id stringlengths 3 9 | source stringclasses 1 value | version stringclasses 1 value | text stringlengths 1.54k 298k | added stringdate 1993-11-25 05:05:38 2024-09-20 15:30:25 | created stringdate 1-01-01 00:00:00 2024-07-31 00:00:00 | metadata dict |
|---|---|---|---|---|---|---|
55938976 | pes2o/s2orc | v3-fos-license | A Linear Interpolation-Based Algorithm for Path Planning and Replanning on Girds
Field D* algorithm is widely used in mobile robot navigation since it can plan and replan any-angle paths through nonuniform cost grids. However, it still suffers from inefficiency and sub-optimality. In this article, a new linear interpolation-based planning and replanning algorithm, Update-Reducing Field D*, is proposed. It employs different approaches during initial planning and replanning respectively in order to reduce the number of updates of the rhs-values of vertices. Experiments have shown that Update-Reducing Field D* runs faster than Field D* and returns smoother and lower-cost paths.
Introduction
In mobile robot navigation, path planning leads a robot from its initial location to some desired goal location.The two most popular techniques for path planning are deterministic algorithms and randomized algorithms [1].Among deterministic algorithms, A* provides heuristic search in static, known environments [2].LPA* combines heuristic search and incremental search [3].D* Lite could replan in unknown environments efficiently [4].When provided with a grid-based representation of environments, these algorithms are limited by the discrete set of possible headings between gird cells.For example, the eight-connected grids restrict the agent's heading changes by multiples of π 4 .As a result, the paths are suboptimal and unrealistic looking.To alleviate this problem, several methods for any-angle path planning have been investigated.A* PS uses post-smoothing to generate any-angle path [5].But it does not always work as is showed in Figure 1 (The path returned by A* PS (in black line) is not the optimal path (in dash line) from S to G). Basic Theta* algorithm and AP Theta* algorithm (Angle-Propagation Theta*) allow the parent of a node to be a node other than its local neighbor [6].AA* (Accelerated A*) can plan a shortest any angle paths fast [7].However, these algorithm are only usable for uniform cost environments.
Based on D* Lite and linear interpolation, Field D* could fast plan and replan any-angle path in grid environments, whatever the environment costs are known or partially-known, uniform or non-uniform [8].Field D* is employed as the path replanner in a wide range of fielded robotic systems.However, Field D* still suffers from two major drawbacks: 1) It plans and replans much slower than D* Lite.
2) The path returned by Field D* is not always the optimal solution.Motivated by these observations, a linear interpolation-based planning and replanning algorithm, Update-Reducing Field D* (URFD*), is proposed in this paper.It reduces the number of updates of the rhs-values so as to speed up the search.It also employs a method of post-smoothing to generate a lower-cost and smoother path.Besides, a heuristic with a variable factor according to the environments is used.As a result, the novel algorithm could efficiently produce a near-optimal path in non-uniform cost grids.
The Idea of Field D*
Field D* stores the rhs-value, a one-step look ahead estimate of the goal distance (by the goal distance of a vertex we mean the cost of an optimal path from this vertex to the goal).For vertex s it satisfies: where s goal is the goal vertex.denotes the set of all neighboring vertices of s.
is the cost of a path between s and s.In classical grid-based methods, it is assumed that s and s are two corner vertices and the path between s and s is a straight line.Field D* relaxes this assumption and takes any point along the boundary of a cell into consideration.To make it possible, Field D* makes an approximation that the path cost of any point s y residing on the edge between two consecutive corner vertices s 1 and s 2 is a linear combination of where y is the distance from s 1 to s y (assuming unit cells).
With the form of the optimal path in a unit cell, Field D* could compute and find the point to move by making where v is the variable on which the path cost depends.
Differences and Inefficiency
A linear interpolation-based replanner performs differently from a classical path replanner such as D* Lite, which leads to its inefficiency.To explain this, we define g s as the cost of the optimal path from vertex s to the goal with respect to the linear interpolation assumption (so it is slightly different from the cost of the actual optimal path).We call g s (or A linear interpolation-based replanner expands vertices in a different way from D* Lite.With a consistent heuristic, a locally overconsistent vertex (whose g-value is lager than rhs-value) becomes locally consistent (the g-value equals rhs-value) after selected for expansion and then remains locally consistent until edge cost changes are detected in D* Lite.It implies that D* Lite expands any locally overconsistent vertex at most once.However, a linear interpolation-based replanner tends to expand a locally overconsistent vertex for many times.Besides, the key values (denoting the priorities of vertices in the priority queue) of the vertices selected for expansion are monotonically nondecreasing over time in D* Lite, while it is not naturally the case in a linear interpolation-based replanner.This can be explained as follows: For some vertex s, the computation of its rhs-value is based on the g-value of one neighboring vertex in D* Lite, but the g-values of two neighboring vertices in a linear interpolation-based replanner.During the planning and replanning process, it is common that at least one of the two neighboring vertices has an inaccurate rhs-value.Relied upon them, s will get an inaccurate rhs-value.And the vertices relied upon s will also be affected, and so on.It is the reason that a linear interpolation-based replanner updates the rhs-values of vertices repeatedly to their final results.Such a phenomenon is easily observed particularly in environment consisting mainly of free space since the g-value of a vertex is very close to those of its neighbors in free space.
After the cost field created, path extraction for linear interpolation-based replanners is also different from that for D* Lite.When backpointers are not recorded, D* Lite can trace back a lowest-cost path from s start to s goal by always moving from the current vertex s, starting at s start , to any neighbor s that minimizes , c s s g s until s goal is reached.So the optimality of the result depends on the accuracy of g s .But in a linear interpolationbased replanner the g-values are not the goal distances exactly, resulting in the sub-optimality of paths.
From the discussion above, we can see that there exist two major drawbacks of Field D*: 1) It plans and replans much slower than D* Lite, especially in the environments consisting mainly of free space.2) The solution path is not the optimal solution.Figure 2 shows the second problem.The path returned by Field D* has unnecessary heading changes even if no obstacle exists (see Figure 2(a)).To extract a smoother path, [9] gives a gradient interpolation method.The result is showed in Fig- ure 2(b), from which we can see that the unnecessary heading changes still exit.Obstacles can also make the interpolation assumption break down so that affects the quality of extracted paths.To alleviate it, [8] uses a onestep look ahead mechanism.But this method checks very limited steps so that cannot avoid generating a pathologic path between vertices a and b in Figure 2(c).So when it is a locally overconsistent vertex, we can use the rhs-value instead of the g-value to make the computation more close to the g * -value.When it is a locally consistent vertex, we can also use the rhs-value because it equals the g-value and thus could get a result at least no poorer than that computed by the g-value.
Update-Reducing
During replanning, if we only encounter cell cost decreases the approach above is still useful.However, when locally underconsistent vertices (whose g-values are smaller than rhs-values) appear, this approach tends to make the algorithm less efficient and even incomplete.It could be explained as follows: When the rhs-value of a vertex becomes larger due to edge cost increases, the old rhs-value is out of date and thus to be abandoned.However, the algorithm does not distinguish between the old and the new so that it is possible for the old rhs-value to be used to compute the rhs-value of another vertex, resulting in "false" relation between these two vertices.is infinite so that does not affect the key value), which possibly makes a expand while b can never be expanded again.Thus the "false" relation has no chance to be corrected, resulting in the incompleteness of the algorithm.
In order to reduce the updates of the rhs-values during replanning, we use a technique similar to which Delayed D* used to speed up D* Lite [10], that is, delaying the processing of locally underconsistent vertices.
During path extraction, A* search, which depends on the accuracy of heuristic less heavily than greedy search does, could avoid errors caused by obstacles (see Figure 2(c)).However, based on the linear approximation, A* search still cannot ensure an optimal path even if no obstacles exist (see Figure 2(b)).And greedy search needs to be kept for checking solution paths for any loops.Note that the limitation of post-smoothing showed in Figure 1 can be overcome if it is already an any-angle path before smoothing.
Combined with the methods above, Update-Reducing Field D* (URFD*) is a modified version of Field D*.It redefines the rhs-values (denoted by rhs'-values to be distinguished from the original) as where notation follows from (1).URFD* calculates the rhs-values of vertices according to (5) during initial planning.During replanning it calculates the rhs-values according to (1), which is similar to Field D*, and delays the propagation of cost increases.It checks the consistency of a path in every path extraction and ends with a post-smoothing step.
Algorithm Description
Figure 4 shows the pseudocode of the URFD* algorithm.During initial planning, URFD* calls ComputeShortest-Path() to expand vertices.ComputeCost() calculates the rhs-value in a way similar to the interpolation-based path cost calculation in Field D*, but every vertex uses rhs-values, instead of g-values, of its neighbors.ComputeState() then computes the rhs-values according to (5) (line 15).During replanning, ComputeState() calls Compute-Cost() to compute the rhs-values according to (1) (line 16).The rhs-values of the start vertex and every vertex immediately affected by the changed edge costs are updated, but only the locally inconsistent start vertex and locally overconsistent vertices are inserted into priority queue U for expansion (lines 42 -45).Then FindRaiseS-tatesOnPath() (FRSOP) is called.FROSP checks whether locally underconsistent vertices are in the vicinity of the node.All the unprocessed locally underconsistent vertices that are adjacent to this node will be added into priority queue U (lines 31 -33).Here vicinity(s) refers to the set of all corner vertices in the vicinity of node s (s is included).When the number of nodes exceeds the given limit maxsteps, or a loop is found, which indicates a potential failure of path extraction, FRSOP stops the extraction to expand locally underconsistent vertices in priority queue U.After path extraction, the solution path is post-processed by a smoothing step.(lines 39, 49).Given two cell boundary nodes along the path, the postsmoothing replaces the solution path between these two nodes with a straight line path if the latter is less costly.It is done with cell boundary nodes along the solution path iteratively.However, some small techniques are used to avoid a large amount of computation: 1) It only performs a single iteration.2) It only smoothes the path between cell corners because necessary and sharp heading changes usually occur on them.3) Before smoothing a path between two cell corners, it checks whether the costs of all grid cells that the original path is through are all same.If they are, the original path is kept.
Experimental Results
We compared the performance of URFD*, Field D* and Delayed D*. 800 different 500 × 500 random grid environments were generated: 400 environments with uniform cost grids and 400 with non-uniform cost grids.For uniform cost grid environments, four different initial percentages of obstacle cells were selected: 10%, 20%, 30% and 40%.For non-uniform cost grid environments, we assigned each traversable cell an integer cost between 1 (free space) and 15, and four different initial percentages of free space cells were selected: 90%, 70%, 50% and 30%, while the rest of the cells each got a cost (infinity or an integer between 2 and 15) randomly with the same probability (namely 1/15).For each environment, the initial task was to plan a path from the lower left corner to a randomly selected goal on the right edge.After that, we altered the costs of cells close to the agent with probability 0.1 (1.6% of the cells in the environment were changed) and had each approach repair its solution path.We use the weighted heuristic, which is described in the previous section, for URFD* and Field D*.We selected the results in three kinds of environments (A: uniform cost grids with 10% obstacle cells.B: uniform cost grids with 30% obstacle cells.C: non-uniform cost grids with 50% free space cells) and showed them in Table 1.Four performance measures were used here: the path cost, the total number of rhs-value updates (that is, updates of the rhs-values), the total number of vertex expansions (updates of the g-values) and the runtime.Each value is a ratio of a performance measure of URFD* (or Field D*) to that of Delayed D* averaged over initial planning (or replanning) episodes.Note that in environments with more free space the runtimes of initial planning and replanning of Field D* drastically increased while those of URFD* increased much more stably.The performance in environments C shows the possibility that the number of updates of the rhs-values during replanning could be slightly larger than that of Field D* in some scenarios.However, since the number of vertices in the priority queue is limited by selectively processing locally underconsistent vertices, making the priority queue operations less expensive, the runtime of URFD* is still shorter than that of Field D* in those scenarios.
Conclusion
We present URFD*, a linear interpolation-based algorithm that plans and replans any-angle paths in dynamic environments with uniform and non-uniform cost grids.It makes efforts in the reduction of updates of the rhsvalues, which contributes to the gain in efficiency.The solution paths returned by URFD* are smooth and nearoptimal.As opposed to Field D*, it performs faster planning and replanning and returns a path with lower cost and fewer heading changes.However, URFD* is not optimal either due to the linear interpolation assumption.
Figure 2 .
Figure 2. Paths returned by interpolation methods.dependsheavily on the number of updates of the rhsvalues of vertices, the key to high efficiency of our algorithm is reducing the number of updates.We use the fringe vertices to refer to the vertices on the fringe of the expanded vertices during initial planning.The fringe vertices have not been expanded yet so that their g-values are not computed (namely infinite), leading the rhs-values computed by the g-values of the fringe vertices to inaccuracy.Then the inaccurate rhs-values along with the infinite g-values affect next vertex expansions.This is the main source of the repeated updates of rhs-values.Note that most of the fringe vertices are just locally overconsistent vertices during initial planning, as is showed in Figure3.(The locally consistent vertices (in light grey), which have been expanded at least once, are almost surrounded by the locally overconsistent vertices (in dark grey).Vertices in black are obstacles.)The rhs-values of locally overconsistent vertices are better informed and thus more accurate than the g-values.So when it is a locally overconsistent vertex, we can use the rhs-value instead of the g-value to make the computation more close to the g * -value.When it is a locally consistent vertex, we can also use the rhs-value because it equals the g-value and thus could get a result at least no poorer than that computed by the g-value.During replanning, if we only encounter cell cost decreases the approach above is still useful.However, when locally underconsistent vertices (whose g-values are smaller than rhs-values) appear, this approach tends to make the algorithm less efficient and even incomplete.It could be explained as follows: When the rhs-value of a vertex becomes larger due to edge cost increases, the old rhs-value is out of date and thus to be abandoned.However, the algorithm does not distinguish between the old and the new so that it is possible for the old rhs-value to be used to compute the rhs-value of another vertex, resulting in "false" relation between these two vertices.For example, there exist two vertices a and b.After initial planning, g(a), rhs(a) and g(b) are all infinite, rhs(b) is 50.Then rhs(a) and rhs(b) are updated due to cell cost increases.When rhs(a) is recomputed, old rhs(b) (namely For example, there exist two vertices a and b.After initial planning, g(a), rhs(a) and g(b) are all infinite, rhs(b) is 50.Then rhs(a) and rhs(b) are updated due to cell cost increases.When rhs(a) is recomputed, old rhs(b) (namely
Figure 3 .
Figure 3.A snapshot during initial planning of Field D*. 50) is used.Then rhs(b) is updated to a new value of 64.Thus, the computation of rhs(a) seems to rely on rhs(b) but in fact this relation possibly dose not exist.Furthermore the wrong rhs(a) leads to a priority error of a (g(a)is infinite so that does not affect the key value), which possibly makes a expand while b can never be expanded again.Thus the "false" relation has no chance to be corrected, resulting in the incompleteness of the algorithm.In order to reduce the updates of the rhs-values during replanning, we use a technique similar to which Delayed D* used to speed up D* Lite[10], that is, delaying the processing of locally underconsistent vertices.During path extraction, A* search, which depends on the accuracy of heuristic less heavily than greedy search does, could avoid errors caused by obstacles (see Figure2(c)).However, based on the linear approximation, A* search still cannot ensure an optimal path even if no obstacles exist (see Figure2(b)).And greedy search needs to be kept for checking solution paths for any loops.Note that the limitation of post-smoothing showed in Figure1can be overcome if it is already an any-angle path before smoothing.Combined with the methods above, Update-Reducing Field D* (URFD*) is a modified version of Field D*.It redefines the rhs-values (denoted by rhs'-values to be distinguished from the original) as | 2018-12-13T10:15:22.424Z | 2012-06-29T00:00:00.000 | {
"year": 2012,
"sha1": "23b173468b0e88d42e0c1249cbfa297d788a8d10",
"oa_license": "CCBY",
"oa_url": "http://www.scirp.org/journal/PaperDownload.aspx?paperID=20473",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "23b173468b0e88d42e0c1249cbfa297d788a8d10",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Mathematics"
]
} |
229212553 | pes2o/s2orc | v3-fos-license | Clinical pharmacology of cefuroxime and cefuroxime axetil in infants and children
Cefuroxime is a second-generation cephalosporin β -lactamase resistant active against gram-negative and gram-positive organisms and the prodrug cefuroxime axetil is the drug of choice for treatment of Lyme disease. Cefuroxime inhibits bacterial cell wall synthesis, is bactericidal, and is one of the few cephalosporins available for oral, intravenous, and intramuscular administration. Cefuroxime is rapidly absorbed when given by mouth, and its blood concentrations remarkably vary among subjects, but the drug has not adverse-effects even when is given at high doses. The gastrointestinal absorption-rate of cefuroxime axetil is 70%, and after hydrolysis it relays cefuroxime. Parenteral dosing-regimens consist of 25 mg/kg in infants and dosing intervals decrease with increasing postnatal age. Cefuroxime oral dosing- regimens are: 10, 20 mg/kg and 250 mg twice-daily in children aged up to one year, 2 to 11 and 12 to 17 years, respectively. Elimination half-live are: 6 hours in the first week of life and 1.8 hours in children aged up to 6 years. Cefuroxime is efficacy and safe for prophylaxis and treatment of respiratory-tract and acute urinary-tract infections, pneumoniae, acute otitis media, and sinusitis in children. Prophylaxis for surgical procedures requires 50 mg/kg before procedure and then 30 mg/kg for up 3 days after intervention. Cefuroxime concentration in cerebrospinal fluid is about 10% of that in plasma and the drug successfully treated meningitis caused by β -haemolytic streptococci, Streptococcus pneumoniae, and Neisseria meningitides, but when Salmonella species are the infective pathogens additional 5 mg dose is given intraventricularly. Some bacteria may become resistant to cefuroxime and antibiotic consumption is associated with increased bacterial-resistance. The aim of this study is to review the published data on cefuroxime and cefuroxime axetil dosing, efficacy, safety, and cefuroxime effects, prophylaxis, treatment, optimization, meningitis, pharmacokinetics and bacterial-resistance.
Introduction
This non-toxic broad-spectrum cefuroxime is one of the few cephalosporins that may be administered parenterally and orally. Cefuroxime is a β-lactamase resistant second-generation cephalosporin active against gram-positive organisms (including group B streptococci and penicillin-resistant staphylococci) and a wide range of gramnegative organisms. It is reasonable active against Haemophilus influenzae and Neisseria gonorrhoeae, but is inactive against Listeria, enterococci, Bacteroides, and Pseudomonas species. It has sometimes been used prophylactically in infants undergoing abdominal surgery Cefuroxime penetrates into the cerebrospinal poorly and coagulasenegative staphylococci are increasingly resistant to this antibiotic. Two-thirds of the lipophilic acetoxyethyl ester dose, cefuroxime axetil, is absorbed when is given orally, and after hydrolysis, it releases cefuroxime. There are no published reports of the use of this formulation in children aged<three months, but it has been widely used to treat otitis media and other respiratory infections in older children. It is just as active as treatment with co-amoxiclav and less likely causes troublesome loose stools. Cefuroxime is largely excreted by the kidney. Elimination half-life in plasma falls from six hours at birth to about three hours at two weeks. Infants aged > one month clear cefuroxime from their plasma almost as fast as adults (half-life one hour), but dosage intervals should be extended in infants with severe renal failure. Toxic adverse-effects are rare but oral treatment does sometimes cause nausea and vomiting and a change in stool frequency, and pseudomembranous colitis has occasionally been reported. Other problems, as with cephalosporin in general, are uncommon. Lyme disease is caused by the spirochete Borrelia burgdorferi and human infection is caused by the bite of an infected animal tick. Illness was once rare in the United Kingdom but does not seem common in Europe and North America. While a migrating annular skin lesion (erythema migrans) is the classic presentation, symptoms are very variable. Foetal infection was first recognized in 1985, and it is now realized that the risk to the foetus is comparable to that of congenital syphilis. Cefuroxime axetil is the antimicrobial agents of choice for early disease in young children. Intravenous ceftriaxone, cefotaxime, or benzylpenicillin are recommended if there are cardiac or neurological complications. The duration of treatment is usually two to four weeks [1,2].
Cefuroxime inhibits bacterial cell wall in a manner similar to that of penicillin [3]. Peptidoglycan is a heteropolymeric component of the bacterial cell wall that provides rigid mechanical stability. β-Lactam antibiotics inhibit the last step in the peptidoglycan synthesis, the cell wall is 50 to 100 molecules thick in gram-negative bacteria and it is only 1 or 2 molecules thick in gram-negative organisms. The peptidoglycan is composed by glycan chains, which are linear strands of two alternating amino sugars (N-acetylglucosamine and N-acetylmuramic) that are cross-linked by peptide chains. Peptidoglycan precursor formation takes place in the cytoplasm. The synthesis of UDP-acetylmuramylpentapeptide is completed with the addition of a dipeptide, D-alanyl-Dalanine (formed by racemization and condensation of L-alanine). UDP-
Literature search
The literature search was performed electronically using PubMed database as search engine and the cut-off point was June 2020. The following key words: "cefuroxime infants effects", "cefuroxime children effects", "cefuroxime infants metabolism", "cefuroxime children metabolism", "cefuroxime infants pharmacokinetics", "cefuroxime children pharmacokinetics", "cefuroxime infants resistance", and "cefuroxime children resistance" were used. In addition, the books Neonatal Formulary [1], The British National Formulary for Children [3], and The Pharmacological Basis of Therapeutics [4] were consulted. The manuscript is written according to "Instructions for Authors".
Cefuroxime efficacy and safety in infants and children
de Louvois [5] administered cefuroxime at a dose of 50 mg/ kg twice-daily for 5 days to 27 infected infants at birth and in early neonatal life. There was a significant improvement of symptoms and infants were discharged from hospital after 5 days of treatment, and cefuroxime was efficacy, safe, and well tolerated.
Powell et al. [6] assessed the efficacy, safety and tolerability of cefuroxime axetil suspension in 36 infants and children, aged 3 months to 12 years, and doses were: 10, 15, or 20 mg/kg twice-daily or thricedaily. Three of 36 patients (8.3%) were withdrawn from therapy because of adverse-effects. Of 32 patients who completed therapy, 9 (28.1%) developed mild-reactions including oral thrust, diarrhoea, or diaper dermatitis, and none were withdrawn from therapy. Complete clinical cure occurred in 28 subjects (87.5%), and 4 subjects (12.5%) were clinically improved but still required an additional antibiotic after one week to complete therapy, and treatment was effective, safety and well tolerated.
Reed et al. [7] studied efficacy, safety and tolerability of cefuroxime axetil suspension in 35 children, aged 3 months to 12 years. The drug was administered at doses of: 10, 15, or 20 mg/kg twice-daily or thricedaily. Three of 35 subjects (8.6%) were withdrawn from therapy because of drug-related reactions. Of 32 subjects who completed therapy, 9 (28.1%) developed mild oral thrush, diarrhoea, or diaper dermatitis; bud none were withdrawn from therapy and this treatment was efficacy, safe, and well tolerated.
Barson et al. [8] treated 36 children, aged 3.5 to 4.7 years, suffering from infections in the face, epiglottis, or buccal cellulitis who were treated with cefuroxime at a dose of 75 mg/kg daily. Blood culture was performed in 22 children (61.1%) and the infective pathogens were: Haemophilus influenzae type b, Streptococcus pneumoniae, and nontypable Haemophilus influenzae, and this treatment was efficacy and safe.
Syrogiannopoulos et al. [9] administered cefuroxime axetil at a dose of 250 mg twice-daily for 10 days to 55 children, aged 5 to 11 years, suffering from acute otitis media. A tympanogram was obtained on enrolment and 2 days after completion of therapy. Complete cure occurred in 74.1% children and clinical improvement was observed in 25.9% children. There was no clinical failures and none children experienced relapse of acute otitis media infection one month after end of therapy.
Olivier [10] observed that treatment of community-acquired pneumonia caused by Mycoplasma, Chlamydia, or Legionella species required a drug which penetrates respiratory tissues effectively such as cefuroxime, and she reviewed literature on treatment of this disease. Cefuroxime axetil has been extensively used in infants aged > 3 months, including those with pleural effusion complications or pneumonia; treatment duration was 24 to 72 hours, and therapy was successful.
Cefuroxime effects in infants and children
Burman et al. [11] observed that Enterobacter cloacae strains dominated aerobic flora in faeces of 79 of 953 infants (8.3%) discharged from neonatal intensive care units. Cefuroxime yielded slight increase of MICs of ampicillin, cephalexin, and cephalothin compered to isolates obtained from untreated infants (P-value=0.02).
Zachariassen et al. [12] investigated cefuroxime pharmacokinetics in infants born to mothers undergoing caesarean section who received cefuroxime for prophylaxis. The drug migrated into infants and its plasma concentration and elimination half-life remarkably varied in infants.
Dierikx et al. [13] assessed the effect of cefuroxime on bacterial composition in gut of infants born to mothers treated with cefuroxime at single dose of 1.5 grams during, prior caesarean section and after umbilical clamping. Cefuroxime migrated into infants and impaired gut intestinal microbiota.
Arnez et al. [14] determined efficacy and adverse-effects caused by cefuroxime axetil administered at a dose of 30 mg/kg daily to 46 children, aged<15 years, suffering from erythema migrans. Efficacy and adverse-effects were compared to those of 44 children treated with phenoxymethyl penicillin at a dose of 100,000 IU/kg daily. Both treatments were equally efficacy, whereas adverse-effect-rates were: 26.1% and 6.8% (P-value=0.0301) following cefuroxime axetil and phenoxymethyl penicillin administration, respectively.
Cefuroxime adverse-effects in children
In literature there is only one study on adverse-effects caused by cefuroxime in children and it was reported by Gold and Rodriguez [15], they stated that cefuroxime is free from adverse-effects.
Cefuroxime and cefuroxime axetil prophylaxis for urinarytract infection and surgery in infants and children
Gurevich et al. [16] observed that recurrent urinary-tract infections occurred in 19.8% infants, aged<1 month, and the first episode of infection occurred in 6 infants (5.9%); peak incidence occurred 2 to 6 months after birth. One-hundred-one infants, aged<2 months, with the first and recurrent episodes of urinary-tract infections were treated with cefuroxime axetil. Infective pathogens were: Escherichia coli, Klebsiella species, Enterococcus species, Morganella morganii, Proteus and Enterobacter species. Prophylaxis significantly reduced the frequency of urinary-tract infection episodes up to 8.3%.
Bitsori et al. [17] determined effects of prophylaxis in 638 uropathogens isolated from the urine of children. Infective uropathogens were: Escherichia coli, Klebsiella species, Pseudomonas aeruginosa, Enterococcus and Proteus species. Bacterial-resistancerate of these organisms to several antibiotics increased following their administrations, whereas that to cefuroxime decreased (P-value<0.016). Continuous surveillance, re-evaluation of empiric dosing regimens, and further assessment of prophylaxis are needed to optimize prophylaxis of urinary-tract infection.
Bi et al. [18] characterized pathogens and their antibiotic susceptibility in 895 isolated obtained from children with urinarytract infections caused by catheter indwelling in order to optimize prophylaxis. Escherichia coli was the most frequent pathogen, followed by Staphylococcus aureus, Staphylococcus epidermidis, and Enterococcus species. Gram-negative species were the predominant infective uropathogens and bacterial-susceptibility patterns must be determined in order to optimize prophylaxis.
Knoderer et al. [19] evaluated the efficacy of two prophylaxis regimens with cefuroxime in 210 children, aged ≤ 18 years, undergoing cardiovascular surgery. Prolonged prophylaxis was performed before surgery and short-term prophylaxis was performed on the first day after intervention. Short-term post-operative prophylaxis did not increased infection outcomes.
Cefuroxime treatment of infants and children
Brogden et al. [20] stated that staphylococci β-lactamase producing and gram-negative bacteria are not resistant to ceftriaxone. This antibiotic was given parenterally to infants and children and effectively cured infections caused by gram-positive and gram-negative aerobes. Cefuroxime was also efficacy to treat respiratory infections caused by Haemophilus influenzae, Streptococcus pneumoniae, Klebsiella pneumoniae cephalosporin-resistant and Enterobacter.
Nelson et al. [21] Administered cefuroxime intravenously at a dose of 25 mg/kg thrice-daily to 100 infants, with mean age of 15 months, suffering from pneumoniae caused by Streptococcus pneumoniae, Haemophilus influenzae type b, Staphylococcus aureus, Neisseria meningitides B and pneumococcus species. All microorganisms were susceptible and had an MIC ≤ 1.25 µg/ml. Mean duration treatment was 3.1 days in afebrile infants and 5.1 days until respiratory symptoms were gone. Eosinophilia occurred in only 10 infants (10.0%), and cefuroxime successfully treat pneumonia.
Pichichero [22] stated that antibiotic-resistance, caused by antecedent antibiotic treatment, is increasing in pathogens causing acute otitis media infections in infants. Amoxicillin is the first choice antibiotic to treat acute otitis media although a high dose of 80 mg/kg is needed to eradicate resistant-pathogens. Oral cefuroxime is a secondline antibiotic to treat this disease.
Shafina et al. [23] enrolled 101 infants and children, aged 1 day to 13 years, suffering from acute urinary-tract infections caused by Escherichia coli, Klebsiella and Enterococcus species which were treated with ampicillin. The resistance-rate to ampicillin was high for Escherichia coli, Klebsiella and Enterococcus whereas that to cefuroxime was low and cefuroxime is an appropriate antimicrobial agent for treatment of acute urinary-tract infections caused by these bacteria.
Wald [24] stated that Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis are the commonest organisms causing acute and subacute sinusitis in infants and children. Amoxicillin is the appropriate treatment for infections caused by pathogens producing β-lactamase. If a patient does not respond to this treatment cefuroxime-axetil should be used as it has a broad-spectrum of activity.
Cefuroxime administered to mothers before caesarean section causes alteration of microbiota composition in gut of their infants
Kamal et al. [25] assessed alteration of microbiota in infant gut caused by cefuroxime prophylaxis performed in pregnant women undergoing caesarean section. Faecal samples were collected 10 days after birth and also after 9 months. Composition of gram-positive cocci and Enterobacteriaceae was assessed in infant gut born to these mothers. No alteration of microbiota composition was observed 10 days after birth whereas bacteria composition was altered 9 months later.
Optimization of cefuroxime and cefuroxime axetil treatments in children
Carlier et al. [26] optimized treatment with cefuroxime at a dose of 750 mg thrice-daily in 20 critically ill children who had creatinine clearance<20 ml/min. Optimization of treatments were performed according to different cefuroxime dosing-regimens and also to creatinine clearance of: 50, 100, 200 and 300 ml/min. Five-hundred children were simulated by Monte Carlo method for optimizing dosing-regimen which was defined as the blood concentration > MIC for at least 65% of dosing interval (65% fT >MIC ). Probability of attainment target-rates (65% fT >MIC ) was calculated for each child and simulations were performed with the following cefuroxime MICs: 0.5, 1, 2, 4, 8 and 16 µg/ml. Simulations revealed that treatment failure occurred for organisms with an MIC=8 µg/ml when the cefuroxime was administered by intermittent infusion to children having renal clearance ≥ 50 ml/min. Continuous infusion, with higher doses than the normal ones, achieved target-rates. However, even continuous infusions of cefuroxime at a dose of ≤ 9 grams daily did not guarantee adequate blood concentrations for children with high creatinine clearance ≥ 300 ml/min when MIC was 8 µg/ml.
Ibar-Bariain et al. [27] optimized dosing-regimens of cefuroxime axetil using Monte Carlo simulations in children. Optimal treatment was defined for cefuroxime axetil plasma concentrations > MIC for at least 60% of dosing interval (fT >MIC ≥ 60%), and also for the cumulative fraction ≥ 90%. Simulations revealed that cefuroxime axetil optimal dosing-regimens were: 10 to 15 mg/kg twice-daily infused over 0.5 hours for an MIC ≤ 0.5 µg/ml. However, for MIC=1 µg/ml, only cefotaxime administered at a dose of 66 mg/kg thrice-daily ensured a probability of therapy success-rate > 90%.
Pichichero et al. [28] used Monte Carlo simulations for optimizing dosing-regimens of 10 β-lactamase resistant-antibiotics and assessed the probability of attaining pharmacodynamic target-rates in children suffering from recent respiratory infections caused by Haemophilus influenzae. Two-hundred-thirty-three isolates of Haemophilus influenzae were obtained from paediatric outpatients suffering from acute otitis media (N=55), sinusitis (N=58), and lower respiratory tract infections (N=120). Five-thousand concentration-time profiles were simulated for 10 antibiotics approved by US FDA including cefuroxime, using pharmacokinetic parameters and weights of male children aged 5 years. Optimal dosing regimen, against Haemophilus influenzae, was defined for cefuroxime unbound concentration in plasma > MIC for 50% of the dosing interval (50% fT > MIC). Sixty-seven isolates (3.3%) were β-Lactamase producing and the infection sites were: acute otitis media 38%, sinusitis 36%, and lower respiratory-tract 21%. Among 10 antibiotic tested, cefuroxime was the most active agent for prophylaxis and cure of children with acute otitis media caused by Haemophilus influenzae.
Cefuroxime axetil treatment of lyme disease in children
Lyme disease is caused by the bit of the spirochete Borrelia burgdorferi. Wormser et al. [29] reported guidelines for treatment of Lyme disease in children. Suggested treatments were cefuroxime axetil at a dose of 15 mg/kg twice-daily (maximum 500 mg) for 14 to 21 days. Recommended duration of treatments was for: erythema migrans (14 to 21 days), acute neurological disease, meningitis, radiculopathy or cranial-nerve (14 to 28 days), cardiac disease 1 st and 2 nd degree (14 to 31 days), 3 rd degree (14 to 213 days), and arthritis (14 to 28 days). Occurrence-rates of the first and multiple lesions of erythema migrans were: 98.3 and 16.8%, respectively, in subjects treated with cefuroxime axetil and the size of the primary lesion was 79+10 cm 2 [30]. Gerber et al. [31] reported the frequency of treatment for infections caused by Borrelia burgdorferi bit and was: fever 24%, fatigue 58%, headache 42%, arthralgia 33%, neck pain 26%, nausea 21%, abdominal pain 17%, diarrhoea 16%, myalgia 16%, sore throat 12%, cough and rhinorrhoea 8%, vomiting and conjunctivitis 5%. Rochette [32] measured the cerebrospinal volume in 54 preterm and term infants, with a postmenstrual age of 30 to 60 weeks, and the volume was 1.97+0.54 ml/kg. The Pearson's rank correlation test revealed that cerebrospinal fluid volume correlated with both the infant weight and the postmenstrual age (r=0.78, P-value<0.0001).
Ceftriaxone penetration into the cerebrospinal fluid of infants and children
Pfenninger et al. [33] enrolled 33 infants, aged > 3 months, suffering from meningitis caused by: Haemophilus influenzae, Neisseria meningitides, and Streptococcus pneumoniae. Cefuroxime was intravenously infused at a dose of 50 mg/kg 4 times-daily, the duration of treatment was 12 to 13 days, and cerebrospinal fluid concentration ranged from 1.1 to 18.8 µg/ml (mean, 7.0).
Edwards et al. [34] studied 7 children, aged 7 months to 9 years, suffering from meningitis caused by Staphylococcus epidermis, Staphylococcus aureus, Staphylococcus hominis, and Staphylococcus sciuri. Cefuroxime was intravenously infused at doses of 67 and 76 mg/ kg thrice-daily for 2 to 14 days; cefuroxime crossed the brain blood barrier easily and achieved concentrations from<2.0 to 22.5 µg/ml in cerebrospinal fluid. The wide interindividual variability of cefuroxime concentration reflected both the site of local inflammation and the time after dosing.
Kuzemko and Walker [35] measured cefuroxime concentration in cerebrospinal fluid of 6 children, aged 1.5 to 6 years, with meningitis caused by Neisseria meningitides (MIC=0.025 µg/ml), Haemophilus influenzae (MIC=0.5 µg/ml), and Streptococcus pneumoniae (MIC was not determined). Cefuroxime was intravenous infused at a dose of 25 mg/kg twice-daily in three children and other 3 children received this drug by lumbar puncture. Cefuroxime concentration ranged from 1.5 to 6.0 µg/ml in cerebrospinal fluid and was > MICs for Neisseria meningitides and Haemophilus influenzae indicating that cefuroxime is an appropriate antimicrobial agent to cure meningitis caused by these organisms.
Ceftriaxone treatment of bacterial meningitis in infants and children
Sirinavin et al. [36] intravenously infused cefuroxime at a dose of 50 mg/kg daily for 14 to 15 days to 7 infants and children, aged 25 days to 12 years, suffering from meningitis caused by β-haemolytic streptococci, Streptococcus pneumoniae, Neisseria meningitides, Salmonella krefeld and Salmonella typhimurium. Cerebrospinal fluid and blood specimens were sampled 2 hours after dosing after the fourth or successive treatment days. Other 6 infants were suffering from meningitis caused by Haemophilus influenzae and all were cured. Five infants of these were initially treated with cefuroxime co-administered with penicillin bud cefuroxime treatment was discontinued after two days because etiological agents were: hospital-acquired Pseudomonas cepacia and Pseudomonas aeruginosa strains which are cefuroximeresistant. Cefuroxime concentration was measured in cerebrospinal fluid and serum of other 9 subjects on the second treatment day and was: 6.4+1.7 and 35.5+9.5 µg/ml, respectively, and the ratio was 18.8+3.0%. Cefuroxime concentration was also measured in cerebrospinal fluid and plasma of 6 subjects on the 14th treatment day and was: 3.6+2.2 and 32.2+23.3 µg/ml, respectively, and the ratio was 12.1+2.5%. In children without meningitis, cefuroxime concentrations were: 1.3+1.3 and 32.5+14.7 µg/ml, in cerebrospinal fluid and serum, respectively, and the ratio was 3.8+2.4%. Cerebrospinal fluid concentration was > 10fold higher than the MICs of β-haemolytic streptococci, Streptococcus pneumoniae, Haemophilus influenzae, Streptococcus pneumoniae, and Neisseria meningitides, whereas MICs of Salmonella strains were similar to those of cefuroxime. Meningitis caused by β-haemolytic streptococci, Streptococcus pneumoniae, Neisseria meningitides and Haemophilus influenzae type b, was cured with 50 mg/kg cefuroxime given 4 timesdaily, whereas meningitis caused by Salmonella species required an additional dose of 5 mg cefuroxime injected intraventricularly for 4 days. Other ten subjects without meningitis received single cefuroxime dose of 50 mg/kg and cerebrospinal fluid concentration was 3 to 6 times lower than those measured in subjects with meningitis, thus inflamed meninges improve cefuroxime penetration-rate into cerebrospinal fluid.
Del Rio et al. [37] evaluated susceptibility of Haemophilus influenzae, Neisseria meningitides, and Streptococcus pneumoniae to cefuroxime in 38 infants and children, aged 1.3+1.3 years (range, 4 weeks to 6.5 years), suffering from meningitis. Twenty-seven subjects (71.0%) had meningitis caused by Haemophilus influenzae type b, (β-lactamase negative and β-lactamase positive), Haemophilus influenza type e β-lactamase-negative, Streptococcus pneumoniae, Neisseria meningitides, and group B Streptococcus, whereas 4 subjects were not infected. Cefuroxime plasma concentration was measured after single daily intravenous doses of 50 and 75 mg/kg of cefuroxime for a minimum of 5 days in 5 children. Table 1 shows susceptibility of bacteria to cefuroxime. Cefuroxime plasma concentration correlated with the protein content following a dose of 50 mg/kg (P-value=0.01) and 75 µg/ml (P-value=0.039), and the concentration was higher following 75 mg/kg at each time interval. Mean cefuroxime exposition in cerebrospinal fluid, expressed as the AUC in cerebrospinal fluid to that in plasma ratio, ranged from 6.4 to 10.0% following 50 and 75 mg/ kg doses, respectively. Cefuroxime concentration (µg/ml) was measured in cerebrospinal fluid on the first and 6 th doses at various times after administration and was respectively: 0.96 and 2.1 at 0.5 hours, 17.3 and 14.5 at 1 hour, 2.5 and 3.4 at 2 hours, and 2.1 and 0.94 at 4 hours. Both MICs and MBCs tested for these 5 organisms were lower than cefuroxime concentrations in cerebrospinal fluid suggesting that this drug is an appropriate antimicrobial agent to treat bacterial meningitis caused by these organisms.
Cefuroxime plasma and serum concentrations in infants
Philipson and Stiernstedt [38] measured cefuroxime plasma concentration in infants at birth, born to mothers who received cefuroxime at a dose of 750 mg, and it ranged from 1.4 to 3.6 µg/ml. Reference 5 is above reported under the heading cefuroxime efficacy and safety. de Louvois [5] measured cefuroxime serum concentration in 28 preterm and term infants with a body weight of 2.0+0.85 kg who received cefuroxime parenterally at a dose of 25 mg/kg twice-daily. Median concentrations were: 35, 45, 42, 35, 26, and 10.5 µg/ml at: 0.25, 0.5, 1, 3, 5, 12 hours after dosing, and the median elimination half-life was 6.0+2.1 hours.
Cefuroxime pharmacokinetics in infants undergoing cardio pulmonary bypass
Gertler et al. [39] studied cefuroxime pharmacokinetics in 42 infants, aged<1 year, of whom 36 had congenital heart defect and underwent cardio pulmonary bypass, whereas 6 did not. Cefuroxime single dose of 50 mg/kg was intravenously infused before cardio pulmonary bypass and 75 mg/kg was injected in the prime volume. The impact of age (Fmat) on total body clearance had a sigmoid maturation function and was calculated according to the following equation: Peak cefuroxime total and unbound plasma concentrations were: 271+62 and 229+52 µg/ml, respectively, before cardio pulmonary bypass (dose=50 mg/kg), and those after the second dose (75 mg/kg) were: 404+102 and 341+86 µg/ml, respectively. Table 2 shows cefuroxime pharmacokinetic population parameters. The peripheral distribution volume is larger than the central distribution volume indicating that cefuroxime well distributes in body tissues.
Cefuroxime pharmacokinetics in infants and children
Reference 37 is above reported under the heading cefuroxime penetration into the cerebrospinal fluid. Del Rio et al. [37] investigated cefuroxime pharmacokinetics in plasma of 5 infants and children on the first and 6th treatment days. Cefuroxime was intravenous infused at doses of 50 and 75 mg/kg. Following a dose of 50 mg/kg, plasma concentrations (µg/ml) were: 105+36 at the end of infusion and 4.4+4.5 6 hours later, and after a dose of 75 mg/kg, these concentrations were: 152+43.5 and 5.4+3.0, respectively. Table 3 summarizes pharmacokinetic parameters obtained in 5 subjects following multiple cefuroxime dosing of 50 mg/kg, on the first and the 6 th administration days. Absorption and elimination half-live, distribution volume, and the total body clearance are not statistically different according the two sampling days suggesting that cefuroxime does not accumulate in blood.
Powell et al. [40] assessed cefuroxime axetil pharmacokinetics in 28 infants and children, aged 3 months to 12 years (mean, 1.2 years) with bacterial infection, and were treated with a single cefuroxime axetil suspension at doses of 10, 15, or 20 mg/kg. Table 4 shows that cefuroxime axetil serum peak concentrations were not statistically different according to the 3 doses. In addition, Tmax, AUC, and elimination half-life are not statistically different according the dosages indicating that cefuroxime axetil pharmacokinetic parameters are not dosing-dependent.
Cefuroxime pharmacokinetics in children undergoing cardio pulmonary bypass
Knoderer et al. [41] assessed cefuroxime pharmacokinetics in 16 children, aged 1.2+0.3 years with a body-weight of 9.7+2.8 kg, requiring cardio pulmonary bypass for cardiac surgical repair. Cefuroxime was intravenously infused at a dose of 25 mg/kg one hour before skin incision and another dose of 12.5 mg/kg was injected in the cardio pulmonary bypass prime fluid. The distribution volume of the peripheral compartment was similar to that of the central compartment and this contrasts with the results obtained in infants undergoing cardio pulmonary bypass in whom peripheral distribution volume is larger than the central distribution volume [39]. Likely, total protein content is lower in plasma of infants than that of children, thus the extent of cefuroxime bound plasma proteins is lower in infants with consequent higher migration-rate into body-tissues causing a larger peripheral distribution volume.
Mechanisms bacterial-resistance to cefuroxime
Källman et al. [42] determined cefuroxime susceptibility by discdiffusion and by E-test methods. The ability of bacteria to grow in Källman et al. [43] studied whether efflux or down-regulation of porins contributes to cefuroxime resistance in Klebsiella pneumoniae. Transcription-rate of acrA, ompK35, ramA, and soxS was determined by quantitative RT-PCR. All clinical isolates had similar non-susceptibility to cefuroxime. Phenylalanine arginine β-naphthylamide did not increase susceptibility to cefuroxime. Increased acrA transcriptionrate and decreased ompK35 transcription-rate was seen in all strains. Multidrug-resistant phenotype of Klebsiella pneumoniae is associated with increased acrA and ramA gene transcription-rate and decreased ompK35 gene transcription-rate. Since cefuroxime resistance was not reversed by β-naphthylamide, this is attributable to decreased levels of OmpK35 rather than to efflux.
Gunduz and Altun [45] evaluated causative pathogens and their antibiotic-resistance in 850 positive strains isolated from the urine of children aged 3.0+3.7 years. Commonest causative pathogens were: Escherichia coli (64.2%) followed by Klebsiella pneumoniae (14.9%). Overall resistance-rate to three antibiotics ranged from 62.6 to 29.8% and that to cefuroxime was 28.7%. Escherichia coli was the commonest bacteria isolated which had the highest resistance-rate to cefuroxime. Several authors observed that antibiotic consumption including cefuroxime was associated with measurable and statistical significant increase in the resistance-rate [46][47][48].
Discussion
Cefuroxime is a second-generation cephalosporin β-lactamase resistant which is active against gram-positive and most gramnegative organisms, and bacteria highly sensitive to cefuroxime are: Haemophilus influence (even strains ampicillin-resistant), Neisseria meningitides, and Streptococcus pneumoniae. Cefuroxime is available for oral, intravenous, and intramuscular administration; it is rapidly absorbed when given by mouth, the gastrointestinal absorption-rate of the prodrug cefuroxime axetil is 70%, and after hydrolysis it releases cefuroxime [1,3]. Cefuroxime and cefuroxime axetil are efficacy and safe in infants and children [5][6][7][8][9][10] and adverse-effects do not occur even when high doses are given [15]. Parenteral dosing-regimens consist of 25 mg/kg, twice-daily, thrice-daily or 4-times-daily and dose intervals decrease with infant maturation [1]. Oral dosing-regimens are: 10, 15 mg/kg and 250 mg for children aged up to 1 year, 2 to 11 and 12 to 17 years, respectively-tract [3]. Prophylaxis reduces the frequency of infection episodes of urinary-tract [3,23] and respiratory infections [20], surgery [16][17][18][19], pneumonia [21] acute otitis media [22], and sinusitis [24], and it has been used to decrease bacteria burden before surgery [6][7][8][9][10][11][12][13][14][15][16][17][18][19]. Cefuroxime axetil is the drug of choice for treatment of erythema migrans [14] and Lyme disease [29]. Treatment with cefuroxime has been used to treat generalized infections [20], urinarytract infections [22,23], and also pneumonia [21], and sinusitis [24], caused by gram-negative and gram-positive organisms. Optimization of dosing-regimens consists in maintaining cefuroxime plasma concentrations above MICs of infective pathogens for at least 50% to 60% of dosing intervals [26][27][28]. Cefuroxime crosses the brain-bloodbarrier easily and concentration in cerebrospinal fluid is about 10% of that in plasma [32][33][34][35], and successfully treated meningitis caused by β-haemolytic streptococci, Streptococcus pneumonia, and Neisseria meningitides [36,37], whereas meningitis caused by Salmonella species requires an additional dose of 5 mg injected intraventricularly. Cefuroxime pharmacokinetics have been assessed in infants [39], in infants and children [37,41] and those, following cefuroxime axetil administration have been reported in infants and children [40]. Cefuroxime half-life is 1.8 hours in children aged up to 6 years [37] and 6 hours in infants during the first week of life [38]. Following three doses of cefuroxime axetil, cefuroxime Cmax and AUC 0-∞ were not statistically different indicating that cefuroxime does not accumulate in blood [40]. Consistent results were obtained following cefuroxime administration on the first and 6 th dose, thus cefuroxime does not accumulate in blood even for prolonged treatment [37]. Distribution volume of cefuroxime is 0.60 L/kg in infants and children; thus it migrates into body tissues in relevant amounts and diffuses in more than one half of the body volume [37]. Escherichia coli [42,44,45] and Klebsiella pneumoniae [43,45] may become cefuroxime-resistant; efflux pump [42] is responsible for increased-resistance of Escherichia coli [42] whereas resistance of Klebsiella pneumonia is caused by increased Half-life (hours) 19.1+3.51 Absorption half-life (hours) 0.037+0.006 Peak concentration obtained at the first dose (µg/ml) 328+102 Table 5. Cefuroxime pharmacokinetic parameters are obtained in 16 children, aged 1.2+0.3 years (body-weight=9.7+2.8 kg), undergoing cardio pulmonary bypass. Cefuroxime was intravenously infused at a dose of 25 mg/kg 1 hour before skin incision and another dose of 12.5 mg/kg was given in the cardio pulmonary bypass prime solution. Figures are the mean+SD, by Knoderer et al. [41] transcription-rates of acrA, ramA genes and decreased transcriptionrate of the gene OmpK35 [43]. Multidrug-resistant of Enteric extendedspectrum β-lactamase producing organism to cefuroxime has been observed in about 20% of isolates. Resistance is caused by genetic exchange mechanisms, consequently disseminate multidrug-resistance, and requires immediate attention [44]. Cefuroxime consumption is associated with increased bacteria resistance [46][47][48], continuous surveillance of resistance-rates and antibiotic consumption are of utmost importance for optimizing antibiotic dosing strategy.
In conclusion, cefuroxime is a second-generation cephalosporin β-lactamase resistant; it inhibits bacterial cell wall synthesis of grampositive and gram-negative organisms, and is rapidly bactericidal. Administration forms are available for oral and parenteral administrations, cefuroxime is rapidly absorbed when given orally, and the absorption-rate of the prodrug cefuroxime axetil is about 70%. Cefuroxime axetil works by realising cefuroxime after hydrolysis. Oral administration of cefuroxime and cefuroxime axetil yields high blood concentrations ensuring eradication of pathogenic microorganisms. These drugs, even given at high doses, do not cause adverse-effects. Cefuroxime axetil is the drug of choice to treat erythema migrans [14] and Lyme [29] diseases. Cefuroxime has been found efficacy against infections caused by gram-negative and gram-positive bacteria. Cefuroxime elimination half-live are: 6 hours in infants during the first week of life and 1.8 hours in children aged up to 6 years. Cefuroxime penetrates into the cerebrospinal fluid where achieves a dose of about 10% of that in plasma concentration, and successfully treated meningitis caused by β-haemolytic streptococci species, Haemophilus influenzae, and Neisseria meningitides, whereas meningitis caused by Salmonella species requires an additional dose of 5 mg. Some bacteria, including Escherichia coli and Klebsiella pneumoniae, may became cefuroximeresistant, and antibiotic consumption, including cefuroxime, is associated with increased resistance, and judicious use of this antibiotic ensures efficacy against bacteria. | 2020-11-12T09:06:47.962Z | 2020-01-01T00:00:00.000 | {
"year": 2020,
"sha1": "9bd879a9865392186ebf9572e4e325d16254fb93",
"oa_license": "CCBY",
"oa_url": "https://www.oatext.com/pdf/CMI-5-212.pdf",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "2d60d9c30d7689a7fd4af087d9cb41c9cb061aed",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
118083591 | pes2o/s2orc | v3-fos-license | Quantum semigroup structures on quantum families of maps
These are notes accompanying lectures at 7th ILJU School of Mathematics"Banach Spaces and Related Topics"in Gyeongju, Koera. The lectures are devoted to exposition of the theory of quantum families of maps with emphasis on the study of quantum semigroup structures appearing in this context.
is a C * -algebra with the same norm and algebraic structure as described for C(X). However, C 0 (Y ) does not have a unit unless Y is compact (ant then C 0 (Y ) = C(Y )). The algebras C(X) and C 0 (Y ) are commutative, i.e. f g = gf for all f, g in C(X) (resp. C 0 (Y )).
Example I. 1.3. Let H be a Hilbert space and consider the Banach space K(H ) of all compact operators acting on H . Then with composition of operators as multiplication and Hermitian conjugation as the involution K(H ) is a C * -algebra. It is unital (i.e. has a unit) if and only if H is finite-dimensional. Unless dim H = 1, K(H ) is not commutative.
Let A and B be C * -algebras. A map Φ form A to B is called a * -homomorphism if Φ is a homomorphism of algebras A → B and Φ(a * ) = Φ(a) * for all a ∈ A. It turns out that any *homomorphism is automatically a contraction for the norms of A and B. A * -isomorphism is a bijective * -homomorphism (which is automatically isometric).
It turns out that the C * -algebras presented in Examples I.1.2 and I.1.3 are to some extent general. In the context of Example I.1.2 we have the famous theorem of Gelfand which will be discussed in detail below.
Theorem I.1.4 (I.M. Gelfand). Let A be a commutative C * -algebra with unit. Then there exist a compact space X and a * -isomorphism of C(X) onto A. The space X is unique up to homeomorphism.
By a simple trick of adjoining the unit one can extend Theorem I.1.4 to the case of non-unital C * -algebras. More importantly, morphisms between C * -algebras can be characterized in a similar way. In order to be more precise we will make a brief digression and define morphisms of C *algebras.
I.1.1. Multipliers and morphisms. Let A be a C * -algebra. The multiplier algebra of A is a canonically defined unital C * -algebra M(A) such that A is embedded as an ideal in M(A) which is essential, i.e. any other non-zero ideal of M(A) has an non-zero intersection with A. The defining property of M(A) is that if C is a unital C * -algebra and we are given an embedding of A into C as an essential ideal, then there exists an injective * -homomorphism of C into M(A) extending the identity map A → A (considered as a map from a subset of C onto a subset of M(A)). It can be shown that M(A) = A if and only if A is unital.
Example I.1.5. Let Y be a locally compact space. Then the multiplier algebra M C 0 (Y ) of C 0 (Y ) is canonically isomorphic to the C * -algebra C b (Y ) of all bounded continuous functions on Y . This algebra can be further shown to be isomorphic to the algebra of all continuous functions on the Stone-Čech compactification of Y .
Example I.1.6. Let H be a Hilbert space. Then the multiplier algebra M K(H ) of K(H ) is canonically isomorphic to the C * -algebra B(H ) of all bounded operators on H .
The notation Φ(A)B stands for the linear span of all products of the form Φ(a)b with a ∈ A and b ∈ B. Recall that B is an ideal in M(B) so such products are automatically elements of B. The set of all morphisms from A to B will be denoted by Mor (A, B).
Any Φ ∈ Mor(A, B) admits a unique extension to a * -homomorphism Φ of unital algebras M(A) → M(B). It follows that if C is another C * -algebra and Ψ ∈ Mor(C, A) then we have a well defined map Φ•Ψ : C → M(B). It turns out that this is a morphism form C to B. In particular morphisms can be composed.
With help of the notion of morphisms of C * -algebras we can re-state the Gelfand theorem in its fuller version: Theorem I.1.7.
(1) Let A be a commutative C * -algebra. Then there exists a unique up to homeomorphism locally compact space X such that A is isomorphic to the C * -algebra C 0 (X). (2) Let A and B be commutative C * -algebras and let X and Y be locally compact spaces such that B = C 0 (X) and A = C 0 (Y ). Then Thus any statement about locally compact spaces and continuous maps between them can be formulated as a statement about commutative C * -algebras and their morphisms.
Before continuing let us note the following: let A and B be C * -algebras. Among many norms on the algebraic tensor product A ⊗ alg B there is always the smallest norm satisfying (the appropriate analog of) (I. 1.2). This norm is usually called the minimal C * -norm on A⊗ alg B and the completion of A ⊗ alg B with respect to this norm, denoted by A ⊗ B is called the minimal tensor product of A and B. In case either A or B is commutative (and also in many other cases) this is, in fact, the only C * -norm on A ⊗ alg B. Moreover we have Proposition I.1.8. Let X and Y be locally compact spaces. Then the minimal tensor product
I.2. Quantum spaces and their morphisms
The Gelfand theorem tells us that the category of locally compact spaces and continuous maps is dual to the category of commutative C * -algebras with morphisms. But the class of all (not necessarily commutative) C * -algebras with morphisms is also a category (with the commutative ones forming a full subcategory). Can we treat all C * -algebras as algebras of functions? Of course, a non-commutative C * -algebra cannot be isomorphic to an algebra of functions on a space, but it can be regarded as one. Thus we arrive at the following definition: Definition I.2.1. A quantum space is an object of the category dual to the category of all C *algebras with morphisms.
Thus any C * -algebra A is thought of as the algebra of functions on a quantum space X. From now on we will write C 0 (X) instead of A and treat the virtual object X in a way reminiscent of studying a locally compact space. We will call a quantum space X compact if the C * -algebra C 0 (X) is unital and in this case we will use the symbol C(X) to denote this C * -algebra. Following this analogy we will say that a quantum space M is finite if C 0 (M) is a finite-dimensional algebra (in this case it is automatically unital, so according to the remarks above we will denote it by C(M)).
The concept of a map between quantum spaces is also quite clear. If X and Y are quantum spaces (i.e. we are given two C * -algebras which we denote C 0 (X) and C 0 (Y)) then a continuous map from X to Y is, by definition, an element of Mor C 0 (Y), C 0 (X) .
It has to be stressed that we are not developing a theory of some new kind of mathematical object. Any theorem about quantum spaces and their maps is a theorem about C * -algebras and their morphisms. We end this section with the following warning: motivated by Proposition I.1.8 one is tempted to define the Cartesian product of two quantum spaces X and Y as the quantum space E such that C 0 (X) ⊗ C 0 (Y) = C 0 (E). In fact we will almost do this. More precisely we will discuss various objects and phenomena in the category of quantum spaces in which the tensor product of C * -algebras will play a role analogous to the role played by the Cartesian product of topological spaces. However this should not be taken as a strict definition because the categorical properties of the tensor product are not quite the same in the category of commutative algebras (where the tensor product is a categorical sum) and in the category of all C * -algebras.
I.3. Compact quantum groups and quantum semigroups
Definition I.3.1 (S.L. Woronowicz). A compact quantum group is a compact quantum space G such that there exists a ∆ ∈ Mor C(G), The map ∆ form Definition I.3.1 is called the comultiplication or coproduct while the density conditions in point (2) are referred to as the cancellation properties. The reason for this is that if G (as in Definition I.3.1) is a classical space (i.e. the C * -algebra C(G) is commutative) then writing G instead of G we obtain form ∆ a continuous associative two-argument operation µ : G × G → G on the compact space G and the density conditions are equivalent to the fact that µ obeys cancellation from the right and from the left. It is a known fact that a compact semigroup whose multiplication obeys cancellation properties is a compact group, and so any compact group provides an example of a compact quantum group.
Compact quantum groups have been studied for many years now and many interesting examples were discovered. However, our main focus in these notes will be on more general objects. An example of such a more general structure is that of a compact quantum semigroup: a compact quantum space S is a compact quantum semigroup if it is equipped with a coassociative ∆ ∈ Theorem I.4.1 (J.R. Jackson). Let X, Y and E be topological spaces such that X is Hausdorff and E is locally compact. For ψ ∈ C(X × E, Y ) define σ(ψ) as the mapping from E to C(X, Y ) given by Then σ is a homeomorphism of C(X×E, Y ) onto C E, C(X, Y ) with all spaces of maps topologized by their respective compact-open topologies.
In other words a continuous family of maps X → Y parametrized by points of E is encoded in a single map E → C(X, Y ) and vice versa.
