospective analysis suggest that tumor oxygenation may play an important role during adjuvant radiotherapy of endometrial carcinomas. Patients with normal hemoglobin levels during therapy (according to the definition of the EORTC = 12.0 g/dl) have a substantially better overall and recurrence free survival (Figure 7 ) [ 139 ]. Due to this strong impact on patients' health, these aspects require further studies. Figure 7 Comparison of recurrence-free survival between in patients undergoing adjuvant radiotherapy with respect to anemia. Log Rank = 9.1; df = 1; p < 0.003.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC506786.xml
314481
Researchers Add to Proteomics Toolbox
null
Genes use a simple language—written in the molecules of DNA—to build thousands of proteins in a dizzying variety of sizes and shapes. With only four different nucleotide building blocks, DNA codes for the 20 different amino acids (each with their own structures and properties) that provide the foundation for the enormous diversity of protein form and function. This diversity makes the systematic study of all the proteins of a given organism (called proteomics), a challenging enterprise. Interactions between proteins underlie nearly every fundamental process within the cell. They can form higher-order multiprotein complexes like those involved in transcription and replication, help transport proteins to their proper location in the cell, and participate in signaling pathways. Because of their importance, disruption of these interactions can have disastrous consequences. For example, the loss of the ability of a normal cellular protein called Src to bind to certain other proteins can be associated with cancer progression. The determinants of these interactions are poorly understood, but in many cases these interactions are mediated by small pieces of the proteins, which are called peptides. Peptides serve as the starting point for the novel strategy reported in this issue. Gianni Cesareni and colleagues have added to the repertoire of proteomic analysis by devising a global strategy to investigate protein–protein interactions on an organismal level using yeast as a model organism. The authors select a protein of interest from yeast, which can be thought of as the “bait” for which they wish to identify protein-binding partners. They start by looking at a number of different previously identified peptides that bind the bait protein. Commonalities between the sequences of these peptides form the “consensus” binding sequence, a base framework of protein sequence from which many possible variations can be derived. Since the protein sequences of all proteins (the proteome) in yeast can be deduced from the sequenced genome, the authors can scan the proteome for proteins that contain the consensus, or a closely related, sequence. These proteins could potentially bind the bait protein. Peptide sequences from these identified proteins are synthesized chemically and arrayed on a membrane, which is bathed in a solution containing the bait protein. After washing off the excess bait protein, they can figure out where it remains on the membrane and therefore tell which peptides the bait protein has bound. The proteins corresponding to these peptides are candidate binding proteins that are validated by further experimentation. The protein–protein interactions identified by this approach can be used to extend the network of known interactions in the proteome. This will enable researchers to draw functional linkages between proteins, whether they are involved in a basic biological process or in human disease. By examining whole families of proteins, it may also aid in elucidating the underlying determinants of binding specificity, which would provide clues to the biomechanisms underlying cell processes. These insights could lead to methods for manipulating these interactions, perhaps even in cases of human disease, as in the case of Src and cancer. This approach can readily be applied to the proteomes of more complex organisms like humans and adds to the growing number of experimental strategies available to researchers in proteomics. Protein interaction network
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314481.xml
512283
Immobilized probe and glass surface chemistry as variables in microarray fabrication
Background Global gene expression studies with microarrays can offer biological insights never before possible. However, the technology possesses many sources of technical variability that are an obstacle to obtaining high quality data sets. Since spotted microarrays offer design/content flexibility and potential cost savings over commercial systems, we have developed prehybridization quality control strategies for spotted cDNA and oligonucleotide arrays. These approaches utilize a third fluorescent dye (fluorescein) to monitor key fabrication variables, such as print/spot morphology, DNA retention, and background arising from probe redistributed during blocking. Here, our labeled cDNA array platform is used to study, 1) compression of array data using known input ratios of Arabidopsis in vitro transcripts and arrayed serial dilutions of homologous probes; 2) how curing time of in-house poly-L-lysine coated slides impacts probe retention capacity; and 3) the retention characteristics of 13 commercially available surfaces. Results When array element fluorescein intensity drops below 5,000 RFU/pixel, gene expression measurements become increasingly compressed, thereby validating this value as a prehybridization quality control threshold. We observe that the DNA retention capacity of in-house poly-L-lysine slides decreases rapidly over time (~50% reduction between 3 and 12 weeks post-coating; p < 0.0002) and that there are considerable differences in retention characteristics among commercially available poly-L-lysine and amino silane-coated slides. Conclusions High DNA retention rates are necessary for accurate gene expression measurements. Therefore, an understanding of the characteristics and optimization of protocols to an array surface are prerequisites to fabrication of high quality arrays.
Background The generation of reliable gene expression data with cDNA microarrays requires fabrication of quality arrays. This task encompasses the amplification of adequate amounts of concentrated PCR product for use as probe from the cDNA clone, followed by ordered arraying of the probes onto coated glass slides. The glass slide is a key variable in either spotted cDNA or oligonucleotide array fabrication since it must possess: 1) a uniform surface that yields spots of consistent shape and size, 2) low background fluorescence, and 3) high DNA retention capacity. Since the array is clearly a source of experimental variability, we have developed a novel three-color array approach where it is possible to directly visualize either cDNA or oligonucleotide arrays prior to hybridization [ 1 - 3 ]. For cDNA arrays, the probes are easily tagged with a third, Cy3/Cy5 compatible, fluorescent dye (fluorescein) during amplification. After purification of PCR products, which includes removal of unincorporated oligonucleotide primer, the detected fluorescein fluorescence represents deposited cDNA probe on the array. This three-color approach allows for assessment of slide fabrication independent of hybridization, thereby enabling 1) direct visualization of array/element morphology, 2) quantification of probe deposition and retention on the slide surface and 3) ultimately a means for array quality control prior to hybridization. By labeling the array itself with a third color, we have observed that arrays fabricated together are not equivalent in terms of a number of measurable physical parameters, including the amount of DNA probe deposited and retained and the amount of background arising from probe solublized and re-deposited during post-processing. In prior studies, we observed that these pre-hybridization array-based variables play a direct and significant relationship in replicate consistency, and that microarray data quality can be improved through prehybridization slide selection based upon these quality parameters [ 1 , 2 ]. As a result of these studies, we identified putative slide acceptance criteria: array fluorescein mean element intensity >5000 RFU/pixel, coefficient of variation (CV) in intensity <10%, mean signal to noise score (signal/signal + noise; S/S+N) >0.85, and CV in spot size <20%. In this report, using known input ratios of in vitro transcript we experimentally correlate the quantity of support bound probe to measured expression ratios, in order to validate our quality control threshold for array acceptance. We then utilize our three-color array platform to evaluate the characteristics of in-house prepared poly-L-lysine coated slides and 13 additional commercially available coating surfaces, in terms of background auto-fluorescence, spot morphology, and DNA retention. Results and Discussion The relationship between support bound probe and measured ratio reliability It has been assumed that the amount of cDNA probe deposited and retained on the array surface would have a nominal effect on observed differential expression ratios due to the competitive nature of two channel fluorescent hybridizations [ 4 ]; however this assumption has been shown to be false [ 1 , 5 ]. Yue et al., using unlabeled Saccharomyces cerevisiae probes and complementary Cy5 and Cy3 labeled cDNA targets derived from in vitro transcripts, indirectly demonstrated this by printing yeast probes at increasingly dilute concentrations (<50 ng/ul) and observed elimination of the measured dynamic range to where input transcript ratios of 30:1 or 1:30 were both detected as output ratios close to 1:1, illustrating that limiting bound probe results in an underestimation or failure to detect differential gene expression [ 5 ]. To expand upon these observations and place them in context with the quality control standards of our three-color array platform, we conducted similar experiments using Arabidopsis probes and transcripts. Total thymus RNA extracted from the DR+/+ and DR lyp/lyp [ 6 ] BioBreeding rats was directly labeled through reverse transcription reactions possessing cyanine dyes; these labeling reactions were spiked with known input ratios (30:1, 10:1, 1:1, and 1:0) of Arabidopsis gene in vitro transcript and hybridized to 18,000 probe rat cDNA arrays possessing serially diluted fluorescein-labeled Arabidopsis probes (cellulose synthase, chlorophyll a/b binding protein, and ribulose-1,5-bisphosphate and triosphosphate isomerase, 1:2 dilution series printed at 200 ng/ul to 6.25 ng/ul). This approach allowed comparison of known RNA input ratio to measured output ratio, enabling a direct and quantitative measure of the relationship between the amount of support-bound probe and ratio data compression (Figure 1A and 1B ). Using Matarray [ 7 , 8 ], spots possessing low hybridized image quality ( q com ) were filtered; these spots were either saturated or possessed high background. Spots with low hybridization intensities, which would normally be flagged by Matarray, were intentionally retained to study ratio compression due to low amounts of support-bound probe. After filtering, 896/1536 data points from 16 different arrays were available for analysis. Plotted on the y-axis of Figure 1b , is the measured Arabidopsis in vitro transcript output log ratio divided by the log ratio of transcript actually introduced into the Cy3 and Cy5 labeling reactions. In this analysis, a perfect measurement is represented by a value of "1". On the x-axis, is plotted the spot fluorescein intensity. When the spot fluorescein intensity falls below 5000 RFU/pixel, the data variability and data compression (underestimation of differential gene expression) dramatically increase. These results recapitulate our previous observations where replicate consistency was found to decrease when the array average spot fluorescein intensity dropped below 5000 RFU/pixel, whereas arrays possessing average fluorescein intensities above 5000 RFU/pixel were found to generate equally good data. These results further demonstrate that use of measurable array characteristics are effective quality markers for printed arrays (judged by their effect on the hybridization data) and serve to validate our array intensity quality control threshold of >5000 RFU/pixel. Figure 1 Evaluation of measured output ratio of spiked Arabidopsis in vitro transcript at known input ratios. A. Total thymus RNA extracted from the DR+/+ and DR lyp/lyp [6] BioBreeding rats spiked with known input ratios of Arabidopsis gene in vitro transcript and hybridized to 18,000 probe rat cDNA arrays possessing serially diluted fluorescein-labeled Arabidopsis probes. B. Evaluation of data compression as a function of support-bound probe. On the x-axis is plotted the average pixel fluorescein intensity per spot plotted against the Arabidopsis transcript measured output log ratio/actual input log ratio. As spot intensities fall below 5000 RFU/pixel, ratio measurements become increasingly compressed. Impact of poly-L-lysine cure-time on DNA retention capacity Clearly, the amount of immobilized probe on the coated glass surface is a critical array fabrication variable, therefore factors that affect the amount of retention characteristics, such as surface chemistry, probe concentration, spotting buffer, spotting conditions, cross-linking and blocking conditions are important to understand. Protocols for coating glass microscope slides with poly-L-lysine are readily available on-line and reasonably simple to perform (for example: ; ; ). Although most available protocols are quite similar, some recommend the curing of slides for two weeks prior to spotting, while others state that coated slides are not stable for extended periods of time and recommend not printing onto slides that are greater than 4 months old. To investigate slide coating age as a potential variable in retention capacity, we fabricated more than 1,000 rat cDNA arrays (18,000 element/slide) using in-house poly-L-lysine coated slides ranging in age from 3 to 12 weeks. These slides were coated in 26 independent sessions and utilized over 12 different print runs. After printing all arrays were post-processed [ 9 ] and imaged under standardized conditions as previously described [ 1 , 2 ]. A significant loss of DNA retention capacity is observed (Figure 2 ) when the average array spot fluorescein intensity is plotted against the coating age at the time of printing (R 2 = 0.84; p < 0.0002). An average array fluorescein intensity of >15,000 RFU/pixel was observed when printing on slides 4 weeks old or less, however a nearly 50% reduction in retention capacity is observed when printing on poly-L-lysine coating greater than 10 weeks old. We speculate that the poly-L-lysine may become oxidized thereby losing its positive charge and ability to initially electrostatically interact with the negatively charged DNA. Irregardless of the type of degradation occurring to the surface coating over time, these results indicate that, at least for in house fabricated poly-L-lysine coated slides, shelf-life is a significant variable in the fabrication of quality arrays capable of yielding reliable gene expression measurements. Figure 2 Relationship between DNA retention and poly-L-lysine cure time. Retention capacity is lost as the poly-L-lysine cure time increases (R 2 = 0.84; p < 0.0002). Analysis includes 999 arrays printed over 12 different print runs. Each print run consisted of 100 arrays, printed onto poly-L-lysine slides from 2 or more coating lots. Investigation of probe retention characteristics of commercially available coated slides Given the potential time-dependent variability of in-house prepared poly-L-lysine coated slides, we investigated the retention characteristics of commercially available coated slides. Our objective was to identify a surface with consistently higher retention characteristics than our "fresh" in-house slides without having to change the spotting buffer (1.5 M betaine/5% DMSO) or the nonaqueous post-processing protocol [ 1 , 2 , 9 ], since these methods were previously found to yield high quality results on poly-L-lysine coated slides prepared in-house. We obtained examples of 13 different vendor-supplied slides for evaluation that possessed either poly-L-lysine, aminosilane, or undisclosed surface chemistries. Prior to printing, background auto-fluorescence in the fluorescein, Cy3, and Cy5 channels was evaluated. Fluorescein auto-fluorescence was observed on all poly-L-lysine slides except for those produced in-house, as well as 6 of the aminosilane slides (Asper Biotech, Corning, Erie Scientific, Genetix, Telechem), and the proprietary surface from Full Moon Biosciences. Cy3 auto-fluorescence was observed on all 3 commercial poly-L-lysine slides but not those prepared in-house, however none was observed on any of the aminosilane slides. Insignificant background in the Cy5 channel was only observed on 2 commercial poly-L-lysine slides (Electron Microscopy Sciences, Polysciences Inc.). To study retention characteristics, a single 9600 element human cDNA array was spotted onto each slide in 1.5 M betaine/3%DMSO. The in-house poly-L-lysine slides were less than 6 weeks old and all vendor-supplied slides were unpacked from any special packaging immediately before printing. Five replicate arrays for each slide type were generated. The five replicates were evenly distributed over the arrayer deck (capacity 100 slides) by arranging the slides into 5 groups of 18 to account for any variance introduced through print rank order (ie first versus last), since we previously identified this as a variable that influences array average array fluorescein intensity [ 1 ]. Fluorescein images were again obtained, under strict standardized conditions, immediately after printing and again after post-processing to measure DNA deposited and retained (Figure 3A ; Table 1 ) [ 1 , 2 ]. The average amount of DNA deposited per element varied considerably among slides of different sources (n = 5 per source) ranging from a low of 2,300 +/-300 RFU/pixel to a high of 20,700 +/-3,300 RFU/pixel. Among the slides evaluated, the poly-L-lysine coated slides yielded larger spot sizes, perhaps due to a lower hydrophobicity than the aminated surfaces. The average DNA retained per element after blocking/post-processing ranged from a low of 1,400 +/-200 RFU/pixel to a high of 11,100+/-2,700 RFU/pixel on poly-L-lysine slides prepared in-house (Figure 3B ). These results, combined with our previous observations [ 1 , 2 ], indicate that DNA concentration, choice of printing buffer, slide position on the arrayer deck and slide surface chemistry all influence the amount of DNA deposited and ultimately retained, which can be effectively monitored by our three-color approach. We have previously shown that the amount of probe solublized and redistributed over the slide during post-processing is an important quality control parameter [ 2 ]. Eight of the surfaces tested generated fluorescein signal to noise values (signal/signal + noise; S/S+N) ≥ 0.90 after post-processing, a value that we have previously shown to sufficient to generate high replicate reproducibility [ 2 ]. However, the majority of vendor-supplied coated slides, did not meet or exceed our established average array element intensity value of 5,000 fluorescein RFU/pixel under the printing and post-processing methods optimized for our in-house coated slides. These results point towards the possibility that array fabrication should be optimized to the specific surface selected for use and that even the same surface chemistry from different sources may perform differently. Figure 3 A: Fluorescein images of 18,000 element rat cDNA arrays on in-house poly-L-lysine coated slide after printing (A1) and array after non-aqueous post-processing (A2). Fluorescein images of simultaneously 18,000 element rat cDNA arrays on Full Moon Biosystems coated slide (undisclosed chemistry) after printing (A3) and array after non-aqueous post-processing (A4). Note differences in amount of DNA deposited and retained. (White spots are saturated). B: Comparison of retention capacity of 14 different coating surfaces using human 9,600 probe cDNA arrays. Tabulated measurements are based upon 5 replicates slides (~48,000 elements) for each slide type evenly distributed over the arrayer deck (ie 5 slides of a given type did not occupy 5 adjacent positions on the arrayer deck). Slides 1–4 are poly-L-lysine: MCW in-house, Electron Microscopy Sciences, Polysciences, and Cel Associates, respectively. Slides 5–13 are aminated: Asper Biotech, Apogent, Bioslide, Erie Scientific, Genetix, Corning Ultra GAPS, Corning GAPS II, Sigma, and Telechem Super Amine, respectively. Slide 14 is an undisclosed chemistry offered by Full Moon Biosystems. The graph represents the average spot fluorescein intensity RFU/pixel (burgundy) +/- standard deviation (yellow). Table 1 Retention Studies on Commercial Coated Slide Surfaces Vendor Chemistry RFU/Pixel deposited × 10 3 RFU/pixel retained × 10 3 Percent Retention Spot Diameter Processed S/(S+N) MCW in-house Poly-L-lysine 19.0+/-7.7 11.1+/-2.7 58.4% 114+/-9 0.90+/-0.00 Electron Microscopy Sciences Poly-L-lysine 15.3+/-4.9 2.12+/-1.1 13.9% 100+/-28 0.85+/-0.04 Polysciences Poly-L-lysine 11.5+/-6.3 1.3+/-1.0 11.3% 125+/-6 0.86+/-0.06 Cel Associates Poly-L-lysine 5.7+/-3.6 2.3+/-0.5 40.4% 100+/-35 0.79+/-0.09 Asper Biotech Aminated 6.8+/-2.4 3.2+/-1.1 47.1% 98+/-22 0.93+/-0.05 Apogent Ezrays Aminated 6.0+/-2.9 2.8+/-0.3 46.7% 109+/-20 0.92+/-0.02 Bioslide Aminated 3.3+/-0.9 1.5+/-0.5 45.5% 93+/-7 0.87+/-0.02 Erie Scientific Aminated 11.5+/-4.1 3.2+/-0.9 27.8% 81+/-15 0.90+/-0.01 Genetix Aminated 4.1+/-1.2 1.2+/-0.2 29.3% 94+/-23 0.86+/-0.02 Corning Ultra GAPS Aminated 2.3+/-0.3 1.5+/-0.2 65.2% 103+/-13 0.92+/-0.01 Corning GAPS II Aminated 3.9+/-0.9 1.4+/-0.2 35.9% 117+/-18 0.92+/-0.01 Sigma Aminated 20.7+/-3.3 5.4+/-1.0 26.1% 93+/-8 0.79+/-0.01 Telechem Super Amine Aminated 6.7+/-2.1 1.7+/-0.5 25.4% 107+/-19 0.80+/-0.0 Full Moon Biosystems Proprietary 6.4+/-1.4 2.1+/-0.2 32.8% 78+/-7 0.90+/-0.1 It has been reported that the amount of UV irradiation may be an important array fabrication variable since the amount of hybridization signal from spotted 70-mer oligonucleotides has been found to be dependent on the amount of cross-linking [ 10 ]. In this previous report, different optimal cross-linking intensities for attachment of spotted 70-mer oligonucleotides were observed for different slide coating chemistries (poly-L-lysine, aldehyde, aminosilane, epoxide) [ 10 ]; furthermore, different cross-linking optima for probe attachment were also observed for slides with the same or similar slide chemistry from different vendors. This variable was not explored in our evaluation of vendor-supplied surfaces and may account for some of the performance differences observed. The report by Wang et al., [ 10 ] prompted us to revisit this parameter for our in-house slides and we have observed approximately 20% better DNA retention by increasing the UV cross-linking energy from 60 mJ/cm2 to 200 mJ/cm2 independent of coated slide lot. Fabrication of high quality spotted arrays is a daunting task possessing a high number of variables. The vendor supplied slides tested here were done so under conditions that have been optimized for our in-house prepared poly-L-lysine coated slides, although our optimized protocol is not drastically different than those used by other laboratories nor drastically different from any of the vendor provided protocols. Our observations, as well as the observations of others, suggest that optimization of ones protocol to a surface chemistry is an essential first step to generating reliable global gene expression measurements using in-house spotted microarrays. Methods A sequence-verified human library (Research Genetics, Huntsville, AL), consisting of 41,472 clones or a 36,000 clone rat cDNA library obtained from the University of Iowa was used as a source of probe DNA. Cultures were grown in 150 ul Terrific Broth (Sigma, St. Louis, MO) supplemented with 100 mg/ml ampicillin in 384 deep-well plates (Matrix Technologies, Hudson, NH) sealed with air pore tape sheets (Qiagen, Valencia, CA) and incubated with agitation for 14–16 hr. Clone inserts were amplified in duplicate in 384-well format from 0.5 ul bacterial culture or from 0.5 ul purified plasmid (controls only) using 0.26 μM of each vector primer (SK865 5'-fluorescein-GTC CGT ATG TTG TGT GGA A-3' and SK536: 5'-fluorescein-GCG AAA GGG GGA TGT GCT G-3' [ 5 ]) (Sigma-Genosys, The Woodlands, TX) in a 20 μl reaction consisting of 10 mM Tris-HCl pH8.3, 3.0 mM MgCl 2 , 50 mM KCl, 0.2 mM each dNTP (Amersham, Piscataway, NJ), 1 M betaine [ 11 , 12 ], and 0.50 U Taq polymerase (Roche, Indianapolis IN). Reactions were amplified with a touchdown thermal profile consisting of 94°C for 5 min; 20 cycles of 94°C for 1 min, 60°C for 1 min (minus 0.5° per cycle), 72°C for 1 min; and 15 cycles of 94°C for 5 min; 20 cycles 94°C for 1 min, 55°C for 1 min, 72°C for 1 min; terminated with a 7 min hold at 72° [ 13 - 15 ]. PCR reactions were analyzed for single products by 1% agarose gel electrophoresis. Products from replicate plates were pooled and then purified by size exclusion filtration using the Multiscreen 384 PCR filter plates (Millipore, Bedford, MA). Forty wells of each 384-well probe plate were quantified by the PicoGreen assay (Molecular Probes, Eugene, OR) according to the manufacturers instructions. After quantification, all plates were dried down, and reconstituted at 125 ng/μl in 3% DMSO/1.5 M betaine. It has been shown that betaine normalizes base pair stability differences, increases solution viscosity, reduces evaporation rates [ 11 ], and enhances probe binding to surfaces such as poly-L-lysine or aminosilane [ 1 , 9 ]. We have observed higher probe retention at much lower DNA concentrations (150–200 ng/ul) in the presence of betaine versus the typically required 4–500 ng/ul when using conventional printing solutions [ 2 , 3 ]. Arabidopsis thaliana PCR product and in vitro transcript were purchased from Stratagene (La Jolla, CA) as part of the SpotReport ® -10 Array Validation System. Arabidopsis thaliana PCR product was cloned into the pCRII vector using the TA cloning kit (Invitrogen, Carlsbad CA) and fluorescein-labeled PCR products for photosystem I chlorophyll a/b-binding protein, RUBISCO activase, ribulose-1,5-biphosphate carboxylase/oxygenase, lipid transfer protein 6 lipid transfer protein 5, papain-type cysteine endopeptidase, root cap 1, and triosphophate isomerase were generated using vector-specific primers essentially as described above. Products were purified, quantified, and a 1:2 dilution series (200 ng/ul to 12.5 ng/ul) was prepared and printed in duplicate onto each array. Poly-L-lysine coated slides were prepared in-house as previously described [ 16 ] on Corning (Corning, NY) pre-cleaned 75 × 25 mm glass micro slides. Nine different commercially available aminosilane coated slides (Apogent Discoveries, Waltham, MA; Asper Biotech, Redwood City, CA; Bioslide Technologies, Walnut, CA; Corning Inc, Corning NY; Erie Scientific, Portsmouth, NH; Genetix, St. James, NY; Sigma, St. Louis, MO; Telechem International Inc, Sunnyvale, CA) and 3 different commercially available poly-L-lysine coated slides (Cel-Associates, Pearland, TX; Electron Microscopy Sciences, Fort Washington, PA; Polysciences Inc., Warrington, PA) were obtained for evaluation. Additionally, slides coated with a proprietary chemistry (Full Moon Biosystems, Sunnyvale, CA) were obtained. Microarrays possessing a density of 9,600 human probes/slide were printed onto coated slides using a GeneMachines Omni Grid printer (San Carlos, CA) with 16 Telechem International SMP3 pins (Sunnyvale, CA) at 40% humidity and 22°C. To control pin contact force and duration, the instrument was set with the following Z motion parameters, velocity: 7 cm/sec, acceleration: 100 cm/sec 2 , deceleration: 100 cm/sec 2 . All slides were post-processed using the previously described non-aqueous protocol [ 9 ] using 60 mJ/cm 2 UV cross-linking energy. This protocol has yielded more favorable fluorescein post-blocking signal-to-noise values (signal/signal+noise; S/S+N) as compared to blocking in aqueous solutions[ 2 ]. Image files on all slides were collected prior to printing to establish background fluorescence (fluorescein, Cy3 and Cy5), after printing (fluorescein), and after blocking (fluorescein), with a ScanArray 5000 (GSI Lumonics, Billerica, MA). Array image files were analyzed with the Matarray software [ 7 , 8 , 17 ]. Isolation of mRNA, labeling, and hybridization were performed as described previously . Known input ratios of photosystem I chlorophyll a/b-binding protein (30:1); RUBISCO activase (10:1); ribulose-1,5-biphosphate carboxylase/oxygenase (5:1); lipid transfer protein 6 (1:1); 0.7 lipid transfer protein 5 (1:1); papain-type cysteine endopeptidase (1:5); root cap 1 (1:10); and triosphophate isomerase (1:30) were spiked into Cy3 and Cy5 RNA labeling reactions, respectively. After hybridization, arrays were scanned with a ScanArray 5000 (GSI Lumonics, Billerica, MA) and image files were obtained. Again, array image files were analyzed with the Matarray software [ 7 , 8 , 17 ]. Authors' Contributions MJH and XW conceived of the study, its design and coordination, and drafted the manuscript. LM, JT, and SM carried out the array fabrication and gene expression studies. XW performed the statistical analysis. All authors read and approved the final manuscript.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC512283.xml
523843
T Cells Cause Lung Damage in Emphysema
null
T lymphocytes may have an important role in the pathogenesis of smoking-related emphysema, according to a new study by researchers from Houston, Texas, United States. “We now know that T cells are not only present in chronic obstructive pulmonary disease [COPD], but are harmful,” comments Steven Shapiro from Brigham and Women's Hospital , Harvard Medical School, who was not involved in the study. “We also now have a pathway that could be interrupted to prevent lung destruction in COPD.” Farrah Kheradmand and colleagues took lung samples from 28 ex-smokers who had been admitted to hospital for lung resection: 18 patients had moderate to severe COPD as well as evidence of emphysema, and ten patients had none. The researchers isolated lung lymphocytes from the samples and used two-color flow cytometry to phenotypically characterize the cells. They found that lymphocytes taken from patients with emphysema expressed more CCR5 and CXCR3 receptors, which are associated with a particular type of T cell called T helper 1 (Th1), than did those from control individuals. By contrast, expression of CCR4 receptors, which are found on T helper 2 (Th2) cells, was very low in both control and emphysema groups. CT image of the lung of subjects with end-stage emphysema next to a photomicrograph of their resected lung stained with H&E In a separate experiment, Kheradmand's team showed that lung lymphocytes taken from patients with emphysema secreted more of three other proteins—interferon gamma, monokine induced by interferon (MIG), and interferon-inducible protein 10 (IP-10)—than control patients. MIG and IP-10 are known to be produced by injured epithelial cells and are ligands for CXCR3 receptors, which are expressed by Th1 cells. Importantly, the researchers were also able to show that isolated peripheral lung macrophages secreted matrix metalloproteinase-12 (MMP12), an enzyme that degrades elastin—a protein important for lung elasticity—in the lungs, in response to IP-10 and MIG. Together these findings, say the authors, indicate that Th1 cells, but not Th2 cells, are required for producing the elastin-destroying lung environment of emphysema. The researchers now intend to investigate the antigens that drive the Th1-based inflammation that underlies emphysema. “Ultimately, we seek to understand the biochemistry of tobacco smoke that triggers inflammation in the first place, and whether such insight might explain other environmentally triggered lung diseases,” explains Kheradmand. “To understand such detailed immune mechanisms, we really need an improved experimental model of disease, and this we are currently working on.”
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC523843.xml
544953
A swimming robot actuated by living muscle tissue
Biomechatronics is the integration of biological components with artificial devices, in which the biological component confers a significant functional capability to the system, and the artificial component provides specific cellular and tissue interfaces that promote the maintenance and functional adaptation of the biological component. Based upon functional performance, muscle is potentially an excellent mechanical actuator, but the larger challenge of developing muscle-actuated, biomechatronic devices poses many scientific and engineering challenges. As a demonstratory proof of concept, we designed, built, and characterized a swimming robot actuated by two explanted frog semitendinosus muscles and controlled by an embedded microcontroller. Using open loop stimulation protocols, the robot performed basic swimming maneuvers such as starting, stopping, turning (turning radius ~400 mm) and straight-line swimming (max speed >1/3 body lengths/second). A broad spectrum antibiotic/antimycotic ringer solution surrounded the muscle actuators for long term maintenance, ex vivo. The robot swam for a total of 4 hours over a 42 hour lifespan (10% duty cycle) before its velocity degraded below 75% of its maximum. The development of functional biomechatronic prototypes with integrated musculoskeletal tissues is the first critical step toward the long term objective of controllable, adaptive and robust biomechatronic robots and prostheses.
Background Many technological barriers exist for the implementation of life-like mobility in robotic and prosthetic systems. Included among these barriers are (1) the availability of high-energy density storage media, (2) the availability of adequate muscle-like actuators, and (3) the availability of biologically inspired sensory technologies. As a possible resolution to these challenges, we consider in this investigation the use of living muscle tissue as a viable actuator for synthetic devices. Although important research has been conducted to advance a synthetic actuator technology with muscle-like properties, engineering science has not yet produced a motor system that can mimic the contractility, energetics, scalability and plasticity of living muscle tissue [ 1 , 2 ]. Muscle has several important advantages in addition to favorable dynamic characteristics [ 1 - 6 ]. In its function as a motor, muscle acts to provide positive mechanical work at a considerable aerobic transduction efficiency, or 1000 Joules of work per gram of glucose consumed [ 7 ]. It is a "smart material", having integrated sensors for the detection of displacement and rate of displacement (muscle spindles) as well as force (Golgi tendon organs). It can repair itself when damaged, and can functionally adapt to an increase in the demands of the environment by undergoing hypertrophic and hyperplastic growth [ 8 ] as well as fiber type transformations [ 9 - 12 ]. Muscle has integrated series-elastic components, which are thought to give rise to many of the "life-like" characteristics of animal movement [ 13 ], and the fuel that it consumes is a renewable resource, while the waste products produced are environmentally compatible. In this investigation, we examine the feasibility of using animal-derived muscle as an actuator for artificial devices in the millimeter to centimeter size scale. Perhaps researchers in the past did not consider muscle tissue a viable mechanical actuator because of tissue maintenance and control difficulties. The objectives of this study are to identify, and to begin to address, the many technical challenges related to maintaining and controlling explanted muscle tissues in the context of a robotic platform. To this end, we construct a hybrid swimming robot comprising a synthetic elastomeric tail actuated by a single pair of whole muscle explants from frog semitendinosus muscle. We anticipate that basic swimming maneuvers such as straight-line swimming and turning can be performed by alternately modulating electrical signals to each muscle actuator across two electrode pairs, one on each muscle near the neuromotor junction. We further anticipate that a multi-day robotic maintenance or lifespan can be achieved by surrounding the muscle actuators with a specific bath of amphibian ringer's solution comprising antibiotic and antimycotic agents. To test these ideas, we construct two robotic build-ups, each comprising a freshly dissected pair of explanted semitendinosus muscles. For each build-up, pilot data are collected to characterize the robot's swimming mechanics and lifespan. Methods Muscle Removal and Maintenance The surgical removal of muscle specimens designated for robotic actuation were performed according to procedures approved by the Committee on Animal Care, Northeastern University (Approval #0402-025-05). Briefly, adult frogs ( Rana pipiens ) were pithed, and both semitendinosus muscles were dissected free and removed with tendons intact. Before removal of the tissues from the animal, the length of each muscle belly was measured at an equilibrium or rest length. The resting length measurement was conducted on the intact muscle specimen with the limb positioned at an anatomically neutral position (see Table I for muscle lengths). After removal from the animal, the muscle, including its intact tendons, was weighed (see Table I for muscle mass). Each tendon was manipulated via tightly secured silk suture (size 5-0). Each muscle was then pinned at its rest length in a 100 mm Petri dish with a previously prepared SYLGARD (Dow Chemical) polydimethylsiloxane (PDMS) substrate. Table I Muscle actuator parameters and swimming robot performance parameters (mean values, N = 4) at the maximum forward swimming speed for robotic build-ups, B1a and B1b. Robot Muscle Mass (g) Muscle Len. (mm) Peak Muscle Strain Muscle Shortening Vel. (mm/s) Tail-beat Freq. (Hz) Tail Amp. (mm) Max. Robot Speed (mm/s) Wave Speed (mm/s) Wave Len. (mm) Slip B1a 0.31 31 6.5% 25 3.1 16 31 121 39 0.26 B1b 0.34 31 6.5% 25 3.1 16 45 140 45 0.32 Shortly before harvesting the muscles, two fresh liters of amphibian ringer solution were prepared according to a protocol specifically designed for frog organ culture [ 14 , 15 ]. The amphibian ringer comprised: NaCl, 83.89 mM; NaHCO 3 , 28.11 mM; KCL, 1.5 mM; KH 2 PO 4 , 1.2 mM; MgSO 4 , 1.2 mM; CaCl 2 Dihydrate, 1.3 mM; Glucose, 10 mM; MEM Amino Acid Mixture, 1:50 dilution (GIBCO #1130051); MEM Vitamin Mixture, 1:100 dilution (GIBCO # 1120052); Creatine, 1 mM; DL-Carnitine, 1 mM; Ferric Chloride, 0.9 μM; Human Serum Transferrin, 1.35 μM; Insulin, 1 mU/ml; L-Glutamine, 1:100 dilution; Sigma Chemical #A9909, 1:50 dilution (an antibiotic/antimycotic). A broad-spectrum, antibiotic/antimycotic was added out of necessity for long-term maintenance of the muscles, ex vivo. We observed, for periods greater than 24 hours, septic degradation of the muscle specimens in the absence of the antibiotic/antimycotic agents. After each muscle was placed within a Petri dish, a small volume of ringer solution was used to surround each muscle, the balance being used in the test tank for the swimming robot evaluations. The total amount of time between muscle removal from the animal to finalizing the muscle installation into the robotic swimmer was approximately 1 hour. Test Tank Construction and Swimming Robot Design The test tank was constructed from 6 mm (1/4") thick cast acrylic sheet, welded together with methylene chloride, with silicone fish tank adhesive being applied to form a water-tight seal at each joint. The test tank was 30 cm square and 6 cm deep. The robotic platform (Figure 1 ) was specifically designed to accommodate the frog semitendinosus muscles. The actuators were a single pair of whole muscle explants from frog semitendinosus muscle, arranged as antagonists on either side of the robot in an open-frame architecture. This open-frame architecture exposed the explanted tissues to the amphibian ringer solution during robot operations. The robotic platform mass before installation of the muscle actuators was 12.15 g, and the overall length (L) was 12 cm. Of this total length, the fore or anterior 7 cm section comprised a rigid frame machined from acetyl (Delrin) with nylon threaded fasteners, while the aft or posterior 5 cm section comprised a compliant cast silicone tail. A closed-cell Styrofoam float was affixed to the rigid forward section to provide positive buoyancy. The compliant tail had a narrow rectangular section between the mounting flange and the insertion to the rigid Delrin backbone. This compliant segment (Figure 1 ) served as a hinge for single degree-of-freedom actuation, permitting mediolateral oscillations of the tail. This narrow compliant section also provided a restoring force to return the tail to its neutral position when no muscle force was applied. Figure 1 The Biomechatronic Robotic Platform. The top image is a photograph (side view) of the device (robot B1a) shortly after initial testing. The bottom image is a schematic (to scale) with the float and embedded controller removed, showing the main components of the system: semitendinosus muscles (M), suture attachments (s), Styrofoam float (F), electrode wires (w), cast silicone tail assembly (T), rigid Delrin backbone (D), rigid Delrin head piece (H), lithium batteries (B), compliant hinge segment (k), cylindrical tail mounting boss (a), encapsulated microcontroller, infra-red sensor, and stimulator unit (C). The single part silicone RTV (Dow Corning type 734 flowable silicone) tails were cast using a 5-part virgin Teflon mold machined to form a single solid tail assembly with all of the features shown in Figure 1 . Casting of one-part silicones was accelerated by the addition of ~1 drop of water-based food coloring per 10 ml of silicone elastomer. This technique allowed tails of different mechanical properties to be readily color-coded during casting, and allowed the elastomer to be fully polymerized and set throughout the entire cross section within 15 minutes of initial mixing. Castings of this sort are not biocompatible for several days due to the emission of acetic acid. If placed in an aqueous environment too quickly with a living tissue, tissue damage would inevitably result. Thicker sections require longer waiting periods, but we found that storage on the shelf for at least one week prior to use was sufficient to achieve biocompatibility with no noticeable effects on the explanted tissues. The cylindrical mounting boss permitted different tail assemblies to be inserted or removed, simply by pressing the boss into a cylindrical receptacle in the Delrin spine. A 0.07 mm diametric interference fit was used. The tail mold allowed different tail lengths and base thicknesses to be cast by simply changing the two Teflon plates that formed the sides of the triangular mold cavity, allowing easy adjustment of the tail compliance. The final tail geometry resulted in sufficient compliance to allow the tail to assume a sigmoidal shape, with a wave traveling caudally when actuated in water at frequencies above ~2 Hz. After design iterations, the spring constant of the compliant tail was 0.42 Newton*cm/radian, and the stiffness remained the same throughout all subsequent experimental sessions. The onboard electronics were based upon a previously published design for an implantable muscle stimulator [ 16 ], and thus the circuit architecture will not be reproduced here. Several minor modifications were made to the circuit hardware. The MAX630 DC-DC converter was not used. The system was powered by two 3 Volt, 48 mAh tabbed lithium batteries (Panasonic # BR1225-1VC) connected in series. The actual operating voltage of the batteries was ~2.8 V [ 16 , 17 ]. The embedded microprocessor (PIC16C54A, SSOP package), was operated from only the first battery in the series, at 2.8 VDC with a 40 kHz crystal oscillator to minimize the power consumption of the device [ 16 , 17 ]. The stimulator output buffer was powered by both lithium batteries in series and was constructed using logic level HEXFETs (International Rectifier # IRF7105) to provide capacitive discharge square pulse stimulation to each actuator at ~5.6 V. The pulse was sufficient to elicit a sub-maximal contraction of each semitendinosus muscle. To minimize the size of the on-board control electronics, a PC board was not used, rather each component was soldered by hand directly to the leads of each IC chip with jumper wires added as necessary. Stimulation was controlled remotely via a unidirectional infra red (IR) link from a hand-held command module. The on-board fixed stimulation parameters were: amplitude = 5.8 V (alternating bipolar) [ 16 ], frequency = 80 Hz, pulse width = 100 μsec. The remote command module allowed for manual control of the onset of stimulation, the train duration (0 to 2550 ms, in 10 ms increments), the dwell time (time between stimulus trains (0 to 2550 ms, in 10 ms increments), and a setting to control either alternating stimulation between the antagonistic actuators for forward motion, or continuous one-sided muscle activation for steering control. The electrodes were fashioned from medical grade TFE coated 40 AWG stainless steel multi-strand electrode wire (Cooner Wire). The distal ends were stripped to allow the electrode wire to be wrapped around each muscle, as described previously [ 16 ]. The finished on-board control modules were encapsulated using electronic grade epoxy, followed by 6 coats of Dow silicone elastomer #734 dispersed with toluene, according to the method described previously [ 17 ]. During robotic swimming operations, the fuel sources were a glucose-bearing ringer solution (~2 g/L glucose), and lithium batteries to power the embedded microcontroller and stimulator system. Due to the micro-power electronic design, the estimated battery life for the system was ~21 days (assuming a 10% stimulation duty cycle) [ 16 , 17 ]. The semitendinosus muscle was selected primarily due to its convenient size and tendon anatomy. It is easily dissected with both proximal and distal tendons attached. The proximal tendons of each semitendinosus muscle were sutured to the rigid Delrin head piece, and the distal tendons were sutured to the lateral mounting flange on each side of the tail base using 5-0 braided silk suture (Figure 1 ). The muscles were mounted symmetrically on opposite sides of the robotic platform to act as antagonists, providing a single degree-of-freedom reversible actuator for the base of the compliant tail. Muscle length was adjusted manually during installation by sliding the sutures through the tail flange to achieve the desired muscle length. Both muscle lengths were adjusted to set each muscle at rest length when the tail was in its neutral position. With no muscle force applied, the restoring torque of the silicone hinge-joint returned the tail to the neutral position, thus both muscles were at their rest length when neither was activated. This important feature is essential for muscle maintenance, as muscles maintained at stretched lengths are known to degenerate more rapidly than muscles held at lengths corresponding to the ascending limb of the length-tension curve [ 14 ]. Robotic Experiments and Performance Characterizations Two robotic platforms were evaluated in terms of muscle actuator performance, swimming efficiency and locomotory maneuverability. Each robotic platform was designated "B1x", where "x" indicated the build-up, serialized as "a, b, c, ..." for each subsequent pair of explanted frog muscles. Two build-ups were constructed, B1a and B1b, each with a separate pair of freshly explanted frog semitendinosus muscles. Prior to swimming evaluations, two liters of ringer solution (ringer composition in Methods: Muscle Removal and Maintenance ) were poured into the test tank, providing a fluid depth of approximately 2.1 cm, enough for the robot to swim without touching the bottom of the tank. The tank temperature was measured but not controlled, and was allowed to stabilize at room temperature, approximately 22°C for the duration of each experiment. The ringer solution was aerated with unfiltered room air using 4 standard porous stone fish tank aerators, one placed at each corner of the tank, and connected to an aquarium aeration pump via silicone tubing. Aeration was discontinued briefly before each test run to minimize turbulence in the test tank. For each robotic build-up, or for each pair of explanted semitendinosus muscles, the test tank ringer solution was not replaced or replenished for the entirety of the robotic experimental session. Muscle installation was carried out with the robotic platform partially immersed in ringer solution using #5 forceps (Fine Science Tools). After installation was complete, the muscles were allowed to acclimate for a period of approximately 5 minutes before stimulation. The robot was manually placed to allow forward motion through the bath, and muscle stimulus parameters, specifically stimulus train duration and dwell time, were varied manually until the maximum swimming velocity was achieved. To increase swimming speed, dwell period was decreased and train duration was increased until further decreases in dwell time or further increases in train duration did not result in additional increases in forward swimming speed. During experimentation, swimming speed was determined by measuring the amount of time required for the robot to swim across a known, fixed distance. Once the maximum swimming speed was achieved, the ventral view of the swimming robot was filmed (Sony Model #DCR-TRV820; 30 frames/sec), and the film was then digitized to determine tail-beat frequency, tail amplitude, and the wave speed and wave length of the propulsive body wave. In addition to forward straight-line swimming, muscle stimulus parameters were varied to investigate turning maneuvers. At a maximum forward swimming speed, the robot's open loop, alternating stimulation pattern between the antagonistic actuators was switched to a continuous one-sided muscle activation for steering control, causing the robot to turn in the direction of the single stimulated muscle (a medial turn resulting from one-sided medial muscle stimulation). Here again, the ventral view of the swimming robot was filmed, and the film was then digitized to determine the maximum turning radius. For each tail-beat period, at least 10 video frames were captured, separated in time by 33 ms, depending on the swimming speed of the robot. A customized software program was used to digitize 10 points on each side of the outline of the ventral silhouette of the robot, for a total of 20 points for each image. A series of cubic spline functions were used to draw the best-fit line along these points [ 18 , 19 ], and a midline was constructed. Tail-beat frequency was measured by tracking a digitized point on the tail tip from the ventral view over the course of one tail-beat cycle and dividing by the elapsed time. Tail amplitude was determined by measuring the tip-to-tip linear distance at the two extremes of tail excursion and then dividing by two. As described by [ 20 ], mean propulsive wavelength was measured directly from the reconstructed midlines as the distance between two successive peaks present on the robot's body. Propulsive wave speed was calculated by dividing the distance between the anterior most point of the body exhibiting undulation and the tail tip by the time required for the crest of the wave to pass through these points. To estimate the overall mechanical swimming efficiency of each robotic build-up, we calculated the robot's slip value, a dimensionless velocity [ 21 ]. A high slip value indicates a larger contribution to rearward, thrust-producing forces than lateral forces. Slip was calculated by dividing the robot's steady state swimming velocity by its propulsive wave speed. To estimate muscle actuator performance at the maximum swimming speed, muscle strain and shortening velocity were estimated using the tail-beat frequency and amplitude measurements taken from the digitized films. After the swimming experiments were finalized, the change in linear distance between the robot's muscle attachment points was measured when the robot's tail was re-positioned from a neutral, straight position to the tail amplitude posture measured during straight-line swimming. As an estimate of peak muscle shortening strain, this linear-distance change was then divided by the muscle's resting length, or the muscle belly length when the tail was held straight (resting length measurement protocol defined in Methods: Muscle Removal and Maintenance ). Still further, to estimate muscle-shortening velocity at the maximum swimming speed, the measured linear-distance change between muscle attachment points was divided by the time required for the tail to re-position from a neutral, straight position to the tail amplitude posture measured during straight-line swimming. This time period was measured from the digitized films and was equal to approximately one quarter of a tail-beat period. For the turning maneuvers, the turning radius was estimated from the ventral video images by tracking the spatial trajectory of a point midway between the tail tip and the nose of the robot, a distance 6 cm from the tail tip along the midline of the robot when the tail assumed a neutral, straight orientation. The turning radius was the radius of a circle with an arc curvature equivalent to the midpoint trajectory curvature. Semitendinosus Contractile Experiment: Maximum Shortening Velocity To estimate the contractile efficiency of the robotic muscle actuators at the maximum swimming velocity, a separate experiment was conducted to determine the maximum shortening velocity of freshly dissected semitendinosus muscles of comparable size and rest length to that of the muscles employed in robotic build-ups, B1a and B1b. Six freshly dissected semitendinosus muscles were placed in a muscle characterization apparatus (Aurora Model 305B) and isotonic contraction experiments [ 22 ] were conducted to measure the muscles' maximum shortening velocity. The contractile experiment was conducted at the same temperature as the robotic experiments, or 22°C. Results Robotic Performance Characterizations For the B1a and B1b robotic swimmers, the locomotory performance parameters at maximum swimming velocity are summarized in Table I . Table I also includes the muscle actuator mass and rest length for each robotic build-up. For both robot B1a and B1b, the total muscle mass did not exceed 6% of the total mass of the robot (B1a = 4.8%; B1b = 5.3%). Even with such a low relative actuator mass, swimming robots B1a and B1b achieved top speeds greater than 1/4 and 1/3 body lengths per second, respectively (here the robot's total length, 12 cm, was used as the normalization factor). For both robotic swimmers, forward swimming speed was readily controllable simply by decreasing the dwell period or by increasing the train duration. The maximum steady state, forward swimming speed was achieved with alternating actuator contractions of 110 ms train duration, with 40 ms dwell periods between each stimulus train, resulting in 3.1 tail-beats per second. Further increases in the stimulus train duration or further decreases in the stimulus dwell time did not result in additional increases in forward swimming speed. Each robotic build-up was capable of the following controlled maneuvers: forward accelerations, decelerations, steady state gliding, and turning to the right or left. The robot was capable of surface swimming only, so all maneuvers were restricted to 2-dimensions. Turning was accomplished after forward momentum had been established by continuously activating only one actuator. The minimum gliding turn radius was 400 mm as estimated from the digitized video images of the robot's midpoint trajectory. After swimming the full length of the test tank, the robot was manually repositioned to the opposite end of the tank where it began, once again, to swim across the tank width. Typically, a period of swimming activity (~3 min) was followed by a period of swimming inactivity (~30 min). Due to muscle fatigue, periods of inactivity were required to restore the robot's peak swimming velocity to at least 75% of its maximum value measured during the first session of robotic swimming (first 10 minutes of the robot's lifespan). Robot B1a swam for a sum total of 45 minutes over a 7.5 hour lifespan (10% duty cycle), after which its swimming velocity degraded below 75% of its maximum value even after a 30 minute period of swimming inactivity. In distinction, robot B1b swam for a much longer period – a sum total of 4 hours over a 42 hour lifespan (10% duty cycle) before its velocity degraded below 75% of its maximum value following a 30 minute period of swimming inactivity. To compare the overall swimming efficiency of each robotic build-up, we calculated the propeller efficiency using the measure of slip (swimming velocity/ propulsive wave speed) (Table I ). In a steady-state condition, at the maximum forward swimming speed, slip values for robotic build-ups, B1a and B1b, were 0.26 and 0.32, respectively. By comparison, slip values generally increase with swimming speed in fish, ranging from 0.2 to 0.7 in most fish [ 19 , 20 ]. The mechanical swimming efficiency of robots B1a and B1b, as determined by their respective slip values, were within the biological efficiency range. Maximum Shortening Velocity and the V/V max Ratio at Maximum Swimming Speed In a separate experiment from the robotic investigations, six freshly dissected semitendinosus muscles (mass = 0.34 ± 0.04 g; rest length = 30 ± 1 mm; Mean ± S.E., N = 6 muscles) produced a maximum shortening velocity, V max , of 78 ± 3 mm s -1 (Mean ± S.E., N = 6 muscles) in isotonic contractions. At the maximum swimming speed, the muscle actuators within robots B1a and B1b experienced a shortening velocity of 25 mm s -1 (Table I ), giving a V/V max ratio of 0.32, an intermediate contraction velocity where muscle typically produces peak power and efficiency [ 7 ]. Discussion Although a great deal of research has been conducted to advance a synthetic actuator technology with muscle-like properties, engineering science has not yet produced a motor system that can mimic the contractility, energetics, scalability and plasticity of living muscle tissue [ 1 , 2 ]. In this investigation, we examine the feasibility of using animal-derived muscle as an actuator for artificial devices. We construct a simple robotic platform powered by explanted living amphibian muscle and controlled by an embedded microcontroller via an infra red data link. Using an open loop control and a simple interface design, we present preliminary data that suggests that living muscle might one day be employed as a practical, controllable actuator. Hybrid robot B1b remained active for up to 42 hours, and during that time, performed basic swimming maneuvers such as starting, stopping, turning and straight-line swimming at speeds exceeding 1/3 body lengths per second. The muscle-actuated swimming robot also offered a reasonable swimming efficiency, as indicated by a slip value of 0.32 (see Table I ). Muscle Fiber Type and Control The frog semitendinosus muscles employed in the robot were comprised predominantly of fast-twitch muscle fibers, and therefore provided higher mechanical power, at the expense of being considerably more fatigable, than would have been achievable using a slow-twitch muscle of comparable size. Ideally, a biomechatronic swimming robot would incorporate several muscle fiber types to permit both explosive as well as low-power locomotion and maneuvering. For the robotic platform of this investigation, it is important to note that the stimulation was non-physiologic in many ways. Each muscle was stimulated in bulk, with all fibers being subjected to approximately the same electric field. In living muscle in vivo, individual motor axons innervate one or more muscle fibers, establishing the fundamental neuromotor functional unit: a motor unit . In a sophisticated biomechatronic system, a motor-unit level of control would be desirable (fast vs. slow), both for controllability and for tissue phenotype maintenance. Tissue Failure Modes In this study, the performance of the muscle actuators eventually degraded to the point where they were no longer effective mechanical actuators. Several factors contributed to the observed tissue degradation. To begin with, explanted muscle generally has a very finite functional life expectancy [ 14 , 15 ], usually less than one day. Excluding such transient failure modes as metabolic muscle fatigue, the major failure modes of muscle in vitro generally fall into one of the following categories: (1) core necrosis due to lack of oxygenation/capillary perfusion and large diffusion distances, (2) sepsis, (3) exogenous toxicity, (4) electro-chemical damage resulting from excessive electrical stimulation, (5) accumulated contraction-induced injury, (6) sarcomeres heterogeneity leading to loss of thick and thin filament overlap in regions of muscle fibers (exacerbated by prolonged periods at or above the optimal length for force generation), and (7) direct mechanical damage to the muscle from external sources, such as the robot frame, attachment hardware, or electrodes. For the tissue-actuated device of this investigation, several design considerations were made to minimize many of these failure modes. The bath was aerated to assist oxygen delivery to the tissues, although this strategy would only be helpful to the outer shell of muscle fibers no greater than ~200 μm from the surface. In addition, the level of muscle cell depolarization was kept to a minimum in order to limit electro-chemical damage [ 16 ]. Still further, the muscle actuators were attached to the robot frame at rest length in order to minimize the risk of excessive muscle strains and sarcomere heterogeneity. Clearly, when looking to the future, other failure modes must be considered when very long periods of ex vivo tissue maintenance are necessary. These include loss of muscle excitability and mass, phenotypic drift, and de-differentiation of the muscle from desired adult muscle phenotypes. Muscle Actuator Source: Engineered Muscle versus Explanted Tissue Even though organogenic mechanisms are poorly understood, it is nonetheless possible to engineer functional muscle organs from individual cells in culture [ 23 - 26 ], but currently these tissue constructs have several practical limitations that limit their usefulness as living actuators. Among these limitations are: (1) low contractility, similar to that during early stages of muscle development, (2) low excitability, thus requiring large amounts of electrical energy to adequately stimulate the tissue to contract, (3) the lack of perfusion, which limits the tissue cross section to a maximum radius of approximately 200 μm, and (4) the lack of suitable tissue interfaces, both neural and mechanical. Given such technological limitations, we chose in this study to employ explanted muscle tissues for robotic actuation. However, once these technical hurdles are overcome, engineered muscle actuators might offer important advantages to the construction of biomechatronic robots. Future Work The results of this investigation, although preliminary, suggest that some degree of ex vivo robustness and longevity is possible for natural muscle actuators if adequate chemical and electromechanical interventions are supplied from a host robotic environment. Clearly, an important area of future research will be to establish processes by which optimal intervention strategies are defined for a given hybrid-machine task objective. Another important area of research will be tissue control. It has been established that natural muscle changes in size and strength depending on environmental work-load, and when supplied with appropriate signals, changes frequency characteristic or fiber type [ 9 - 11 ]. Hence, an important area of future work will be to put forth strategies by which muscle tissue plasticity can be monitored and controlled. Finally, strategies must also be devised to control the force and power output of muscle, in the context of robotic systems, through the modulation of electrical pulses to the muscle cell. To achieve the long-term objective of functional, muscle-actuated robotic and prosthetic devices, we feel controlling machine movements through electrical stimulation, harnessing muscle tissue plasticity, and maintaining ex vivo contractility are critical areas for future research. Conclusion In this paper, we ask whether muscle tissue explants can be employed as mechanical actuators for robots in the millimeter to centimeter size scale. Using a very simple control and interface design, we present preliminary data that suggests that living muscle might one day be employed as a practical, controllable actuator. The robot of this investigation remained active for up to 42 hours, and during that time, performed basic swimming maneuvers such as starting, stopping, turning and straight-line swimming at speeds exceeding 1/3 body lengths per second. It is our hope that this work will lead to further studies of tissue actuated robots and prostheses that will result in an even wider range of biomechatronic machine capabilities.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544953.xml
516026
Chloroplast DNA rearrangements in Campanulaceae: phylogenetic utility of highly rearranged genomes
Background The Campanulaceae (the "hare bell" or "bellflower" family) is a derived angiosperm family comprised of about 600 species treated in 35 to 55 genera. Taxonomic treatments vary widely and little phylogenetic work has been done in the family. Gene order in the chloroplast genome usually varies little among vascular plants. However, chloroplast genomes of Campanulaceae represent an exception and phylogenetic analyses solely based on chloroplast rearrangement characters support a reasonably well-resolved tree. Results Chloroplast DNA physical maps were constructed for eighteen representatives of the family. So many gene order changes have occurred among the genomes that characterizing individual mutational events was not always possible. Therefore, we examined different, novel scoring methods to prepare data matrices for cladistic analysis. These approaches yielded largely congruent results but varied in amounts of resolution and homoplasy. The strongly supported nodes were common to all gene order analyses as well as to parallel analyses based on ITS and rbcL sequence data. The results suggest some interesting and unexpected intrafamilial relationships. For example fifteen of the taxa form a derived clade; whereas the remaining three taxa – Platycodon , Codonopsis , and Cyananthus – form the basal clade. This major subdivision of the family corresponds to the distribution of pollen morphology characteristics but is not compatible with previous taxonomic treatments. Conclusions Our use of gene order data in the Campanulaceae provides the most highly resolved phylogeny as yet developed for a plant family using only cpDNA rearrangements. The gene order data showed markedly less homoplasy than sequence data for the same taxa but did not resolve quite as many nodes. The rearrangement characters, though relatively few in number, support robust and meaningful phylogenetic hypotheses and provide new insights into evolutionary relationships within the Campanulaceae.