Before stating the definition of a quantum family of maps let us dwell for a little while on the universal property of the space C(X, Y ). Assume that X is locally compact and let Γ be the It is known that Γ is continuous. Moreover if E is a locally compact space and ψ : X × E → Y is continuous then there exists a unique continuous map λ : Indeed, the map λ is simply σ(ψ), where σ is the homeomorphism from Jackson's theorem. The uniqueness follows from the fact that σ is one-to-one. This universal property of the pair C(X, Y ), Γ characterizes it uniquely. We will use this property to define the analog of C(X, Y ) for quantum spaces.
I.4.2. Definition of a quantum family of maps.
Definition I.4.2. Let X, Y and E be quantum spaces. A continuous quantum family of maps X → Y parametrized by E is an element Let us analyze Definition I.4.2 in more detail. The morphism Ψ describes a continuous map from the quantum space corresponding to C(X)⊗ C(E) to Y. In view of Jackson's theorem and the fact that tensor product corresponds to Cartesian product of spaces (cf. Proposition I.1.8) we can interpret this as a family of maps X → Y parametrized by the quantum space E. Specifying the definition of a quantum family of maps to the "classical case", i.e. when all considered C * -algebras are commutative, we obtain the standard notion of a continuous family of maps: More generally, if only E is a classical space (which means that C 0 (E) is commutative) then there exists a unique family {Ψ e } e∈E of morphisms from C 0 (Y) to C 0 (X) parametrized continuously (in a natural topology on the set of morphisms) by points of E.
Since a quantum family of maps is simply a morphism from a C * -algebra to a tensor product of two C * -algebras, it is not our aim to derive a general theory of such objects, as it would not likely give non-trivial results. We will rather focus on particular examples of quantum families, especially those possessing certain universal properties (see Lecture II).
Before we deal with quantum families of maps with such additional properties let us define the composition of quantum families of maps. This is a direct generalization of the notion of composition of classical families of maps.
Definition I.4.4. Let X 1 , X 2 , X 3 , D 1 and D 2 be quantum spaces. Consider families of maps The composition of Ψ 1 and Ψ 2 is the quantum family of maps Note that the composition of morphisms in (I.4.1) is defined via extending ( It is easy to see that if the considered quantum families are classical (i.e. the algebras C(D 1 ) and C(D 2 ) are commutative) then composition of quantum families corresponds exactly to the family of all compositions of * -homomorphisms from the families Ψ 1 and Ψ 2 . The operation of taking a composition of quantum families is associative in the following sense: let X 1 , X 2 , X 3 , X 4 , D 1 , D 2 and D 3 be quantum spaces and let .
From now on we will concentrate on compact quantum spaces (the considered C * -algebras will be unital). The reasons for this are the following: • it is technically easier, in particular morphisms become simply unital * -homomorphisms, • in all considered examples the interesting C * -algebras will be unital, • the universal properties formulated in the category of compact quantum spaces in fact hold in the category of all quantum spaces and the technical details are not at all difficult, while the notation is much heavier. Therefore we will from now on even refrain from writing Ψ ∈ Mor(A, B) favoring the notation Φ : A → B.
I.5. Quantum families of all maps, existence question Definition I.5.1. Let X and Y, E be quantum spaces and let Φ : C(Y) → C(X)⊗C(E) be a quantum family of maps. We say that • Φ is the quantum family of all maps from X to Y and • E is the quantum space of all maps from X to Y if for any quantum space D and any quantum family Ψ : C(Y) → C(X) ⊗ C(D) there exists a unique Λ : C(E) → C(D) such that the diagram commutes.
The motivation for Definition I.5.1 comes from the universal property of the space C(X, Y ) described at the end of Subsection I.4.1. Indeed, upon specializing to the classical situation (X, Y, E are taken to be classical spaces as well as all possible spaces D), Definition I.5.1 turns out to be exactly the definition of the space C(X, Y) with its compact open topology). This result is not entirely trivial.
The immediate question now is whether, given two quantum space X and Y the quantum space of all maps from X to Y (and the corresponding quantum family) exists. In the next theorem we use the notion of a finitely generated unital C * -algebra. By this we mean an algebra which is a quotient of the full group C * -algebra of a free group on finite number of generators.
Theorem I.5.2. Let X and Y be quantum spaces such that C(X) is finite dimensional and C(Y) is finitely generated and unital. Then the quantum space of all maps X → Y exists. Moreover this quantum space is compact.
One way to make sure the assumptions of Theorem I.5.2 are satisfied is to assume that both C(X) and C(Y) are finite dimensional. Indeed many very interesting examples come from the situation when X and Y are the same finite quantum space M. Contrary to the classical situation, examples of such spaces can be very intriguing.
Example I.5.3. Let M be the classical two point space: C(M) = C 2 . Clearly the classical space of all maps M → M has four points. However, the quantum space of all maps M → M is much more interesting. It turns out to be a quantum space E such that The quantum family of all maps M → M is Φ : The infinite dimensionality of C(E) justifies saying that the quantum space of all maps from a two point space to itself is infinite.
The explanation of these formulas is quite simple: in order to determine a * -homomorphism from C 2 to C 2 ⊗ A, where A is some unital C * -algebra, we only need to define its value on [ 1 0 ]. This value can be written as , and a, b ∈ A must be projections because [ 1 0 ] is. The C * -algebra C(E) described above is simply the universal C * -algebra generated by two projections without any relations.
Lecture II
Idea of proof of Theorem I.5.2. Begin by proving the theorem for C(Y) = C * (F n ). If C(X) has the decomposition then we define E n so that C(E n ) is generated by be unitary. The we define a map Φ n as then the unique * -homomorphism where {v 1 , . . . , v n } are the generators of F n . The pair C(E n ), Φ n ) has the universal property that for any quantum space D and any quantum family Ψ : C * (F n ) → C(X) ⊗ C(D) there exists a unique Λ n : C(E n ) → C(D) such that the diagram commutes. Now we take Y such that C(Y) is a quotient of C * (F n ) and let π : C * (F n ) → C(Y) be the quotient map. Define E so that C(E) is the quotient of C(E n ) by the ideal generated by the set (ω ⊗ id)Φ n (x) x ∈ ker π, ω ∈ C(X) * and let λ : C(E n ) → C(E) be the quotient map. Now it can be shown that there exists a map Φ : The pair C(E), Φ will be the quantum space and family of all maps X → Y. The universal property follows from the fact that if D is a quantum space and Ψ : C(Y) → C(X) ⊗ C(D) is a quantum family of maps then by the first step of the proof, there exists a unique Λ n : C(E n ) → C(D) such that Ψ•π = (id ⊗ Λ n )•Φ n .
Then it can be shown that there exists a unique Λ : C(E) → C(D) such that Λ n = Λ • λ (this follows form the construction of C(E) and λ). Therefore the diagram commutes. It follows that the diagram is simply a part of (II.0.2).
The fact that the quantum space of all maps X → Y might not exists should rather be interpreted as our inability to see it. More precisely, the approach to non-commutative topology we have adopted (via C * -algebras) should be viewed more as "non-commutative topology of locally compact quantum spaces". It is clear that the space of continuous mappings with compact open topology often fails to be locally compact. This happens also in the classical case.
is commutative. Moreover the * -homomorphism ∆ is coassociative: and there exists a unique character ǫ of C(E) such that In other words the quantum space of all maps M → M has a natural structure of a compact quantum semigroup. The character ǫ corresponds to the unit element of this semigroup and is called a counit.
Example II.1.2. Let, as in Example I.5.3, C(M) = C 2 and let E be the quantum space of all maps M → M. We know that C(E) is generated by the two distinguished elements P and Q (see Example I.5.3). It is easy to find that the quantum semigroup structure on E is given by the comultiplication: Also ǫ(P ) = 1 and ǫ(Q) = 0.
It is natural to expect that the quantum semigroup of all maps M → M is not a quantum group. Let X be a quantum space and let ω be a state on C(X) (i.e. a positive linear functional of norm 1 -an analog of a probability measure on M). We say that a quantum family of maps Ψ : for all a ∈ C(X). This is a direct generalization of the standard notion of a family of maps preserving a given (probability) measure. preserving ω there exists a unique Θ : C(W) → C(D) such that The quantum semigroup W of all maps M → M preserving ω q can be described as follows: C(W) is generated by three elements β, γ and δ such that The comultiplication ∆ ωq : C(W) → C(W) ⊗ C(W) is given by The counit ǫ maps γ and δ to 0 and β to 1. The proof of the above facts is mainly computational. One first observes that M 2 (C) is the universal C * -algebra generated by one element n such that n 2 = 0, nn * + n * n = 1.
(in fact we can specify n = 0 1 0 0 ). Then the mapping Φ ωq will map n to a matrix α β γ δ of elements of C(W). The relations between matrix elements are derived from 1+q 2 1. In particular it follows from the relations that α = −q 2 δ, so be two quantum families of maps. We say that Ψ 1 and Ψ 2 commute if where σ is the flip Theorem II. given by Ψ(m) = ψ(m) ⊗ 1. Let U be the quantum commutant of Ψ. The C * -algebra C(U) is generated by α, β and γ with β = β * , γ = γ * and α * α + γ 2 + αα * + β 2 = 1, α 2 + βγ = 0, α * β + γα * + αγ + βα = 0, αβ + βα * = 0, γα + α * γ = 0. The comultiplication ∆ Ψ : C(U) → C(U) ⊗ C(U) acts on generators as It can be shown that U is not a compact quantum group (with the comultiplication ∆ Ψ ). II.3. Further topics: quantum space of maps from a finite quantum space to a quantum semigroup and its structure So far we have concentrated on subsemigroups of the quantum semigroup of all maps from a given finite quantum space to itself. In classical topology there is also another way to form semigroups of maps. Namely one considers the space of all continuous maps form a fixed topological space to a topological semigroup. This time the group operation is not composition of maps, but a pointwise multiplication.
A non-commutative analog of this construction was investigated first by M.M. Sadr. His idea is contained in the following theorem: The crucial remarks are: • Upon substituting a classical semigroup S for S and taking the classical space of all maps K → S instead of H, we obtain exactly the structure of the semigroup of all maps K → S. • As stated, Theorem II.3.1 can only be considered for classical spaces K. Indeed, if we take a non-commutative C * -algebra instead of C(K) the multiplication map µ is no longer an algebra homomorphism. With a more detailed analysis of the structure of C(H), ∆ H shows the following: Proposition II.3.2. Let K be a finite set and let S be a quantum semigroup as in Theorem II.3.1. Let H be the quantum space of all maps K → S with the quantum semigroup structure as described in Theorem II. 3
.1. Then H is a quantum group if and only if S is.
This fact is quite puzzling in view of the following example: Example II.3.3. Let K be the quantum space defined by C(K) = M 2 (C) and let S be the classical group Z 2 . Let H be the quantum space of all maps K → S. Then C(H) is isomorphic to the universal unital C * -algebra generated by three elements p, q and z with the relations p = p * , p = p 2 + z * z, zp = (1 − q)z, q = q * , q = q 2 + zz * .
It can be shown that this C * -algebra does not admit a compact quantum group structure. Therefore one cannot hope for an analog of Proposition II.3.2 when K is not a classical space (since Z 2 is a group, hence a quantum group, and H does not admit a quantum group structure). | 2013-07-15T17:26:16.000Z | 2013-07-15T00:00:00.000 | {
"year": 2013,
"sha1": "8e5f626a26009a3c2408b7f61e6bac6025836a65",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "8e5f626a26009a3c2408b7f61e6bac6025836a65",
"s2fieldsofstudy": [
"Mathematics",
"Physics"
],
"extfieldsofstudy": [
"Mathematics"
]
} |
158738553 | pes2o/s2orc | v3-fos-license | Shifting Ground : Alcohol Liability and the Normative Content of Tort Law in Australia
An Australian High Court decision has found a publican not liable in tort for damage due to a patron's decision to drive whilst intoxicated. This decision is the most recent of a series of similar cases, and legislative intervention in all states, which attribute liability on the basis of fault, consistently with traditional negligence actions. However, the exposition of normative grounds in negligence actions is often inchoate. This paper attempts to explicate alternative moral bases for negligence actions which shift loss away from the primarily responsible plaintiff to a secondarily responsible defendant, considering in particular whether there is a regressive tendency in shifting responsibility for self-preservation. It will consider whether moral responsibility is diminished by the state of intoxication and whether the decision to drink to the point of intoxication thus becomes the basis of moral censure. Finally, it considers whether the normative content of the tort of negligence is capable of withstanding nonnormative decisional claims.
Introduction
T HE ROLE OF the law of torts in common law jurisdictions has been subject to significant philosophical critique; in particular, the 'tension between the corrective justice and the utilitarian or wealth maximization conceptions of tort law' has been subject to extensive debate. 1 The school of thought that characterises the aims of tort law in economic or utilitarian terms has many adherents whose theses characterise public debate.However, there remains a significant residual claim that the law of torts is grounded in morality.The elements of the law of negligence in particular are situated in the moral duties and sensibilities of the 'reasonable person'.This article traces the remaining moral aspect of the law of negligence particularly as it manifests itself in claims relating to third party alcohol liability.In other words, it addresses issues arising when damage consequential upon alcohol abuse is attributed to someone other than the intoxicated party.In this way, risk is shifted from one who is more 'morally' reprehensible to one who is less so.This area of law has, in Australia, been subject to legislative intervention, and cases subsequent to the legislative change have recently retraced the steps that have allowed recovery from third parties.This paper considers the relative moral positions of the defendant and the plaintiff to assess the perspective from which the law of negligence is currently considered.This account does not make any normative claims -it merely assesses the apparent basis upon which the law of negligence is currently situated.
The Normative Content of Tort Law
'In tort, the most familiar divide today is that between the law-and-economics camp that focuses on efficient deterrence, and the philosophical camp that tends to focus on corrective justice.'2In the tort of negligence as it applies to cases involving alcohol liability, this division can be considered from the point of view of two claims: 1.That the objective of compensation to attain the most efficient allocation of resources, so that liability achieves deterrence in the form of 'primary accident cost avoidance'.
As Calebrese observes/notes: '[A] pure market approach to primary accident cost avoidance would require allocation of accident costs to those acts or activities (or combinations of them) which could avoid the accident costs most cheaply.'32. That at objective of compensation is to achieve the objective of distributive justice, so that loss is shifted from the injured person on the basis of the moral 'blameworthiness' of the defendant; although sometimes, when there is an element of contributory negligence, the issue is the comparative blameworthiness of the parties.
The law and economics school is essentially a realist position, and does not speak to the origins of tort law; in any event it is my position that there is sufficient residual moral force in tort law to prompt an articulation of comparative moral claims in the shifting of loss.This is a vexed issue in tort law: [it] is initially puzzling from the point of view of distributive justice.It protects persons and property against injury and invasion without regard to the distributions it upholds.It focuses on misfortunes that one person brings on another, leaving equally devastating losses to lie where they fall.Why have an institution charged with figuring out whose problem it is when things go wrong-that answers its questions not by looking to some idea of distribution, but instead to norms of conduct governing relations between private parties, invoking duties and standards of care, questions of remoteness, proximity and causation?4 The elements of individual torts act as proxies for the moral content of the tort, and potentially disguise the presence of any moral coding.This paper focuses in particular on the role of the law of negligence, which has as basic premise the provision of compensation to those injured by the failure of another to comply with a duty of care.The law of negligence in Australia has evolved three components: a duty to take care, a breach of that duty, and damage caused by the breach which is not too remote from the breach.Of these principles, the principle of duty had, in its origin, the most obviously moral overtones.However it suffers from significant conceptual and definitional difficulties.
Early conceptualisations of duty were expressly compared with the moral position.In a sweeping re-evaluation of pre-existing particular rules, the classic formulation of duty in Donoghue v Stevenson 5 Lord Atkin said: 'in English law there must be, and is, some general conception of relations giving rise to a duty of care, of which the particular cases found in the books are but instances.'You owe a duty to your 'neighbour' only: acts or omissions which any moral code would censure cannot in a practical world be treated so as to give a right to every person injured by them to demand relief.In this way rules of law arise which limit the range of complainants and the extent of their remedy.The rule that you must love your neighbour becomes in law, you must not injure your neighbour; and the lawyer's question, Who is my neighbour?receives a restricted reply.You must take reasonable care to avoid acts or omissions which you can reasonably foresee would be likely to injure your neighbour.Who, then, in law is my neighbour?The answer seems to be -persons who are so closely and directly affected by my act that I ought reasonably to have them in contemplation as being so affected when I am directing my mind to the acts or omissions which are called in question. 6rd Atkin's reference to 'the lawyer's question' reflects the influence of the Judeo-Christian tradition.It contains an allusion to the Christian Gospel: thou likewise. 7 Although the Atkinian analysis directly eschewed the applicability of the higher moral standard, the comparison with religious code and the moral imperative 'ought' import into this analysis direct moral force.The impracticality of the duty analysis as a generalisation is, however, problematic; it fails to import certainty, so plaintiffs and defendants alike cannot accurately manage risk or determine liability in the event of injury.As the law of negligence has expanded to regulate an increasing number of commercial realms, negligent misstatement and mass torts the desire to impose certainty has battled with the reluctance to arbitrarily minimise recovery by injured parties.
The quandary in Australia is articulated by the High Court in Sullivan v Moody: 8 As Professor Fleming said, 'no one has ever succeeded in capturing in any precise formula' a comprehensive test for determining whether there exists, between two parties, a relationship sufficiently proximate to give rise to a duty of care of the kind necessary for actionable negligence. 9 As a result of this conceptual uncertainty, various jurisdictions have trialled alternative formulations, rooted in the Atkinian test, but teasing out [potentially] more definable elements, such as foreseeability, proximity, public policy, vulnerability, and general or specific reliance.
In the United Kingdom, a three stage test has been applied which requires analysis of whether it is 'fair, just and reasonable' to impose a duty of care, 10 although the Australian High Court has eschewed this formulation as 'capable of being misunderstood as an invitation to formulate policy rather than to search for principle.' 11he inchoate moral component to the duty question is less apparent in relation to the question of breach; although questions of duty and the content of that duty are often inextricably intertwined.The question of breach is generally considered to be a question of fact: in his Concluding unscientific postscript in Berrigan Shire Council v Ballerini 12 Callaway JA noted that the imposition of general duties of care in negligence has no sure foundation in legal principle.It involves trade-offs and value judgments that may have been better left to the legislature in 1932 and again in 1963 and 1976. 13 The search for a principle is worse than looking for a needle in a haystack.The needle is not there.The second observation is that, however the difficulties are to be resolved, duty is a question of law and breach is a question of fact.The discipline of civil juries, which we still have in Victoria, However, this is a little misleading: pragmatically, there are two steps to determining the question of breach: to show that there has been a falling below of the standard, you must first define the standard that is required to be met -the standard of care.This standard is one expected by society at a given time and place, as defined by tort law -it is measured by what the reasonable person in the circumstances would have done, which is an objective test.This is a question of law, and involves an analysis of the trade-offs a reasonable person would make when considering the precautions required to protect against negative consequences of their actions.The common law articulation of the test is typified by that of Mason J in Wyong Shire Council v Shirt: 15 In deciding whether there has been a breach of the duty of care the tribunal of fact must first ask itself whether a reasonable man in the defendant's position would have foreseen that his conduct involved a risk of injury to the plaintiff or to a class of persons including the plaintiff.If the answer be in the affirmative, it is then for the tribunal of fact to determine what a reasonable man would do by way of response to this risk.The perception the reasonable man's response calls for a consideration of the magnitude of the risk and the degree of the probability of its occurrence, along with the expense, difficulty and inconvenience of taking alleviating action and any other conflicting responsibilities which the defendant may have. 16is test, with minor variations, has been incorporated into the Wrongs Act 1958 (Vic) s.48 and similar legislation in other Australian states.Articulations of comparative social utility of activities have potential moral overtones; although, of course, they could be analysed by economic criteria.
Potentially moral questions assert themselves in the analysis of the damage requirement in the tort of negligence; typically at the points where the law requires a policy judgment to draw the line between recovery and non-recovery.Thus, although the concept of causation is said to have policy (not moral) content, considerations of what is the 'cause' of another act is capable of moral direction.On appeal in March v E & MH Stramare Pty Ltd 17 Mason J noted that [c]ommentators subdivide the issue of causation in a given case into two questions: the question of causation in fact -to be determined by the application of the "but for" test -and the further question whether a defendant is in law responsible for damage which his or her negligence has played some part in producing: …It is said that, in determining this second question, considerations of policy have a prominent part to play, as do accepted value judgments: see Fleming, p 173.However, this approach to the issue of causation (a) places rather too much weight on the "but for" test to the exclusion of the "common sense" approach which the common law has always favoured; and (b) implies, or seems to imply, that value judgment has, or should have, no part to play in resolving causation as an issue of fact.As Dixon C.J., Fullagar and Kitto JJ. remarked in Fitzgerald v. Penn (at p 277): "it is all ultimately a matter of common sense" and "(i)n truth the conception in question (i.e., causation) is not susceptible of reduction to a satisfactory formula": at p 278. …That said, the 'but for' test, applied as a negative criterion of causation, has an important role to play in the resolution of the question.… The principles on causation have been set out in the legislation, separating the normatively neutral question of fact from the value laden question of appropriateness, and also requiring that policy considerations dealing with remoteness of damage are considered expressly. 18n Victoria, for example amendments to the Wrongs Act 1958 (Vic) introduce the notion of 'scope of liability', which is a reference to remoteness of damage.Section 51 states: 1.A determination that negligence caused particular harm comprises the following elements-(a) that the negligence was a necessary condition of the occurrence of the harm (factual causation); and (b) that it is appropriate for the scope of the negligent person's liability to extend to the harm so caused (scope of liability).
Remoteness of damage is a matter of judgment and degree, and is heavily influenced by policy.This is underlined by s 51(4), which states that '[f]or the purpose of determining the scope of liability, the court is to consider (amongst other relevant things) whether or not and why responsibility for the harm should be imposed on the negligent party.'It is clear that these sections are designed to compel the courts to clearly articulate and justify their choice of policy factors, but the section gives no guidance on how appropriateness should be determined, so it is likely that the settled common law principles for testing remoteness -reasonable foreseeability -will continue to apply.
In sum, moral considerations have the potential to be applied at all stages of the negligence analysis, either as a direct application of a moral duty to a 'neighbour', or as an inchoate manifestation of shared moral principles informing 'policy'.
Alcohol Liability
Alcohol misuse is provocative in public policy because it is at once a pressing social problem and an articulation of consumer choice; at once a source of taxation revenue and a burden on the public health system.The imposition of liability on an individual for damages incurred whilst intoxicated is consistent with the principles espoused by tort law from both a moral and an economic perspective.Alcohol liability has attracted significant commentary in a number of jurisdictions.
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
There are circumstances in which the consumption of alcohol is a problem in itself -for instance, where the imbiber is underage.There are jurisdictions in which criminal charges may be brought against adults involved in the supervision of premises on which alcohol is consumed by underage drinkers.In tort law, however, a problem will typically arise where there is alcohol abuse.Excessive consumption of alcohol can lead to the following easily conceded consequences: 1. Physical injury or death to the person who has consumed the alcohol as a result, for instance, of alcohol poisoning or suffocation in his or her vomit. 20a. Physical injury or death may result as a result of secondary consequences of alcohol consumption, such as impaired faculties: drink driving leading to an accident, falls, cuts by broken glass, and so on. 21Problems have also occurred with the provision of band entertainment, with the presence of alcohol-and possibly drug-affected patrons, security guards who are frequently external contractors, loud noise, possibly strobe lights, dancing and stage-diving, and a confused, highly stimulating environment which could lead to epileptic seizure. 222. Physical injury or death to a third party as a result of the actions of the intoxicated person, whether they are accidental or deliberate. 233. Damage to property belonging to the person who has consumed the alcohol -for instance, a damaged vehicle; 4. Damage to property belonging to a third party.5.More obliquely, an uncontrolled environment of alcohol consumption and abuse may result in dangerous, threatening or harassing conditions.
In all cases, the danger arises as a consequence of the consumption of alcohol, and it could be asserted that the consumer is the party primarily responsible for the consequences.However, there are frequent circumstances in which liability can shift -or be shared -with a third party, typically the server of the alcohol.Alcohol liability is one of the increasingly broad terrains in tort law in which liability may be 'peripheral' -that is, the tortfeasor is not directly involved in the action giving rise to the
FRANCINE ROCHFORD
loss, but is in a relationship of control or influence over the person who is directly involved, or where there are other factors which may increase the scope of the duty of care.Liability, therefore, is primary, rather than vicarious, but in an ordinary causal sense the tortfeasor is not directly responsible for the action.The response to this question addresses a more general concern with the liability of peripheral parties which is 'a growing feature of modern tort law'. 24Whereas the question extends into a number of situations, 'the most noteworthy and troublesome form of this phenomenon is the increase in the past twenty years in the number of novel claims alleging that the defendant had negligently failed to control or deter the active injuring party or warn the plaintiff about the latter'. 25This species is considered to be most problematic because 'there can very often be a mismatch who is being sued and the perceived relative unimportance of that party's role in producing the damage: in other words, very often, the party who is alleged to have failed to control the third party is peripheral in a causal sense '. 26 To impose liability in such a case carries two dangers.First, it is often true that a plaintiff could plausibly implicate the omissions of very many parties to control the third party who in fact injured him or her.Hence to hold liable all those whose omissions were so implicated would signal a vast inhibition on freedom.Secondly, imposing liability on a peripheral party who had merely failed to control a third party deflects attention away from the party or parties directly and principally responsible for the damage. 27cordingly, the circumstances in which a person has a duty to control another's actions to prevent harm to strangers are the exception rather than the rule. 28In certain circumstances a person may be held liable in negligence for the actions of third parties causing harm; however, those circumstances are limited.Nevertheless, it is well established that a person may owe a duty of care to prevent someone under their control or on their premises from harming a third party, although there are limits on the circumstances in which this would be the case.For instance, Smith v Leurs 29 sets out the limits of a parent's liability to control children; Dorset Yacht Co Ltd v Home Office 30 discusses the liability of a prison authority to landowners in the vicinity of an institution for acts of escaped inmates; Hogan v Gill 31 deals with the situation in which a six year old gained access to a gun which discharged, causing injury and Curmi v McLennan 32 arose when 16 or 17 year old children gained access to guns which discharged, causing injury.The occupier of premises to which the public is admitted may be under a duty to prevent one patron from causing injury to another. 33lcohol liability cases have had mixed results, and hinge on their facts.In Chordas v Bryant (Wellington) Pty Ltd 34 the plaintiff was struck by another patron at a bar and claimed that the negligence of the hotel owner had resulted in the provision of alcohol to the already drunk patron, and that it was unsafe so to do.However, the patron had not been causing any problems prior to his altercation with the plaintiff, and the assault was unforeseeable.It seems likely that the result would have been different if the patron had been visibly intoxicated and had been causing a disturbance prior to the assault.Conversely, in Wormald v Robertson 35 hotel staff were aware that the patron was causing a disturbance, but did not act.The plaintiff intervened when the intoxicated patron grabbed a woman at the bar.The patron smashed a glass beer jug in the plaintiff's face.In that case the court held that the hotel owner had breached its duty of care by failing to eject the intoxicated patron. 36he Supreme Court of Queensland caused a groundswell of dissent in Johns v Cosgrove 37 by awarding damages to a man struck by a car on his way home after being served to a state of 'gross intoxication' in a Surfers Paradise hotel.Whereas Justice Des Derrington specified some limitations on the circumstances in which a duty of care would be owed, the decision still urged caution upon those who provide alcohol or allowed the service of alcohol.He said that, while it was not negligent to serve a patron to the point of drunkenness, it was negligent if it was likely the person would be endangered by being drunk.In this case, '[t]he plaintiff was served up to the state of gross intoxication when he was a danger to himself for it was known to the bar staff that he would then go home by a route that would take him into close proximity to a busy highway'. 38Justice Derrington noted that the hotel owner would not have been held liable merely because the patron was served alcohol to the point of intoxication.The fact which led to liability was the knowledge of the hotelier that the patron would be placed in a dangerous situation by the consumption of alcohol.The case was heard on appeal in the Queensland Court of Appeal on another point.
Until the recent decision in the High Court in Tandara, the most significant case had been South Tweed Heads Rugby League Football Club Limited v Cole. 39The New South Wales Court of Appeal found that a club was not liable for injury to a patron caused when the patron was struck by a vehicle whilst making her way home from a day of drinking, and the High Court agreed, holding that the club had not breached any duty it owed to the woman. 40s Cole had been at the Club from around 9:30 am, firstly taking advantage of free Spumante at a Champagne breakfast, then purchasing bottles of Spumante.She was refused service in the afternoon, on the grounds of intoxication, and was ejected at around 5:30 pm.She was offered the services of a taxi and a bus and driver by the management, but declined.She was hit by a car after leaving the club and suffered serious injuries from the accident and has continuing disabilities.A majority of the High Court held that the club was not liable on the basis that no duty was owed.Chief Justice Gleeson noted that a successful claim would give rise to 'both an unacceptable burden upon ordinary social and commercial behaviour and an unacceptable shifting of responsibility for individual choice.'He also considered 34 (198834 ( ) 91 ALR 149. 35 (1992) ) Aust Torts Reports 81-180 (Qld CA). 36See also Speer v Nash (Unreported) NSW Supreme Court, 17 December 1992. 37Unreported, Derrington J, Supreme Court of Queensland, 12 December 1997. 38Johns v Cosgrove (Unreported, Derrington J, Supreme Court of Queensland, 12 December 1997. 39[2004] HCA 29.
FRANCINE ROCHFORD
that she was not so intoxicated as to be held not legally responsible for her own actions.Justices Gummow and Hayne did not find it necessary to determine whether a duty was owed, considering that if there was a duty, then it had been discharged.Justice McHugh, in dissent, found that an occupiers' duty applied to the club, whereas Kirby J in dissent considered that '[t]he club, that began the day with a breakfast with the supply of free bottles of spumante in large quantities and thereafter tolerated the continued presence on its premises of a patron … who was obviously drunk …cannot say that it owed no duty of care to her.' The most recent statement in the High Court as to the liability of a server of alcohol to an intoxicated patron occurred in CAL No 14 Pty Ltd t/as Tandara Motor Inn v Motor Accident Insurance Board (Tandara). 41In that case the Court allowed an appeal by the owners and licensees of a hotel against the Motor Accidents Insurance Board, the Tasmanian statutory body which provides compulsory insurance against the personal injury as a result of traffic accidents in that state.Scott, a motorcyclist, was drinking at the appellant's premises and was intoxicated when he attempted to ride home.He suffered fatal injuries when he ran off the road close to his home.In proceedings in negligence, it was asserted that a duty of care was owed to Scott by the licensee, that the relevant duty was breached, and that as a consequence of that breach Scott was killed.The Supreme Court of Tasmania found that no duty was owed to Scott, as did the High Court on Appeal.The Court also found that if there had been a duty, it had not been breached, and if it had been breached it had not been shown that the breach had caused the death.
The matter which is potentially morally relevant, and the one at issue in the High Court, is the existence of a duty of care owed by the licensee to the motorcyclist.Justices Gummow, Heydon and Crennan, with French J and Hayne J concurring, noted that a range of duties were clearly owed by the licensee to the motorcyclist, including a duty to ensure that the premises were safe, and to ensure that the equipment did not injure the drinker.However, no duty arose to prevent the motorcyclist from riding home after drinking, 42 or to call his partner to ensure that she collected him. 43The motorcyclist's position was distinguished from 'exceptional' cases in which, for instance, 'a person is so intoxicated as to be completely incapable of any rational judgment or of looking after himself or herself, and the intoxication results from alcohol knowingly supplied by an innkeeper to that person for consumption on the premises'. 44Nor was the case analogous to other exceptions, such as may arise in the case of 'intellectually impaired drinkers, drinkers known to be mentally ill, and drinkers who become unconscious.' 45The court, without extending the influence of the principle, left open the potential for distinguishing between plaintiff recovery of damages on the basis of their capacity to make coherent and autonomous decisions, where the defendant would know of the incapacity.
The Moral Trajectory in Australian Cases
Tort law is immersed in the lifeworld.It is grounded in customary law and shifts and alters with the redefinition of socially contingent concepts, such as reasonableness.However, a system of law as a cultural system does not necessarily constitute a coherent, ordered set of rules.A designation as the sum of all rules enacted by a valid legislative body in a particular social order would mislead, because laws and institutions change over time, whereas the legal system is characterised by continuity.
However where, as in this case, there is confusion in the background assumptions of the society informing the lawmaker, the apparent moral legitimacy of a law is in question, without necessarily altering the validity of the legal system.If society's views have altered, and the law has not reflected that shift, or the law has changed, and society's views do not now converge with the law, tension may arise.
The obligatory nature of a legal norm makes the problem of choice between two established norms seem clear on its face; however a certain degree of tension must arise because the background presumption of the legitimacy of the legal system is challenged by this evidence that the norms systematized in the legal system are not directly or clearly informed by the norms arising in the course of cultural exchange.At times the distinction between a legal and a social norm is by no means clear, given the degree of interaction between the two.
It could be argued that, in the case of the law of torts, the perception of a 'crisis' was manifestation of a lack of convergence between legal and moral norms.There were, in fact, multiple systemic failures which contributed to the crisis, but the perception of crisis inflamed a form of 'moral panic' 46 in relation to certain types of tortious claim.In claims involving injury to parties considered to be themselves at fault, the language of crisis invokes fear and the scapegoating of unpopular litigants delivers personal vindication.In Australia, proponents of the tort law crisis had some characteristics of moral entrepreneurs. 47s it correct to characterise this as a 'moral' question?The characteristic of these actions which speaks to moral behaviour is the sublimation of personal autonomy and hence responsibility -refusing to take the consequences of blameworthy behaviour.In this case, the personal decision to drink to excess is morally blameworthy, however when that decision delivers undesirable consequences the costs of those consequences are shifted to another, who is or may be less blameworthy.This invokes a reaction of outrage, facilitated by the agency of newspapers.For instance, after the decision of the Supreme Court of Queensland in Johns v Cosgrove 48 the media attention focused on the case as a threat to the current social order. 49he decision was consistent with existing authority, but much of the disquiet prompted by the case was caused by the fear that it was a manifestation of the extension of duty and standard of care engineered by insurance companies seeking to shift their own liability.
As already indicated, the notion of 'duty of care' in negligence has, at its core, a moral aspect.This is typically framed from the perspective of the defendant.Whereas it is not equivalent to a moral duty, the legal duty has at its base a moral component.In dissent in Neindorf v Junkovic 50 Justice Kirby considered in some detail the notion of duty.The analysis indicates the value judgments to be made in the determination of breach, and the contestability of the idea of reasonable risk, but he goes further to note that To the extent that the Court turns away from the earlier principles, in my respectful view it endorses notions of selfishness that are the antithesis of the Atkinian concept of the legal duty that we all owe, in some circumstances, to each other as 'neighbours'.This is a moral notion, derived originally from Scripture, that has informed the core concept of the English law of negligence that we have inherited and developed in Australia.It is the notion that, in the past, encouraged care and attention for the safety of entrants on the part of those who invite others onto their premises.(It also encouraged such persons to procure insurance against risk).To the extent that these ideas are overthrown, and reversed, this Court diminishes consideration of accident prevention.(It also reduces the utility and necessity of insurance).From the point of view of legal policy, these are not directions in which I would willingly travel. 51om The Individual's Perspective, the Moral Framework would be Constructed Thus: The freedom of a person to do X (say, play loud music, or smoke, or drink to excess) must be balanced against the obligations X owes to the community not to so play or smoke or drink in a manner which will impede the freedom of others.The right of X to expression of musical tastes, inhalation of legal substances or eat or drink is balanced by the duty owed 50 Neindorf v Junkovic [2005] HCA 75 (Kirby J [53]).When dealing with non-primary or peripheral liability, however, the question is more complex.The right of a licensee to sell intoxicating beverages is balanced by the duty to the community not to sell it in such a way as to cause adverse consequences in the community.These correlative rights and duties are enforced by penal statutes.The imposition of tortious liability imposes a secondary rights/duties corollary -the right to sell intoxicating beverages may be balanced by a duty to the drinker and perhaps to third parties and by a shifting of the financial risk so that the risk of adverse outcomes shifts from the drinker or alternatively is shared by the drinker and the server.The policy reasons for liability under penal legislation seem clear; the policy priorities in the case of tortious liability are a mix of utilitarian and moral motives.Servers of alcohol are in a position to monitor the drinking of patrons, have the putative capacity to control the level of drinking -by refusing service -and are potentially in a position to monitor and control the consequences of intoxication -by the provision of in-house security, or by providing alternative mechanisms of transport for patrons who are drivers.Moreover, it is argued that the cost should fall upon the server, as the one profiting from the provision of alcohol.
However, a more nuanced approach which considered the comparative moral position of the drinker and the server.The High Court in Tandara, considering the duty owed by the licensee to the drinker, noted that 'the duty conflicts with [the drinker's] autonomy.The duty on the Licensee would have prevented Mr Scott from acting in accordance with his desire to ride his wife's motorcycle home52 .'If however, (absenting a universal traffic insurance scheme), the drinker had collided with and injured a pedestrian, or another motorist, would the moral autonomy afforded to the drinker necessitate him taking financial as well as moral responsibility for his actions?The unarticulated moral stance would apparently find no duty owed to the proximate drinker who had responsibility for his or her actions; does that necessarily mean that no duty was owed to the entirely innocent third party, injured as a consequence of the server's desire for revenue untrammelled by risk?
The idea of autonomy is problematic when intoxication is at issue.Intoxication strips the subject of autonomy, as it strips him or her of decision-making capacity.At what point is moral autonomy to be exercised?When the decision is made to drink the first drink?When the decision is made to drink the alcohol which tips the drinker into the state of intoxication?When the decision in made to carry out the injurious act whilst intoxicated?Problematically, unless the drinker has set out to become intoxicated at the outset, the drinker's moral agency is compromised by intoxication.Even if the drinker has set out to become intoxicated, does that necessarily import responsibility for all actions taken whilst intoxicated?Other areas of law, notably contract law, recognise the distinct position of the intoxicated person -if so intoxicated that their reason is impaired.The court's consideration in Tandara of the notion of autonomy made a distinction between the existence of a duty when the defendant was aware of their impaired capacity -thus their impaired autonomy -and the position of the drinker before that state was reached.However, the extent of that duty was not relevant to the case and accordingly not clearly articulated.
Morality and Modernity
In alcohol liability the odd corollary to modernity has been the utilisation of what could be called 'moral panic' strategies to achieve pecuniary ends, by drawing on assumptions of a declining moral trajectory to argue for limits on a purportedly burgeoning liability.
The perception of increasing potential liability became a matter of legislative concern as a result of a perception of a 'crisis' in tort litigation and liability insurance 53 which was overcompensating victims or shifting loss in circumstances where it should have been allowed to lie where it fell.Legal commentators noted that the argument that the tort system had become increasingly burdensome had not been verified, 54 and there had, in fact, been a judicial shift towards a more restrictive attitude to tort actions. 55However the perception was supported by a largely unquestioning media.
The Commonwealth government appointed an expert panel in May 2002 56 to review the law of negligence, however its mandate was limited by Treasury so that certain premises remained undisputed.The report of the panel ('Ipp Report') 57 prompted a range of legislative reforms on a state by state basis.Some of the changes to the legislation are in the form of restatements or clarification of the common law position; some make wide ranging changes to the law.The report resulted in widespread legislative reform resulting in major limitations on the ability of plaintiffs to recover damages for loss. 58anifestations of a shifting moral trajectory in liability for injury due to alcohol abuse can be seen most obviously in the legislative provisions themselves.Part 6 of the Civil Liability Act 2002 (NSW), for example, deals specifically with intoxication.Section 49(1) states that a person is not owed a duty of care merely because intoxicated, and intoxication does not of itself affect the standard of care.Section 50 applies where a plaintiff in civil proceedings was intoxicated, and where injury or death occurred.The Court is not to award damages to the plaintiff unless the injury would have occurred regardless of the intoxication.These provisions were inserted by the Civil Liability (Personal Responsibility) Act 2002 (NSW), signalling the assertion of principles of moral autonomy.The legislation relevant in Tasmania, the Civil Liability Act 2002 (Tas) contains in s.5 a presumption of contributory negligence where a person is intoxicated.
The dominant theme in the legislation has been that persons should take responsibility for themselves, and if they choose to become intoxicated, and as a consequence are injured, then the loss will be more likely to lie where it falls, rather than be shifted through the agency of the law of negligence.Realistically, this will have the consequence that the loss will remain with the statutory insurer (compulsory vehicle accident insurance schemes in Australian states are typically no-fault schemes) or, where a motor vehicle is not involved, will fall upon the intoxicated injured party.
Conclusions
The tort of negligence has significant residual moral content.Although alternative justifications for the functions of tort law exist -most notably by articulation of the economic costs of alternative risk allocation -when tort allocation mechanisms have appeared to deliver morally disagreeable outcomes the legislature has intervened to require the judiciary to alter the moral trajectory of tort law.In alcohol liability this is manifested in an instruction to take account of the moral autonomy of the intoxicated person by requiring them to take the consequences of their actions.This appears to rebalance the perception that individual autonomy is compromised by allowing the individual to shift responsibility for their voluntary actions to a less morally compromised defendant.This decision may be more consistent with a moral conceptualisation of tort liability than one based on wider social interests, such as the desire to spread risk, or the desire to place the obligation to manage drinkers on the one with the financial interest in doing so.However, the general moral trajectory of tort law simultaneously ascribes moral autonomy to the injured intoxicated, and also allows the financial repercussions of risky behaviour to remain with the risk-taker.
About the Author
Dr. Francine Rochford Dr Francine Rochford belongs to the Bendigo campus of La Trobe University in Victoria, Australia, where she teaches in the Law of Contracts, Law of Torts, Constitutional Law and Business Law subjects.Her research interests include jurisprudence, the law and policy relating to rural water delivery, and the law of education.Previous research relating to jurisprudence has considered Habermas and the discursive resolution of norms.Previous research relating to the law of torts includes papers on the tortious liability of higher education institutions, the tortious liability of water supply authorities for contaminated water, and the standard of care in torts.In 2010 she is a visiting researcher at Lewis and Clark Law School in Oregon.
51 Ibid [85].234 THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES to those injured by the manifestation of those rights.Every freedom/right is balanced by a duty/obligation.
Drinkers, Drivers and Bartenders Chicago, University of Chicago Press, 2000; Clair McIvor, 'Liability in respect of the intoxicated' (2001) 60 Cambridge Law Journal 109; Parliament of Victoria, Drugs and Crime Prevention Committee, Trends in Negligence and Public Liability: The evolving liability of licensees and servers of alcohol to their patrons and third parties Occasional Paper March 2001; Erika Chamberlain, 'Alcohol provider liability in Canada and the United Kingdom: Legal and Cultural Influences' (2004) 33 Arts Law Review 129; Mandy Shircore, 'Drinking, driving and causing injury: the position of the passenger of an intoxicated driver' (2007) 7 QUTLJJ 375; E Chamberlain and R Solomon, 'The role of social host liability in reducing impaired driving and underage drinking in Canada'; Peter Gordon, 'Gearing up for litigation against alcohol manufacturers' presented at the Thinking Drinking -from problems to solutions' Conference, 26th February 2007 [available at www.adf.org.au/downloads/ThinkingDrinkingII_2007/GordonP.pdf;FrankASloan,EmilyMStout,Kathryn Whetten-Goldstein, Lan Liang,22See, for instance, Downunder Rock Café Pty Ltd v Roberts[1991]VicSC 101, where a patron of nightclub injured by falling lighting grid after Tarzan leap by a number of patrons.23As in Wormald v Robertson (1992) Aust Torts Reports 81-180 (Qld CA). | 2019-05-07T12:12:08.912Z | 2010-01-01T00:00:00.000 | {
"year": 2010,
"sha1": "b54ed52b2304cbcc090810a79606f85d4aaa6840",
"oa_license": "CCBY",
"oa_url": "https://figshare.com/articles/journal_contribution/Shifting_Ground_Alcohol_liability_and_the_normative_content_of_tort_law_in_Australia/22200358/1/files/39458323.pdf",
"oa_status": "GREEN",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "b54ed52b2304cbcc090810a79606f85d4aaa6840",
"s2fieldsofstudy": [
"Law"
],
"extfieldsofstudy": [
"Political Science"
]
} |
55354407 | pes2o/s2orc | v3-fos-license | Influence of phenological stage on the antioxidant activity of Thymus zygis s . l . essential oil
When a species has proven antioxidant capacity and possible commercial application it is important to know whether or not this activity varies with the phenological stage of the raw material commodity. The present work aims to assess this topic for the case of Thymus zygis s.l. We analyzed the yield and the chemical composition (gas chromatography-mass spectrometry) of its essential oil in five wild populations from the Southwest of Iberian Peninsula at flowering and fruiting stage. We measured their TEAC (Trolox equivalent antioxidant activity) values. The main components were p-cymene (flowering: 33.6-16.2%; fruiting: 43.2-30.3%), γ-terpinene (flowering: 13.5-5.78%; fruiting: 4.67-1.96%) and thymol (flowering: 46.45-22.2%; fruiting: 54.24-13.9%); the latest, especially high (flowering: 46.45%; fruiting: 54.24%) in one of the populations, identified as Thymus zygis ssp. gracilis (Boiss.) R. Morales, which showed the highest antioxidant activity as well. In the set of samples TEAC oscillated between 16.12-30.52 mmol Trolox L–1 (flowering) and 10.96-40.24 mmol Trolox L–1 (fruiting). No significant differences between antioxidant activity of both stages were found, but yields were significantly higher at flowering time. These levels of antioxidant activity are close to those achieved as highly antioxidant food products, as coffee or wine. This species can be a good raw material for agribusiness use, specially the population of Thymus zygis ssp. gracilis. Additional key words: antioxidant capacity; chemical composition; Iberian Peninsula; Lamiaceae; phytogenetic resources; thyme.
in the fruiting stage.It was dried in an airy room in darkness, and conserved for two months in paper bags.Two extractions of the essential oil were carried out for each sample (hydrodistillation; European Pharmacopoeia, Council of Europe, 1996); one with the whole plant (WP) and another with flowers and leaves (FL).Essential oil yield and its percentage composition were calculated.
We found that the yields of the populations studied oscillated between 0.5 and 1.73% for the whole plant (WP) and between 0.96 and 3.42% for the flowers and leaves (FL) (Table 2).The high yields of this plant have made it one of the most comercial Spanish thymes (Jordán et al., 2009).Our results may suggest that it is much more advantageous the extraction in flowers and leaves than in the whole plant, because yields are significantly higher in the flowering (WP = 0.9-1.73%;FL = 1.43-3.42%)than fruiting (WP = 0.35-0.93%;FL = 0.96-1.92)[Wilcoxon test, (WP), n = 4, *, p = 0.043; (FL), n = 4, *, p = 0.043].
Introduction
Thymus zygis s. l., is a plant endemic to the Iberian Peninsula and northern Morocco, which is a pioneer scrub on all types of soils.Its traditional use is well known in its native range (Vázquez, 2008;Morales, 2010) and its essential oil has been chemically characterized using material from North Africa, Spain and Portugal (Blanco, 2005;Figueiredo et al., 2008).Some authors (Jiménez et al., 1993;Jordán et al., 2009) have pointed out its antioxidant power.This capacity makes it a plant of possible use in the food industry, important to inhibit oxidation, one of the major causes of food spoilage.It prevents rancidity and/or deterioration of the nutritional quality, colour, flavour and texture (Antolovich et al., 2002).
When a species has proven antioxidant capacity and possible commercial application it is important to know whether or not this activity varies with the phenological stage of the raw material commodity.Studies addressing this question have not been made and it is a highly interesting topic from the applied point of view.Thus, the present work aims to assess the influence of the phenological stage on the antioxidant activity of Th. zygis.
Table 2 summarizes chemical composition of essential oils studied.Chemical composition of thyme essential oil may vary throughout the growing season (Blanco et al., 2010).Our results offered only significant differences between the flowering and fruiting season in three of the components identified (Kruskal-Wallis test, p-cymene: n = 10, *, p = 0.028; limonene, n = 10, *, p = 0.015; γ-terpinene: n = 10, **, p = 0.009: see Table 2), without impact on the final antioxidant activity, although two of them did not represent residual amounts in the total mixture.It is evident that the total antioxidant activity is due to the sum of the activities of each component, many of them with demonstrated antioxidant capacity have been tested individually by Dorman et al. (2000a).
Finally, it is also noteworthy that this essential oil has a high proportion of thymol.This fact gives a remarkable quality of the product, given the particular organoleptic characteristics of this phenolic derivative, top rated by industry (Sotomayor, 1998), and much more competitive from a business perspective than others as carvacrol, present in related species but with unpleasant smelling, irritating properties, and a clearly lower value.56, 24.6, 17.21, 22.97, and 24.4 mmol Trolox L -1 .Fruiting stage: 37.69, 15.35, 11.98, 14.49, and 22.81 mmol Trolox L -1 .
Table 1 .
Provenance of the study material, indicating locality (all in Badajoz province), altitude, UTM coordinates, collection date, voucher specimens in the HSS Herbarium (Herbario del Suroeste de España, Badajoz, Spain), and population taxon (subspecies) sylvestris Short communication.Antioxidant activity of Thymus zygis essential oil Even though, for an industrial extraction process, it seems more logical to use the whole plant, because the separation of flowers and leaves consumes many hours of factory work.
Table 2 .
Percentage composition (whole plant, WP) and yield (WP and flowers and leaves, FL) of the essential oil of five Thymus zygis populations (P) 3 RTM = Retention time in GLC-MS (min). | 2018-12-07T21:57:14.730Z | 2012-05-03T00:00:00.000 | {
"year": 2012,
"sha1": "b04d1208908642b41fc0bd91ba429d369d921086",
"oa_license": "CCBY",
"oa_url": "https://revistas.inia.es/index.php/sjar/article/download/2469/1672",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "b04d1208908642b41fc0bd91ba429d369d921086",
"s2fieldsofstudy": [
"Agricultural And Food Sciences"
],
"extfieldsofstudy": [
"Biology"
]
} |
235670195 | pes2o/s2orc | v3-fos-license | Effects of metallicity on mode switching in Cepheids
The mode switching in Cepheids is studied using the methods of the nonlinear theory of stellar pulsation, depending on the main sequence mass $M_0$ and the abundance of elements heavier than helium $Z$. The grid of evolutionary and hydrodynamic models of core--helium burning Cepheids is represented by 30 evolutionary sequences of stars with initial masses $5.7M_\odot\le M_0\le 7.2M_\odot$ and $Z=0.014$, 0.018, 0.022. For considered values of $Z$ the periods of the fundamental mode and the first overtone at the oscillation mode switching are shown to depend on the mean density of the stellar matter. The upper limit of the period of the first overtone decreases witn increasing $Z$ from $\approx 6.9$ day for $Z=0.014$ to $\approx 4.1$ day for $Z=0.022$. The theoretical period--radius relation is independent of $Z$ and agrees well (within 2.5\%) with recent measurements of Cepheid radii based on the Baade--Wesselink method. The fundamental parameters of the short--period Cepheid CG Cas were derived with application of observational estimates of the period and the rate of period change. This star is shown to be the first--overtone pulsator.
. One of the possible solutions based on the Fourier decompostion of the Cepheid light curve was proposed by Antonello and Poretti (1986) but it can only be applied in the case of high precision photometric measurements.
The periods of the fundamental mode Π 0 and the first overtone Π 1 corresponding to oscillation mode switching are described with good accuracy as a function of mean densitȳ ρ = M/( 4 3 πR 3 ), where M and R are the mass and the radius of the Cepheid (Fadeyev 2020). This conclusion is based on results of consistent stellar evolution and nonlinear stellar pulsation calculations for Cepheid models with main sequence masses 5.1M ≤ M 0 ≤ 6.1M and the abundance of elements heavier than helium Z = 0.02. However recent estimates of the galactic abundance gradient obtained from spectroscopic observations of Cepheids are Therefore, the dependencies Π 0 (ρ) and Π 1 (ρ) for other values of Z remain uncertain.
The goal of the present study is to determine the dependencies Π 0 (ρ) and Π 1 (ρ) for Cepheids with metal abundancies Z = 0.014, 0.018 and 0.022. The initial conditions necessary for solution of the equations of radiation hydrodynamics and time-dependent convection describing nonlinear stellar pulsations are obtained from the grid of evolutionary sequences computed from the zero age main sequence to the stage of helium exhaustion in the stellar core. Evolutionary sequences were computed for stars with initial masses 5.7M ≤ M 0 ≤ 7.2M with the mass step ∆M 0 = 0.1M . The initial abundance of helium is assumed to be Y = 0.28.
evolutionary sequences of cepheids
As in the previous work of the author (Fadeyev 2020) the initial conditions for equations of radiation hydrodynamics describing stellar pulsations were determined from selected stellar models of evolutionary sequences computed with the program MESA version 12778 (Paxton et al. 2019). Convective mixing was treated according to the standard theory (Böhm-Vitense 1958) with mixing length to pressure scale height ratio α MLT = Λ/H P = 1.6. Extended convective mixing (overshooting) at convective boundaries was calculated using the prescription of Herwig (2000) with parameters f = 0.016 and f 0 = 0.004. Effects of semiconvection were taken into account according to Langer et al. (1985) with efficiency parameter α sc = 0.1 in the expression for the diffusion coefficient D sc .
To inhibit irregular growth of the convective core accompanied by appearance of spurious loops in the HR-diagram the evolutionary calculations of the core helium burning stage were carried out with option 'conv_premix_avoid_increase'. The sufficiently small amplitude of central helium abundance jumps was reached owing to significant increase of the mass zone number of the evolutionary model (N ∼ 4 × 10 4 ) together with reduction of the time step up to ∆t ev = 10 3 yr.
The rate equations for nucleosynthesis were solved for the reaction network consisting of 26 isotopes from hydrogen 1 H to magnesium 24 Mg with 81 reactions. Calculation of reaction rates was carried out using the JINA Reaclib data (Cyburt et al. 2010). Mass loss due to the stellar wind was computed by the Reimers (1975) formula with parameter η R = 0.3.
The role of metallicity in evolution of Cepheids is illustrated in Fig. 1 In Cepheid models considered in the present study the radius of the node of the first overtone is r n ≈ 0.80R, where R is the radius of the upper boundary of the hydrostatically equilibrium model. The necessary condition for pulsations in the first overtone is that the helium ionizing zone driving pulsation instability should locate in layers with radal distance r > r n . This condition is not fulfilled for models of the evolutionary sequence Z = 0.022, M 0 = 6.2M during the third crossing of the instability strip so that the Cepheid evolves between the blue and red edges of the instability strip as the fundamental mode pulsator (see Fig. 1).
results of stellar pulsation calculations
In this study we computed 30 evolutionary sequences for stars with initial metal abun- The main goal of this work is to determine the periods of the fundamental mode Π 0 and the first overtone Π 1 at the pulsation mode switching for evolutionary sequences with specified values of Z and M 0 . Results of calculations are shown in Fig. 3, where periods Π 0 and Π 1 are plotted as a function of the mean stellar densityρ. As is seen, these dependencies are independent of Z within the margin of error and are given by relations The plots in Fig. 3 illustrate the fact that periods of the fundamental mode and the first overtone at the mode switching increase with increasing mass and increasing radius of the Cepheid (i.e. with decreasingρ) irrespective of Z. Decrease of the stellar mean density is accompanied by extension of the helium ionizing zone so that pulsations in the first overtone become impossible once the lower boundary of the ionizing zone reaches the node of the first overtone. As seen in Fig. 3, the interval of existence of oscillations in the first overtone (i.e. with period Π 1 < Π 1 ) reduces with increasing Z.
For the sake of clarity, properties of the models at the upper limit of existence of the first overtone oscillations are listed in Table 1. First three columns of Table 1 give the metal abundandance Z, initial stellar mass M 0 and the number of the instability strip crossing i. (Fadeyev 2013). Together with period of the principal mode we evaluated also the period of the secondary mode. In the fundamental mode pulsator the role of the secondary mode is played by the first overtone, whereas in the case of the first overtone pulsator the role of the secondary mode belongs to the fundamental mode. The period ratio of the fundamental mode and the first overtone was found to be Π 0 /Π 1 ≈ 1.50 and independent of the locus with respect to the mode switching.
The period-radius relations obtained from the present evolutionary and hydrodynamic calculations are shown in Fig. 4. As in Fig. 3, the plots in Fig. 4 do not show any dependence on Z, so that the periods and radii locate near the straigth lines f 0 and h 1 which are given by and where the pulsation period is expressed in days.
For hydrodynamic models considered in the present study the lower limit of the fundamental mode period is Π 0 = 5.59 day, whereas the upper limit of the first overtone is Π 1 = 6.57 day.
In the range of these two values the fundamental mode to first overtone period ratio evaluated from (3) and (4) derived by Lazovik and Rastorguev (2020). In Fig. 4
relation (5) is shown by the dotted line.
For fundamental mode pulsators with period Π = 5.6 day the difference between (3) and (5) does not exceed ≈ 2.6%, whereas for periods Π ≈ 15 day the difference between the theory and observations becomes as small as ≈ 0.2%, so that the theoretical and empirical plots become almost the same.
fundamental parameters of the cepheid cg cas The O − C diagram of CG Cas spans somewhat more than a hundred years and is fitted by the parabolic dependence indicating the secular period growth with rateΠ = 0.170 s/yr (Turner et al. 2008). Thus, there is an opportunity to evaluate the fundamental parameters of this star on the basis of consistent stellar evolution and nonlinear stellar pulsation calculations. However, the earlier attempt to determine the fundamental parameters of CG Cas for the evolutionary sequences with metal abundance Z = 0.02 failed because the upper limit of the first overtone period was found to be shorter than the period of CG Cas (Fadeyev 2020).
As seen in Fig. 3 conclusions Dependence of the stellar opacity on Z significantly affects the gas density in the stellar envelope and location of the pulsation driving zone with respect to the node of the first overtone.
Increase of the mass and the radius of the Cepheid is accompanied by decreasing mean density and increasing width of the helium ionizing zone responsible for excitation of oscillations. Pulsations in the first overtone become impossible when the inner boundary of the helium ionizing zone reaches the node of the first overtone during pulsation motions. Increase of periods Π 0 and Π 1 corresponding to the mode switching is due to decrease of the gas density in the envelope of the pulsating star. Existence of the upper limit Π 1 implies that the necessary condition for oscillations in the first overtone is not fulfilled for Π 1 > Π 1 .
The Extention of the Cepheid grid to lower Z will allow us to consider the stars with galactocentric distances greater that the solar galactocentric distance and may lead to somewhat larger values of Π 1 . Enlargerement of Π 1 due to decrease of Z can be evaluated from additional evolutionary and hydrodynamic computations. references | 2021-06-30T01:16:26.492Z | 2021-06-01T00:00:00.000 | {
"year": 2021,
"sha1": "fd52be7c21417dadc56ea9f44f9f03d74af3a63a",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/2106.15383",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "fd52be7c21417dadc56ea9f44f9f03d74af3a63a",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
226965958 | pes2o/s2orc | v3-fos-license | Conservation analysis of SARS-CoV-2 spike suggests complicated viral adaptation history from bat to human
Abstract Background The current coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome (SARS)-CoV-2, has become the most devastating public health emergency in the 21st century and one of the most influential plagues in history. Studies on the origin of SARS-CoV-2 have generally agreed that the virus probably comes from bat, closely related to a bat CoV named BCoV-RaTG13 taken from horseshoe bat (Rhinolophus affinis), with Malayan pangolin (Manis javanica) being a plausible intermediate host. However, due to the relatively low number of SARS-CoV-2-related strains available in public domain, the evolutionary history remains unclear. Methodology Nine hundred ninety-five coronavirus sequences from NCBI Genbank and GISAID were obtained and multiple sequence alignment was carried out to categorize SARS-CoV-2 related groups. Spike sequences were analyzed using similarity analysis and conservation analyses. Mutation analysis was used to identify variations within receptor-binding domain (RBD) in spike for SARS-CoV-2-related strains. Results We identified a family of SARS-CoV-2-related strains, including the closest relatives, bat CoV RaTG13 and pangolin CoV strains. Sequence similarity analysis and conservation analysis on spike sequence identified that N-terminal domain, RBD and S2 subunit display different degrees of conservation with several coronavirus strains. Mutation analysis on contact sites in SARS-CoV-2 RBD reveals that human-susceptibility probably emerges in pangolin. Conclusion and implication We conclude that the spike sequence of SARS-CoV-2 is the result of multiple recombination events during its transmission from bat to human, and we propose a framework of evolutionary history that resolve the relationship of BCoV-RaTG13 and pangolin coronaviruses with SARS-CoV-2. Lay Summary This study analyses whole-genome and spike sequences of coronavirus from NCBI using phylogenetic and conservation analyses to reconstruct the evolutionary history of severe acute respiratory syndrome (SARS)-CoV-2 and proposes an evolutionary history of spike in the progenitors of SARS-CoV-2 from bat to human through mammal hosts before they recombine into the current form.
INTRODUCTION
A vigorous outbreak of a novel coronavirus in Wuhan, China, in early 2020 has brought a new wave of concern on the potential risk of coronavirus virulence. Coronaviruses are single-stranded RNA viruses that are capable of imposing health threats to human. Among four genera of coronavirus, namely, alpha-, beta-, gamma and deltacoronavirus, betacoronavirus has the highest medical significance, as it is responsible for not only diseases in those with mild symptoms but also some of the most serious epidemics since the beginning of this century: severe acute respiratory syndrome (SARS) caused by SARS-CoV in 2003, middle-east respiratory syndrome (MERS) caused by MERS-CoV in 2012, and the recent COVID-19 outbreak caused by SARS-CoV-2 in early 2020 [1]. On the other hand, coronaviruses such as human coronavirus (HCoV) OC43 and HKU1, along with alphacoronaviruses HCoV-NL63 and HCoV-229E, cause mild common cold.
The adaptability of coronavirus in mammals and birds gives rise to its diversity and potential virulence to human. Coronaviruses have a wide range of hosts, as suggested by an examination on the deposit in NCBI Genbank that features coronavirus sequences extracted from mammalian animals, such as feline, canine, camels, rodents, swine, cattle, bats, as well as birds. Human coronaviruses have their origins mostly from bats or rodents. Rodent-originated human coronaviruses include HCoV OC43 and HKU1 [2], while others, including HCoV NL63, 229E, SARS-CoV and MERS-CoV, are probably from bat [3]. Bat coronavirus usually passes to human via an intermediate host, as observed in the outbreaks of SARS and MERS in which civet and camel are found to be the respective intermediate hosts [3]. During cross-species transmission, the virus needs to adapt the new host such that it can infect and replicate in the new environment.
There have been numerous studies on the origin and possible transmission pathway of SARS-CoV-2. Bat coronavirus (BCoV) RaTG13 from Rhinolophus affinis bat, sharing 96% identity with SARS-CoV-2 genome, is a frequently cited candidate of direct progenitor, while other studies identify pangolin coronavirus (PCoV) as direct progenitor based on the similarity to the receptor-binding domain (RBD) in coronavirus strains in pangolins from wildlife smuggling [4][5][6][7]. In general, phylogenetic studies use full-genome sequence, spike and Rdrp (RNA-dependent RNA polymerase) to reconstruct phylogenetic relationship between different CoV strains, especially Rdrp based on its vital importance in replication and thus conservation across species. Nonetheless, the spike gene should also be given more attention, given the crucial role it plays on viral infection.
Viral entry of coronavirus is mediated by a series of molecular interactions between viral surface proteins and host cell receptors. Spike glycoprotein is the major surface protein responsible for host recognition. Spike protein is a type I viral fusion protein, which is primarily characterized by the perpendicular projection of fusion proteins on viral membrane, having an RBD near the N terminus of primary sequence and alpha-helical coiled coil as major secondary structure of fusion subunit [8]. Table 1 summarizes receptor specificities for noted coronaviruses [9,10]. Mature spike monomers assemble into homotrimers on viral membrane surface that gives the distinctive crown-like structure.
Studies have shown that spike RBD itself carries all necessary information for receptor binding in a host-specific manner, as recombinant bat SARSr-CoV Rp3 with replaced RBD gains ability to enter human ACE2-expressing cells [11,12]. Given the role of spike protein in coronavirus infection, the analysis of the sequences of spike proteins and the characterization of their evolutionary changes help reveal the history of the recent SARS-CoV-2 and identify the source of infectability. In this study, we carried out phylogenetic and sequence analyses to characterize the nucleotide and amino acid sequences of spike proteins in SARS-CoV and SARS-CoV-2 and related CoV strains from public databases, such that we may get insights on the development of human-susceptible coronaviruses from the view of the adaptation of spikes.
MATERIALS AND METHODS
Viral sequences were downloaded primarily from NCBI Genbank, with pangolin coronavirus genomes obtained from GISAID EpiCoV database by 25 March 2020. The accession numbers of the dataset used can be found in Supplementary Data 1 and 2. Annotated nucleotide sequences of spike were extracted from Genbank feature annotations and were translated into amino acid sequences using BioPython and validated with Genbank record. Multiple sequence alignment was done by MAFFT with default settings except that two guiding trees were built [13]. Phylogenetic tree was constructed using neighbor-joining method, maximum composite likelihood model with Gamma-distributed site rate and 500 bootstrap replications in MEGA version 10.1.7.
Sequence conservation analysis was performed using R package RPHAST 1.6.9 [14]. Nucleotide sequences of spike from 13 CoV strains were selected to generate a comparison of CoV strains between SARS-CoV clade and SARS-CoV-2 clade. Spikes of SARS-CoV-2 were aligned with each of the remaining 12 CoV with MAFFT in setting described above. A neutral model was computed using Jukes-Cantor model (JC69) as a substitution model. The expected length of conserved element was set as 500 and the target coverage as 0.99 to accommodate this dataset. Plots were generated with RPHAST utilities.
Codon adaptation index was calculated according to Sharp and Li [15] in R, using human ribosomal protein as the reference codon set [16]. The relative adaptiveness of the i th synonymous codon of the j th amino acid, denoted as w i , is defined as where X ij is the observed count for the i th synonymous codon of the j th amino acid and maxðX j Þ is the count of the most abundant codon of the j th amino acid. A minimum of w ¼ 0:01 was assigned to codons that were not observed in the reference set. The codon adaptation index (CAI) of a given coding domain sequence (CDS), partitioned as a series of L codons with relative adaptiveness scores w 1 , w 2 , w 3 , . . ., w L , is thus defined as the geometric mean of all scores.
Codon counts of 1984 human ribosomal protein CDS were obtained from codon usage database of Kazusa Institute (https://www.kazusa.or.jp/codon/). Relative adaptiveness for each codon can be found in Supplementary Data 3. The calculation was done using in-house R script.
Sequence similarity analysis was done with Simplot version 3.5.1 [17]. Spike nucleotide sequence from AY572034 for civet CoV, DQ022305 for BCoV-HKU3, EPI_ISL_410539 for PCoV-17, EPI_ISL_410721 for PCoV-19, AY345986 for SARS-CoV, MN996532 for BCoV-RaTG13, MG772933 for BCoV-ZC45 and MG772934 for BCoVZXC21 were used to compute sequence similarity against MN996527 for SARS-CoV-2, using a window size of 200 bp, a step size of 20 bp and the Kimura distance model. Bootscan analysis employed the same window size and step size, in addition to neighbor-joining tree with 500 bootstrap replications.
Variant analysis of receptor binding domain was done using a Hamming distance-based clustering approach. We referred to currently available studies and identified 24 RBD binding site residues in SARS-CoV-2, which are either involved in direct contact or in proximal distance to ACE2 receptor during the process of host recognition [18][19][20]. Multiple sequence alignment was then performed on the 13 CoV strains. Homologous residues as well as the corresponding codons in the 12 CoV strains from SARS-CoV-2 clade and SARS-CoV clade were identified by manual inspection. The sum of Hamming distance of codons was calculated for each pair of CoV strains, then hierarchical clustering was performed and visualized with heatmap and principle coordinate analysis.
RESULTS
Whole-genome phylogenetics reveals a young clade of Sarbecovirus strains.
Phylogenetic analysis using the whole genome of 995 coronaviruses categorizes them into alphacoronavirus and betacoronavirus ( Fig. 1). HCoV 229E and HCoV NL63 with their respective related viruses share much closer relationship between each other than those in beta-CoV clade, as shown by the close branching points and short branch lengths [21]. In contrast, beta-CoV viruses demonstrates higher variability between each other. From the most recent common ancestor of beta- CoV in this comparison, the first bifurcation produces the last common ancestor (LCA) of lineage A (Embecovirus) and the LCA of lineages B and C. LCA of Embecovirus further evolves and bifurcates into HCoV-HKU1 and HCoV-OC43, respectively [21,22]. LCA of lineage B-C undergoes another bifurcation event that results in subgenera Merbecovirus (MERS-CoV and related strains) and Sarbecovirus (SARS-like CoV) [22]. MERS-CoV deviates in a greater extent from the LCA of Merbe-Sarbecovirus and yet it presents high similarity within the clade. Finally, SARS-CoV and SARS-CoV-2 with their respective related CoV in animal hosts branch away from their LCA in relatively recent time, as suggested by their proximity to each other at the tip. Further inspection on the Sarbecovirus clade shows detailed splitting patterns among closely related CoV. A smaller phylogenetic tree is constructed with only MERS-CoV, SARS-CoV and SARS-CoV-2 with related species, with MERS-CoV set as an outgroup ( Fig. 2A, outgroup not shown). From the phylogram two main clades can be identified, representing the SARS-CoV lineage (b) and SARS-CoV-2 lineage (a). In the SARS-CoV-2 group, bat SARS-like CoV strains ZC45 and ZXC21 are probably the earliest to separate from LCA and presumably have the highest similarity to LCA [23]. Two subsequent bifurcation events follow, which gives rise to two groups of pangolin-CoV (PCoV). The earlier one (c) produces PCoV from Guangxi, China, and the later one (d) produces PCoV from Guangdong, China. The final bifurcation (e) yields bat-CoV RaTG13, as agreed by numerous studies to be the closest relative of SARS-CoV-2 [7,24,25]. The major branch bifurcations in SARS-CoV-2 clade have branch support values higher than 99%, suggesting a highly confident splitting pattern for the major subclades within SARS-CoV-2-related strains. On the other side, SARS-CoV is related to a larger number of bat SARS-like CoVs in this dataset, with smaller differences to each other suggested by shorter branch lengths. Civet is the most probable intermediate host of SARS-CoV passing from bat to human. Within the SARS-CoV lineage, civet SARS-CoV is nearly identical to SARS-CoV, leaving all bat CoVs as the outgroup. However, in the case of SARS-CoV-2, instead of PCoV being the closest relative, it demonstrates that bat CoV RaTG13 has out-competed all PCoVs and become the closest relative to SARS-CoV-2. Table 2 summarizes the collection dates, submission dates and first publication of coronavirus strains in SARS-CoV-2 clade.
SEQUENCE ANALYSIS ON SPIKE
The S gene encodes the spike protein responsible for host recognition, this suggests that its nucleotide and amino acid sequences, implicating its three-dimensional structure, is species-specific and possibly serves as a better candidate in characterizing the adaptation of CoV in different hosts. We extracted nucleotide sequences of the spike from SARS-CoV-2, SARS-CoV, MERS-CoV and bat SARS-like CoV genomes. Since pangolin CoV does not have predefined gene feature annotation, we performed multiple sequence alignment with SARS-CoV-2 and identified the homologous spike regions in pangolin CoV genomes, then translated the identified CDS. We observe a slightly different branching patterns for SARS-CoV-2-related lineage on nucleotide and amino acid sequences. When only spike sequences are considered, in both cases of nucleotide and amino acid, BCoV-RaTG13 has the highest similarity to SARS-CoV-2, a consistent observation with the whole-genome phylogenetic analysis. However, the linking relationships of PCoV-17 and PCoV-19 to SARS-CoV-2 are different between the nucleotide and amino acid trees. For nucleotide sequence, bat-CoV-RaTG13 and SARS-CoV-2 share an LCA, which then shares another LCA with PCoV-19 ( Fig. 2C), and this is the same pattern observed in the whole-genome comparison, with branch support value of 98%. On the contrary, in the case of amino acid sequence, the closest PCoV to SARS-CoV-2-BCoV-RaTG13 clade is PCoV-17 (Fig. 2B), though this bifurcation has a much lower branch support value of 83%, which suggests a less robust splitting (<90%). Since its parent branch point and child branch points have >90% branch support values, we still consider the branching between PCoV-17 and SARS-CoV-2-BCoV-RaTG13 clade as credible. In addition, we confirm the insertion of tetrapeptide 681 PRRA 684 right before the putative S1/S2 cleavage site, as previously reported, to be SARS-CoV-2-exclusive, which is caused by the insertion of 23568 CTCCTCGGCGGG 23579 that produces an synonymous TCA->TCT mutation of S 680 followed by 681 PRRA 684 [26]. [24,25]. This suggests a recombination event that alters the composition of amino acid sequence roughly between Y 421 to P 527 , covering more than half of the RBD and the full extent of receptor binding motif (RBM) [27]. The Bootscan plot provides a clearer representation of possible recombination events (Fig. 3B). The region between 1260 and 1580 bp has the highest probability of being originated from PCoV-19. Similar but weaker signals can also be observed in other loci, such as the peaks around 1720, 2520 and 3580 bp.
CONSERVATION ANALYSIS ON SPIKE
In addition, we also observe that BCoV-ZXC21 and BCoV-ZC45, though being classified into the SARS-CoV-2 clade, have strikingly low average similarity in this region (43.13% and 48.38%, respectively), even lower than SARS-CoV and civet-CoV in SARS-CoV clade (67.5% and 67.75%, respectively). Thus, this short sequence may come from an LCA with SARS-CoV in more recent time before it returns to the SARS-CoV-2 clade by homologous recombination.
A more sophisticated conservation analysis was performed to examine these claims. Nucleotide sequences of spike of the 12 CoV strains from SARS-CoV and SARS-CoV-2 clade were aligned with SARS-CoV-2 spike sequence, and conservation scores were calculated and visualized along the alignment to demonstrate the degree of conservation between SARS-CoV-2 and the query sequences. In parallel, conservation scores were also calculated for whole genome alignment (Fig. 3C). The whole-genome conservation analysis demonstrates that the Sarbecovirus species are well conserved in general, especially at the region between 10 000 and 21 000 bp, which corresponds to the end of Orf1a and the full extent of Orf1b in both SARS-CoV-2 clade and SARS-CoV clade species. For SARS-CoV-2 clade, the complete genome sequence is generally conserved except for the S locus, revealing the highly variable nature of spike sequence for host adaptation.
For spike conservation, the results agree with sequence similarity analysis about the close relationship between BCoV-RaTG13 and SARS-CoV-2, with a gap spanning around 1500 bp, which reflects the recombination event mentioned previously. Conserved element plot demonstrates the coherency of the phylogenetic reconstruction of SARS-CoV-2 clade CoV (Fig. 3D). Nucleotide sequence downstream of around 2300 bp, corresponding to the S2 subunit, is largely conserved among SARS-CoV-2 clade species. The extent of conservation in S1 subunit is then observed to be correlated to the phylogenetic relationship. It is worthy to mention that the first 1000 bp of SARS-CoV-2 spike sequence, corresponding to the N-terminal domain (NTD) region, has low conservation scores when compared with PCoV-19, but relatively higher with PCoV-17. This observation can be reaffirmed by sequence similarity analysis, in which PCoV-17 rather than PCoV-19 has higher similarity next to BCoV-RaTG13 in the NTD (76.32% vs 51.10%). Since it is not evident that the spike NTD has undergone homologous recombination with PCoV-17, we propose that the similarity is a result of cumulative mutations in the course of natural selection. In summary, there is a higher probability that SARS-CoV-2 and BCoV-RaTG13 share the most recent LCA than any other CoV strains in SARS-CoV-2 clade in our current study. The NTD of spike may have evolved separately from the latter part of the gene body, then recombined with the LCA of BCoV-RaTG13 and SARS-CoV-2.
CODON ADAPTATION IN SPIKE
CAI is a widely used metric to infer the overall adaptiveness of a coding sequence given a reference set of codon usage [15]. The CAI of a coding sequence is defined as the geometric mean of relative adaptiveness of codons, which is in turns defined as the frequency of that codon divided by the frequency of the most abundant codon of the same amino acid in a reference set. The reference codon usage is usually derived from perpetually translated proteins, which reflects the optimal tRNA abundance of the species. As viral translation depends on host cell machineries, human ribosomal proteins are used as the reference. arranged according to collection time, plots are created to demonstrate the difference between CAI for spike and CAI for orf1ab. Lower: Like the middle plot, but CAI for each year is averaged as geometric mean time. For the spike, the CAI scores of SARS-CoV, bat SARSr-CoV and MERS-CoV are the highest among all and occupy the top region in the plot, while the CAI scores of the four HCoV species occupy a lower region (Fig. 4, middle and lower). The CAI scores of SARS-CoV-2 and pangolin CoV strains sit between these two groups, closer to HCoV and overlapping with HCoV 229E. For the CAI scores of orf1ab, a similar pattern for SARS-CoV and MERS-CoV can be observed. One important feature of these plots of CAI scores against time is that the scores remain constant regardless of the type of CDS (non-structural/ structural) and species. Considering that all CAI scores were calculated using human reference, it suggests that for either SARS-CoV-2 or SARS-CoV and their related strains, a change in the host does not affect the range of CAI. Instead, we observe that the coronavirus strains with the highest CAI scores are those resulting in higher mortality rate. The CoV strains causing common cold have lower CAI scores, SARS-CoV and MERS-CoV have higher, and SARS-CoV-2 sits in the middle of them.
In summary, this result suggests that SARS-CoV-2 has lower usage of the most abundant codons in human than SARS-CoV. The CAI scores do not appear to increase or decrease significantly through time, implying that the optimization of codon usage in a new host is not a main driving force of the evolution of the coronaviruses, and also not a significant factor for the prevalence of the coronaviruses. The result implies that CAI scores are more correlated to the phylogenetic grouping of the coronaviruses (Table 3).
DISTANCE ANALYSIS ON RBD BINDING SITES
Twenty-four sites on the RBD of SARS-CoV-2 are identified from literature and they are compared across closely related CoV as well as SARS-CoV and SARS-like CoV (Tables 4 and 5) [18][19][20]. We observe that only 8 out of 24 residues (33.3%) are identical between SARS-CoV and SARS-CoV-2, which is lower than that in bat-CoV-RaTG13 (15, 62.5%), PCoV-17 (16,66.7%) and PCoV-19 (22,91.7%). On the other hand, SARS-CoV and civet SARS-CoV have 20 (83.3%) identical residues, which is the same as some other CoVs, such as bat-SL-CoV-Rs4874 and WIV16 and WIV1 (20, 83.3%), and is distant from the remaining related CoVs like bat-SL-CoV-Rs4231 (11, 45.8%) and HKU3 (5, 20.8%). Considering the codons encoding the residues, we use the codons at each site in each species to compute a distance matrix, in which the distance represents the Hamming distance of all codons in one species to all codons in another species (Fig. 5A). Three clusters can be observed in the heatmap, one for SARS-CoV-2 and pangolin-CoV-2019, one for SARS-CoV, civet SARS-CoV, bat SARS-like CoVs WIV1, WIV16 and Rs4874, Table 3. The number of sequences being assessed (N), mean, standard deviation and standard error of codon adaptation index in 7 CoV strains for spike and orf1ab coding domain sequences. . It suggests that considering only receptor binding sites in RBD, pangolin-CoV-19 has the largest possibility to be able to recognize human ACE2 as well as SARS-CoV-2. As previously mentioned, SARS-CoV-2 has largely conserved spike sequence with BCoV-RaTG13 except for RBD, which probably comes from PCoV-19. Moreover, bat-SL-ZC45 and ZXC21, which are always grouped together with SARS-CoV-2 in contiguous spike sequences, are now displaying a different clustering pattern, that they cluster with BCoV-HKU3 in SARS-CoV clade. In reference to the hierarchical clustering in Fig. 5A, we propose that BCoV ZXC21, ZC45 and HKU3 may be related to the early variant of humaninsusceptible bat CoV derived from the LCA of all Sarbecovirus. Since SARS-CoV and SARS-CoV-2, both recognizing human ACE2, are clustered into a larger subgroup while having distant relationship within the subgroup, it implies that the humansusceptibility in the two CoVs evolves independently, thus having a larger edit distance with each other. Nonetheless, it may be speculated that any of the CoV strains within the subgroup will be capable of evolving into human-susceptible RBD binding sites.
DISCUSSION
We have shown with phylogenetic analyses different clustering patterns of SARS-CoV-2-related species using either whole genome or spike sequences. The relationship between the CoV strains in SARS-CoV-2 clade is a disputable matter. Jaimes et al. [26] produced a tree using SARS-CoV-2 spike protein sequences, it agrees with the bifurcation order of BCoV ZC45, ZXC21, and RaTG13 with SARS-CoV-2 but they did not include pangolin CoV in their tree. One study used the concatenated genomic sequence of ORF1ab, S, E, M and N genes to produce a tree, which included all pangolin CoVs used in this study, and it gives the same branching pattern in SARS-CoV-2 clade as our tree Bat-SL-CoV
SARS-CoV
Civet-SARS-CoV Table 5. Table of TTT TAT TAT --TAT TAT TAT TAT TAT TAT -1357 TAT TAC TAT TAT TAT TAC TAT TAT TAT TAT TAT TAT TAC 1363 TTG CTC TTA TTG TCT TCT TAT TAT TGG TCT TCT TCT TCT 1366 TTT TTT TTT TTT CAC TAC TAC TAT TAC TAC TAC CAC TAC TAC TAC TAT using unedited whole-genome sequence [28]. Moreover, Lam et al. compared the phylogenetic trees using two regions (Regions 4 and 5) covering the end of ORF1ab and a portion of S1 subunit of the spike. Both trees agree that BCoV-RaTG13 has the highest similarity to SARS-CoV-2 in the compared regions. However, the tree for region 5, corresponding to the RBD region, demonstrated that bat CoV ZXC21 and ZC45 group with bat SARS CoV HKU3 and Rf1, both belonging to SARS-CoVrelated clade as shown in this study. This clustering pattern can also be observed in our mutation analysis of RBD, we thus reaffirm that this region is highly sensitive to host adaptation and that the RBD in BCoV-ZXC21 and ZC45 is highly similar to the early form of Sarbecovirus RBD. Our mutation analysis on the RBD region focuses exclusively on the contact sites of SARS-CoV-2 spike protein with the human ACE2 protein. The result demonstrates that pangolin CoV, rather than bat CoV RaTG13, has the highest similarity to SARS-CoV-2. In comparison, Lam et al. [28] have also produced two similar trees with SARS-CoV-2 related CoVs using nucleotide sequences of RBD with lengths at around 570 bases (i.e. 190 residues), one tree for full sequence and one tree for synonymous sites only. Neither tree has the exact branching as our tree, but it also demonstrated that SARS-CoV-2 and PCoV-19 has the highest similarity when compared using all bases in the RBD sequence. The authors suggested that the similarity of RBD between pangolin CoV and SARS-CoV-2 is caused by convergent evolution rather than recombination. The tree with synonymous sites represents a neutral selection model without recombination, and it shows that the spike of SARS-CoV-2 has the closest relationship to the one in bat CoV RaTG13. Nonetheless, the result of our RBD analysis demonstrates that the evolution of spike to become human-susceptible probably occurs inside an intermediate pangolin host.
The phylogenetics of both whole genome and spike sequences suggest an apparently direct passage of SARS-CoV-2 from bat to human, but the distance analysis on RBD rejects this pattern. Based on the analysis of spike sequences, it is reasonable to believe that pangolins or related mammals are the intermediate host, but the transmission takes more complicated steps. The existence of a conserved element in the NTD in BCoV- RaTG13 and SARS-CoV-2 but not elsewhere implies independent lineages of pangolin CoV and SARS-CoV-2/BCoV-RaTG13 since early time. The pangolin lineage may be originated from BCoV-ZXC21/ZC45 via cross-species transmission and it becomes a sister lineage of SARS-CoV-2/BCoV-RaTG13. The LCA of pangolin CoV may have bat-like RBD, but subsequently a pangolin-specific RBD will be developed. The NTD in BCoV-RaTG13 may have resulted from a recombination with an undocumented bat Sarbecoviruses such that it does not cluster with BCoV-ZXC21/ZC45. PCoV-17 may have emerged earlier than PCoV-19, then undergone a recombination to obtain NTD from BCoV-RaTG13 lineage through an intermediate bat CoV (A, Fig. 6) with modified NTD sequences. This would explain the similarity in NTD between BCoV-RaTG13, PCoV-17 and SARS-CoV-2, while keeping PCoV-17 in pangolin lineage. In the final stage, BCoV-RaTG13, which provides most of the spike sequence except the RBD, must recombine with the RBD of PCoV-19. The phylogenetics of NTD, RBD and S2 subunit imply that the recombination only replaces RBD from BCoV-RaTG13 spike sequence. The final intermediate host is not necessarily pangolins, but they would be susceptible to both the infection of BCoV-RaTG13 and PCoV-19. This hypothesis assumes that (i) CoV transmits from bats to intermediate mammals, and then to human in a single-way fashion, i.e. without backward transmission and (ii) cross-species transmission does not significantly alter the spike sequence except for the RBD region, and any substantial changes in sequence other than RBD region are the result of recombination. This hypothesis predicts the existence of a bat SARS-like CoV of which the NTD of spike will cluster with BCoV-RaTG13 better than with SARS-CoV-2, while in whole-genome phylogenetics it will be the sister species of BCoV-ZXC21/ZC45. Further investigation and analyses will be required to test the hypothesis.
acknowledgements
The authors would like to thank Professors Ge and Deng for their managerial support and thank ICTO for their support in computing facility in University of Macau.
supplementary data
Supplementary data is available at EMPH online. | 2020-10-29T09:04:23.299Z | 2020-11-05T00:00:00.000 | {
"year": 2020,
"sha1": "308901dd0fc4a080acbaba86daac362856b3a869",
"oa_license": "CCBY",
"oa_url": "https://academic.oup.com/emph/article-pdf/2020/1/290/34824330/eoaa041.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "28920a1cc667d3eab1eeb9169af134de3ea9d4ca",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
258968939 | pes2o/s2orc | v3-fos-license | Production performance and rumen bacterial community structure of Hu sheep fed fermented spent mushroom substrate from Pleurotus eryngii
This study aimed to investigate the effect of fermented spent mushroom substrate from Pleurotus eryngii (SMPE) supplementation on production performance, meat quality and rumen bacterial community structure of Hu sheep. 120 2-month-old Hu sheep with average body weight [(13.50 ± 3.10) kg] were selected and randomly divided into 4 groups with 3 replicates per group and 10 sheep per replicate. The control group (RL1) was fed a total mixed ration (TMR), and group RL2, RL3 and RL4 were fed the basal diets supplemented with 15%, 30% and 45% fermented SMPE, respectively. The pretest period lasted for 10 days and the test period lasted for 150 days. The results showed that: (1) Difference (p < 0.05) was observed in average daily feed intake (ADFI) and feed conversion ratio (FCR) between RL2 and RL4 groups. The eye muscle area (EMA) and grade rule (GR) values in RL2 and RL3 were significantly higher than those in RL1 and RL4 groups (p < 0.05). (2) The contents of threonine, valerine, leucine, lysine, histidine, essential amino acids, flavor amino acids, aspartic acid, serine, glutamic acid and arginine of the longissimus dorsi muscle in RL2 and RL3 groups were significantly higher than RL1 and RL4 (p < 0.05). (3) A total of 1,202,445 valid sequences were obtained from rumen of Hu sheep fed different amounts of fermented feed, and the valid sequences were clustered into 9824 Operational Taxonomic Units (OTUs). (4) α diversity analysis showed that the richness and diversity of rumen bacterial communities in Hu sheep in RL1, RL2, RL3 and RL4 groups were significantly higher than RL0 (raw materials of fermented SMPE) group (p < 0.05). β diversity analysis showed that the bacterial community structure was the most different between RL0 and RL3. (5) At the genus level, compared with RL1, the relative abundance of Christensenellaceae R-7 in RL3 group decreased significantly by 33.59%, the relative abundance of Prevotellaceae UCG001 in RL2, RL3 and RL4 decreased significantly by 50.41%, 62.24% and 49.17%, respectively, and the relative abundance of Ruminococcaceae NK4A214 in RL2 group increased significantly by 35.01% (p < 0.05). In summary, the addition of fermented SMPE to TMR can significantly improve the production performance, meat quality and rumen bacterial community diversity and abundance of Hu sheep.
SMPE is the remaining media residue after harvesting, which has a wide range of sources and low price. SMPE has shown great potential as animal feed, its protein content and amino acid composition are close to corn, and the crude fiber content is close to roughage, which is an excellent "neutral" feed 4 . Studies have shown that spent mushroom substrate is rich in crude fiber, crude protein, polysaccharides, crude fat, calcium, phosphorus and other active nutrients, and is a new type of high-quality feed raw material for livestock and poultry 5,6 . However, spent mushroom substrate is not fully utilized in the actual production process at present, only a very small part is developed for animal feed, most of them are directly incinerated or discarded as waste, which not only wastes biological resources but also causes serious environmental pollution problems. Therefore, the rational development utilization technology of spent mushroom substrate as feed can not only turn waste into treasure and protect the environment, but also alleviate the problem of shortage of feed resources and improve the economic benefits of animal husbandry industry, which has important ecological significance and broad market prospects.
The main components of SMPE are agricultural byproducts of high fibre raw materials such as bagasse, wheat bran, corn flour, etc., resulting in low nutritional value, poor palatability, low digestion and utilization 7,8 . In addition, SMPE are prone to mildew or breeding pathogenic bacteria due to the high content of moisture, loose and porous raw material characteristics 9 . However, the nutritional value and palatability of SMPE can be improved through microbial fermentation. After microbial fermentation, macromolecular substances such as cellulose and hemicellulose are degraded into carbohydrates, amino acids, vitamins and other nutrients that are easily digested and absorbed by the animal, improving palatability and prolonging the preservation time. Lactic acid bacteria (LAB) and yeast are the most commonly used probiotics, they can not only convert the fermentation matrix into bacterial protein to improve the nutritional value, but also can produce flavor substances include acids, alcohols, esters and other aromatic substances to improve feed palatability 10 . LAB can produce organic acids, bacteriocins, hydrogen peroxide and other metabolites with bacteriostatic activity to inhibit the growth of other harmful bacteria 11 . Saccharomyces cerevisiae helps lactic acid bacteria due to use of substrates such as lactic acid and organic acids 12 . In addition, Bacillus subtilis has high protease and amylase activity 13 . Therefore, the construction of mixed-strain fermentation has been widely concerned.
Previous studies have shown that mushroom substrate feed is of great significance in the development and utilization of unconventional feed for ruminants 14 . Rumen is a unique digestive organ of ruminants, and there are a large number of microorganisms inside it, mainly including bacteria, fungi, archaea, protozoa, etc., of which anaerobic bacteria are the dominant microorganisms 15 . Rumen bacteria are closely related to ruminant production performance and meat quality 16 . The feed is fermented and decomposed under the action of microorganisms after entering the rumen, which is conducive to the efficient absorption of nutrients by the animal. Henderson et al. 17 and Maga et al. 18 found that feed is the dominant factor affecting the change of rumen microbial community structure in ruminants, which in turn affects the digestion and absorption of nutrients and energy supply. Therefore, understanding the microbial community composition structure of the rumen is the key to promote feed digestion and absorption and improve animal production performance.
We hypothesized that the supplementation of fermented SMPE with appropriate proportion might positively affect Hu sheep productivity, meat quality and rumen bacterial community. To test this hypothesis, the objectives of the present study were to determine the effects of different amounts of fermented SMPE on the production performance, meat quality and rumen bacterial community structure of Hu sheep.
Results
Effects of fermented spent mushroom substrate on production performance and meat quality of Hu sheep. According to Table 1, ADFI of RL1 and RL2 groups was significantly higher than that of RL4 (p < 0.05). FCR of RL2 was the lowest and significantly different from RL3 and RL4 groups (p < 0.05), while RL4 group was significantly higher than other groups (p < 0.05). No difference (p > 0.05) was observed in IBW, FBW, WG and ADG between groups. The EMA and GR values in the RL2 and RL3 were significantly higher than those in the RL1 and RL4 (p < 0.05). Compared with RL1, the EMA of RL2 and RL3 increased by 37.23% and 42.30%, and the GR value increased by 60.00% and 66.67%, respectively. No difference (p > 0.05) was observed in PSW, CW, SR, ST and back fat thickness (BFT) between RL1 and RL2 ( Table 2). In addition, there were no significant differences in MCP, pH of the longissimus dorsi muscle at 1 h and 24 h after slaughter, DR at 24 h and 48 h after slaughter (p > 0.05) ( www.nature.com/scientificreports/ Among the essential amino acids (EAAs), the contents of threonine, valerine, leucine, lysine, histidine and total amount of essential amino acids of the longissimus dorsi muscle in RL2 and RL3 were significantly higher than those in RL1 and RL4 (p < 0.05) (Table 3), the content of isoleucine was significantly higher than RL4 (p < 0.05), and the contents of methionine and phenylalanine were not significantly different between samples (p > 0.05). Compared with RL1, the contents of threonine, valerine, leucine, lysine, histidine and total essential amino acids in the RL2 significantly increased by 20.81%, 16 Among the non-essential amino acids, the contents of aspartic acid, serine, glutamic acid and arginine of the longissimus dorsi muscle in RL2 and RL3 were significantly higher than those in RL1 and RL4 (p < 0.05), and the tyrosine content was significantly higher than that in the RL4 (p < 0.05). Compared with RL1, the contents of aspartic acid, serine, glutamic acid and arginine in RL2 were significantly increased by 18.64%, 23.05%, 17.77% and 22.92%, respectively, and by 17.80%, 15.23%, 17.49% and 21.30% in RL3, respectively. For flavor amino acids (FAAs), the contents in RL2 and RL3 were significantly higher than those in RL1 and RL4 (p < 0.05).
OTU composition and structure of bacterial community. After quality control of data obtained by the IonS5 TM XL sequencing platform, a total of 1,202,445 valid sequences were obtained from raw materials and rumen liquid of Hu sheep, with an average valid sequences of 80,163 per sample. Valid sequences were clustered as 9824 OTUs at a sequence similarity threshold of 97%. Among them, RL0 had an average of 417 OTUs, RL1 www.nature.com/scientificreports/ was 980, RL2 was 1081, RL3 was 1165, and RL4 was 1061. The number of OTUs in RL0 was significantly lower than others (p < 0.05). The dilution curve directly reflects whether the extracted optimized sequence depth is reasonable and indirectly reflects the species richness in the sample. It can be seen from Fig. 1 that the number of sequences extracted reaches more than 30,000, and the curves tend to be flat, indicated that the sequence reservoir volume measured by different samples can better reflect the number of bacterial community species, and the amount of sequencing data was basically reasonable. The Venn diagram can visualize the differences and overlaps in the OTU composition of bacterial communities in different samples (Fig. 2). The results of Venn analysis showed that at the OTU level, specific bacterial OTU accounted for 14.71% (359) of the total OTU sequence number in RL0, specific bacterial OTU accounted for 8.85% (216) of the total OTU sequence number in RL1, specific bacterial OTU in RL2 accounted for 3.07% (75) of the total OTU sequence number, and specific bacterial OTU in RL3 accounted for 9.63% (235) of the total www.nature.com/scientificreports/ OTU sequence number. Specific bacterial OTUs in RL4 accounted for 6.06% (148) of the total OTU sequence number. In addition, the number of bacterial OTUs shared by RL0, RL1, RL2, RL3, and RL4 was 277 (11.35%).
Analysis of bacterial community diversity. The species richness and uniformity of rumen bacterial
communities treated with different amounts of fermented spent mushroom substrate were evaluated by α diversity index in the samples (cutoff = 37,136). Table S.2 showed that α diversity index of rumen bacterial community in RL1, RL2, RL3 and RL4 was significantly higher than RL0 (p < 0.05), indicated that the species diversity of rumen bacterial community was significantly higher than raw materials. However, with the increase of the addition of fermented SMPE, the differences in Observed species index, Shannon index, Simpson index, Chao1 index and ACE index were not significant (p > 0.05). The β diversity index could measure the degree of divergence in species diversity between two samples by Unweighted Unifrac distance. The smaller the value, the smaller the difference in species diversity between two samples. As can be seen from Fig. 3, the smallest distances of Unweighted Unifrac were RL2 and RL4, with a value of 0.343, and the largest distances were RL0 and RL3, with a value of 0.763. It can be seen that the difference in bacterial community structure between RL2 and RL4 was the smallest, and the difference between RL0 and RL3 was the largest.
PCoA analysis of bacterial community. The results of Principal Co-ordinates Analysis (PCoA) of rumen bacterial community with different treatments based on OTU were shown in Fig. 4, and principal component 1 (PC1) and principal component 2 (PC2) explained 80.45% and 7.71% of the variance of the variables, respectively, and the cumulative contribution rate was 88.16%. PC1 clearly distinguished the bacterial community in RL0 from RL1, RL2, RL3 and RL4, and PC2 clearly distinguished between four groups, indicated that there were great differences in the bacterial community structure of raw materials and the rumen liquid of Hu sheep with different treatments.
Changes in bacterial community composition and structure. A total of 10 phyla were detected in raw materials and rumen liquid of Hu sheep treated with different amounts of fermentation feed, the dominant taxa were as follows: Firmicutes (42.86-78.73%), Bacteroidetes (8.54-48.56%), Proteobacteria (0.71-7.49%) and Fibrobacteres (0.18-6.86%) (Figure S.1). Compared with RL1, the relative abundance of Firmicutes in the RL4 was significantly increased by 9.36% (p < 0.05), the relative abundance of Bacteroidetes in the RL3 was increased by 2.10%, but the difference was not significant (p > 0.05), and the relative abundance of Fibrobacteres in the RL4 was significantly increased by 68.24% (p < 0.05).
The relative abundance of Lactobacillus and Bacteroides in RL1, RL2, RL3 and RL4 decreased significantly compared with RL0 (p < 0.05), and the relative abundance of Prevotella 1, Christensenellaceae R-7, Ruminococcaceae NK4A214, Fibrobacter, Rikenellaceae RC9 and Prevotellaceae UCG001 increased significantly (p < 0.05). Compared with RL1, the relative abundance of Christensenellaceae R-7 in RL3 decreased significantly by 33.59%, the relative abundance of Prevotellaceae UCG001 in RL2, RL3 and RL4 decreased by 50.41%, 62.24% and 49.17%, respectively, and the relative abundance of Ruminococcaceae NK4A214 in RL2 increased significantly by 35.01% (p < 0.05). In addition, the analysis of bacterial community heat map based on genus level also showed that the Relationships between production performance and meat quality and rumen bacterial of Hu sheep. To further identify potential correlations between changes in production performance and meat quality and rumen bacterial of Hu sheep, we calculated Spearman's correlation coefficients between them. Firstly, we analysed the correlation between production performance and meat quality (Fig. 6). Both FAAs and EAAs showed positive correlations with PSW, CW, BFT, EMA and GR, Meanwhile, FAAs was positively correlated with EAAs. In addition, both PSW and CW were positively correlated with SR, BFT, but negatively correlated with pH (1 h) and pH (24 h), PSW was positively correlated with CW. BFT exhibited a positive correlation with SR, EMA, GR, whereas was negative correlated with pH (24 h). EMA and pH (1 h) were positively correlated with GR and pH (24 h), respectively. In contrast, DR (24 h) and DR (48 h) were negatively correlated to MCP and ST, respectively. Then, we analysed the correlation between rumen bacterial and production performance, meat quality. As shown in Fig. 7, Prevotellaceae UCG001 showed a positive correlation with DR (24 h) and proline.
Bacteroides showed a positive correlation with ST and FCR and negatively correlated with DR (48 h). Similarity, Ruminococcus was positively correlated with glycine and negatively correlated with EMA, GR and arginine. In addition, Fibrobacter showed a negative correlated with DR (48 h).
Discussion
Effects of fermented feed on production performance and meat quality of Hu sheep. Production performance of animals affects the economic benefits, the improvement of animal ADG and the reduction of FCR are particularly important in the evaluation of feed value. Slaughter performance is one of the important indicators reflecting animal production performance, among which PSW, CW and SR are important factors affecting animal economic value. Gao et al. 4 found that the supplementation of 30% fermented spent mushroom substrate from Pleurotus eryngii feed had higher body weight gain, ADG and dry matter intake, and lower FCR of Matou goat. Chu et al. 9 showed that adding 30% fermented mushroom by-product from Flammulina velutipes to the diet of growing-fattening pigs can significantly improve the carcass weight and quality of pigs. The results of this study showed that the group fed with 15% of SMPE had high ADFI compared with RL4 and low FCR compared with RL3 and RL4. Moreover, the overall production performance of Hu sheep improved at 15% and 30% compared to 0% of fermented SMPE, this may be due to the beneficial bacteria and bioactive substances contained in fermented SMPE to improve the rumen environment and promote digestion of feed. Firstly, the conversion rate of mineral elements was improved in fermented SMPE inoculated with probiotics, such as LAB, Saccharomyces cerevisiae and Bacillus subtilis. Those mineral elements might combine with pro- www.nature.com/scientificreports/ teins, amino acids, polysaccharides, etc. in the bacteria to form as absorbable organic compounds, which had high utilization rate and biological activity 4 . Secondly, the beneficial bacteria in fermented SMPE could multiply in the rumen after eaten by animals, thus produced a variety of digestive enzymes to enhance the degradation of macromolecular substances such as fiber. Meanwhile, the endocrine secretion of digestive enzymes in the rumen were induced by the beneficial bacteria, so as to improve the conversion rate of fermented SMPE 19 , thereby improving the conversion rate of fermented feed and improving slaughter performance. Besides, the proportion of fermented SMPE should not be too high. Adding too much fermented feed will lead to a decrease in slaughter www.nature.com/scientificreports/ performance, which may be due to the large amount of cellulose contained in the substrate, which stimulates rumen peristalsis, accelerates the circulation rate of chyme, so that the nutrients in the chyme are not completely absorbed and are excreted, reducing the apparent digestibility of feed nutrients 20 . It can be seen that Hu sheep fed with 15% of fermented SMPE had the best slaughter performance. Amino acids are important raw materials for protein synthesis, providing a material basis for host growth, metabolism and maintenance of vital signs, and their type and content are important factors affecting muscle flavor and nutritional value, and essential amino acid content is an important indicator to measure muscle quality 21 . When Boutry et al. 22 fed newborn pigs a leucine-rich diet, they found that the rate of protein synthesis in the muscles of pigs was accelerated, and the signaling pathways related to protein synthesis were significantly enriched. The results of this study showed that the total amount of essential amino acids in RL2 and RL3 was significantly higher than that in RL1, indicated that the addition of 15% and 30% of fermented feed could significantly increase the amino acid content of mutton and improve the nutritional value of meat quality. Mushroom mycelium residue in the SMPE has high protein concentration, therefore, was used as a protein source for ruminants 23 . In addition, the content of essential amino acids and total amino acids of SMPE were significantly higher than those of raw materials after mixed-strain fermentation, increased by 15.45% and 25.50%, respectively 24 . The content of flavor amino acids, including aspartic acid, glutamic acid and glycine etc., was higher in SMPE raw materials 24 . After fermentation, the microorganisms convert non-protein nitrogen into bacterial proteins, and some microorganisms have the function of secreting proteases, the amino acid content of SMPE is further increased, thereby increasing the amino acid content of muscle. Aspartic acid, glutamic acid and glycine are umami amino acids in muscles, and their content is an important indicator that affects the flavor of meat, of which aspartic acid and glutamic acid can also be used as drugs to treat some diseases, mainly used for the treatment of liver diseases, digestive tract diseases, encephalopathy, cardiovascular diseases, respiratory diseases and for improving muscle vitality, pediatric nutrition and detoxification. In this trial, the contents of aspartic acid, glutamic acid and glycine in RL2 and RL3 were significantly higher than those in RL1. It can be seen that the addition of 15% and 30% fermented SMPE to the diet of Hu sheep had a significant effect on the amino acid content in muscle, which could improve the flavor and nutritional value of mutton.
Effects of fermented feed on the structure of rumen bacterial community in Hu sheep. The microorganisms in ruminant rumen mainly include bacteria, fungi and protozoa, the rumen microbial composition structure is essential for maintaining environmental homeostasis in the rumen, promoting feed digestion and absorption, and benefiting animals. The structure of rumen microbial community is affected by many factors, of which diet types, structure and feeding mode are the most important factors 25 . For example, Liu et al. 26 found that the addition of yeast culture supplementation to diets changed the bacterial community composition of sheep's rumen in the house feed. In this experiment, the dominant phylum of Hu sheep rumen was Firmicutes, Bacteroidetes and Fibrobacteres, which were basically consistent with previous studies. Previous studies have shown that the Bacteroidetes and the Firmicutes are the dominant phylum in ruminant rumens 27,28 . Evans et al. 29 found that Bacteroides plays an important role in the degradation of non-fibrous substances, while Firmicutes is mainly involved in the decomposition of fibrous substances. The phylum Fibrobacteres are recognized as major bacterial degraders of lignocellulosic material in the herbivore gut 30 . After microbial fermentation, the nutritional value of SMPE was significantly improved, but due to the limitation of the raw material, the cellulose content remained at a high level. Ruminants do not produce cellulase enzymes themselves but depend on bacteria and fungi in the rumen to break down cellulose. Therefore, fed TMR diets contained fermented SMPE promote the growth and proliferation of Firmicutes and Fibrobacter in the rumen of Hu sheep. With the increase of the addition of fermented SMPE, the relative abundances of Firmicutes and Fibrobacteres in RL4 were significantly increased compared with those in RL1. The results showed that the addition of fermented SMPE to TMR of Hu sheep could promote the growth and proliferation of Firmicutes and Fibrobacter, and was conducive to the degradation of fibrous substances in the diet.
Previous studies have shown that Prevotella is the dominant genus of ruminant rumen 31,32 . Thoetkiattikul et al. 33 discovered by high-throughput sequencing of 16S rDNA that the dominant bacteria of ruminant rumens were Prevotella and Flavobacterium. In this experiment, the dominant genera in the rumen of Hu sheep were Prevotella 1, Christensenellaceae R-7, Ruminococcaceae NK4A214, Fibrobacter, Rikenellaceae RC9, Saccharofermentans and Prevotellaceae UCG001. It was not completely consistent with the results of previous studies, and it was speculated that it might be caused by differences in animal breed, age, feed structure, feeding management, etc. Prevotella is considered to be a degrading genus with a strong ability to decompose hemicellulose 34 , and plays an important role in the degradation of crude proteins, starches, xylan, and pectin 33,35 . The results of this experiment showed that the relative abundance of Prevotella 1 in the rumen of Hu sheep was the highest, which was consistent with the previous results, but the difference between the groups was not significant. This may be due to the fact that the diet designed for this study had similar crude protein and energy levels, so it did not have a significant effect on the relative abundance of Prevotella. Ruminococcaceae include ruminococcus flavefaciens and ruminococcus albus and are the main fibro degrading bacteria in the rumen and can produce a large amount of cellulase, hemicellulase and xylanase to degrade cellulose and hemicellulose in forage. In this study, the relative abundance of Ruminococcaceae NK4A214 in RL2 was significantly higher than RL1. The results showed that the addition of fermented SMPE to TMR of Hu sheep could promote the growth and reproduction of rumen and improve the degradation rate of fiber in the diet.
Diet plays an important role in shaping the rumen microbes in ruminants, alters the rumen's environment and metabolism 36,37 . Then alters the muscle metabolism and meat quality. For example, TMR containing palm kernel meal at 18% of Tibetan sheep increased the abundance of Christensenellaceae R-7, Ruminococcaceae UCG-013, Lachnospiraceae UCG-002, and Family XIII AD3011 in the rumen but decreased the abundance of Prevotella 1, www.nature.com/scientificreports/ the above bacteria groups regulate meat quality by regulating rumen metabolites (succinic acid, DL-glutamic acid, etc.) 38 . In this study, RL2, RL3 and RL4 decreased the abundance of Prevotellaceae UCG001 compared with RL1. Correlation analysis shown that Prevotellaceae UCG001 was positive correlation with DR (24 h) and proline. Indicated that the addition of fermented SMPE to diets may regulate meat quality of Hu sheep by affecting microorganisms in the rumen, since rumen microbes might affect the production of VFA and microbial protein by microbial fermentation of feedstuff 39 , the above functional metabolites subsequently alters the deposition of metabolites in muscles. Therefore, further research is needed to measure the end products of the fermentation process to determine how SMPE feed alters internal environment homeostasis and flora composition in the rumen to affect meat quality.
Materials and methods
Preparation of spent mushroom substrate from Pleurotus eryngii. SMPE was provided by Fujian Greenbao Group, the cultivation material formula was as follows: bagasse 13.0%, wood chips 22.2%, corn cob 26.0%, bran 18.0%, corn flour 8.8%, soybean meal 9.0%, lime 1.2%, light calcium 1.8%, which was adapted from previously described 40 . Waste fungus sticks were selected after harvest the mushrooms for 1 time, the mycelium was white, fresh and mildew-free, then unbagged, crushed and set aside. The compound microbial agent "Huojunduo" was bought from Beijing Shengyuda Biotechnology Co., Ltd. Number of viable bacteria: Bacillus subtilis ≥ 100 × 10 6 CFU·g −1 , lactic acid bacteria ≥ 10 × 10 6 CFU·g −1 , Saccharomyces cerevisiae ≥ 100 × 10 6 CFU·g −1 , total bacterial ≥ 210 × 10 6 CFU·g −1 . The formula of spent mushroom substrate mixed fermentation material was: 500 kg of fungus bran, 260 kg of fine bran, 180 kg of barley, 10 kg of brown sugar, 1 kg of fermentation fungus agent and 50 kg of water, which was adapted from the method as described by Gao et al 10 . The microbial agent was evenly sprayed on the raw materials, mixed evenly and divided into sealed bags, and fermented anaerobically at room temperature for 21 days. The nutrients before and after fermentation were shown in Table S.3.
Experimental design. The experiment was conducted at Longyan Green Tao Animal Husbandry Co., Ltd.
in Gaopi Town, Yongding District, Longyan City, Fujian Province, China (24.96• N, 116.86• E, above sea level 310 m). The present study used 120 Hu-breed lambs, the ratio of male to female stands at 1:1, around 60 days of age, the mean body weight was 13.50 kg (SD = 3.10) at the start of the performance test, all the male sheep were castrated. A univariate experimental design was used and randomized into 4 groups with 3 replicates in each group and 10 Hu sheep per replicate. 0 (RL1), 15% (RL2), 30% (RL3) and 45% (RL4) of fermented SMPE were added to TMR respectively, and the diet formula and nutrient level were shown in Table S.4. The trial was carried out by house feeding, sheep were adapted to the diet for a period of 10 days followed by 150 days of growth performance recording. Before the test, the sheep were numbered and dewormed, the sheep house was cleaned regularly, disinfected on time, and raised by special personnel in the well-ventilated sheep house. Feed once a day at 8:00 and 17:00 and drink freely. Feed should be based on a slight surplus in the trough, and the surplus should be collected and measured daily.
Measurement indicators and method.
Chemical composition analysis. The samples were dried in a forced-air oven at 60 °C for 72 h and then ground and sieved through a 1 mm sieve for chemical analysis. Moisture, crude ash, crude fibre (CF), and crude protein (CP) contents were determined using the method from the Association of Official Analytical Chemists 41 . Neutral detergent fibre (NDF), acid detergent fibre (ADF) contents were determined using the procedure reported by Van Soest et al. 42 . Calcium and total phosphorus were determined by the potassium permanganate method and the ammonium molybdate spectrophotometric method, respectively 41 . Metabolizable energy were calculated according to the method of Gao et al. 43 .
Production performance determination. Average daily feed intake (ADFI) was measured based on the difference between the offered and residual feed. The body weight of individual animals was measured weekly, and the weight gain (WG) was calculated based on the difference between the initial body weight (IBW) and final body weight (FBW). The feed conversion ratio (FCR) was determined cumulatively through the collected data.
At the end of the experimental, 3 test male sheep were randomly selected in each group (1 sheep per replicate), fasting for 24 h and water for 2 h before slaughter. Slaughter before weighing, cut neck and bleed, peel off fur, remove head, hooves and internal organs, and measure slaughter performance indicators, including carcass weight (CW), slaughter rate (SR), skin thickness (ST), eye muscle area (EMA), back fat thickness (BFT), grade rule (GR), etc. CW: pre-slaughter live weight to remove the weight of the head, fur, hooves, tail, and internal organs (retaining the kidneys and surrounding fat); SR(%) = CW/ weight before slaughter × 100; EMA: the crosssectional area of the superior spine-ophthalmic muscle between the 12th and 13th ribs, the cross-sectional contour of the eye muscle was depicted with sulfuric acid paper, and then the contour area was calculated by the accumulation meter (QCJ-2000, Shandong); GR value: The thickness of the tissue between the 12th and 13th ribs was measured 11 cm from the midline of the dorsal spine with a vernier caliper.
Meat quality determination. The left longissimus dorsi muscle was taken after slaughtered to determine the meat quality indicators such as meat cooking percentage (MCP), pH and drip rate (DR). MCP: peel off the muscle-attached fat, weigh with a balance with a sensitivity of 0.1 g (count as W1), put the meat sample into a polyethylene packaging bag, steam in an aluminum steamer for 30 min, take it out, and put in cool water to cool down for 1 h, dry the surface moisture with paper, weigh (count as W2), MCP(%) = W2/W1 × 100%. pH: The pH was measured 45 min after slaughter with a pH meter, and the average value was taken as the final result. DR: Cut 2 pieces of meat with a length of 5 cm, width of 3 cm and thickness of 2 cm, put them in plastic bags www.nature.com/scientificreports/ (meat samples must not contact the wall of plastic bags), tie the mouth of the bag tightly and hang them in a refrigerator at 4 °C, take them out after 24 h and 48 h, absorb the moisture on the surface with absorbent paper and weigh, record as the final weight. DR (%) = (initial weight-final weight) / initial weight × 100%. The amino acid composition of the longissimus dorsi muscle was determined. After treatment with a freeze dryer before testing, smashed and passed through a 0.425 mm sieve. Referring to the method of GB 5009.168-2016, the Hitachi L-8900 automatic amino acid analyzer (Japan) was used for measurement.
Sample determination and data analysis. DNA extraction and PCR amplification of rumen liq-
uid. After slaughter, the rumen contents were immediately collected from the rumen of 3 sheep in each treatment, transferred to sterile containers, stored in dry ice, and transported to the laboratory at − 80 °C for DNA extraction. The total DNA of the microbial genome of the samples was extracted by CTAB. DNA concentration was determined by Nanodrop and DNA quality was detected by 1.2% agarose gel electrophoresis. Dilute DNA to 1 ng·μl −1 using sterile water. The V4 region of the bacterial 16S rDNA gene was amplified using universal primers 515F (5′-GTG CCA GCMGCC GCG GTAA-3′) and 806R (5′-GGA CTA CHVGGG TWT CTAAT-3′). PCR products were detected using a 2.0% agarose gel for electrophoresis and bands of interest were recovered by the Gene JET Gel Recovery Kit (Thermo Scientific). The recovered products were sent to the IonS5 TM XL platform of Beijing Novogene Technology Co., Ltd. for high-throughput sequencing.
Statistical analysis. After the sequencing data were quality controlled by Fast QC, Cutadapt (V1.9.1) was used to filter short sequences (< 200 bp) and low-quality sequences (q < 25). Operational Taxonomic Units (OTUs) are assigned to valid sequences with 97% similarity with the SILVA database by the Mothur method. The sample diversity index was calculated using Qiime (Version 1.9.1) and principal component analysis was performed using R. All experiments were performed in triplicate and the data obtained were expressed as mean ± standard deviation (SD). For determinations of growth performance, slaughter performance, amino acid content of mutton, meat quality, relative abundance and diversity index of rumen bacterial community of Hu sheep, one-way analysis of variance (ANOVA) was done followed by Fisher's least significant difference (LSD) at p = 0.05 using SPSS 22.0 for Windows (SPSS Inc, Chicago, USA).
To identify putative correlations between production performance and meat quality and rumen microbial composition of Hu sheep, Spearman's rank correlation coefficients for pairwise comparisons were calculated and visualized with the corrplot package in R.
Conclusion
Overall, the production performance of Hu sheep was superior while fed with TMR contained fermented SMPE at 15% relative to the other groups. RL2 and RL3 resulted in an increase in EAAs and FAAs content. For rumen bacteria, the relative abundances of Firmicutes and Fibrobacteres increased in RL4, and Ruminococcaceae NK4A214 increased in RL2. However, RL2, RL3 and RL4 decreased the abundance of Prevotellaceae UCG001 compared with RL1. The correlation analysis indicated that the addition of fermented SMPE to TMR may improve the production performance and meat quality by affecting microorganisms in the rumen of Hu sheep.
Ethics statement
All of the procedures involving animals were approved by the Animal Care and Use Committee and Ethics Committee at the Agriculture Ecology Research Institute, Fujian Academy of Agricultural Sciences (NO. PZCAS-FAAS21022). We obtained written informed consent to use the animals in this study from the owners of the animals. All of the experiments were performed in accordance with the recommendations in the Regulations for the Administration of Affairs Concerning Experimental Animals of the State Council of the People's Republic of China. All of the experiments were carried out in compliance with the ARRIVE guidelines.
Data availability
The raw datasets generated during the current study are available in the NCBI (PRJNA944903) repository. | 2023-05-31T06:16:29.636Z | 2023-05-29T00:00:00.000 | {
"year": 2023,
"sha1": "a9c10bd8eb57c7a14daa530ba8114e7ab4160200",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "80ba7e5de59c1b72299fb41bab09ac565298f3e0",
"s2fieldsofstudy": [
"Agricultural and Food Sciences"
],
"extfieldsofstudy": [
"Medicine"
]
} |
6258995 | pes2o/s2orc | v3-fos-license | Reduced Exposure to Calcineurin Inhibitors in Renal Transplantation
n engl j med 358;23 www.nejm.org june 5, 2008 2518 Achilles’ heel of drug-eluting stents as compared with bare-metal stents, before our study, there was little information about the shortand longterm comparative safety and efficacy of these devices for off-label indications. Our study has filled the short-term knowledge gap with data from 1 year of follow-up, and with the recent receipt of funding from the National Heart, Lung, and Blood Institute Dynamic Registry for continued follow-up of these patients, it is our goal to help fill the knowledge gap regarding the long-term safety and efficacy of these devices. Oscar C. Marroquin, M.D. Suresh R. Mulukutla, M.D.
Reduced Exposure to Calcineurin Inhibitors in Renal Transplantation
To the Editor: Ekberg et al. (Dec.20 issue) 1 report the results of the Efficacy Limiting Toxicity Elimination (ELITE)-Symphony trial, in which they evaluated reduced exposure to calcineurin inhibitors in patients undergoing renal transplantation.The authors' putative conclusion is that a quadruple immunosuppressive regimen of daclizumab, "low-dose" tacrolimus, mycophenolate mofetil, and corticosteroids should be considered the standard in renal transplantation.We have concerns, however, about the equipotency of the four immunosuppressive strategies used.The trough levels of sirolimus that were achieved in the calcineurininhibitor-free group have not been demonstrated to be clinically effective and are, in our view, probably too low to protect grafts from acute rejection early after transplantation.In addition, the use of mammalian target of rapamycin (mTOR) inhibitors such as sirolimus immediately after transplantation is well known to be associated with a variety of acute postsurgical complications -for example, wound healing problems -as seen in this study.Therefore, most clinicians and current study designs strongly favor the delayed introduction of mTOR inhibitors.The high rates of side effects and drug discontinuation seen in the calcineurin-inhibitor-free group may be due largely to an ill-timed early initiation of sirolimus after transplantation.With these study-design limitations, we believe that clinicians should use caution when considering whether to abandon other potentially valuable immunosuppressive regimens on the basis of the conclusions from this study.