Background The Campanulaceae sensu stricto are a nearly cosmopolitan angiosperm family consisting of latex-bearing, primarily perennial herbs or occasional subshrubs that typically have alternate leaves, sympetalous corollas, inferior ovaries, and capsular fruits. Allied to the Campanulaceae are the Lobeliaceae, Cyphiaceae, Cyphocarpaceae, Nemacladaceae, Pentaphragmataceae, and Sphenocleaceae; at times, all of these taxa have been included in the Campanulaceae at varying taxonomic rank by different authors (Table 1 ). Taxonomic treatments lack consensus (Table 1 ) and phylogenetic work has only recently been attempted. Campanulaceae in the strict sense are recognized as 600 [ 1 ] to 950 [ 2 ] species distributed among 35 [ 1 ] to 55 [ 2 ] genera. Generic circumscription and intrafamilial classification vary widely according to author. Within the family as few as two [ 3 ] and as many as 18 [ 4 ] tribes have been recognized (Table 1 ). Fedorov's more recent work [ 5 ] recognized eight tribes (Table 1 ), but only included taxa present in the former Soviet Union. Although Kolakovsky's treatment of Old World Campanulaceae [ 4 ] is the most recently published attempt to produce a more complete intrafamilial classification of the Campanulaceae (Table 1 ), the scope of the work is limited compared to that of either A. de Candolle [ 3 ] or Fedorov [ 5 ]. In all treatments, the Campanuleae and Wahlenbergieae (at whatever rank) are typically the largest, most inclusive taxa, with segregate tribes consisting of only one to a few genera. Table 1 Classification systems of Campanulaceae. All major intrafamilial subdivisions are included (level of subdivisions indicated by number of dashes) but only those genera sampled in this study are included. If sampled genera are not listed, the genus was not recognized by the author but, rather, was subsumed into one of the listed genera. A deCandolle AP deCandolle Schönland Federov Takhtajan Kovanda Kolakovsky 1830 [3] 1839 [53] 1889 [45] 1972 [5] 1987 [2] 1978 [1] 1987 [4] Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae -Subtribus I -Wahlenbergieae - Lobelioideae -Sphenocleoideae - Cyanantheae - Campanulinae - Prismatocarpoideae Jasione Jasione -Cyphioideae -Campanuloideae Cyananthus Campanula Prismatocarpus Codonopsis Platycodon -Campanuloideae -- Campanuleae -Wahlenbergieae Symphyandra Roella Platycodon Codonopsis --Pentaphragmeae Campanula Wahlenbergia Legousia - Canarinoideae Wahlenbergia Wahlenbergia --Sphenocleae Symphyandra Edraianthus - Wahlenberginae - Wahlenbergoideae Prismatocarpus Prismatocarpus --Campanuleae --Peracarpeae Jasione Wahlenbergia -- Wahlenbergieae Roella Roella Campanulinae --Ostrowskieae Codonopsis Codonopsis Jasione -Subtribus II Edraianthus Symphyandra --Michauxieae Merciera Cyananthus Wahlenbergia Petromarula -Campanuleae Trachelium --Phyteumateae Roella Roella Codonopsis Campanula Petromarula Campanula Asyneuma Prismatocarpus Edraianthus Platycodon Trachelium Campanula Wahlenberginae Legousia -Platycodoneae Jasione Cyananthus Symphyandra Trachelium Cyananthus -- Wahlenbergieae Platycodon -Platycodinae --Azorineae Musschia Symphyandra Jasione Codonopsis Musschia Platycodon -- Musschieae -incertae sedis Musschia Prismatocarpus --Edraiantheae - Campanuleae Musschia Merciera -Merciereae Merciera Edraianthus Campanula -- Echinocodoneae Merciera Edraianthus -- Jasioneae Legousia -- Annaea Wahlenbergia Jasione Triodanis --Muehlbergelleae Codonopsis - Michauxieae --Theodorovieae Roella -Phyteumaeae --Gadellieae Platycodinae Asyneuma --Ostrowskieae Platycodon Trachelium - Campanuloideae Musschia Petromarula -- Campanuleae -Peracarpeae Campanula Symphyandra Trachelium --Phyteumateae --Peracarpeae --Sergieae --Michauxieae -- Neocodoneae Asyneuma Legousia -- Edraiantheae Edraianthus --Sachokieleae --Mzymteleae The most comprehensive treatment of the Campanulaceae remains the monograph of A. de Candolle [ 3 ], who recognized two groups corresponding to the Wahlenbergieae and Campanuleae (Table 1 ). Simple basal leaves and simple, alternate or occasional whorled, cauline leaves that are often different in shape than the basal leaves, characterize the Campanuleae in de Candolle's sense. Flowers are solitary or borne in cymes or racemes, and have five corolla lobes that are mostly fused proximally. The inferior ovary usually has 3–5 carpels and develops into a capsule that mostly dehisces by lateral pores (rarely a berry). The Wahlenbergieae are mostly perennials characterized by simple, alternate, cauline leaves. Flowers are solitary or borne in cymes or heads, and petals may be free, proximally fused, or distally fused. The ovary is inferior, semi-inferior, or superior, and consists of two, three, or five carpels. The fruit is generally a capsule dehiscing by apical pores or valves (rarely a berry). Both groups have five stamens with filaments that are often proximally dilated and anthers with introrse dehiscence; nectaries are generally present, and many ovules are attached to axile placentae. The entire family is characterized by secondary pollen presentation in which protandry is combined with a close association of anthers around the style and introrse pollen discharge onto the style for presentation to pollinators. This syndrome is similar to that found in Lobeliaceae and Asteraceae, but invaginating stylar hairs are unique to the Campanulaceae. Capsule characters vary considerably and provide the basis for most intrafamilial classification schemes. Campanuleae typically include taxa with capsules dehiscing by lateral pores, whereas Wahlenbergieae usually include taxa with capsules dehiscing by apical valves. Ovary characters, such as carpel number and position, have also been important in traditional classifications. For example, the monotypic tribe Platycodoneae [ 6 ] or subtribe Platycodinae (Table 1 ) is sometimes segregated. It is defined by carpels that are equal in number to and alternate with the calyx lobes, whereas in Campanuleae and Wahlenbergieae the carpels are often fewer than the calyx lobes, or if the same in number then opposite them [ 1 , 7 , 8 ]. Little correlation appears to exist among diagnostic features; therefore there is considerable taxonomic disagreement among classifications. In certain instances it is difficult to discern the rationale behind tribal placement of individual genera. The high level of disagreement among both inter- and intrafamilial classifications of the Campanulaceae indicates that phylogenetic assessment of the family is needed. Cosner, in her thesis [ 9 ], included an early version of a portion of the work described here, and Eddie, in his thesis [ 10 ] developed phylogenetic hypotheses based on ITS sequence data and morphology. An expanded version of the ITS work has been published [ 11 ] but leaves some major lineages unsampled and the relationships among some major groups are unresolved or poorly supported. Further phylogenetic work is clearly warranted. The chloroplast genome has proven to be a useful tool for phylogenetic reconstruction. Chloroplast DNA (cpDNA) of land plants is highly conserved in nucleotide sequence as well as gene content and order; its relatively slow rate of evolution makes it an excellent molecule for phylogenetic and evolutionary studies [ 12 ]. Chloroplast genomes of photosynthetic angiosperms average about 160 kilobase pairs (kb) in size; the circular chromosome is divided by two copies of a large (in angiosperms usually about 25 kb) inverted repeat (IR) into large and small single copy regions (LSC and SSC, respectively) [ 13 , 14 ]. Restriction site mapping, gene sequencing, and analysis of gene order rearrangements have been used to study cpDNA variation for phylogenetic investigations [ 12 ]. Here we use the distribution of gene order changes in the chloroplast genomes of the Campanulaceae to estimate phylogenetic relationships in the family. Generally, major gene order changes are rare. Therefore, when they occur, such mutations are extremely useful as phylogenetic markers because they are readily polarized and typically lack homoplasy [ 15 - 17 ]. Four categories of cpDNA gene order rearrangements have been proposed: 1) inversions, 2) insertions or deletions, 3) IR expansion or contraction or loss, and 4) transpositions; all of which may have occurred during chloroplast genome evolution in the Campanulaceae [ 18 ]. When rearrangements have been discovered elsewhere, they are generally few and easily characterized. The distributions of such characters make effective markers of monophyletic groups. For example, both the loss of one copy of the IR and inversions are extremely useful characters in legume phylogeny [ 19 , 20 ], defining large clades within the family. Other examples of phylogenetically informative inversions are found within Asteraceae [ 21 ], Ranunculaceae [ 22 , 23 ], ferns [ 24 , 25 ], and vascular plants [ 26 ]. Many other examples could be cited. The earlier work of Cosner [ 9 , 18 ] and Knox [ 27 , 28 ] characterized some chloroplast genomes of the Campanulales and identified a number of rearrangements relative to the consensus gene order of angiosperms found in tobacco. Members of the Lobeliaceae exhibit multiple rearrangements but are less rearranged than the Campanulaceae. Three rearrangements may be shared between the two families – a loss of the accD gene, the expansion of the inverted repeat into the small single copy region, and, perhaps, an inversion of the region corresponding to tobacco probes 40–44. Then, within the Campanulaceae, more than 40 inversions, more than eight putative transpositions, two additional gene losses, additional IR expansion or contraction events and 18 large insertions greater than 5 kb in size may have contributed to observed differences among the chloroplast genomes sampled [ 9 ]. Due to this unprecedented number of gene order mutations, it is not possible to unambiguously determine the evolutionary order of most events or in some cases to even define the events themselves. This complex situation poses special problems for using these rearrangements to estimate phylogenetic relationships. In this paper we develop alternative character codings for the data and compare the results of parsimony analyses of the different data sets. In addition, we compare the ability of the gene order data to support robust phylogenetic hypotheses to that of sequence data from rbcL and ITS. Finally, the phylogenetic implications of the cpDNA rearrangement data for the Campanulaceae are discussed. Results Our data indicate that the eighteen mapped Campanulaceae chloroplast genomes (Table 2 ) are drastically rearranged relative to those of other land plants (Fig. 1 ). The tobacco cpDNA gene order represents the consensus gene order for angiosperms [ 13 , 15 ]. Therefore rearrangements in Campanulaceae chloroplast genomes were identified relative to tobacco. Because characterizing specific mutational events was not always possible three different coding methods (Matrix 1, 2 and 3) were developed. Matrix 1 coded all gene order changes as endpoints (derived adjacencies, relative to tobacco, were identified and scored for presence/absence). Matrix 2 and 3 involved recoding some endpoint characters to recognize 31 specific mutations. Matrix 2 and 3 were analyzed with and without weighting. See Methods for additional details on character encoding and analyses. Table 2 Species of Campanulaceae mapped for chloroplast DNA structural rearrangments. Species Source Voucher a Adenophora confusa Nannf. R.C. Haberle 179 TEX Asyneuma virgatum (Labill.) Bourm. Berlin-Dahlem b 0104 Campanula elatines L. T. Ayers 88–287 BH Codonopsis viridis Wall. T. Ayers 88–229 BH Cyananthus lobatus Wall. ex Benth. M. Cosner 179 OS Edraianthus graminifolius (L.) A.DC. T. Ayers 88–195 BH Jasione heldreichii Boiss. & Orph. T. Ayers 88–208 BH Legousia falcata (Ten.) Fritsch ex Janch. Berlin-Dahlem b 0143 Merciera tenuifolia (L.f.) A. DC. K. Steiner 2445 OS Musschia aurea Dumort T. Ayers 88–274 BH Petromarula pinnata (L.) A. DC. T. Ayers s.n. c BH Platycodon grandiflorus (Jacq.) A. DC. T. Ayers 88–216 BH Prismatocarpus diffusus (L.f.) A. DC. K. Steiner 2448 OS Roella ciliata L. T. Ayers s.n. c BH Symphyandra hofmannii Pant. T. Ayers 88–225 BH Trachelium caeruleum L. M. Cosner 173 OS Triodanis perfoliata (L.) Nieuwl. M. Cosner 178 OS Wahlenbergia gloriosa Lothian T. Ayers 88–217 OS a, abbreviations for herbaria: BH = Bailey Hortorium (Cornell University, Ithaca, NY); OS = Ohio State University Herbarium (Columbus); TEX = University of Texas Herbarium (Austin) b, Botanischer Garten and Botanisches Museum, Berlin-Dahlem c, s.n. = sin numero (no number assigned by collector) Figure 1 Linearized cpDNA maps for 18 species (Table 2) of Campanulaceae (plus tobacco) showing order in which the consecutively numbered tobacco probes hybridized. Lines under maps indicate location and extent of IR. Asterisks indicate the position of the putative 23S rDNA duplicative transposition; parenthetical asterisk (*) indicates partial deletion/divergence of the 23S rDNA transposition. Size and location of large insertions designated by "i" followed by size in kb (insertions less than 5 kb not shown). Seventy-nine variable characters were included in the endpoints only matrix (Matrix 1). Forty-two of the derived character states were unique to a single taxon and 37 were phylogenetically-informative. Six trees of 97 steps were obtained with consistency indices of 0.81 with autapomorphies included and 0.67 with autapomorphies excluded (CI = 0.81/0.67). Ten nodes were common to the six shortest trees (Fig. 2 ). Eight of those ten nodes have bootstrap values (BS) greater than 50, but BS exceeded 90 for only three nodes. Figure 2 One of six shortest trees obtained in the maximum parsimony analysis of Matrix 1. Tree length is 97 steps; consistency index is 0.81 (with autapomorphies, 0.66 without). The number of character changes is given above the branches and bootstrap values (where greater than 50) are given below. Arrows indicate nodes that collapse in the strict consensus of all six shortest trees. To construct Matrix 2 and Matrix 3, we interpreted endpoints as events where possible. Under our interpretation, several types of rearrangements contributed to cpDNA evolution in the family, including multiple inversions (scored primarily as endpoints), five IR expansion or contraction events, eight transpositions, two deletions, and 14 large insertions greater than 5 kb in size (Table 3 ). Although transposition probably does occur, at least occasionally, in the chloroplast genome [ 29 ], it is not a common mechanism of rearrangement. Still, in some instances transposition could explain rearranged gene orders with fewer steps than multiple inversions and so we hypothesized transposition events in some cases. Matrix 2 and 3 each were composed of 84 variable characters of which thirty-one and thirty-four, respectively, were parsimony informative. Table 3 List of chloroplast DNA rearrangement characters. Numbers refer to tobacco cpDNA hybridization probes. Endpoints are given as novel probe adjacencies. Other rearrangement types are indicated as: T = transposition (T' = secondary transposition [14] of most of 53–56); I = inversion; i = insertion; D = deletion (or divergence); IRc and IRe = IR contraction or expansion, respectively (followed by single copy region affected). Characters marked with asterisks (*) are those rescored in Matrix 3 relative to Matrix 2. *1. 11/60 22. T (53,54) 43. 40/56 64. i (15) *2. 56/53 23. T (53–56) 44. 39/37 65. IR e (LSC) *3. 49/37 24. 98/28 45. 37/44 66. T (5–8) 4. 40/35 25. 53/98 46. 56/61 67. T (6–9) *5. 28/11 26. T' (53–56) 47. 76/96 68. i (9) 6. 15/76 27. 49/39 48. 77/106 69. i (18) 7. 60/27 *28. 37/40 49. i (5) 70. l (60–61) *8. 26/44 29. 56/39 50. 39/25 71. T (28) 9. 41/47 30. 53/40 51. 9/37 72. T (16–17) 10. 48/36 31. 37/28 52. 48/56 73. i 7 11. 35/25 32. 40/26 53. 49/61 74. IR e (SSC) 12. 16/90 33. 27/44 54. 5/29 75. i (8) 13. T (93) 34. 8/40 55. 50/28 76. i (9) 14. 84/76 35. 39/26 56. 26/50 77. i (9) 15. 84/90 36. 27/11 57. 40/16 78. i (15) 16. 10/49 37. 56/36 58. 25/10 79. i (16) 17. 53/28 38. 44/10 59. 15/56 80. D (93) 18. 37/60 39. 9/53 60. IRc (LSC) 81. i (5) 19. 56/11 40. 49/36 61. IR e (SSC) 82. i (6) 20. I (16–17) 41. 8/36 62. IR e (LSC) 83. i (19) 21. 56/27 42. 9/40 63. D (45–46) 84. i (5) The unweighted analysis of Matrix 2 produced 241 equally parsimonious trees of 93 steps (CI = 0.90/0.79). The strict consensus of the 241 trees includes six resolved nodes (Fig. 3a ) all six of which were supported by BS values of at least 50 and five nodes were supported at 90% or above. The weighted analysis of Matrix 2 (Fig. 3b ) resulted in 12 equally parsimonious trees of 125 steps (CI = 0.93/0.82). The strict consensus of the twelve trees retains ten resolved nodes. Seven of the ten nodes have BS values over 50 and for five nodes BS ≥ 90. Both analyses of Matrix 3 generated the same two equally parsimonious trees (Fig. 4 ). The lengths of the two trees were 87 steps (CI = 0.97/0.92) or 118 steps (CI = 0.97/0.93) depending on whether unweighted (Fig. 4a ) or weighted (Fig. 4b ) analyses were conducted. Only three endpoint characters are homoplasious in the Matrix 3 analyses (Fig. 5 ). The strict consensus of these two trees retains nine resolved nodes, all nine of which are supported with BS ≥ 50. Six (or five in the weighted analysis) nodes received strong support (BS ≥ 90). Figure 3 Trees obtained in unweighted (a) and weighted analyses (b) of Matrix 2. Fig. 3a shows one of 241 shortest trees of 93 steps; consistency index is 0.90 (including autapomorphies, 0.78 without). Fig. 3b shows one of 12 shortest trees of 125 steps; consistency index is 0.93 (including autapomorphies, 0.80 without). The number of character changes is given above the branches and bootstrap values are given below. Arrows indicate nodes that collapse in the strict consensus of all the shortest trees. Figure 4 The two shortest trees obtained in both the unweighted and weighted analyses of Matrix 3. The trees are 87 steps long without weights and 125 steps when weights are applied. The consistency index of the trees in the unweighted analysis is 0.97 (including autapomorphies, 0.92 without); in the analysis with weights applied the CI is 0.97 (including autapomorphies, 0.93 without). Values given on the upper tree (Fig. 4a) pertain to the unweighted analysis, values on the lower tree (Fig. 4b) to the weighted analysis. The number of character changes is given above the branches and bootstrap values are given below. Arrows indicate nodes that collapse in the strict consensus of the shortest trees. Figure 5 Strict consensus tree of the two equally parsimonious trees from the analysis of Matrix 3 (Fig. 4) showing character changes. Number and type of each change are indicated by e = endpoint, IV = inversion, IS = insertion >5 kb, T = transposition, and D = deletion/divergence. Three of the 84 characters (18, 31 and 37) exhibit homoplasy; each character is an endpoint that changes twice within the tree. Homoplastic changes are shown by the character number below the branch upon which the change occurs. To the right of the tree, brackets and letter/number designations indicate major clades discussed in the text. Symbols at the ends of the terminal branches indicate the geographic distribution of the taxa: □ = eastern Asia; ● = Europe (includes North Africa); ◆ = Americas (primarily North America); * = Southern Hemisphere (mainly Southern Africa). All results (Figs. 2 , 3 , 4 , 5 ) indicate that Codonopsis , Platycodon , and Cyananthus are basal within the family. Analyses on Matrix 2 and 3 support a Codonopsis + Cyananthus sister group relationship and a monophyletic basal clade whereas the Matrix 1 analysis supports a Codonopsis + Platycodon sister group and a paraphyletic basal grade. Neither outcome is very well supported; the alternative scenarios each require only a single additional step in the other data set. Within the fifteen derived taxa some of the relationships are not resolved or resolved but weakly supported. However, some groupings are well supported in all analyses. The South African taxa, Merciera , Prismatocarpus and Roella , form a clade (BS = bootstrap value = 98 - 100). Wahlenbergia is the sister to these three taxa in all analyses with varying levels of support (BS = 82, 60, 69, 99, 99, in the five analyses based on gene order changes). Other groupings include a Symphyandra + Edraianthus clade (BS = 86-91) and Legousia + Asyneuma + Petromarula + Triodanis (BS = 94-100). The five analyses had somewhat different characteristics (Table 4 ). For example, Matrix 1 and Matrix 3 analyses generated fewer equally-parsimonious trees than Matrix 2. The Matrix 1 analyses resolved the most nodes. Matrix 3 analyses exhibited the lowest amounts of homoplasy and supported the highest number of nodes BS ≥ 90. Comparing all results, no nodes with high bootstrap values (BS ≥ 90) were conflicted by other nodes of equally high value. However, there were three instances of incongruence involving nodes of lesser support – Matrix 1 and 3 analyses supported Campanula + Adenophora , whereas Matrix 2 supported Adenophora + Jasione ; Matrix 1 supported Codonopsis + Platycodon (BS = 50), whereas Matrix 2 and 3 supported Codonopsis + Cyananthus (BS = 94-99); and Matrix 1 supported (weakly, BS = 57) the placement of Cyananthus at the base of the derived clade, whereas Matrix 2 and 3 analyses supported the monophyly of the basal group (BS = 56-68). One clade, Legousia + Asyneuma (BS = 87), was recovered only by the Matrix 1 analysis within a clade not further resolved by the other gene order analyses. Table 4 Comparison of characteristics from the different analyses, best values for each characteristic shown in bold. Matrix 1 Matrix 2, no weights Matrix 2 weighted Matrix 3, no weights Matrix 3 weighted rbcL ITS* Number of MP trees 6 241 12 2 2 9 1 Number of resolved nodes in consensus of MP trees 10 6 10 9 9 14 13 Nodes retained in consensus of all trees to 1% longer 6 5 5 6 7 5 2 CI (with / without autapomorphies) 0.81/0.67 0.90/0.79 0.93/0.82 0.97/0.93 0.97/0.94 0.77/0.66 0.69/0.60 Number of nodes BS ≥ 50 8 7 7 9 9 13 10 Number of nodes BS ≥ 90 3 5 4 6 5 5 6 Average bootstrap value of resolved nodes 72 91 72 85 86 78 74 Number characters (PI) 37 31 34 116 195 Total homoplastic characters (number/percent of PI) 15/40.5% 8/25.8% 3/8.8% 63/54.3% 141/71.9% Homoplastic characters with one excess change 12/32.4% 7/22.6% 3/8.8% 48/41.4% 82/41.8% Homoplastic characters with two excess changes 3/8.1% 1/3.2% 0 14/12.1% 41/20.9% Homoplastic characters with three or more excess changes 0 0 0 1/0.1% 18/9.2% *the ITS analysis includes 3 fewer taxa than the others We included sequence data here mainly to allow for a comparison with gene order data in terms of phylogenetic utility. The rbcL data from the same eighteen taxa (Table 5 ) provided 116 parsimony-informative characters that, when analyzed, yielded nine shortest equally-parsimonious trees of 338 steps (C = 0.77/0.66). The strict consensus of the nine trees retained fourteen nodes (fig. 6a ), thirteen of which had BS ≥ 50 and four of which were supported BS ≥ 90. The ITS data of Eddie et al [ 11 ] from taxa equivalent to fifteen of the eighteen mapped taxa (Table 5 ) provided 196 parsimony-informative characters from which a single most parsimonious tree of 716 steps (fig. 6b ) was generated (CI = 0.69/0.60). The tree contains thirteen resolved nodes of which ten had BS ≥ 50 and four had BS ≥ 90. The two sequence data sets had lower CI values than any of the gene order analyses and a higher percentage of homoplastic characters (Table 4 ). The ITS data had especially high levels of homoplasy; the ITS data had a higher percentage of characters that change three or more times in excess than the Matrix 3 analyses had for total homoplastic characters (Table 4 ). In the Matrix 3 analysis only three characters (endpoints) are required to change more than once over the most parsimonious tree; each has one excess change. With the inclusion of the sequence-based analyses, there were additional instances of incongruence between weakly supported nodes: 1) The placement of Musschia and Jasione varies between the rbcL and ITS results (the placement of these taxa is largely unresolved by the gene order data); 2) In both sequence-based trees, Campanula and Adenophora are separate lineages (rather than sister taxa) basal to the Legousia - Asyneuma - Triodanis - Petromarula clade, whereas in the gene order analyses they are allied to Symphyandra - Edraianthus , and Trachelium ; 3) Matrix 1 supports a Legousia - Asyneuma clade, whereas a Legousia - Triodanis clade occurs in the sequence-based trees; and 4) Matrix 1 and ITS support a Codonopsis - Platycodon grouping within the basal clade, whereas rbcL and Matrix 2 and 3 analyses support Codonopsis - Cyananathus. Among these instances of disagreement between weakly supported nodes, there is no general pattern of disagreement between the sequence data and the gene order analyses. And among strongly supported nodes, again, there is complete agreement, among all analyses-sequence and gene order. Table 5 Taxa for which rbcL and ITS data were analyzed. Gene Order/rbcL Species GenBank Accession rbcL ITS "equivalent" taxon GenBank Accession ITS [11] Lobelia cardinalis AY655144 Lobelia tenera AF054938 Adenophora confusa AY655145 Adenophora divaricata1 AY322005 & AY331418 Asyneuma virgatum AY655146 Asyneuma japonica AF183437 & AF18343 Campanula elatines AY655147 Campanula lusitanica AY322025 & AY331438 Codonopsis viridis AY655148 Codonopsis lanceolata AY322048 & AY331461 Cyananthus lobatus L18795 [40] Cyananthus lobatus AY322050 & AY331463 Edraianthus AY655150 Edraianthus AY322052 & AY331465 graminifolius graminifolius Jasione heldreichii AY655151 Jasione crispa AY322059 & AY331472 Legousia falcata AY655151 Legousia speculum- AY322065 & AY331478 veneris Merciera tenuifolia AY655153 No equivalent NA Musschia aurea AY655154 Musschia aurea AY322067 & AY331481 Petromarula pinnata AY655155 Petromarula pinnata AY322069 & AY331482 Platycodon grandiflorus AY655156 Platycodon grandiflorus AY322074 & AY331487 Prismatocarpus diffusus AY655157 No equivalent NA Roella ciliata AY655158 Roella ciliata AY322074 & AY331487 Symphyandra hofmanni AY655159 Symphyandra hofmanni AY322076 & AY331489 Trachelium caeruleum L18793 40 Trachelium caeruleum AY322078 & AY331491 Triodanis perfoliata AY655160 Triodanis leptocarpa AY322079 & AY331492 Wahlenbergia gloriosa AY655161 No equivalent NA Figure 6 Trees obtained from sequence data. The values below nodes are bootstrap percentages; the values above the nodes indicate the number of changes that occur on that branch. Fig. 6a. The one shortest tree (716 steps) based on ITS sequence data. The CI is 0.69 with autapomorphies and 0.60 without. Fig. 6b. One of the nine shortest trees (338 steps) based on rbcL sequence data, CI = 0.77/0.66. The node that collapses in the strict consensus of the nine trees is marked with an arrow. Discussion Phylogenetic analysis of cpDNA rearrangements The relatively large number of gene order mutations that have occurred in the Campanulaceae chloroplast genomes causes difficulties when interpreting their phylogenetic significance. The phylogenetic analysis of such a complex set of cpDNA rearrangements within a group of plants is without precedent. The first problem was simply defining individual mutational events. Although the ideal way to analyze rearrangement data is to determine presence or absence of specific events, in the