To the Editor: Ekberg et al. suggest that treatment with a combination of low-dose tacrolimus with daclizumab, mycophenolic acid, and corticosteroids is superior with regard to renal function, renal allograft survival, and acute rejection.However, the primary end point, the calculated glomerular filtration rate at month 12, depends heavily on the completeness of the data.Both the group that received standarddose cyclosporine and the group that received low-dose sirolimus were at a disadvantage because of higher numbers of patients with missing values (37 and 38 patients in these two groups, respectively, vs. 25 and 24 in the lowdose cyclosporine and low-dose tacrolimus groups, respectively, with the missing value arbitrarily calculated as 10 ml per minute).Fourteen missing calculations of the glomerular filtration rate result in a mean glomerular filtration rate that is estimated to be 1.75 ml per minute lower for the whole group.Along these lines, when the measured glomerular filtration rate was used, differences between groups were much smaller.The authors should provide an explanation for this imbalance with regard to missing values.How would the calculated mean glomerular filtration rates be affected if imputed values and the last observation carried forward were not used for missing values?What was the timing of the last observation carried forward in the different groups?What value was used for the glomerular filtration rate in patients who had graft loss, died, or both?To the Editor: On the basis of the results of the ELITE-Symphony study, Ekberg et al. conclude that a regimen containing daclizumab, mycophenolate mofetil, corticosteroids, and low-dose tacrolimus may offer an advantage over other immunosuppressive regimens after renal transplantation.However, the design of the study did not include a regimen containing standard-dose tacrolimus without daclizumab induction, whereas such a triple regimen is currently the standard in many transplantation centers.Moreover, the additional drug costs of about $6,500 (U.S.) incurred by the use of daclizumab are not discussed.Another strategy to reduce the adverse effects of calcineurin inhibitors is to discontinue their use in immunologically low-risk patients.Such an approach has been associated with favorable long-term results 1,2 and is probably more cost-effective.
The authors reply:
The objective of the ELITE-Symphony study was to identify a regimen of low toxicity and high efficacy.The choice of trough levels was based on commonly used long-term maintenance levels that we used from the day of transplantation and that we hypothesized to be equipotent.The results indicate that they were 1.
2.
The New England Journal of Medicine Downloaded from nejm.org at RADBOUD UNIVERSITEIT NIJMEGEN on November 8, 2012.For personal use only.No other uses without permission.
Copyright © 2008 Massachusetts Medical Society.All rights reserved.
not -a concern raised by Guba and Jauch; the sirolimus-based regimen was less effective but also had the highest toxicity and the greatest number of premature withdrawals.We concur that sirolimus toxicity may be reduced if its use after transplantation is delayed.Krüger and colleagues suggest that reduced exposure to mycophenolic acid in the cyclosporine group because of a drug-drug interaction may, in part, explain our results.We agree, but for two reasons we do not believe this is a major issue.First, although mycophenolic acid exposure is greater when mycophenolate mofetil is given in combination with sirolimus 1 or tacrolimus 2 than when it is administered with cyclosporine, the results for the sirolimus group were discordant with those for the tacrolimus group.Second, a larger proportion of patients in the tacrolimus and sirolimus groups received doses of mycophenolate mofetil that were lower than the specified dose (2 g per day) over the first 6-month period after transplantation, which would have reduced the difference in mycophenolic acid exposure between these groups and the cyclosporine groups.Regarding the comment about new-onset diabetes after transplantation, our study was designed in 2002, when there was no established consensus definition for this condition, and before publication of the American Diabetes Association-World Health Organization guidelines.New-onset diabetes after transplantation in the ELITE-Symphony study was defined as adverse-event reports that included the term "diabetes" or "hyperglycemia."However, only a small fraction of patients became insulindependent.
We agree with Krämer and colleagues that the calculated glomerular filtration rate can be influenced by the degree of completeness of the data.We conducted several sensitivity analyses for the mean glomerular filtration rate with different imputations and use of the last-observation-carriedforward method.These analyses yielded the same pattern of results as those reported in our article, thus confirming the robustness of the results.Finally, van den Hoogen and Hilbrands suggest that our study should have included a triple regimen containing standard-dose tacrolimus without daclizumab induction.However, standarddose cyclosporine was considered the benchmark when our study was designed.Regarding the comment about additional drug costs associated with daclizumab, health economics was not part of the scope of our article, nor were long-term changes in the maintenance treatment.
Nasal CPAP for Very Preterm Infants
To the Editor: In the trial reported by Morley et al. (Feb.14 issue), 1 a significant reduction in the use of surfactant in the group treated with early continuous positive airway pressure (CPAP) as compared with the intubation group (38% vs. 77%) was perhaps the most striking finding.In the CPAP group, the median time for intubation was 6.6 hours (interquartile range, 2.2 to 19.3), and we inferred from this that surfactant was probably given as a rescue treatment.Since the timing of surfactant therapy is likely to affect outcome measures, 2 perhaps the advantages of early CPAP balanced out the advantages of early surfactant treatment in the intubation group.To address this question, a detailed comparison of the timing of surfactant treatment in both groups would be of interest.
Germany bernhard.kraemer@klinik.uni-regensburg.deDrs.Krämer and Banas report having participated in clinical trials sponsored by Astellas, Novartis, Roche, and Wyeth.Dr. Krämer reports receiving lecture fees from Astellas, Novartis, and Wyeth; being a member of advisory or safety boards for Astellas, Novartis, and Wyeth; and receiving research grants from Astellas and Novartis.No other potential conflict of interest relevant to this letter was reported.
Since the publication of the article, Dr. Halloran reports receiving consulting fees and grant support from Astellas.No further potential conflict of interest relevant to this letter was reported. | 2018-04-03T04:35:44.560Z | 2008-12-31T00:00:00.000 | {
"year": 2008,
"sha1": "f3ccd5875810470ce9b0261172535eaef0e7a9e6",
"oa_license": "CC0",
"oa_url": "https://diposit.ub.edu/dspace/bitstream/2445/134284/1/567541.pdf",
"oa_status": "GREEN",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "cf9a327718e1b10494e10584342e2cea9066537f",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
250508034 | pes2o/s2orc | v3-fos-license | Cervical cancer screening programmes and age-specific coverage estimates for 202 countries and territories worldwide: a review and synthetic analysis
Summary Background Cervical cancer screening coverage is a key monitoring indicator of the WHO cervical cancer elimination plan. We present global, regional, and national cervical screening coverage estimates against the backdrop of the 70% coverage target set by WHO. Methods In this review and synthetic analysis, we searched scientific literature, government websites, and official documentation to identify official national recommendations and coverage data for cervical cancer screening for the 194 WHO member states and eight associated countries and territories published from database inception until Oct 30, 2020, supplemented with a formal WHO country consultation from Nov 27, 2020, to Feb 12, 2021. We extracted data on the year of introduction of recommendations, the existence of individual invitation to participate, financing of screening tests, primary screening and triage tests used, recommended ages and screening intervals, use of self-sampling, and use of screen-and-treat approaches. We also collected coverage data, either administrative or survey-based, as disaggregated as possible by age and for any available screening interval. According to data completeness and representativeness, different statistical models were developed to produce national age-specific coverages by screening interval, which were transformed into single-age datapoints. Missing data were imputed. Estimates were applied to the 2019 population and aggregated by region and income level. Findings We identified recommendations for cervical screening in 139 (69%) of 202 countries and territories. Cytology was the primary screening test in 109 (78%) of 139 countries. 48 (35%) of 139 countries recommended primary HPV-based screening. Visual inspection with acetic acid was the most recommended test in resource-limited settings. Estimated worldwide coverage in women aged 30–49 years in 2019 was 15% in the previous year, 28% in the previous 3 years, and 32% in the previous 5 years, and 36% ever in lifetime. An estimated 1·6 billion (67%) of 2·3 billion women aged 20–70 years, including 662 million (64%) of 1·0 billion women aged 30–49 years, had never been screened for cervical cancer. 133 million (84%) of 158 million women aged 30–49 years living in high-income countries had been screened ever in lifetime, compared with 194 million (48%) of 404 million women in upper-middle-income countries, 34 million (9%) of 397 million women in lower-middle-income countries, and 8 million (11%) of 74 million in low-income countries. Interpretation Two in three women aged 30–49 years have never been screened for cervical cancer. Roll-out of screening is very low in low-income and middle-income countries, where the burden of disease is highest. The priority of the WHO elimination campaign should be to increase both screening coverage and treatment of detected lesions; however, expanding the efforts of surveillance systems in both coverage and quality control are major challenges to achieving the WHO elimination target. Funding Instituto de Salud Carlos III, European Regional Development Fund, Secretariat for Universities and Research of the Department of Business and Knowledge of the Government of Catalonia, and Horizon 2020. Translations For the French, Spanish translations of the abstract see Supplementary Materials section.
S1.1. Search terms
The systematic review included an initial search in official websites of countries that publish information on cancer control plans, screening policies and coverage statistics (e.g., health departments and national epidemiological centres) followed by a global review using web search engines to look for additional web-based materials. Search terms included specific country names and the following keywords "HPV", "screening", "cervical cancer", "cytology", "pap smear", "papanicolau", "VIA", "VILI", "coverage", "uptake", including the abbreviated versions and its variations (e.g. "VIA", Visual Inspection with Acetic Acid, early detection of cancer…) The algorithm also included a systematic search in PubMed using a combination of MESH terms and keywords relevant to "cervical cancer", "screening", and "coverage" for each specific country.
S1.2. Data extraction and translations
Six reviewers were involved in the search and data extraction. This process was done directly, without the help of translation, for publications in Romance languages or English. Eleven professional translators assisted investigators in the search of information and the interpretation of the data for 37 countries in which official languages are Arabic, Russian, Ukrainian, Armenian, Serbian, Croatian, Bosnian, Macedonian, Montenegrin, Slovenian, Polish, Romanian, Greek, Czech, Slovak, Hindi, Urdu, Persian, and Standard Chinese.
Methods S2. Data pre-processing.
Official cervical cancer screening recommendations and coverage data (for at least one of the five screening intervals: previous year, previous two years, previous three years, previous five years, ever in lifetime) were identified in 139 and 164 out of 202 WHO member states and associated countries and territories, respectively (Tables S1-S2). A global database with 10 302 single ages from 20 to 70 years and 202 countries was constructed including the identified information on screening recommendations and coverages. Below we describe the process of data selection and transformation.
S2.1.1. Country-specific corrections and assumptions for coverage data
For five countries, the coverage of an interval different from those included in the analysis was assigned as the country coverage (Table MS2. 1.1.1). For Sweden, Papua New Guinea, and Norway, we assigned the previous eight-, 10-and 25-year coverage, respectively, as the ever in lifetime coverage. For France and England-United Kingdom, we assigned the previous 3.5 years coverage as the previous three years coverage. For United Kingdom we did not have a global national coverage, but four regional coverages (England, Scotland, Wales, Northern Ireland). We calculated the national coverage applying the proportion of population that each region contributes to the national coverage (82.3% England, 8.1% Scotland, 4.7% Wales, 2.8% Northern Ireland). For ages 20-24 years and 65-70 years, and for screening intervals of three-and five-years, we only had screening coverage from England, so we assumed this coverage as the national one.
For each singe-age and country, we verified that no original coverage exceeded that of its upper screening interval (coverage previous year < coverage previous two years < coverage previous three years < coverage previous five years < coverage ever in lifetime). This rule was not satisfied in seven countries (Table MS2. 1.1.2). For Germany (cov 1-year> cov 3-year and ever, age group 20-24 years), Uruguay (cov 1-, 3-, 5-year > cov ever, age 20 years; cov 1-year > cov 3-year, age 70 years), Belgium (cov 2-year> cov ever, age group 20-24 years), Estonia (cov 3-, 5-year > cov ever, age group 20-24 years), and Czechia (cov 5-year > cov ever, age group 20-24 years) the aggregated coverage for the mentioned age groups (in brackets) systematically included ages outside 20 to 70 years and therefore we decided to not use this data when discrepancies appeared and treat this coverage as missing. For Iceland and Kenya, the rule was satisfied when the disaggregated ever in lifetime distribution was applied to disaggregate the 1-, 3-, and 5year aggregated coverages (see section S2.2). y: year; NA: Not Available; Data disag: Data disaggregated, Sec: Section.
S2.1.2. Country-specific assumptions in programme data:
To perform the statistical analysis, we had to assign a missing screening interval for the recommendations in 10 countries. When the primary screening test was VIA (Guinea, Madagascar, Mozambique, and Timor-Leste) or cytology (Dominica, Cyprus, Vanuatu, Cook Islands, and Syria), we assumed a 5-year interval. Particularly for Sri Lanka, we assumed a 10-year interval because the screening recommendations indicate that cytology should be performed at ages 35 and 45 years.
For 11 countries (Belgium, Cyprus, Portugal, Romania, South Africa, Paraguay, Vanuatu, China, Syria, Timor-Leste, and Cook Islands) we assumed that there was no screening invitation as no data on screening invitation was identified in the original sources.
In eight countries, the antiquity of the screening programme (years since the introduction of the screening programme in the country) was missing (Cyprus, Guinea, Antigua & Barbuda, Grenada, St Lucia, Vanuatu, Timor-Leste, Cook Islands), and we assumed less than 10 years for all of them as a more conservative decision.
In eight countries (United Arab Emirates, Belgium, Canada, Italy, Portugal, Spain, and United Kingdom) more than one programme recommendation was registered (e.g., different regional recommendations, or recommendations only in some regions of the country). We selected the one considered more representative of the country population.
In 19 countries (American Samoa, Bermuda, Bolivia, Brunei, Chile, Colombia, Ecuador, El Salvador, Guatemala, Mexico, Myanmar, Malaysia, Nicaragua, Panama, Paraguay, Peru, Puerto Rico, USA and, South Africa) more than one screening test was recommended for a specific age group (e.g., cytology or VIA for women aged 30-49 years). We selected the most used test or the one preferred in national recommendations.
S2.2. Data transformation
Screening coverages were reported aggregated by age groups of five years, ten years, or more. Coverages were transformed into single age datapoints by assigning the same coverage to all ages in the reported age group.
S2.2.1. Country-specific disaggregation of coverage data by age
For seven countries (Belgium, Iceland, Israel, Kenya, Peru, Slovenia, and Turkey) with reporting coverages for wide age range groups (e.g., 30-65 years), we modified the aggregated data using the age disaggregated distribution from other screening intervals within the same country ( For one country (Gabon), we modified the aggregated out of programme data coverage using a corrected factor based on disaggregated data from countries with same income and antiquity of screening programme.
For Gabon we had aggregated data for the previous year and previous three years coverage among women aged 20-70 years. The screening programme recommends to screen women between 25-65 years, so ages 20-24 years and 66-70 years are out of programme. We calculated a correction factor including the mean disaggregated data of the previous year and previous three years intervals from those countries with the same income and antiquity of screening programme (Georgia, Jamaica, Romania). In summary, to calculate the new disaggregated out of programme data for the previous year and previous three years coverage intervals, we applied the following formula: New disaggregated previous year and previous three years coverage for 20-24 years = Original aggregated previous year and previous three years coverage for 20-24 years * mean (previous year and previous three years disaggregated coverage for 20-24 years from Georgia, Jamaica, Romania) / mean (Original previous year and previous three years aggregated coverage for 20-70 years). We also recalculated the data for those ages included in the screening programme (25-65 years) to obtain the same mean aggregated data for 20-70 years.
Methods S3. Statistical procedures.
To produce global estimates of screening coverage we developed a multi-step algorithm to impute missing datapoints based on the closest available data (Table S4). This algorithm included the following statistical procedures:
S3.1. Linear interpolation
Linear interpolation is a mathematical method to estimate new datapoints within the range of a set of known datapoints using linear polynomials 1 . Linear interpolation between two known points is the straight line between these points. We can interpolate as many points as needed between the two known points.
Application
In our stepwise algorithm, imputation by linear interpolation between screening intervals within the same country was applied first (Table S4: Step 2). We selected ages with at least two coverage datapoints in different screening intervals, and then imputed missing values between them by linear interpolation. This is a conservative strategy that imputes plausible data assuming linear functions between the known coverage datapoints of the different screening intervals.
S3.2. Predictive Mean Matching using Multiple Imputation
Predictive Mean Matching (PMM) is a semi-parametric statistical imputation method for missing values [2][3][4] . Imputation by PMM uses an observed value from one value with a similar predictive mean. Compared with standard methods based on linear regression and normal distribution, PMM produces imputed values that are much more like real values sampled from the data. The aim of this process is to reduce the bias introduced in a dataset through imputation. Compared to other imputation methods, it usually imputes less implausible values (e.g., negative incomes) and takes heteroscedastic data into account more appropriately.
The PMM algorithm builds a small subset of observations where the outcome variable matches the outcome of the observations with missing values. Multiple imputation is recommended, repeating the process iteratively at least five times, although it is advantageous to set higher iterations, ranging from 20 to 100 5-6 . It creates several different plausible imputed datasets and combines the results obtained from each of them.
Application
In our stepwise algorithm, the PMM was applied second when linear interpolation was not possible (Table S4: Step 3a). PMM includes as many relevant covariates in the model as possible to obtain better estimates of missing data. Overfitting (e.g., making it more dependent on your data) is not real a problem and any available associations, small or large, are used.
We included the following covariates: - -Coverage characteristics: single age (20 to 70 years), range in years of the coverage figure (i.e., if coverage was reported for the age group of 25-29 years old, this corresponds to a range of five years), estimation year (year of estimation of the original coverage).
-Programme characteristics: existence of screening recommendations (dichotomous yes/no), use of personal invitation to screening (dichotomous yes/no), whether age was included in the recommended screening ages (dichotomous yes/no), whether the year of original reported coverage corresponds to the date of the current screening recommendations (dichotomous yes/no), antiquity of the programme (dichotomous 10 years/>10 years).
We built several models including a combination of the covariates. The final PMM model included the following covariates: single age (20 to 70 years), range in years of the age group in which the coverage was reported, existence of screening recommendations (dichotomous yes/no), whether the age was included in the recommended screening ages (dichotomous yes/no), individual invitation to screening (dichotomous yes/no), whether the year of the reported coverage corresponds to the date of the current recommendations (dichotomous yes/no), HDI (as quantitative), UN geographical subregion (22 categories), whether the country was under FCS (dichotomous yes/no), whether the country is an SST (dichotomous yes/no). Some covariates, such as country, could not be included in the model because the PMM model detected multi-collinearity and results could be instable.
Multiple imputation was applied, creating 40 imputed datasets from the interim dataset obtained after steps 3b and 3c (Table S4) in which all countries with at least one observed coverage datapoint had all datapoints filled. Resulting databases without missing data included the 202 (countries) * 51 (single ages from 20 to 70) records, resulting on 10 302 records in each dataset (a total of 412 080 records).
S3.3. Last observation carried forward / Next observation carried backward
Single imputation methods replace a missing datapoint by a single value, and analyses are conducted as if all the data were observed. In Last Observation Carried Forward (LOCF) and Next Observation Carried Backward (NOCB) single imputation methods, the single value used to fill in the missing datapoint comes from the observed values in the same subject [7][8] . LOCF imputes the last measured value before the missing value and carries it forward. NOCB is a similar approach to LOCF but works in the opposite direction, by imputing the first observation after the missing value and carrying it backward. These techniques are common statistical approaches to the analysis of longitudinal repeated measures where some follow-up observations may be missing.
Application
In our stepwise algorithm, LOCF or NOCB were applied third, if linear interpolation was no applicable and PMM was not used because we do not have any coverage for those ages (Table S4: Step 3b). Within countries with single ages with missing data, we carried forward or backward the mean of the last/first five observations as long as the ages had the same screening recommendations. Its use was restricted to a maximum carrying of 5 datapoints to avoid overrepresentation of one specific imputed datapoint.
S3.4. Bootstrapping to calculate confidence intervals (CIs)
It is important to consider inference of the estimates after multiple imputation. When estimates are calculated using non-parametric models (e.g., PMM method), it is unclear how to obtain valid inference, meaning obtain standard errors, and therefore confidence intervals). Bootstrap estimation may be an option 9 .
In general, we can distinguish between two approaches for bootstrap inference when using multiple imputation: M imputed datasets are created and bootstrap estimation is applied to each of them, or B bootstrap samples of the original dataset (including missing values) are drawn and in each of these samples, the data is imputed. Taking this into account, there are four different approaches to combine results and calculate confidence intervals using bootstrap: -Boot Multiple Imputation: B bootstrap samples, including missing data, are drawn and each of them is imputed M times. There are M imputed datasets associated with each bootstrap sample. They can be used to obtain the corresponding point estimates, and the set of ordered estimates can be used to construct the confidence interval.
Application
In our stepwise algorithm, we used the first approach of Multiple Imputation Boot (using pooled sample) to calculate the confidence intervals of our estimates. We had to decide in advance the ages and regions for which we wanted to produce the screening coverage estimates (i.e., worldwide among women aged 30-49 years, by income among women aged 25-65 years).
Therefore, from the complete database that included the 40 imputations (412 080 records), we calculated, for a previously defined age group and region, the number of women screened for each of the 40 imputed datasets. Then, we computed, for the 40 estimated numbers, 3 000 bootstrap estimations, we combined these 3 000 estimated numbers of screened women from the 40 imputations, and we computed the confidence interval using the 2.5% and 97.5% quantiles. The same process was done for each screening interval, and for each selected age group and region.
Methods S4. Treatment of missing coverage data.
As previously mentioned, coverage data was available for at least one of the five screening intervals for 164 out of 202 WHO member states and associated countries and territories, covering a combination of 7 324 single age datapoints: 3 239 point estimates were available for the previous year, 1 110 for the previous two years, 3 451 for the previous three years, 1 580 for the previous five years, and 5 472 for the ever in lifetime.
From a total of 51 510 point estimates (202 countries x 51 single ages from 20 to 70 years x 5 screening intervals), we had available information for 14 852 (28.8%) point estimates. Coverage data was unavailable for 38 countries, corresponding to 1 938 single age datapoints and 9 960 point estimates (Table MS4).
S4.1. Stepwise algorithm to impute missing specific single age data coverage.
A multi-step algorithm was developed to impute missing datapoints based on the closest available data.
The imputation algorithm was different in countries with at least one coverage datapoint and in countries without any coverage datapoints.
S4.1.1. Algorithm in countries with at least one coverage datapoint
Imputation by linear interpolation between screening intervals was firstly applied (Table S4: Step 2). 3 606 single ages had at least two coverage datapoints reported with 6 976 missing coverage datapoints. 4 669 missing datapoints could be imputed by linear interpolation because coverage data was available in upper and lower screening intervals. 2 307 datapoints remained as missing because they could not be imputed by linear imputation (i.e., missing on coverage in the previous year).
For single ages with at least one coverage datapoint and when linear interpolation was not possible, we imputed missing data using the PMM method using the covariates explained in Methods S3 (Table S4: Step 3a). Using the PMM model, 40 imputed databases were created derived from the multiple imputation process. In total, 17 099 coverage datapoints without data from each imputed database were calculated, obtaining complete single ages for the five screening intervals for the 164 countries with at least one coverage datapoint. For each singe-age specific imputation, it was verified that no coverage exceeded that of its upper screening interval (coverage 1-year <coverage 2-year <coverage 3-year <coverage 5-year <coverage ever). When this rule was not met, coverage was recalculated as follows: original data was preserved, imputed coverage datapoints were reordered assigning the minimum imputed value to the lowest screening interval and the maximum imputed value to the greatest screening interval, intermediate coverage datapoints without original data were assigned as missing (Table S4: Step 5). When the minimum or maximum imputed value assigned to a coverage datapoint was greater or lower respectively than the original coverage datapoint, the original coverage datapoint was assigned to the missing values. Finally, the remaining missing datapoints were imputed by linear interpolation between screening intervals. With this strategy 116 154 single ages were corrected among the 40 imputed datasets.
For single ages without coverage datapoints, last observation carried forward or next observation carried backward techniques were applied within the last five observations before or after the missing data (Table S4: Step 3b). This approach was only applied when programme characteristics were the same between single ages. To avoid overrepresentation of one specific imputed datapoint, we carried forward or backward the mean of the last/first five observations. With this strategy, 58 400 coverage datapoints were imputed among the 40 imputed datasets (1 460 datapoints in each imputation).
For the rest of single ages without coverage datapoints, a ponderation rate was applied. This ponderation rate was calculated using the imputed coverage database for each country and five-year age group (20-24, 25-29, …, 60-64, 65-70) from countries with same income and same single ages included in the screening recommendations (Table S4: Step 3c). With this strategy, 149 600 coverage datapoints were imputed among the 40 imputed datasets (3 740 datapoints in each imputation).
Again, for each singe-age specific imputation, it was verified that no coverage exceeded that of its upper screening interval. When the rule was not met, coverage data was reordered, and linear interpolation was applied as previously explained (Table S4: Step 5). A total of 3 544 single ages had to be changed among the 40 imputed datasets.
S4.1.2. Algorithm in countries without coverage datapoints
Missing screening coverage datapoints for the 38 countries without any original coverage data were imputed by the PMM method, directly applied to the original coverage dataset (Table S4: Step 4). This model included the same covariates used for countries with at least one coverage datapoint available.
Using the PMM method, 40 imputed databases were created derived from the multiple imputation process. A total of 9 690 datapoints (38 countries x 51 single ages from 20 to 70 years x 5 screening intervals) from each database were imputed, obtaining complete single ages for the five screening intervals for the 38 countries without any original coverage data.
Again, for each singe-age specific imputation, it was verified that no coverage exceeded that of its upper screening interval. When the rule was not met, coverage data was reordered, and linear interpolation was applied as previously explained (Table S4: Step 5). A total of 61 814 single ages were rectified among the 40 imputed databases. Figure S3 shows the final coverage dataset by country and imputation method.
S4.2. Validation of the algorithm
Missing data imputation has many assumptions and may produce bias in the screening coverage estimates. Several sensitivity analyses and validations were performed to address and validate our methodological approach to treat missing data.
S4.2.1. Simulations a) Comparison between original and imputed data selected randomly
Fifty simulations were performed taking a random sample of 200 original coverage data for each simulation. These 200 original coverages included different age groups and age ranges (e.g., original coverage data for the previous year was taken for one country for the age group 25-29 years, but for another country the original coverage data for the previous year was taken for the broader age group 25-64 years). After this, the stepwise algorithm to impute missing data was run again treating this data as missing. The 50 simulations performed resulted in 66 672 single age coverages to be imputed. When comparing the imputed data with the original coverage data not used, the correlation coefficient was 0.89 including all simulations combined ( Figure S4). In 6 917 out of 66 672 (10.4%) imputed data, the difference with the original data was equal or greater than 15%. Among the rest of imputed data, 10.1% showed a difference between 10-15%, 23.6% a difference between 5-10%, and around 55.9% showed a difference lower than 5%.
b) Comparison between original and imputed data, excluding one by one each country
All the original coverages from the 164 countries with at least one datapoint were removed and treated as missing data, country by country. Missing data was imputed using the multi-step algorithm, including the imputation of the original removed data for each country. Screening coverages were then estimated worldwide, by income level and by region, obtaining almost identical estimation rates ( Figure S5). Differences ranged from 0% to 1% in almost all countries. Despite this, differences between 5-10% depending on the screening interval were observed when data from countries with largest populations were put as missing and imputed (e.g., China, India).
S4.2.2. Sensitivity analyses for the imputation algorithm
Additionally, sensitivity analyses of the most uncertain assumptions of the stepwise algorithm were performed (Table S4).
a) Restriction of coverage data by age and region
First, we calculated the coverage estimates for the age group 30-49 years using three different agescenarios. Scenario one used the complete database with individual coverages for single ages from 20 to 70 years, while scenario two and three restricted the database to individual coverages for the age group from 25 to 65 years and from 30 to 49 years, respectively. Estimations of screening coverage were very similar to the final estimates of our study, worldwide and by income and region (Table S6).
Second, we calculated the coverage estimates for the age group 30-49 years using two region-scenarios. Scenario one included the complete database with coverage data from all regions, while scenario two only included data for countries corresponding to each income level or SGD region. Coverage estimates were very similar compared to the final estimates of our study (Table S7). By income level, imputations including only upper middle-income countries showed differences in the coverage estimates in the previous year close to 7% (from 11% to 18%) compared to imputations including all countries. Differences close to 3% were also observed in all screening intervals when low-income countries were validated. By SGD regions, differences ranged from 2% to 5% in the estimation coverages for some screening intervals in
b) Sensitivity analyses for steps 1 to 4 of the imputation algorithm
We carried out a sensitivity analysis of each step of the imputation algorithm, including modification of aggregated and out-of-programme coverage, linear interpolation, LOCF and NOCB techniques, and application of a ponderation rate. Global estimates derived from the non-application of these four steps of the algorithm (Table S4) were consistent compared to the reference ones, with differences close to 1-2% compared to our reference estimates (Table S8). The main differences in the screening coverages were observed when linear interpolation was not used (scenario 3) and PMM was applied instead, where disparities of up to 10% were observed. Specifically, by income, the non-application of linear interpolation resulted in differences in screening coverages ranging from 2 to 4% in low-income countries, although results were similar for the other income regions. By region, the greatest differences were observed in Western Africa (close to 10%) and ranging from 4 and 5% in Western Asia and Australia & New Zealand.
We also did a sensitivity analysis considering a group of scenarios where all coverages from countries without data were set to 0%, 25%, 50%, and 75-100%. We applied these values to each screening interval. In the most extreme scenarios, global estimates were 1% lower when missing data was set to 0 (scenario 2) and 6% higher when missing data was set to 75-100% (scenario 5) ( Table S9). The impact of these assumptions was more prominent in low-income countries, with estimates between 2 to 6% lower (scenario 2) or between 33 to 39% higher (scenario 5). This impact was almost negligible in high-and upper-middle-income countries. Differences by income could be explained by a higher number of missing data in lower-middle-and low-income countries compared to high-and upper-middle-income countries. By region, we observed similar patterns, with estimates 56 to 60% higher (scenario 5) in Sub-Saharan Africa (excluding Southern Africa), 19 to 23% higher (scenario 5) in Polynesia and 8 to 9% higher (scenario 5) in Western Asia.
S4.2.3. Impact of imputations in the results
The impact of the imputation algorithm in the global estimates varied substantially depending on the population of the country for each single age between 20 and 70 years old (Table S5). We could obtain coverage data from original sources for 7 324 single ages, corresponding to 14 852 datapoints considering all the screening intervals (29%) (Table S3). When population weights are considered, the 14 852 datapoints represented 59% of the targeted women instead of 29%, meaning that more than half of coverage data used to estimate the number of screened women came from the original sources.
Regarding the imputed datapoints (71%), 9% were derived from linear interpolation, representing 13% of targeted population (Step 2). 33% of datapoints were imputed by the PMM method, corresponding to 36% of targeted population (Step 3a). Around 10% of datapoints were imputed using the approach of LOCF and NOCB, or a ponderation rate based on other country coverage datapoints with similar characteristics, representing less than 0.1% (Steps 3b and 3c) of targeted population.
Step 4, that included PMM imputation for countries without data, was the one with more associated uncertainty because imputation of coverages came from the rest of countries. About 19% of datapoints were imputed through this methodology, representing 3% of targeted population.
Methods S5. WHO official country consultation.
Following WHO's quality standards for data publication an official consultation round with WHO member states and associated countries was conducted from November 27th, 2020, to February 12th, 2021, to review, comment and provide insight on the estimates. Countries were presented with draft estimates and sources of data. The country consultation was responded by 83 countries, resulting in an update of screening policies and coverages in 33 and 42 countries, respectively. Figures S1.1, S1.2 and S1.3. Distribution of original coverage data by age and screening interval.
To assess whether the pooled coverage data for one screening interval could be disaggregated using the distribution from other screening intervals with more disaggregated data within the same country, we compared the distribution of the original data among screening intervals in the 63 countries with original disaggregated data for more than one screening interval. We specifically assessed whether the ever in lifetime distribution was comparable to the other screening intervals distribution. Figure S1.2), and only one country had different distribution for all age groups (The Netherlands) ( Figure S1.3). Considering this, we assumed that age-distribution was similar among screening intervals. Statistical tests could not be performed to verify this assumption because coverage data was not independent among screening intervals and the number of datapoints was too small.
Figures S1.1, S1.2 and S1. 3 show the distribution of ever in lifetime coverage (y-axis) compared with other screening intervals, by single age (x-axis) and screening interval (lines).
Figure S1.2. Distribution of original coverage data by single age and screening interval among countries with different pattern in older ages for the ever in lifetime coverage compared to other screening intervals.
Blue line: previous year; Pink line: previous two years; Green line: previous three years; Black line: previous five years; Red line: ever in lifetime.
Figure S1.3. Distribution of original coverage data by single age and screening interval among countries with different pattern for the ever in lifetime coverage compared to other screening intervals.
Black line: previous five years; Red line: ever in lifetime.
Figures S2.1 and S2.2. Density plots of missing data using Predictive Mean Matching.
We plotted below the density plots of original and imputed data using the Predictive Mean Matching method for each screening period ( Figures S2.1 and S2.2). In blue we drew the original data and in red we drew the 40 imputed data. As displayed in the plots, PMM imputes lower coverages to the original data for all screening intervals except for the ever in lifetime coverage, suggesting that screening data was missing in countries with small coverages.
Figure S3. Imputation method used to estimate missing data by age and country
The figure below presents the imputation method used to estimate missing data for the previous year, previous three years, previous five years and ever in lifetime among women aged 20-70 years by country. Green squares represent the original coverage data (Step 0), and the rest of estimations are coming from the multi-step imputation algorithm (Steps 1 to 4) (Table S4).
Legend:
Step 0 Original data coverage
Countries with at least one coverage datapoint
Step 1
Modification of aggregated/out of recommendations data
Step 2
Linear interpolation
Step 3a Predictive Mean Matching method to records with at least one datapoint Step 3b Last observation carried forward / next observation carried backward technique Step 3c Ponderation rate
Countries without coverage datapoints
Step 4 Predictive Mean Matching method to all records of the database Figure S4. Correlation between imputed data and original data for 50 simulations.
One of the strategies to validate the imputation algorithm was to run 50 simulations in which in each simulation we selected a random sample of 200 age groups with original coverage data to be treated as missing, and we imputed the data following the stepwise algorithm (Table S4). The resulting plot of all simulations combined shows a correlation coefficient of 88.8%. 6 917 out of 66 672 single age datapoints imputed (10.4%) had distances equal or greater than 15% (coloured in red in the plot).
Figure S5. Differences between imputed data and original data, excluding one by one each country.
Another strategy to validate the imputation algorithm was to select one by one each country with original coverage data (164 countries), treat the country as missing and impute the data following the stepwise algorithm. The boxplot shows screening estimates worldwide and by income level, for each screening interval. Global estimates worldwide show differences between 1-4% and of maximum 11% by income level and region. The higher differences came from countries with larger populations, such as China, India, Brazil, or USA. Of the 164 countries for which coverage data was available, most had data for the previous year, the previous three years, and/or ever in lifetime (46%, 48%, and 74% respectively). Coverage was available for at least one of the five screening intervals for a combination of 7 324 / 10 302 (71%) single age datapoints (Table S3.1). For the 30-49 age group, data completeness was higher with 3 260 / 4 040 (81%) datapoints available (Table S3.2). * Combination of at least one datapoint for any of the screening intervals. | 2022-07-24T05:11:09.768Z | 2022-07-12T00:00:00.000 | {
"year": 2022,
"sha1": "4d52a7691b0f0db6b63581cd35814a3edb00fa51",
"oa_license": "CCBYNCND",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "74dcb2297e101762d1928e807a0d092880a67c35",
"s2fieldsofstudy": [
"Medicine",
"Political Science"
],
"extfieldsofstudy": []
} |
232422641 | pes2o/s2orc | v3-fos-license | Early-Responding Civil Society and a Late Coming State: Findings from Turkey during the Pandemic
: By building on available literature and interview notes with the NGOs working mainly on refugees and irregular migrants, this study examines the experiences of civil society in Turkey in the first 6 months of the COVID-19 pandemic. One of the main findings is that NGOs responded quite early and with agility to the sufferings of people before the state entered into the field and implemented its own rules and methods to deal with the crisis. The study also reveals that the pandemic affected the focus of the NGOs, their services, finances, and working methods in the field. The organizations we examined adapted to a new online working environment which increased demand for their services and aggravated difficulties in assuring clients access to needed healthcare services. The study finds that relations with the local and central governments remain difficult, and during the pandemic, the NGOs were compelled to act very cautiously. While suffering from their society ’ s political polarization, they report that xenophobia continues to rise and has been worsened by the impact of the pandemic.
Introduction
Turkey was stricken by COVID-19 just a bit later than were most other European countries. Even before the first case was diagnosed, Turkey's health authorities announced that the country was ready to fight with the virus. Though the health ministry has been criticized for the reliability of its statistical methods for keeping and reporting records on the pandemic, the health system seemed not have experienced an overload until the pandemic more fully established itself. At the time of this writing in late 2020, the number of confirmed cases and deaths is on the increase in the country, and Turkey has joined the global health crisis.
As the UNDP (2020: 2) notes, COVID-19 is much more than a health crisis; "it has the potential to create devastating social, economic, and political crises that will leave deep scars." In such times of crisis, NGOs and their activities become critical in reaching out to vulnerable people. As presented in this volume, NGOs are often at the forefront of COVID-19 efforts; they work to tackle the pandemic and support fragile and vulnerable communities. Throughout the pandemic, migrants have found themselves subject to particular stress. As noted by IOM (2020) and OECD (2020), migrants are more vulnerable than more settled persons in times of crisis because of the effects upon them of a confluence of personal, social, legal, and structural factors. They have socio-economic difficulties and difficulties in accessing healthcare, and they have to struggle with cultural and linguistic barriers, as well. They are reported to be more affected by the spread of COVID-19 in societies in which the pandemic is spreading. Moreover, in these times of crises, they may experience increased stigmatization and social exclusion (IOM 2020).
Turkey has been a land of diverse migrations (İçduygu and Kirişçi 2009). In the late 1970s, irregular migration became the main flow of migration to the country (Genç 2018). Since then the country has received waves of asylum seekers, the most recent wave has brought a mass migration of Syrian refugees since 2011. According to the website of Turkey's Directorate General for Migration Management, the number of Syrian refugees who are registered and under temporary protection is more than 3.6 million. 1 In addition there are many Syrian refugees who remain unregistered and are not provided any proper legal status. Such irregular migrants lack access to any kind of formally provided services, including health.
Refugees and irregular migrants were among the first to be struck by the pandemic in Turkey. NGOs struggled to reach out to them and alleviate their problems. In line with the questions of this special issue, the main focus of this study is directed toward the study of the early impact of pandemic on NGOs working with migrants in Turkey. By building on in-depth interviews, the available literature, statistics and reports, the study examines the impact of the pandemic on the services, finances, and working methods of NGOs serving migrants in Turkey, and their relations with local and central government authorities. Following and introductory section on NGOs and civil society in Turkey, the study presents the analysis of our interview process. Heper (2011) and Mardin (2006: 29) observe that a civil society in the Western senseas "a part of the society that could operate independently of central government and based on property rights" did not emerge in the late Ottoman/early Turkish polity. In line with this view, Turkey has generally been described as a country whose civil society is relatively undeveloped. If one refers to a secular modern civil society in which property rights are guaranteed by the law and which is organized around associations that raise funds by means of membership fees and collect regular donations, this approach might be accepted as true. However, if we include traditional non-governmental groups which are organized around traditional institutions taking pre-modern values as reference, and which conduct their activities around pre-modern issues and with nonsecular motivations, one comes to the realization that Turkey's civil society is not all that weak. In terms of faith-based, charity-based, hometown-based or ideology-based NGOs, civil society in Turkey begins to demonstrate itself with modern forms and in richness and vibrancy. In Turkey, this type of civil society organization easily finds empty spaces in which it may conduct its work, as both local and central government authorities are not sufficiently capable, nor have they the adaptative capacity to assure prompt resolution of problems as they present themselves. It takes considerable time before the state addresses the problem area and develops comprehensive and structural solutions. The state in Turkey, like in other countries, learns first and then reacts after some period of time. In the pursuit of their organizational and programmed goals, it is important to note that civil society organizations defer to the positions and policies of state authorities. 2 In terms of legal framework, we can distinguish six main forms of NGOs in Turkey: (1) associations (dernek), (2) cooperatives (kooperatif), (3) chambers (oda), (4) foundations (vakıf), (5) trade unions, and (6) civil servant trade unions. NGOs and non-profit organizations do not have a common legal structure but are usually established as either an association or a foundation (Kuzmanovic 2012: 8). According to Turkey's Ministry of Interior, there are 120.668 active associations (www. siviltoplum.gov.tr) and 5774 foundations (www.vgm.gov.tr) in Turkey. The majority of these associations perform activities involving social solidarity, religious services, education, and humanitarian aid, while foundations focus their activities on serving children (education included), the elderly, women, disabled persons, and the poor. Despite these forms, history, and numbers, NGOs in Turkey have struggled to survive and retain their autonomy vis a vis a strong and centralized state which did not provide much room for autonomous groups to take root and act outside its authority (Heper 2011;Keysan 2019;Keyman and İçduygu 2003). Social groups were provided a greater space only when they accommodated to the center.
NGOs and Civil Society in Turkey
In the 1920s, NGOs were strictly restricted by several laws such as the Law for Maintenance of Public Order (1925) or the Labour Law of 1934. In the 1930s, the single party regime became more repressive and all autonomous societal forces, including NGOs have been suppressed. With the adoption of a more restrictive law in 1938, the state's repressions and restrictions on civil society continued unless compliance is demonstrated to 'the center' (Ahmad 1993;Anheier and Toepler 2009;Zihnioğlu 2011). During these early decades of the Turkish Republic, traditional faith-based entities were specifically targeted and closed, and pious foundations were strictly controlled by the state through the hands of Directorate General for Foundations. As the space of civil society was narrowed, many faith-based communities like religious orders retreated from the public space and became less visible.
By mid-1940s, 'the center' started to be balanced by the growing private sector and local leaders, who joined opposition against the single-party regime. In addition to these, the end of WWII brought new international dynamics which would contribute to the end of the single-party regime in Turkey and the creation of a pluralist, multi-party system (Ahmad 1993, pp. 201-214). In this context, civil society in Turkey found liberation. Following the amendment of the Law of Associations in 1946, many associations were developed, another law recognized the right of industrial workers to organize in unions in 1947, and in 1950, the transition to multi-party regime provided a fertile ground for all civil society organizations (Ahmad 1993;Zihnioğlu 2011). Although free organization of people had taken form by the early 1950s, the state, now controlled by the Democratic Party, continued to keep control of associations, foundations, labor unions, opposing media, and all other kinds of civil society actors, and it did not let them become active within the political arena (Mardin 2006). Restrictive attitudes and repressive measures returned in the second half of 1950s; Opposition, press, and universities were suppressed while the space for civil society organizations was severely limited (Ahmad 1993).
The first decade of the multi-party system in Turkey ended with military takeover in 1960. Following economic and political instabilities during the 1960s and 1970s, two other military interventions would follow this first one, and in each of them, except the one in 1960, military regimes would restrict and repress civil society. According to Doğan (2013: 61) the military regime between 1980 and 1983 "could be perceived as the ground zero, where all democratic institutions were banned" including most of the NGOs. In a similar vein, Şimşek (2004) shares numbers, following the 1980 coup, indicating that more than 20.000 of the existing 38.354 NGOs were closed down by the military regime. In line with the restrictive and repressive approach of the military regimes, Turkish people associated civil society and the NGOs with negative meanings and they shied away from becoming active participants within civil society in the 1990s (World Values Survey 2007).
NGOs have had a rocky evolution in the Turkish society. The space allocated to them has been widened with every democratization effort, and then narrowed in every authoritarian turn. Anheier (2017, 19) shows that this is not something peculiar to Turkish society. A similar pattern has been reported in other countries with authoritarian regimes and/or illiberal democracies. Anheier's assessment of recent regulations and current proposals addressing civil society organizations in G20 countries clearly shows the increasing regulatory tendencies among these countries. In particular, it is clear that the illiberal democracies or authoritarian members of the G20 have been most restrictive on their regulations about the civil society. However, like in other authoritarian countries, civil society organizations in Turkey are tolerated as long as they provide welfare services and do not counter the hegemony of the state (Lewis 2013, 332;Spires 2011, 36).
Yet global, regional, national, and local developments in the late 1980s triggered rapid social/cultural, economic, and political change in the Turkish society. As Keyman and İçduygu (2003) note, one of the most important sites of change was civil society. While 'the strong state' was in a legitimacy crisis, civil society had been growing since the 1980s and especially during the 1990s, and it contributed to the development of a more liberal and democratic society in Turkey. Kuzmanovic (2012, p. 6) argues that the revival of the civil society in Turkey should be discussed within the context of the general revival of civil society across the globe in the 1990s. The end of Cold War advanced neo-liberalism, which presented the empowerment of the civil society as the main route to turn societies into liberal democracies. International policy actors and donors, such as the EU, the UN, and the World Bank, embraced this international policy strategy of democratization as the main way to eliminate authoritarian state tradition and to reduce poverty in Turkey (ibid). According to many civic activists, the UN Habitat II Conference, which brought together hundreds of NGOs in Turkey in 1996, was a milestone for its civil society development (Kuzmanovic 2012;Zihnioğlu 2011). Moreover, the Marmara Earthquake activated many NGOs in 1999 (Kubicek 2002). And, in the same year Turkey was declared a candidate for EU membership. As EU candidacy necessitated the introduction of a liberal legal framework for civil society, it was expected to have a transformative impact on state-society relations. In addition, in order to empower NGOs, the EU provided financial and technical assistance during the 2000s (Zihnioğlu 2011). As it is noted before, in addition to the EU, IMF and World Bank also supported NGOs. As a result, 2000s witnessed a large increase in the numbers of NGOs in Turkey.
It should also be noted that Turkey's EU candidacy and the later consolidation of political power in the hands of the Justice and Development Party (Adalet ve Kalkınma Partisi, AKP) from 'the periphery' challenged the center and provided a more liberal legislative environment for associations and foundations (Kuzmanovic 2012, 12). The new laws -the Law on Associations in 2004 and the Law on Foundations in 2008adopted by the AKP governments transformed the restrictive legal framework on civil society and provided a greater degree of freedom for NGOs. By 2000, the number of associations reached 72.378, it became 84.994 in 2010, and 120.632 in 2020 (www.siviltoplum.gov.tr, 2020). However, that increase in quantity did not transform itself in quality, and the impact of NGOs in making a real difference was limited (Kubicek 2002;Şimşek 2004). It is none the less important to recognize that civil society in Turkey underwent a substantial process of transformation in the 2000s.
NGOs in Turkey may be categorized on different criteria like their political stance, motivations, source of income, and relation to modernity. In line with these characteristics, there exist pro-government NGOs, and NGOs acting with opposition parties; faith-based NGOs such as pious foundations and religious orders, charities, secular modern NGOs; and hometown associations in the country. If we focus on their incomes we see government funded NGOs as well as foreign donor funded NGOs as well as NGOs that are supported directly and over a wide popular base in society. If we focus on their stance vis-à-vis modernity, we can identify NGOs such as women's associations, youth associations, and consumer protection associations among others, which focus their attention on specific issues and problems of modern societies. Religious orders and hometown associations, on the other hand, would be categorized as NGOs which do not appear to be directly motivated in modern directions. These non-governmental bodies frequently refer in their work to premodern values and institutions. However, these apparently historical references combine with operations in modern times and within modern institutional settings. Moreover, these associations are called upon to address effects of modernity at play within their communities. Therefore, their modes of operation can appropriately be described as hybrid.
Method
The first focus of our study is directed toward how NGOs functioned within the context of Istanbul's urban poverty during the pandemic. When we started to communicate with the NGOs that provide services for the urban poor, we found that refugees and irregular migrants had joined the ranks of most vulnerable and fragile people in the Turkish society, and that many NGOs were directing increased attention to their problems. In line with this, we directed our attention to NGOs working with migrants in Turkey. We addressed their experiences in the first 6 months of the pandemic in Istanbul. The impact of the pandemic on the focus of their services, finances, working methods, and relations with local and central government authorities make up the foci of this study.
Our study builds its analysis from an examination of interview notes, available literature, relevant sector reports, and news. We adopted qualitative research techniques and collected data through in-depth interviews. We employed purposive sampling and chose to interview representatives of NGOs which provide services to migrants (refugees, irregulars, regulars, foreign students) and urban poor. We contacted more than 15 NGOs, and found that six of them -Tarlabaşı Dayanışma, Derin Yolsulluk Ağı, Mülteciler ve Göçmenlerle Dayanışma Derneği, Mazlum-Der, Şefkat-Der, and İnsani Yardım Vakfıresponded with willingness to be interviewed. We conducted the interviews in September 2020.
We coded interview notes independently by conducting complete coding and identified broad patterns (Braun and Clarke 2013;Saldaña 2015). After several rounds of discussions, data-derived codes were clustered together according to the similarity and the patterns they formed, and the following themes, which are presented in the analysis, were created: mismatch in the field, digitalization of the NGOs, tense relationship with the central government, neglect by local governments, political polarization as an obstacle, difficulty of accessing healthcare, and xenophobia fueled by the pandemic.
COVID-19 in Turkey: Initiatives of Neighborhoods and Governments
As the number of positive cases increased, the initial response of the Turkish government was the declaration of a curfew between 10 and 12 April 2020. This first curfew decision was followed by others. The aim was to decrease the pace of the spread of virus in the society. This measure was a new experience for the majority of the society. The curfew created new and unexpected problems, which aggravated the negative effects of the pandemic. Its side effects would be recognized soon. People who were 65 and older were placed within the risk category and were not at all allowed to leave their places of residence. They were therefore not able to meet their own most basic needs. Neighborhood initiatives were initiated and volunteers in these loose neighborhood networks offered to do errands for their elderly neighbors, such as shopping or paying bills. This initiative created positive community spirit and widely appreciated as a positive element among all negativities of the pandemic. Traditional solidarity networks of Turkish society also became activated, but in the more urban locations they met only a fraction of the actual need. Clearly, a need existed for new and innovative solidarity activities.
The society expected the central government and local governments to act in this troublesome period, to provide new services and heal the wounds of the pandemic. The central government initiated a nationwide aid campaign to raise funds that would be spent for people in need. This campaign was supported by a number of pro-government business groups, the Turkish Chamber of Commerce, and some other entities affiliated with the government. Opposition groups criticized this initiative, observing that government is the established tax collector. These groups observed that asking taxpayers to donate the government, even though for the purpose of aiding people, was something odd, providing an unrealistic expectation of voluntary financial and physical efforts. The aid campaign "Biz bize yeteriz" (We are self-sufficient, Turkey) raised 1.960.239.000 TL (around 275.000 USD) and distributed aid to 1.960.239 families (1000 TL per family) (https://bizbizeyeteriz.gov.tr/).
Local governments are not customarily considered to be part of civil society. In the Turkish case, however, they are significant actors in the search for autonomy from the central government. Although they are kept under the tutelage, monitoring and control of the strong centralized government, they harbor centrifugal forces, groups, and ideas. In other words, they provide venues for groups at the periphery to search for alternatives from the strong-center (Erder and İncioğlu 2008). One political party, the Justice and Development Party (Adalet ve Kalkınma Partisi, AKP) has been ruling the country since 2002. In the 2010s, AKP governments have taken steps to further centralize the state, and later transformed the parliamentary system into a heavily centralized executive-style presidency in 2018. Despite this establishment of centralized power, for the first time since 2002, the AKP lost power in three major cities in 2019 local elections: Istanbul, Ankara, and İzmir, which constitute about half of Turkey's GDP (Sancak 2019). Not only in these cities, but control of municipalities of other major cities such as Adana and Antalya were also passed to opposition parties. The first year in office of these mayors, especially in Istanbul and Ankara, has been marked by tensions with the central government under President Erdoğan (Cagaptay and Yüksel 2020). On top of this, before their first year at office was over, COVID-19 was confirmed to have reached Turkey on 11 March 2020.
The opposition mayors, who now governed the most populous cities of Turkey, called for immediate containment measures. They 'spearheaded social efforts to battle the pandemic, launched fundraising campaigns, and encouraged citizens to wear masks in public' (Cagaptay and Yüksel 2020). As they were not rich with resources, nor were they able to levy new taxes, their fund-raising campaigns were to a certain extent considered normal compared to the central government's campaign. According to Moubayed (2020), the central government did not cooperate with opposition-run municipalities and it tried to limit their ability to distribute social assistance. In line with these efforts, the central government's response to calls for pandemic containment was delayed and fundraising campaigns launched by opposition mayors were declared illegal and banned on March 30. State banks blocked donation accounts, while President Erdoğan asserted that only the presidency had the right to raise funds (Butler 2020). The following day, Erdoğan introduced the central government's fundraising initiative (We are self-sufficient, Turkey -Biz bize yeteriz).
The opposition mayors resisted these limitations and pressures and spearheaded solidarity initiatives to provide assistance for individuals affected by the lockdown. One of these solidarity initiatives involved the campaign titled 'paying a neighbor's bill' (Askıda Fatura). After being blocked by the central government, the Istanbul Metropolitan Municipality, under Mayor Ekrem İmamoğlu's rule, started this initiative, in which a digital payment portal is provided for those citizens who want to pay water and gas bills of fellow burghers who had lost their livelihoods because of the pandemic. The municipality guaranteed donors that the invoices belonged to the people in need, developing a need evaluation system to provide this assurance. With this practice the municipality also increased its revenues as unpaid city water and gas bills were paid. Thus, was accomplished a triple win, not only for citizens and government, but later on, families in need, as the Istanbul Metropolitan Municipality added 'family support' and 'mother-kid support' packages to this initiative. In the period between May 4, 2020 and September 14, 2020, 180.174 bills amounting to more than 3.8 million dollars were paid in this way (https://askidafatura.ibb.gov.tr, Sept 15, 2020).
The Ankara Metropolitan Municipality launched a 'hot-meal for everyone campaign' on April 2, 2020, and started to deliver hot meals to people older than 65, health-care personnel, and people in need in the capital city. As the municipality's fundraising campaign was blocked by the central government, the opposition mayor, Mansur Yavaş, called upon companies, philanthropists, and other municipalities to make in-kind contributions to the municipality's food bank. Another initiative of Yavaş was 'An iftar from me'. Islam's holy month of fasting, Ramadan, was in May, and the municipality acted as a bridge between the needy and philanthropists willing to buy an iftar (fast-breaking evening meal in the Ramadan) for a person in need. Citizens could send an iftar by paying 15 TL (around 2 dollars) online to a website launched by the municipality: www.iftarver.com. Though the campaign started in Ankara, people from different cities started to make donations and the campaign was extended to all Turkey in a very short period of time. According to the web site of the municipality, 35.087 philanthropists funded 264.542 iftars in Ankara, and in total 73.787 people sent an iftar to more than a half a million people in all cities of Turkey (https://www.iftarver.com, Sept 16, 2020).
A limitation of these imaginative and innovative donation programs involved the very small (if any) number of refugees and irregular migrants who were able to benefit from these initiatives. The priority of local governments was their own citizen burghers. Moreover, it should also be noted that opposition parties, and therefore opposition mayors, have been critical of the central government's migration policies, and we believe that this contributed to the reflection of economic and social distress in the form of negative attitudes about migrants in the society (Erdoğan 2019).
NGOs during the Pandemic
According to the Third Sector Foundation of Turkey (Türkiye Üçüncü Sektör Vakfı, TÜSEV), a support network of over 100 associations and foundations, the spread of COVID-19 has adversely affected the activities of NGOs. In order to determine the problems experienced in this period, the network conducted a survey on the impact of COVID-19 outbreak on NGOs operating in Turkey (TÜSEV 2020). Their answers show that the main impact of the COVID-19 was on the activities organized by the NGOs: a majority of them had to cancel seminars, conferences, and performances; many noted that they had canceled training activities; and 43% of the respondents reported that they could not provide services because they had to end their field work and close down service centers. About one-third of the NGOs reported that they had to stop all of their activities, while 83% of the respondents reported an increase in demand for their services. Almost half of them noted that pursuing activities in other areassuch as healthcare needs, healthcare rights, online working systems training, poverty, and violence against women-came into their agenda. The pandemic affected aids and donations to the NGOs, too. Half of the respondents reported a decrease in the aids and donations they received (ibid.). When we complement this survey's findings with in-depth interviews with representatives from NGOs, several recurrent themes about the experiences of NGOs during the pandemic in Istanbul become clear. These themes are discussed in the following paragraphs.
Firstly, the interviewees reported a mismatch in the field: the number of people who asked for their services increased immensely. While the number of potential recipients has increased 'at least more than 10 times', the level of funding available decreased. According to the interviewees, refugee mothers and their children were desperate. In addition to refugees and asylum seekers, irregular migrants started to ask for their services. They asked for food, shelter, and help to access to healthcare.
We know how Turkish citizens suffer from the pandemic. If the Turkish citizens are experiencing these difficulties, think about the immigrants, refugees, asylum seekers, especially those who are not registered. They are the most vulnerable group, they do not have any legal status, they are completely devastated. We can say that their number has increased many times. (Şefkat-Der, September 21, 2020).
The interviewees noted that irregular migrants were not among the people who asked for their services before the pandemic. In line with the studies, they described them as young and healthy people, who had been employed informally in different sectors. When the pandemic hit Turkey, curfew affected all sectors of economy including the informal, and it left many irregular migrants unemployed. According to the interviewees, under these circumstances many irregular migrants returned to their countries (if they were allowed and able), and many among those who deliberately stayed or found themselves stuck in Turkey began to ask for services.
After the pandemic, the numbers have seriously changed. I can say that the number of undocumented people who asked for help has increased by 10,000 percent. The number has increased a lot. This means that these people were really desperate. Among these, the number of young [healthy] people with strong physique who applied to us increased. People who are between the ages of 18-35 turned to us and asked for support. This age group does not appeal, this is very rare. (Tarlabaşı Dayanışma, September 23, 2020).
According to the interviewees, organizational budgets decreased as traditional donation flows dried up as well. Some former donors found that they had been transformed into people who themselves needed help.
Turkish citizens knew what to do during the curfew. They stayed in their houses, and while the government offices are not working, their applications have decreased. But there was no decrease in the applications of foreigners. Their problems continued and got worse. [On the other side] businesses were [also] affected. Donors could not donate, they could not support or sponsor to the activities of Mazlum-Der. Sources decreased, our expenses have also decreased. This way, we could survive. (Mazlum-Der, September 22, 2020).
Secondly, the interview notes revealed the digitalization of the NGOs. All of the interviewees expressed that they had to rely on online communication channels in order to serve their community during the lockdown. All have explained that social media channels have been used intensively. They set up WhatsApp groups, and collected needs through them. The networked organizations shared awareness-raising messages about COVID-19 via their Facebook and Instagram pages and updated their websites. Their beneficiaries asked for help through WhatsApp, and NGOs tried to link them with the experts and assistance through these channels. In several cases, they organized counseling or meeting sessions via Zoom. Though one of the interviewees asserted that the effectiveness of their work/service decreased when they worked online, it appeared to us that they did their best and adapted to a new online working environment very quickly.
Thirdly, almost all interviewees, except the pro-government ones, explained that they had a tense relationship with the central government. All interviewees have noted that they were not able to receive any funding or any kind of contribution from the central government during the lockdown. Several interviewees noted that local branches of the central government ignored their effectiveness in the field.
We explained this through social media and other media channels, we went to the district governors. The governors did not give us an appointment. District governors have insulted us. We did not want to seem to be doing great things, we did not say such a thing. We just tried to give recommendations about these people, they are desperate. (Tarlabaşı Dayanışma, September 23, 2020).
None of the interviewees, including the pro-government ones, reported a close relationship with the central government and they described the relationship between NGOs and the central government as a rocky one. We understand that following the transformation of the parliamentary system into a heavily centralized executive-style presidency in 2018, the state became more authoritarian, and the legal and regular venues for the flexible development of NGOs have shrunk. In referring to this development, interviewees underlined the difficulties they came across in their fundraising activities. Although they had the right to raise funds for their services under law, according to several of them the central government controlled, limited, and sometimes prohibited their fundraising activities.
If we hadn't been careful, they might have stopped our work somehow. You know that the Metropolitan Municipality started a fund raising initiative, then the central government stopped it. Perhaps my field experience helped us. I tried to act cautiously, [I told that] we should act very slowly, let's not be visible, we are a network of solidarity. (Derin Yoksulluk Ağı, October 3, 2020).
One of the interviewees explained that when an NGO publicly criticizes the government for its policies, visits from Ministry of Interior inspectors to examine their papers, and possibly even prosecution, may be expected.
The target should be zero in violence. Other than that, everything else is a failure.
[…] I criticized the government and I was put on trial for that. I mean, I was sued and I was even found guilty. When you criticize the officials, the Directorate of Associations and the Ministry of Internal Affairs send their inspectors immediately. But we talk a lot, I criticize it because this is one of the most important features of our association: justice. (Şefkat-Der, September 21, 2020).
Moreover, in line with the centralized strong-state tradition, the interview notes revealed that acting ahead of the state seems unacceptable to the central government. We concluded that, whenever this occurs, even though the organization may be providing services to vulnerable and fragile people that the state fails to serve, the central government moves to deactivate the NGO because it is seen to cause the perception that the state is incapable of providing those services: Even for those associations which are not political can have problems. Officials do not like an atmosphere of 'we are doing better than the state'. (Şefkat-Der, September 21, 2020).
In line with their experiences in the field, interviewees criticized the central government for not striving enough to reach out to migrants and refugees. Though Turkey's open door policy for the Syrian refugees is widely celebrated within the NGO community, most of its members criticize aspects of this policy-the temporary protection status, detention centers, the decision to open the borders for migrants to cross into the European Union, and repatriation of irregular migrants and some Syrians: Turkey gives refugees a special status. This is not in compliance with international law. Statuses such as temporary protection, actually these statuses are made up to bypass the refugee status. Groups, which were once granted special statuses for certain political considerations, can suddenly become targets [of anti-immigrant attitudes, political discourses] when the conjuncture changes. Mazlum-Der is trying to prevent deportation of migrants from Detention Centers. We are trying to help Syrians for their residence permits […] (Mazlum-Der, September 22, 2020).
These critical views held that migration policies of the government became more impractical during the lockdown as they failed to provide services to refugees and migrants in the field. In addition, government agencies working on migration were temporarily closed. Interviewees explained different outcomes of this temporary closure. According to the interviewee from Göçmen Dayanışma Derneği, the closure/lockdown of the government agencies paralyzed the services provided to the refugees. 3 Differently from this, the interviewee from Mazlum-Der explained that refugees and irregular migrants' rights could not be violated as the government and law enforcement agencies were inactive. As he explained as an example, the irregular migrants could not be deported during the lockdown. 4 When the interviewees were asked about their relationship with the local governments before and during the pandemic, they reported that they were neglected by them. All interviewees except one explained that they did not receive any support from local government. The interviewee from the Göçmen Dayanışma Derneği (Migrant Solidarity Network) stated that they cooperated and were supported by the local government in their operations and activities. 5 But other than this, they did not receive any funding from the local authorities. When we asked about the Istanbul and Ankara Metropolitan Municipalities' campaigns -'paying a neighbor's bill' and 'hot meal for everyone', interviewees noted that these campaigns, 'paying a neighbor's bill' in particular, asked personal details of the people in need. Therefore irregular migrants were afraid and did not apply for the assistance in paying their bills. 6 Municipalities wanted to consult with us once. We shared our practices, but similar to the NGOs, municipalities in Turkey also do not have the infrastructure, the legal bases, and the understanding to provide solutions. This may happen because of the bureaucracy, or because the municipality is governed by the opposition party. They are not effective in the field. […] For example 'paying a neighbor's bill' campaign asked for too much details. The person is already a refugee and he is afraid. If I asked this person to fill the online form with all his information, he would be suspicious of me. (Tarlabaşı Dayanışma, September 23, 2020).
The interviewee from Tarlabaşı Dayanışma explained their model of organization in the field and how they linked philanthropists with irregular migrants and refugees in need. He explained that they shared this model with the local government but found that their input was neglected. 7 The other interviewees also shared similar experiences about the attitude of the local governments. Differently from all interviewees, the interviewee from Derin Yoksulluk Ağı stated that their model to match donors and the people in need was later adopted by the Istanbul Metropolitan Municipality. 8 But, in general, the interviewees reported that local governments failed to provide services to the people including the migrants: Let me say this, I am not trying to discredit the mayor, but all municipalities, including ours, including İstanbul Metropolitan Municipality, failed. I am saying this frankly. Because the hunger of a baby cannot wait for the bureaucracy, for the legislation. You throw that legislation to the garbage and do something. The person calls me and says that she can give only water with sugar to her baby. This is hunger. […] After the pandemic all local governments will need to think about what is wrong with their social services legislation. They should question what is wrong [so] that they could not provide any services. But they do not do that, they say they did a great work! (Derin Yoksulluk Ağı, October 3, 2020).
Interviewees reported political polarization as a main obstacle for their operations. Polarization has been one of the key characteristics of Turkish politics for decades. However, recent public opinion surveys report an alarming level of polarization in the Turkish society (KONDA Barometresi Kutuplaşma 2019: 11; KONDA 2018). 9 As the interview notes reveal, polarization has been counterproductive for the NGOs. The interviewees stated that they had to act very cautiously and that they tried to remain out of politics. Otherwise they were suppressed by the central government, their applications delayed, their fundraising campaigns declared illegal, and their usage of local government facilities restricted, etc. Two interviewees expressed that when they criticized the government, their accounts were examined by the inspectors from the Ministry of Interior. Even the interviewees from NGOs known as pro-government claimed that whenever they criticized government or state institutions, they felt they were being ostracized by the relevant state authorities and their networks. Similarly, the NGOs which are affiliated with the opposition shared similar experiences with the local governments governed by opposition parties. Their statements revealed that political polarization narrowed their space and affected their operations negatively. In the end, all interviewees underlined that they wanted to be autonomous from politics and asserted their wish to provide services for their target groups without being affected by the political polarization in the country.
Another issue related to the high degree of politicization of their activity area was identified by the interviewees as the difficulty of accessing healthcare for migrants during the pandemic. They noted that access to healthcare was already very difficult and problematic for the refugees and irregular migrants in Turkey, and that the pandemic problematized it further. 10 The biggest risk lies with the undocumented migrants right now. Their situation is worse than terrible. I cannot find a word to describe. Here we are trying to do something with the special efforts of some NGOs and some individuals, but it is not possible to catch up with their needs. I cannot explain it. I think about their state of health, their sicknesses. I know that many of them cannot even go to the hospital because they hesitate.
[…] Undocumented migrants are already afraid to go to hospital, they cannot take the PCR test. The doors are closed. They go only if it is a matter of life and death (Şefkat-Der, September 21, 2020).
Though Syrian refugees, who are registered and under temporary protection, possess the right to access health services, unregistered Syrians and irregular migrants do not have access to regular health care. As the interviewees noted, the governorship of Istanbul took two important decisions for the Syrian refugees in Istanbul in July 2019: it stopped the registration of Syrians in Istanbul and removed those registered in other cities of Turkey to their original registration cities. These decisions paralyzed many Syrians who have been trying to settle in Istanbul. As they were already in fear about being sent to other cities, they were already experiencing difficulties blending into society before the pandemic. They tended to access healthcare only through informal channels such as informal clinics set up by Syrian doctors and sometimes through private clinics and private hospitals which turned a blind eye toward them. Similarly to the situation of the Syrian refugees, who are not registered in Istanbul, other irregular migrants do not have legal access to healthcare. When it is urgent, they seek health in private clinics and private hospitals. 11 Interviewees reported that they prepared information leaflets in different languages, sent awareness-raising messages through WhatsApp groups, distributed masks, hand sanitizers, and sanitary gloves. Moreover, these NGOs tried to find informal solutions to the health needs of the people who asked for their services. They directed them to the private hospitals with which they had been cooperating. They organized Zoom sessions with doctors. In cases where online solutions did not work, they directed persons in need of service to doctors in their informal networks. When we asked whether there were any COVID-19 cases among the people who asked for their services, none of the interviewees responded. But they all asserted that refugees, and irregular migrants in particular, were so afraid that they would never go to a hospital if they thought that they carried the symptoms. 12 Lastly, almost all interviewees have noted that the pandemic fueled xenophobia in Turkey. They underlined that anti-immigration attitudes have been on the rise in Turkey in general and in Istanbul in particular. They shared different anecdotes about the xenophobic incidents that refugees and migrants have been experiencing in the form of anti-immigrant attitudes. According to the interviewees, due to the economic crisis in the country, xenophobia was already on the rise before the pandemic. However, in their view, as the economic crisis worsened in the onslaught of the pandemic, anti-immigrant attitudes increased and xenophobia also rose. 13 We were led to understand that this development did not only affect the fundraising activities of the NGOs, but also their operations in the field. Several of the interviewees told of experiencing verbal attacks directed toward them while they were in the field distributing food and other needs to irregular migrants during the lockdown: We feel it in the field. Once we were in a street, we were talking, motivating irregular migrants, who are African, a lady from the next building shouted us 'go away, you scum' […] Then police came. She called them. There are many others, they threaten us, they do not want us to visit their neighborhoods, they warn us about migrants because 'they steal their jobs'. Numerous messages were sent from social media. […] In addition, during the pandemic, landlords expelled migrants from their homes, by saying that "go away you scum, you will infect us". People were thrown into the streets. They were paying their rents. They give no harm. They do not even leave the house. (Tarlabaşı Dayanışma, September 23, 2020).
In addition, two of the interviewees shared the xenophobic messages written to their Facebook pages with us. The words of the interviewee from Şefkat-Der described the position of the majority of the Turkish society about immigration in Turkey and also their approach to the operations of NGOs working in this field: I have been working in the civil society for more than twenty-five years. For the first time in the last fourfive years, the responsiveness of the society has fallen behind the responsiveness of the state. Turkish people used to pity migrants, refugees, war victims, Africans, etc. and they pressured the state to act. They used to be more responsive than the state. Now they see Syrians and Afghans as the reasons of their economic and other problems. And unfortunately, [there is] discrimination, exclusion, and otherization of the migrants. 14
Conclusion
In the light of our analysis, we conclude that, the pandemic and the overlapping economic crisis have deeply affected Turkey's lower classes, underclass, and invisiblesits refugees and irregular migrants. NGOs which are providing services to these groups have been overwhelmed by the number of people who sought their services and by the tasks asked from them. In response to these challenges, with decreasing funds in hand, they have been required to change their modus operandi and adopt new strategies, such as using technology to collect needs, disseminate information about the pandemic, raise funds, and link donors with recipients. In so doing, they have not been able to attract visible support from central and local government authorities. They find themselves neglected.
In times of crisis as those of pandemic, pluralist and social democratic visions tend to fade, and governments seek to take advantage of emergencies to further strengthen their positions vis-a-vis civil society by further advancing authoritarian forms and structures. In the Turkish case, in which the state is already marked in the 'strong state' tradition and by the recent transformation to a heavily centralized executive-style presidency, the room for civil society maneuver has already been narrowed. As the interview notes reveal, the pandemic put the latest nail in the coffin. The central government declared fundraising campaigns of municipalities illegal, and tried to control the distribution of social assistance for persons affected by the lockdown. Though expected to play an implicit, invisible role in general, pro-government NGOs sought to act with the authorities during the pandemic. The work of almost all NGOs has been governmentally neglected, except these pro-government ones. This differential approach to the NGOs is closely linked with the political polarization in the country but all interviewees, regardless of the ideological position of their NGO, complained in a varying degree that they were not able to operate fully in the field due to the polarization.
The NGOs whose work is central to this paper have diligently and persistently struggled throughout the pandemic with the pre-existing problem of facilitating access to healthcare among Turkey's migrants. They devised new strategies and acted as bridges between refugees and irregular migrants in need and government authorities and health system. Our study indicates, by noting deficiencies of the migration policies, the steady rise of xenophobia in Turkey. The testimony of our respondents indicates how fully the economic crisis had already affected Turkish attitudes about migrants, and this situation has now been considerably worsened with the impact of the pandemic. Our interviewees describe a crisis in which their efforts to serve a population in need becomes increasingly difficult. Both central and local government authorities have been and continue to be slow to respond to the crises occasioned by both migration and poverty. These governmental institutions remain distant from populations that desperately need their service. Whatever their party, they have failed to comprehend obligations to assist their national and municipal constituents and residents in need. And, all the while, NGOs from diverse ideological backgrounds continue with their efforts to lessen the destructive effects of the pandemic in the field, the resources dedicated to their efforts in steady decline.
When we put Turkish experience into a global context, we note that in line with the experiences of civil societies of similar regimes, Turkish civil society suffers from authoritarian practices of individual and institutional government authorities in particular and the political system in general. How the NGOs respond to these restrictive and repressive regimes has been described by Anheier (2017) and others, and. the Turkish case generally follows the global pattern. However, this study notes that Turkey has a big, vibrant, strong and nominally less politicized civil society which is mainly preoccupied with practical problems. Those problems are addressed very slowly by the state, which requires a long learning and adaptation process. We think that this opens a window of opportunity for civil society actors who act as pioneers in a problematic area until the state learns and adapts itself. When the state inserts itself into the field, civil society actors tend to withdraw. And over time, the state narrows down the space that was once filled by civil society. The Turkish case shares this scenario with other similar political regimes. Still, the unique solutions, reaction speed, and style of Turkish civil society organizations in the Covid-19 experience may offer lessons to other countries. | 2021-02-06T14:09:53.696Z | 2021-01-01T00:00:00.000 | {
"year": 2021,
"sha1": "a4818e919233f086b427c7d5fac2c47c0bcd582b",
"oa_license": "CCBY",
"oa_url": "https://www.degruyter.com/document/doi/10.1515/npf-2020-0049/pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "6281bfffbc3d938552d1adcdc9d3ea49b8343bdf",
"s2fieldsofstudy": [
"Political Science"
],
"extfieldsofstudy": [
"Political Science"
]
} |
18950629 | pes2o/s2orc | v3-fos-license | Growth hormone (GH)‐transgenic insulin‐like growth factor 1 (IGF1)‐deficient mice allow dissociation of excess GH and IGF1 effects on glomerular and tubular growth
Abstract Growth hormone (GH)‐transgenic mice with permanently elevated systemic levels of GH and insulin‐like growth factor 1 (IGF1) reproducibly develop renal and glomerular hypertrophy and subsequent progressive glomerulosclerosis, finally leading to terminal renal failure. To dissociate IGF1‐dependent and ‐independent effects of GH excess on renal growth and lesion development in vivo, the kidneys of 75 days old IGF1‐deficient (I −/−) and of IGF1‐deficient GH‐transgenic mice (I −/− /G), as well as of GH‐transgenic (G) and nontransgenic wild‐type control mice (I +/+) were examined by quantitative stereological and functional analyses. Both G and I −/− /G mice developed glomerular hypertrophy, hyperplasia of glomerular mesangial and endothelial cells, podocyte hypertrophy and foot process effacement, albuminuria, and glomerulosclerosis. However, I −/− /G mice exhibited less severe glomerular alterations, as compared to G mice. Compared to I +/+ mice, G mice exhibited renal hypertrophy with a significant increase in the number without a change in the size of proximal tubular epithelial (PTE) cells. In contrast, I −/− /G mice did not display significant PTE cell hyperplasia, as compared to I −/− mice. These findings indicate that GH excess stimulates glomerular growth and induces lesions progressing to glomerulosclerosis in the absence of IGF1. In contrast, IGF1 represents an important mediator of GH‐dependent proximal tubular growth in GH‐transgenic mice.
Introduction
During the last decades, a significant role of the growth hormone (GH) and insulin-like growth factor 1 (IGF1) system in kidney development and function, and in the pathogenesis of chronic kidney diseases (CKD) and their sequelae, has become apparent (Wolf et al. 2000;Schrijvers et al. 2004;Mak et al. 2008;Kumar et al. 2011;Kamenicky et al. 2014;Bach and Hale 2015). In different entities of CKD, including diabetic nephropathy (DN), progressive glomerulosclerosis is the common histopathological feature, and is regarded as the determining mechanism driving the progressive loss of func-tioning nephrons and subsequent renal scarring, finally leading to terminal renal failure (Klahr et al. 1988;Ichikawa 1989, 1991;el Nahas 1989). There is strong evidence for an involvement of the GH/IGF1 axis in induction, maintenance, and progression of glomerular and podocyte hypertrophy and damage, which represent the pathogenetic key lesions triggering development of progressive glomerulosclerosis (Fogo and Ichikawa 1991;Wanke et al. 2001;Kriz 2002;Pavenstadt et al. 2003;Reddy et al. 2007;Kumar et al. 2011). The proceeding enlargement of glomeruli and podocytes induces a characteristic sequence of alterations, finally leading to obsolescence of the glomerulus, atrophy of the corresponding nephron, and subsequent tubulointerstitial lesions (Fogo and Ichikawa 1991;Wiggins 2007). Glomerular hypertrophy is characterized by increased numbers of glomerular mesangial and endothelial cells and is commonly associated with increased matrix deposition in the glomerular mesangium (Schrijvers et al. 2004;Thrailkill et al. 2009). Unlike mesangial and endothelial cells, visceral epithelial glomerular cells (podocytes) are postmitotically fixed and therefore not capable of hyperplastic growth responses (Kriz 1996;Wanke et al. 2001;Wiggins 2007). Along with the progressive enlargement of the glomerular volume, the mean podocyte volume increases and a characteristic sequence of podocyte alterations subsequently develops, including podocyte foot process effacement, impairment of the glomerular filtration barrier, albuminuria, detachment of podocytes from the glomerular basement membrane (GBM), synechia formation between the capsule of Bowman and denuded GBM areas of the glomerular tuft, capillary and extracapillary hyalinosis, collapse of glomerular capillaries, and misdirected filtration into the periglomerular interstitium with subsequent inflammatory and fibrotic tubulointerstitial lesions and atrophy of nephrons (Kretzler et al. 1994;Kriz 1996Kriz , 2002Kriz , 2012Kriz et al. 1996;Wanke et al. 2001;Kriz and LeHir 2005;Wiggins 2007). Much of our understanding and knowledge of the pathogenetically relevant processes in development of glomerulosclerosis was gained by studying the nephropathy of GH-transgenic mice (Doi et al. 1988(Doi et al. , 1990Wanke et al. 1991Wanke et al. , 2001Wolf et al. 1993;Yang et al. 1993;von Waldthausen et al. 2008). GHtransgenic mice exhibit permanently elevated systemic concentrations of GH and IGF1. They display a typical spectrum of GH excess associated alterations, such as disproportionately stimulated body and organ growth (Wanke et al. 1991(Wanke et al. , 1992Le Roith et al. 2001) and diverse characteristic organ lesions (Doi et al. 1990;Wanke et al. 1991Wanke et al. , 1992Wanke et al. , 2001Stefaneanu et al. 1993;Wolf et al. 1993;Wolf and Wanke 1997;Dirsch et al. 1998;Miquet et al. 2008;Blutke et al. 2014). In the kidneys, GH-transgenic mice reproducibly develop the full spectrum of the lesions delineated above, from induction of overall renal growth and glomerular hypertrophy to age-related development of progressive glomerulosclerosis and tubulointerstitial lesions, finally leading to terminal renal failure (end-stage renal disease) (Wanke et al. 1991(Wanke et al. , 2001Wolf et al. 1993). However, the exact roles of excess GH and of subsequently elevated IGF1 in stimulation of renal and glomerular growth in GHtransgenic mice are yet not completely clarified. Since GH-transgenic mice have been and will continue to be used as valuable models enabling the study of diverse aspects of different human nephropathies, including the essential involvement of the GH/IGF1 axis in the development of glomerulosclerotic alterations (Reddy et al. 2007; Kumar et al. 2010Kumar et al. , 2011Grunenwald et al. 2011) and the testing of new therapeutic strategies, it is of crucial importance to further improve the understanding of the contribution of single factors of the GH/IGF1 axis to the pathomechanisms underlying the nephropathy in the model of the GH-transgenic mouse. Therefore, the present study used a unique collective of IGF1-deficient GH-transgenic mice (I À/À/ G) (Blutke et al. 2014) to dissociate IGF1-dependent and IGF1independent excess GH-stimulated morphologic growth effects on different renal compartments, nephron structures, and cell types in vivo.
Morphometrical and quantitative stereological parameters
Volume densities (fractional volumes)
Equations-General parameters
Volume densities Area density of the structure X in the reference compartment Y ∑A (X) /∑A (Y) Quotient of the cumulative section area of the structure X in all examined kidney sections per case and the cumulative section area of the reference compartment Y in the same sections ∑Pt (X) /∑Pt (Y) Quotient of the total number of points hitting section profiles of the structure X in all examined sections per case and the total number of points hitting the reference compartment Y in the same sections Pt Pt(X/Y) Point density of the structure X in the reference compartment Y
Volumes
V (X, Y) Total volume of the structure X in the reference compartment Y V V(X/Y) Volume density of the structure X in the reference compartment Y V (Y) Total volume of the reference compartment Y in both kidneys Numerical volume densities Cumulative number of all counted elements (Q À ) of the structure X in all examined disectors per case h Disector height, that is, distance between the reference and the "look-up" section ∑A (Y) Cumulative area of the reference compartment sections in all examined disectors per case (for each disector, the mean of the reference compartment section areas in the reference and the "look-up" section is used for calculation) h 9 ∑A (Y) Cumulative volume of all disectors examined per case f s Linear tissue shrinkage factor (0.91 for GMA/ MMA-, and 0.95 for Epon-embedded murine kidney tissue) Numbers N (X, Y) Number of individual elements of a distinct structure X in the reference compartment Y N V(X/Y) Numerical volume density of elements of the structure X in the reference compartment Y V (Y) Volume of the reference compartment Y
Ethics statement and animal housing
All experiments were approved by the author's institutional committee on animal care, were carried out in accordance with the German Animal Protection Law, and conformed to international guidelines on the ethical use of animals. All efforts were made to minimize the number of animals used and their suffering. All animals investigated in the present study were maintained under specified pathogenfree conditions in a closed barrier system on a 12:12-h light-dark cycle and had free access to a standard rodent diet (V1534; Ssniff, Soest, Germany) and tap water.
Urine analyses
Spot urine samples were taken weekly (from week 4 to 11) between 2:00 and 3:00 PM and immediately stored at À80°C until assayed. Per week, at least five urine samples were collected from mice of each investigated genotype and sex. Urine creatinine concentrations were determined using an automated analyzer technique (Hitachi, Merck, Darmstadt, Germany). For SDS-PAGE, urine samples were diluted to a creatinine content of 1.5 mg/dL. Urine proteins were temperature denatured (Thermoblock TB1, Biometra, Germany) and separated using a 12% SDS-PAGE gel (Protean III, Bio-Rad, Munich, Germany) together with a broad molecular weight standard (Bio-Rad) and a mouse albumin standard (Biotrend, Cologne, Germany), as described previously (Herbach et al. 2009). Coomassie blue staining of gels was performed according to a standard protocol.
Body and kidney weights, kidney processing, histology, and electron microscopy At 11 weeks of age, I À/À (5/3), I À/À /G (3/4), I +/À (5/5), I +/À /G (5/5), I +/+ (5/5), and G (5/5) (n = male/female, if not stated differently) mice were sacrificed. After determination of body weight (Blutke et al. 2014), mice were killed by cervicocranial dislocation and immediately perfused with neutrally buffered 2.5% glutaraldehyde solution through the heart, as described previously (Wanke et al. 2001;Herbach et al. 2009). After immersion fixation in 2.5% glutaraldehyde solution, the kidneys were removed carefully, separated from adjacent tissues, blotted dry, and weighed to the nearest weight (mg). The density of the glutaraldehyde-fixed kidney tissue was determined according to the method described by Scherle (Scherle 1970) and consistently corresponded to 1.05 g/ cm³. Next, the kidneys were cut perpendicular to the longitudinal axis (20 AE 5 approximately 1-mm thick slices in both kidneys), routinely processed and embedded in plastic for light microscopy and transmission electron microscopy (TEM) as described previously (Hermanns et al. 1981;Hoeflich et al. 2002). For TEM, three 1-mm³ samples of the renal cortex per animal were selected by systematic random sampling (Nyengaard 1999), postfixed in 1% osmium tetroxide, and routinely embedded in Epon resin. Per case, 10 consecutive 0.5-lm thick semithin sections were cut from each Epon block and stained with toluidine blue O and safranin. For TEM, ultrathin sections (70-80 nm) were cut, stained with uranyl citrate and lead citrate, and examined (EM10, Zeiss, Eching, Germany). The remaining kidney slices were then embedded in plastic, containing hydroxymethylmethacrylate and methylmethacrylate (GMA/MMA, Sigma-Aldrich Laborchemikalien GmbH, Seelze, Germany), as described previously (Hermanns et al. 1981). For qualitative and quantitative morphological analyses, plastic sections with a nominal section thickness of 1.5 lm were cut on a Reichert-Jung 2050 rotary microtome (Leica, Wetzlar, Germany) and stained with hematoxylin and eosin, periodic acid Schiff (PAS), and periodic acid silver methenamine (PASM). For counting and sizing of glomeruli, 140 consecutive sections with a nominal section thickness of 1 lm were additionally cut from each GMA/ MMA block per case, using a Microm HM 360 rotary microtome (Microm, Germany). From each section series, every 20th section was systematically randomly selected (6 AE 2 kidney sections per case) and stained with PAS.
Quantitative stereological analyses
Design-based stereological methods, such as the physical disector and point counting (Sterio 1984;Howard and Reed 2004) were used for determination of the volumes of distinct renal zones and nephron segments, of the total number of glomeruli, the number and volume of proximal tubular epithelial cells and of different glomerular cell types, and of glomerular capillary volumes and lengths. Plastic resin-embedded samples of kidney tissue were used for quantitative stereological analyses, and stereological estimates were corrected for embedding-related tissue shrinkage.
Total volumes of the kidney, renal zones, and distinct nephron segments
The kidney volume was obtained by dividing the kidney weight by kidney density. The cross-sectional areas of the kidney (Kid) and of individual renal zones in the kidney (Kriz and Koepsell 1974;Kriz and Bankir 1988), that is, the cortex, the outer stripe of the medulla (OSM), and the inner stripe of the medulla together with the inner zone of medulla (ISM/IZM) were planimetrically determined in micrographs of PASM-PAS-stained GMA/MMA kidney sections (20 AE 5 per case), using a Videoplan TM image analysis system (Zeiss-Kontron, Eching, Germany). The volume fractions of the cortex (V V(Cortex/Kid) ), the OSM (V V(OSM/Kid) ), and the ISM/IZM (V V(ISM/IZM/Kid) ) in the kidney were calculated according to the principle of Delesse (Howard and Reed 2004). The absolute volumes of the individual renal zones (V (Cortex, Kid) , V (OSM, Kid) , V (Cortex+OSM, Kid) , V (ISM/IZM, Kid) ) were calculated by multiplication of the volume fractions of the respective zones in the kidney with the total kidney volume. The fractional volumes of the glomeruli in the cortex (V V(Glom/Cortex) ) of distinct nephron segments within the cortex and OSM, including proximal tubules (V V(PT/Cortex+OSM) ) and other tubules (V V(OT/Cortex+OSM) ), and of the proximal tubular epithelium (PTE) within the cortex and OSM (V V(PTE/Cor-tex+OSM) ) were determined by point counting (Weibel 1979) (370 AE 89 points per case) in 11 AE 2 systematically randomly selected areas (Nyengaard 1999) of PASM-PASstained GMA/MMA sections, as principally described earlier (Hoeflich et al. 2002). The absolute volumes of the respective nephron segments (V (PT, Cortex+OSM) , V (PTE, Cortex+OSM) , V (Glom, Cortex) ) were calculated from their corresponding volume fraction and the absolute volume of the reference space (cortex, OSM).
Mean glomerular volume and total number of glomeruli
The mean glomerular volume ( v (Glom) ) and the total number of nephrons (glomeruli) in both kidneys (N (Glom, Cortex) ) were estimated, using the physical disector method (Sterio 1984;Howard and Reed 2004), in combination with systematic point counting, as principally described in previous publications of our group (Hoeflich et al. 2002;Herbach et al. 2009). Per case, 6 AE 2 PAS-stained serial GMA/MMA sections pairs were systematically randomly selected (nominal section thickness = 1 lm, disector height = 20 lm). Using an automated stereology system (VIS-Visiopharm Integrator System TM Version 3.4.1.0 with newCAST TM software, Visiopharm A/S, Denmark), each 35 AE 10 corresponding locations of the renal cortex were then systematically randomly sampled at 40 9 final magnification in both sections, automatically congruently aligned, and digitally superimposed with unbiased counting frames (344,017 lm² area) and 9 9 9 point test grids. The cross-sectional area of cortical kidney tissue present within each sampled counting frame (A (Cortex) ) was determined by point counting. In the subsequent disector analyses, 62 AE 12 Q À (glomeruli) were counted per case. The numerical volume density of glomeruli in the renal cortex was calculated as: embedded murine kidney tissue (0.91) (Herbach et al. 2011). N (Glom, Cortex) was then calculated as the product of N V(Glom/Cortex) and V (Cortex, Kid) . Subsequently, the mean glomerular volume was estimated as v To validate the precise disector heights used for the calculation of the numerical volume densities of glomeruli in the kidney, PTE cells in the renal cortex and glomerular cells in the glomeruli, the thicknesses of GMA/MMA and Epon sections were controlled, using an orthogonal resectioning technique, as described previously (Hoeflich et al. 2002). From the GMA/MMA and Epon section series, a section that was not sampled for the disector analyses was randomly selected and re-embedded in Epon resin, perpendicularly to its original section plane. The thickness of the re-embedded sections was then measured in image prints of subsequently prepared AzurII/Safraninstained semithin Epon sections taken at 1000 9 magnification, using a semiautomatic image analysis system (Videoplan â , Zeiss-Kontron, Germany). Per case, four measurements were made. Since the measured mean thickness of re-embedded Epon sections was 0.501 AE 0.005 lm and 1.037 AE 0.035 lm for GMA/ MMA sections, a section thickness 0.5 lm, respectively of 1.0 lm was consistently used for calculation of the disector volumes.
Number and volumes of glomerular cells and of PTE cells
The mean volume and total number of PTE cells, the mean numbers of distinct glomerular cell types per glomerulus (C: All glomerular cells; Pod: Podocytes, M-E: Mesangial and endothelial cells), and the mean podocyte volume were unbiasedly determined, applying the physical disector method (Sterio 1984;Howard and Reed 2004), principally as described above. For this, the numerical volume density of PTE cells in the cortex and of glomerular cells in the glomerulus was unbiasedly estimated in consecutive serial semithin sections of Epon-embedded cortical kidney tissue samples; per section series, 2-3 pairs of sections with 1.5-lm distance between these sections, that is, the nth and the n + 3rd section from a series cut with a nominal section thickness of 0.5 lm, were systematically randomly sampled (Howard and Reed 2004). Per case, the corresponding profiles of 9 AE 1 systematically randomly sampled glomeruli were photographed at 400 9 magnification in both sections of a section pair, using a Leica DFC 320 camera (Leica, Germany) connected to a microscope (Orthoplan, Leitz, Germany). Images including a size ruler were printed and overlaid with a plastic transparency with 576 equally spaced test points. The areas of the glomerular cross-sections (A (Glom) ) were planimetri-cally measured, using a Videoplan TM image analysis system (Zeiss-Kontron, Germany). The volume fraction of podocytes in the glomeruli V V(Pod/Glom) was estimated from the fraction of points hitting podocyte section profiles, and points hitting the corresponding reference space, that is, the glomerular cross-section profile (Howard and Reed 2004). On the average, 559 AE 273 points were counted per case. All glomerular cell nuclei profiles (C, Pod, M-E) sampled within a glomerular cross-section in the first (=reference) section, which were not present in the corresponding glomerular section profile in the second (=look-up) section, were counted (Q À ). The operation was then repeated by interchanging the roles of the lookup section and the reference section, increasing the efficiency of cell nuclei counting by factor two. On the average, 228 AE 48 glomerular cell (C) nuclei (Q À ) were counted per case (Pod: 69 AE 17; M-E: 160 AE 41). The numerical volume density of glomerular cells in the glomerulus (N V(C/Glom) , N V(Pod/Glom) , N V(M-E/Glom) ) was then calculated from the number of Q À counted per cell type and the respective disector volume, defined by the cumulative areas of analyzed glomerular section profiles and the distance (disector height = 1.5 lm) between the examined section pairs: 5 lm) and f s = linear tissue shrinkage factor for Epon-embedded murine kidney tissue (0.95) (Herbach et al. 2009). The mean number of cells per glomerulus (N (C,Glom) ) and of podocytes per glomerulus (N (Pod,Glom) ) was calculated by multiplying the numerical volume density of the respective glomerular cells in the glomerulus with the mean glomerular volume (refer to section 2.6.2.2). The mean podocyte volume ( v (Pod) ) was calculated dividing V V(Pod/Glom) by N V(Pod/Glom) .
The mean volume ( v (PTE cell, Cortex+OSM ) and the total number of PTE cells (N (PTE cells, Cortex+OSM) ) were assessed in an analogous manner. Digital microscopic images (taken at 400 9 magnification) of systematically randomly selected fields (11 AE 3 per case) within the sampled disectors were digitally superimposed with an unbiased counting frame (Howard and Reed 2004) of known area (12,543 lm 2 ), and a grid of equally spaced test points (68 points per counting frame). The volume density of the PTE in the renal cortex and the OSM (V V(PTE/Cortex+OSM) ) was estimated by point counting, as the fraction of points hitting PTE cell section profiles, and points hitting the renal cortical tissue section area. On the average, 727 AE 331 points were counted per case. The numerical density of PTE cells in the renal cortex (N V(PTE cells/Cortex+OSM) ) was calculated from averagely 66 AE 17 PTE cells (Q À ) counted within the unbiased counting frames of all investigated disectors per case, and the corresponding cumulative disector volumes per case (product of the disector height (1.5 lm) and the cumulative area of the cortical kidney tissue within the counting frames in all examined disectors): N V(PTE cells/ Cortex+OSM) = (∑Q À (PTE cells) /h 9 ∑A (Cortex+OSM) ) x f s ³ with h = disector height (1.5 lm) and f s = linear tissue shrinkage factor for Epon-embedded murine kidney tissue (0.95) (Herbach et al. 2009). N (PTE cells, Cortex+OSM) was calculated as the product of N V(PTE cells/Cortex+OSM) and V (Cortex+OSM, Kid) . The mean volume of PTE cells ( v (PTE cell, Cortex+OSM) ) was obtained by dividing V V(PTE cells/Cor-tex+OSM) by N V(PTE cells/Cortex+OSM) .
Mesangial and glomerular capillary volumes
The volume densities of capillaries, and of the mesangium within the glomeruli (V V(Cap/Glom) and V V(Mes/Glom) ) were determined by point counting (fraction of points hitting capillary section profiles, or the mesangial area in PASstained sections, respectively, and points hitting the glomerular section profile) in 36 AE 5 systematically randomly sampled glomerular profiles in PAS-stained GMA/ MMA sections at 200 9 magnification. Per case, 420 AE 58 points were counted. The mean mesangial and capillary volumes per glomerulus ( v (Mes, Glom) and v (Cap, Glom) ) were calculated as the product of the respective numerical volume density in the glomerulus and the mean glomerular volume ( v (Glom) ). The total volume of glomerular capillaries in the kidneys was calculated as V (Glom-cap, Kid) = V V(Cap/Glom) 9 V (Glom, Cortex) .
Mean length of the capillaries in a glomerulus
The mean length of the capillaries in a glomerulus (L (Cap, Glom) ) was determined in the same glomerular profiles sampled in the serial semithin Epon sections used for estimation of glomerular cell numbers. The number of glomerular capillary section profiles (ΣQ (Cap) = 671 AE 155 per case) present within the examined glomerular section profiles was counted. The cumulative glomerular tuft profile area of all examined glomerular section profiles (ΣA (Glom) ) was measured planimetrically, as described above. The length density of capillaries per glomerulus was calculated as L V(Cap/Glom) = 2 9 Q A(Cap/ Glom) 9 f s 2 , with Q A(Cap/Glom) = ΣQ (Cap) /ΣA (Glom) and f s = 0.95. L (Cap, Glom) was then obtained as: L (Cap, Glom) = Lv(Cap/Glom) 9 v(Glom) (Nyengaard 1999;Howard and Reed 2004). The total length of glomerular capillaries in the kidneys (L (Glom-cap, Kid) ) was calculated as: L (Glom-cap, Kid) = L (Cap, Glom) 9 N (Glom, Kid) .
Glomerular basement membrane (GBM) thickness and filtration slit frequency (FSF)
The thickness of the GBM was determined by the orthogonal intercept method as described earlier (Ramage et al. 2002;El-Aouni et al. 2006;Herbach et al. 2009). Per case, ultrathin sections of six glomeruli, that were sampled from semithin sections, were prepared for TEM, and peripheral glomerular capillary loops were photographed (7 AE 1 pictures per case) in a predetermined manner (by half turns of the stage handle). Photographs were developed to a final print magnification of 54,324 9 , and covered by a transparent 2.5-cm 2 grid. Where gridlines transected the GBM, the shortest distance between the endothelial cell membrane and the outer lining of the lamina rara externa underneath the cell membrane of the epithelial foot processes was measured, using a logarithmic ruler template with dimensions and midpoints, as described in detail earlier (Ramage et al. 2002). The true harmonic mean thickness (T h(GBM) ) of the GBM was estimated as: T h(GBM) = (8/3p) 9 (10 6 /M) 9 l h(GBM) , with l h (GBM) (apparent harmonic mean GBM thickness) = ∑ Number of observations/∑ (Midpoints 9 number of observations), and M = final print magnification. On the average, 46 AE 7 (range: 31-66) intercepts per animal were measured. The FSF was determined in the same electron micrographs, by counting the number of epithelial filtra- Numbers of examined animals (75 days old) are given in brackets. Data are means AE SD. One-way ANOVA with LSD post hoc test. * Statistically significant differences (P ≤ 0.05) between male and female mice of the identical genotype. † Statistically significant differences (P ≤ 0.05) between sex-matched I À/À versus I À/À /G, I +/À versus I +/À /G, and I +/+ versus G mice.
Statistical analyses
All data are presented as means AE SD. Data were analyzed by one-way ANOVA with LSD post hoc tests (IBM SPSS Statistics, Version 18), taking the effects of sex and genotype into account. P ≤ 0.05 were considered significant. To depict effects of GH-transgene expression on I À/ À , I +/À , and I +/+ mice, sex-matched GH-transgenic versus non-GH-transgenic mice with corresponding Igf1 genotypes were compared: I À/À versus I À/À /G, I +/À versus I +/ À /G, and I +/+ versus G, if not stated otherwise. Genderspecific differences were analyzed by comparison of male versus female mice of identical genotypes.
Results
Body weight, kidney weight, and kidney volume IGF1-deficient mice (I À/À /G and I À/À mice) displayed approximately 60% lower body and kidney weights and kidney volumes, as compared to mice with intact or heterozygously disrupted igf1 alleles (Tables 1 and 2). GH-transgenic mice consistently displayed significantly higher body weights than non-GH-transgenic mice with identical Igf1 status (G vs. I +/+ ; I +/À /G vs. I +/À ; and I À/ À /G vs. I À/À ; Table 1). Similarly, I À/À /G, I +/À /G, and G mice exhibited higher kidney weights and kidney volumes than I À/À , I +/À , and I +/+ mice, respectively. However, the differences of kidney weights and volumes did not reach statistical significance in male I +/À /G versus I +/À , and in female I À/À /G versus I À/À mice (Tables 1 and 2). Except for male I +/À /G versus I +/À mice, the relative kidney weights (% of body weight) did not differ significantly between GH-transgenic and non-GHtransgenic mice of corresponding igf1 allele status (Table 1).
Histological and ultrastructural findings I À/À mice displayed remarkably smaller glomerular profiles than mice of all other genotypes. In contrast, the glomerular profiles of GH-transgenic mice (I À/À /G, I +/À /G and G mice) consistently appeared greatly enlarged, as compared to non-GH-transgenic mice with the same Igf1 status (I À/À , I +/À , and I +/+ mice) (Fig. 1). Histopathological examination revealed glomerulosclerotic lesions present in all GHtransgenic genotypes (I À/À /G, I +/À /G, and G mice), comprising glomerular mesangial expansion, matrix accumulation and hypercellularity, multifocal capillary hyalinosis, capillary collapse, formation of synechiae between the glomerular tuft and the capsule of Bowman, crescent formation (Figs. 1 and 2), as well as tubulo-interstitial lesions, such as mild inflammatory cell infiltration, interstitial fibrosis, and multifocal tubular atrophy, along with signs of Numbers of examined animals are given in brackets. Data are means AE SD. One-way ANOVA with LSD post hoc test. * Statistically significant differences (P ≤ 0.05) between male and female mice of the identical genotype. † Statistically significant differences (P ≤ 0.05) between sex-matched I À/À versus I À/À /G, I +/À versus I +/À /G, and I +/+ versus G mice. proteinuria, that is, accumulation of protein droplets in proximal tubular epithelial cells, and tubular protein casts. These alterations were most marked in G mice, followed by I +/À /G, and I À/À /G mice. In contrast, mice of non-GHtransgenic genotypes (I À/À , I +/À , and I +/+ ) did not show glomerulosclerotic alterations. Ultrastructurally, the glomerular podocytes of mice of GH-transgenic genotypes (I À/À /G, I +/À /G, and G) appeared enlarged, as compared to non-GH-transgenic mice and displayed podocyte foot process effacement (Fig. 3).
Urine analyses
Mice of all GH-transgenic genotypes (I À/À /G, I +/À /G, and G) demonstrated albuminuria from approximately the fifth week of age onward. The areas and staining intensities of albumin bands (~64 kDa) in SDS-PAGEbased urine protein analyses were markedly increased in G mice, as compared to I À/À /G mice. Typically for male GH-overexpressing mice (Norstedt and Palmiter 1984), male GH-transgenic mice of all examined Igf1 genotypes (I À/À /G, I +/À /G, and G) consistently displayed drastically reduced major urinary protein (MUP) bands (<20 kDa), as compared to male non-GH-transgenic mice (Fig. 4).
Volumes of distinct renal zones and nephron segments
In GH-transgenic mice versus non-GH-transgenic mice of the identical Igf1 status (G vs. I +/+ , I +/À /G vs. I +/À , and I À/À /G vs. I À/À ), the volume fractions of the cortex (V V(Cortex/Kid) ), of the outer stripe of the medulla (V V (OSM/Kid) ), and of the inner stripe of the medulla together with the inner zone of medulla (V V(ISM/IZM/ Kid) ) in the kidneys showed only minor, slightly varying (3-8%) differences (Table 2). The absolute volumes of the cortex (V (Cortex, Kid) ) and of the outer stripe of the medulla (V (OSM, Kid) ), however, were significantly increased in G versus I +/+ mice, and in female I +/À /G versus I +/À mice, but not in I À/À /G versus I À/À mice (Table 2). Similarly, the total volumes of proximal tubules, and of other tubules in the renal cortex and the OSM (V (PT, Cortex+OSM) and V (OT, Cortex+OSM) ) were significantly increased in G versus I +/+ , and female I +/À / G versus I +/À mice as well, but not in I À/À /G versus I À/À mice (Table 5). In contrast, the total volumes of the glomeruli in the kidneys (V (Glom, Cortex) ) were significantly increased in all GH-transgenic mice versus non-GH-transgenic mice of the identical Igf1 status (Table 3). Figure 2. Glomerular histology. I À/À /G, I +/À /G, and G mice display glomerulosclerosis with glomerular mesangial expansion and matrix accumulation, capillary hyalinosis, and crescent formation. Note the larger glomerular section profile areas in I À/À /G, I +/À /G, and G mice, as compared to I À/À , I +/À , and I +/+ mice (glomerular hypertrophy). Female mice at 11 weeks of age. Silver-PAS staining, GMA/MMA sections. Bars = 50 lm. Total number of glomeruli and mean glomerular volume GH-transgenic mice displayed significantly lower numerical volume densities of glomeruli in the renal cortex (N V (Glom/Cortex) ) than non-GH-transgenic mice of identical Igf1 status, except for I +/À /G versus I +/À male mice. The total numbers of glomeruli in the kidneys (N (Glom, Cortex) ), however, did not differ significantly between I À/À /G versus I À/À , I +/À /G versus I +/À , and G versus I +/+ mice (Table 3). However, corresponding to their significantly smaller kidney and cortex volumes (Table 2), the total numbers of glomeruli in IGF1-deficient (I À/À /G and I À/ À ) mice were approximately 25% lower than in mice with intact (I +/+ and G mice), or heterozygously disrupted igf1 alleles (I +/À and I +/À /G mice). The mean glomerular volume v (Glom) , as well as the mean glomerular volume related to body weight ( v (Glom) /BW) was significantly increased in mice of all GH-overexpressing genotypes, as compared to non-GH-transgenic mice of identical Igf1 status (Table 3, Fig. 5). On the average, I À/À / G, I +/À /G, and G mice displayed a 2.7-fold higher mean glomerular volume, as compared to I À/À , I +/À , and I +/+ mice. Except for female G versus I +/+ mice, also the mean glomerular volume related to kidney weight ( v (Glom) /KW) was significantly increased in GH-transgenic mice, as compared to non-GH-transgenic mice of identical Igf1 status (Table 3).
Mean mesangial and capillary volumes and mean capillary lengths per glomerulus and total glomerular capillary lengths and volumes in the kidneys
Mice of all GH-transgenic genotypes (I À/À /G, I +/À /G, and G) consistently displayed significantly higher volume densities of the mesangium in the glomeruli (V V(Mes/ Glom) ), as well as significantly higher mean mesangium volumes per glomerulus ( v (Mes, Glom) ), as compared to non-GH-transgenic mice of identical Igf1 status (I À/À , Figure 3. Transmission electron microscopy of peripheral glomerular capillary walls. Mice of GH-transgenic genotypes (I À/À /G, I +/À /G, and G) consistently exhibit podocyte foot process effacement and reduced filtration slit frequency, as compared to mice of non-GH-transgenic genotypes (I À/À , I +/À , and I +/+ ). Male mice at 11 weeks of age. Bars = 500 nm. Figure 4. SDS-PAGE urine protein analysis. Representative gel of 11-week-old male mice. Spot urine samples were diluted to identical creatinine concentrations (3 mg/dL) and run on a 12% SDS-PAGE. Coomassie staining. The I À/À /G and G mice display albuminuria (bands at approx. 63 kDa, arrow). Also note the drastically reduced intensity of the major urinary protein (MUP) bands at approximately 18 kDa in urine samples of male I À/À /G and G mice, as compared to male I À/À and I +/+ mice. I +/À and I +/+ , respectively). The capillary volume per glomerulus (V (Cap, Glom) ), the total volume of the glomerular capillaries in the kidneys (V (Glom-cap, Kid) ), the length density of capillaries in the glomeruli (L V(Cap/ Glom) ), the mean length of the capillaries per glomerulus (L (Cap, Glom) ), and the total length of the glomerular capillaries in the kidneys (L (Glom-cap, Kid) ) were also significantly increased in I À/À /G versus I À/À , I +/À /G versus I +/À , and G versus I +/+ mice (Table 4).
Number of cells per glomerulus and mean podocyte volumes
Due to a significant increase of the numbers of endothelial and mesangial cells per glomerulus, the mean total numbers of cells per glomerulus in I À/À /G, I +/À /G, and G mice were significantly increased, as compared to I À/À , I +/À , and I +/+ mice, respectively. In contrast, the mean numbers of podocytes per glomerulus were not significantly different between I À/À /G and I À/À , I +/À /G and I +/À , or G and I +/+ mice (Fig. 6). However, the mean podocyte volume was significantly increased in GH-transgenic mice, as compared to non-GH-transgenic mice of the identical Igf1 status, except for male I +/À /G versus I +/À mice (Fig. 7A).
Volume densities and total volumes of proximal tubular epithelial cells in the cortex and the OSM, and total numbers and mean volumes of PTE cells The volume densities of proximal tubular epithelial (PTE) cells in the renal cortex and the OSM (V V(PTE/Cortex+OSM) ) did not differ significantly between I À/À /G and I À/À , I +/À / Numbers of examined animals are given in brackets. Data are means AE SD. One-way ANOVA with LSD post hoc test. * Statistically significant differences (P ≤ 0.05) between male and female mice of the identical genotype. † Statistically significant differences (P ≤ 0.05) between sex-matched I À/À versus I À/À /G, I +/À versus I +/À /G, and I +/+ versus G mice. Figure 5. Mean glomerular volume at 11 weeks of age. I À/À /G, I +/À / G, and G mice display significantly higher mean glomerular volumes, as compared to I À/À , I +/À , and I +/+ mice. Data are means AE SD. Oneway ANOVA with LSD post hoc test. Statistically significant differences (P ≤ 0.05) are indicated by different superscripts. The numbers of examined animals are given in brackets. G and I +/À , or G and I +/+ mice (Table 5). The total volume of PTE cells in the cortex and the OSM (V (PTE, Cor-tex+OSM) ) were significantly increased in G versus I +/+ mice. In I À/À /G versus I À/À , and in I +/À /G versus I +/À mice the V (PTE, Cortex+OSM) was increased, but the differences did not reach statistical relevance (Table 5). The total numbers of PTE cells in the kidneys (N (PTE cells, Cor-tex+OSM) ) were significantly increased in G versus I +/+ and in I +/À /G versus I +/À mice, but not in I À/À /G versus I À/À mice, whereas the mean volume of PTE cells ( v (PTE cells, Cortex+OSM)) was virtually unaltered between GH-transgenic and non-GH-transgenic mice of corresponding Igf1 status (Fig. 8).
Discussion
Although the formal pathogenesis of progressive glomerulosclerosis in GH-transgenic mice has been well characterized, the exact mechanisms inducing stimulation of renal and glomerular growth, glomerular and podocyte hypertrophy, mesangial and endothelial hyperplasia, and the well-known subsequent renal lesion patterns seen in GH-transgenic mice are yet not fully understood. Several of the various biological actions of GH, such as stimulation of somatic growth, are indirectly mediated by GHinduced IGF1 (Le Roith et al. 2001). Although GH stimulates the release of predominantly liver-derived IGF1 into the circulation, elevated systemic IGF1 levels are not essential for mediation of GH-induced IGF1-dependent growth-promoting actions (Yakar et al. 1999;Le Roith et al. 2001), since several, if not the majority, of the growth-stimulating effects of IGF1 are actually mediated by locally produced IGF1, acting in an autocrine or paracrine fashion (Le Roith et al. 2001;Stratikopoulos et al. 2008). The exact roles of excess GH and of IGF1 in stimulation of renal and glomerular growth in GH-transgenic mice, however, are still unclear. Since different glomerular cell types simultaneously express both GH and IGF1 receptors (Doi et al. 2000;Tack et al. 2002;Karl et al. 2005;Reddy et al. 2007;Vasylyeva and Ferry 2007;Bridgewater et al. 2008), a clear differentiation of excess GH and IGF1 actions promoting glomerular hypertrophy and glomerulosclerosis in vivo is complicated. Almost 25 years ago, it was demonstrated that both GH-transgenic and Igf1-transgenic mice do develop glomerular hypertrophy. However, the magnitude of glomerular enlargement displayed by Igf1-transgenic mice is much lower than the glomerular hypertrophy exhibited by GHtransgenic mice, and in contrast to GH-transgenic mice, Igf1-transgenic mice do not develop glomerulosclerosis (Doi et al. 1988(Doi et al. , 1990. These observations indicate that both excess GH-dependent and GH-independent elevation of systemic IGF1 levels stimulate glomerular growth. However, stimulation of glomerular growth by systemically elevated IGF1 levels in Igf1-transgenic mice without concurrently elevated systemic GH levels is obviously not sufficient to trigger the development of glomerulosclerosis (Doi et al. 1990). These findings of course do not elucidate the potential role of excess GH-stimulated, locally produced glomerular IGF1 for the development of glomerular hypertrophy and glomerulosclerosis in GHtransgenic mice with systemically elevated GH levels.
To dissociate the relative contributions of IGF1mediated and IGF1-independent effects of GH excess on Figure 6. Number of cells per glomerulus of 11-week-old mice. The total glomerular cell numbers, as well as the numbers of mesangial and endothelial glomerular cells per glomerulus are significantly higher in I À/À versus I À/À /G, I +/À versus I +/À /G, and I +/+ versus G mice. In contrast, the number of podocytes per glomerulus does not significantly differ between GH-transgenic mice and non-GH-transgenic mice with identical Igf1 status. Data are means AE SD. One-way ANOVA with LSD post hoc test. Statistically significant differences (P ≤ 0.05) are indicated by different superscripts. The numbers of examined animals are given in brackets. Figure 7. Mean podocyte volume (A) and filtration slit frequency (B) at 11 weeks of age. I À/À /G, I +/À /G, and G mice display significantly higher mean podocyte volumes, as compared to I À/À , I +/À , and I +/+ mice. The number of podocyte-foot process filtration slits per mm of the glomerular basement membrane (GBM) is significantly reduced in I À/À versus I À/À /G, I +/À versus I +/À /G, and I +/+ versus G mice. Data are means AE SD. One-way ANOVA with LSD post hoc test. Statistically significant differences (P ≤ 0.05) are indicated by different superscripts. The numbers of examined animals are given in brackets.
stimulation of renal growth and subsequent development of progressive glomerulosclerosis in vivo, the present study applied model-independent stereological methods on kidney samples of a unique collective of GH-transgenic IGF1-deficient mice. Since permanently excessively elevated levels of GH in G mice induce a simultaneous increase in endogenous IGF1 levels, none of the phenotypical effects observed in G mice can clearly be attributed to GH or IGF1 alone by comparing G mice to I +/+ mice. For the same reason, a comparison of G mice and I À/À /G mice is also not suitable to identify IGF1-independent effects of excess GH on renal growth and pathology. Additionally, IGF1 is a crucially important factor for embryonal and fetal renal growth and nephronogenesis, and therefore the initial "kidney status" (characterized by, e.g., kidney volume, nephron number, initial mean glomerular or podocyte volume) is not comparable between Igf1-knockout mice and mice with intact igf1 alleles. Therefore, a meaningful dissociation of IGF1-independent effects of excess GH from GH-induced IGF1mediated effects can only be performed by comparing I À/ À /G mice and I À/À mice. Correspondingly, groups of Table 5. Quantitative stereology of renal tubules.
Parameter
n (m/f) mice with identical IGF1 genotypes were compared throughout the present study (I À/À vs. I À/À /G, I +/À vs. I +/ À /G, and I +/+ vs. G). Implementation of studies on Igf1-knockout mice is generally hindered by the low perinatal survival rates of IGF1-deficient mice, restricting the numbers of animals available for investigation (Liu et al. 1993;Liu and LeRoith 1999;Lupu et al. 2001;Blutke et al. 2014). Although the numbers of examined I À/À and I À/À /G mice were limited to 7-8 mice per genotype, the results of exhaustive quantitative stereological analyses were capable to clearly identify common and dissociated growth-promoting effects of GH and IGF1 on the kidney, distinct kidney structures, and cell types.
Confirming the essential role of IGF1-mediated GH-stimulated postnatal somatic growth (Powell-Braxton et al. 1993;Liu and LeRoith 1999;Le Roith et al. 2001;Blutke et al. 2014), I À/À and I À/À /G mice reached only~35% and~57% of the body weight of sex-matched WT control animals at 75 days of age, although the body weights of I À/À /G mice were significantly higher than those of I À/À mice (Blutke et al. 2014).
In the kidneys, a disproportional stimulation of renal growth by GH-overexpression, as indicated by increased relative kidney weights in elder (>120 days) mice of various GH-transgenic mouse lines (Brem et al. 1989;Wanke et al. 1991), was yet not present in 75 days old GH-transgenic mice versus non-GH-transgenic mice of the same igf1 allele status. Interestingly, the effect of GH overexpression on the absolute volumes of the kidneys and of distinct kidney compartments, including the cortex, the outer stripe of the medulla (OSM), and the proximal tubules, differed in GH-transgenic Igf1-knockout mice, as compared to GH-transgenic mice with intact igf1 alleles. On the average, the total kidney weights and volumes of G mice were significantly increased by 86%, as compared to sex-matched I +/+ mice (93% in male mice and 78% in female mice). In contrast, the nonsignificant increase of kidney weights and volumes in I À/À /G versus I À/À mice accounted merely for 37% (20% in male mice and 54% in female mice), supporting the finding that stimulation of overall renal growth by GH excess is largely but not exclusively mediated by IGF1 (Doi et al. 1990). The marked increase of the kidney volume in G versus I +/+ mice was caused by a significant, on the average 95% increase in the volume of the renal cortex (V (Cortex, Kid) ), and by a significant, on the average 78% increase in the OSM (V (OSM, kid) ) in male and female G versus I +/+ mice. The increased volumes of the renal cortex and the OSM mainly resulted from a significant, on the average 97% enlargement of the volume of proximal tubules (V (PT, Cortex & OSM) ). The latter was found due to a significant increase in the total volume of the proximal tubular epithelia (PTE) in the renal cortex and the OSM. Quantitative stereological analyses revealed a hyperplastic growth of PTE cells to account for the increased total PTE volume in G versus I +/+ mice. Although the mean cellular volume of PTE cells was virtually unaltered in all GHoverexpressing versus non-GH-transgenic mice of identical Igf1 status, the total number of PTE cells in G versus I +/+ mice was significantly increased by 76%. In contrast, the volumes of the renal cortex, the OSM, the proximal tubules, the PTE, and the total PTE cell numbers in I À/À / G versus I À/À mice and in I +/À /G versus I +/À mice were throughout increased by just~30-40%. However, due to the limited number of Igf1-knockout mice of 75 days of age available for analysis, these differences did not (yet) reach the level of statistical significance. The observed differences in stimulation of tubular growth between G, I +/ À /G, and I À/À /G mice show that IGF1 is a major mediator of PTE hyperplasia induced by GH excess, consistent with previously reported proliferative/mitogenic effects of IGF1 on proximal tubular epithelial cells in vitro (Blazer-Yost et al. 1992).
In addition to their dwarf phenotype, and their proportionally reduced kidney sizes, the complete deficiency of IGF1 also essentially affected renal development in I À/À and I À/À /G mice. The numbers of nephrons (N (Glom,Kid) ) in the small kidneys of I À/À and I À/À /G mice were considerably lower than in mice with intact or heterozygously deleted igf1 alleles, while Igf1-haploinsufficiency in I +/À and I +/À /G mice did not detectably affect nephron numbers, as compared to G or I +/+ mice. Furthermore, GH overexpression did also not significantly influence the number of nephrons in I À/À /G versus I À/À , I +/À /G versus I +/À , and G versus I +/+ mice, indicating that GH overexpression does not affect nephronogenesis in GH transgenic mice.