Campanulaceae, this was not possible in many cases given our present knowledge. Where multiple overlapping rearrangements have occurred between genomes, the two specific endpoints that define a particular inversion may not be determinable. Because of the inherent complexity of the data, we felt a new method of character analysis of the rearrangement data was warranted. Our approach involved coding endpoints, along with more easily defined rearrangements, as characters for different cladistic analyses. Endpoints were defined as two non-contiguous tobacco regions that are now adjacent in the genomes of one or more species. Using endpoints as characters is advantageous. It allows for the incorporation into the analysis of data that could not be used if only unambiguously interpreted events were included. However, using endpoints as characters has several drawbacks, including the inadvertent weighting of certain events over others. Inversions necessarily produce two endpoints, and transpositions three, whereas gene losses and IR boundary changes produce a single endpoint. Therefore inversions would be included twice and transpositions three times if scored as "independent" endpoints rather than events. Plus, if both endpoints of an inversion are still intact in a genome, the inversion is scored twice, if only a single endpoint remains the inversion is counted only once, and if both endpoints have been lost (through further mutation) the inversion will not be included at all. This may represent a problem in the Campanulaceae analyses because there appears to be a mixture of event types and at least some endpoint reuse [ 18 ]. Our inclusion of transposition as a possible mechanism for gene order mutation in the Campanulaceae chloroplast genomes is problematic. Definitive evidence supporting the occurrence of transposition in the plastid genome is lacking. Transposition has been invoked to explain chloroplast DNA rearrangements, for example in "subclover" [ 30 ] and wheat [ 31 , 32 ]. In these cases, transposition has been supported using parsimony arguments (one transposition explaining a change with fewer steps than three inversions) or using the existence of inverted- and direct-repeat sequence motifs near the boundaries of rearrangements [ 33 ]. In Campanulaceae, some lines of evidence in addition to parsimony suggest the possibility of transposition as a mechanism. First, the abundance of rearrangement events within the family suggests some mechanism that facilitates gene order mutation; transposition is one such process. Second, the segment of the genome defined by tobacco probes 53–56 is now located, in most of the derived taxa, within the inverted repeat. The region from which it has been removed appears otherwise undisturbed. In Asyneuma , the 53–56 region has been secondarily removed from the IR and returned to near its original location leaving behind small portions of 53 and 56 in the IR, detectable using southern hybridization. In Wahlenbergia , Merciera , Prismatocarpus and Roella , the 53, 54 portion of the 53–56 block has moved from the IR back to the LSC. One explanation for the high level of rearrangement apparently associated with this segment is that the region contains a transposable element. Third, a possible duplicative transposition is suggested (Fig. 1 ) in Trachelium [ 18 ] and five other taxa [ 9 ]. In addition to a full-length (presumably functional) copy of the 23S rRNA gene, a partial copy is located within ycf1 . Transposition is one manner in which segments of DNA can be both copied and moved within a genome. None of our data are definitive. The observed rearrangements could have taken place as the result of multiple inversions. Therefore, it is important to note that if transposition is not active in the Campanulaceae genome, our phylogenetic results will not be greatly affected. Events coded in Matrix 2 and 3 as single transpositions would be underweighted inversions if incorrectly interpreted. The fact that the analysis of Matrix 1 yields results compatible with those of the matrices that include transpositions suggests that, if our interpretation is erroneous, it does not affect the phylogenetic conclusions. Our three methods of character scoring did yield largely compatible results in our analyses. Relationships that were strongly supported in one analysis were found in all analyses. Events make more desirable characters but they will only improve analyses if the postulated events are the correct ones. Comparing analyses that include event interpretations with endpoint only analyses is one way to determine the phylogenetic effects of the hypotheses of events used. Endpoint only analyses also allow studies that minimize a priori assumptions about the evolutionary events. It is possible that more complex evolutionary scenarios occurred, in which some inversions evolved in parallel, or in which similar gene orders resulted from a different set of inversions. The parsimony analyses may underestimate the number of inversions shared between primitive and advanced genera, because evidence of shared inversions may have been lost. Although we have attempted to produce the simplest evolutionary schemes, it is very possible that longer, more complicated scenarios actually occurred, especially given that the Campanulaceae seem predisposed to cpDNA rearrangements. However, given the congruence of the results among our various analyses, we feel our phylogeny is a reasonable estimate of relationships within the family. Elsewhere, we have analyzed a reduced subset of characters and taxa for the Campanulaceae cpDNA data set using endpoint scoring and constructing trees using breakpoint distances among other methods (e.g., [ 34 ]). Other computational biologists have also used this reduced data matrix to test different methods of phylogeny reconstruction based on gene order data (e.g., [ 35 , 36 ]). These various studies produced trees that are largely congruent with those generated in this paper suggesting that the Campanulaceae cpDNA gene order data are providing a consistent estimate of phylogenetic relationships given any logical method of scoring and analysis. Although the rearrangements in Campanulaceae are complex, the phylogenetic utility of the gene order data is evident. In most previous examples of phylogenetic use of rearrangements the small number of events allowed for the circumscription of only very broad groups [ 15 ]. Because there are so many rearrangements in the Campanulaceae, smaller groups can be identified. This has resulted in the most highly resolved phylogeny as yet developed based entirely on cpDNA rearrangements. Not only do these data support a well-resolved phylogeny but they provide robust support of several nodes. Matrix 3 supports as many nodes at BS ≥ 90 as ITS and more than rbcL . In matrix 3, only three endpoint characters (8.8% of parsimony-informative characters) are homoplastic, each changing one extra time over the tree. In contrast, within ITS, 18 characters (9.2% of parsimony-informative characters) change three or more extra times over the tree, and 71.9% of characters are homoplastic. Presumably because of this high level of homoplasy, only two nodes are retained in the consensus of all ITS trees from the shortest to 1% longer, whereas seven nodes are retained in matrix 3 trees "to 1% longer" – the highest number of any of the analyses. Matrix 3, the matrix in which characters are most interpreted as mutational events, is especially strong in its performance, exceeding both sequence data sets in average bootstrap value per resolved node and CI (in addition to those characteristics just discussed). This suggests that the closer we can get to scoring the actual mutations the stronger gene order data will perform. Although the endpoint only matrix provides useful insights on relationships, we would argue that the extent to which these gene order characters cannot recover the phylogeny is directly related to our ability to define individual mutational events. Phylogenetic implications of the rearrangement data Most traditional classifications of the Campanulaceae are based mainly on capsule dehiscence and ovary position and arrangement. As Kovanda [ 1 ] and Thulin [ 8 ] recognized, classification of the Campanulaceae based on capsule characters alone brings together otherwise radically different taxa. Neither the Campanuleae nor Wahlenbergieae (at whatever taxonomic rank) are monophyletic based on cpDNA rearrangements (Fig. 5 ). Likewise, no traditional classification (Table 1 ) suggests that Codonopsis , Platycodon , and Cyananthus are basal in the family as supported by both gene order and sequence data. Takhtajan's system [ 2 ] is something of an exception among traditional classifications; however, he suggested only Cyananthus (in its own tribe Cyanantheae) as the most primitive member of the family, placing Platycodon and Codonopsis in other tribes (Table 1 ). In contrast, studies of pollen ultrastructure have indicated that Platycodon , Codonopsis , and Cyananthus are basal members of Campanulaceae [ 37 , 38 ]. These taxa have colpate to colporate apertures, whereas the remaining family members (as surveyed here) have porate grains [ 37 - 41 ]. The evolutionary scheme based on pollen morphology presented by Dunbar [ 38 ] suggests that Cyananthus (colpate) and Codonopsis (colpate) are more closely related to each other than either is to Platycodon (colporate), which is also supported by the gene order tree (Clade B, Fig. 5 ). Thulin [ 8 ] believed that pollen morphology should constitute a key part of any modern reassessment of relationships in the Campanulaceae. He suggested that all taxa with elongated apertures should be removed from Campanuleae and Wahlenbergieae, and those with porate grains removed from Schönland's Platycodinae. Following the removal of colpate and colporate taxa, Campanuleae sensu Schönland are comprised of Northern Hemisphere genera, whereas Wahlenbergieae contain Southern Hemisphere taxa, with the exceptions of Edraianthus and Jasione (although Jasione occurs in North Africa as well as Europe). The gene order data indicate that the affinities of Jasione and Edraianthus lie with Northern Hemisphere species rather than with Wahlenbergieae. The gene order data also are compatible with other available nucleotide data in addition to those reported here [[ 10 , 11 , 42 ], L. Raubeson, A. Oestriech and R. Jansen, unpublished data], a morphology-based cladistic study [ 10 ] and are also largely congruent with a serological study of the Campanulaceae [ 43 ]. Although the gene order and serological studies differed somewhat in the taxa sampled, both included a group containing Trachelium and Campanula . They also agreed in the grouping of Asyneuma and Petromarula . The only discrepancy was in the placement of Legousia ; the serological study placed this genus basal to all others surveyed [ 43 ]. The groups delimited by cpDNA rearrangements also exhibit geographical integrity. Wahlenbergia is primarily a Southern Hemisphere Old World genus [ 44 ]; W. gloriosa , mapped for this study, is Australian [ 44 ]. Roella , Merciera , and Prismatocarpus are all endemic to South Africa [ 45 - 47 ]. The nine genera in the Trachelium and Legousia clades are primarily European to Eurasian, although Triodanis is endemic to North America and Campanula has a few North American representatives [ 5 , 48 - 50 ]. Musschia is endemic to the island of Madeira [ 51 ]. There has been considerable debate regarding the relationships among the four centers of taxonomic diversity of the Campanulaceae: Asia, Europe (especially the Mediterranean), South Africa, and western North America. Bentham [ 52 ] hypothesized a northern origin for Campanulaceae but he did not specify a particular region. Takhtajan [ 2 ] suggested a basal position of the Asian genus Cyananthus . Studies of pollen ultrastructure indicated that the Asian genera Codonopsis , Cyananthus , and Platycodon are basal members of the Campanulaceae [ 37 , 38 ]. Recent studies of the Campanulales [ 42 , 53 , 54 ] indicate that the order consists of several families, including the Campanulaceae, Cyphiaceae, Cyphocarpaceae, Lobeliaceae, Nemacladaceae, and Stylidaceae. Several of these families are restricted to the Southern Hemisphere (all but Nemacladaceae from North America and Campanulaceae which is cosmopolitan), implying that the Southern Hemisphere may be the ancestral area for the Campanulaceae [ 54 ]. Phylogenies based on rbcL sequence data position the Campanulaceae sister to the North American family Nemacladaceae [ 42 , 54 ]. Our cpDNA phylogeny based on genome rearrangements (Fig. 5 ) provides strong support for the basal position of the three examined Asian platycodonoid genera, suggesting that the early radiation of the family may have occurred in Asia rather than Africa. The genera from the Southern Hemisphere ( Merciera , Prismatocarpus , Roella , and Wahlenbergia ) are in a much more derived position in the cpDNA tree. In addition, the gene order data suggest affinities of several controversial genera (Fig. 5 ). Schönland [ 48 ] united Musschia and Platycodon as Platycodinae, clearly incompatible with both our results and pollen evidence. Musschia is placed in the derived clade (A), although its exact placement varies among all the analyses, including rbcL and ITS. De Candolle [ 3 ] was unsure of Merciera 's taxonomic position because its four basal ovules and single-seeded (by abortion) unilocular capsule [ 43 ] are unique in the Campanulaceae [ 55 ]. This genus was later recognized as a separate tribe, Merciereae [ 56 ], but is allied with other southern African genera in the cpDNA analysis (Fig. 5 ). Takhtajan [ 2 ] placed Merciera with Wahlenbergia , Roella and Prismatocarpus in his Wahlenbergieae but also included other genera forming a polyphyletic group according to our results. Adenophora and Symphyandra have been segregated from Campanula based on the presence of a conspicuous tubular nectariferous disc and connate anthers, respectively. Adenophora and Campanula are sister taxa in the gene order analyses (except those based on Matrix 2) and Adenophora' s chloroplast genome is derived relative to Campanula' s (Fig. 5 ). Further sampling within Adenophora and the large genus Campanula will be necessary to determine if this is a general result. Symphyandra is more closely related to Edraianthus than Campanula but all are within the A3 Clade (Fig. 5 ). Edraianthus has traditionally been considered close to Wahlenbergia [ 3 ] but this is not supported by any of the results reported here or by morphological studies of Hilliard and Burtt [ 57 ]. Much controversy surrounds the taxonomy of the genera Triodanis and Legousia . In some treatments, both genera were included under the illegitimate name Specularia (e.g. [ 3 , 48 ]). McVaugh [ 58 , 59 ] and Fernald [ 60 ] disagreed regarding the circumscription of the genera; Fernald felt that Triodanis as a genus is very weak and should be merged with Legousia . McVaugh [ 58 ] argued that the two genera should either remain separate or both be subsumed into Campanula . In his system, both species studied here ( T. perfoliata and L. falcata ) belong to Triodanis . As expected, Triodanis and Legousia belong to the same cpDNA clade (A2), united by an unusual mutation that transferred a large segment of the large single copy (LSC) region to the SSC region [ 9 ]. However, Legousia has a putative transposition not found in Triodanis , whereas Triodanis has a unique large insertion [ 9 ]. Conclusions Despite the difficulties in interpreting such a complex set of rearrangements, the systematic utility of chloroplast DNA in the Campanulaceae is evident. Our results support the division of the family into two groups previously unrecognized in any taxonomic treatment. In addition, numerous groupings within the larger, derived clade are strongly supported. The data indicate that traditional classifications based on fruit and ovary characters are unnatural, and suggest affinities of several difficult genera. Additional sampling within large genera, such as Campanula and Wahlenbergia , will be necessary to fully elucidate relationships among chloroplast genomes. It is likely that intrafamilial relationships can be further resolved by including other genera in rearrangement analyses. Although homoplasy is not absent in our data, it is low and considerably lower than some sequence data such as ITS. Although any reasonable scoring method for the gene order data generates results that are largely compatible among the different analyses, the more that the gene order data can be interpreted as actual mutational events (and the presence or absence of those events used as characters) the stronger will be the results. Even in cases such as this, with high levels of gene order complexity, cpDNA gene order mutations make excellent phylogenetic markers. Methods Total DNA was isolated from one species in each of 18 genera in the Campanulaceae (Table 2 ) according to the CTAB method of Doyle and Doyle [ 61 ]. DNAs were digested with the restriction endonucleases Bam HI, Bgl II, Eco RI, Eco RV, Hind III, and Sst I, and double digests were carried out using Hind III and the remaining five enzymes. Hybridization probes consisted of 106 small tobacco cpDNA probes (average size 1.2 kb) provided by J. Palmer [ 62 ]. Twenty-one cloned Hind III cpDNA fragments from Trachelium caeruleum of the Campanulaceae were also used as hybridization probes [ 18 ]. Complete single and double digest restriction site maps were constructed for 16 of the 18 taxa, and nearly complete maps were constructed for the remaining two taxa, Jasione and Roella [ 9 ]. It was not possible to map the small single copy (SSC) region of Roella because hybridization signals became increasingly weak in later rounds of hybridization. In Jasione , rearrangements involving the IR/SSC junction and SSC region prohibited the complete resolution of the map. The restriction site maps were then interpreted as linear "number" maps representing the hybridization patterns of 106 consecutively numbered tobacco cpDNA probes for the 18 taxa (Fig. 1 ). Rearrangements were recognized as any change in the order of gene segments relative to the order observed in tobacco. The recognition of such disruptions is straightforward; the interpretation of the disruptions as actual mutational events can be quite complicated. As a hypothetical example, the ancestral order in a region may be 1-2-3-4-5-6; while the order 1-2-5-3-4-6 may be observed in a rearranged genome. In the rearranged genome 2-5, 5-3, and 4-6 are adjacencies that are derived relative to the ancestral order. But what set of events is responsible for the change? A simple transposition of 5 to the position between 2 and 3 can account for the difference in a single event. Alternatively, two inversions with one shared endpoint may be responsible or two inversions with unique endpoints followed by a transposition can explain the differences. Additional explanations would also be compatible with these data. On what basis do we choose among multiple scenarios? As an actual example, the chloroplast genome of Platycodon could have evolved from a tobacco-like ancestor by two different models each involving seven inversions (Fig. 7 ); not one inversion is common to the two scenarios. Thus in our initial approach to data analysis (generating Matrix 1) we did not define events, but utilized endpoints only. In the hypothetical example 2-5, 5-3, and 4-6 are "endpoints" -derived adjacencies absent in the ancestral gene order. Taxa with genomes that exhibit the derived adjacencies are coded as 1 for those characters and those with the ancestral condition as 0. Figure 7 Two alternative models of seven inversions each to explain the evolution of Platycodon cpDNA structure from a tobacco-like ancestor. Numbers in parentheses show order of hybridized tobacco cpDNA probes. Inversion endpoints are shown by the arrows. Locations of regions represented by probes 6-7, 8, and 9 are believed to be the result of transposition [14]; these events are required in addition to the inversions to completely explain the new gene order. We constructed two additional matrices that did include events since some endpoints (or combination of endpoints) seemed readily interpretable. For example, if a region of the genome was simply reversed in order (i.e., 1-4-3-2-5 relative to 1-2-3-4-5) we assumed that an inversion had taken place to result in the different arrangements of gene segments. Likewise if genomes differed in content of the IR, we assumed that single duplication or loss events were responsible. Making such inferences, we constructed Matrix 2 that is composed of 31 events and 53 endpoints. We then went further and constructed hypotheses of rearrangement events to account for the differences among the genomes of the three major clades delimited among the fifteen derived taxa [ 9 ]. If these scenarios indicated that an inversion likely was shared between two or more genera, the taxa were coded as having an endpoint even if the endpoint has been lost due to disruption by subsequent events. We were conservative in our application of this approach and only six endpoint scorings were modified in Matrix 3 compared to Matrix 2. To summarize, we produced three data matrices that represented increasing levels of interpretation of endpoints as actual events. Cladistic analyses were performed on each of the three data sets using equal weighting of all included characters. The second and third matrices were also analyzed giving weights of two to all non-endpoint characters. This weighting represents an attempt to compensate for the unintentional weight given to inversions in which both endpoints are present. This, of course, results in down-weighting inversions in which only one endpoint remains and fails to include inversions whose endpoints are both absent. Finally, to allow for a direct comparison of performance between gene order data and sequence data over the same taxa, we conducted maximum parsimony analyses of ITS and rbcL data. The ITS sequences were generated and aligned by Eddie et al [ 11 ]. We determined taxa equal or equivalent to our taxa and performed analyses on just those taxa from the Eddie matrix (Table 5 ); only fifteen of our eighteen taxa were represented. We generated rbcL data to add to the two taxa already available [ 42 ] so that we had rbcL sequence data from all of the eighteen mapped taxa. Exactly the same DNAs were used. In generating the rbcL sequences, we PCR-amplified about 1370 bp of the gene in 50 μl reactions containing: 1 μl unquantified total genomic DNA, 0.2 mM each dNTP, 2.5 mM MgCl2, 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.4 μM each primer, and 1 unit Taq polymerase. Cycling conditions were as follows: 1 95°C denaturation step for 3 minutes 30 seconds, 30 cycles of 1 minute at 95°C, 1 minute at 55°C, and 1 minute 30 seconds at 72°C, and finally a 7 minute 72°C step. The PCR primers plus two internal primers were used for sequencing; the forward amplification primer and two internal primers were designed by G. Zurawski (his Z-1, Z-427 and Z-895). The Zurawski primer commonly used as the reverse amplification primer did not work in many Campanulaceae; we designed an alternative: 5'-GTATCCATTGCGCAAACTC-3'. For sequencing, two successful PCR reactions were combined and then cleaned (and concentrated) using the Qiagen QIAquick PCR Purification Kit (catalog number 28104). Depending on the concentration of the recovered product, 0.5–2 μl of this template was cycle sequenced and resolved on an ABI Prism 377 Automatic DNA Sequencer. Electropherograms were inspected, and then sequences were edited and assembled using Sequencher, vers 3.1 (Gene Codes Corp.) The sequences have been deposited in GenBank (accession numbers in Table 5 ). Alignment was performed by Sequencher and adjusted manually. Alignment of the rbcL sequences was very straightforward. For all parsimony analyses, searches were conducted using the branch and bound algorithm in PAUP* 4.0b10 (PPC) [ 63 ]. Tobacco was used as the outgroup for the gene order data since it has the ancestral chloroplast genome gene order for the angiosperms [ 15 , 26 ] and Lobelia was used as the outgroup for the sequence data. See the additional file - data file 1 – for the Nexus file used in the PAUP analyses. This file includes the three gene order matrices and the rbcL alignment. The ITS alignment of Eddie et al [ 11 ] is available online [ 64 ]. The strength of the support, in each data set, for monophyletic groups was evaluated by calculating bootstrap values [ 65 ] using 10,000 heuristic (TBR, multrees option) replicates. In addition, for each matrix, analyses were performed to generate all trees from the shortest to one percent longer. We used a percentage rather than an equal number of steps in an attempt to make an equivalent comparison among the different sized data sets. A consensus of these trees was determined and the number of nodes retained "to 1% longer" was calculated. Authors' contribution MEC performed the DNA isolations and Southern hybridizations, mapped the genomes, performed character codings and analyses for Matrix 2 and Matrix 3, and wrote up her work as a chapter of her Ph.D. thesis. LAR confirmed genome maps and character codings, updated MEC's analyses, added MATRIX 1 and its analysis, generated the rbcL sequence data, analysed ITS and rbcL data, and modified the thesis chapter for publication. RKJ assisted in all aspects of the work. Supplementary Material Additional File 1 One additional file is provided – data file 1. This file is in NEXUS format and includes data for Matrix 1, Matrix 2, and Matrix 3 encodings and the rbcL data, concatenated for each taxon. PAUP statements in the file identify each individual data matrix and the characters in Matrix 2 and 3 that are weighted in some analyses. Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516026.xml
516783
Real time electrocardiogram QRS detection using combined adaptive threshold
Background QRS and ventricular beat detection is a basic procedure for electrocardiogram (ECG) processing and analysis. Large variety of methods have been proposed and used, featuring high percentages of correct detection. Nevertheless, the problem remains open especially with respect to higher detection accuracy in noisy ECGs Methods A real-time detection method is proposed, based on comparison between absolute values of summed differentiated electrocardiograms of one of more ECG leads and adaptive threshold. The threshold combines three parameters: an adaptive slew-rate value, a second value which rises when high-frequency noise occurs, and a third one intended to avoid missing of low amplitude beats. Two algorithms were developed: Algorithm 1 detects at the current beat and Algorithm 2 has an RR interval analysis component in addition. The algorithms are self-adjusting to the thresholds and weighting constants, regardless of resolution and sampling frequency used. They operate with any number L of ECG leads, self-synchronize to QRS or beat slopes and adapt to beat-to-beat intervals. Results The algorithms were tested by an independent expert, thus excluding possible author's influence, using all 48 full-length ECG records of the MIT-BIH arrhythmia database. The results were: sensitivity Se = 99.69 % and specificity Sp = 99.65 % for Algorithm 1 and Se = 99.74 % and Sp = 99.65 % for Algorithm 2. Conclusion The statistical indices are higher than, or comparable to those, cited in the scientific literature.