In GH transgenic mice, progressive glomerular and podocyte hypertrophy, subsequent podocyte damage, and albuminuria are pathogenetic key lesions in the development of glomerulosclerosis (Wanke et al. 2001;Pavenstadt et al. 2003;Wiggins 2007). The mean glomerular volume and the mean podocyte volume, as the relevant parameters for assessment of glomerular and podocyte hypertrophy, were significantly increased in all GH-transgenic mice versus their non-GH-transgenic controls of identical Igf1 status. In parallel, mice of all investigated GH-transgenic genotypes displayed glomerulosclerotic kidney lesions, significantly reduced podocyte filtration slit frequencies, and albuminuria. The glomerulosclerosis indices of the mice examined in the present study have already been reported in a previous publication (Blutke et al. 2014). Mice of all GH-transgenic genotypes (I À/À /G, I +/À /G, and G) consistently display significantly higher glomerulosclerosis indices than non-GH-transgenic mice (I À/À , I +/À , and I +/+ mice). Since glomerulosclerotic lesions were exclusively observed in GH-transgenic mice, these lesions must be regarded as a consequence of permanently and excessively elevated levels of GH (and GHdependent IGF1, if present). In contrast, the elevated endogenous GH secretion with a pulsatile secretion pattern that occurs as a feedback reaction to IGF1 deficiency (Blutke et al. 2014) is apparently not sufficient to induce development of comparable glomerular alterations in (non-GH-transgenic) Igf1-knockout mice. Since I À/À /G mice and I +/À /G mice display significantly lower glomerulosclerosis indices than G mice (Blutke et al. 2014), IGF1 deficiency apparently slightly ameliorates the severity, respectively, the velocity of progression of excess GHinduced glomerulosclerosis in GH-transgenic mice. On one hand, these findings indicate that excess GH-induced IGF1 contributes to the development of glomerulosclerosis in GH-transgenic mice. On the other hand, although the glomerulosclerosis indices in GH-transgenic mice with intact igf1 alleles are significantly higher than in IGF1-deficient mice, the present data also prove that GH excess can cause glomerular and podocyte hypertrophy sufficient to induce glomerulosclerosis, independently of IGF1.
Indicating an excess GH-stimulated glomerular growth even in the absence of IGF1, the relative mean glomerular volumes of GH-overexpressing mice were on the average three times as large as in sex-matched non-GH-transgenic mice of identical Igf1 status, with male G mice exhibiting the greatest average glomerular volume increase. Distinct glomerular growth processes contributing to the enlargement of the glomerular tuft in GH-transgenic mice with or without intact igf1 alleles were extensively characterized by quantitative stereological analyses. Thus, glomerular hypertrophy resulted from an increased mesangial volume per glomerulus, hyperplasia of mesangial and endothelial glomerular cells, a higher glomerular capillary volume due to an increased capillary length per glomerulus, as well as from hypertrophy of podocytes. G mice consistently displayed the highest absolute magnitudes in the different glomerular compartment-and cell-type growth parameters, whereas the relative increases in these parameters were in the same ranges in mice of all GHoverexpressing genotypes, as compared to non-GH-transgenic mice of the same Igf1 status. Therefore, the relative increase in the mean glomerular and podocyte volume in GH-transgenic mice, as compared to non-GH-transgenic mice of the same Igf1 status appears to be a critical step in the development of glomerulosclerosis in GH-transgenic murine nephropathy models, rather than excess of a distinct absolute glomerular or podocyte size.
Podocyte hypertrophy, as an initial step in the development of podocyte damage subsequently leading to glomerulosclerotic lesions, is regarded as a common reaction of postmitotic podocytes to progressive glomerular enlargement in diverse experimental nephropathy models (Kretzler et al. 1994;Wolf and Wanke 1997;Wanke et al. 2001;Wiggins 2007). Additionally, it has been demonstrated that podocyte hypertrophy can also precede glomerular hypertrophy in murine models of diabetic nephropathy (Herbach et al. 2009) and that the glomerular podocyte is a target of direct GH action (Reddy et al. 2007;Kumar et al. 2010Kumar et al. , 2011. Given a probable contribution of direct GH-mediated growth stimulation to podocyte hypertrophy in the pathogenesis of glomerulosclerosis in GH-transgenic mice, the significantly increased mean podocyte volumes in I À/À /G versus I À/À mice indicate that the growth stimulatory effect of excess GH on podocytes can as well occur independently of IGF1. However, the exact molecular pathways of potential direct GH actions governing and modulating growth responses of different renal/glomerular cell types in the pathogenesis of glomerular hypertrophy and glomerulosclerosis will require further investigations. In parallel, the finding that the highest increases of the mean glomerular volume and of the mean podocyte volume were observed in G versus I +/+ mice, confirms that excess GH-induced IGF1 also substantially contributes to the development of the pathogenetically relevant alterations leading to glomerulosclerosis in GH-transgenic mice with intact igf1 alleles. In summary, the data presented in this study demonstrate dissociated effects of GH and IGF1 in excess GHstimulated tubular and glomerular growth in vivo, and indicate that (1) IGF1 is not necessary for mediation of the effects of GH-overabundance causing progressive glomerulosclerosis in GH-transgenic mice; and (2) IGF1 is an important mediator of excess GH-induced proximal tubular hyperplasia. In light of the currently "reawaken" interest in the role of GH in the pathogenesis of human kidney diseases as diabetic nephropathy, and potentially deducible therapeutic strategies for prevention and treatment of these diseases (Kumar et al. 2011), the results of the present study strongly encourage further investigations in this field. | 2017-08-17T13:08:17.661Z | 2016-03-01T00:00:00.000 | {
"year": 2016,
"sha1": "f804b307aa2b71c3507ff3174f8b9d9fe289f478",
"oa_license": "CCBY",
"oa_url": "https://physoc.onlinelibrary.wiley.com/doi/pdfdirect/10.14814/phy2.12709",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "87d762bc8c4160b4153d76c26c4d6eb8ed40d87d",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Chemistry",
"Medicine"
]
} |
219457206 | pes2o/s2orc | v3-fos-license | Interfacing AI with Social Sciences: the Call for a New Research Focus in HCI
. We provide arguments for the necessity of broadening the engagement of HCI in translating knowledge created in social sciences to a major force that can drive AI and direct the ways in which it will impact various aspects of our world. We begin to sketch the outline of this engagement as a research agenda within HCI in reference to some of the definitional manifesto on the HCI’s foundational role [4] as an action science [3]. Also, a part of our own research that scrutinizes some of the major AI projects [1][2] informs the presented arguments.
Introduction
Similar to many other domains of research and design, the future of HCI is increasingly bound with the advances in Artificial Intelligence (AI).This is attributed to the opportunities it can create to enhance the processes and methods of research, but also to the challenging questions it raises such as "explainability", "agency", "trustworthiness", "ethics", and so forth.
However, the discussion that we would like to instigate in this paper takes a different standpoint and seeks to pronounce a new responsibility, a crucial and urgent one, that is embedded in the HCI's special scientific placement.Departing from a widely acknowledged observation that despite the substantial and multifaceted impact of AI on human life social sciences are far from being at a leading position, we argue that HCI can play a key role in rectifying this disconnect.The task, in a nutshell, is to bring the knowledge created in various areas of social science to the format and position that can effectively shape the development of non-human intelligent actors and inform the policies governing whether and how they should be adopted.The eventual objective is to ensure that AI as such will take us to a "better future" where human values and priorities are advanced, and prevent it from falling into the other paths that empirical findings and discourses already can forecast the ramifications.In the interest of clarity, we constrain the discussion of AI to what we refer to as "major AI projects", projects such as Smart City or Personalized Learning that are initially backed by certain technological possibilities and produced a wave of corporate investments and academic investigations.This is our primary interest in AI.Instead of discussing AI as a broad (and rather ambivalent) concept, the objective is to provide a platform in which major AI projects can be meticulously analyzed using the frameworks constructed in the relevant domains of social science.
We develop our argument is three steps: 1) we start by highlighting a set of cross-domain concerns stemming from the current mechanisms through which major AI projects have been conceptualized and advancing.Then we describe 2) why we think HCI holds the methodological capabilities to rectify those problems, and why it is only HCI that can do so.In the end, we 3) try to extend this discussion beyond abstract reflections and begin to draft a research agenda that can reify some of the described objectives.Furthermore, we support our arguments throughout this text with one specific example of Autonomous Vehicles within the framework of Smart Cities as one major AI project.
2 The problem of tech-oriented market-led AI Until social sciences produce a "sociological conception of AI" [6], the most conceivable forecast for the proliferation of AI relies on the existing unrivaled trend, in which it is the market-led tech sector that determines what AI is, how it should be invested in, and in what shape and function will reach the societies.In this model, our cities, our homes, our everyday social interactions, our education system and many fundamental aspects of our world will be subject of changes that are founded within the tech industry and, at its best, steered by the insights confined within the scope of computer and data sciences.
The problem of such a model is beyond the fact that data and computer sciences lack the necessary knowledge.One should, more importantly, see the principles embedded at the core of data science that are in contradiction with some of the human values.For example, the inherent desire for clustering or classification of data has shown that can engender division, discrimination, and segregation in social realms-the problem that has been already seen in the context of social media [12], but also the same has begun to manifest itself in the physical social environments [7] [11].
To our observation, across different domains, there is no shortage of critical scholarship that identifies the risks in the currently dominant formulation of AI projects.As an example, for the case of Smart City, beginning with technology and not with urban knowledge is rigorously criticized.Within the urbanism research community, Hollands describes his concerns for the obfuscation of the negative effects of IT on cities by it's business motivated promoters [9], Greenfield warns a return to the failed utopias of 20th century high modernism [15], Kitchen forecasts the rise of technocratic governance coming with the wave of smart city [10], Vanolo criticizes the current conceptions of smart city for the creation of discrimination against the "non-smart" citizens [13], Datta demonstrates, from the analysis of currently existing smart cities, the justification for regimes and processes of land dispossession [8], and Wiig criticizes the priorities given to the attraction of global business against local urgencies in dominant conception of smart cities [14].
The domain of urbanism and the case of Smart City is one example.To similar extents, one can observe critical discourses of the tech-oriented and market-led AI across many strands of social science.
The perennial mission of HCI
"HCI manages innovation to ensure that human values and human priorities are advanced, and not diminished through new technology.This is what created HCI; this is what led HCI onto and then off the desktop; it will continue to lead HCI to new regions of technology-mediated human possibility."[4].It has been widely recognised, as exemplified in the above statement, that one of the perennial roles of our community is to search for, develop and apply various means of assessing technological innovations and their societal impacts.What we would like to highlight in this contribution is the necessity to found a special focus within HCI that investigates a specific type of projects, namely the highly invested and deeply impactful projects that capitalise on the advance in data and computer science to create non-human intelligent actors and integrate them into social contexts that they both utilise and modify.Projects such as selfdriving vehicles, which are often rushed to performance test and production before understanding the changes they bring with themselves in terms of urban individual and collective experiences.
The distinguishing attributes of major AI projects-e.g.speed of proliferation and propagation-we believe, entail multidisciplinary analytical scholarship that is proactive in terms of guiding the trends and preventive of futures in which "human values and priorities" would be compromised".
In order to understand whether or not HCI is capable of carrying out the task of interfacing social sciences with AI projects, one should consider the record of HCI research and design in embarking upon multidisciplinary endeavors.HCI has a successful history of intertwining with reflections and concepts in the various domain of social science, adopting and re-appropriating research methods from those domains, and creating situations of mutual learning.It also has shown that can speak the technical language of computer and data science and be visionary in the realm of technologies.
At the end, we would like to go one step further and claim that the onus falls uniquely on us, the HCI research and design practitioners.This is not only an interdisciplinary involvement; HCI should rather take the lead and commit itself with the responsibility of proactively checking AI agenda with social sciences and ensuring that AI takes us to a better world.This is justified by the intermediary placement of HCI between computing power and human values.The evolution of HCI methods, concepts, and reflections originally to interface computing systems with humans, has situated our domain in a unique position to interface computer science with humanities.This a unique position for HCI, which implies a unique responsibility.The other fields with similar placements, such as Digital Human-ities (DL) and Science and Technology Studies (STS), define themselves within more confined scopes and consequently are limited in terms of methodological capacities needed for the described research focus.Loosely termed, DL applies digital tools to the study of humanity without concerning much with the impact of various technologies on humanity, and in STS the rise and adoption of technologies are examined as a social process, retaining the focus on the technology or science and rather than their effect on societies.
A new HCI research agenda
Without aspiring to a complete or a final picture, in this section, we would like to briefly mention some of the constituents of a research program that can be conducive to the mentioned objectives: -Rigorous analytical discussions should be developed that scrutinise major AI projects.The analytical frameworks are borrowed from the relevant domain of social science corresponding to the project to be discussed, which provide the researchers with a list of agreed-upon subject matters in the specific context of the study.For example, a critical discussion of self-driving cars would be structured based on topics of interest when accounting for car mobility in urban design.Topics such as spatial justice, public health and well-being, sustainability, congestion, urban forms, and so forth.The immediate questions to be answered are whether the current formulation of the AI project would have a positive or negative impact on each of those topics.Moreover, the list of topics can be complemented by questions on the new interactive experiences that the AI project itself would impose to its users.For example, questions surrounding interaction between pedestrians and the self-driving vehicles and how to build up the sense of trust towards them.-The outcome of such analysis may be in the form of proposals for modifications or amendments to the current formulation of the project or generate alternative narrations, new perspectives, and new schematisms for creativity and development.For example, in the context of smart mobility, one may see the value in redirecting the attention to how AI can enhance active mobility (walking and cycling).The same strand of multidisciplinary work in the first step continues here, this time in creating a new vision of what urban scientists see as enhancing human priorities and what HCI researchers may foresee as trustful intelligent mobile actors in the city and a realistic adoption scenario (that matches the constraints of the cities and avoids the mess and myth of technological pledges [5]).-To be able to become an effective force that guides the evolution of AI projects, the resulting discourses should extend their reach beyond the borders of academic environments and interact with people and public policymakers.Therefore, an essential part of the task is to translate the created visions and insights into the appropriate format that can offer the public opportunities for collecting alternative perceptions of, for example, what a smart city, smart mobility, and eventually smart mobile citizens can be.
These are some of the interrelated steps that together can support the engagement of different stockholder of AI projects to question technological innovations and to seriously contribute in shaping their future-ideas that we believe should be scaffolded within the scope of HCI research and design related to the underpinning characteristics of our field as briefly sketched previously.
Concluding remarks
The discussion presented in this contribution is grounded in the fact that AI as such can lead our world to various futures.It can lead our cities to be even more car-dependent (through the promotion of autonomous vehicles); it can instead enhance active mobility (walking and cycling), make possible sustainable use of urban spaces and create human-scale public spaces.City is one example; such alternative futures extend to many aspects of life.The discourse that we initiate in this paper foregrounds the responsibility of HCI in studying such futures and providing directions for AI adoption policies. | 2020-05-21T09:10:16.237Z | 2020-05-07T00:00:00.000 | {
"year": 2020,
"sha1": "71a404d47c9a1599d51927e2f4b8b022a623478b",
"oa_license": "CCBY",
"oa_url": "https://cardiffuniversitypress.org/site/chapters/10.18573/book3.z/download/4271/",
"oa_status": "HYBRID",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "5b892c0689bb0a64e2aed40b0b5867ca3ffd6b74",
"s2fieldsofstudy": [
"Art"
],
"extfieldsofstudy": [
"Engineering"
]
} |
14993720 | pes2o/s2orc | v3-fos-license | Biologically inspired protection of deep networks from adversarial attacks
Inspired by biophysical principles underlying nonlinear dendritic computation in neural circuits, we develop a scheme to train deep neural networks to make them robust to adversarial attacks. Our scheme generates highly nonlinear, saturated neural networks that achieve state of the art performance on gradient based adversarial examples on MNIST, despite never being exposed to adversarially chosen examples during training. Moreover, these networks exhibit unprecedented robustness to targeted, iterative schemes for generating adversarial examples, including second-order methods. We further identify principles governing how these networks achieve their robustness, drawing on methods from information geometry. We find these networks progressively create highly flat and compressed internal representations that are sensitive to very few input dimensions, while still solving the task. Moreover, they employ highly kurtotic weight distributions, also found in the brain, and we demonstrate how such kurtosis can protect even linear classifiers from adversarial attack.
Introduction
Deep Neural Networks (DNNs) have demonstrated success in many machine learning tasks, including image recognition (Krizhevsky et al., 2012), speech recognition , and even modelling mathematical learning (Piech et al., 2015), among many other domains. However, recent work has exposed a remarkable weakness in deep neural networks (Szegedy et al., 2014) (see (Warde-Farley & Goodfellow, 2016) for a survey), namely that very small perturbations to the input of a neural network can drastically change its output. In fact, in image classification tasks, it is possible to perturb the pixels in such a 1 Neurosciences PhD Program, Stanford University 2 Department of Applied Physics, Stanford University. Correspondence to: Aran Nayebi <anayebi@stanford.edu>, Surya Ganguli <sganguli@stanford.edu>.
way that the perturbed image is indistinguishable from its original counterpart to a human observer, but the network's class prediction is completely altered. These adversarial examples suggest that despite the above successes, machine learning models are not fundamentally understanding the tasks that they are trained to perform.
Furthermore, the imperceptibility of these adversarial perturbations to human observers suggests that these machine learning algorithms are performing computations that are vastly different from those performed by the human visual system. This discrepancy is of particular scientific concern as deep neural networks now form foundational models in neuroscience for the visual processing stream (Yamins et al., 2014;McIntosh et al., 2016;Kriegeskorte, 2015). So their susceptibility to adversarial perturbations that are imperceptible to us suggest our models are missing a fundamental ingredient that is implemented in the brain. However, the existence of adversarial examples is also of particular technological concern in machine learning, as these adversarial examples generalize across architectures and training data, and can therefore be used to attack machine learning systems deployed in society, without requiring knowledge of their internal structure (Papernot et al., 2016a;Warde-Farley & Goodfellow, 2016).
It is important to note that adversarial examples of this form are not limited to deep networks but are also an issue even in linear high dimensional classification and regression problems. A plausible explanation for the existence of these adversarial examples lies in the idea that any algorithm that linearly sums its high dimensional input vectors with many small weights can be susceptible to an attacker that adversarially perturbs each of the individual inputs by a small amount so as to move the entire sum in a direction that would make an incorrect classification likely. This idea lead to a fast method to find adversarial examples which could then be used to explicitly train neural networks to be robust to their own adversarial examples .
However, it is unclear that biological circuits explicitly find their own adversarial examples by optimizing over inputs and training against them. Therefore, we are interested in guarding against adversarial examples in a more biologically plausible manner, without explicitly training on ad-arXiv:1703.09202v1 [stat.ML] 27 Mar 2017 versarial examples themselves. Of particular interest is isolating and exploiting fundamental regimes of operation in the brain that prevent the imperceptible perturbations that fool deep networks, from fooling us. In this paper, we take inspiration from one fundamental aspect of single neuron biophysics that is not often included in artificial deep neural networks, namely the existence of nonlinear computations in intricate, branched dendritic structures (Stuart et al., 2016;Koch, 1999;London & Häusser, 2005). These nonlinear computations prevent biological neurons from performing weighted sums over many inputs, the key factor thought to lead to susceptibility to adversarial examples. Indeed, the biophysical mechanism for linear summation in neurons corresponds to the linear superposition of trans-membrane voltage signals as they passively propagate along dendrites. These voltage waves can linearly sum synaptic inputs. However, there is also a high density of active ionic conductances spread through the dendritic tree that can destroy this linear superposition property in purely passive dendrites, thereby limiting the number of synapses that can linearly sum to O(10)−O(100). These active conductances lead to high threshold, nonlinear switch like behavior for voltage signalling. As a result, many parts of the dendritic tree exist in voltage states that are either far below threshold, or far above, and therefore saturated. Thus biological circuits, due to the prevalence of active dendritic processing, may operate in a highly nonlinear switch-like regime in which it is very difficult for small input perturbations to propagate through the system to create large errors in output.
Rather than directly mimic this dendritic biophysics in artificial neural networks, here we take a more practical approach and take inspiration from this biophysics to train artificial networks into a highly nonlinear operating regime with many saturated neurons. We develop a simple training scheme to find this nonlinear regime, and we find, remarkably, that these networks achieve state of the art robustness to adversarial examples despite never having access to adversarial examples during training. Indeed we find 2-7% error rates on gradient-based adversarial examples generated on MNIST, with little to no degradation in the original test set performance. Furthermore, we go beyond performance to scientifically understand which aspects of learned circuit computation confer such adversarial robustness. We find that our saturated networks, compared to unsaturated networks, have highly kurtotic weight distributions, a property that is shared by synaptic strengths in the brain (Buzsáki & Mizuseki, 2014). Also, our networks progressively create across layers highly clustered internal representations of different image classes, with widely separated clusters for different classes. Furthermore we analyze the information geometry of our networks, finding that our satu-rated networks create highly flat input-output functions in which one can move large distances in pixel space without moving far in output probability space. Moreover, our saturated networks create highly compressed mappings that are typically sensitive to only one direction in input space. Both these properties make it difficult even for powerful adversaries capable of iterative computations to fool our networks, as we demonstrate. Finally, we show that the highly kurtotic weight distributions that are found both in our model and in biological circuits, can by themselves confer robustness to adversarial examples in purely linear classifiers.
Adversarial Example Generation
We consider a feedforward network F with D layers of weights W 1 , . . . , W D and D + 1 layers of neural activity vectors x 0 , . . . , x D , with N l neurons in each layer l, so that x l ∈ R N l and W l is an N l × N l−1 weight matrix. The feedforward dynamics elicited by an input x 0 are where b l is a vector of biases, h l is the pattern of inputs to neurons at layer l, and φ is a single neuron scalar nonlinearity that acts component-wise to transform inputs h l to activities x l . We take y to be the class indicator vector generated from x D . We also denote by x D = F (x 0 ) the network's composite transformation from input to output.
For such networks, the essential idea underlying adversarial examples is to start with a test example x 0 that is correctly classified by the network with class indicator vector y, and transform it through an additive perturbation ∆x 0 into a new input x 0 + ∆x 0 that is incorrectly classified by the network F as having a "goal" class label y G = y. Moreover, the perturbation ∆x 0 should be of bounded norm so as to be largely imperceptible to a human observer. This idea leads naturally to an optimization problem: However, as this is a complex optimization, many simpler methods have been proposed to efficiently generate adversarial examples (e.g. Miyato et al., 2016;Papernot et al., 2016a)). In particular, the fast gradient sign method of Goodfellow et al. (2015) is perhaps the most efficient method. Motivated by the notion that adversarial attacks can arise even in linear problems in high dimensional spaces, Goodfellow et al. (2015) linearized the input-output map F around the test example and searched for bounded l ∞ norm perturbations that maximize the network's cost function over the linearized network. More precisely, suppose the cost function of the network is C 0 = C(F (x 0 ), y), then its linearization is where J is the Jacobian of F . Then the bounded l ∞ norm optimization that maximizes cost has the exact solution (3) If a network can be susceptible to these gradient-based adversaries, then we can choose to be small enough for the given dataset so it is imperceptible to human observers yet large enough for the network to misclassify. For MNIST, Goodfellow et al. (2015) took = 0.25, since each pixel is in [0, 1]. We follow this prescription in our experiments.
With efficient methods of generating adversarial examples (3), Goodfellow et al. (2015) harnessed them to develop adversarial training, whereby the network is trained with the interpolated cost function: As a result, the network is trained at every iteration on adversarial examples generated from the current version of the model. On maxout networks trained on MNIST, Goodfellow et al. (2015) found that they achieved an error rate of 89.4% on adversarial examples, and with adversarial training (where α = 0.5), they were able to lower this to an error rate of 17.9%.
We now turn to ways to avoid training on adversarial examples, in order to have the networks be more intrinsically robust to their adversarial examples. Papernot et al. (2016b) suggested knowledge distillation, which involves changing a temperature parameter T on the final softmax output in order to ensure that the logits are more spread apart. However, the authors do not try their approach on adversarial examples generated by the fast gradient sign method, nor does this approach address the broader criticism of Goodfellow et al. (2015) that models susceptible to gradient-based adversaries operate heavily in the linear regime. We develop a method that strongly departs from the high dimensional linear regime in which adversarial examples abound. The basic idea is to force networks to operate in a nonlinear saturating regime.
Saturating Networks
A natural starting point to achieve adversarial robustness is to ensure that each element of the Jacobian of the model, J = ∂F/∂x 0 , is sufficiently small, so that the model is not sensitive to perturbations in its inputs. Jacobian regularization is therefore the most direct method of attaining this goal; however, for sufficiently large networks, it is computationally expensive to regularize the Jacobian as its dimensions can become cumbersome to store in memory.
An immediate alternative would be to use a contractive penalty as in Gu & Rigazio (2015), whereby the Frobenius norm of the layer-wise Jacobian is penalized: where each λ l ∈ R. For element-wise nonlinearities, Rifai et al. (2011) show that this penalty can be computed in O max l x l × x l−1 time, where |·| denotes the length (number of units).
While indirectly encouraging the activations to be pushed in the saturating regime of the nonlinearity, this contractive penalty can nonetheless be practically difficult to compute efficiently for networks with a large number of hidden units per layer, and also tends to limit the model's capacity to learn from data, degrading test set accuracy.
Saturating autoencoders were introduced by Goroshin & LeCun (2013) as a means of explicitly encouraging activations to be in the saturating regime of the nonlinearity, in order to limit the autoencoder's ability to reconstruct points that are not close by on the data manifold. Their penalty takes the following form for a given activation h = Wx+b and λ ∈ R, where the complementary function is defined as: and reflects the distance of any individual activation to the nearest saturation region. Not only is this penalty simple, but it can be cheaply computed in O(|h|) time.
Experiments and Results
Here we adapt the above regularization, originally designed for autoencoders, to protect against adversarial examples in supervised classification networks. We found that applying this regularization to every network layer, including the readout layer prior to the softmax output, worked best against adversarial examples generated by the fast gradient sign method. Thus, our penalty took the following form: Observe that for a ReLU function, the complementary function in (7) is itself, so φ c (z) = max{0, z}. While (7) can also be intricately extended to differentiable functions (as is done in (Goroshin & Le-Cun, 2013)), for a sigmoid function we can simply take We used TensorFlow for all of our models (Abadi et al., 2015), and we trained both 3 layer multilayer perceptrons (MLPs) with sigmoid and ReLU nonlinearities, as well as convolutional neural networks (CNNs) on 10-class MNIST. For comparison, we trained the adversarially trained networks as in (4), finding that α = 0.5 gave the best performance. Each network was optimized for performance separately, and we varied the number of hidden units for the MLPs to be between 200-2000 to choose the architecture that provided the best performance. Our CNN architecture is detailed in Table 2, and we used the stronger penalty f (z) = z only at the last layer of the CNN. We used Adam (Kingma & Ba, 2015) as our optimizer.
In order to effectively train with the saturating penalty in (8), we found that annealing λ during training was essential. Starting with λ min = 0, this was progressively increased to λ max = 1.74 in steps of size 0.001 for the sigmoidal MLP, λ max = 3.99 × 10 −8 in steps of size 10 −10 for the ReLU MLP, and λ max = 10 −5 in steps of size 10 −5 for the CNN. We ultimately found that the CNN was easier to find an annealing schedule for than the MLPs, further suggesting the viability of this approach in practice.
We list above our results in Table 1. As can be seen, for each model class, we are able to maintain (with little degradation) the original test set accuracy of the network's vanilla counterpart, while also outperforming the adversar-ially trained counterpart on the adversarial set generated from the test set. We now turn to analyzing the source of adversarial robustness in our networks.
Internal Representation Analysis
We now examine the internal representations learned by saturating networks (in particular the MLPs) and compare them to those learned by their vanilla counterparts, to gain insight into distinguishing features that make saturating networks intrinsically robust to adversarial examples. In Figure 1, we compare the weight distributions of the vanilla MLP to the saturating MLP. The saturating MLP weights take on values in a larger range, with a tail that tends to extreme values in the saturating regime of the nonlinearity. For the sigmoid, this leads to extreme weight values on both ends, while for the saturating ReLU MLP, this leads to extreme negative values. A particularly dramatic change in the weight distribution is a much larger positive excess kurtosis for saturating versus vanilla networks. In- deed, high kurtosis is a property shared by weight distributions in biological networks (Buzsáki & Mizuseki, 2014), raising the question of whether or not it plays a functional role in protection against adversarial examples. In §8, we will demonstrate that highly kurtotic weight distributions can act as a linear mechanism to protect against adversarial examples, in addition to the nonlinear mechanism of saturation.
Moreover, in Figure 2, we see that the pre-nonlinearity activations at each layer across all 10,000 test examples also tend to extreme values, as expected, validating that these models are indeed operating in the saturating regime of their respective nonlinearities.
Beyond examining the weights and the activations, we also examine the global structure of internal representations by constructing, for each network and layer, the representational dissimilarity matrix (RDM) of its activities (Kriegeskorte et al., 2008). For each of the 10 classes, we chose 100 test set examples at random, and computed the pairwise squared distance matrix, between all pairs a and b of the 1000 test examples. Here h l,a and h l,b are the hidden unit activations at layer l on inputs x 0,a and x 0,b , respectively. Figure 3, a distinguishing feature emerges between the RDMs of the vanilla network and the saturated network. At every layer, while within class dissimilarity, in the diagonal blocks, is close to zero for both networks, between class dissimilarities in the off-diagonal blocks are much larger in the saturated network than in the vanilla network. Moreover, this dissimilarity is progressively enhanced in saturating networks as one traverses deeper into the network towards the final output. Thus while both net- works form internal representations in which images from each class are mapped to tight clusters in internal representation space, these internal clusters are much further apart from each other for saturating networks. This increased cluster separation likely contributes to adversarial robustness because it necessitates larger norm input perturbations to move representations in deeper layers across decision boundaries, not only in the output layer, but also in intermediate layers.
The geometry of saturating networks
While the RDM analysis above showed increased cluster separation in internal representations, we would like to understand better the geometry of the network input-output map and how it contributes to adversarial robustness. To this end, we seek to understand how motions in input space are transformed to motions in the output space of probability distributions over class labels.
To do so, we rely on the framework of information geometry and Fisher information (Amari & Nagaoka, 1993). In particular, the network output, as a probability distribution over class labels, is endowed with a natural Riemannian metric, given by the Fisher information. We can think of the 10 dimensional vector of inputs h D in the final layer, as coordinates on this space of distributions (modulo the irrelevant global scaling h D → λh D ). In terms of these coordinates, the actual probabilities are determined through the softmax function: . The Fisher information metric on the space h D i is then given by In turn, this metric on h D induces a metric on input space x 0 via the pullback operation on metrics. The resultant metric G in on input space is given by where J = ∂h D ∂x 0 is the Jacobian from input space to layer D. Geometrically, if one moves a small amount from x 0 to x 0 + dx, the resultant distance dl one moves in output probability space, as measured by the Fisher information metric, is given by Thus the metric assigns lengths to curves in input space according to how far they induce motions in output space. Also, the Jacobian J is of independent geometric interest. As a local linearization of the input-output map, the number of non-trivially large singular values of J determine how many directions in input space the network's input-output map is locally sensitive to. To explore the geometric structure of both vanilla and saturating deep network maps, we move continuously in input space between the most confident images in a given source class, x 0 S , and a target class, x 0 T along a simple linear interpolation path in pixel space: As we move along this path, in Figure 4, we plot the length element in (12), the induced trajectory in output probability space, and the spectrum of singular values of the Jacobian J. As expected, the length element increases precisely when the output trajectory in probability space makes large transitions. At these points, one or more singular values of J also inflate.
Several distinguishing features arise in the geometry of vanilla versus saturated networks in Figure 4. The first is that the length element is more smooth and continuous for the vanilla network, but locally flat with sharp peaks when class probabilities transition for the saturating network. Thus for saturating networks, one can move long distances in input space without moving much at all in output space. This property likely confers robustness to gradientbased adversaries, which would have difficulty traversing input space under such constant, or flat input-output maps.
A second distinguishing feature is that, in vanilla networks, at probabilistic transition points, multiple singular values inflate, while in saturating networks, only one singular value does so. This implies that vanilla networks are sensitive to multiple dimensions, while saturating networks perform extremely robust and rapid transitions between distinct probabilistic outputs in a way that sensitivity to input perturbations in all directions orthogonal to the transition are strongly suppressed. This property again likely confers robustness to adversaries, as it strongly constrains the number of directions of expansion that an adversary can exploit in order to alter output probabilities.
Finally, it is interesting to compare the geometry of these trained networks to the Riemannian geometry of random neural networks which often arise in initial conditions before training. An extensive analysis of this geometry, performed by Poole et al. (2016), revealed the existence of two phases in deep networks: a chaotic (ordered) phase when the random weights have high (low) variance. In the chaotic (ordered) phase the network locally expands (contracts) input space everywhere. In contrast, trained networks flexibly deploy both ordered and chaotic phases differentially across input space; they contract space at the center of decision volumes and expand space in the vicinity of decision boundaries. Saturating networks, however, do this in a much more extreme manner than vanilla networks.
More powerful iterative adversaries
One can construct more powerful adversaries than the fast sign gradient method by iteratively finding sensitive directions in the input-output map and moving along them. How robust are saturating networks to these types of adversaries? From the information-geometric standpoint de-scribed above, given the local flatness of the input-output map, as quantified by our Riemannian geometric analysis, an iterative gradient-based adversary may still encounter difficulty with the saturated network, especially since the number of directions of expansion are additionally constrained by the compressive nature of the map.
We first created adversaries via iterative first order methods. For each chosen source image x S and its associated correct source class y S , we chose a target class, y T = y S . We then attempted to find adversarial perturbations that cause the network to misclassify x s as belonging to class y T . To do so, starting from x (0) adv = x S , we iteratively minimized the cross entropy loss via gradient descent: This procedure adjusts the adversarial example x (t) adv so as to make the incorrect label y T more likely. We used Adam (Kingma & Ba, 2015) so that the learning rate α t would be adaptive at each iteration t. For a given source class, we started with the source image the network was either least or most confident on.
Although we were able to get the vanilla network to misclassify in either case (usually within less than 10 iterations), there were several cases (such as when the source class was a 3 and the target class was a 7) where we were unable to get the saturated network to misclassify, even in the most extreme case where we ran Adam for 30 million iterations. Although the image was changing at each iteration and the mean pixel distance from the starting image was steadily increasing and converged, the resultant image did not cause the saturated network to misclassify. As a result, we moved onto second order adversaries, as Szegedy et al. (2014) had similarly considered. Thus, we considered quasi-Newton methods such as L-BFGS where we would minimize the cross entropy loss as follows: where B t is the approximate Hessian at iteration t and the learning rate α t is obtained by performing a line search in the direction p t where B t p t = −∇ x (t) adv (y S , y T ).
In Figure 5, we ran L-BFGS for 1000 iterations on both the vanilla network and the saturated network, starting with a source image that each network correctly classified but had the lowest softmax probability in that class (lowest confidence). In Figure 8 in the Supplementary Material (SM), we also include the same analysis, but starting with the most confident source image in each class.
Regardless of whether we start with a source image with the least or highest confidence in that class, we can always find an adversarial image to fool the vanilla network to misclassify as the intended target class (and usually within 1-2 iterations). However, for the saturated network, even starting with the least confident source image, we were, in the majority of cases, unable to fool the network. Moreover, as depicted in Figure 8 in the Supplementary Material (SM), it was even more difficult to fool the saturated network with the most confident source image, resulting in only 5 such cases, even after 1000 iterations.
8. Role of weight kurtosis: a linear mechanism for robustness to adversaries As we observed in §5, saturating networks had high kurtosis weight distributions in every layer when compared to their vanilla counterparts. Indeed, such kurtotic weight distributions are prevalent in biological neural networks (Buzsáki & Mizuseki, 2014). Here we demonstrate that high kurtosis weight distributions can act as a linear mechanism to guard against adversarial examples.
Indeed, sensitivity to adversarial examples is not unique to neural networks, but arises in many machine learning methods, including linear classification. Consider for example a classification problem with two cluster prototypes with weight vectors w 1 ∈ R n and w 2 ∈ R n . For simplicity, we assume w 1 and w 2 lie in orthogonal directions, so w 1 · w 2 = 0. An input x is classified as class 1 if w 1 · x > 0, otherwise x is classified as class 2. Now consider a test example that is the cluster prototype for class 1, i.e. x = w 1 . Let us further consider an adversarial perturbation w 1 + ∆x. This perturbed input will be misclassified if and only if (w 2 − w 1 ) · (w 1 + ∆x) > 0. Following the fast sign gradient method, we can choose ∆x to be the maximum perturbation under the constraint ∆x ∞ < in (3). This optimal perturbation is ∆x = sgn(w 2 − w 1 ).
In order to have this bounded l ∞ norm perturbation cause a misclassification, we must then have the condition Here, recall we are assuming w 1 · w 2 = 0 for simplicity. Thus if the l 1 norm in the denominator is small, then the network is adversarially robust in the sense that a large perturbation is required to cause a misclassification, whereas if the l 1 norm is large, then it is not. Now in high dimensional spaces, l 1 norms can be quite large relative to l 2 norms. In particular for any unit l 2 norm vector v, we have 1 ≤ v 1 ≤ √ n, where the upper bound is realized by a dense uniform vector with each entry 1 √ n and the lower bound is realized by a coordinate vector with one nonzero entry equal to 1. Both these vectors are on the l 2 ball of radius 1, but this l 2 ball intersects the circumscribing l 1 ball of radius √ n at the former vector, and the inscribing l 1 ball of radius 1 at the latter vector. This intersection of l 1 and l 2 balls of very different radii in a high dimensional space likely contributes to the prevalence of adversarial examples in high dimensional linear classification problems by allowing the denominator in (14) to be large and the numerator to be small.
However, we can avoid the bad regime of dense uniform vectors with large l 1 norm if the weights are sampled from a kurtotic distribution. In this case, we may then expect that the numerator w 1 2 2 in (14) would be large as we are likely to sample from extreme values, but that the denominator w 2 − w 1 1 would be small due to the peak of the distribution near 0. To test this idea, we sampled unit norm random vectors of dimension 20000, so that w 1 · w 2 ≈ 0. We sampled their values iid from a Pearson Type VII distribution, with density function given by f (x; γ 2 ) = c(γ 2 ) 1 + , B is the Euler Beta function, and γ 2 denotes the excess kurtosis of the distribution. In Figure 8, we computed the ratio in (14) and scaled it by the input intensity, which is given by the average absolute value of a nonzero component of w 1 . The resultant scaled ratio was then computed for each value of γ 2 . Note that a standard Gaussian has an excess kurtosis of 0, which serves as a baseline. Hence, increasing the excess kurtosis via, for example, a Pearson Type VII density, increases the scaled ratio by almost 40% from the Gaussian baseline. In fact, if we sample, via inverse transform sampling, from the weight distribution of the saturated network at a given layer, then the scaled ratio can increase by as much as 300% from when we sample from the distribution of the weights in that same layer for the vanilla network. Thus even in the case of linear classification, kurtotic weight distributions, including the weight distributions learned by our saturating networks, can improve robustness to adversarial examples.
Discussion
In summary, we have shown that a simple, biologically inspired strategy for finding highly nonlinear networks operating in a saturated regime provides interesting mechanisms for guarding DNNs against adversarial examples without ever computing them. Not only do we gain improved performance over adversarially trained networks on adversarial examples generated by the fast gradient sign method, but our saturating networks are also relatively robust against iterative, targeted methods including secondorder adversaries. We additionally move beyond empirical results to analyze the sources of intrinsic robustness to adversarial perturbations. Our information geometric analyses reveal several important features, including highly flat and low dimensional internal representations that nevertheless widely separate images from different classes. Moreover, we have demonstrated that the highly kurtotic weight distributions found both in our networks and in our brains, can act as a linear mechanism against adversarial examples. Overall, we hope our results can aid in combining theory and experiment to form the basis of a general theory of biologically plausible mechanisms for adversarial robustness. The red 'X' denotes that an adversarial image was not found in order to cause the network to misclassify for that target class. The source image was for the most confident image for that source class. | 2017-03-27T17:45:07.000Z | 2017-03-27T00:00:00.000 | {
"year": 2017,
"sha1": "600a5d60cb96eda2a9849413e747547d70dfb00a",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "600a5d60cb96eda2a9849413e747547d70dfb00a",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Mathematics",
"Computer Science",
"Biology"
]
} |
216274068 | pes2o/s2orc | v3-fos-license | Understanding the Challenges Faced by Child Focused Civil Society Organizations in Ho Chi Minh City
The presence of an active civil society is one of the most integral parts of progressive and productive community development. Within the past two decades civil society organizations have grown in size and scope throughout Viet Nam. This has been in an attempt to offset the divestment of international aid networks and match the rapid economic development occurring in urban centers like Ha Noi and Ho Chi Minh City (HCMC). The prospect of high paying jobs resulting from increased domestic and foreign investment in urban areas has encouraged rapid rural-urban migration within HCMC. As a result, urban poverty is on the rise in HCMC and the children of these vulnerable populations face unfavorable odds in educational, emotional, and legal support. The civil society organizations discussed in this paper aim to fill these gaps by supporting children in need in HCMC. However, this research finds that organizations lack sustainable funding networks, autonomy, and human resources. To ensure that their practices persist, government policy must be implemented to encourage a culture of domestic charity from individuals and businesses, as well as assist these organization's capacity building measures . • I NTRODUCTION of their successes and challenges. This report was produced to analyze a series of in-depth, qualitative Since the implementation of Doi moi ('Renovation') accounts from a sector of civil society that works policies in 1986, a formal civil society has grown with disadvantaged children in Ho Chi Minh City exponentially. In the early 1990s, international (HCMC). Findings from this research highlight the development organizations and charities entered growing importance of local community based Viet Nam to act as a catalyst for economic and social organizations (CBOs) that work with disadvantaged development. The government began introducing children in HCMC as well as the challenges they face not-for-profit science and technology research in an undefined space. Conversations with leaders organizations to work alongside existing mass, of local CSOs emphasize valuable information state-sponsored organizations such as the Farmers concerning the demographics of their clientele, Association. As the economy expanded and the network of funding, disposal of human resources, advent of internet and market economics brought programmatic focus, structure of governance, greater civilian independence, the demand for and relationship with the community, donors, and a comprehensive civil society grew. Community government. The conclusions reached in this paper members began generating ideas for community will illustrate that in order to ensure the longevity based organizations that would champion human and strength of these organizations, there must be rights and partner with better endowed international changes made to government policy that encourage organizations (IOs). Over the years, civil society charity from individuals and businesses, and improve organizations (CSOs) achieved greater autonomy their organizational capacity. These accounts are and reach in their communities. Unfortunately, the not representative of the entire network of CSOs recent increase in rural to urban migration and the in HCMC but they do provide an important socio withdrawal of international aid has put stress on the political context. sustainability of these organizations, begging the question: what challenges do CSOs currently face? BACKGROUND There has been little in-depth research Def i n i n g C iv i l Soc i ety i n V ie t N a m conducted that details the evolution of civil society in Viet Nam that also highlights first-hand accounts Civil society organizations are entities that 61 THE GW UNDERGRADUATE REVIEW • VOLUME 3 UST OF ABBREVIATIONS CBO Commun ity Based Organ ization CSR Corporate Soc ia l Respons ib i l i ty CPV Commun ist Pa rty of Vietnam cso Civ i l Society Organ ization DOLISA Department of Labour, I nva l ids , and Soc ia l Affa i rs FO i Fore ign D i rect I nvestment HCMC Ho Ch i Minh City I DA I nternationa l Deve lopment Associat ion INGO I nternationa l Non-governmental Agency 1 0 I nternationa l Organ ization MDG M i l l en i um Deve lopment Goa l MOUSA M in istry of Labour, I nva l ids , and Soc ia l Affa i rs NGO Non-governmental Organ izat ion ODA Offic ia l Deve lopment Assistance OECD Organ ization for Economic Co-operation and Deve lopment SDG Susta inab le Deve lopment Goal USO Un ited States Do l l a r UN ICEF Un ited Nat ions I nternationa l Ch i l d ren 's Emergency Fund VUSTA Vietnamese Un ion of Science and Techno logy Associat ions
• I NTRODUCTION of their successes and challenges. This report was produced to analyze a series of in-depth, qualitative Since the implementation of Doi moi ('Renovation') accounts from a sector of civil society that works policies in 1986, a formal civil society has grown with disadvantaged children in Ho Chi Minh City exponentially. In the early 1990s, international (HCMC). Findings from this research highlight the development organizations and charities entered growing importance of local community based Viet Nam to act as a catalyst for economic and social organizations (CBOs) that work with disadvantaged development. The government began introducing children in HCMC as well as the challenges they face not-for-profit science and technology research in an undefined space. Conversations with leaders organizations to work alongside existing mass, of local CSOs emphasize valuable information state-sponsored organizations such as the Farmers concerning the demographics of their clientele, Association. As the economy expanded and the network of funding, disposal of human resources, advent of internet and market economics brought programmatic focus, structure of governance, greater civilian independence, the demand for and relationship with the community, donors, and a comprehensive civil society grew. Community government. The conclusions reached in this paper members began generating ideas for community will illustrate that in order to ensure the longevity based organizations that would champion human and strength of these organizations, there must be rights and partner with better endowed international changes made to government policy that encourage organizations (IOs). Over the years, civil society charity from individuals and businesses, and improve organizations (CSOs) achieved greater autonomy their organizational capacity. These accounts are and reach in their communities. Unfortunately, the not representative of the entire network of CSOs recent increase in rural to urban migration and the in HCMC but they do provide an important socio withdrawal of international aid has put stress on the political context. sustainability of these organizations, begging the question: what challenges do CSOs currently face?
BACKGROUND
There has been little in-depth research D efi n i n g Civi l Soci ety i n Viet N a m conducted that details the evolution of civil society in Viet Nam that also highlights first-hand accounts Civil society organizations are entities that The Vietnamese government has consistently amended the law to establish control over the autonomy of the growing non-profit sector. The most recent issuance of Decree 45 confirms these motivations, making certain provisions stricter, especially those pertaining to policy criticisms. In the end, state agencies and CSOs have an 'ask-give' relationship. The state decides if an organization can become a CSO and implement their projects. While organizations may be classified as non-governmental, local officials intently monitor the activities of NGOs. The government is not often involved in the planning, funding, or execution of many of these organizations. However, it wields the power to suspend organizations or activities it does not condone. In recent years, a Law of Associations was proposed to the National Assembly to formalize and simplify the registration process. However, this proposal was turned down in 2016 because it included unconstitutional provisions and restricted international aid (Viet Nam News, 2016).
The B u s i n ess Secto r a n d CSOs The business sector has the weakest relationship with CSOs. A survey given to 115 members of civil society groups in Viet Nam cited a weak relationship between their organizations and the business sector; see Figure 1 (Le, Nguyen, Pham, & Phuong, 2018, p.55). This lack of cooperation stems from the close connection between business and the state that is reminiscent of Viet Nam's former command economy. Many Vietnamese businesses dedicate significant capitaltosustainties with government officials, who in turn vouch for businesses. Businesses protect these relationships by refraining from donating to organizations the government may not condone.
There are also very few initiatives in written law that encourage businesses to donate to a charitable organization. PricewaterhouseCoopers found that "donations are generally non-deductible, except certain donations for education, health care, natural disasters, building charitable homes for the poor,
History of I nternational Aid in Viet Nam
Since opening to global involvement, Viet Nam has been a major recipient of development aid from bilateral and multilateral donors. IOs like the World Bank and UNICEF and international non-governmental organizations like World Relief and Save the Children became drivers of a capitalist market economy in adherence to the UN's Millennium Development Goals (MDGs). By 2015, all of Viet Nam's MDGs had been met and it remains one of the strongest of the World Bank's borrowers. Within 30 years, Viet Nam's poverty rate experienced a sharp decline, from 70% to 6% (World Bank, 2019). Investment remained relatively strong, even as Viet Nam took the spot as the fastest growing economy in Southeast Asia ("Vietnam fastest-growing economy", 2019). Decades after the first round of foreign aid crossed its borders, the OECD found that Viet Nam was the fourth largest recipient of development assistance-receiving 2.308 billion USD in 2017 (Bony-Cisternes, 2019, p.2).
While international aid remains a robust part of civil society, investment numbers have begun to decline.
The classification of Viet Nam as a middle-income country in 2012 was monumental, implying that adequate economic strength, independence, and capacity building was achieved. As a result, international donors like the International Development Association -an arm of the World Bank-began withdrawing. Not only has aid been dwindling, but the aid that enters Viet Nam is starting to be allocated towards a new set of global interests. For example, the EU decided to allocate its 2014-2020 allowance (400 million EUR) to the protection of the environment and climate change (Bony-Cisternes, 2019, p.4). As climate change is addressed, funding for existing non-environmental pursuits is abandoned. Consequently, CSOs are forced to explore alternative, sustainable funding methods often within the domestic private sector. However, CSOs face a major obstacle in a lack of financial and human capital necessary for building their organization's capacity. Accounts from the non-profits leaders discussed in this research mention that in the past NGOs did not have to focus time, energy, and staff to generate funds for their organizations. Today, innovative forms of fundraising need to be generated, such as crowdfunding, corporate responsibility, public events, and grant writing; otherwise, civil society will recede.
Im p ortance of Ch ild -Foc u sed CSOs in HCMC
Ho Chi Minh City has attracted rural migrants in large credence to weak rural development initiatives within the Mekong Delta and Central Viet Nam. Climate change has threatened the present and future profitability and sustainability of agricultural production and tourism. Therefore, residents seek out higher-paying positions in factories or the informal sector in hopes of delivering remittances to families left behind in rural communities. As a result, HCMC is among the world's fastest growing cities, expanding from a population of 5 million in 1999, to about 13 million in 2019 (World Bank, 2019). In 2015 UNICEF found that 36% of HCMC's population were temporary residents, suggesting that there is a great proportion of inter-provincial migrants (2017, p.35).
Many migrants with families choose to bring their children along during the move to keep their families united and provide better educational opportunities for their children. UNICEF, in collaboration with the HCMC People's Committee, created a report titled the Situation Analysis of Children in HCMC in 2017 in response to the influx of migrant children in HCMC. This report outlined the city's goal of establishing a Child Friendly City (CFC) with the aim of protecting children's rights as outlined by the Convention on the Rights of the Child (CRC). The city's planning and budgeting focused on the CRC's four pillars: "every child survives and thrives, every child learns, every child is protected, and every child participates" (Kurukulasuriya, Nguyen, Phan, & Tran, 2017, p.16). This report found that while the city is making progress in these areas, disparities continue to grow, particularly for disadvantaged groups, such as: children with disabilities, impoverished children, and the children of migrants. A lack of birth documentation, temporary housing, and migrant status creates barriers to entry in primary and secondary education. This can influence a child's physical and emotional health, as well as their access to adequate and sustainable shelter. In the 1990s, international and local CSOs began pop-up around the city to support disadvantaged children such as: street children, child laborers, migrants, and those living with HIV/ AIDS. These organizations provide educational, physical, and emotional support to disadvantaged populations in urban areas, funded by predominantly international networks.
Understanding the history of civil society in Viet Nam and how it is situated within the law is important in comprehending many of the hierarchical and structural challenges that CSO leaders currently face. Just as well, understanding the role that international investment has played in response to the adoption of a market economy and growing demand for child-focused support services highlights the motivation behind this research.
Resea rch M et h o d s a n d Sa m p l i n g
This study was designed to answer the question: what financial and organizational challenges do child-focused civil society organizations currently face and where do they derive from? To help answer this question, research was gathered from primary and secondary sources over the course of a month in HCMC. Because it is an urban center that receives a high number of economic migrants, there is a large network of CSOs that exist to support these populations. As a result, HCMC was the most pragmatic place of study. CSOs in HCMC tend to be comprised of smaller-scale community organizations focused on services for disadvantaged groups like children and migrants, whereas CSOs in Hanoi tend to be institutes and centers that are more closely linked with the government. Through the advisor the interviewer was connected to a series of leaders within various local civil society organizations in HCMC that work with vulnerable populations of children. A variety of large civil society organizations, such as UNICEF and Save the Children, were also contacted but no response was received. To gain various perspectives on the issue, the interviewer conducted five semi-structured interviews guided by a set of questions tailored to the professional positionality of the interviewee (see Appendix A for a full list of interviewees and Appendix B for a sample of interview questions). Interviews were determined to be the best form of data collection given that each organization had a unique story that would best be told in an informal dialogue. The interviewer would be able to ask follow-up questions that offered more detail about the organization, it's history, and present day challenges. All of the interviewees were the primary leaders of their respective organizations, so they were able to speak extensively on the history, mission, financials, challenges, etc, of their organizations. The interview questions prompted discussions on the importance of their organization; the structure and administrative hierarchies that exist; their financial histories; and the challenges they are currently facing. These questions helped guide the qualitative research affording the interviewer a better understanding of the challenges non-governmental support networks face in today's climate. Interviews were conducted in-person and tailored to the availability of the interviewee. Most of the interviews were conducted in English, however one required the use of a local translator. Due to the limited time and in-country connections, convenience and snowball sampling were used to garner interviewees. A short description of the organizations successfully visited are located in the section below.
C o l l a b o rati n g O rg a n izati o n s The civil society organizations spoken with all exist as non-profit social protection services for disadvantaged children in HCMC. They provide educational opportunities, case work, shelter, food, legal aid, job skill building, counseling, family reintegration, etc. The details of each of the CSOs structure and mission are stated below.
Children's House, District 8
The Children's House was started by the Department and Ministry of Labour, Invalids, and Social Affairs in 1993 to support orphans and street children in District 8 in HCMC. Their mission is to care for and educate disadvantaged boys under the age of 14 so they may reunite with their families or build a successful and independent life with the skills they've built at the Children's House. They currently house, feed, educate, and emotionally support 14 boys, while also giving them the necessary legal support to obtain personal identity papers. While the CSO was initiated by the Vietnamese government, they do not financially support the organization; they must rely on a variety of international and local donors, such as Chayra (Germany) and AtDta Foundation (Switzerland).
Thao Dan Social Protection Center
Thao Dan is a social protection center that was founded in 1992 to support street children, ages 6-18, in HCMC. At its inception, it supported a few dozen children; today they help an average of 160 children per year. Currently, children experiencing homelessness, those that are economically disadvantaged, or child laborers are referred to this organization for support. The mission of Thao Dan work is to provide opportunities for children to grow up happily in a safe and healthy environment. They have created a safe space for children to share their experiences and offer emotional support through social workers; provide legal assistance to aid educational and housing pursuits; and offer lunch and a space to play. They rely on a system of volunteers to help carry out programming and case work. Financial support comes from large international donors like UNICEF, to help maintain programming and a small staff.
Cau Han Project
Cau Han Project is one branch of the HCMC Child Welfare Association. The mission of the Cau Han Project is to provide education to disadvantaged children in District 7. This represents those who cannot enter the public-school system because they lack proper identification papers. The organization provides students 5-grade levels of free education, partial scholarships to attend public or private schools, social support services like counseling, and legal services. Their organization is supported by the government, but they must rely on international donors like Terre des Hammes (Germany) to carry out their programming.
Green Bamboo
The Green Bamboo project is another branch of the HCMC Child Welfare Association. A group of social workers opened Green Bamboo in 1994 to support street children in District one. In 1998 they began housing, providing meals, educating, and providing counseling/ legal support for disadvantaged boys in the community. In the past they relied on international organizations to support their programming and staff, however international divestment has posed financial challenges for the sustainability of the organization.
Eth i cs
The topic in question was not inherently exploitative or invasive in nature. However, several precautions were taken to ensure the research was conducted in an ethical manner. The main concern was protecting the identity and emotions of the interviewees. The interviewer explicitly asked for verbal and written consent for permission to use names, titles, and organization names before conducting interviews. The interviewer would also state who they were and the purpose of the research prior to the conversation. Interviewees were invited to ask for clarification at any point and reminded of their ability to abstain from answering any questions. All interviews were conducted with the convenience of the interviewee in mind. At the conclusion of the interview the interviewer provided interviewees with their contact information to ensure the interviewer could be reached at any point with questions, comments, and concerns.
Empathy and gratitude were also maintained throughout the interview. While the interviewees were not compensated monetarily, they and translators were given a small gift to thank them for the time they spent aiding the research.
L i m i tations
There are several limitations that affected the authority and breadth of this research. First and foremost, the language barrier was a significant roadblock in research. Though interviews had primarily been conducted in English, the language barrier was still posed as a burden. This limited the breadth of the research because there were not as detailed, in-depth conversations as would have been ideal. Rewording or discarding certain questions were also necessary because of the language gap. Just as well, interviewees often gave answers that were only tangentially related or completely unrelated to the question asked which furthered the disconnect in understanding. The language barrier may also have affected data collection as inaccurate interpretations of verbal responses, even after repetition, may have been recorded. With more resources and/or time, this barrier could have been reduced.
The second consequence of the language barrier was the inability to gather and interpret data conducted in Vietnamese. The advisor and several interviewees had suggested several relevant studies written in Vietnamese that would have helped guide the research. As these could not be translated due to time constraints, reliance was placed on research conducted by international aid organizations like the World Bank, UNICEF, the United Nations, and other foreign research. This ultimately affected the thoroughness and bias of the research.
The limited amount of time dedicated to conducting research was also a significant barrier. While the advisor offered suggestions on several potential interviewees, a significant amount of time was spent waiting for replies to interview inquiries rather than on interviews themselves. This limited the scope and depth of the research.
The fact that the interviewer was not from Viet Nam also affected the course of the research. Heavier reliance was placed on the advisors to offer connections to different interviewees. As a result, snowball and convenience sampling was used to establish connections across HCMC. This meant that options were limited to the ability of the advisor to establish connections and for those contacts to recruit more interviewees. Random sampling would have had more legitimate and generalizable data.
There was also difficulty in gaining the perspective of each relevant stakeholder. A more in-depth analysis of the challenges CSOs face would have included the perspective of the government, large INGOs, businesses, and/or the children who use CSO services. In an effort to not exploit children in these vulnerable situations that population was not surveyed. However, there was an attempt to get the government's (DOLISA) perspective on child support services and the role of CSOs. However, given the foereign status of the researcher, there were a series of unexpected red tape necessary to get an interview. This process could not be afforded given scarce time and resources. In all, conducting research as a foreigner requires careful attention towards language and cultural barriers, as well as local knowledge.
Rural-urban Migration
An inevitable challenge for many Vietnamese citizens is the ominous threat posed by climate change. In agricultural areas such as central Viet Nam and the Mekong Delta, climate change is affecting the salinity of the water and soil, making it harder for farmers to raise these crops. Just as well, homes and businesses are being destroyed in Central Viet Nam because of flooding and storms. The consequences of climate change are approaching too fast for communities and the government to instill resiliency measures. Representatives from Thao Dan and Cau Han confirmed that many children come to their organizations because their families are victims of poor agricultural yields in the Mekong Delta. Struggling families look towards big cities like HCMC for opportunity, hearing stories of wealth and prosperity. When asked if migration and the need for their services will slow in the coming years, all my interviewees stated, 'no' , noting that any future in coastaljagricultural areas in Viet Nam is bleak.
Even provinces outside of HCMC that may not be impacted by climate change look to the city to achieve financial freedom. Dr. Nguyen from the University of Economics argued that the population increase in the city can be attested to growing inequality among regions/provinces. People only think of metropolitan areas to obtain economic mobility, causing economic migrants to flood to areas in and around HCMC. He also argued that land policy, such as tenure and quantity capacity, has forced families to relocate. A representative at Thao Dan confirmed this migrant profile, stating that many parents of the children who use their services were economic migrants, many of which now work long hours in Viet Nam's developing industrial sector. This changing economic system explains why their organization opened in the mid-1990s. Not only was this the period that the government opened to CSOs, but migration networks increased because of increased opportunity that resulted from foreign direct investment and a competitive market economy. More migrants mean more children that face challenging odds regarding educational and emotional support, requiring CSOs to build a larger capacity to serve these populations.
Lack of Documentation
One of the greatest challenges for children who utilize these CSOs is a lack of personal documentation. Many of the children that use these services do so because they lack the appropriate identification papers to attend public school. In speaking with Mr. Linh at the Children's House, he says that because the boys are orphaned or have migrated from other provinces, they lack birth documentation that is necessary to register for public school. Instead these boys must attend school sessions for undocumented children that are shorter in length and less developed. Chi Huong from Cau Han discussed why so many children are undocumented in the provinces; over the years they found that families that give birth in hospitals leave before obtaining their child's birth certificate in order to avoid hospital costs. Families may also not be aware of the implications of not registering their children, and don't have the time or resources to pursue it later on. Many CSOs in HCMC, including the four I spoke with, devote considerable resources and time to help children obtain identification papers. These papers are critical for a child trying to attend school or receive basic healthcare services.
Weak Fundraising Capacity
One of the most striking findings from conversations with CSOs was realizing the lacking capacity for fundraising. For over two decades these organizations have had large, stable international donors to rely on. Now these networks are fading away, and organizations are left without the capacity to independently seek out funds. The capacity to fundraise was not built into the staffing or programming of these organizations, therefore strong networks of domestic donors do not exist. Chi Lieu commented on this sentiment, stating that before, "[her] organization just used money, now [they] have to get money." She furthered stated that her skills are in social work, not fundraising. The decline of human capital in these small CSOs makes it almost impossible to dedicate the appropriate amount of time to find grants and sustainable donors to build a local network of financial support. Organizations like the LIN Center have entered the space to aid capacity building within these organizations through grant opportunities and consulting. Unfortunately, LIN does not have the human and financial capital necessary to aid such a large list of non-profit organizations in need of capacity building in HCMC.
Chi Huong said the future of financial freedom starts with emphasizing corporate social responsibility. As discussed earlier in the report, corporations have no responsibility to CSOs that support community development. She cited that the government has not instituted any policy to encourage or incentivize this behavior. One of the policies that she thinks could encourage this growth is tax incentives and advocacy work. The rise of internet mediums has improved transparency, inspiring a new relationship of trust between donors and CSOs. Chi Ngan said that local people were willing to contribute to the welfare of others through philanthropy but find distaste in the lack of transparency. Through sites like Facebook, organizations can post about their donations and activities, improving the transparency and reactivity of donations. Grand scale change will not be possible without policy to back it. This begins with a better relationship developing between CSOs and the government.
Autonomy
As previously discussed, the current laws governing civil society limit the autonomy of organizations. The government can cancel programming that doesn't align with their beliefs, or prevent the launch of certain CSOs altogether. For most of the organizations I spoke with it has not presented a major problem because their work is less controversial, however Cau Han does have experience with government intervention. The inaugural project at Cau Han in 1997, sexual safety programming, had to change gears after three years and provide job skill training because the government had not wanted to promote prostitution. In the future, policy broadening the autonomy and rights of organizations should be proposed to ensure that the needs of the community are being met.
Non-profit Law & Tax Incentfoes
In places like the United States, CSOs often benefit from 501(c)3 status. This status, as defined in tax law, allows those who donate to the registered non-profit to write off any of those donations on their taxes. This encourages donations from people with smaller personal wealth, as well as businesses. No such status exists in Vietnamese law. Chi Ngan explained that people within the community or neighborhood donate to their efforts through in-kind donations, such as notebooks, food, or clothing; Chu Linh told a similar narrative. There is not one reason for this decision, however people are not incentivized to even give the smallest of monetary donations because there is no government policy incentivizing action. She also noted that people who donate to organizations like to see the immediate impact or trajectory of their donation. By donating in-kind, such as notebooks or clothing, you can control the destiny of your investment. By streamlining the registration process for CSOs and restructuring CSO law, a system for tax-deductible donations could theoretically be built.
Attitude Towards Migrants
Chi Ngan also explained her frustration with the government's inattention towards migrants. She says that the government has sponsored reports citing a significant decrease in the population of street children/ families in HCMC over the past decade. These studies were conducted by research institutions that must operate within the framework of the government. Her organization conducted research of their own, finding a higher number of disadvantaged peoples in HCMC. She said that the government lacks local knowledge of migrant settlement patterns, and therefore conducts inaccurate research. The consequences of inaccurate data are that it signals to international donors that their investment in CSOs might not be needed. Dr. Doan discussed that the city is constantly trying to paint HCMC as a rich and prosperous city; it would not be in the government's interest to publish data that would challenge those goals. They are in the business to encourage investment and tourism to places like HCMC. In the future, independent research institutions that utilize local knowledge will have a greater role to play in development policy and CSO strength.
CONCLUSION
Over the course of a month, I conversed with a series of changemakers that have dedicated their lives to advocate on behalf of and improve the lives of disadvantaged children in HCMC. Twenty years ago, these local civil society organizations arose out of a need to supp ? rt vulnerable children affected by the retrenchment penod post Doi M6i. While Viet Nam has achieved relative economic prosperity over the past twenty years, the need for these services remains robust. Climate change, growing income inequalities, and a rapidly changing economy have urged a new wave of rural-urban migra � ion. The capacity to support these vulnerable populations is being challenged by the divestment of international aid; government policy; and a lacking financial capac � t � . The sustainability of this vital sub-sector of civil society is contingent on the active role of g � vernm � nt policy change that would encourage domes � ic chan � y; demand CSR from the business sector; and aid capacity building measures for CSOs. The testimonies that this research has compiled hope to inspire and be utilized by community members in Viet Nam to demand action from the government. Just as well, this country's history is not unique. Communities with similar histories plagued by imperialism and colonization have often received streams of international aid that have provided momentary relief, while undermining the sustainability and capacity of the organizations they are trying to aid.
Fu rther Researc h
Over the course of my research period I came across ideas that were beyond the scale of this research but would be interesting extensions on the topic. First and foremost, it would have been beneficial to meet with more local CSOs to obtain a larger sample size of the support networks available to disadvantaged children in HCMC. With a greater sample size, I would have gained a broader perspective of the challenges, successes, and ideas for the future of sustainable civil society capacity building.
Next, I would have liked to speak with larger INGOs in the area, such as UNICEF and Save the Children, to understand their reach and relationships in HCMC. What is the future ofiIOs in Viet Nam, given Viet Nam's increased economic status and growing civil society? Are I Os trying to aid the capacity building of these organizations or are they merely providing short-term solutions? Are there elements of Vietnamese law that limit the reach or breadth of their programming?
Finally, as I met with interviewees and did more research on the role of corporate responsibility, I realized the growing importance of the business sector in the sustainability of civil society organizations. Speaking with large businesses in HCMC would uncover why they would or would not donate to different CSOs. Are businesses looking to be incentivized by government policy? What kind of CSOs would they be willing to donate to?
ACKNOWLEDGE MENTS
This research project would not have been possible without the guidance and support of my advisors: Duong Van Thanh and Nguyen Thi Minh Chau. I am extremely grateful for their unwavering support and dedication to help me pursue this research.
I would also like to thank all of the individuals that graciously agreed to take the time out of t � eir bus y _ schedules to share their stories, thoughts, and opm10ns: Le Thi Ngan, Nguyen Doan, Nguyen Tri Linh, Tong Thi Hu � ng, and Nguyen Thi Bich Lieu. Gaining these perspectives was a critical part of the research and afforded me a new outlook on community development that I plan to carry with me into my academic and professional pursuits. The complete index of organizations and individuals can be found in Appendix A. I would also like to thank the network of local students for their eagerness and willingness to aid our translation needs during the research process. | 2020-04-02T09:31:25.004Z | 2020-01-01T00:00:00.000 | {
"year": 2020,
"sha1": "8471183c85e6d6aa7d27378e901513af5b6c8cd0",
"oa_license": null,
"oa_url": "http://scholarspace.library.gwu.edu/downloads/pn89d7377",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "29384ce03390a900cd275b08846d065b0969ab0b",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Political Science"
]
} |
236469820 | pes2o/s2orc | v3-fos-license | Sequential development of ANCA‐associated vasculitis and anti‐GBM disease: A report of two cases
Abstract In case of AAV with kidney involvement, physicians should explore anti‐GBM antibodies and be aware of the possible sequential development of AAV, especially with MPO‐ANCA, and anti‐GBM glomerulonephritis. This sequential disease history is associated with a poor renal outcome, highlighting the need for urgent diagnosis and management.
| INTRODUCTION
Double-positive vasculitis for ANCA and anti-GBM antibodies simultaneously is well described. Conversely, cases of sequential development of anti-GBM disease after ANCA-associated vasculitis are exceptionally reported. We describe 2 cases of ANCA-associated vasculitis followed by anti-GBM disease, suggesting that glomerular damages due to ANCA-associated vasculitis could induce an anti-GBM glomerulonephritis.
Antineutrophil cytoplasm antibodies (ANCA) targeting myeloperoxidase (MPO) are commonly found in antiglomerular basement membrane (GBM) disease. The association of small vessel vasculitis double positive for ANCA and anti-GBM antibodies occurring simultaneously is well described. 1,2 In contrast, sequential development of ANCAassociated vasculitis (AAV) followed by anti-GBM disease is rarely reported. Recently, evidence suggests that glomerular damages due to ANCA-associated glomerulonephritis could reveal sequestered epitopes of the GBM, inducing an anti-GBM immune response.
A survey across tertiary centers for the management of vasculitis affiliated to the French Vasculitis Study Group allowed us to identify two cases of AAV followed by biopsy-proven anti-GBM disease: a 60-year-old man with eosinophilic granulomatosis with polyangiitis (EGPA) and a 23-year-old woman with granulomatosis with polyangiitis (GPA). Despite prompt management, adequate induction therapy, and plasma exchanges, the first case reached endstage renal disease and the second case experienced a relapsing anti-GBM glomerulonephritis. The sequential occurrence of the two diseases is exceptionally reported in the literature, affecting preferentially elderly males with MPO-ANCA and a poor renal prognosis.
| CASE 1
A 60-year-old man with a history of asthma and chronic sinusitis presented a chronic nonproductive cough, recurrent fever, and limbs' neuropathic pain. Blood tests revealed persistent eosinophilia with mild inflammatory syndrome. Renal function and urine sediment were normal. Chest-computed tomography showed a diffuse interstitial lung disease with micronodules. Electroneuromyography revealed multiple mononeuropathy. Infectious serological tests and parasitological investigations were negative. Serum complement levels were normal. Antinuclear antibodies and serum cryoglobulins were negative. ANCA were positive, identified as MPO-ANCA. The patient was diagnosed with EGPA. Highdose glucocorticoids led to clinical improvement, allowing a slow tapering and withdrawal three years later.
Five months after glucocorticoids weaning, he developed an acute renal failure (creatinine serum level 7.2 mg/ dl from 1.1 mg/dl previously) together with hematuria, mild proteinuria, elevated inflammatory parameters, and normal eosinophils count. High titers of MPO-ANCA (>200 UI/ ml, N < 3.5 UI/ml) and anti-GBM antibodies (>200 UI/ml, N < 20 UI/ml) were detected. There was no alveolar hemorrhage on chest computed tomography. Kidney biopsy revealed a necrotizing and crescentic glomerulonephritis without rupture of Bowman's capsule, with IgG linear staining along the GBM on immunofluorescence (Figure 1), consistent with anti-GBM glomerulonephritis. He was treated with plasma exchanges, high-dose glucocorticoids, combined with rituximab, or cyclophosphamide (according to a double-blind randomized controlled trial). Anti-GBM antibodies were cleared, while MPO-ANCA remained detectable. Azathioprine was administered as maintenance therapy, replaced by rituximab because of digestive side effects. Unfortunately, despite treatment, kidney failure progressed to end-stage renal disease requiring dialysis.
| CASE 2
A 23-year-old woman with a history of chronic sinusitis, developed central diabetes insipidus with an enlarged pituitary gland on magnetic resonance imaging, bilateral renal pseudotumors and MPO-ANCA (6.3 UI/ml, N < 3.5 UI/ml) suggesting GPA. Renal function and urine sediment were normal. A biopsy of renal pseudotumor revealed a necrotizing glomerulonephritis with granulomatous inflammation, extracapillary proliferation, and fibrinoid necrosis-confirming GPA ( Figure 2). Immunofluorescence was negative. In order to preserve her fertility, rituximab was administrated as induction therapy, with glucocorticoids, followed by an azathioprine-based maintenance therapy. Azathioprinerelated gastrointestinal toxicity required replacement by methotrexate. The outcome was favorable with regression of renal masses and ANCA titers normalized. However, central diabetes insipidus persisted and was treated with desmopressin. Methotrexate combined with prednisone was continued as maintenance therapy for three years.
Four months after the end of maintenance therapy, she was admitted with fever, dyspnea, nonproductive cough, and weight loss. Biological findings showed an acute kidney injury (creatinine serum level 5.2 mg/dl from 0.9 mg/ dl previously) with hematuria and mild proteinuria. Chestcomputed tomography was normal. Renal ultrasound did not show any renal mass. Autoimmune panel showed high titers of anti-GBM antibodies (> 200 UI/ml) and transiently positive ANCA antibodies with a cytoplasmic staining pattern at a titer of 1/80 neither identified as PR3 nor MPO-ANCA. Kidney biopsy revealed a crescentic glomerulonephritis with IgG linear deposits along the GBM on immunofluorescence (Figure 3), establishing the diagnosis of anti-GBM disease.
The patient was treated with plasma exchanges, high-dose glucocorticoids, and rituximab. Clinical status and renal function improved. Unfortunately, six months later she experienced a relapse of anti-GBM glomerulonephritis with an acute renal failure (creatinine serum level 2.2 mg/dl from 1.4 mg/dl previously) and high titers of anti-GBM antibodies (148 UI/ml). ANCA were negative. A second line of immunosuppressive therapy by high-dose glucocorticoids and intravenous cyclophosphamide was initiated, with iterative plasma exchanges. Anti-GBM antibodies titers decreased progressively but remained detectable. Treatment combining progressively taperedprednisone and oral cyclophosphamide was pursued without further renal function deterioration. Her actual stabilized glomerular filtration rate is 52 ml/min/1.73 m² (creatinine serum level 1.3 mg/dl).
| DISCUSSION
We report two cases of MPO-AAV characterized by the development 3 years later of anti-GBM glomerulonephritis: a first case of EGPA without initial renal involvement and a second case of atypical GPA with renal pseudotumors and a pauci-immune crescentic glomerulonephritis. Since the initial clinical presentation of these 2 cases was suggestive of AAV, anti-GBM disease was not considered at first and anti-GBM antibodies were not searched at that time. However, circulating anti-GBM antibodies are found in approximately 10% of cases of AAV. 2 In addition, low titers of anti-GBM antibodies can be found years before the onset of any clinical feature of anti-GBM disease. 3 The sequential occurrence of these vasculitis suggests a pathophysiologic involvement of ANCA in the development of anti-GBM disease. Glomerular inflammation related to AAV could induce damage of the GBM, revealing sequestered epitopes of type IV collagen leading to a phenomenon of epitope spreading with the production of anti-GBM antibodies and the development of anti-GBM glomerulonephritis. 1,2 This sequence beginning by an overt flare of AAV should be distinguished from the entity of small vessels vasculitis double positive for ANCA and anti-GBM antibodies occurring simultaneously. Subclinical elevated ANCA before the onset of anti-GBM disease have been reported. In a case-control study involving serum samples (some of which collected more than five years before the diagnosis of anti-GBM disease) of 30 patients diagnosed with anti-GBM disease, Olson et al. found a significantly higher percentage of disease subjects with detectable ANCA (>1 UI/ml) in multiple serum samples over time before the onset of anti-GBM disease compared with matched healthy controls for age, gender, and race. 3 By contrast, the sequential development of the two diseases has exceptionally been reported previously and limited to 4 other cases (Table 1). [4][5][6][7] They reveal a male predominance with an advanced age at AAV diagnosis and a MPO-ANCA subtype predominance, which is characterized-in the setting of isolated AAV-by a higher frequency of kidney involvement. 8 Renal prognosis is poor after biopsy-proven anti-GBM disease development, since only one patient remained dialysis-free. In this latter case, kidney function remained stable despite the persistence of anti-GBM antibodies and recurrent anti-GBM glomerulonephritis, which is infrequent in contrast to AAV. Given the patient's young age, the concern of avoiding gonadal failure, and despite the lack of evidence of noninferiority of this induction therapy, rituximab was preferred to cyclophosphamide. In case of double-positive vasculitis, a renal involvement is systematically observed and the kidney function seems to be more severely impaired. Serum creatinine level is significantly higher than in isolated AAV in which renal improvement is more frequently observed through immunosuppressive therapy. 9 According to different studies comparing the prognosis of these double-positive patients to patients with isolated anti-GBM antibodies, we can observe conflicting results. Some studies suggest a better outcome in terms of renal prognosis 10,11 but other studies have failed to demonstrate similar results. 1,12 In 1996, Heeringa et al. demonstrated an aggravation of anti-GBM nephritis in case of immunization with MPO-ANCA in a murine model. 13 In 2017, McAdoo et al. reported more chronic kidney damages with sclerotic glomeruli, interstitial fibrosis, and tubular atrophy in cases with double-positive vasculitis compared with isolated anti-GBM disease. Nevertheless, the prognosis does not seem to be worse. Indeed, they observed a better renal outcome. 10 In addition, these doublepositive patients may have had an atypical presentation leading to a delayed diagnosis, and thus delayed initiation of an immunosuppressive therapy and plasma exchange. Furthermore, renal prognosis of double-positive vasculitis presenting concomitantly ANCA and anti-GBM antibodies and sequential occurrence of the 2 vasculitis have to be distinguished from each other. Anti-GBM antibodies should be monitored since the diagnosis of AAV and during the follow-up of the patients, guiding toward the onset of an associated anti-GBM disease and leading to a prompt management which could improve renal survival. The onset of these 2 uncommon distinct diseases, occurring distinctly by 3 to 4 years delay, cannot be the result of chance only, assuming a causal relation between the 2 diseases. However, further work is necessary to understand the involvement of AAV in the occurrence of anti-GBM disease, including its immunopathogenic mechanism.
| CONCLUSION
In case of diagnosis of AAV or anti-GBM disease, the presence of both ANCA and anti-GBM antibodies should be routinely explored. The few reported cases-including ours-highlight the existence of sequential occurrence of AAV and anti-GBM disease, in which an anti-GBM response following kidney damage due to the AAV is suspected. They are characterized by exclusively MPO-ANCA, old age, and poor renal outcome, characteristics not shared by our second patient. A close monitoring of ANCA and anti-GBM antibodies and the maintenance of a long-term immunosuppressive therapy are important to prevent the potential development of an anti-GBM glomerulonephritis and a dialysis-dependent renal failure. | 2021-07-29T05:27:10.432Z | 2021-07-01T00:00:00.000 | {
"year": 2021,
"sha1": "3e6ccefddd97f18da6cdfc310c5ef5fdbe70463d",
"oa_license": "CCBYNCND",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ccr3.4553",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "3e6ccefddd97f18da6cdfc310c5ef5fdbe70463d",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
12823397 | pes2o/s2orc | v3-fos-license | Boerhaave's Syndrome: Still a Diagnostic and Therapeutic Challenge in the 21st Century
Boerhaave's syndrome is a rare but potentially fatal condition characterised by a transmural tear of the distal oesophagus induced by a sudden increase in pressure. Diagnosis is challenging as the classic triad of vomiting, abdominal or chest pain, and subcutaneous emphysema is absent in many patients. Management is multidisciplinary and relies on rapid, distinct, and repeated imaging. Treatment has not been standardised and may be conservative, endoscopic, or surgical. We present a typical case which illustrates possible diagnostic pitfalls and the therapeutic conundrum surrounding management of the syndrome. Based on time of presentation and eventual presence of sepsis, a therapeutic algorithm is proposed.
Introduction
Boerhaave's syndrome, first described in the 18th century by the Dutch physician Herman Boerhaave, refers to an oesophageal tear caused by an abrupt rise of intraluminal pressure [1]. It is a rare but life-threatening condition that requires urgent diagnosis and treatment. If treatment is delayed, severe and potentially lethal complications such as mediastinitis, pleural empyema, septic shock, and multiple organ failure may develop. A high index of suspicion is imperative for timely diagnosis and to assure well-selected radiological and endoscopic investigations. Prompt and adequate therapy reduces mortality. Some centers advocate early and extensive surgery as the cornerstone of treatment, yet others prefer a more conservative, endoscopic, or less invasive surgical approach. Such experience-based difference in attitude likely explains the lack of well-established treatment guidelines.
Case Report
A 41-year-old man with an extensive medical history, including Child-Pugh B liver cirrhosis and chronic pancreatitis, consulted his family doctor with progressive dyspnea, retrosternal pain radiating to the back, repeated vomiting, and fever for 3 days. On admission at the emergency ward, the patient appeared confused and agitated. Core temperature was 38.5 ∘ C, heart rate was regular at 150 beats/min, and respiratory rate was 40 breaths/min. Physical examination revealed a supple but tender abdomen, normal peristalsis, and muffled breath sounds over the right lung. Blood analysis showed macrocytic anemia, 18500 leukocytes/mm 3 , normal enzymes, a C-reactive protein (CRP) of 303 mg/L, and a lactate level of 7.4 mmol/L. Chest X-ray showed a silhouette sign over the right heart border and small mediastinal radiolucent streaks of air ( Figure 1). A contrastenhanced computed tomography (CT) scan of the thorax confirmed the pneumomediastinum (Figure 2(a)) and showed a dilated oesophagus with a distal tear, bilateral pleural effusions, and heterogeneous retro-and paracardiac collections (Figure 2(b)). The diagnosis of Boerhaave's syndrome was made. Blood cultures were taken; intravenous antibiotic therapy ( -lactam + aminoglycoside) and ample fluid resuscitation were initiated. The patient was transferred to the ICU. Two hours after ICU admission, the patient became increasingly oxygen-and vasopressor-dependent. A new chest X-ray revealed a rapidly evolving right pleural effusion ( Figure 3). Subsequently, the patient was intubated and mechanically ventilated. A chest tube was inserted in the right pleural space. A brown, foul-smelling fluid was evacuated. The pleural fluid did not contain food particles, pH was 7, and amylase was 685 U/L. Gastroscopy confirmed an oesophageal tear in the lower one-third of the oesophagus. At that time, a severe concomitant mucosal inflammation precluded further endoscopic treatment because of a too high risk of oesophageal rupture. A control CT scan demonstrated a left-sided paravertebral encapsulated air-fluid collection and a left pleural effusion that had substantially increased in size. A pigtail catheter was inserted into the left pleural space. CT guided drainage of encapsulated pleural collections was performed. Four days after ICU admission, a selfexpandable oesophageal stent was endoscopically placed. Despite stenting, continuous bilateral pleural drainage, and broad antimicrobial covering, the patient remained septic with hectic fever and persistently high leukocyte counts and CRP levels. Video-assisted thoracoscopic surgery was performed with drainage and rinsing of the mediastinum and pleural cavities. Thereafter, fever subsided and clinical and biological parameters progressively normalised. After an ICU stay of 33 days, the patient was discharged.
Discussion
Boerhaave's syndrome is a barogenic tear of the oesophagus caused by a sudden rise of intraluminal pressure in its distal end. It accounts for approximately 15% of all cases of oesophageal rupture and has a mortality rate up to 40%. In the majority of cases, the rupture is located in the left posterolateral wall of the distal third of the oesophagus [2].
Boerhaave's syndrome is readily suspected in a patient with a history of overindulgence in food or drinks who, after severe or repeated vomiting, experiences excruciating chest pain and develops subcutaneous emphysema [3]. However, up to one-third of patients have atypical symptoms or are admitted with severe respiratory distress and/or shock. The differential diagnosis of Boerhaave's syndrome includes a variety of acute thoracic and abdominal conditions including myocardial infarction, pulmonary embolus, dissecting aorta, ruptured aortic aneurysm, perforated peptic ulcer, Mallory-Weiss syndrome, pancreatitis, pneumonia, and spontaneous pneumothorax [4]. On physical examination, patients generally appear acutely ill with tachycardia and tachypnea. Fever may be present or not. Auscultation occasionally reveals decreased breath sounds on the perforation side. When mediastinal emphysema is present, Hamman's sign (i.e., a mediastinal "crackling" accompanying every heart beat) may be heard in left lateral decubitus position [2].
Blood laboratory tests are of little help to support the diagnosis, except for excluding other pathologies (e.g., myocardial infarction, pancreatitis). Presence of food particles, a high amylase content, and a low pH in pleural fluid are either highly confirmatory or very suggestive of oesophageal perforation [2,3].
Imaging is of key importance for diagnosis. Plain chest X-rays may reveal subcutaneous and/or mediastinal emphysema, mediastinal widening, pleural effusion(s), pneumothorax, hydrothorax, and intrathoracic air-fluid levels or masses [5]. In 20% of cases, a "V sign" is noticed which appears as a radiolucent streak of air dissecting the retrocardiac fascial planes [6]. A rapidly developing or evolving effusion requires urgent investigation. Importantly, up to 15% of patients have normal chest roentgenograms. CT scanning of thorax and mediastinum has largely replaced the former gastrografin swallow tests. CT indeed allows a more detailed assessment of the lungs, mediastinum, pleura, and aorta. CT also is more sensitive than plain radiography for detection of small amounts of paraspinal or pleural air-fluid collections [7]. Finally, endoscopy enables direct visualisation of the location and extent of the perforation but must be performed with caution since it may potentially aggravate the oesophageal tear.
Management of Boerhaave's syndrome remains a controversial issue. Basically, three levels of treatment are distinguished: a conservative, an endoscopic, and a surgical approach [8,9]. Conservative treatment consists of cessation of oral intake, administration of fluids and parenteral nutrition, broad-spectrum antibiotics, H 2 -blockers, and eventual mediastinal, pleural, or abscess drainage. It can be offered to selected, nonseptic patients with a small or well-contained perforation [10]. Endoscopic therapy is increasingly used in patients whose perforation is diagnosed early without widespread contamination and sepsis. Endoluminal placement of a self-expandable metallic stent to bridge an oesophageal tear has shown encouraging results. Yet, endoscopic stenting in Boerhaave's syndrome has a rather undetermined and center-dependent success rate and may not be devoid of side effects such as enhanced mediastinal or pleural contamination and accidental stent migration [11]. Finally, surgical treatment ranges from a less invasive approach consisting of debridement and drainage of mediastinum and pleural cavities to extensive resection of the thoracic oesophagus [8,9,12]. Factors determining surgical intervention are the extent of the perforation, eventual concomitant pathologies that require synchronous management, the degree of mediastinal or pleural contamination, and the presence of sepsis. Abbas et al. advocate primary surgical repair when patients present with sepsis, large uncontained leaks, and extensive contamination [13]. Based on a literature review, de Schipper et al. propose to perform surgery (i.e., open thoracotomy with resection, hemifundoplication, and pleural/mediastinal drainage) in patients diagnosed early (<48 h), regardless sepsis is present or not, and in patients diagnosed beyond 48 h who remain or become septic under conservative treatment [14]. Experience with less invasive surgical procedures is scarce. Haveman et al. compared open thoracotomy with video-assisted thoracoscopic surgery (VATS) and found similar success rates [15]. More prospective studies are warranted to establish the potential advantage of VATS in treating patients with Boerhaave's syndrome. To deal with single incident cases, a treatment algorithm is composed and presented in Figure 4. According to this algorithm, it could be argued that early surgical intervention might have been the preferred optimal first-line treatment in our patient.
Conclusion
Boerhaave's syndrome still represents a diagnostic and therapeutic challenge. Timely recognition shouldered by repeated radiological imaging is an important prognostic determinant. The choice of treatment depends upon critical and continuous assessment of the patient's clinical status, the extent and duration of the leak, and the presence of sepsis. Minimally or less invasive treatment options have been applied successfully, but their impact must be evaluated in prospective trials. | 2018-04-03T01:01:44.202Z | 2013-06-09T00:00:00.000 | {
"year": 2013,
"sha1": "75638bbf7801f2fcb4c0833bbecbdd42e60cb1dd",
"oa_license": "CCBY",
"oa_url": "http://downloads.hindawi.com/journals/cricc/2013/161286.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "75638bbf7801f2fcb4c0833bbecbdd42e60cb1dd",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
235251537 | pes2o/s2orc | v3-fos-license | Approaches to the formation of an optimal personal investment portfolio in Ukraine
ARTICLE INFO Liudmyla Gavatiuk, Maksym Karvatskyi, Alina Korbutiak, Natalia Sokrovolska and Eduard Yurii (2021). Approaches to the formation of an optimal personal investment portfolio in Ukraine. Investment Management and Financial Innovations, 18(2), 91-105. doi:10.21511/imfi.18(2).2021.08 DOI http://dx.doi.org/10.21511/imfi.18(2).2021.08 RELEASED ON Thursday, 06 May 2021 RECEIVED ON Thursday, 19 November 2020 ACCEPTED ON Tuesday, 27 April 2021
INTRODUCTION
The culture of consumption and saving and the relative basic principles of money management are an integral part of social life in Ukraine. The core rules to money management are as mandatory for every citizen as traffic rules or computer/phone skills.
Population's skill to manage and accumulate financial resources plays an essential role in economic processes. The proportion, which represents how the available finances are distributed for consuming and saving, influences not only the economy directly through consumer demand, but also the speed of fixed capital accumulation, thereby providing a long-term perspective for the country's economic growth.
With the development of market relations, investment activity of citizens, including Ukrainians, is becoming more important for stability of national economies. Financial resources of the population make up a significant sector of financial relations, since they accumulate a greater part of social savings. Under conditions of market relations, savings have a distinct economic growth in a form of investment capital. Despite the fact that citizens when investing are guided by personal reasons, their role in providing the national capital gain is quite substantial. Involving the citizens in active investment allows ampli-However, globalization and innovation in financial markets make it increasingly difficult for citizens to make personal investment decisions. For an average person with no relative experience and knowledge, the investment is often a challenging and complicated process. That is why an attempt to gain an understanding of main reasons, instruments and criteria for making diversified investment by a natural person is quite topical.
The starting point for personal investment is the formation of its own investment portfolio, assessment of investment opportunities, investment attractiveness of the external environment. There are also many factors (for example, economic crises, inflation, financial illiteracy, exchange rate jumps, stock market cataclysms) affecting the safety and return on investment. Therefore, the question of how to make an investment portfolio is so important.
The approach to the formation of an investment portfolio deals with the desire to diversify risks as much as possible while maintaining profitability. For each individual, the optimal option is when the maximum effect is achieved in the form of return on investment, and the risk of loss is reduced to zero. However, this situation is not possible under market conditions, so the person should always consider the restrictions on risk or return.
Given this, the main motives, tools and criteria of diversification of individual investment portfolio are relevant today.
LITERATURE REVIEW
Nowadays, the ability to manage and accumulate personal resources plays an important role in economic processes, since the financial resources of the population form a significant sector of financial relations in society, and under market conditions, these savings acquire a form of investment capital. Therefore, it is important to take into account the current experience of behavioral models of individuals in making investment decisions.
Notable scientific achievements in economic decision-making by individuals and its influence on economic processes belong to Kahneman (2002). He determined the influence of psychological factors on making investment and purchase decisions. Application of psychological methods to economic analysis allows claiming that investors' revenue on a stock market depends not only on risky operations and faulty decisions, but also reaction to incoming data, evaluation of which depends on psychological perception. The statement about the fact that financial market functioning and investment services spreading considerably depend on people's behavior, preferences and aspirations is no longer hypothetical.
Standard economic models come from a concept that people seek their maximum benefit and minimum expenses. The supporters of behavioral theory (Kahneman et al., 1982) disconfirm some of the traditional dogmas by showing that people often make decisions based on assumptions, emotions, intuition and empirical rules, but not on analysis of revenues and expenses; that markets "suffer from disease of group thinking and herding behavior"; and that investment decision can often be influenced by how the suggested decisions are made.
The global economic crisis has revealed the inefficiency of investment decisions and fallibility of decision-making by separate individuals and financial organizations in the process of their activities, which has led to global economic consequences (Kizyma, 2009). Hence, a dilemma of optimal investment decisions and economic choice as to investment instruments is still topical and needs further research.
"One of the main ideas of behavioral economics borrowed from psychology is a wide spread of excessive confidence. People do the things, which they should have done at all, because they believe in their success", says Kahneman in his Nobel lecture. He calls it "unrealistic optimism", which is one of the driving forces of capitalism because many people are unaware of risks they take. This statement of the scientiest is also represented in the book by N. Taleb "The Black Swan", where it is mentioned that people do not pay attention to very possible consequences of rare but extensive destructive events, which affect our assumptions about the true future.
So, individual investment decision can have a direct or indirect influence on other people, and most of the time these decisions are not economically supported and based on personal experience or rational expectations, rather than common sense and mathematical calculations. Such a conclusion plays an important role in practical investment portfolio as people owing to their optimism or being underinformed can make wrong financial decisions, which can negatively affect them or their surroundings. Traditional economic theory states that individuals reach decisions on reasonable grounds. However, in real life people often tend to be irrational when assessing the potential alternatives, since they make decisions based on limited amount of information, which is, moreover, not always reliable.
In this regard, "the prospect theory", which is based on the theory of "unrealistic optimism", is considered valuable as it provides a more profound explanation to the rational expectations theory. Therefore, according to "the prospect theory" by Kahneman and Tversky (1982), an average person cannot correctly assess prospective benefits in absolute terms. The results of this research appear to be quite practical when applied to investment decision-making in the middle or long term. However, irrational behavior is not dominant in investment decision-making, since the creation of an investment portfolio requires a deliberate choice of investment instruments and means.
Another obstacle encountered when creating a high-income investment portfolio is asymmetric information. The problem of decision-making under asymmetric market information was examined by Akerlof (1970), the Nobel Prize laureate in Economics. He discovered that when entering into a transaction different subjects have different in-formation regarding the object of the transaction, which influences their choices This conclusion is of an immense value for the research of decision-making in the sphere of personal investment under limited information. Insufficient awareness and expertise lead to wrong decisions in personal investment that decreases investment efficiency, and in imperfect market structures can even lead to a limited investment choice by the large masses of people (Akerlof, 1970).
Insufficient awareness, self-assurance, false or out-dated information and stereotypical thinking have a considerable influence on decision-making. One of such manifestations, which is innate to the modern society, is a phenomenon of "herding instinct", which was introduced into economics by Keynes (1973). The expressions of herding instinct include consumer sentiment, deception, illusions regarding money value and general assumptions about how economy works. One of the fundamental factors determining the structure of investment for citizens is their main reasons for savings and potential expectations. The latter can often foster self-assurance, which blunts the sense of danger and leads to excessive investment in unreliable or quite risky financial instruments. People believe that shares will invariably bring them huge revenue in time that buildings will rise in price every year. It is a kind of a multiplier: more trust means more economic activity. Monthly consumer sentiment index, which is calculated by the University of Michigan in partnership with Reuters and is considered to be the most famous barometer of sentiment, increased in April 2019 to 97.2 index points compared to the previous index 96. 9. Moreover, the index is at the level much higher to its historical maximum (111,4 in 2000) than its minimum (51,7 in 1980), which reflects the strong confidence in tomorrow against all social and political problems (World Economic Forum, 2019).
To analyze the process of creating investment portfolio, it is worth considering the thought of Shiller (2005) about the fact that people's narration about past economic event can also influence consumers' decisions regarding some financial instruments or institutions. He states that "a serious crisis strikes down the optimism, and the consequences can be perceptible for decades" (Shiller, 2005).
From the standpoint of measuring the institutional quality of the social sector and, hence, the financial capacity of individuals to form their investment portfolio, the sudy by Vasilieva et al.
(2018) on assessing the institutional quality of the social sector is of interest. The authors studied the functioning of the social sphere in the conditions of establishing a market economy in Ukraine by assessing institutional changes and comparing them with the best practices based on existing indicators; they studied the relationship between institutional quality and economic development, assessed the quality of social institutions through the prism of human capital development, which is currently one of the strategic competitive advantages of the country.
When deciding to buy consumer goods, save financial resources or invest them in various types of assets, people form and make investment decisions related to the choice of priority goals and optimization of financial risks. However, this choice is often limited by the financial capabilities of the individual, his level of awareness or selection of optimal investment instruments. In this context, it is interesting to study the standard of living (living comfort) and working environment of the population. Mishchuk and Grishnova (2015) found significant effects of a number of economic and non-economic factors on migration. And this, in turn, affects the ability of individuals to save, acquire real or financial assets (invest), and participate in the formation of financial funds of the state (through tax payments).
The previously referenced theoretical works bear an important practical meaning for realizing the principles of investment made by natural persons and creation of personal investment portfolio since in most financial decisions an individual takes risks and cannot be assured of the results of his/her activity because of being influenced by the postulates of the above-mentioned theories.
In economics, savings are closely connected to investments, in other words, savings are the ground to investment. In this context, personal finance is considered as the primary link of the financial system and as a category that has an important impact on the development of financial processes in the state (Kutsyak, 2012). The source of economic growth in developed countries is the gross national savings -state, business and personal savings of citizens. But in conditions of economic instability and cyclical fluctuations in the economy, it is the financial assets of the population that are the main potential for social development. The finance of individuals is the primary element of the financial system, which reflects and actively influences the economic processes in the country. They mediate personal consumption and investment in the public and corporate sectors of the economy, thus creating conditions for their development. Thus, the result of the formation of personal finances is the accumulation of resources in the funds of personal consumption and savings and their further investment.
The relationship between investment and savings is significantly influenced by factors such as the type of market competition, favorable market conditions, the development of stock and financial markets, and the presence of deferred demand or future expectations of the population.
Sidorchuk (2016) investigates the impact of micro-and macroeconomic preconditions on the formation of personal investment portfolio and investment decision-making. Thus, microeconomic preconditions are due to the ratio of different forms of household wealth during its life cycle and the loss of value for money over time if not used. Macroeconomic preconditions are based on the approach according to which the amount of savings in the country's economy is equal to the investment of all subjects of the economic system, while personal savings become a source of investment for economic entities. Such an investment-saving mechanism has a multiplier effect on the long-term development of the country's economy and is considered as the main method of increasing overall welfare.
The savings generated by the population can be attracted as investment in the domestic market and become a substitute for borrowing by the state in foreign markets. Many researchers are inclined to believe that such savings are a reserve resource of the national economy, which only slightly enters the country's financial market and takes the form of an investment resource. Obstacles in this way are low financial literacy, distrust of the coun-try's financial system, high poverty, underdeveloped stock market, difficulty opening an account and buying securities, a limited range of financial instruments.
In this context, research is on the investment of individuals as an additional resource for economic growth of regions and the economy as a whole. In particular, Chechetova and Chechetova-Terashvili (2019) consider personal finance as a reserve for the economic growth of regions. According to them, the investment function of cash flows of the population is of great importance for the region's economic growth. Activation of investment processes by attracting such funds is possible subject to the improvement of financial instruments and the formation of incentives from state and regional authorities. To competently and effectively carry out investment activities, it is necessary to know about the possible tools for investing free funds, their features, main advantages and disadvantages. At the same time, in order not to make a mistake when choosing an investment instrument, you need to take into account not only your wishes, but also environmental factors.
There is a close relationship between personal savings and investment. The transformation of savings into investment will provide additional income, which will lead to increased consumer demand and improve the quality of life. Growing consumer needs will encourage businesses to increase production, which will lead to increased payments to the budget, in addition, tax liabilities of individuals will increase due to the increase in the object of taxation, which will also increase the revenue side of the budget (Nosova, 2015).
An interesting approach is to study the investment capacity of the population through the processes of de-shadowing of the economy and the development of the system of non-banking financial institutions (Lavryk, 2018). The investment opportunities of the population are developing in close connection with the national investment trends, and their growth is constrained by many problems. But positive changes have been made towards reformatting the population into an active strategic investor, which, in turn, will lead to a reduction in the shadow economy. Among the main areas in which investment activities can be intensified, raising the general standard of living of the population, raising real wages, and creating additional jobs were recognized that. Particular attention is paid to non-bank financial institutions, which have a strong potential to return shadow capital to the economy, provided that they are trusted and ensure the reliability of their work.
It is summarized that in developed countries, the main areas of investment for individuals are deposits in corporate securities and insurance funds, and in Ukraine -deposit accounts of commercial banks or lending to housing construction (Lavryk, 2018).
GENERALIZATION OF THE MAIN STATEMENTS
The dynamics of investment sentiments of the population reflects the economic, political and social situation of the country, as this sector of the economic system is most sensitive to the development of favorable or, conversely, negative phenomena in the economy and society (Figure 1).
Analyzing the dynamics of people's asset increment, one can notice that the most dramatic decline was in 2014 and in 2019, which was a period of political and economic crisis. The decline in the financial and real sectors of investment happened simultaneously. A greater correlation with the political situation was represented by financial assets, which during the crisis decreased from UAH 110,888 million to UAH 27,868 million, which is almost 4 times, and in 2019 decreased from 35,805 to UAH -95,017 million, while non-financial assets gradually declined. The decrease in the growth of financial assets is due to inflation in the economy, the devaluation of the national currency and the crisis in the banking system of Ukraine, as commercial banking services are most popular with the population and other financial market instruments are less popular due to low financial literacy. This situation allows formulatimg a hypothesis that investments in the financial economy are more dependent on the social and political situation in a country than investments in non-financial assets.
According to the dynamics of people's assets increment in the period of stability, there is an inversely proportional dependence between correlating the financial and non-financial instruments of investment.
The year of 2017 clearly shows a noticeable increase in financial investments and a decline in non-financial operations, while the next year illustrates the inverse correlation of these two indicators. Thus, a conclusion can be drawn that the population in a crisis period prefers cash in national or foreign currency.
Quite the opposite is the situation during periods of the economic growth and recovery of business activity. Some percent of the country's population wishing to prevent savings depreciation and profit from accumulated financial resources, invests in different instruments according to personal convictions and the type of sector that develops on the country's financial market. Under these conditions, individual investment activity includes investment of temporarily free funds in a great amount of investment objects generating a certain type of their diversified combination that enables evening out the decline in profitability of some assets by the increment in profitability of other ones. Such investment system is traditionally called an investment portfolio. Figure 2 shows the structure of investment instruments most used by Ukrainians.
The maximum percentage of gross capital formation by households goes to deposits and cash (49%); this indicates that in general the population finds it easier to invest savings in cash (mostly in foreign currency and short-term deposits). Investment in fixed capital (31%) is popular due to high reliability and greater profitability compared to deposit accounts.
Due to a number of circumstances, other financial instruments are not fully used by Ukrainian citizens for investment. According to statistics, financial instruments are used by less than 20% of respondents, while in case of availability of excessive funds, 49% of respondents prefer saving money in cash at home. Such situation clearly demonstrates the problematic aspects in this field. Domestic citizens do not have a financial opportunity or sufficient information for investment (that option was chosen by 76% of respondents), while 14% of respondents mentioned the lack of trust in financial institutions as a main obstacle to invest.
Thus, in market conditions, individuals, when saving their money, have a variety of financial instruments. The formation of organized savings is a clear indicator of a stable economy and confidence in the country's financial system, because the accumulation of cash from the population does not allow attracting this resource into the financial system. When deciding on the forms and methods of investment, an individual is influenced by many both internal and external factors (micro-and macroeconomic preconditions). This makes it impossible to form a universal solution for all individual investors. However, trends and the optimal structure of a personal investment portfolio may have common features for investors, which are ranked by income level, age, national traditions, etc.
In this regard, the results reveal the fragmentation of investment behavior of the population according to people's revenue level. In the "poor" countries, people turn to instruments that are more conservative; for example, bank deposits that reach the mark of 63%, while for the "rich" countries the same instrument constitutes only 28% (2019). Furthermore, a great share of investment in "rich" countries is aimed at securities, insurance and pension programmes (Figure 3).
In general, the world population tends to invest their free funds in securities (35%), bank depos-its (33%) and insurance properties (30%). Other means of investment are represented by only 2%.
One of the reasons for low investment activity in the non-banking sector of the economy in Ukraine and some other European countries is the fact that people do not invest their savings owing to non-confidence in financial institutions, financial illiteracy or undeveloped financial market of these countries. That is why the study of "hidden" savings can be a marketing possibility to economic growth.
Apart from classical financial investments, in "poor" countries, a purchase of house by natural persons can be considered investment expenditure as they are operated largely by savings and can form added value. According to experts, investing in real estate (construction of housing, office space) was and remains the most attractive area for temporarily free funds of private investors.
For example, in the context of unstable national currency and economic and political crisis in Ukraine, estate property is one of the most attractive objects to invest for private investors. In terms of risk, investment in real estate is low-risk, because in the case of economic growth in the country, housing prices are gradually rising, and real estate can always be sold or rented. Thus, the same as quotation of shares and bonds, the value of estate property can decrease under some conditions. However, it can never reach zero, and in real life it means that an owner of the estate property will never be left with some depreciating bonds on hand.
The extent to which the commercial real estate market is developed in a certain country can be judged by the number and activity of investment funds as they are main owners of functioning objects of commercial estate property that bring regular income.
At the same time, real estate market is characterized by the lack of convenient investment instruments. A high input threshold of this market makes it almost inaccessible to direct private investment. In this situation, the presence of real estate investment funds is quite topical. This new instrument of indirect investment in estate property is transparent, efficient and attractive in terms of taxation and liquidity, so it provides an essential increase in investment flows from natural persons into a real estate market.
It is worth mentioning that besides financial factors, people's lack of knowledge in savings man-agement has a considerable influence on investment disproportions. Moreover, the population hardly ever turn to investment facilities, i.e. institutions, which, judging from their functioning, could significantly simplify the process of efficient savings management for average citizens.
According to the data on global competitiveness provided by the World Economic Forum in 2018, Ukraine lost its position and held 81 place among 140 countries. Defining competitiveness as a set of institutions, policies and factors, a country's rating is determined on the basis of a combination of publicly available and own specialized data on 12 dimensions -components of competitiveness, which together are a comprehensive assessment of economic competitiveness. The underdevelopment of financial markets is the second reason after macroeconomic instability, which causes such low ratings of the country. The main regression, according to the WEF report, was recorded in the financial system, in which Ukraine's rating is 136th. That is why the lack of a proper institutional environment limits people's choice to invest.
In case of people forming some savings, especially in the context of instable market conditions and globalization, it is worth considering non-financial investments into "human capital assets". This is the total cost of investment in a person, which is aimed at the formation, accumulation, implementation of its knowledge, abilities, skills and abilities, which provides value growth and income. The peculiarity of investing in human capital is that they depend on human life cycles and are subject to rational and irrational influences on the economic behavior of the individual, family, firm, state as institutional investors in human capital. Research on the role of human capital in shaping the financial potential of households showed that one of the main and priority tasks of the concept of human capital development management in Ukraine should be the recognition of its great social value and driving force of economic growth. This will allow forming a high level of well-being of the population and will positively affect their financial potential. The level of reproduction of national human capital depends on the investment choice (Kizima, 2009;Nosyk, 2015).
Based on this, the study analyzes investments that can improve people's financial condition in future and bring them extra benefit: 1) investment in health (It is worth allocating savings to health improvement, for example, disease prevention or purchasing a gym membership. It is important because not only is it a high priority for a living being, but also, in case of health state deprivation, unexpected expenditures can considerably aggravate a person's financial situation); 2) investment in education (Acquisition of new knowledge requires both time and money. Education expenditures are justifying only if obtained knowledge can be applied in real life). Thus, for many people under 30, it is more beneficial to invest in oneself rather than in financial instruments; 3) innovations (investment in technologies is investment of financial resources in decisions that allow one to shorten daily expenditure on life of both one person and a whole family); 4) investment in assets (i.e. purchasing certain objects, which can benefit, including monetarily, in time.
Creating an individual investment portfolio, one needs to develop one's own strategy of actions and specify the channels for investment. Firstly, it is necessary to focus on investors' activities on a financial market, in other words investing in financial resources, which are riskier and require a higher level of financial knowledge, but can still generate greater revenue.
A generally accepted distribution of investor types and main strategies of their behavior is as follows: conservative, moderate and aggressive. It is believed that along with an increase in the investor experience level, there is a transition from one type to another, and therefore a change in investment strategy. Nevertheless, such possibility tends to appear with an increased level of person's financial literacy, his/her expertise and past positive investment experience.
However, for individuals involved in investment decision-making, this study shows that the person's level of revenue, acceptable risk and the purpose of the investment are decisive. Table 1 represents an average return on investments of separate financial instruments, which can be seen as determinant in creating a certain type of investment portfolio. Furthermore, when choosing the most acceptable source of investment, people have to consider other criteria, amongst which the most important are their liquidity, stability, riskiness and fiscal characteristics.
Financial instrument Return
Based on Table 1 and the above-mentioned criteria, it is possible to design a typical investment portfolio of an individual with low, mean and high income.
The data sown in Figure 4 confirm that low-income people invest in the least risky instruments such as bank deposits and government securities, which provide a relatively high return for people with this level of income. Less efficient and accessible for such people, investment in foreign currency, shares, investment facilities, gold and real estate are considered. For a person with mean income, it is more beneficial to invest in bank deposits. Such people have an opportunity to invest their funds in shares almost doubled. They are prone to invest in state securities, since they are more reliable; however, their profitability is considerably less that of shares and deposits. Less attractive forms of investment are foreign currency, real estate, gold (however, it is still more attractive than for the poor), and investment facilities. Highincome people generally invest in shares, which accounts for almost a third of funds in bank deposits, which is approximately 22%, in gold constituting 19% and in real estate being 10%. Investments in investment facilities, foreign currency and state securities for people with such income are 3-7%.
DISCUSSION
With free finance at their disposal, almost everyone can choose the most useful investment option, but it may seem that investment portfolio man-agement is an extremely difficult and incomprehensible task. Most investors falil to achieve their goals, because their actions are not consistent. To regularly get rich on investing, a potential investor must master the principles of sound management of personal finances and use them in practice.
The analysis of scientific papers on this issue allows systematizing the basic principles associated with the effective management of personal finances. When planning to invest, it is necessary to study their alternatives, and when making investment decisions, it is necessary to deeply understand the features of the market segment chosen for investment. It is necessary to analyze the relationship between the risk and return in detail − the higher the probable return, the greater the risk. It is necessary to diversify investments, try not to invest all your savings in one, even the most reliable, financial instrument. Build your own investment strategy to form a contingency fund in case of unforeseen circumstances, and never invest in "doubtful" instruments (Kizima, 2010).
At the same time, the specialists of the investment company LBLV (2020) recommend the following rules when forming a personal investment portfolio and making investments, in order to achieve Shares success and minimize current risks: invest only free finance, clearly define your goals and intentions, determine the sequence of stages investment portfolio, reinvest earnings, diversify investments.
Given the above approaches, the empirical data on the formation of savings of specific categories of citizens, as well as scientific publications on this issue, an individual, when forming a personal investment portfolio, must take into account and adhere to the following criteria: • Collect information about the investment object. It is unwise to invest money somewhere under the influence of emotions. If you want to invest, it is necessary to learn about the basic principles and functioning of a financial instrument, consult with experts, review ratings, learn basic information about the activities of the company or organization with which you plan to cooperate.
• Diversify an investment portfolio. In the process of creating an investment portfolio, it is necessary to determine the interconnection between risk and profitability. Any attempt to gain income presupposes risk. The higher the potential revenue, the higher the risk. A fear of risks is an emotional reaction to perceiving the chances to gain revenue and suffer financial losses. Risk decline can be provided by collecting information and lowering a degree of the situation uncertainty, by accumulating investment experience, by diversifying investments in different instruments: deposits, unit investment trusts, real estate, securities, which are the most reasonably to be placed on the markets of different countries. It is reasonable to divide investment resources into two levels of investment: the first level -distribution of investment into different instruments, for example, 25% goes to securities, 25 % to bank deposits, 40% to real estate, and 10% to reserve fund; the second level -diversification in frames of chosen instruments: for example, a half out of 25% for securities is invested directly into shares and another half goes to unit investment trusts.
• Form financial reserves. Investments must not concern the primary needs of a citizen or his/ her family, that is why it is necessary to form reserves for at least two months in advance in order to provide a proper amount of funds to pay for the primary needs of a person or a family.
• Minimize the use of borrowed resources. Personal investment is a method of maintaining one's own financial condition in the first place, and only in the second place its task is to multiply one's own wealth. Therefore, the use of speculative methods of enrichment can often lead to unpredictable loss, especially in the context of current Ukrainian realities, like a lack of real planned macroeconomic indicators, exchange rate movements or a change in creditor behavior strategy, it can change the correlation between received and paid interests towards not debtor's benefit.
• Develope your own investment strategy. Market conditions of management require markets specificity, their development through innovation and increasing efficient functioning; therefore, they are quite variable. Under such conditions, it is necessary to constantly learn something new, gain experience in financial decision-making and set one's own priorities that correspond to one's own financial abilities, individual desires and the level of one's reasons satisfaction as a result of such activity. According to personal needs and reasons, a person can determine a strategy of saving money, an opportunity to accumulate funds at low risks or readiness to take risks for potential high revenue.
• The correct choice of the location of the investment object. In the context of globalization, a correct choice of investment object location determines its profitability and stability. Portfolio investments with heterogeneous asset classes and in different countries is complicated and difficult to manage business. Purchasing objects located in different jurisdictions means that one will need to gain an understanding of the peculiarities of local taxation systems and find different contractors to manage the objects. It is reasonable to invest funds in not more than two different assets classes in one or two countries. For a purchase of profitable objects, it is worth choosing the countries with a reliable political system and strong economy. For example, Austria, Great Britain, Germany, the USA and France can be referred to as such countries.
Given the specifics of personal portfolio investment, it should be noted that, in addition to the amount of income, the structure of the investment portfolio, according to investment strategies, is significantly influenced by financial goals, which are often determined by the investor's age. A young person is more at risk, so the share of stocks in its investment portfolio, even with a low income, will be higher than that of people with the same income, but older. This is due to the "time reserve" of the younger generation to compensate for possible losses and much less social obligations compared to the older generation. If a person is in pre-retirement age, the investment portfolio should consist more of investments with a fixed rate of return, because older people need a stable income to ensure a decent standard of living in old age.
Of course, it is undeniable that a significant part of the population, due to limited sources of income, low financial literacy, and limited and difficult access to financial instruments, are not able to implement their own investment strategies. However, globalization processes are changing this. Research on activity and quality of management of the population by the personal finances, revealing of objective laws in character and intensity of investment by the population is especially actual in an unstable financial and economic environment in the country, since it allows using the received data for development of measures to stimulate financial activity of the population, as well as in the event of a crisis − to develop measures to protect the population's financial savings.
CONCLUSION
As a result of the study, specific investment instruments have been proposed that can be used in modern conditions to ensure the optimal ratio of return and risk of an investment portfolio. The formation of a personal investment portfolio is a complex process that includes not only financial aspects, but also psychological motives, socio-cultural status of the person. So, financial decision-making is exposed to a number of risks that should be considered and minimized in the process of shaping an investment portfolio through its diversification, formation of reserves, localization of investment instruments, minimization of credit resources and self-development, as well as training and practical investment experience.
Thus, a conclusion can be drawn that the population, when depositing savings, has a fairly wide choice of channelling investment. However, this choice is often limited to financial possibilities of an individual, his/her level of awareness (financial literacy) or availability of instruments for individuals in different countries. Consequently, the study of factors influencing investment choice and the selection of optimal instruments, which enable to use person's available resources in the most efficient way, appear to be important.
Thus, compliance with the above recommendations will lead to economic activity of the population and an increase in their real incomes, reduce the differentiation of the population by income level, and increase the economic culture and welfare of Ukrainian society as a whole. In addition, there are many unresolved issues in the formation and management of personal finances. A significant part of citizens found themselves in a situation of poverty and survival, and only the wealthiest part of the population became a full-fledged subject of the financial market. In this regard, further research should focus on finding ways to increase income and, ultimately, on the formation of the optimal structure of an individual's investment portfolio, which will have a multiplier effect on the financial system as a whole. | 2021-05-30T15:41:17.772Z | 2021-05-06T00:00:00.000 | {
"year": 2021,
"sha1": "1ebb145a8d769fb2f27e4d397877f51653b052ce",
"oa_license": "CCBY",
"oa_url": "https://www.businessperspectives.org/images/pdf/applications/publishing/templates/article/assets/14985/IMFI_2021_02_Gavatiuk.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "1ebb145a8d769fb2f27e4d397877f51653b052ce",
"s2fieldsofstudy": [
"Business"
],
"extfieldsofstudy": [
"Economics"
]
} |
14712028 | pes2o/s2orc | v3-fos-license | Evolution of primordial black holes in Jordan-Brans-Dicke cosmology
We consider the evolution of primordial black holes in a generalyzed Jordan-Brans-Dicke cosmological model where both the Brans-Dicke scalar field and its coupling to gravity are dynamical functions determined from the evolution equations. The evaporation rate for the black holes changes compared to that in standard cosmology. We show that accretion of radiation can proceed effectively in the radiation dominated era. The black hole lifetime shortens for low initial mass, but increases for high initial mass, and is thus considerably modified compared to the case of standard cosmology. We derive a cut-off value for the initial black hole mass, below which primordial black holes evaporate out in the radiation dominated era, and above which they survive beyond the present era.
INTRODUCTION
The Jordan-Brans-Dicke (JBD) (Brans & Dicke 1961) theory is one of the earliest and most well-motivated alternatives to the Einstein theory of gravitation. The value of the gravitational "constant" is set by the inverse of a time-dependent classical scalar field with a coupling parameter ω. General relativity is recovered in the limit of ω → ∞. Solar system observations impose lower bounds on ω (Bertotti et al. 2003;Will 2006). Generalized JBD models are obtained in the low energy limit of higher dimensional theories. String theoretic (Gasperini & Veneziano 2007) and Kaluza-Klein (Appelquist et al. 1987) models after compactification of the extra dimensions yield several variants of JBD models (or general scalar-tensor models) in which the the scalar field coupling ω may become dynamical, and also models with potential terms for the JBD scalar field.
JBD models have been used for tackling several cosmological problems pertaining to different eras of evolution of the Friedman-Robertson-Walker (FRW) universe. The scenario of extended inflation (Mathiazhagan & Johri 1984;La & Steinhardt 1989) was proposed within the context of JBD cosmology. Generalized JBD models resulting from the compactification of higher dimensional actions (Majumdar & Sethi 1992;Majumdar et al. 1993;Majumdar 1997) enable efficient transition to the post-inflationary radiation dominated phase (or reheating). The JBD scalar ⋆ Email: archan@bose.res.in field has also been incorporated in quintessence scenarios for obtaining the present acceleration of the universe (Bertolami & Martins 2000;Banerjee & Pavon 2001;Sen & Seshadri 2003). A more general class of JBD theory where the coupling parameter ω is an arbitrary function of the scalar field (Bergmann 1968;Wagoner 1970;Bergmann 1970) has interesting consequences on cosmic evolution (Sahoo & Singh 2002, 2003 and has also been applied in the context of obtaining singularity free cosmology (Kalyana Rama 1997) and relic gravitational waves (Barrow et al. 1993;Singh & Sahoo 2004).
Primordial black holes (PBHs) are potentially fascinating cosmo-archeological tools (Carr 2003). It is possible for PBHs to impact through their evaporation products diverse cosmological processes such as baryogenesis and nucleosynthesis (Carr 1976;Zeldovich et al. 1977;Miyama & Sato 1978;Chechetkin et al. 1982;Barrow et al. 1991;Majumdar et al. 1995;Liddle and Green 1998), the cosmic microwave background radiation (Chapline 1975;MacGibbon & Carr 1991), and the growth of perturbations as well (Afshordi et al. 2003). PBHs could act as seeds for structure formation (Mack et al. 2006) and could also form a significant component of dark matter through efficient early accretion in braneworld scenarios (Majumdar 2003;Guedens et al. 2002;Majumdar and Mukherjee 2005). Mechanisms for growth of supermassive black holes by PBHs accreting dark energy have been proposed (Bean & Magueijo 2002), though recent results by Harada & Carr (2005a) and Harada et al. (2006) indicate the lack of self-similar growth of black holes by accreting quintessence.