Background The QRS complexes and ventricular beats in an electrocardiogram represent the depolarization phenomenon of the ventricles and yield useful information about their behavior. Beat detection is a procedure preceding any kind of ECG processing and analysis. For morphological analysis this is the reference for detection of other ECG waves and parameter measurements. Rhythm analysis requires classification of QRS and other ventricular beat complexes as normal and abnormal. Real-time ventricular beat detection is essential for monitoring of patients in critical heart condition. Correct beats recognition is impeded by power-line interference, electromyogram noise and baseline wander often present in the ECG signal. In long-term monitoring electrode impedance can increase considerably, resulting in very low signal-to-noise ratio, which can make detection practically impossible in a single lead. Therefore, usually two or three leads are used for monitoring [ 1 ]. Friesen et al. [ 2 ] have presented a comparison of nine QRS detection algorithms, based on: i) amplitude and first derivative, ii) first derivative only, iii) first and second derivative, and iv) digital filtering. Daskalov et al . [ 3 ] applied these algorithms to selected signals containing records with pronounced baseline drift. The results were unsatisfactory, which was probably due to the use of fixed detection thresholds, whereas adaptive ones would be more appropriate. Poli et al . [ 4 ] used a generic algorithm for QRS detection. The complexes were emphasized with respect to the rest of the signal by polynomial filters and compared to an adaptive threshold. The authors reported 99.60 % sensitivity (Se) and 99.51 % specificity (Sp) with the MIT-BIH Arrhythmia Database. The method is inapplicable in real-time. Afonso et al. [ 5 ] proposed hardware filter banks for ECG signal decomposition, where several parameters were independently computed and combined in a decision rule. The authors reported Se = 99.59 % and Sp = 99.56 % for their real-time, single-channel beat detection algorithm tested with the MIT-BIH Arrhythmia Database. Dotsinsky and Stoyanov [ 6 ] developed a heuristic, pseudo-real-time algorithm for ventricular beat detection for single-channel ECG, based on steep edges and sharp peaks evaluation criteria. They reported Se = 99.04% and Sp = 99.62%, obtained with two channel recordings from AHA and MIT-BIH Arrhythmia Database Moraes et al . [ 1 ] combined logically two different algorithms working in parallel – the first has been taken from the work of Englese and Zeelenberg [ 7 ] and the other was based on Pan and Tompkins [ 8 ], and Ligtenberg and Kunt [ 9 ]. Moraes et al . [ 1 ] reported Se = 99.22 % and Sp = 99.73 % after having excluded records of patients with pacemaker. After excluding a few more recordings 108, 200, 201 and 203, containing high amplitude noise (according to the authors), the statistical indices rises to Se = 99.56 % and Sp = 99.82 %. Li et al . [ 10 ] have used wavelet transforms for detection. They reported 0.15 % false detections out of 46 files from the MIT-BIH Arrhythmia Database, but with exclusion of files 214 and 215. In addition, we found some errors in their Table II. After correction, the reported accuracy slightly decreased. The large variety of QRS detection algorithms, and the continuous efforts for their enhancement, proves that universally acceptable solution has not been found yet. Difficulties arise mainly from the huge diversity of the QRS complex waveforms and the noise and artifacts accompanying the ECG signals. ECG databases All 48 ECG recordings of MIT-BIH Arrhythmia database were used, without exception. Each one has a duration of 30 min and includes two leads – the modified limb lead II and one of the modified leads V1, V2, V4 or V5 [ 11 ]. The sampling frequency is 360 Hz with resolution 5 μV/bit. Two cardiologists have annotated all beats. Approximately 70 % of the beats are annotated as Normal . Four of the records are from patients with pacemakers. The American Heart Association (AHA) database was also considered, during the evaluation of the method, mostly due to the fact that it contains patients with premature ventricular beats of contraction of R-over-T type, some of them very difficult to detect because of their closeness to the previous complex. Statistical indices for this database are not derived, because they can be compared with limited number of articles working with AHA. The database consists of 80 recordings: 2-leads, 250 Hz sampling rate and 5 μV/bit resolution. Method The differentiated and summed signals from L leads are compared to the absolute value of a threshold MFR = M + F + R – a combination of three independent adaptive thresholds, where: • M – Steep-slope threshold; • F – Integrating threshold for high-frequency signal components; • R – Beat expectation threshold. Two algorithms were developed: Algorithm 1 detects at the current beat. Algorithm 2 Pseudo-real-time detection with additional triggering of potentially missed heart beat in the last interval by RR interval analyses. The algorithms are self-adjusting to the thresholds and weighting constants, regardless of resolution and sampling frequency used. They operate with any number L of ECG leads, self-synchronize to QRS or beat slopes and adapt to beat-to-beat intervals. Preprocessing • Moving averaging filter for power-line interference suppression: averages samples in one period of the power-line interference frequency with a first zero at this frequency. • Moving averaging of samples in 28 ms interval for electromyogram noise suppression – a filter with first zero at about 35 Hz. • Moving averaging of a complex lead (the sintesis is explained in the next section) in 40 ms intervals – a filter with first zero at about 25 Hz. It is suppressing the noise magnified by the differentiation procedure used in the process of the complex lead sintesis. Complex lead The algorithm operates with a complex lead Y of several primary leads L . In cases of 12-standard leads, synthesis of the three quasi-orthogonal Frank leads is recommended first [ 3 , 12 ], thus determining the complex lead as a spatial vector. The complex lead is obtained as: where Xj(i) is the amplitude value of the sample i in lead j , and Y(i) is the current complex lead. The above formula (except the normalizing coefficient 1/L and the absolute value) was initially adopted from the work of Bakardjian [ 13 ]. Operating with unsigned (absolute) values proved convenient when dealing with QRSs and extrasystoles having different, for example positive (in one lead) and negative (in the other lead) deflections. Adaptive steep-slope threshold – M • Initially M = 0.6* max(Y) is set for the first 5 s of the signal, where at least 2 QRS complexes should occur. A buffer with 5 steep-slope threshold values is preset: MM = [ M 1 M 2 M 3 M 4 M 5 ], where M 1 ÷ M 5 are equal to M . • QRS or beat complex is detected if Yi ≥ MFR , • No detection is allowed 200 ms after the current one. In the interval QRS ÷ QRS +200 ms a new value of M 5 is calculated: newM 5 = 0.6* max(Yi) The estimated newM 5 value can become quite high, if steep slope premature ventricular contraction or artifact appeared, and for that reason it is limited to newM 5 = 1.1* M 5 if newM 5 > 1.5* M 5 . The MM buffer is refreshed excluding the oldest component, and including M 5 = newM 5 . M is calculated as an average value of MM . • M is decreased in an interval 200 to 1200 ms following the last QRS detection at a low slope, reaching 60 % of its refreshed value at 1200 ms. • After 1200 ms M remains unchanged. The thresholds definitions are presented in more detail with the help of several examples. Two ECG leads are shown in Fig. 1a . Detected QRSs are marked with 'red O' on Lead 1. The summary lead and the steep-slope threshold are represented in Fig. 1b . Figure 1 Adaptive steep-slope threshold Adaptive integrating threshold – F The integrating threshold F is intended to raise the combined threshold if electromyogram noise is accompanying the ECG, thus protecting the algorithm against 'erroneous beat detection'. Initially F is the mean value of the pseudo-spatial velocity Y for 350 ms. With every signal sample, F is updated adding the maximum of Y in the latest 50 ms of the 350 ms interval and subtracting maxY in the earliest 50 ms of the interval. F = F + (max(Y in latest 50 ms in the 350 ms interval ) - max(Y in earliest 50 ms in the 350 ms interval ))/150 The way F is updated means that not every sample in the interval is integrated, but just the envelope of the pseudo-spatial velocity Y . The weight coefficient 1/150 is empirically derived. Two ECG leads are shown in Fig. 2a . The pseudo-spatial velocity Y and the integrated threshold are presented in Fig. 2b . The correct detection is due to the rise of F (hence of MFR ) with about 0.2 mV. The beat complex is included in the integration process (note the high rise of F after any of the complexes), thus making almost impossible a close detection to the previous complex. Figure 2 Adaptive integrating threshold Adaptive beat expectation threshold – R The beat expectation threshold R is intended to deal with heartbeats of normal amplitude followed by a beat with very small amplitude (and respectively with very small slew rate). This can be observed for example in cases of electrode artifacts. Conversely to the integrating threshold protecting against erroneous QRS detection, R is protecting against 'QRS misdetection'. A buffer with the 5 last RR intervals is updated at any new QRS detection. Rm is the mean value of the buffer. • R = 0 V in the interval from the last detected QRS to 2/3 of the expected Rm . • In the interval QRS + Rm * 2/3 to QRS + Rm, R decreases 1.4 times slower then the decrease of the previously discussed steep slope threshold ( M in the 200–1200 ms interval). • After QRS + Rm the decrease of R is stopped. The time-course of the beat expectation threshold R is shown in Fig. 3 . The decrease of R (respectively MFR ) with about 0.2 mV at the fourth QRS allows its detection, despite the lack of complex in Lead 2, which leads to a two-fold decrease of the summary lead amplitude Y (Fig. 3b ). Figure 3 Adaptive beat expectation threshold Combined adaptive threshold – MFR The combined adaptive threshold is a sum of the adaptive steep-slope threshold, adaptive integrating threshold and adaptive beat expectation thresholds. (Fig. 4 ) Figure 4 Combined adaptive threshold MFR = M + F + R Algorithm 2: pseudo-real-time detection with additional triggering of eventually missed heart beat in the last detected RR interval All previous considerations relate to Algorithm 1, which detects a beat at its occurrence. Additional checking for an eventually missed heartbeat is performed by Algorithm 2. Its function is explained by the signal in Fig. 5 . The fourth complex at the 15.2 s in Fig. 5b should be missed due to the fact that, MFR is greater then the summary lead Y . Figure 5 Pseudo-real-time detection with additional triggering of eventually missed heart beat in the last RR interval. Let's mark the previous RR interval with t1 and the last – with t2 (Fig. 5a ). If t1 is not shortened, which is tested by logic OR of the 2 conditions t1>Rm OR Rm-t1<0.12*Rm AND in the same time t2 is quite long to fulfill the condition abs(t2-2*Rm)<0.5*Rm, the interval is subjected to check for a missed complex. A test is performed on each of the primary leads where a sharp peak is searched (defined as a product > 4 μV of two signal differences having one central and two lateral points 8 ms apart). If the test is passed, a second one is carried out for the amplitude of the summary lead at that point, which should be bigger then 1/3 of the mean value of the buffer MM , in order to define this point as a missed QRS complex. Results and discussion Normally the statistical indices Se and Sp are derived from the following parameters: correctly detected beats TP (true positive), falsely detected beats FP (false positive) and undetected beats FN (false negative). In addition, we used two parameters, adopted from Dotsinsky and Stoyanov [ 6 ], as described below. SP – shifted positive error was introduced in order to explain cases like the one shown in Fig. 6 . Here the algorithm made a false positive error before the 3rd QRS and missed the following QRS. Formally, this is a false positive error, immediately followed by a false negative. However, if the time interval between these two detections is reasonably short, for example ≤ 200 ms, we accepted this as one error only, labeled as Shifted False Positive Error (SP). Figure 6 Shifted positive error at the P wave Another example of SP error as a result of artifacts just before the normal complexes is shown in Fig. 7 . Figure 7 Shifted positive errors, false positive + false negative twins SN – shifted negative error was assumed by the same principle as SP, but in the opposite sense. It also included twin FN+FP error occurring within 200 ms. The first incoming FP or FN error of the shifted is defining it as SP or SN. The logic of using shifted errors (instead of FP and FN or FN and FP in cases when they appear within 200 ms of each other) is that thus the total number of beats in a record retains its value. Otherwise it would change depending on the type and number of errors and thus impede correct computation of Se and Sp. The sensitivity Se is calculated by summing FN SN, while the specificity Sp – by summing FP+SP. The method was developed in Matlab. All 48 recordings from the MIT-BIH Arrhythmia database, without any exception, were used for testing the two algorithms. The processed files containing detection marks were automatically compared with the original MIT-BIH annotated beats by specially designed software. It shows all cases where the annotation and detection marks differ by more than 60 ms. These cases were examined by an independent expert, thus excluding possible author's influence. The results for the two algorithms are presented in Table 1 . Table 1 Statistical results for the two algorithms File Annotated beats Algorithm 1 Algorithm 2 TP FN FP SN SP TP FN FP SN SP 100 2273 2273 0 0 0 0 2273 0 0 0 0 101 1863 1862 1 4 0 0 1862 1 4 0 0 102 2187 2187 0 0 0 0 2187 0 0 0 0 103 2084 2062 2 54 11 9 2065 0 58 12 7 104 2212 2211 1 0 0 0 2211 1 0 0 0 105 2567 2543 2 35 8 14 2544 2 36 8 13 106 2027 2017 1 1 0 9 2018 0 1 0 9 107 2137 2135 2 0 0 0 2137 0 0 0 0 108 1763 1664 2 40 3 94 1674 1 42 3 85 109 2532 2521 11 1 0 0 2527 5 0 0 0 111 2124 2124 0 0 0 0 2124 0 0 0 0 112 2539 2539 0 0 0 0 2539 0 0 0 0 113 1797 1797 0 0 0 0 1797 0 0 0 0 114 1879 1879 0 0 0 0 1879 0 0 0 0 115 1953 1951 0 4 1 1 1952 0 4 0 1 116 2412 2389 22 2 0 1 2392 19 2 0 1 117 1535 1535 0 0 0 0 1535 0 0 0 0 118 2275 2275 0 0 0 0 2275 0 0 0 0 119 1987 1987 0 0 0 0 1987 0 0 0 0 121 1863 1863 0 0 0 0 1863 0 0 0 0 122 2476 2476 0 0 0 0 2476 0 0 0 0 123 1518 1516 2 0 0 0 1516 2 0 0 0 124 1619 1617 2 0 0 0 1619 0 0 0 0 200 2601 2549 9 39 18 25 2552 6 41 20 23 201 1963 1902 60 0 0 1 1902 60 0 0 1 202 2136 2130 6 0 0 0 2130 6 0 0 0 203 2978 2901 71 13 3 3 2911 62 27 3 2 205 2656 2652 4 0 0 0 2652 4 0 0 0 207 1862 1860 2 0 0 0 1862 0 1 0 0 208 2954 2937 14 7 2 1 2939 11 7 2 2 209 3004 3004 0 1 0 0 3004 0 1 0 0 210 2647 2591 56 1 0 0 2603 44 1 0 0 212 2748 2748 0 0 0 0 2748 0 0 0 0 213 3551 3548 3 0 0 0 3550 1 0 0 0 214 2260 2258 1 1 1 0 2256 4 1 0 0 215 3362 3362 0 0 0 0 3362 0 0 0 0 217 2208 2204 3 0 0 1 2205 2 0 0 1 219 2154 2153 1 0 0 0 2153 1 0 0 0 220 2048 2048 0 0 0 0 2048 0 0 0 0 221 2427 2426 1 0 0 0 2426 1 0 0 0 222 2483 2480 2 0 0 1 2482 0 0 0 1 223 2595 2585 10 0 0 0 2590 5 0 0 0 228 2053 2053 0 0 0 0 2053 0 1 0 0 230 2256 2256 0 0 0 0 2256 0 0 0 0 231 1886 1886 0 0 0 0 1886 0 0 0 0 232 1767 1766 0 12 0 1 1766 0 12 0 1 233 3076 3073 3 0 0 0 3074 2 0 0 0 234 2753 2753 0 0 0 0 2753 0 0 0 0 Sum 110050 109548 294 215 47 161 109615 240 239 48 147 Of all 110050 annotated beats ('unknown' or 'questionable' were excluded from the study), true detected are 109548 for Algorithm 1 and 109616 for Algorithm 2. The statistical indices are: Algorithm 1 : Se = 99.69 %, Sp = 99.66 %; Algorithm 2 : Se = 99.74 %, Sp = 99.65 %. The standard way of Se and Sp calculation, considering the joint SP and SN errors as two separate errors gives the following results: Algorithm 1 : Se = 99.54 %, Sp = 99.61 %; Algorithm 2 : Se = 99.60 %, Sp = 99.60 %. Algorithm 2 improved the sensitivity by 0.05 % (0.06 % for the standard evaluation) as a result of decreased number of undetected beats. This result can be observed for example in recordings 109,203, 210 and 223, where the additionally detected beats are respectively 6, 9, 12 and 5. The performance of both algorithms was especially tested with the file A5001 from the AHA containing R-over-T premature ventricular complexes, very close to the previous normal QRS complex (Fig. 5a ). An improvement of 74 undetected by Algorithm 1 R-on-T complexes was observed. The detection of such premature ventricular complexes occurring at the time of ventricular repolarization was considered important, having in mind possible risk of ventricular fibrillation triggering by R-on-T events. Conclusions The proposed algorithms for real-time and pseudo-real-time implementation are adaptive, independent of thresholds and constants values. They are self-synchronized to the QRS steep slope and the heart rhythm, regardless of the resolution and sampling frequency used. Due to the integration threshold, the algorithms are practically insensitive to electromyogram and similar high-frequency noise. The algorithms can operate with one, two or more leads, using a combined lead signal derived from the sum of absolute values of the differentiated lead signals. The statistical indices are higher than, or comparable to those, cited in the scientific literature.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516783.xml
518976
Valgus and varus deformity after wide-local excision, brachytherapy and external beam irradiation in two children with lower extremity synovial cell sarcoma: case report
Background Limb-salvage is a primary objective in the management of extremity soft-tissue sarcoma in adults and children. Wide-local excision combined with radiation therapy is effective in achieving local tumor control with acceptable morbidity and good functional outcomes for most patients. Case Presentation Two cases of deformity after wide-local excision, brachytherapy and external beam irradiation for lower-extremity synovial cell sarcoma are presented and discussed to highlight contributing factors, time course of radiation effects and orthopedic management. In an effort to spare normal tissues from the long-term effects of radiation therapy, more focal irradiation techniques have been applied to patients with musculoskeletal tumors including brachytherapy and conformal radiation therapy. As illustrated in this report, the use of these techniques results in the asymmetric irradiation of growth plates and contributes to the development of valgus or varus deformity and leg-length discrepancies. Conclusions Despite good functional outcomes, progressive deformity in both patients required epiphysiodesis more than 3 years after initial management. There is a dearth of information related to the effects of radiation therapy on the musculoskeletal system in children. Because limb-sparing approaches are to be highlighted in the next generation of cooperative group protocols for children with musculoskeletal tumors, documentation of the effects of surgery and radiation therapy will lead to improved decision making in the selection of the best treatment approach and in the follow-up of these patients.
Background Limb-salvage is an important treatment objective for adults and children with extremity soft tissue sarcoma and often requires the use of limited surgery and irradiation [ 1 ]. Limiting the extent of resection balances the need for radical excision with the need to preserve the functional and structural integrity of the limb and tissues adjacent to those involved with tumor. Radiation therapy has been proven to compensate for incomplete resection or limited resections with involved, close or indeterminate margins as long as the dose and volume are adequate [ 2 - 5 ]. Excellent rates of local control have been achieved for adults and children with extremity soft tissue sarcoma using limb sparing approaches [ 6 - 11 ]. Little is known about the long-term morbidity of the combined effects of limited surgery and irradiation on bone and soft tissue in the pediatric population. Surgery and radiation therapy both have the potential to cause significant morbidity including loss of function and deformity [ 6 , 11 - 15 ]. Tumor resection often requires removal of normal tissue compartments and structural elements, even in a limb-sparing approach. This places the patient at risk for complications including destabilization and abnormalities in growth and function. Additive are the effects of high-dose irradiation, which is often required in the treatment of these tumors, and which may compound the effects of resection. The use of chemotherapy, when indicated, may also add to the combined effects of treatment. The timing of surgery and radiation therapy, the operative approach and the selection of the specific radiation treatment modality often depends on a number of important clinical factors including the size and type of tumor, site of involvement, prior surgical manipulation, extent of resection and the potential for as good functional outcome. Individual cases of valgus and varus deformity after limited surgery and irradiation for extremity soft-tissue sarcoma are presented and discussed to identify factors that may be responsible for these treatment complications. Pre-existing orthopedic problems, multiple attempts at resection, post-operative infection, the use of chemotherapy and the addition of brachytherapy to external beam radiation therapy appear to be contributory. Because limb-sparing approaches will be an important component of the next generation of cooperative group studies for extremity soft tissue sarcoma in children, the incidence and severity of this and other treatment-related complications should be documented as well as efforts to limit the effects of these treatments and identify solutions for established problems. Case Presentations Case 1 At the time of diagnosis, this patient was an 8 year-old male with an approximate 12–24 month history of mild pain and swelling in the left popliteal region. There was no complaint of fever or decreased range of motion. He presented to a local orthopedist (April 1995) and was found to have a palpable abnormality on the posterior aspect of the knee consistent with a Baker's cyst. Aspiration was unsuccessful and the patient was treated with ibuprofen for a 10-day course. Nearly one year later (April 1996) the family sought a second opinion and an MR study was ordered that revealed a cystic structure [Figure 1 ]. The patient returned to the original orthopedist and was found to have a painful and enlarging mass in the left popliteal region. He underwent resection of a solid and cystic mass (November 1996) measuring 7.0 × 6.0 × 3.0 cm. The tumor was described as a high-grade synovial cell sarcoma. The extent of resection was incomplete with gross residual tumor remaining about the lateral aspect of the knee and external to the joint capsule which also appeared to be the site of origin. The patient was transferred one month later to St. Jude Children's Research Hospital for further evaluation and treatment. Figure 1 MRI at diagnosis (case one). At the time of his evaluation after referral, he had strong popliteal pulses. There was a 7.0 × 4.5 cm area of swelling and numbness in the left popliteal region. The deep-tendon reflexes were brisk and the motor exam and gait were normal. MR showed residual abnormality consistent with tumor lateral to the joint capsule. Metastatic work-up including nuclear bone scan and CT scan of the chest was negative. Tumor bed re-excision with placement of afterloading catheters was performed in December 1996. All visible residual abnormality was removed without significant disruption of underlying ligaments and tendons. The walls of the tumor bed were biopsied to map the extent of microscopic residual disease. Microscopic residual disease was anticipated given the site of involvement and the limited ability to operate beyond the extent of the abnormal appearing tissues. Eleven afterloading catheters were placed in a parallel array to cover the tumor bed [Figure 2 ]. Radio-opaque clips were placed at the site of the biopsies and to demarcate the extent of the tumor bed for brachytherapy planning. The final pathology confirmed the presence of residual tumor in the operative specimen and microscopically involved margins at the central and superomedial aspects of the tumor bed. Figure 2 Afterloading catheters and dosimetry (case one). Five days after surgery, the 11 catheters were loaded with a total of 135 seeds representing 408 millicuries of I 125 . The dwell time of the implant was 64 hours and the patient received a total implant dose of 2560 cGy delivered at 40 cGy/hr. Three weeks later, the patient began external beam irradiation at 180 cGy per day and received a total external beam dose of 4860 cGy using 6 MV photons with treatment delivered in a parallel-opposed beam arrangement using a CT based treatment plan [Figure 3 ]. Radiation therapy was completed in February 1997. Figure 3 Treatment ports and dosimetry (3D). A decision was made to initiate chemotherapy three weeks into the course of external beam irradiation based on the perceived high-risk nature of his case – longstanding history of symptoms, known residual tumor and size of tumor at presentation. Chemotherapy included vincristine, ifosfamide, and adriamycin was eventually administered for a total of four cycles. There was central dehiscence of the wound prior to the completion of radiation therapy. The wound was colonized with Enterococci sensitive to ampicillin and managed with antibiotics, whirlpool treatment and daily dressing changes. There was a one-week treatment break during the external beam portion of the treatment. At the completion of chemotherapy (June 1997) the patient underwent excision of scar tissue with rotation flap of gastrocnemius and skin and Z-plasty of the semi-membranous and semi-tendinous tendon for a non-healing ulcer in the operative region. There was also contracture of the knee joint without signs of abscess or cellulitis. One month later, a second procedure was required to debride and irrigate the left popliteal fossa wound at the site of previously irradiated tissue and contracture release with muscle and fasciocutaneous flap closure. One year after the completion of all therapy (June 1998) the patient reported full range of motion and softening of previously fibrotic tissue. He was actively playing baseball and had no imposed limitations. Nearly two years after completion of treatment (March 1999), the patient was noted to have Trendelenburg gait after prolonged walking ascribed to poor endurance of weakened hip abductors bilaterally. Hip hiking was also noted on the contralateral side due to leg length and ASIS height discrepancy. A difference of 2.5 cm was noted when measured from the umbilicus to the medial malleolus. He had grown 7.0 cm since the time of diagnosis. He was instructed to stretch the heel cord and strengthening his weak hip abductors. Arrangements were made to provide a shoe lift to accommodate the leg length difference. The discrepancy improved to less than 1 cm over a three month period of time (March-June 1999). He was carefully monitored for growth discrepancy and 6 months later, nearly 3 years after the initiation of treatment, the discrepancy returned to its original value of greater than 2 cm. He also had difficulty with ambulation with slight left-sided limp. Orthopedic surgery was re-consulted. The discrepancy was followed and treated with additional shoe lift. The patient continued to engage in normal activities including sports and reported full range of motion and normal strength. More than four and a half years after initiating treatment (August 2001) [Figure 4 ], x-ray scanogram revealed estimated lengths of 49.0 cm and 46.5 cm for his right and left femora respectively, and 40.7 cm and 37.0 cm for his right and left tibiae respectively [Figure 5 ]. There was a clinically noticeable varus deformity of the left knee. All growth plates were still open radiographically. He was taken to surgery five years after the initiation of definitive therapy (December 2001) [Figure 6 ] for a panepiphysiodesis of the right leg. Using an image intensifier to locate the distal femoral growth plate, medial and lateral incisions were made and a wire-guided cannulated reamer was used to obliterate the plate. The same procedure was done to the proximal tibial growth plate. There were no complications from surgery and the wounds healed appropriately. He was started on physical therapy including quad sets to maintain adequate range of motion. There has been no evidence of tumor recurrence or metastatic disease nearly seven years after treatment. There are no limitations regarding activities and there has been no progressive angular growth deformities. Figure 4 Serial plain films (case one). Figure 5 Scanogram 4/2001 (case one). Figure 6 Pre-Op MR and photograph (case one). Case 2 At the time of diagnosis this patient was a 9 year-old female with a one year history of pain and swelling about her left knee. She had experienced a fall and related all symptoms to the fall. She was seen in her local emergency room by her family physician; there was no diagnosis or treatment. Approximately one month prior to her representation, she was struck in the left knee by a basketball and developed worsening pain. She was seen by an orthopedic surgeon (December 1999) and was noted to have a valgus posture of both lower extremities, exaggerated on the left by external rotation and she walked with a mild limp. The left knee had no effusion but was hypersensitive to light touch over the lateral aspect where there was soft tissue swelling just below the knee. There was no obvious mass in the area, although firm palpation was difficult because of patient discomfort. Plain films were normal and an MR was ordered that revealed an apparent meniscal cyst in the lateral aspect of the left knee [Figure 7 ]. Biopsy of the cystic structure was performed (November 1999) that revealed a high-grade synovial cell sarcoma. Metastatic work-up consisting of nuclear bone scan and CT of the chest were negative. Amputation was offered by the local care team that included a radiation oncologist because of their concern about possible contamination of the joint space and uncertain functional outcome. The patient was referred to St. Jude Children's Hospital for further evaluation and treatment. Figure 7 MR Imaging (case two). At the time of her evaluation after referral (January 2000), there was a well healed scar with no excessive swelling. There was mild tenderness on the lateral aspect of her left knee. Additional imaging studies showed abnormality at the site of prior surgery equivocal for residual tumor. There was no evidence of abnormality in the joint space. The tumor bed was explored. There was no physical evidence of compromise at the level of the joint space. She underwent wide local excision with placement of afterloading catheters. Six catheters were placed in a parallel array with 1 cm spacing. Radio-opaque clips were placed to delineate the tumor bed and assist in brachytherapy planning [Figure 8 ]. The margins of the resection were involved with tumor, as demonstrated by field biopsies and assessment of the margins of resection. Satellite tumor nodules were present in the resection specimen. Four days after surgery the six afterloading catheters were loaded with 82 seeds representing 302 mCi of I 125 (Figure). The dwell time of the implant was 62 hours and the patient received a total implant dose of 2480 cGy delivered at 40 cGy/hr. Two weeks later the patient began external beam irradiation at 180 cGy/day and received a course of treatment and total external beam dose of 5040 cGy using 6 MV photons with treatment delivered with two beams using a CT based treatment plan [Figure 9 ]. Radiation therapy was completed in March 2000. Figure 8 Brachytherapy films and dosimetry. Figure 9 External Beam films and dosimetry. The patient suffered moist desquamation corresponding to the radiation therapy portal that was predicted based on the treatment and the use of a tissue equivalent bolus material which was placed on the wound on alternating days during her course of external beam irradiation. She was able to return home on the last day of treatment. On routine follow-up, only 4 months after treatment, left leg appearing to be slightly longer than her right leg by less than 1 cm. No corresponding gait problems were reported. Nearly one year after treatment (January 2000) physical examination showed good range of motion at the left knee; however, there was significant valgus angulation. An MR study was reviewed by Orthopedic Surgery and was noted to show growth arrest laterally and predominantly involving the distal femoral physis [Figure 10 ]. Based on these findings, the family was informed that an epiphysiodesis of the distal femoral physis would likely be required to prevent additional deformity. Due to the angulatory deformity, an osteotomy of the distal femur would be required. Because of high-dose irradiation and concerns about bone healing, osteotomy and epiphysiodesis were deferred until the three year evaluation was performed. At that time, the patient had a significant valgus deformity. The morbidity of the deformity was such that ambulation was difficult. The patient underwent a closing wedge correcting osteotomy, which was fixed with a contour plate. The patient subsequently fractured the plate secondary to early and unprotected weightbearing (against medical advice). She was placed in a cast and ultimately healed her osteotomy. She continues to have a significant limb length discrepancy and will require future lengthening procedures. She remains without evidence of disease nearly 4 years after treatment. Figure 10 MR Imaging and photography. Conclusion Due to its high propensity for local recurrence and metastasis, aggressive treatment of synovial sarcoma is imperative. While many different methods have been used in the treatment of these high-grade tumors, including mono-bloc soft part resection and amputation, the current standard includes local excision and radiation therapy when feasible [ 16 ]. Wide local excision with adjuvant radiation therapy is known to achieve a satisfactory rate of local control and good functional outcome [ 1 - 15 ]. Because these tumors commonly arise near tendon sheaths or joint capsules, treatment plans intending to achieve limb conservation may injury the epiphyseal growth plate affecting normal growth and development. Efforts should continue to improve our ability to delineate the tumor, achieve resection with microscopically negative margins and irradiate the region at risk in a manner that minimizes the effect on normal tissues [ 17 ]. Both patients in our study underwent two surgical procedures and were treated with brachytherapy and external beam irradiation. Brachytherapy was used to confine the highest doses to the region at risk and minimize the dose received by normal tissues. The use of brachytherapy shortens the overall treatment time and increases the rate of local control in the setting of involved margins of resection. CT-based treatment planning was used to define the volume of irradiation and to spare normal tissue structures. By reducing the amount of radiation dose delivered to normal tissues, the probability of growth deformity, radio-chemotherapy interactions, and even the hypothetical risk of second tumor formation may be lowered. No effort was made to symmetrically irradiate the physes, which would hypothetically lead to symmetrically diminished growth without the added effect of angular deformity. Because of concerns about the effects of total joint irradiation and its possible effects on functional outcome, the inhomogeneous and asymmetric approach was taken. Despite the valgus and varus deformity experienced by these children and the need for intervention, both children and parents were completely satisfied with their functional outcome and indicated that they would chose the same course of treatment if presented again with the same options. The rationale for radiation therapy The importance of achieving local control with aggressive surgery and high-dose irradiation cannot be overemphasized. Local control is crucial to long-term survival and avoiding the morbidity of local tumor progression. Local control, even in the setting of metastatic disease, is an important endpoint. Radiation therapy is highly successful in achieving local control in soft tissue sarcoma and is standard in the care of children with high-grade tumors such as those reviewed in this report. At our institution high-grade tumors with resection margins of 1 cm has an observed local control rate of 72% (5 of 7 patients) in the absence of radiation therapy and 100% (7 of 7 patients) when radiation therapy was given postoperatively. Among 20 unirradiated high-grade tumors that were completely resected with margins > 1 cm only 15 (75%) were locally controlled for an extended period [ 18 ]. Our policy is to use external beam irradiation or brachytherapy alone for high-grade tumors that are completely excised, regardless of age or other considerations including anatomic location. We also recommend brachytherapy combined with external-beam irradiation for high-grade tumors with involved, close or indeterminate margins, regardless of size or anatomic location [ 19 ]. Low-grade tumors are treated with external-beam radiation therapy or brachytherapy only when the risk of recurrence and re-resection morbidity is high, or at the time of recurrence. These policies apply even to patients with metastatic disease who are likely to survive for an extended period of time after aggressive multimodality therapy including metastasectomy [ 19 , 20 ]. Exceptions may be considered for small, superficial tumors in very young patients when resection can be performed with adequate margins, generally < 5 mm although prospective studies demonstrating the appropriateness of this approach are limited [ 21 ]. Bone growth and development and the effects of various conditions and treatments Despite efforts to achieve local control and minimize the effects of treatment on normal tissues, damage to bone and soft tissues may be unavoidable. Synovial cell sarcoma commonly arises near tendon sheaths and joint capsules of adolescents and young adults and may be in close proximity to an epiphyseal growth plate during a time of rapid growth. The situation is made worse if the tumor is located around the distal femur or proximal tibia. Among the four epiphyseal plates in the lower extremity that contribute to the growth of the limb, those around the knee make the most significant contribution, with the distal femur and proximal tibia accounting for 50–90% and 57% of limb growth, respectively depending on age [ 22 ]. In our study, Case 1 was eight years old at the time of diagnosis and had a tumor lateral to the joint capsule of the left knee. Case 2 was nine years old at the time of diagnosis and had a tumor lateral to the left distal femur and in close proximity to the joint space. It was less than one year after diagnosis and definitive management that both patients developed a clinically significant angular deformity and leg length discrepancy. There are two non-congenital mechanisms that are known to interfere with growth of the physis: direct trauma and environmental change around the plate. Trauma includes acute injury to the growth plate in manner that affects all or partial growth and results in premature closure or the formation of a physeal bar. Even if the region has retained its ability to grow it is hampered by solid bone formation across the plate [ 23 ]. Environmental change is less common and poorly understood. Roberts [ 24 ] discussed the disturbance of epiphyseal growth in the knee of infants with osteomyelitis and suggested that damage to the epiphysis might be due to an abscess or ischemia following occlusion of the blood supply. Infection is known to produce more severe leg length discrepancy problems than trauma, because the patients are typically younger at onset [ 23 ]. Tumors can contribute to leg length discrepancy either by direct invasion or by originating from the cartilage cells of the physis, thereby stealing growth potential from the plate [ 25 ]. Vascular malformations adjacent to the physis have been known to both inhibit and stimulate growth [ 26 , 27 ]. Paralysis is also known to cause of growth inhibition, although the mechanism is poorly understood. Proposed contributors include reduced muscle activity, which indirectly alters the blood supply, and abnormal vasomotor control [ 23 ]. Avascular necrosis of the epiphysis can involve the growth plate, which obtains its blood supply from epiphyseal circulation, causing growth inhibition. Peterson described a case in which premature closure of the distal tibial physis occurred in an infant after a temporary but significant episode of vascular insufficiency during surgery to correct developmental dislocation of the right hip [ 28 ]. Rogalski et al. [ 27 ] observed that the proximity of vascular abnormalities to the epiphyseal growth plate was associated with growth disturbance. In his series, 11 out of 41 patients with extremity angiodysplastic lesions developed either hypertrophy or leg length discrepancy. Although vascular malformations have been associated with both undergrowth and overgrowth, all of the patients in this study with leg length discrepancies had overgrowth of the involved limb. These same authors postulated that increased oxygen uptake and increased flow often associated with such vascular malformation contributed to the alteration in growth. It is also conceivable that repeated surgeries to the lateral aspect of the knee in both patients played a contributive role. A change in the environment surrounding the physis, such as muscle atrophy following prolonged bed rest, paralysis, or limb-sparing surgery with muscle loss, is known to cause a significant slowing of growth [ 23 ]. Although efforts were made to keep scarring from surgery to a minimum, re-resection of tumor combined with the damaging effects of radiation to soft tissues would inevitably cause a decrease in tissue vascularity potentially leading to subclinical or clinical necrosis, causing a continual mechanical compression of the physis and retarding growth. Such damage has been documented in heat-related injuries to extremities in which circumferential eschar causes a prolonged ischemia to the physis and subsequent growth inhibition [ 29 ]. The patients included in this report had many of the above noted contributions to abnormal growth and development including a pre-existing condition, vascular compromise due to multiple surgeries, loss of muscle mass, limited use of the extremity for a defined period of time, infection and chemotherapy. Chemotherapy may temporarily reduce bone growth through systemic effects that include the direct effects of specific agents or the indirect effects resulting from systemic infections and abnormalities in metabolism and nutrition. The effects of irradiation on bone growth It has been known for almost a century that radiation therapy at sufficient levels can affect growing bone. Several factors contributing to the severity of effect including the total dose, dose per fraction, dosimetry (asymmetry and inhomogeneity) and age at the time of irradiation [ 24 , 30 , 31 ]. Probert and Parker [ 32 ] studied the standing and sitting height of 44 children who underwent total spinal irradiation for Hodgkin's disease, medulloblastoma or acute lymphoblastic leukemia. Among the patients receiving more than 3,500 rads of spinal irradiation, 8 out of 29 (28%) had a sitting height more than 2 standard deviations below the mean for age. Among those receiving less than 2,500 rads, 6 out of 15 (40%) had a sitting height more than 2 standard deviations below the mean for age. They concluded that doses in excess of 2,000 rads affect vertebral body growth in children. They further noted that children less than six years of age or those undergoing puberty experienced the most significant damage, suggesting that there is an increased sensitivity of bone to irradiation during specific developmental periods. The conclusions of Gonzalez and Breur [ 33 ] were slightly different. In their study, that included 22 patients who experienced growth retardation of long bones as the result of radiotherapy in childhood, definitive limb shortening was strongly dependent on the age of the patient when the irradiation treatment began. When the growth remaining after irradiation was taken into account, no differences in radioresponsiveness were apparent. Their results suggested that despite, a temporary decrease in growth rate, irradiated bone will eventually grow at a similar rate to unirradiated bone. The total dose administered had a major influence on limb shortening as higher doses produced a greater effect. The authors noted that a "saturation dose" was apparent at 40 Gy because higher doses did not appear to produce further considerable increase in shortening. The epiphyseal growth plate is the area of the developing skeleton most sensitive to the effect of radiation due in part to its rapidly proliferating stem cell population. Even low doses of radiation have been shown to cause histologic changes including temporary swelling, fragmentation, and degeneration of chondrocytes [ 31 ]. When higher doses are given, permanent changes including necrosis and premature closure of the physis become evident. Such was the case in our study, as both patients received a total radiation dose of over 50 Gy regionally and nearly 75 Gy focally, which undoubtedly contributed to the observed leg length discrepancies. Furthermore, the angular deformity can be attributed to the unequal dose distribution across the physis as depicted in Figures 3 and 9 . Orthopedic intervention for valgus deformity Creating a treatment plan for patients with leg deformities who have undergone radiation therapy with or without chemotherapy can pose a difficult challenge. Because the osteocytes of neighboring bone are also destroyed, it may take years for the bone to revascularize and repopulate with healthy osteocytes [ 23 ]. The absence of healthy osteoblasts and precursors make lengthening procedures difficult due to unpredictable healing. Radiation damage to regional soft tissues is also an important consideration when planning a lengthening procedure. For these reasons, we decided to delay intervention to correct both the angular deformity and leg length discrepancy, and try to minimize further progression of leg deformity. At the time of surgery, Case 1 had a leg length discrepancy measuring 6 cm clinically. It was felt that epiphysiodesis of the growth plates in the healthy right knee should be the initial treatment, as it would halt progression of discrepancy and allow for some degree of correction, albeit unpredictable because of radiation damage to the growth plates of the left knee. Because of significant angular deformity, case 2 required an osteotomy to correct the defect. Future procedures are planned to address the anticipated leg length inequalities. Historically, there have been two primary treatments for patients with angular limb deformities: epiphysiodesis and stapling. Both methods seek to achieve the same result while offering different sets of advantages and disadvantages. Partial epiphysiodesis of the knee to correct angular deformity was first described by Phemister in 1933 [ 34 ]. It has been used for the correction of idiopathic genu valgum or varum in the adolescent patient. Bowen [ 35 ] described a common surgical technique in which a bone block, centered on the physeal line, was removed through a 2 cm incision, rotated 90 degrees, and reset. This procedure causes growth arrest on the treated side and allows for continued growth and self-correction on the opposite side. Advantages to partial epiphysiodesis include a good assessment of further growth using the Green and Anderson technique [ 36 ], small surgical scar and high predictability of self-correction due to permanent physeal ablation. Disadvantages involve its confinement for use in adolescents due to the irreversibility of the physeal ablation. Also, estimation of skeletal maturity is difficult and unreliable. Physeal stapling, an alternative method, was first reported by Blount and Clark in 1949[ 37 ]. They used stainless steel staples to produce reversible growth retardation and it remains the only reversible means of manipulating growth. This procedure has traditionally been used for adolescents [ 38 ]. Newer radiation delivery techniques including high-dose rate brachytherapy, intraoperative radiation therapy [ 39 , 40 ] and the spectrum of conformal external beam radiation therapy planning and delivery techniques [ 17 ] seek to confine the prescription dose to the region at risk and minimize the dose received by normal tissues. Computerized treatment planning technology and the use of 3-dimensional imaging permits the delineation of both target and normal tissue structures to the extent that the dosimetry for a defined normal tissue structure, such as bone or soft tissue, may be known with a high degree of precision. This information can be used relatively to compare different treatment plans for a given patient. Prospectively assessed, this information may be used as a clinical variable to correlate treatment dosimetry to abnormalities in growth and development including their time to onset and severity [ 41 ]. Until more complete knowledge is available regarding the effects of 3-dimensional dosimetry on bone and soft tissue, the full benefit of these newer treatment techniques will not be realized. We are concerned about the effects of newer treatment technology and the use of more focal irradiation. More focal treatment is likely to result in inhomogeneity and asymmetric irradiation of growth elements in bone. Prospective assessment of the use of these techniques is required. Limited surgery and irradiation may result in growth abnormalities and deformity. These effects may have minimal or significant impact depending on functional outcome and the value attached to limb preservation for a particular patient. As mentioned in this report, both patients and families were queried about their decisions regarding treatment and both reported satisfaction with outcome recognizing that side effects were anticipated. Both families attached a high value to limb preservation. Competing Interests None declared. Authors' Contributions DTF reviewed the patient records and drafted the manuscript. WCW contributed to drafting the manuscript. MDN contributed to drafting the manuscript. TEM conceived of the study, participated in the review of the data, and helped draft the manuscript. All authors read and approved the final manuscript Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC518976.xml
546188
Patients' request for and emergency physicians' prescription of antimicrobial prophylaxis for anthrax during the 2001 bioterrorism-related outbreak
Background Inappropriate use of antibiotics by individuals worried about biological agent exposures during bioterrorism events is an important public health concern. However, little is documented about the extent to which individuals with self-identified risk of anthrax exposure approached physicians for antimicrobial prophylaxis during the 2001 bioterrorism attacks in the United States. Methods We conducted a telephone survey of randomly selected members of the Pennsylvania Chapter of the American College of Emergency Physicians to assess patients' request for and emergency physicians' prescription of antimicrobial agents during the 2001 anthrax attacks. Results Ninety-seven physicians completed the survey. Sixty-four (66%) respondents had received requests from patients for anthrax prophylaxis; 16 (25%) of these physicians prescribed antibiotics to a total of 23 patients. Ten physicians prescribed ciprofloxacin while 8 physicians prescribed doxycycline. Conclusion During the 2001 bioterrorist attacks, the majority of the emergency physicians we surveyed encountered patients who requested anthrax prophylaxis. Public fears may lead to a high demand for antibiotic prophylaxis during bioterrorism events. Elucidation of the relationship between public health response to outbreaks and outcomes would yield insights to ease burden on frontline clinicians and guide strategies to control inappropriate antibiotic allocation during bioterrorist events.
Background The September 11, 2001 terrorism events and the ensuing anthrax attacks in the United States were associated with widespread psychological trauma [ 1 ]. In response to the outbreak, public health officials disseminated recommendations for Bacillus anthracis postexposure prophylaxis and clinical management of patients with anthrax disease [ 2 , 3 ]. For asymptomatic persons, antimicrobial prophylaxis was indicated only for confirmed or suspected aerosol anthrax exposure documented by public health or law enforcement [ 4 - 6 ]. Public health officials assured the public that emergency drug supplies would be delivered to clinical settings as needed from the national antimicrobial stockpile [ 7 ]. Despite such assurances and caution against personal stockpiling and self-medication, media reports of increased demand for ciprofloxacin indicated a potential public health problem [ 8 ]. However, it was not clear to what extent individuals with self-identified risk of anthrax exposure approached physicians for antimicrobial prophylaxis. Additionally, the response of frontline clinicians to these requests had not been described. To address these questions, we used data from a large public health survey of Pennsylvania emergency physicians following the 2001 terrorist attacks [ 9 ]. Methods A total of 250 potential study subjects were randomly selected from the 2001 membership database (n = 1,060) of the Pennsylvania Chapter of the American College of Emergency Physicians (ACEP). Information in the database, such as contact telephone numbers and location of practice, is provided during enrollment or during membership renewals, and no active verification is done. Because the accuracy of the database was unknown, we used an 80% estimate based on a previous study of emergency physicians [ 10 ]. During November 2001-January 2002, we conducted a telephone survey with items designed to assess requests for and prescriptions of antibiotic prophylaxis for anthrax. For the types of antibiotics prescribed, the survey choices were ciprofloxacin, doxycycline, amoxicillin, penicillin VK, other, or none. Emergency departments with post-graduate training programs were categorized as "academic" while all others were categorized as "non-academic." Details concerning survey instrument design and data collection are described elsewhere [ 9 ]. Population and administrative data maintained by the Pennsylvania Department of Health's Bureau of Health Statistics were used to stratify respondents by geographic location. Counties with a population density of ≥ 450 persons per square mile were defined as "urban" while all others were categorized as "rural." "Eastern" Pennsylvania was defined as the Southeast and Northeast Districts of the Pennsylvania Department of Health. All other locations were considered "western." We used one-way frequency analyses to describe distributions of responses for all categorical items. Associations were quantified using odds ratios (ORs) with associated 95% confidence intervals (CIs). Statistical analyses were performed using SAS software (SAS Institute Inc., Cary, NC). Results Forty-three of the 250 physicians in the sample were excluded from the study because of insufficient contact information or because the physician was no longer practicing medicine in the state. Of the remaining 207 physicians (24% of the estimated population of 848 subjects with accurate information), 97 were interviewed (47% response rate). Sixty-four (66%) of the 97 respondents had received patient requests for antimicrobial prophylaxis against anthrax; of the physicians who received requests, 16 (25%) prescribed antibiotics. Physician setting (urban vs. rural; eastern vs. western) was not associated with either patient request for, or physician prescription of, antimicrobial prophylaxis (Table 1 ). Of the 52 respondents in urban areas, 38 (73%) had received requests for antibiotics while 26 (59%) respondents in rural counties had received such requests (odds ratio [OR], 1.9; 95% confidence interval [CI], 0.8 – 4.4). Similarly, the type of institution (academic versus non-academic) was not associated with requests for or prescriptions of antibiotics. Table 1 Physicians' response to patients' requests for anthrax prophylaxis. * Characteristic Setting no (%) † State location no (%) ‡ Emergency department type no (%) § Variable Urban Rural OR (95% CI) Eastern Western OR (95% CI) Academic Non academic OR (95% CI) Received requests for antibiotics 38 (73) 26 (59) 1.9 (0.8–4.4) 33 (70) 31 (63) 1.4 (0.6–3.2) 21 (68) 41 (66) 1.1 (0.4–2.7) Prescribed antibiotics for anthrax 10 (20) 6 (14) 1.5 (0.5–4.7) 6 (13) 10 (21) 0.6 (0.2–1.8) 7 (23) 9 (15) 1.8 (0.6–5.4) Requested testing for anthrax 18 (35) 11 (24) 1.6 (0.7–4.0) 15 (32) 14 (28) 1.2 (0.5–2.9) 8 (26) 21 (33) 0.7 (0.3–1.8) Abbreviations: OR: Odds ratio; CI: Confidence interval. * The ninety-seven survey respondents were used in the analysis. † Fifty-two physicians responded from practices located in one of the eleven "urban" counties that had a population density of ≥450 persons per sq mile in 2001. ‡ Forty-seven respondents were from eastern Pennsylvania, defined as the Northeast and Southeast Districts. § Thirty-one physicians responded from practices that were considered "academic" based on presence of training for residents or fellows. Fifteen physicians prescribed antibiotic prophylaxis to 23 patients; nine physicians prescribed to one patient, four prescribed to two patients, and two physicians prescribed to three patients. One physician did not answer the question regarding number of patients prescribed prophylactic antibiotics. Ten physicians (63%) prescribed ciprofloxacin while 8 physicians (50%) prescribed doxycycline (Figure 1 ). Fourteen (88%) of the respondents that prescribed antibiotics to patients also referred the patients for diagnostic tests for anthrax exposure. Figure 1 Types of antibiotics prescribed by emergency physicians. A total of 16 physicians prescribed various antibiotics shown above. Fifteen physicians prescribed these types of drugs to 23 patients. One physician did not respond to the question on the number of patients the physician had given anthrax prophylaxis. Discussion Following the September 11 terrorist attacks and the ensuing anthrax outbreak, the majority of the emergency physicians we surveyed in Pennsylvania had received patient requests for anthrax prophylaxis; a quarter of these physicians prescribed antibiotics for these patients. Physicians that reported patients' requests for antimicrobial prophylaxis were distributed across the state, suggesting that patients' search for protection against anthrax was widespread. Our results are consistent with data on specimens submitted to Pennsylvania public health laboratory officials for B. anthracis analysis. During October-December 2001, the Pennsylvania Bureau of Laboratories received approximately 1400 specimens including white powder, environmental swabs, and letters from counties in eastern and western Pennsylvania districts. Of these, 27 (or about 18 requests per 100,000 population) came from Cambria County in the western district (PA DOH: unpublished data). There were 11,063 anthrax-related telephone inquiries received from October 8 to November 11, 2001 by the Centers for Disease Control and Prevention's (CDC) emergency operations centers; queries originated from all states and one US territory. Most of these calls were from members of the public concerning anthrax vaccines (≈58%), suggesting that search for protective measures against anthrax was widespread across the United States [ 11 ]. Other studies have documented an increased use of antimicrobial agents that was temporally related to the anthrax outbreak; this use could not be ascribed to that recommended by the CDC. For example, a recent national study reported that approximately 160,000 more ciprofloxacin and 96,000 more doxycycline prescriptions were written in 2001 compared to 2000 [ 12 ]. When other investigators compared ciprofloxacin utilization in 2001 with 2000, they found that it declined for all months except October 2001 when ciprofloxacin utilization increased 9.8%. They also found that the increase was not limited to areas where anthrax cases had occurred, suggesting that many Americans sought antibiotic prophylaxis [ 13 ]. Of the physicians reporting that they had prescribed antibiotics for anthrax prophylaxis, the majority used ciprofloxacin. This finding is consistent with other studies [ 12 , 13 ] and is likely because the initial CDC guidelines recommended ciprofloxacin prophylaxis for B. anthracis exposure until susceptibility results were known [ 4 , 5 ]. When tests showed that B. anthracis isolates recovered from patients involved in the anthrax attacks were susceptible to other antibiotics, public health officials indicated that doxycycline might be preferable over ciprofloxacin [ 2 ]. While both drugs are approved for postexposure prophylaxis [ 14 ], the rationale for favoring doxycycline was to prevent ciprofloxacin resistance in more common bacteria. Unfortunately use of antibiotics has inherent risks and costs and optimizing benefits is especially difficult in the midst of bioterrorist events. Consequences of antibiotic treatment of unexposed individuals include adverse drug reactions, increased risk of antimicrobial resistance, depletion of antibiotics, and monetary costs [ 15 ]. Furthermore, use of emergency departments for sporadic distribution of prophylactic antibiotics to persons presenting with self-identified risk appears inefficient. It is unclear whether these persons can be adequately managed in emergency departments without the support of public health and law enforcement officials. Public health response likely influences demand for and outcomes associated with antibiotics requests during bioterrorism attacks. When we asked physicians to suggest on what health departments could do to reduce the influx of patients to the emergency departments, they cited official communications to make the public "less worried." In Illinois, a surge in environmental samples received by public health officials for anthrax tests was associated with both media reports of anthrax cases in other states and a specific announcement on October 29, 2001 by the US attorney general and the FBI director. The announcement asked US citizens and law enforcement agencies to be on the "highest alert" based on "credible information" [ 16 ]. Lessons learned from the 2001 anthrax attacks in New Jersey suggest that communities in which the public health sector and clinicians have a strong working relationship are better prepared to meet mass prophylaxis needs [ 17 ]. Similarly, lessons learned in New York City during the same outbreak demonstrated the benefits of advance logistical planning for mass postexposure prophylaxis including an antibiotics distribution site and clear eligibility criteria [ 18 ]. We acknowledge some limitations to our results. First, as in any survey, these data are subject to non-response bias. But the 47% response rate is comparable with other telephone surveys conducted among physicians in general [ 19 ]. In addition, responders and non-responders had similar baseline characteristics, suggesting that these groups were comparable [ 9 ]. Second, the study is limited to types of antibiotics prescribed and cannot be used to estimate dosage, number of pills allotted to these patients, costs, or compliance to treatment for perceived or real anthrax exposure. Third, the indications for prophylaxis were not studied. While it is not certain, it is likely that at least the vast majority of the antibiotic perscriptions found in this study were outside indications described in public health guidelines. It is plausible that some patients sought prescriptions for storage; it is also likely that they at least initiated the antibiotic course. Conclusions Taken together with other recent studies of antibiotic utilization [ 12 , 13 ], these data suggest a need for public health response to increased demand for prophylaxis against perceived bioterrorism exposures. Experiences in areas where mass prophylaxis was delivered and suggestions made by the respondents in this study suggest that improved public health communications might reduce the influx of patients to emergency departments. In addition, as demonstrated by New York City's exemplary response to the West Nile virus outbreak in 1999 and the 2001 anthrax attacks, prior logistical plans and strong working relationships among public health officials and clinicians are essential [ 18 , 20 ]. Elucidation of the relationship between public health response to outbreaks and outcomes may offer lessons to reduce inappropriate demand for and prescriptions of antibiotics during bioterrorist events. List of abbreviations ACEP: American College of Emergency Physicians CDC: Centers for Disease Control and Prevention PA DOH: Pennsylvania Department of Health Competing interests The author(s) declare that they have no competing interests. Authors' contributions NMM conceived of the study and its design, and coordinated revision of the manuscript drafts with all authors. KGJ participated in data collection, study design, and contributed to the study's critical review. ARK participated in the design of the study and performed statistical analyses. RCA secured funding and participated in the study's critical review. JTR participated in the study design and coordination. EL contributed to the design of the study and participated in its critical review. All authors read, commented on, and approved of the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546188.xml
523857
Apolipoprotein M
Apolipoprotein M (apoM) is a 26-kDa protein that is mainly associated with high-density lipoprotein (HDL) in human plasma, with a small proportion present in triglyceride-rich lipoproteins (TGRLP) and low-density lipoproteins (LDL). Human apoM gene is located in p21.31 on chromosome 6 (chromosome 17, in mouse). Human apoM cDNA (734 base pairs) encodes 188-amino acid residue-long protein. It belongs to lipocalin protein superfamily. Human tissue expression array study indicates that apoM is only expressed in liver and in kidney and small amounts are found in fetal liver and kidney. In situ apoM mRNA hybridization demonstrates that apoM is exclusively expressed in the hepatocytes and in the tubule epithelial cells in kidney. Expression of apoM could be regulated by platelet activating factor (PAF), transforming growth factors (TGF), insulin-like growth factor (IGF) and leptin in vivo and/or in vitro . It has been demonstrated that apoM expression is dramatically decreased in apoA-I deficient mouse. Hepatocyte nuclear factor-1α (HNF-1α) is an activator of apoM gene promoter. Deficiency of HNF-1α mouse shows lack of apoM expression. Mutations in HNF-1α (MODY3) have reduced serum apoM levels. Expression of apoM is significantly decreased in leptin deficient ( ob/ob ) mouse or leptin receptor deficient ( db/db ) mouse. ApoM concentration in plasma is positively correlated to leptin level in obese subjects. These may suggest that apoM is related to the initiation and progression of MODY3 and/or obesity.
Cloning and characterization of human apoM Human apolipoprotein M (apoM) was found and initially isolated from chylomicrons by Xu and Dahlbäck in 1999 [ 1 ]. When they performed SDS-PAGE for delipidated human triglyceride-rich lipoprotein (TGRLP) and sequenced protein bands ranging from 6–45 kDa, one of sequences identified as the N-terminal sequence of MFHQIWAALLYFYGI. No homologous protein was identified in public databases, but several human expressed sequence tags (EST) were found similar to these N-terminal amino acid sequence. Based on these sequences, full-length cDNA of the novel protein was obtained with 188 amino acids [ 1 ]. Rabbit antibodies were raised against five synthetic peptides based on the protein sequence. The pooled antisera were used to analyze distribution of the protein among various lipoprotein subclasses using Western blotting. Under reducing conditions, a 26-kDa band was particularly predominant in high density lipoprotein (HDL) but was also observed in low density lipoprotein (LDL) and TGRLP. A less pronounced band (approximately 23-kDa) was observed, which corresponded in size to the non-glycosylated protein [ 1 ]. As majority of the protein is associated with lipoprotein in plasma, it fulfills the criteria for classification as an apolipoprotein. And this novel protein was named apolipoprotein M (apoM) [ 1 ] as the last previously identified apolipoprotein was called apoL [ 2 ]. Gel filtration of plasma showed that apoM was associated with HDL-sized particles in wide-type and apoA-I deficient mice and with HDL- and LDL-sized particles in LDL receptor-deficient mice, whereas it was mainly found in VLDL (very low density lipoprotein)-sized particles in high-fat, high-cholesterol-fed apoE deficient mice [ 3 ]. These data suggest that apoM mainly associates with HDL in normal mice, but also with the pathologically increased lipoprotein fraction in genetically modified mice. Gene location and amino acid sequence of apoM The identified human apoM cDNA (734 base pairs) encoded 188-amino acid residue-long protein. The 5'-untranslated region was 33 nucleotides and the 3'-untranslated region 120 nucleotides, not including the poly (A) tail. Southern blot analysis of different species gave positive signals in all mammalian genomes but not in DNA from chicken and yeast [ 1 ]. Human apoM gene is located in p21.31 on chromosome 6 (Fig. 1 ) (chromosome 17, in mouse). The genomic sequence of this region was determined and the human apoM gene identified (GenBank accession number AF118393). In human genome, the apoM gene is surrounded by BAT4 and NG34 on one side and BAT3 on the other. Both mouse apoM gene and its human counterpart are predicted to contain 6 exons enclosed in a 1.6-kb genomic region, which is consistent with the results of Southern blotting. The calculated molecular mass of the protein was 21,256. There is one potential site for N-linked glycosylation at Asn-135 (Asn-Glu-Thr), whereas Asn-148 (Asn-Arg-Ser-Pro) is less likely to be glycosylated because Pro-151 follows Ser-150. The amino acid sequences of human and mouse apoM are 79% identical (82%, human and rat apoM) (Fig. 2 ), and just like human apoM the mouse sequence predicts the presence of a signal anchor, as there is no predicted signal peptidase cleavage site. The amino acid sequence of apoM contained six cysteines, which may involve in the formation of three disulfide bridges. Figure 1 ApoM gene location in chromosome 6 . ApoM gene is located in chromosome 6 p21.31 . Figure 2 Comparison of apoM amino acid sequence of rat, human and mouse . Dots indicate residues that are identical to the top line (rat). One potential site for N-linked glycosylation site (Asn-Glu-Thr) is indicated by a large dot above the sequence of human apoM (○). Two sequences underlined indicate the typical lipocalin motifs. Protein structure of apoM Based on sensitive sequence searches, it is proposed that apoM is related to the lipocalin protein superfamily (Fig. 2 ) [ 1 ]. Subsequently, Duan et al. [ 4 ], used computer protein modeling of two lipocalins, mouse major urinary protein (MUP) and human retinol binding protein (RBP) as initial templates to build apoM protein structure, which demonstrated that apoM has the same structure of lipicalin protein superfamily. ApoM retains an uncleaved N-terminal signal peptide that most likely anchors the molecule into single layer lipids on HDL [ 4 ]. The major phospholipids in HDL is phosphatidylcholine, which has a positively charged choline group exposed to the solvent. Two electronegative regions are striking in the apoM model and are located around the N-terminus and the opening of the binding pocket. In this three-dimensional model, characterized by an eight-stranded anti-parallel β-barrel, a segment including Asn-135 could adopt a closed or open conformation. ApoM presents three disulfide bridges, which would make it a member of the lipcalin subgroup of proteins with three s-s bonds [ 4 ]. Tissue distribution and cellular expression of apoM Northern blot analyses of multiple tissues (including spleen, thymus, prostate, testis, ovary, small intestine, colon, leukocytes, heart, brain, placenta, lung, liver, skeletal muscle, kidney, pancreas, stomach, thyroid, spinal cord, lymph node, trachea, adrenal gland and bone marrow) showed that apoM was mainly expressed in kidney and liver [ 1 ]. Furthermore, human tissue expression array study indicated that apoM is only expressed in liver and in kidney and small amounts were found in fetal liver and in fetal kidney [ 5 ]. To elucidate whether and when apoM is expressed, Zhang et al. investigated apoM expression patterns during mouse and human embryogenesis [ 6 ]. ApoM transcripts were detectable in mouse embryos day 7.5 to day 18.5. It was expressed at low levels at day 7.5, increased significantly at day 9.7 and decreased at day 10.5, and then increased continually up to day 18.5. ApoM-positive cells appeared mainly in liver of day 12 embryos as detected by in situ hybridization. In day-15 embryos, apoM was expressed in both liver and kidney. During human embryogenesis, apoM was strongly expressed in livers of 3–5 month-old human embryos and continued to be strongly expressed throughout embryogenesis. In the kidney, apoM expression was highest in 5–9 month-old embryos. There was some expression of apoM in small intestine, particularly in later stages of embryogenesis. In skeletal muscle, minute apoM expression was found in 3–5 months-old embryos, and some apoM expression was found in stomach in earlier stages of embryogenesis [ 6 ]. These finding suggest that apoM has high organ specificity and strongly indicate that the physiological function of apoM must be related with liver and kidney. Both immunohistochemical staining and in situ apoM mRNA hybridization demonstrated that apoM is exclusively expressed in the hepatocytes and in the tubule epithelial cells in human kidney [ 5 ]. Thus, apoM may have specific function in vivo , which may be related to the hepatic lipid and/or lipoprotein metabolism. Regulation of apoM expression In vitro , several biological factors have been tested to examine their influences over the transcription and secretion of apoM in hepatic cell line (HepG2 cells). Like apoB, apoM is highly hydrophobic and must co-circulate with lipoprotein particles in the blood stream. It has been demonstrated that apoB could be down-regulated by transforming growth factor-beta (TGF-β) [ 7 , 8 ]. Xu et al reported that TGF-β could also down-regulate apoM expression and secretion in HepG2 cells [ 9 ]. It suggests that apoM, similar to apoB, may involve in the hepatic lipoprotein metabolism in vivo . In another study, Xu et al demonstrated that platelet-activating factor (PAF) could up-regulate apoM expression in HepG2 cells, whereas, lexipafant, a PAF-receptor antagonist significantly suppressed the mRNA levels and the secretion of apoM in HepG2 cells in a dose-dependent manner. Neither tumor necrosis factor-α (TNF-α) nor interleukin-1α (IL-1α) influences apoM expression or secretion in HepG2 cell cultures [ 10 ]. It indicates that apoM may relate to the host defense response because apoM gene is located in histocompatibility complex III (HMC-III) region on chromosome 6. Many genes in this region are related to the immune response, and the apoM gene is very close to the TNF-α gene and lymphotoxin genes. Thus, apoM may also be related to the immune response system, or regulated by cytokines or other inflammatory factors. Administration of adrenocorticotropic hormone (ACTH) has beneficial effects on plasma lipoproteins [ 11 - 15 ]. A consistent decrease of plasma total cholesterol and LDL cholesterol by 20–40% is seen during ACTH treatment [ 14 - 17 ]. It has been demonstrated that pronounced hypolipidimic effects of ACTH might be related to the inhibition of apoB synthesis in hepatic cells [ 18 ]. However ACTH didn't influence apoM expression and secretion in vivo and in vitro [ 18 , 19 ], indicates that apoM may have somewhat difference from apoB on lipid and/or lipoprotein metabolism in vivo . Richter et al. reported that apoM gene expression could be regulated by HNF-1α. Mutant HNF-1α-/- mice completely lack expression of apoM in liver and kidney. Serum apoM levels in HNF-1α+/- mice are reduced by 50% compared with wild-type animals. By analyzing the apoM promoter and identifying a conserved HNF-1 binding site, they showed that HNF-1α is a potent activator of apoM promoter, that a specific mutation in the HNF-1 binding site abolished transcriptional activation of apoM gene. HNF-1α protein can bind to the HNF-1 binding site of apoM promoter in vitro [ 20 ]. Liang and Tall reported that leptin up-regulated mRNA level of apoM in ob/ob mice [ 21 ], suggesting that leptin could stimulate hepatic cells to produce apoM. Faber et al. reported that plasma concentration of apoM was similar in wild-type, LDL receptor-deficient and apoE deficient mice but was reduced by 33% in apoA-I-deficient mice compared with the wide-type mice, which suggest a connection between apoM and apoA-I metabolism [ 3 ]. Xu et al., found that in both liver and kidney, expression of apoM was significantly lower in leptin deficient ob/ob mice and in leptin-receptor deficient db/db mice than in control mice. Furthermore, leptin administration significantly increased plasma apoM levels and apoM mRNA levels in liver and in kidney in ob/ob mice [ 22 ]. It is concluded that both leptin and leptin-receptor are essential for the apoM expression, indicating that leptin is a physiological regulating factor on apoM synthesis in vivo . Physiopathology and potential clinical importance of apoM Xu et al. investigated the relationship between plasma apoM levels and leptin levels, body mass index (BMI), fasting glucose, fasting insulin as well as lipoprotein concentrations in females displaying a wide range in BMI (18.9–57.1 kg/m 2 , n = 51). In univariate analysis, apoM correlated significantly to leptin (r = 0.54, P < 0.001), BMI (r = 0.70, P < 0.001), fasting insulin (r = 0.33, P = 0.025), total cholesterol (r = -0.41, P = 0.016), and LDL-cholesterol (r = -0.39, P = 0.018). The correlations between apoM and cholesterol and between apoM and leptin remained significant after adjustment for the influence of BMI. Forward stepwise multiple regressions when leptin, BMI, insulin and cholesterol were entered in a model as independent variables and apoM as the dependent variable showed that cholesterol and leptin were independent predictors of circulating apoM. These two parameters yielded an r 2 of 0.28, thereby explaining approximately 30% of the variance in apoM. Hence, apoM is positively correlated to leptin and negatively correlated to cholesterol levels in humans [ 23 ]. Richter et al. measured apoM levels in the serum of nine HNF-1α /maturity-onset diabetes of the young (MODY3) patients, nine normal matched control subjects (HNF-1α +/+), and nine HNF-4α /MODY1 subjects. Serum levels of apoM were significantly decreased in HNF-1α /MODY3 subjects when compared with control subjects as well as with HNF-4α /MODY1 subjects, indicating that HNF-1α haploinsufficiency rather than hyperglycemia is the primary cause of decreased serum apoM protein concentrations. Thus, serum levels of apoM may be a useful serum marker for the identification of MODY3 patients [ 20 ]. Alzheimer's disease (AD) is a complex, multifactor disorder, probably resulting from an interaction between environmental and genetic factors [ 24 - 26 ]. Increasing evidence points to a link between cholesterol turnover and AD [ 27 , 28 ], suggesting that genes implicated in brain cholesterol homeostasis may be potential candidate genes for AD. It is well known that apoE genotype and apoE receptor are related to AD [ 28 , 29 ]. With this background, Kabbara et al examined association of apoM with the risk of developing AD. It is excluded apoM as a genetic determinant of AD in a large French case control population [ 30 ]. Conclusion In conclusion, apoM is a novel HDL apolipoprotein. Like apoB apoM could be regulated by several cytokines in vivo and in vitro . HNF-1α is one of the most important activator of apoM gene promoter. Plasma apoM concentration is positively correlated to leptin levels and negatively related to plasma cholesterol levels. Both leptin and leptin receptor are essential for apoM expression in vivo . Plasma apoM levels may be used as the marker for identification of MODY3. The detailed relationship between apoM and MODY3 as well as obese needs further investigation.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC523857.xml
544600
Arrays of ultraconserved non-coding regions span the loci of key developmental genes in vertebrate genomes
Background Evolutionarily conserved sequences within or adjoining orthologous genes often serve as critical cis -regulatory regions. Recent studies have identified long, non-coding genomic regions that are perfectly conserved between human and mouse, termed ultra-conserved regions (UCRs). Here, we focus on UCRs that cluster around genes involved in early vertebrate development; genes conserved over 450 million years of vertebrate evolution. Results Based on a high resolution detection procedure, our UCR set enables novel insights into vertebrate genome organization and regulation of developmentally important genes. We find that the genomic positions of deeply conserved UCRs are strongly associated with the locations of genes encoding key regulators of development, with particularly strong positional correlation to transcription factor-encoding genes. Of particular importance is the observation that most UCRs are clustered into arrays that span hundreds of kilobases around their presumptive target genes. Such a hallmark signature is present around several uncharacterized human genes predicted to encode developmentally important DNA-binding proteins. Conclusion The genomic organization of UCRs, combined with previous findings, suggests that UCRs act as essential long-range modulators of gene expression. The exceptional sequence conservation and clustered structure suggests that UCR-mediated molecular events involve greater complexity than traditional DNA binding by transcription factors. The high-resolution UCR collection presented here provides a wealth of target sequences for future experimental studies to determine the nature of the biochemical mechanisms involved in the preservation of arrays of nearly identical non-coding sequences over the course of vertebrate evolution.
Background Comparative genome sequence analysis, often termed phylogenetic footprinting, has proven successful for the identification of cis -regulatory regions[ 1 , 2 ]. Recent computational and experimental studies have identified a small number of large, highly conserved enhancers, or 'global control regions', associated with the regulation of important developmental genes such as DACH [ 3 ], SOX9 [ 4 ], Dlx bigene [ 5 , 6 ], and HOX-D [ 7 , 8 ] clusters. These regulatory regions can act at distances of several hundred kilobases from their target genes, while at the same time conferring an equivalent expression pattern to reporter genes over much shorter distances (e.g. [ 3 ]). A recent computational analysis proves that such highly conserved elements (termed ultra-conserved elements (UCRs)) are occurring far more often than expected [ 9 ]. In the study by Bejerano et al ., UCRs are defined as regions perfectly conserved between human and mouse longer than 200 base pairs (bp). The study reports a significant association of a non-transcribed subset of those elements with DNA-binding proteins; an equivalent observation has been made independently by Boffeli et al .[ 10 ] for a limited number of most highly conserved elements between human and pufferfish. The stringent criteria for conservation applied in the two studies miss many known enhancer elements that are shorter than 200 bp, and highly conserved across all vertebrates. For instance, in a recently published study, Sabarinadh et al . [ 11 ] described a number of non-transcribed regions flanking the genes of HoxD gene cluster that are highly conserved across vertebrate genomes. In this paper, we define a set of UCRs using high-resolution criteria that detect segments conserved between the human, mouse and pufferfish genomes. Analysis of this set provides insights into a previously unrealized organizational structure of UCRs in vertebrate genomes. We conclusively show that clusters of UCRs are globally associated with many of the genes that act as master regulators during vertebrate development. The clustered distribution of these regions along chromosomes and, importantly, around their presumptive target genes suggests that gene regulation involves the coordinated action of numerous, widely dispersed elements. Results Definition and genomic environment of ultra-conserved non-coding regions (UCRs) We initiated this study by applying comparative genomics to identify putative regulatory regions for a number of evolutionary conserved homeodomain transcription factors that control neural cell fate determination [ 12 , 13 ]. When we examined the genomic landscapes surrounding homeodomain gene loci, we consistently found non-coding regions that exhibited a striking degree of sequence conservation between human and mouse over a minimum of 50 bp. Many of these regions are at least partially conserved over extended periods of evolution. The observed nucleotide identities between human and mouse sequences exceed even those of exon sequences encoding identical proteins. Such striking sequence conservation has previously been anecdotally associated with long-range enhancers for several developmental genes [ 3 - 8 ]. To test whether the association of UCRs with regulatory genes reflected a global genomic trend, we identified a comprehensive set of human/mouse/pufferfish UCRs for detailed analysis. We defined minimum requirements for a UCR (see Methods) and performed a genome-scale computational analysis that retrieved 3583 human/mouse/pufferfish UCRs. Since one of the requirements is that the UCRs are not overlapping actively transcribed genomic regions, they would belong to type II UCRs defined by Bejerano et al . [ 9 ]. The median UCR length was 125 bp, but extreme lengths (>1000 bp) were observed. Qualitative assessment of "genescapes", the gene structures, surrounding UCRs revealed them to be present either in introns, in dense clusters around a group of genes or in 'gene deserts' (up to several thousands kilobases from known genes). There appeared to be a strong association between locations of our set of UCRs and genes encoding transcription factors – even stronger than that reported by Bejerano et al .[ 9 ] [see Additional file 1 and 2 ]. This observation will be proven in the subsequent analysis. UCRs are strongly associated with DNA-binding proteins To quantitatively assess the characteristics of genes proximal to UCRs, we analyzed the over-representation of gene annotations. We retrieved the InterPro [ 14 ] domain annotation for all genes adjacent to or containing UCRs. A statistical assessment (Fisher's exact test) of the observed domain biases for these genes was performed to assess the probability that the domain distributions were the same for the UCR genes as compared to the set of all genes. Even with a conservative (Bonferroni) correction for multiple testing [ 15 ], structural domains of transcription factors are significantly over-represented (P-value 9.33e-66) within the gene annotations (Table 1 ) [all domains are listed in Additional file 3 and 4 ]. In order to obtain robust results, we chose the four domains from Table 1 present in the highest number of proteins (homeobox, C2H2 zinc finger, forkhead and nuclear steroid receptor). We examined the extent to which all known genes containing each of these four transcription factor domains co-localize with UCRs (Figure 1 ). We found that a high proportion of these genes (163/1084; P-value 7.33e-11) are in genomic neighborhoods (<8 kb) of UCRs: more than 30% of all homeodomain-encoding genes have an UCR within 8 kbp (90/237; p-value 8.67e-11), and more than 55% have one within 100 kb (133/237, P-value 7.78e-11). The UCR association rates (the fraction of genes with an UCR closer than 8 kb, compared to the expected value) for genes encoding forkhead (8/31, P-value 6.6e-11), nuclear steroid receptor (9/38, P-value 2.81e-9) or zinc finger domains (56/751 P-value 8.12e-11) were noted as significant as well. These data provide strong evidence that the UCRs are spatially associated with genes encoding regulatory proteins. Table 1 Over-representation of protein domains in genes flanking UCRs. Bonferroni-corrected and uncorrected Fisher Exact Test p-values are shown for the 16 most over-represented InterPro domains. Typical transcription factor domains (DNA binding domains) are indicated in bold. A full list of all InterPro domains with P-values is given in [Additional file 3]. Domain description INTERPRO ID Fisher test P value Corrected P value HTH_lambrepressr IPR000047 6.40E-20 5.36E-17 Homeobox IPR001356 1.60E-12 1.34E-09 Antennapedia IPR001827 1.37E-10 1.15E-07 Paired_box IPR001523 2.39E-05 2.00E-02 HLH_basic IPR001092 2.40E-05 2.01E-02 POU_domain IPR000327 3.06E-05 2.56E-02 Homeo_OAR IPR003654 3.08E-05 2.58E-02 TF_Fork_head IPR001766 6.15E-05 5.15E-02 Znf_C4steroid IPR001628 7.45E-05 6.23E-02 Hormone_rec_lig IPR000536 1.06E-04 8.86E-02 HMG_12_box IPR000910 1.81E-04 1.51E-01 Stdhrmn_receptor IPR001723 2.63E-04 2.20E-01 COUP_TF IPR003068 7.62E-04 6.38E-01 LIM IPR001781 1.10E-03 9.18E-01 RtnoidX_receptor IPR000003 1.28E-03 1.07E+00 FN_III IPR003961 2.57E-03 2.15E+00 Figure 1 Spatial correlation of transcription factor gene families to UCRs in the human genome . A . Cumulative distribution of distances to the closest UCR for selected subsets of genes. Distance to the closer end of the transcript mapping (either 3' or 5'). Majority of major classes of transcription factors are closer to UCRs than random genes. B, C . Occurrence of UCRs around selected subsets of genes. This plot summarizes the distribution of distances to all UCRs on the same chromosome for each gene in the subset. There is a visible over-representation of UCRs up to 300 kb from homeobox genes, and up to 150 kb from C2H2 zinc finger genes. UCRs clusters encompass the entire gene loci of key developmental genes In order to visualize the distribution of UCR locations across the human genome, we generated a UCR density map for each chromosome [see Additional file 5 ]. Figure 2a shows such a map for chromosome 2. Visual inspection reveals an obvious qualitative tendency of UCRs to occur in large clusters, which was validated by a quantitative comparison of the distributions of nearest-neighbor distances between UCRs and a neutral background model (P-value 8.02e-16; Kolmogorov-Smirnov test). There is no observed correlation between regions of high gene density and UCRs, consistent with previously reported observations that larger conserved regions can be located in gene deserts [ 3 ]. As previously noted, many of the UCRs are adjacent to homeobox protein-encoding genes (Figure 1a , Figure 2b ). It is interesting to note that the over-representation of UCRs near homeobox genes extends up to 300 kbp away from the transcription start site (Figure 1b ). This is consistent with numerous observations that control regions need not be proximal to targeted genes, but can be located hundreds of kilobases from the transcription start site [ 3 , 7 , 16 ]. A similar trend is observed for UCRs near C2H2 zinc finger genes, with over-representation of UCRs extending up to 150 kbp away (Figure 1c ). Large clusters of UCRs can span regions of several hundred kilobases around inferred target genes. For the 50 largest UCR clusters we generated comprehensive views of the chromosomal neighborhood (Figure 3 ). We find that 41 of the 50 clusters span one or more genes known to be expressed in embryonal development, including fundamental master regulator genes ( i.e . the HoxD cluster, Nkx6.1, Nkx2.2 and Pbx3) [for detailed annotated lists of genes associated with UCR clusters, see Additional files 6 and 7 ]. To provide access to the entire set of UCRs, we have implemented a basic UCR browser with links to the UCSC genome browser [ 17 ]. Figure 2 Genomic distributions of UCRs and transcription factor genes . A . Distribution of UCRs on human chromosome 2 is shown in yellow, and total gene density along the chromosome is shown in blue (top track). Note the lack of correlation between gene density and UCR density. Positions of homebox-domain containing genes locations are marked in red, and generally coincide with local maxima of UCR density. The remaining UCR density peaks coincide with genes for transcription factors belonging to structural classes other than homeobox. B . Close-up of a UCR cluster coinciding with the HoxD gene cluster. The HoxD cluster coincides with one of the larger UCR density peaks on chromosome 2, and is associated with nine UCRs. UCR locations are shaded in yellow. Figure 3 Genomic landscape surrounding the most prominent UCR clusters in the human genome . UCRs were counted by sliding a 500 kb window along the chromosomes. Overlapping UCR-containing windows were merged into a single cluster span. Each of the regions shows a 4 MB region around the corresponding UCR cluster. The cluster span coordinates correspond to the human genome NCBI build 33 (UCSC hg15, April 2003). Transcription factor genes are colored according to structural class. UCR clusters are visibly correlated with transcription factor genes; other developmental regulators that do not contain any of the probed protein domains were located manually (boxed), such as the autism susceptibility gene (chromosome 7, number 37) and the DACH gene (chromosome 13, number 10). The numbers correspond to annotations in [Additional file 6 and 7]. The figure was created with the help of the Bio::Graphics Perl library[27]. Rare duplications of UCRs across evolution We performed a global pairwise comparison of all UCRs, in order to determine if UCR duplication was common across evolution. We discovered only five sets of duplicated UCRs, all of which are adjacent to corresponding duplicated genes. For example, duplicated UCRs are present in the introns of SOX5 (on chromosome 12) and SOX6 (on chromosome 11), two highly similar genes involved in chondrocyte differentiation [ 18 ]. Of special interest is the conservation of UCRs in the Iroquois ( IRX ) gene clusters. IRX genes are situated in two clusters of three genes each, present on human chromosomes 5 and 16 [ 19 ]. Similarly positioned arrays of UCRs are present in each of the four intergenic regions between the IRX genes (Figure 4 ). The great majority of UCRs, while conserved across vertebrate evolution, show no similarity between the clusters within the species. An intriguing exception is the set of four UCRs that are highly similar in both cluster position and nucleotide sequence. Figure 4 Sets of UCRs sharing high sequence similarity are involved in regulation of related genes: the case of Iroquois gene clusters . Four similarly positioned UCRs are located within the two Iroquois gene clusters at chromosomes 5 and 16. Block arrows indicate significant sequence similarity. The arrow width is inversely proportional to the alignment BLASTN E-value. There are additional shorter blocks of similarity between the two three-gene clusters; however, most UCRs have diverged between the two clusters, while still preserved across vertebrates. Discussion The human genome contains numerous ultra-conserved regulatory sequences that are shared broadly across vertebrates. These UCRs occur in arrays of highly conserved regulatory elements spanning large chromosomal regions . The clusters are co-localized with genes encoding key proteins for the regulation of development, with a particular correlation with genes encoding transcription factors. The strength of association between UCRs and diverse classes of DNA binding transcription factors validates that a relatively simple definition of UCRs captures a biologically meaningful set of functional sequences. The presence of non-coding UCRs is predictive for the presence of genes implicated in development, differentiation and malignancies. The list presented in [ Additional file 6 ] hints at potentially crucial roles of currently uncharacterized transcription factor genes, while the collection of reported UCRs provides a wealth of regulatory locations for further study. Exceptional mechanisms are brought to bear to retain UCRs over hundreds of millions of years of parallel evolution. UCRs are more strongly conserved than sequences encoding identical proteins, and exhibit sequence identity exceeding essentially all known cis-regulatory sequences. The retention properties suggest that UCRs have important functions in the vertebrate genome. The observed UCRs could fall into multiple functional categories, including enhancers of transcription, regulators of chromatin structure and unknown genes for non-coding transcripts. A small subset of UCRs have been identified previously as enhancers of transcription [ 7 , 3 ]. The high conservation and length of UCRs compared to binding sites for single transcription factors suggests that the mode of regulation must involve more than the binding of small number of transcription factors. Homeotypic clusters of binding sites, as seen in developmental genes in Drosophila melanogaster [ 20 ], represent one regulatory mechanism that could explain the occurrence of long, conserved non-coding regions. However, as transcription factors tolerate considerable variation between functional binding sites, a homeotypic cluster of binding sites as such cannot warrant the extreme level of conservation observed in UCRs. Alternatively, the recent emergence of the role of microRNAs in regulation suggests that there could be additional non-coding genes in the human genome, perhaps at the sites of ultra-conservation. The clustering of UCRs suggests that UCR-mediated transcriptional regulation may involve molecular events on a greater scale, possibly involving chromatin structure. This potential link to chromatin structure is suggested by the striking pattern of UCRs in the IRX gene clusters. Most of the UCRs have no similarity between the two clusters, with the exception of a set of four UCRs that have retained both mutual sequence similarity and spatial position (Figure 4 ). It is tempting to assume that the retention of their mutual similarity is a consequence of IRX cluster co-regulation, the mechanism of which remains unknown. Based on the preservation of nearly identical sequences over ~450 million years of vertebrate evolution, it is reasonable to postulate the influence of exceptional biochemical mechanisms. Numerous hypotheses could account for the observed data, broadly falling into two categories – active mechanism(s) resulting in the decrease of mutational frequency in UCRs, or negative pressure consistent with evolutionary selection against such mutations. Given the breadth of possibilities, we leave postulation until further data emerges. Conclusion Since Bejerano et al .[ 9 ] focused on larger regions (200 bp) of perfect nucleotide identity compared to our more permissive settings (95% sequence identity over 50 bp), the genomic arrangement of UCR-containing regions with respect to their presumptive target genes was not fully realized. Our findings include critical new information about UCR clusters, particularly with regards to patterns of conservation, their genomic organization, and the insights they provide into potential chromatin regulating mechanisms. These mysterious regions retained over hundreds of millions of years of evolution appear to contribute to a novel mechanism of developmental regulation. Detailed studies of UCRs that will ensue from the discoveries reported here promise to advance our understanding of vertebrate development. Methods Definition of UCRs applied in this study We defined UCRs as non-protein coding genomic regions having a sequence identity > 95% over a 50 bp sliding window of length in human/mouse comparison (based on the tight alignments track from the UCSC genome browser database[ 17 ], using human and mouse assemblies hg15 and mm3, respectively). As a further constraint, an UCR must overlap with sequences conserved between the human and pufferfish genomes, as defined in the UCSC genome browser databases (a BLAT [ 21 ] alignment between human and pufferfish with a minimum BLAT score of 20). In order to avoid inclusion of coding sequence, we required that a UCR must not overlap a mouse or human cDNA mapped to the genome (based on cDNA tracks from from the UCSC genome browser database[ 17 ]) or overlap putative coding regions predicted by GenScan [ 22 ]. Calculation of UCR and gene distributions The distribution of UCRs in the genome was calculated by counting the number of UCRs within a 500 kilobase (kb) window which was progressively slid over each chromosome in 100 kb intervals. The same approach was used to estimate the gene density; specifically by summing the number of bases within the window that aligned with human mRNA (from the UCSC Genome Browser database). Gene-UCR distance calculation Distances between a given gene and UCR on the same chromosome were defined as the shortest distance between the starting points and/or endpoints of UCR and gene in the human genome (UCSC assembly hg15), using EnsEMBL [ 23 ] gene annotation. Genes based solely on ESTs or computational predictions were not included. Estimation of significance of Gene-UCR distances The distances from genes within a set (for instance, all forkhead domain-containing genes) to the closest UCRs were calculated as above. The expected fraction of gene-UCR distances smaller than 8 kb was estimated by simulation: UCR genome coordinates were randomly chosen and distances measured as above. The simulation process was repeated 1000 times and the average fraction reported. In order to estimate if the observed distribution was significantly different from the expected, we used the chi-squared test. Estimation of domain over-representation in genes closest to UCRs For each UCR, the closest upstream and downstream gene within 2 Mbp was identified (UCRs inside introns of genes were analyzed separately). EnsEMBL InterPro [ 14 ] domain annotation was used to tabulate a contingency table consisting of the positive sample counts (number of genes in the set containing a certain domain), negative sample counts (number of remaining genes in the set), background positives (number of genes containing the same domain in the genome) and background negatives (remaining genes). For clarity, a given gene was only counted once, and multiple occurrences of the same domain within the same protein were not counted. For each domain found in the UCR-proximal genes, we tested the null-hypothesis that the sample and background sets are drawn from the same population versus the alternative hypothesis that the sample set has a higher frequency of the domain, using Fisher's Exact Test [ 24 ] from the R statistical package . Since the number of tests is considerable, we corrected for multiple sampling using the conservative Bonferroni method [ 15 ], in which the number of tests is multiplied with the P-value from the Fisher test with the number of unique domains tested (837). An analogous analysis was performed with genes containing one or more UCRs within their introns [see Additional file 4 ]. Estimation of clustering tendency We used the distances between consecutive UCRs as a statistic indicating clustering. A neutral background distance distribution was created by assigning UCRs genome coordinates randomly, and subsequently measuring distances between consecutive UCRs. This process was repeated 1000 times. We compared the distance distribution between naturally occurring UCRs and the background using the Kolmogorov-Smirnov test [ 25 ], which assigns a probability that two distributions are similarly shaped. UCR sequence similarity analysis All possible pairs of UCRs were aligned using NCBI BLASTN [ 26 ] with standard settings. For any pair to be reported as near-identical, we required an HSP of at least 50 bp and a pairwise sequence identity exceeding 75%. Abbreviations UCR – ultraconserved non-coding region; bp – basepairs; kbp – 10 3 base pairs Authors' contributions AS collected the data and performed most steps of bioinformatic and statistical analysis presented in the paper. He produced all the Figures in the paper and Table 1 , and co-wrote the manuscript. PB made initial analyses of putative regulatory elements on selected genes involved in neural tube development. He discovered a number of super-conserved regions in the process, which helped create the rules for their genome-wide computational detection. He also co-wrote the first versions of the manuscript. SB participated in the annotation of the gene set and in the creation of software for the visualization of results. PE prepared genome sequence and annotation data for human, mouse and pufferfish. He and AS designed the statistical tests applied in the study. JK participated in the initial analyses and data extraction with PB. WWW participated in result interpretation, design of statistical tests, and writing later versions of the manuscript. JE initiated and co-supervised the study, which has the roots in his research in developmental neurobiology. He also co-wrote the manuscript. BL designed and supervised the bioinformatic study, developed the initial framework for the analysis of the genomic sequences, made an independent observation about high incidence of clustering of super-conserved regions around genes encoding DNA-binding proteins, and annotated the UCR clusters with co-localizing genes. He also co-wrote the manuscript. Supplementary Material Additional File 1 Genescape around 50 randomly selected UCRs . Selected UCRs are shown as yellow triangles, other UCRs as light yellow triangles. Genes are colored after domain (red = Homeobox, green = C2H2 Zink fingers in green, pink = Nuclear receptors, Blue = forkhead). Click here for file Additional File 2 Genescape around 50 randomly selected genes . UCRs are shown as as light yellow triangles. Color coding of genes as above. Click here for file Additional File 3 Complete list of protein domains in genes flanking UCRs . Each tested domain is listed along with corrected and uncorrected P-value as in Table 1 . Click here for file Additional File 4 Complete list of protein domains in genes with UCR(s) in intron(s) Each tested domain is listed along with corrected and uncorrected P-value as in Table 1 . Click here for file Additional File 5 UCR distribution in the human genome UCR density (pink) and gene density (blue) is shown for each chromosome. Densities are calculated as described in Methods. Click here for file Additional File 6 Genes associated with enumerated UCR clusters from Figure 3 . UCRs were counted by sliding a 500 kb window along the chromosomes. Overlapping UCR-containing windows were merged into a single cluster span. The cluster span coordinates correspond to the human genome NCBI build 33 (UCSC hg15, April 2003). A more exhaustive list is found in [ Additional file 7 ] Click here for file Additional File 7 Extended list of UCR clusters An extended, but less annotated, version of in [ Additional file 6 ] Click here for file
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544600.xml
544947
Insulin expressing hepatocytes not destroyed in transgenic NOD mice
Background The liver has been suggested as a suitable target organ for gene therapy of Type 1 diabetes. However, the fundamental issue whether insulin-secreting hepatocytes in vivo will be destroyed by the autoimmune processes that kill pancreatic β cells has not been fully addressed. It is possible that the insulin secreting liver cells will be destroyed by the immune system because hepatocytes express major histocompatibility complex (MHC) class I molecules and exhibit constitutive Fas expression; moreover the liver has antigen presenting activity. Together with previous reports that proinsulin is a possible autoantigen in the development of Type 1 diabetes, the autoimmune destruction of insulin producing liver cells is a distinct possibility. Methods To address this question, transgenic Non-Obese Diabetic (NOD) mice which express insulin in the liver were made using the Phosphoenolpyruvate Carboxykinase (PEPCK) promoter to drive the mouse insulin I gene (Ins). Results The liver cells were found to possess preproinsulin mRNA, translate (pro)insulin in vivo and release it when exposed to 100 nmol/l glucagon in vitro . The amount of insulin produced was however significantly lower than that produced by the pancreas. The transgenic PEPCK-Ins NOD mice became diabetic at 20–25 weeks of age, with blood glucose levels of 24.1 ± 1.7 mmol/l. Haematoxylin and eosin staining of liver sections from these transgenic NOD PEPCK-Ins mice revealed the absence of an infiltrate of immune cells, a feature that characterised the pancreatic islets of these mice. Conclusions These data show that hepatocytes induced to produce (pro)insulin in NOD mice are not destroyed by an ongoing autoimmune response; furthermore the expression of (pro)insulin in hepatocytes is insufficient to prevent development of diabetes in NOD mice. These results support the use of liver cells as a potential therapy for type 1 diabetes. However it is possible that a certain threshold level of (pro)insulin production might have to be reached to trigger the autoimmune response.
Background Genetic alteration of non-pancreatic cells in a diabetic person to synthesise, store and secrete insulin in the same manner as a pancreatic β cell is a potential therapy of type 1 diabetes. The hepatocyte has been suggested as a suitable target cell for such gene therapy [ 1 - 9 ]. Such cells made in vitro are capable of synthesizing and storing pro(insulin) and can secrete this peptide in response to a physiological challenge with glucose [ 1 , 2 ]. Moreover, when transplanted into diabetic mice these cells can lower blood glucose levels to the normal range [ 2 ]. Further, Ferber et al. [ 9 ] has recently showed that adenovirus-mediated in vivo transfer of the PDX-1 transgene to the mouse liver results in the conversion of a hepatocyte subpopulation to the beta cell phenotype. Ferber also showed that this population of trans-differentiated liver cells was induced to produce the prohormone convertases PC1/3 and PC2 leading to complete processing of proinsulin. The amount of insulin produced was sufficient to ameliorate streptozotocin-induced hyperglycaemia in the mice. For this gene therapy approach to be viable, the insulin producing liver cells must not be destroyed by the immune system or else this could lead to liver damage. This fundamental question has not been addressed by previous studies with (pro)insulin producing hepatocytes as diabetes in these models was induced chemically rather than by autoimmune means [ 6 - 9 ]. The autoimmune destruction of insulin-producing hepatocytes is a possibility since hepatocytes express major histocompatibility complex (MHC) class I molecules [ 10 ], and constitutively express Fas [ 11 ], moreover the liver cells have antigen presenting activity [ 12 ] and are attacked in autoimmune diseases such as chronic active hepatitis. Furthermore (pro)insulin appears to be an autoantigen in the development of type 1 diabetes [ 13 , 14 ]. Evidence for this comes from the presence insulin and Glutamic Acid Decarboxylase (GAD) autoantibodies in the sera of people with prediabetes [ 13 ]. It also comes from studies that show that diabetes can be adoptively transferred to normoglycaemic mice by the introduction of insulin-specific T cell clones [ 14 ]. These data taken together suggest that hepatocytes which produce (pro)insulin might be targeted and destroyed by the same autoimmune processes responsible for the destruction of pancreatic β cells. To address this question in vivo , transgenic Non-Obese Diabetic (NOD) mice that express insulin in the liver were created using the Phosphoenolpyruvate Carboxykinase (PEPCK) promoter [ 15 ] to direct expression of the mouse insulin I gene. This promoter had previously been used to create a transgenic C57BL/6 human insulin secreting mouse line with hepatic insulin expression [ 7 ]. However these mice do not develop autoimmune diabetes and would be unsatisfactory to address the issue of insulin autoantigenicity. The liver of the transgenic PEPCK-Ins NOD mice were characterised with respect to insulin mRNA transcription, (pro)insulin content and (pro)insulin release. The blood sugar levels of the animals were monitored and the livers of the animals were analysed for any evidence of immune cell infiltration. Methods Materials The PEPCK promoter was generously donated by Dr R. W. Hanson (Case Western Reserve University, Cleveland, Ohio, USA). DNA modifying enzymes and competent bacteria for transformation were purchased from Promega (Madison, Wisconsin, USA). RPMI 1640, α-MEM and fetal bovine serum (FBS) were purchased from Trace Biosciences (Castle Hill, Sydney, Australia). Hybond-N+ nylon membrane for the Southern Blot and ribonuclease protection assay (RPA) was from Amersham International (Bucks, UK). Human insulin and rat insulin standards for the in-house radioimmunoassay (RIA) were obtained from Novo Nordisk (Bagsvaerd, Denmark). Transient transfection of HEPG2 cells The transfection of the HEP G2 cells with the PEPCK-Ins transgene was carried out using Lipofectamine (GibcoBRL, Madison, Wisconsin, USA). Cells were trypsin harvested and seeded 24 hr prior to transfection at a density of 200,000 cells/well in 24 well tissue culture plates (Becton Dickinson, New Jersey, USA) to achieve 60–80% confluence at the time of transfection. 10μl of Lipofectamine was added to 300μl of serum free media and incubated at room temperature for 30 minutes. After the incubation, 300μl of DNA solution (3–4μg) was added and the lipid:DNA complex solution incubated at room temperature for 15 minutes. The cells were washed once with 1 ml of serum free medium prior to the transfer of 300μl of the lipid:DNA complex solutions to the cells. The cells were then incubated for 5 h at 37°C in an air incubator. The DNA/lipid complex media was removed and replaced with 2 mL of α-MEM 10% FBS which contains Antimycotic and Antibiotic solution (10μL/mL) (Sigma, St Louis, Missouri, USA). In addition, for the PEPCK-Ins transgene, a final concentration of 100 nM glucagon was added to the media to upregulate the PEPCK promoter. The cells were incubated at 37°C in an air incubator for 48–60 hr. After the incubation, the conditioned culture medium was collected and assayed for mouse (pro)insulin. Insulin Radioimmunoassay (RIA) (Pro)insulin was measured by an in-house RIA using rat insulin standard (Novo Nordisk Laboratories, Bagsvaerd, Denmark). The lowest value on the rat insulin standard curve used in the assay was 0.25 ng/ml. The intra and interassay coefficients of variance were <5% and 10% respectively. Generation of transgenic mice To direct the expression of the mouse insulin I gene to the liver of the transgenic mice, the 5' flanking sequences (-460 bp to +73 bp) of the rat PEPCK gene [ 15 ] were used to drive the expression of the genomic mouse insulin I gene (-25 bp to +557 bp) [ 16 ]. The construct was digested with Xba I and Pst I (Fig 1A ) and the PEPCK-Ins transgene was isolated from the agarose gel using the Supelco GenElute Spin Column (Sigma, St Louis, Missouri, USA). Standard procedures were followed to generate transgenic mice [ 17 ]. Fertilized mouse eggs were flushed from the oviducts of superovulated NOD mice 6–8 h after ovulation. Male pronuclei of the fertilized eggs were injected with 2pl of the DNA solution (2 ng/μl) and the embryos that divide to the 2 cell stage were implanted in the oviducts of pseudopregnant mice. After birth the animals were tested for the presence of the transgene by PCR and Southern blot of the DNA from tail tip samples taken after weaning at 3 weeks of age. Figure 1 A-C . The PEPCK-Ins transgene. Schematic representation ( A ). Example of PCR screening of PEPCK-Ins transgenic mice ( B ). Lanes 1–5 and lanes 10–14 are PCR reactions with genomic DNA extracted from mice that developed from microinjected mouse eggs. Primers specific for the PEPCK-Ins transgene were used for lanes 1–9 (30 cycles) while primers specific for mouse insulin 1 were used for lanes 10–18 (30 cycles). Lanes 6 and 15 are PCR negative controls (no DNA added). Lanes 7 and 16 are PCR reactions to demonstrate the specificity of the primers used (wild type NOD mouse DNA added). Lanes 8, 9, 17 and 18 are half copy spiked and plasmid controls respectively (positive controls). Example of Southern Blotting for the PEPCK-Ins transgene ( C ). Lane 1 One copy spiked sample (100 ng Balb/c genomic DNA + 102fg PEPCK-Ins plasmid), lane 2 genomic DNA from tail tip of F2 PEPCK-Ins mouse, lanes 3 and 4 are DNA from pups which died before weaning and lanes 5–7 are DNA from still born pups. Each lane was loaded with 15μg of genomic DNA that was digested with Xba I and Pst I to release the transgene. Treatment of the animals The mice were fed ad libitum with a standard diet and kept under a light-dark cycle of 12 h in compliance with the animal ethics of our institution. The facilities used to breed and maintain the mice were specific pathogen free, with air passaged through a HEPA filter. The transgenic PEPCK-Ins NOD mice and the wild type littermates were monitored for natural development of diabetes. The diabetic PEPCK-Ins transgenic NOD mice were maintained by daily insulin injections and sacrificed 4 weeks after they became diabetic. Blood glucose measurements of mice Blood glucose levels of the wild type as well as the transgenic NOD mice were performed using the Medisense Precision QID Blood Glucose Sensor System (Bedford, Massachusetts, USA). The mice were bled by pricking the tail vein and 5μl of the blood placed on to the Precision Plus blood glucose electrode. A mouse was classified as being diabetic if a reading of ≥ 15 mmol/l was obtained on more than one occasion. Insulin mRNA analysis The total RNA from liver and pancreas was isolated using TRIzol (Life Technologies, Grand Island, New York, USA) [ 18 ]. The detection of mouse insulin I mRNA by RT-PCR was performed using Superscript II RNase H - Reverse Transcriptase from GibcoBRL (Grand Island, New York, USA) to synthesize the first strand cDNA as recommended by the manufacturer. The cDNA sample was then used as a template in PCR amplification (30 cycles) using primers specific for mouse insulin I mRNA. The sequence of primers used were: (Forward) TAA CCC CCA GCC CTT AGT GAC CAG CTA TAA (Reverse) AAA GTT TTA TTC ATT GCA GAG GGG TGG GGC The PCR products were run on a 1.2% Tris Acetate EDTA agarose gel, photographed and analysed using the Gel Doc 1000 system (Biorad, Hercules, California, USA). Insulin mRNA in the samples was also visualised using the ribonuclease protection assay (RPA) method. The plasmid used as the template to transcribe the riboprobe for mouse insulin I was a generous gift from Chirgwin J. M. (San Antonio, Texas, USA). The Digoxigenin (DIG) labelled insulin riboprobe was synthesized by the in vitro transcription method using the DIG RNA labeling kit (SP6/T7) from Boehringer Mannheim (Mannheim, Germany), which utilises DIG-dUTP, according to the manufacturer's recommendations. Hybridisation and RNase treatment were performed using the RPA II kit from Ambion according to the manufacturer's recommendations. Next the samples were loaded and run on a 6% polyacrylamide/8 M urea denaturing mini-gel in Tris borate EDTA buffer. The RNA bands on the gel were transferred onto a nylon membrane by using an electro-gel transfer apparatus and the membrane was fixed by using an UV cross linker (Hybaid, UK). For the visualisation of the RNA bands on the nylon membrane, the DIG Wash and Block Buffer Set and the DIG Chemiluminescent Detection Kit, both of which were purchased from Boehringer Mannheim, were used. The membrane was then exposed to a standard Kodak X-ray film for 30 minutes. The film was subsequently analysed using the Gel Doc 1000 system. Organ culture Pancreatic and hepatic tissue organ cultures were performed using organ culture dishes from Becton Dickinson (New Jersey, USA) [ 19 ]. Briefly, the organs were taken from mice, and an aliquot of 10–30 mg removed from each organ. The tissue was then diced into 1 mm 3 explants and spread on a filter paper placed on a sterile wire grid in the inner well of the organ culture dish such that the tissue was exposed to air above and the RPMI 1640 10% FBS medium below. Sterile saline was placed in the outer well to maintain a humidified environment. The tissue in the organ culture dishes were incubated for 24 hr in a humidified 5% CO 2 and air incubator at 37°C. Culture supernatant from all media changes of respective wells were pooled separately and kept at -20°C until the RIA was performed. The results were expressed per mg of tissue. (Pro)insulin content (Pro)insulin was extracted by homogenising a weighed sample of liver or pancreas in acid-ethanol (solution of 0.18 mol/l HCl in 70% ethanol) and incubating overnight at 4°C. The next day, the samples were spun down and the supernatant removed. The (pro)insulin content was determined by RIA and the results expressed per mg tissue. In situ hybridisation of (pro)insulin mRNA To detect the transcription expression of (pro)insulin, in situ hybridisation (ISH) was performed on pancreatic and liver tissues of diabetic transgenic PEPCK-Ins NOD mice and diabetic wild type NOD mice using a modification of a previously described method (20). Briefly, 4μm thick paraffin sections were deparaffinized with xylene, ethanol and air dried. The sections in 0.01 mol/l citric buffer, pH 6.0 were then treated with microwave irradiation for 10 min. Following this procedure the sections were treated with proteinase K (Invitrogen, Carlsbad, CA, USA) 1μg/ml for 15 min at 37°C. They were washed 3 times with phosphate buffered saline for 5 min and dried with ethanol. The sections were then hybridized with [35]-S-labelled riboprobes (1μg/ml) overnight at 57°C. The [35]-S-labelled riboprobes were synthesized by in vitro transcription (SP6/T7) incorporation of [35]-S-dUTP using RNA polymerase (Roche, Mannheim, Germany). The template used was mouse insulin I cDNA that was generously donated by Chirgwin JM (San Antonio, Texas, USA). The hybridisation buffer consisted of a 25μl hybridisation cocktail (labelled riboprobe 500,000 cpm/section, 50% formamide, 0.1% SDS, 0.1% sodium thiosulfate, 0.1 mol/l DTT, 0.3 mol/l NaCl, 20 mmol/l Tris-HCl [pH 7.5], 2 mmol/l EDTA, 20% dextran sulphate, 0.02% sheared salmon sperm DNA, 0.1% total yeast RNA, 0.02% yeast tRNA, 1 × Denhardt's solution). After hybridisation, the slides were soaked in 2X standard saline citrate (SSC), rinsed with deionised water and then treated with RNase A (20μg/ml) for 30 min at room temperature. High stringency washes were then performed on the slides with 2XSSC followed by 0.2XSSC. The slides were then dried by ethanol, air dried for 20 min and the results viewed by NBT2 emulsion autoradiography (Kodak, Rochester, New York). For controls, sense-probe instead of anti-sense was used for hybridisation. The photos were taken using the Olympus IX70 microscope (Melville, New York) with a dark field condenser. Histochemistry Sections of the liver and pancreas from the transgenic mice were washed twice in Phosphate Buffered Saline (PBS), fixed in 10% formalin overnight and embedded into paraffin. Consecutive 5μm sections were cut and placed on poly-L-lysine-coated slides. After dewaxing and serial alcohol rehydration, the tissue sections were treated with H 2 O 2 , and then incubated in PBS containing 10% goat serum or 1% BSA for 20 minutes at room temperature to block nonspecific binding. For insulin staining, primary insulin antibody (Dako Corporation, Via Real, Carpinteria, California, USA) at 1:1250 was added and the sections were incubated overnight. The next morning the sections were washed in PBS and incubated with anti-guinea pig IgG (1:400) at room temperature for 20 minutes. Thereafter the sections were washed again with PBS, and biotinylated anti-rabbit / mouse antibody added to the sections for 15 minutes followed by streptavidin-peroxidase conjugate for a further 15 minutes. Finally, the sections were treated with substrate-chromogen AEC. A standard concentration (0.1%) of haematoxylin was added as a counterstain. The primary antibody was omitted for the negative control. The haematoxylin and eosin staining of sections was performed automatically by the Jung Autostainer XL machine (Leica, New Jersey, USA). Statistical analysis The log rank-test was used to determine whether the occurrence of diabetes in the transgenic mice was significantly different from the wild type NOD mice. The non-parametric t-test (Mann-Whitney U test) was used to determine whether the blood glucose levels of the diabetic transgenic and wild type NOD mice were significantly different. Results Transient transfection of HEP G2 with PEPCK-Ins To confirm that the PEPCK-Ins transgene (Figure 1A ) was functional, 3–4μg of the transgene was transfected into the human hepatoma cell line HEP G2. After the transfection, the cells were incubated in α-MEM supplemented with 10% FBS. Glucagon was added to upregulate the PEPCK promoter expression at a final concentration of 100 nmol/l for 24 hr. At the end of the incubation period, the conditioned culture medium was positive for rodent insulin at a concentration of 0.90 ± 0.04 ng/200,000 HEP G2 cells (n = 8). Generation of transgenic mice From a series of two microinjections, 43 mice were obtained of which 4 were positive for the PEPCK-Ins transgene when analysed by PCR (Figure 1B ) and Southern blot (Figure 1C ). Two of the 4 transgenic founder mice were found to have established germline transmission of the transgene. From the Southern blot analysis, these mice have only a single copy of the transgene (Figure 1C ). Transcription of insulin I mRNA in the liver Of the two lines only one was found to transcribe insulin mRNA in the liver, as determined by mouse insulin I RT-PCR (Figure 2A ) and mouse insulin I Ribonuclease Protection Assay (RPA) (Figure 2B ). No insulin mRNA was detected in the second line despite documented germline transmission of the transgene. The transgenic mice appeared healthy and active, and were normoglycaemic. The blood glucose levels were 6.0 ± 1.1 mmol/l (n = 20) as compared to 6.5 ± 0.5 (n = 20) for wild type littermates. Figure 2 A-B . Mouse insulin 1 mRNA. RT-PCR ( A ). Lanes 1A-6A are RT-PCR reactions using primers specific for mouse insulin 1 mRNA (30 cycles). Lanes 1B-6B are RT-PCR reactions using primers specific for mouse GAPDH mRNA (30 cycles). Lanes 1–4 are total RNA from the liver of progeny from founder T645-18, lane 5 is total RNA from the liver of progeny from founder T647-2 and lane 6 total RNA from the pancreas of progeny from founder T645-18. RPA for mouse insulin 1 mRNA ( B ). Lane 1 pancreatic total RNA, lane 2 liver total RNA from T647-2 progeny and lanes 3–7 liver total RNA from T645-18 progeny. Production of (pro)insulin from liver cells The ability of the liver cells from PEPCK-Ins positive transgenic mice to produce (pro)insulin was examined on liver removed from mice that had been sacrificed. Ideally production of (pro)insulin from the liver would be sought in vivo . This was not possible since any (pro)insulin released from the liver would be indistinguishable from that secreted from pancreatic β cells. Two methods were used to analyse if (pro)insulin was produced from liver tissue removed from the mice. The liver was both extracted for its hormonal content, and placed in organ culture with analysis of the conditioned culture medium for (pro)insulin. In both cases, (pro)insulin was measurable. The (pro)insulin content was 0.5 ± 1 ng/mg (n = 5), which is equivalent to 8% of the (pro)insulin content of the pancreas (Table 1 ) on a weight to weight basis. Organ culture of explants in the presence of 100 nM glucagon resulted in release of 3.4 ± 1 ng/mg (pro)insulin per day, which is 23% of that produced by cultured pancreatic explants. No (pro)insulin could be found either in liver extracts or from cultured liver explants taken from wild type NOD mice. These results show that there was translation of (pro)insulin from preproinsulin mRNA in the liver of transgenic mice. The amount of (pro)insulin present in the liver cells was too small for it to be detected immunohistochemically (data not shown). Table 1 (Pro)insulin content of and release from pancreatic and hepatic tissue of transgenic and wild type NOD mice. PANCREAS LIVER F2 progeny (Pro)insulin content (ng/mg) (Pro)insulin secretion (ng/mg tissue/ day) (Pro)insulin content (ng/mg) (Pro)insulin release in the presence of 100 nmol/l glucagon (ng/mg tissue/ day) Normoglycaemic transgenic (n = 5) 6.3 ± 0.8 15.8 ± 1.6 0.5 ± 0.1 3.4 ± 1.0 Normoglycaemic wild type (n = 3) 6.3 ± 1.0 15.2 ± 2.3 ND ND Diabetic transgenic (n = 5) 0.5 ± 0.1 (n = 3; 2 ND) ND (n = 3) 0.6 ± 0.2 (n = 5) 3.6 ± 0.3 (n = 3) ND – not detectable (<0.025 ng/mg tissue) Data expressed as X ± SD. In situ hybridisation of (pro)insulin mRNA In situ hybridisation confirmed the presence of (pro)insulin mRNA in the pancreatic and liver sections of diabetic PEPCK-Ins transgenic mice (Figure 3A and 3C ) but not in the liver section of the wild type NOD mice (Figure 3E ). The (pro)insulin mRNA was localised to the hepatocytes of the diabetic PEPCK-Ins transgenic mice which appear to be evenly distributed throughout the liver section (Figure 3C ). The sense control sections were negative (Figure 3B,3D and 3E ). Figure 3 A-F . In-situ hybridisation. Pancreatic islets from a diabetic transgenic mouse illustrating insulitis, ( A ) antisense and ( B ) sense. Transgenic liver from a diabetic transgenic mouse ( C ) antisense and ( D ) sense. Wild type liver from a diabetic NOD mouse ( E ) antisense and ( F ) sense. Occurrence of diabetes in transgenic mice One out of 14 males (7%) and four out of 25 females (16%) of the F1 and F2 progeny of the PEPCK-Ins transgenic NOD mice became diabetic (Figure 4 ). The age at which this occurred was 20–25 weeks. Among the wild type NOD mice in our colony, 5 males out of 45 (11%) and 22 out of 51 females (43%) became diabetic (Figure 4 ), at 19–28 weeks of age. The mice were followed for a total of 15 months but no other transgenic mice became diabetic while 55% of the NOD females became diabetic. The incidence of diabetes in the transgenic PEPCK-Ins NOD females (16% diabetic at 30 weeks) differs significantly from the incidence of diabetes in the wild-type NOD females (43% diabetic at 30 weeks) as determined by the log-rank test (chi squared value 16.1, P < 0.001). The overall incidence of diabetes in the transgenic PEPCK-Ins NOD mice (13%) was also significantly lower compared to the wild type (28%) NOD mice as determined by the log-rank test (chi squared value 18.7, P < 0.001). The blood glucose levels of the diabetic transgenic NOD PEPCK-Ins mice (24.8 ± 1.9 mmol/l) were also significantly lower than those of the diabetic wild type NOD mice (>33 ± 2.1 mmol/l) as determined by the Mann Whitney U test, U = 32 and P = 0.004. Figure 4 Incidence of diabetes in F1 and F2 PEPCK-Ins transgenic NOD mice versus wild type F1 and F2 NOD mice. PEPCK-Ins NOD males total n = 14, PEPCK-Ins NOD females n = 25, wild type NOD males n = 45 and wild type NOD females n = 51. The livers of the diabetic transgenic mice were examined histologically to determine if there was cellular infiltration and destruction of hepatocytes. This was not so, the liver having a normal architecture with no evidence of necrosis (Figure 5 ). In contrast, there was a cellular infiltrate in the islets of the diabetic (Figure 6A ) but not normoglycaemic transgenic mice (Figure 6B ) Immunohistochemical staining for (pro)insulin showed a reduced number of β cells (Figure 6C ) compared to normoglycaemic littermates (Figure 6D ). Figure 5 Staining of the liver of diabetic transgenic PEPCK-Ins mice with haematoxylin and eosin. (Black Bar = 20μm). Figure 6 A-D . Haematoxylin and eosin, and insulin staining of pancreatic sections from transgenic PEPCK-Ins NOD mice. H & E staining of a pancreatic islet from a diabetic transgenic mouse illustrating insulitis ( A ), and from a normoglycaemic transgenic mouse ( B ). Insulin staining of a pancreatic islet from a diabetic transgenic mouse showing few remaining β cells ( C ) and from a normoglycaemic transgenic mice ( D ). Black Bar = 20μm for A-C, and 40μm for D. Next, the (pro)insulin content of the liver of the diabetic mice was determined. This was similar to that from the liver of normoglycaemic transgenic mice (Table 1 ). (Pro)insulin release from cultures of liver cells was measured, and again found to be similar to that from the liver cells of non-diabetic transgenic mice (Table 1 ). However, as expected, the (pro)insulin content of the pancreas of diabetic transgenic mice was lower than that of normoglycaemic transgenic or wild type mice (Table 1 ). Indeed, the level was so low as to be comparable to that seen in the liver of either normoglycaemic or diabetic transgenic mice. The amount of (pro)insulin released from pancreatic explants from the diabetic transgenic mice was too low to be detected. Discussion At the end of the study, 5 heterozygous transgenic PEPCK-Ins NOD mice had become diabetic. Nevertheless, preproinsulin mRNA was localised to the hepatocytes by in situ hybridisation and was detected in total RNA extracts of the livers of these transgenic mice by both RT-PCR and RPA. (Pro)insulin was detected in acid ethanol extracts of the livers from the transgenic mice by RIA. Furthermore, (pro)insulin was released from explants of transgenic liver placed in organ culture. These results indicate that the liver cells from these transgenic PEPCK-Ins NOD mice did synthesise and produce (pro)insulin. Histological examination of the livers of the diabetic transgenic mice showed no infiltration of the liver by immune cells even though insulitis was observed in the pancreatic sections. This indicates that hepatic insulin production did not cause the development of tolerance to insulin in these PEPCK-Ins transgenic NOD mice. The lack of infiltration of immune cells in the livers of the diabetic transgenic mice also suggests that the mouse insulin I mRNA transcribing hepatocytes were not targeted or destroyed by the immune system. In addition, as the transgenic mice were sacrificed 4 weeks after the onset of diabetes, this reduces the possibility of delayed destruction of the (pro)insulin expressing hepatocytes. During this time, hepatic insulin expression was at least equal to that found in the pancreas. At the apparently low levels of transgene expression, these results suggest that (pro)insulin is not targeted by the immune system in this transgenic PEPCK-Ins mouse model superimposed on the autoimmune diabetes model of the NOD mouse. Likewise, Lipes et al reported that when pituitary cells in transgenic NOD mice were made to produce insulin, the cells were not targeted by the immune system even though the pancreatic β cells were. No cellular infiltrate also was shown when the pituitary cells that produced (pro)insulin were taken outside the blood brain barrier and transplanted beneath the renal capsule of NOD mice [ 21 ]. Our data and those of Lipes et al suggest that autoimmune destruction of (pro)insulin producing cells in the NOD mouse is specific to the islets. The blood glucose levels of the diabetic transgenic NOD PEPCK-Ins mice (24.8 ± 1.9 mmol/l) were significantly lower (P = 0.004) than those of the diabetic wild type NOD mice (>33 ± 2.1 mmol/l). This is consistent with some (pro)insulin being released from the liver. These results are in agreement with those of Valera et al in which the majority of their high copy number PEPCK human insulin C57BL/6 mice also became diabetic when injected with streptozotocin despite the constitutive release of human (pro)insulin from the livers of the mice (7). Similarly, the transgenic mice which Valera et al made diabetic by streptozotocin injections also had lower blood glucose levels compared to the diabetic wild type NOD mice suggesting that the constitutive hepatic insulin release lowered the blood glucose levels of the diabetic mice but not to normal. This lack of therapeutic effect could be due to the fact that the PEPCK promoter is induced by glucagon and cyclic AMP. The high glucose levels in the diabetic transgenic mice might have resulted in the inhibition of endogenous glucagon release and thus shutting down (pro)insulin production in the liver. However it is possible that in some of our transgenic mice, enough (pro)insulin was produced to prevent diabetes. This would explain the significantly lower incidence of diabetes in the transgenic PEPCK-Ins (13%) compared to the wild type (28%) NOD mice (P < 0.001). One possibility is a protective effect of insulin, whether by allowing exhausted beta cells to rest or by altering the makeup of the T cells in the pancreas destroying the β cells there. The first reason was the basis for the large North American trial in pre-diabetic people in the late 90's with small doses of parenteral insulin [ 22 ]. The second reason was the basis for the large North American trial in pre-diabetic people in the very late 90's with oral insulin [ 23 ]. It is possible that a critical amount of (pro)insulin needs to be produced for an autoimmune effect to be observed. The (pro)insulin content of the transgenic liver cells in normoglycaemic mice was much lower than that of a pancreatic β cell and was also lower than that in the liver of PEPCK-human insulin C57BL/6 transgenic mice produced by Valera et al [ 7 ]. (Pro)insulin was not detected immunohistochemically in the liver of our PEPCK-Ins transgenic mice, whereas it was in the liver of Valera's transgenic mice. Attempts by us to increase production of (pro)insulin by producing homozygous mice failed, with all mice being stillborn (Figure 1C ). This might be because of transient upregulation of the PEPCK gene at birth [ 24 ], resulting in increased production of (pro)insulin that caused hypoglycaemia. Alternatively, it could be due to a transgene integration effect where the transgene disrupted a vital gene; with a double copy deletion being lethal. Another possible explanation for the lack of an autoimmune effect is that presentation of insulin peptides by hepatocytes might be limited by the ability of the liver cells to cleave proinsulin. The enzymes responsible for this in the β cell, prohormone convertases (PC) 1/3 and PC2, are absent from normal liver cells, but the endopeptidase furin is present [ 25 ]. Cleavage of proinsulin by furin however would require genetic modification of the peptide [ 8 ]. In summary, we have shown that transgenic NOD mice that produce (pro)insulin in their liver do not develop cellular infiltration of their liver when autoimmune destruction of pancreatic β cells occur. Furthermore the expression of (pro)insulin in hepatocytes is insufficient to prevent development of diabetes in NOD mice. These results offer hope that eventually liver cells, or a subpopulation of them, may be of value as a therapy for type 1 diabetes.
/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544947.xml