The feasibility of black hole solutions in JBD theory was first discussed by Hawking (1972). The coexistence of black holes with a long range scalar field in cosmology has interesting consequences such as the possibility of energy exchange between the black holes and the scalar field (Zloshchastiev 2005). The black holes themselves could be used to constrain the variation of fundamental constants (MacGibbon 2007). Another interesting issue of gravitational memory of black holes in JBD theory has also been studied (Harada et al. 2002). The evolution of PBHs in JBD cosmology has however remained unexplored in the literature. Since JBD models entail modification of the standard cosmological evolution, it is expected that the evolution of PBHs in JBD models and their associated consequences on various cosmological processes could be substantially modified compared to the case of standard cosmology.
The purpose of this work is to perform a preliminary study of the evolution of PBHs in JBD cosmology. We consider a generalized scalar tensor theory (Bergmann 1968;Wagoner 1970;Bergmann 1970) where the coupling parameter ω(φ) is a function of the scalar field φ, and is thus also time-dependent. We use a set of power-law solutions for the scale factor, the JBD field and the coupling ω obtained in the radiation era (Sahoo & Singh 2002, 2003, and investigate the net energy flux for the PBHs resulting from the processes of Hawking evaporation and the acretion of radiation. We show that accretion of radiation can be effective in the radiation dominated era since the growth of PBHs is always subdominant to the growth of the cosmological horizon in this JBD model. In the matter dominated era, the dynamics of the PBHs is solely governed by the evaporation process whose rate is determined by another set of powerlaw solutions of this JBD model in the matter dominated era (Sahoo & Singh 2002, 2003. The lifetime for PBHs are obtained in terms of their initial masses. Our results outline a viable scenario for the growth and survival of PBHs as constituents of dark matter in JBD cosmology.
GENERALIZED BRANS-DICKE MODEL
We begin with a generalized JBD action given by where the strength of the running gravitational coupling is given by the inverse of φ, i.e., at any time t, G = 1/φ(t), and Smatter corresponds to the action of the relativistic fluid of particles (p = ρ/3) in the flat (k = 0) FRW early universe with scale factor a. The Friedman equation and the equation of motion for the JBD field φ obtained from the above action are given respectively bẏ with the energy conservation equation given byρ For a radiation dominated evolution, one has ρ ≈ ρR ∝ a −4 , and p = ρ/3. Assuming power law solutions for the scale factor a(t), the JBD field φ(t), and also the coupling parameter ω(t) one can obtain (Sahoo & Singh 2002, 2003 the following time-dependences of these quantities: where the subscript i indicates the initial values of the variables. In order to minimize the departure of the evolution of the scale factor from the case of the radiation dominated evolution of standard cosmology (i.e., a(t) ∝ t 1/2 ) we choose |ωi| ≪ 1. Without loss of generality, we set φi = ξM 2 pl , where M pl = G −1/2 is the Planck mass (corresponding to the present value of G). Note that the solution for the JBD field φ indicates the strengthening of gravity with the increase of cosmic time. Such a behaviour for φ arises not just from the specific model (1) used here, but is a rather generic feature of several generalized JBD or scalar tensor models with a potential for the scalar field which rolls down the slope leading to the increase of G with time (Majumdar & Sethi 1992;Majumdar et al. 1993;Majumdar 1997).
EVOLUTION OF PRIMORDIAL BLACK HOLES
Primordial black holes could form due to various mechanisms (Carr 1985) some of which have been studied in detail in the literature (Zeldovich & Novikov 1966;Hawking 1971;Carr & Hawking 1974;Hawking et al. 1982;Kodama et al. 1982). The gravitational collapse of perturbations within the cosmological horizon could lead to sub-horizon sized black hole formation after Jeans crossing. Such a mechanism may not be very effective however, in JBD cosmology where the Jeans length itself could be significantly increased due to the weakening of gravity in certain models by a large value of the JBD field in the early universe. Nevertheless, a strong (first order) inflationary phase transition in extended inflation models driven by the JBD field (La & Steinhardt 1989;Majumdar & Sethi 1992;Majumdar et al. 1993;Majumdar 1997) naturally leads to unnucleated and trapped false vacuum regions and topological defects such as domain walls and wormholes that could easily collapse into black holes, and could also create super-horizon density perturbations. The formation of super-horizon scale PBHs in the expanding FRW background has been studied recently (Harada & Carr 2005b). In the present analysis we will not take recourse to any particular formation mechanism for PBHs, but rather study their cosmological evolution in JBD theory, assuming that there exist PBHs in such scenarios. We now consider the evolution of PBHs in the cosmological background governed by the above solutions (5), (6) and (7). We assume that the PBH density is low enough to ensure radiation domination. For a PBH immersed in the radiation field, the accretion of radiation leads to the increase of its mass with the rate given bẏ where rBH = 2M/φ is the black hole radius, and f is the accretion efficiency. Using the solution for φ given by Eq. (6) with the assumption of |ωi| ≪ 1, and using ρR = [(3M 2 pl )/(32πt 2 )], one obtainṡ with Mi being the initial mass of the PBH at time ti. Since the logarithmic growth rate for a PBH given by Eq. (10) is subdominant to the linear growth of the horizon mass MH ∼ t (since a ∼ t 1/2 ), once a PBH is formed, it is indeed possible for it to grow in size by accreting the radiation energy within its cosmological horizon. For a complete picture of PBH evolution, one needs also to consider the Hawking evaporation process, whose rate is given bẏ where T = φ/(8πM ) is the Hawking temperature, σ the Stefan-Boltzmann constant and g is the effective number of degrees of freedom of the particles emitted by the black hole. The solution for φ given by Eq.(6) leads tȯ where A = (gσξ 2 M 4 pl ti)/(256π 3 ). The complete evolution for the PBH is thus described by the combination of Eqs. (9) and (12): It is apparent from Eq.(13) that for PBHs with initial mass Mi << (A/B) 1/4 , the rate of evaporation exceeds that of accretion. For such a case, accretion soon becomes negligible and black holes lose energy at a rate given effectively by Eq.(12). Note that though the rate of evaporation decreases with time, it is still higher than the corresponding rate in standard cosmology. This is because the Hawking temperature T = φ/(8πM ) is larger for JBD PBHs for large φ.
Hence, a PBH with initial mass Mi < (A/B) 1/4 evaporates out much quicker, with a lifetime tevap given by However, PBHs with initial mass Mi > (A/B) 1/4 experience monotonic growth with accretion dominating over evaporation throughout the period of validity of Eq.(13), i.e., throughout the radiation dominated era. Eq.(13) can be integrated exactly and leads to the following mass-time relationship for the PBHs: with C = A/B. Accretion of radiation can proceed effectively till the universe stays radiation dominated, i.e., up to the era of matter radiation equality teq. This result is qualitatively different from the widely accepted picture in the standard cosmological evolution where accretion of radiation in the radiation dominated era seems to be ineffective (Carr 2003). The domination of accretion over evaporation is observed also in other modified gravity theories, such as in the braneworld scenario (Majumdar 2003;Guedens et al. 2002;Majumdar and Mukherjee 2005). The maximum mass achieved by a PBH of initial mass Mi is given by In the matter dominated era, ρM ∼ a −3 and p = 0. A set of solutions for the JBD cosmological equations (2), (3) and (4) is given by (Sahoo & Singh 2002, 2003 Matching these solutions with those obtained in the radiation dominated era given by Eqs.(5),(6) and (7), one has φ(teq) = ξM 2 pl (ti/teq) 1/2 and ω(teq) = ωi(teq/ti) 1/2 . Substituting the solution for the JBD field φ in Eq.(11) leads to the following evaporation law for PBHs in the matter dominated era: eq . The mass of the PBHs evolve as in the matter dominated era. The lifetime for the PBHs tevap is given by which upon (using Eq.(16)) reduces to Since the availability of background radiation diminishes substantially in the matter dominated era, one can safely neglect any accretion term corresponding to the accretion of radiation by the PBHs. However, other forms of accretion such as of the energy of a quintessence field could, in principle, be effective (Bean & Magueijo 2002), that we do not consider in the present analysis. The rate of evaporation is again faster, as in the radiation dominated era, than in the case of standard cosmology (where tevap ∼ M 3 max ). However, the rate decreases with time as is evident from the r.h.s of Eq.(20) whose counterpart in the standard theory is independent of time.
The overall lifetime of a JBD PBH has an interesting comparison to a PBH lifetime in standard cosmology. A JBD PBH with initial mass Mi < (A/B) 1/4 , evaporates out quicker than a PBH in standard cosmology, being unable to accrete in the radiation dominated era. However, when the initial mass exceeds (A/B) 1/4 , accretion proceeds effectively and dominates over evaporation in the radiation dominated era increasing the PBH mass at the time of matter-radiation equality. In the matter dominated era, though the evaporation rate of JBD PBH is faster, it starts to evaporate out much later than in standard cosmology (without accretion). Thus, the lifetime for a PBH with Mi > (A/B) 1/4 is enhanced in the JBD scenario compared to the case of standard cosmology. Note that in the present analysis we have not considered any possible back reaction of the PBHs on the local background value of the JBD field resulting from the local change of energy density ρ due to the PBHs. The form of the JBD field equation (3) is such that the ρ-dependence drops out in the radiation dominated era. Further, in the late matter-dominated era the value of ω is large enough (in order to satisfy present observational bounds) such that the ρ-dependent term again becomes negligible in Eq.(3). So any back reaction could be effective only in the early matter dominated epoch during which a local rise in energy density could result in a slower rate of decrease for the φ field near the PBHs than what is obtained in Eq.(18). The resultant change in the PBH evaporation law (20) could introduce a minor correction to the PBH lifetime (23) that we have neglected here.
It would be interesting to observe particular examples of evolving PBHs in JBD cosmology. Note that the accretion and evaporation rates depend on the initial value φ of the JBD field through the quantities A and B defined earlier.
Here we fix the value of φi or ξ by setting the present value of φ(tnow) = M 2 pl (corresponding to the present strength of gravity) and evolving φ given by Eqs. (18) and (6) corresponding to the matter and radiation dominated eras, respectively, backwards in time. We thus obtain the value of φ(tEW ) = 10 18 M 2 pl , where tEW denotes the electroweak era since which the JBD evolution equations (3) are assumed to be valid. From Eq.(14) it follows that for a PBH with Mi < (A/B) 1/4 to survive the radiation dominated era, Mi 10 13 M pl . A PBH which forms at time tEW with an initial mass of the order of the cosmological horizon mass at tEW satisfies the condition Mi > (A/B) 1/4 , and accretes radiation. Though the actual growth of mass turns out to be negligible for such a PBH, accretion does play an important role enabling it to survive for much longer (tevap > tnow).
Another interesting consequence of Eq.(23) is that for a PBH to evaporate during the present era, i.e., tevap = tnow one needs to have Mmax ≈ 10 10 M pl . Since accretion cannot cause significant growth, one requires Mi ∼ Mmax ≈ 10 10 M pl . However, a PBH with such a small initial mass cannot even survive the radiation dominated era. Only those PBHs which have Mi > 10 13 M pl enter the matter dominated era and these also survive up to the present era, i.e., tevap >> tnow. Thus there are no end products of black hole evaporation to contend with in JBD cosmology during the matter dominated era and beyond up to present times. All PBHs with initial mass Mi < 10 13 M pl evaporate out in the radiation dominated era itself. On the other hand, PBHs with larger mass can survive as intermediate mass astrophysical black holes contributing to the dark matter density.
CONCLUSIONS
To summarize, in this letter we have studied the evolution of PBHs in JBD cosmology and obtained a cut-off value for the initial mass which decides whether the PBHs could survive up to today. An interesting consequence of our analysis is that no PBHs evaporate during the matter dominated era, and hence there is no distortion of the cosmic microwave background radiation spectrum due to evaporating PBHs. Though our results are in the context of a particular JBD model, the sort of evolution obtained for the JBD field φ is generic to other JBD models as well, in particular those following from higher dimensional theories (Majumdar & Sethi 1992;Majumdar et al. 1993;Majumdar 1997). In order to satisfy solar system constraints (Will 2006) such as the one recently obtained through the frequency shift of radio photons to and from the Cassini spacecraft as they passed near the sun (Bertotti et al. 2003), one requires |ω(tnow)| > 10 4 , which using Eqs. (19) and (7), leads to ωi ∼ 10 −14 , thus strongly validating our approximation of |ωi| ≪ 1 used to derive the black hole evolution equation in the radiation dominated era. With the possibility of future observations of actually finding or ruling out JBD models by determining ω (Acquaviva & Verde 2007), it would indeed be exciting to re-work the standard observational constraints (Carr 2003;Liddle and Green 1998;Carr 1985) on the density of PBHs at several cosmological eras in the JBD scenario. In particular, the faster evaporation of smaller PBHs in the radiation dominated era could impose tighter constraints on the initial mass spectrum from nucleosynthesis bounds (Liddle and Green 1998). Investigation of these issues calls for a more comprehensive analysis by considering a population of black holes in the early universe evolving according to the JBD dynamics. | 2008-01-14T07:17:00.000Z | 2007-09-20T00:00:00.000 | {
"year": 2008,
"sha1": "b9a3d3aa4dc84b1016ace79026538f21f8772f89",
"oa_license": null,
"oa_url": "https://academic.oup.com/mnras/article-pdf/385/3/1467/3294646/mnras0385-1467.pdf",
"oa_status": "BRONZE",
"pdf_src": "Arxiv",
"pdf_hash": "b9a3d3aa4dc84b1016ace79026538f21f8772f89",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
266360493 | pes2o/s2orc | v3-fos-license | Use of bee venom in preventive medicine: An experimental hepatic encephalopathy study in rats
Abstract Objectives Bee venom is used for medicinal purposes, including the treatment of neurological and liver diseases, but its use as a primary health care approach for preventive purposes requires further exploration. The aim of this study was to provide the first investigation into the possible protective effects of bee venom against hepatic encephalopathy, a serious neurodegenerative disease. Materials and Methods An experimental animal study was conducted in which healthy albino Sprague–Dawley rats were randomized into three groups: healthy, control and bee venom groups. All rats were tested for locomotor activity at the beginning and end of the study. No intervention was made in the healthy group, whereas hepatic encephalopathy was induced in the control and bee venom groups by the administration of thioacetamide (TAA) (200 mg/kg/day). The bee venom group also received bee venom (5 mg/kg/day) subcutaneously every day for 14 days prior to the TAA administration. Results The results for the final locomotor activity tests were statistically better in the bee venom group than in the control group, supporting a beneficial effect of prophylactic bee venom application. Blood ammonia levels and liver weights, determined as indicators of inflammation, were lower in the bee venom group than in the control group and were close to levels in the healthy group, but not statistically significant. Conclusions Bee venom administration has protective effects against the development of hepatic encephalopathy and offers a promising therapeutic opportunity in preventive medicine.
INTRODUCTION
Hepatic encephalopathy is an important and serious complication of both acute and chronic liver diseases and causes significant morbidity and mortality.It can present a wide spectrum of neuropsychiatric disturbances, ranging from subclinical fluctuations to marked disorientation, cognitive impairment, confusion and coma (Rose et al., 2020).Its pathogenesis is associated with inflammatory processes, such as systemic inflammation, neuroinflammation and endotoxemia, as well as with ammonia accumulation (Luo et al., 2015).The levels of systemic inflammatory mediators and ammonia increase in liver disease, and these compounds are known to cause morphological changes in the endothelial cells and astrocytes that form the blood-brain barrier and ultimately lead to neuroinflammation and brain oedema with microglial activation (Shawcross et al., 2010).Ammonia also causes inflammation by disrupting amino acid metabolism and energy use in the brain (Luo et al., 2015), and therefore, ammonia is considered the main culprit in the development of hepatic encephalopathy, and all available treatments mostly work on it (Fallahzadeh & Rahimi, 2022).Ammonialowering treatments, such as rifaximin, lactulose, probiotics, bacteria and faecal transplants, aim to reduce ammonia production or increase ammonia excretion (Rose et al., 2020).
Beyond these treatments, one can speculate and expect that modalities with anti-inflammatory effects that can reverse the effects caused by ammonia at the cellular level could stop the pathogenesis of hepatic encephalopathy (Manzhalii et al., 2019).Recent investigations have shown that bee venom, a traditional and complementary medicine treatment widely used since ancient times, has anti-inflammatory, antimicrobial, anti-apoptotic and anti-cancer effects (El-Seedi et al., 2020;Gu et al., 2020;Khalil et al., 2021;Lee & Bae, 2016).In addition, its therapeutic efficacy has been demonstrated in several neurological diseases, such as Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis, and pain, as well as in liver fibrosis (Zhang et al., 2018).
The effectiveness of bee venom as a treatment for both neurological and liver diseases suggests that it could serve as an effective preventive and treatment for hepatic encephalopathy.
The main purpose in primary care is to prevent the development of diseases to reduce morbidity, mortality and even health costs.The aim of our study was to conduct the first exploration of the use of bee venom as a preventive medicine application to establish whether bee venom has neuroprotective and preventive effects in a rat model of hepatic encephalopathy.
Animals
In this study, 18 healthy albino Sprague-Dawley rats weighing 250 ± 25 g were obtained from the Medical Experimental Application and Research Centre of our university.The rats were kept in Plexiglas cages during the study, fed ad libitum and maintained in a 14 h night/10 h day cycle.
Bee venom
Bee venom was administered subcutaneously to the bee venom group rats at a dose of 5 mg/kg.The safety of bee venom has been studied in the literature, and doses below 1200 mg/kg have been found to be safe in rats (El-Tedawy et al., 2020).
Bee venom was purchased from a local supplier and transferred to us through a cold-chain system.The bee venom was collected using a glass bee venom collecting plate and a device that sends an electric current to the plate.The electric current causes the bee venom to stick to the glass plate.The plates were placed vertically in a V shape on top of the beehive.Activation of the device incites the bees to sting, and they leave their venom on the glass plate.The bee venom adheres to the plate and remains stable.After about 20 min of operation, the venom-covered glass plates were carefully removed from the hive without contamination.Bee venom in the form of crystallized powder was scraped from the glass and stored in the refrigerator.
Study protocol
The 18 rats were randomized into a healthy group, a control hepatic encephalopathy group and a bee venom group, each containing 6 animals.All rats were tested for locomotor activities at the start of the study.The healthy group underwent no interventions other than weight monitoring.The control group, after 14 days of weight monitoring, was administered thioacetamide (TAA) intraperitoneally at 200 mg/kg/day for 3 consecutive days to induce hepatic encephalopathy.The bee venom group received bee venom at 5 mg/kg/day every day for 14 days, followed by TAA administration as above for 3 days to induce hepatic encephalopathy.On the 18th day, all rats were retested for locomotor activities, and the study was terminated.A flow chart of the study is presented in Figure 1.
At the end of the study, all the rats were euthanized under ketamine/xylazine (50 mg/kg, 10 mg/kg) anaesthesia.Cardiac blood samples were taken for the measurement of serum ammonia levels to determine the level of hepatic encephalopathy.Possible infection and inflammation in the organs were also evaluated by removing, weighing and comparing the livers, brains and lungs of all rats in the three groups.
Statistical analysis
The results were statistically compared with the Statistical Package for the Social Sciences version 23.0 (IBM).All numerical data are presented as means and standard deviations.The Shapiro-Wilk test was used to determine the normality of the data.The Mann-Whitney U test was used for analysis of the three or more groups when the data were normally distributed, and the Kruskal-Wallis test was used for statistical analysis of the three or more groups when the data did not show a normal distribution.A value of p < 0.05 was considered statistically significant.
RESULTS
The average weights of all rats at follow-up every 3 days are shown in Figure 2. All rats in the healthy and control groups showed normal weight gains, whereas the rats in the bee venom group showed a reduced rate of weight gain due to the application of subcutaneous bee venom every day for 14 days.
All rats included in the study were tested for locomotor activities at the beginning and at the end of the study.No significant differences were detected between the groups in the locomotor activity tests performed at the beginning of the study (p > 0.05); however, significant differences were found in the tests performed after the study, and as requested, the control group had lower locomotory activities than the healthy group (Table 1).The locomotory activity in the bee venom group was intermediate between that of the healthy group and the control group.
The average weights of the brains, livers and lungs of the rats from the three groups are given in Table 2.The organ weights increased in the control group, as expected, due to the presence of infection and inflammation.However, the organ weights, and especially liver weight, were lower in the bee venom group than in the control group, although the differences were not statistically significant (p > 0.05).
The ammonia level in the cardiac blood, examined as an indicator of the development of hepatic encephalopathy, is presented in Figure 3.
The ammonia level was found to be highest in the control hepatic encephalopathy group (168.8 ± 37.8 µg/dL) and lowest in the healthy group (119.0 ± 16.9 µg/dL).The blood ammonia level in the bee venom group (131.8 ± 42.2 µg/dL) was lower than in the control group and close to the value in the healthy group, although the difference was not statistically significant (p = 0.084).
F I G U R E 2 Follow-up of the average weights of the rats in the groups measured every 3 days.
F I G U R E 3
Serum ammonia levels of the study groups measured after euthanasia.
DISCUSSION
Many studies in the literature have investigated the effectiveness of bee venom or its components against diseases.These studies have revealed that bee venom has anti-microbial, anti-protozoal, anticancer, anti-inflammatory and anti-arthritic properties (Khalil et al., 2021).However, until now, no study has investigated bee venom or its components as a treatment to prevent hepatic encephalopathy.
Hepatic encephalopathy is a neurological abnormality caused by liver failure or portal-systemic shunt, and the resulting brain dysfunction is triggered regardless of the underlying liver disease cause (Weissenborn, 2019).Although the hepatic encephalopathy that follows liver transplantation is widely considered reversible, many studies have shown that neuroinflammation and neuronal cell death can be serious consequences of hepatic encephalopathy and can lead to irreversible neurological deficits (Rose et al., 2020).This situation reveals the necessity of effective prevention and treatment of hepatic encephalopathy, independent of the triggering cause.
The present study represents the first investigation to explore the preventive efficacy of bee venom against hepatic encephalopathy.
Our use of a rat disease model confirmed that the rats administered bee venom as a preventive measure showed locomotor activity close to that of the healthy group and had blood ammonia levels lower than those in the control group.The liver weight, used as a sign of inflammation, was also similar in the bee venom group and the healthy group.Taken together, our data indicated that bee venom application has a preventive effect against the onset of hepatic encephalopathy.
Very few studies in the literature have investigated the preventive efficacy of medications against hepatic encephalopathy.One systematic review of studies on the protective activities of L-ornithine L-aspartate, a potential agent for the active removal of ammonia, concluded that the available evidence is insufficient to demonstrate the effectiveness of this compound as a disease preventative (Goh et al., 2018).A randomized controlled study of rifaximin as a hepatic encephalopathy preventative in patients who underwent trans-jugular intrahepatic portosystemic shunt (TIPS) indicated a protective effect (Bureau et al., 2021).The present literature also contains some studies that have targeted gut microbiota treatments and gut barrier function (Alam & Neish, 2018).Some recent studies have explored effective preventive approaches against the hepatic encephalopathy that develops as a direct result of hepatic failure or as a complication of TIPS application for the treatment of portal hypertension in patients with cirrhosis.
One clinical trial protocol aimed at investigating the protective efficacy of the combined use of lactulose and rifaximin in the treatment of hepatic encephalopathy is available in the literature (de Wit et al., 2020).
In hepatic encephalopathy, ammonia removal has been identified as the main mechanism for prevention and prophylaxis (Butterworth, 2021).A recent review aimed at preventing and predicting hepatic encephalopathy revealed two recommendations, energy and protein intake, and pharmacological prevention, which only included two studies, one on lactulose study, the other on probiotics, and the authors stated that no other studies were found as primary prevention goals (Gairing et al., 2023).In our study, our findings for the locomotor activity tests, blood ammonia levels and liver weights revealed that bee venom had positive neurological and hepatic protective effects.Bee venom application may represent a promising medicine with preventive efficacy against the progression of hepatic encephalopathy.
Limitations
Our study was an animal experiment; therefore, human studies are needed to clinically demonstrate the protective effectiveness of bee venom as a preventative against hepatic encephalopathy.Another limitation was that, although we used locomotor activity tests, blood ammonia levels and organ weights to predict the protection level in the liver or brain, the protective efficacy needs to be determined with greater precision at the hepatic or neurological levels by further immunohistochemical and laboratory tests.Further studies are also needed to reveal which of the bee venom components, such as melittin or apanin, are responsible for the protective effects of bee venom.
CONCLUSION
This study provided the first evidence that bee venom administration has preventive efficacy against the development of hepatic encephalopathy, a serious neurodegenerative disease.The data obtained demonstrated that bee venom administration was effective in preventing the disruption of both neurological and liver functions.
These promising results encourage further studies.
Rat locomotor activity tests were measured with the open field activity monitor system (May Act 508) consisting of 4 plastic glass cages equipped with infrared photocells and 15 pairs of emitter and sensor photocells in the side blocks of the cages.The interruption of the photocell rays was recorded with a computer, and the location of the experimental animal was calculated with a sensitive software programme.Complete changes in the location of the experimental animal were termed ambulatory activity, behavioural responses that caused beam interruption but no change in location were termed horizontal activity, upward movements were termed vertical movement and other F I G U R E 1 Experimental procedure flow of the study.movements, such as sniffing, rearing or licking, were termed stereotypical movements.All rats were placed individually in this transparent cage at the beginning and at the end of the study, and their movements were recorded for 10 min to determine changes in locomotor activities.
Baseline and final locomotor activity test results and statistical comparisons of the groups.
a Kruskal-Wallis test.
TA B L E 2 Comparison of weights of organs for inflammation assessment.
a Kruskal-Wallis test. | 2023-12-20T06:17:37.835Z | 2023-12-18T00:00:00.000 | {
"year": 2023,
"sha1": "6f81fe7994091b46dc470fd9aadc82dba53fd001",
"oa_license": "CCBYNCND",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/vms3.1336",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "c5b7a752861ff30f7cebd28ae82260accd98514f",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
248875926 | pes2o/s2orc | v3-fos-license | Ultrastructural changes in porcine liver sinusoidal endothelial cells of machine perfused liver donated after cardiac death
BACKGROUND The machine perfusion (MP) preservation including hypothermic MP (HMP) and midthermic MP (MMP) has been considered as a promising strategy to preserve the functions of liver donated after cardiac death. The importance of understanding liver sinusoidal endothelial cells (LSEC) damage in regulating liver injury during MP has been emphasized. However, the ultrastructural changes in the LSEC and sinusoids around them after MP are unclear. AIM To investigate the ultrastructural changes in the LSEC and sinusoids around them after MP. METHODS Porcine liver grafts undergo a warm ischemia time of 60 minutes perfused for 4 h with modified University of Wisconsin gluconate solution. Group A grafts were preserved with HMP at 8 °C constantly for 4 h. Group B grafts were preserved with a rewarming solution at 22 °C by MMP for 4 h. Then the ultrastructural changes in the LSEC and sinusoids in Group A and B were comparatively analyzed by using osmium-maceration scanning electron microscopy with complementary transmission electron microscopy methods. RESULTS An analysis of the LSEC after warm ischemia revealed that mitochondria with condensed-shaped cristae, abnormal vesicles, reduction of ribosomes and the endoplasmic reticulum (ER) surround the mitochondria appeared. The MP subsequent after warm ischemia alleviate the abnormal vesicles and reduction of ribosomes in LSEC, which indicated the reduction of the ER damage. However, MMP could restore the tubular mitochondrial cristae, while after HMP the condensed and narrow mitochondrial cristae remained. In addition, the volume of the sinusoidal space in the liver grafts after MMP were restored, which indicated a lower risk of pressure injury than HMP. CONCLUSION MMP alleviates the ER damage of LSEC by warm ischemia, additionally restore the metabolism of LSEC via the normalization of mitochondria and prevent the share stress damage of liver grafts.
INTRODUCTION
The shortage of brain-dead donors for liver grafts is a serious problem worldwide [1]. One way to expand the donor organ liver pool is by using grafts with extended criteria including liver donated after cardiac death (DCD) [2][3][4]. However, there is a high risk of acute and chronic liver injury of primary nonfunction caused by ischemia-reperfusion injury after transplantation of DCD liver grafts [5]. Thus, the development of the preservation methods of the liver grafts after cardiac death is required to overcome these problems [1].
The superiority of machine perfusion (MP) preservation to simple cold storage was reported in clinical kidney preservation [6]. Similarly, strategies as MP with oxygen and nutrition-containing solution have also been reported to have numerous advantages to liver transplantation [7,8]. On the other hand, the MP of the DCD grafts has been discussed about the optimal conditions including perfusion temperature, oxygenation, flow rate and pressure, steady or pulsatile flow. Recently, hypothermic machine perfusion (HMP) has been established to preserve the functions of liver grafts, and its application in clinical practice has begun [9].
In addition, the warm MP including midthermic MP (MMP) has also been reported as a promising strategy to maintain the liver graft function to avoid cold ischemic injury and possibly repair the metabolic activity compared to HMP [10,11]. Our previous study also indicated the utilization of MMP reduces the hepatocellular enzyme release to perfusate [12]. We further confirmed that hepatocytes and bile canaliculi of DCD liver grafts after MMP retain a functional ultrastructure compared to HMP, by using scanning electron microscopy after osmium-maceration (OM-SEM) with complementary transmission electron microscopy (TEM) [9,13]. These results showed the practicality of OM-SEM with complementary TEM not only for determining 3D detailed ultrastructure in porcine hepatocytes but in evaluating the function of the transplanted liver which is reflected by ultrastructural characteristics of hepatocytes.
In addition to hepatocyte and bile canaliculi, the importance of liver sinusoidal endothelial cells (LSEC) damage in regulating liver injury during MP has been affirmed [14]. The inferior outcomes following extended criteria donor liver transplantation are thought to be endothelial dysfunctionrelated, however, it ultimately culminates in early graft dysfunction [14]. Thus, in this study, we comparatively analyzed the ultrastructural changes in the LSEC and sinusoids around them at four hours after HMP or MMP by using OM-SEM with complementary TEM methods. As a consequence, the LSEC that regressed one hour after warm ischemia showed a tendency to recover after MP, especially in MMP, suggesting the preventative effects of HMP and MMP on LSEC-related functions in liver grafts.
Antibody
Rabbit anti-ERG monoclonal antibody was purchased from Nichirei Bioscience Corporation (Tokyo, Japan; clone EP111) for chromogenic visualization of viable LSEC in liver grafts and observation by bright field microscopy.
Animals
Domestic female porcine (cross-bred Large White, Landrace, and Duroc pigs; age, 2-3 mo; body weight, approximately 25 kg) from Taisetsusanroku-sya Co., Ltd. (Asahikawa, Japan). The animals were kept in a well-ventilated and temperature/humidity controlled room (in which a light on 12-h cycle), and ad libitum access to food and water. All animal work was performed according to the Guide for the Care and Use of Laboratory Animals at Asahikawa Medical University. All animal studies and procedures were approved by the Institutional Animal Ethics Committee of the Clinical Research Center, Asahikawa Medical University (permit No. 14172).
MP
The harvested porcine liver was perfused with MP systems (Figure 1), as described previously [9,13]. This system consists of 2 separate circulating perfusion circuits for the hepatic artery (HA) and portal vein (PV), where each has a roller pump, a pressure sensor, and a flow meter connected to a separate circulation system via plastic connectors, allowing pulsatile and non-pulsatile flow, respectively. A gas blender attached to the oxygenator mixes air with oxygen is installed to the PV and HA circuit. Both circuits were connected to the hepatic vessels via plastic connectors and oxygen concentration was perfused and maintained at PO2 200-300 mmHg, where they are monitored by using dissolved oxygen meter installed. Waterproof thermocouples were installed in the perfusion systems to measure the solution temperature, and the heat exchanger was installed in the systems where ice-cold water controls the temperature in the organ chamber. As described previously, the flow rate was controlled as 0.22 mL/min/g for the PV and 0.06 mL/min/g for the HA [9,13].
Preparation and preservation of the liver donated after cardiac death
Pigs weighing approximately 25 kg were used as liver graft donors. In this study, the animals were intubated and ventilated under inhalation anesthesia with isoflurane (Forane; Abbott, Japan), and laparotomized. Immediately after laparotomy, the tissue samples were biopsied from the liver surface as a control. Then the animals were injected with potassium chloride intravenously to induce cardiac arrest followed by the removal of ventilation, as described previously [9,13]. The time of the induction of cardiac arrest was set as the point of 0 min of warm ischemia. During warm ischemia, the HA and PV were isolated from the surrounded tissues to connect with organ flush lines. After 60 min of warm ischemia, the tissue samples of the liver were biopsied from distinct regions of the liver surface. Immediately after tissue sampling, the liver grafts that were procured were initially flushed with a Euro-Collins solution that is less viscous via the HA and PV routes at 8 °C as a back table operation. After the operation, the organ flush lines were connected to the perfusion preservation machine systems, and the liver was continuously perfused for 4 h with a modified University of Wisconsin gluconate solution, described previously [9]. The liver grafts were conserved as two groups, A and B. The grafts in group A were perfused at a constant temperature of 8 °C as HMP (n = 3), on the other hand, the grafts in group B were gradually warmed from 8 °C to 22 °C during perfusion as MMP (n = 3), as described previously [13]. After 4 h of MP, liver tissue samples in each group were collected from the well-perfused area of the surface of liver grafts. All the blocks of liver samples were immediately fixed with appropriate fixative for the analysis as described below.
The viability of LSEC in liver grafts preserved for 4 h was evaluated by hyaluronic acid (HyA) levels in perfusate collected from the suprahepatic vena cava as described previously [15]. These results in the text and figures are expressed by the means ± SEM. Comparisons of the significance of differences between each MP group A and B were performed by using unpaired two-tailed t-tests.
Immunohistochemistry
The tissue samples biopsied from the liver of porcine were immediately immersed into 10% buffered formalin subsequently embedded in paraffin. The paraffin-embedded samples were sectioned into 4 μm slices and mounted on microscope glass slides. After the paraffin was removed by xylene, these tissue slices were rehydrated with series of graded ethanol, subsequently transferred into EDTA (pH 9.0) for antigen retrieval. After washing with PBS, the tissue slices were incubated in 0.3% hydrogen peroxide in methanol for 15 min, subsequently transferred into SuperBlock Blocking Buffer for 20 min as blocking. After these treatments, the tissue slices were incubated in a dilution (1:10) of primary antibody for 1 h at room temperature. Then the tissue slices were visualized by the Envision system (DAKO; Glostrup, Denmark), counterstained by using the hematoxylin. After the dehydration with a series of graded ethanol and xylene, the tissue slices were mounted with Malinol. The sections were observed by light microscopy.
TEM
The biopsied tissue samples from the liver of porcine were cut into small pieces and immersed in the fixative mixture of 2% glutaraldehyde (GA)/2% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB, pH 7.4) for 2 h at 4 °C. After washing with 0.1 M PB containing 7.5% sucrose, the samples were post-fixed with 1% osmium tetroxide (OsO4) in 0.1 M PB for 2 h at 4 °C. After washing with 0.1 M PB containing 7.5% sucrose, the samples were dehydrated in a series of graded ethanol solutions (starting with 50% and increasing to 100% ethanol). Then the samples were embedded in epoxy resin (Epon 812). Ultrathin sections (80 nm thick) from the embedded samples in Epoxy resin were cut using a diamond knife. The sections were observed by using TEM (HT7700; Hitachi High Technologies Inc., Tokyo, Japan) without staining by uranyl acetate and lead citrate.
OM-SEM
The osmium maceration method was described previously [9,13]. In brief, the biopsied liver tissue samples of porcine were incised into small pieces and fixed with a fixative mixture of 0.5% GA/0.5% PFA in 0.1 M PB (pH 7.4), for 30 min at 4 °C. Then the tissue samples were fixed with 1% OsO4 in 0.1 M PB for 6 h at 4 °C. After washing thoroughly with 0.1 M PB, the samples were immersed into 25% dimethyl sulfoxide (DMSO) for 30 min and subsequently 50% DMSO for 30 min as cryoprotection. Cryoprotected samples were frozen on a precooled aluminum plate chilled with liquid nitrogen. The frozen tissue samples were fractured into two pieces by using a 2.0 mm flathead screwdriver and a hammer. After freeze-fracture, the samples were transferred into 50% DMSO again for thawing, washed with 0.1 M PB, and then immersed into 0.1% OsO4 in 0.1 M PB solution for 96 h at 20 °C under the light for osmium maceration. After osmium maceration, the tissue samples were further fixed with 1% OsO4 in 0.1 M PB for 1 h at 4 °C. The samples were then washed thoroughly with 0.1 M PB before being conductive stained with 1% tannic acid in 0.1 M PB and subsequently with 1% OsO4 in 0.1 M PB. Then the samples were dehydrated in a series of graded ethanol solutions (starting 70% and increasing to 100% ethanol) and immersed in t-butyl alcohol and dried in a freeze dryer (ES2030; Hitachi High Technologies Inc., Tokyo, Japan). The dried specimens were then mounted onto an aluminum plate with silver paste and received the surface coating with platinum-palladium in an ion sputter coating equipment (E1010; Hitachi High Technologies Inc., Tokyo, Japan). After all the processes said were completed, the specimens were evaluated by using a field emission scanning electron microscopy (FE-May 21, 2022 Volume 28 Issue 19 SEM, S4100; Hitachi High Technologies Inc., Tokyo, Japan).
Localization and ultrastructural characteristics of liver sinusoidal endothelial cells in the porcine liver
In the porcine liver, the interspersed ERG positive sinusoidal endothelial cells were observed ( Figure 2A). Observation of the intercellular ultrastructure of the sinusoidal endothelial cells by TEM showed that the rough endoplasmic reticulum formed narrow networks of lumens ( Figure 2B, arrow), and cytoplasmic vesicles or caveolae occasionally open to the plasma membrane. LSEC had abundant free-ribosome in their cytoplasm and the small sausage-shaped mitochondria ( Figure 2B, green) localized around the nucleus ( Figure 2B, blue). Observation utilized by SEM revealed more detailed ultrastructural characteristics of LSEC. Low magnified observation showed the flat-shaped LSEC lined the sinusoids ( Figure 2C). Also, the space of Disse presented the normal appearance in which apical microvilli of the hepatocytes protruded. In high magnification by SEM, LSEC had the network of rough endoplasmic reticulum and vesicles correspond to TEM observation. In LSEC, mitochondria ( Figure 2D, green) had tubular or plate-formed cristae sparsely arranged identified around the nucleus ( Figure 2D, blue).
Changes of intracellular ultrastructure in LSEC after warm ischemia
Even after warm ischemia, ERG-positive viable LSEC was detected in liver grafts similar to the samples before warm ischemia ( Figure 3A). However, the ultrastructural characteristics of LSEC were substantially changed after warm ischemia. TEM observation revealed that LSEC had a slightly expanded rER, lysosome-like structure, and numerous vesicles. Furthermore, it appears to have a decrease of free ribosomes from the cytoplasm of LSEC ( Figure 3B) and membranous structure ( Figure 3B, arrowhead) surrounding mitochondria ( Figure 3B, green) at the peripheral region from the nucleus ( Figure 3B, blue). Low magnified SEM observation showed the LSEC-covered sinusoid in liver grafts after warm ischemia; however, its lumen was reduced reflecting the loss of blood flow ( Figure 3C). Additionally, the space of Disse between LSEC and hepatocytes ( Figure 3C) was reduced and microvilli were regressed. High magnified SEM observation revealed that a slightly expanded network of rER ( Figure 3D, arrow) and numerous vesicles in LSEC. Around the nucleus ( Figure 3D, blue), small mitochondria with densely distributed tubular cristae ( Figure 3D, green) were surrounded by rER ( Figure 3D and D', arrowhead).
Comparison of ultrastructural changes in LSEC after HMP or MMP
An immunohistochemical study for ERG showed similarly viable LSEC in liver grafts after 4 h for both HMP and MMP ( Figures 4A and 5A). The value of HyA in the perfusate did not indicate a significant difference between HMP and MMP. However, there was a trend of improvement in MMP ( Supplementary Figure 1, 38.33 ± 8.88 HMP_4h vs 26.00 ± 3.46 ng/mL MMP_4h, P = 0.27). Furthermore, the ultrastructural characteristics of LSEC after each MP also showed a difference (Figures 4 and 5). TEM observation of LSEC after 4 h of HMP showed that small globular mitochondria and the slightly expanded network formation of rER localized in the cytoplasm ( Figure 4B, arrow). Besides, intracellular vesicles and free ribosomes in LSEC after HMP were decreased from the point of warm ischemia 1 h. Low magnified SEM observation showed that the lumen of the sinusoids ( Figure 4C) were lined by LSEC, although there were scarcely spaced cavities of Disse between the LSEC and hepatocytes ( Figure 4C). At higher magnification, intracellular expanded rER networks were observed in LSEC. Besides, the spherical mitochondria ( Figure 4D, green) had densely but shrinking tubular cristae and remained localized around the nucleus ( Figure 4D, blue). On the other hand, after 4 h of MMP, expanded rER ( Figure 5B, arrow) and the network of endomembranous structures that had a smooth surface and expanded lumen with high electron density simultaneously appeared in LSEC by TEM observation (Figure 5B, red). Moreover, LSEC had the relatively abundant interspersed free-ribosomes in their cytoplasm and mitochondria with sausage-like formation neighboring to the nucleus ( Figure 5B). Observation by SEM revealed more detailed intrahepatic ultrastructure containing LSEC. In low magnification, sinusoids of liver grafts after 4 h of MMP tend to enlarge and form the cavity of Disse than that of HMP ( Figure 5C). High-magnified SEM observation revealed that LSEC had the expanded rER attached to the surface-smoothed vacuolar structure in the cytoplasm ( Figure 5D). Furthermore, the mitochondria with regularly arranged cristae localized around the nucleus in LSEC take a form of sausage-like structure again ( Figure 5D, green).
DISCUSSION
In this present study, we confirmed the usefulness of the OM-SEM with complementary TEM for the description of the ultrastructure of organelles in LSEC. Based on this, we revealed the intracellular Figure 1D). Previous studies using OM-SEM analyzed the intracellular ultrastructure of hepatocytes including bile canaliculi but did not describe the porcine LSEC [9,13,16]. Our data revealed for the first time that the mitochondria in porcine LSEC have tubular cristae with a relative expansion of matrix. The orthodox-shaped mitochondrial cristae are thought to reflect the state of Low ADP [17].
The distribution of anti-ERG antibody-positive cells which indicate viable LSEC in the liver after 60 min of warm ischemia does not differ from normal, which is consistent with our previous report [18]. Nevertheless, the intracellular ultrastructural characteristics of LSEC were altered after warm ischemia. Our data revealed that mitochondria with condensed-shaped cristae, vesicles, and the ER surround the mitochondria appeared in the LSEC after warm ischemia ( Figure 3D and D'). It is known that LSECs are specifically sensitive to energy deficiency [19]. The condensed-shaped cristae of mitochondria reflect an energy substrate deficiency [17]. During warm ischemia, the lack of energetic substrate interferes with active transmembrane transport in LSEC [20]. ER injury of LSEC also precedes mitochondrial disintegration in the setting of warm ischemia [4]. Damaged ER may cause the appearance of vesicles and vacuoles in LSEC after warm ischemia [21,22]. These suggest that warm ischemia impaired the membrane transport between the ER and mitochondria consequently caused morphological distortion of ER including vesiculation and membranes surrounding mitochondria.
The MP after warm ischemia alleviates the abnormal vesicles and promotes the reduction of ribosomes in LSEC (Figures 4 and 5). In previous studies, it was indicated that LSEC injury reduces liver viability, and MP protects against and alleviates the effects of endothelial injury at the ultrastructural level [23][24][25]. One of the potential protective factors of our machine perfusion systems is the early hypothermic period. The lowest degree of damage on LSEC using electron microscopy was seen at the 4 and 25 °C of machine perfusion groups [26]. Hypothermic state of machine perfusion rescue DCD livers subjected to prolonged warm ischemia via downregulation of LSEC activation [27]. The flow of machine perfusion is also another potential protective factor. The flow stasis leads to acute endothelial dysfunction and apoptosis via rapid loss of shear-stress-dependent KLF2 in LSEC [28]. Machine perfusion which triggers protection of LSEC via upregulation of shear stress-sensitive protective genes [29], consequently may alleviate the ultrastructural abnormalities shown by the present study. Furthermore, the high oxygen availability during machine perfusion also maintains the structural integrity of LSEC [30]. Short-term oxygenated perfusion potentially reduces the risk of a possible release of reactive oxygen species and endothelial shear stress [4]. Consistent with these studies, our present study showed that MP alleviates the ultrastructural abnormalities that appeared in LSEC after warm ischemia.
After warm ischemia, MMP restores the tubular mitochondrial cristae, while after HMP the condensed and narrow mitochondrial cristae remained (Figures 4 and 5). This is consistent with the finding of the previous study that the hepatocyte preserved by HMP has strongly swollen mitochondria, in contrast, MMP could preserve the functional appearance of mitochondria in hepatocytes[13]. These differences of findings rely on the temperature condition during MP. The ultrastructural feature of mitochondria with narrow cristae which appeared after HMP is similar to mitochondria observed in LSEC after cold incubation in UW solution in vitro [31]. During hypothermia, LSEC undergoes a coldinduced decrease in mitochondrial membrane potential and finally loses viability and induces apoptosis [31][32][33]. Similarly, in vivo, the HMP compromises liver grafts by prompting cold-induced damage to LSEC [34]. HMP led to a significant slowdown of mitochondrial respiration rate[33], therefore, interrupting the balance of oxygen supply/demand of LSEC. Furthermore, the reduction of membrane fluidity induced by continuous cool perfusion temperatures of HMP was examined in a previous study [35]. Extended times of HMP may also be subject to disadvantages and limitations concerning endoplasmic stress and alteration of LSEC [16].
On the other hand, the controlled oxygenated rewarming, similar to MMP, was reported to significantly increase gene expression and protein levels of the autophagy-related beclin-1 in liver grafts [11]. Electron-dense endomembranous structure that appears in LSEC after MMP resembles autophagolysosome which suggests that the LSEC is protected by autophagy [36]. The restored tubular orthodox-shaped mitochondrial cristae in LSEC after MMP ( Figures 5B and D), could supply more appropriate energy to match the demand of intracellular function of LSEC. MMP perfusate tends to have lower HyA levels compared to HMP (Supplementary Figure 1) supports the justification of LSEC has more functional integrity after MMP [15].
In addition to hypothermia, fluid shear stress can further aggravate the damage in LSEC after HMP [7,37]. It was observed that the reduction of the space of Disse in the liver grafts after HMP may be the cause of the pressure injury (Figures 4 and 5). Liver sinusoids are also known to be very sensitive to endothelial shear stress because LSEC is easily damaged by high pressure [38,39]. Endothelial injury by shear stress is the main risk of HMP [4,24,37,40]. Vascular perfusion of the liver at 4 °C induces vasoconstriction of the hepatic vasculature [41] and alters LSEC, which is vulnerable to hypothermia [31]. Morphological changes in LSEC structures and increased vascular resistance in the liver graft occurred during HMP, which obstruct the sinusoidal flow [37]. In the present study, the LSEC after HMP was observed as flat-shaped, which consistent with the findings reported by Chai et al [42].
In contrast to the HMP, our previous study showed that the pressure transitions of the HA decrease in the initial stage of MMP [12], which suggests the lower risk of shear stress in LSEC after MMP. In the present study, we showed that the volume of the sinusoidal space in the liver grafts after MMP was restored as control liver, which suggests that the decreased vascular resistance via sinusoidal space enlargement led to reducing the risk of pressure injury in MMP.
This present study has several limitations. First, the condition of MP including temperature during perfusion should be more refined for clinical application. Similar to the previous study of allogenic transplantation of porcine liver grafts under similar conditions as this study [18], the present MMP condition is not suitable for clinical transplantation. Our previous studies suggested that the liver grafts preserved by MMP suffered damage to some extent [13]. The MP with rewarming gave a more promising alternative, as many results preferring these temperature settings were reported in previous studies [43]. Furthermore, previous studies suggest that oxidative stress and endothelial shear stress can be reduced by modulating the perfusion temperature and oxygenation [44]. However, the rewarming velocity of liver grafts, the critical temperature, and the appropriate period of warm temperature needed to adjust metabolic parameters are yet to be sufficiently defined [43]. Therefore, the ultrastructural characteristics of the LSEC preserved by a better-modified MP method should be analyzed in future studies.
However, this study did not evaluate the ultrastructural characteristics of the LSEC in the liver monitored after MP for evaluation of the ischemic reperfusion injury. The most marked destructive changes of ultrastructure in LSEC are seen during reperfusion [45]. Thus, the results of the present study should be further investigated by future studies utilized normothermic reperfusion ex-situ.
CONCLUSION
In conclusion, MP preservation alleviated the ER damage of LSEC caused by warm ischemia. Moreover, MMP temperature conditions restore the metabolism of LSEC via the normalization of cristae of mitochondria and prevent the damage of the liver graft by share stress. In the future, more appropriate MP conditions clinically to preserve the LSEC functions of the liver grafts should be established by May 21, 2022 Volume 28 Issue 19
Research background
Using liver grafts donated after cardiac death (DCD) is one way to solve the shortage of donors for liver transplantation. Machine perfusion (MP) preservation has been considered as a promising, and the optimal conditions of perfusion temperature of MP of the DCD grafts, including hypothermic MP (HMP) and midthermic MP (MMP) has been discussed.
Research motivation
Recent research showed the practicality of scanning electron microscopy after osmium-maceration (OM-SEM) with complementary transmission electron microscopy (TEM) in evaluating the function of the liver grafts which is reflected by ultrastructural characteristics of hepatocytes. This has prompted the application of this novel strategy for the evaluation of liver sinusoidal endothelial cells (LSEC) damage in regulating liver injury during MP.
Research objectives
The present study aimed to establish the usefulness of the OM-SEM with complementary TEM for the evaluation of LSEC damage, and comparatively investigate the ultrastructural changes associated with LSEC at 4 h after HMP or MMP.
Research methods
Female pigs were intubated and ventilated under anesthesia, and their liver tissues were biopsied immediately after laparotomy, as control, and 60 min of warm ischemia. The liver grafts which had warm ischemia of 60 min were perfused for 4 h with modified University of Wisconsin gluconate solution by HMP at 8 °C constantly or MMP rewarming up to 22 °C and biopsied at the endpoint of MP. the LSEC in all biopsied liver samples were analyzed by immunohistochemistry and OM-SEM with complementary TEM. The viability of LSEC in liver grafts preserved for 4 h was evaluated by hyaluronic acid levels in the perfusate.
Research results
Immunohistochemistry showed the interspersed viable ERG positive sinusoidal endothelial cells in all biopsied liver samples. After warm ischemia, the LSEC showed the mitochondria with condensedshaped cristae, abnormal vesicles, reduction of ribosomes and the ER surrounding the mitochondria. Both the HMP and MMP after warm ischemia alleviate the ER damage in LSEC indicated by the abnormal vesicles and reduction of ribosomes. The value of HyA in the perfusate did not indicate a significant difference between HMP and MMP, although there was a trend of improvement in MMP, moreover, only MMP, not HMP, could restore the tubular cristae of mitochondria.
Research conclusions
This research confirmed the usefulness of the OM-SEM with complementary TEM for the evaluation of LSEC damage reflected on the ultrastructure of organelles. MP alleviate the ultrastructural abnormalities indicating ER damage of LSEC caused by warm ischemia. Moreover, MMP temperature conditions restore the metabolism of LSEC via the normalization of ultrastructural characteristics of cristae of mitochondria and prevent the damage of the liver graft by share stress.
Research perspectives
The OM-SEM with complementary TEM is applicable for the detailed evaluation of LSEC damage reflected on the ultrastructure of the MP of various conditions, including different temperatures during perfusion storage, for clinically application of the MP for liver transplantation. | 2022-05-19T15:13:41.861Z | 2022-05-21T00:00:00.000 | {
"year": 2022,
"sha1": "fe063363f4dd3238ba95d0aaed208d7bf782f71b",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.3748/wjg.v28.i19.2100",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "4b53005b086afca7068e6cec90b7b8d6d34a3335",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
237738017 | pes2o/s2orc | v3-fos-license | Use of tamsulosin and dutasteride combination for the management of benign prostatic hyperplasia: a real world practice survey (VELTAM survey)
Non-cancerous enlargement of the prostate is known as benign prostatic hyperplasia (BPH). In BPH bladder wall get thickened due to benign and enlarged prostate that exert a pushing pressure on the urethra. This results in to weakening of the bladder and loss of complete emptying of the bladder, ultimately resulting in urine retention which is known as lower urinary tract symptoms (LUTS). BPH is a common disease affecting men. Prevalence varies as per the age of the men. It ranges from 8% in men with age between 31-40 years to around 80% in men with age more than 80 years. LUTS due to BPH require therapeutic intervention in order to have sustained improvement in clinical symptoms and quality of life (QoL) of the patients. 5-alpha-reductase inhibitors (5ARIs) and alpha-blockers are the two main pharmacological agents used for the management of BPH/LUTS. Dutasteride is a 5-ARI which reduces the ABSTRACT
INTRODUCTION
Non-cancerous enlargement of the prostate is known as benign prostatic hyperplasia (BPH). 1 In BPH bladder wall get thickened due to benign and enlarged prostate that exert a pushing pressure on the urethra. This results in to weakening of the bladder and loss of complete emptying of the bladder, ultimately resulting in urine retention which is known as lower urinary tract symptoms (LUTS). 2 BPH is a common disease affecting men. Prevalence varies as per the age of the men. It ranges from 8% in men with age between 31-40 years to around 80% in men with age more than 80 years. 3 LUTS due to BPH require therapeutic intervention in order to have sustained improvement in clinical symptoms and quality of life (QoL) of the patients. 4,5 5-alpha-reductase inhibitors (5-ARIs) and alpha-blockers are the two main pharmacological agents used for the management of BPH/LUTS. Dutasteride is a 5-ARI which reduces the growth of the cells hence reducing the prostate size by blocking the conversion of testosterone to dihydrotestosterone. 6 On other hand tamsulosin shows its activity by relaxing bladder neck muscles and prostate muscle fibres by being uroselective alpha-blocker. In turn tamsulosin improves urine flow rate. 7 Combination of both dutasteride and tamsulosin has found to be superior at reducing the relative risk of BPH compared to individual component. 8 Real world clinical experiences at the physician level for this combination is lacking. To best to our knowledge this is the first survey where we tried to evaluate the clinical experience and satisfaction level of the physicians using the combination of dutasteride and tamsulosin for the treatment of BPH/ LUTS. Hence, in present study we tried to gather the evidence on real world clinical experience from the physicians using the combination of tamsulosin and dutasteride for the management of BPH.
METHODS
Current study was a retrospective, multicenter real-world study performed between April 2020 to June 2020 at Berhampore, West Bengal, India. In present survey the response from 1571 physicians who were the current prescribers of tamsulosin with dutasteride combination for the management of BPH were collected. A set of nine questions having multiple choice answers were presented to each physician and the response was collected. As it was a real-world survey, questions were designed and approved by the physicians having expertise in the field being investigated.
Statistical analysis
All the responses of above mentioned nine questions were corrected and a excel sheet was maintained. All the data analysis was performed using IBM SPSS ver. 20 software. Frequency distribution was performed to obtain tables. PRISM was used to prepare the graph. All the data is expressed as number and percentage. No further statistical test was performed.
RESULTS
Out of 1571 participants, majority of them considered tamsulosin with dutasteride as the gold standard therapy for the treatment of BPH; 1396 (88.86%). In present study it was found that most common co-morbidity which physician take into consideration while treatment a patients with BPH was cardiovascular (62.61%) followed by asthma (23.04%). Other co-morbid condition which is being considered are CKD, Cancer and DAB in 2.17%, 5.22% and 2.03% respectively. Majority of the participants (85.65%) responded that cardiovascular comorbidity is associated with 30-60% of the patients with BPH. While treating BPH patients with CV comorbidities, majority of the physician consider drug-drug interaction as significant criteriato decide the therapy (87.83%). While treating your BPH patients with CV co morbidities, do you consider drug-drug interaction as significant criteria?
DISCUSSION
Alpha1-selective adrenergic receptor blockers are considered as first-line medical therapy for BPH. However, combination of both tamsulosin and dutasteride is gaining its importance as the first line treatment for BPH. This has been proved in our survey where we collected the response of 1571 physicians who were using the combination of tamsulosin and dutasteride for the management of BPH. It was revealed that majority of the users considered tamsulosin with dutasteride combination as the gold standard therapy for the treatment of BPH (88.86%).
The combination of avodart and tamsulosin (CombAT) study included 4844 men ≥50 years of age with a clinical diagnosis of BPH who were given a combination of tamsulosin and dutasteride found that combination therapy provided significantly greater symptom benefit than either monotherapy and was found to be safe and well tolerable. 8 In present study 67.97% of the physicians using the combination were satisfied with the results and 28.55% of them were extremely satisfied with the combination therapy (Table 1, Figure 1). Finding of current survey is the confirmation for CombAT trial in real world practice for the management of BPH. Further strengthening the findings of present survey, a previous study by Djavan et al evaluating the improvement in symptom and impact of tamsulosin and dutasteride combination on quality of life of patients and reported that combination therapy not only provides better symptom improvement and improved urinary flow rate but is also associated with a more favorable impact on quality of life and patient satisfaction with treatment than monotherapy. 9 Montorsi et al have also found significantly better symptom improvement based on the BPH impact index and International prostate symptom score (IPSS) question 8 with tamsulosin and dutasteride combination than monotherapy. 10 Though BPH is not considered as the life-threatening condition, however, it's impact on the quality of life (QOL) is significant and should be considered while choosing intervention. BPH often occurs along with the age-related comorbidities. Among those most common are cardiovascular disease, hypertension, and erectile dysfunction. 11 In our survey, majority of the physician (85.65%) found that cardiovascular co-morbidity is associated with 30-60% of the patients with BPH. Presence of co-morbid condition along with BPH is a factor determining the choice of the treatment. In present study most common co-morbid condition which physician takes into consideration while treating patients with BPH was cardiovascular diseases followed by asthma ( Table 2). Tamsulosin is more selective for the alfa1A-and alfa1D-AR subtypes and has less affinity for the alfa1B subtype, which regulates blood pressure via vascular smooth muscle contraction. [12][13][14] Hence tamsulosin can safely be used among the patients with cardiovascular diseases as it does not have any impact on the predisposing factor such as blood pressure.
Previous studies have shown that inappropriate drug selection can lead to most frequently associated drugrelated problems (DRPs) such as adverse reactions, drug choice problems, dosing problems, and drug interactions in the management of BPH. Zaman et al reported 24.9% durg-durg interaction related issue among all the DRPs in the management of BPH. 15 In present survey majority of the users responded that while treating BPH patients with CV comorbidities, majority of them consider drug-drug interaction as significant criteria to decide the therapy (87.83%). Previous drug drug interaction studies have shown that tamsulosin has limited drug drug interaction as compared to other alpha blockers. Roehrborn et al found that tamsulosin and dutasteride combination therapy has similar withdrawal rate due to drug-related adverse events as compared to monotherapy. 8 Hence, it seems that this combination has less drug drug interaction in the presence of cardiovascular co-morbidities in the management of BPH. Being a real-world study and lack of control group and randomization; present study findings cannot be applied to whole population. There is a need of a large randomized clinical trial to provide strength to present study findings.
CONCLUSION
Present survey has found that physicians consider tamsulosin and dutasteride combination as the gold standard for the management of BPH as satisfaction level of patients using this combination was high. This combination also provides benefits in patients with cardiovascular co-morbidity due to the antihypertensive effect of tamsulosin. Majority of the physician consider drug-drug interaction as significant criteria to decide the therapy which can be resolved by using tamsulosin in the combination. To conclude tamsulosin and dutasteride combination is the first line choice for the management of BPH. | 2021-09-27T20:55:40.906Z | 2021-07-23T00:00:00.000 | {
"year": 2021,
"sha1": "1004f76dd48e4230101b56f44806b358764510cc",
"oa_license": null,
"oa_url": "https://www.ijmedicine.com/index.php/ijam/article/download/3000/2079",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "a03c12bd4a53fb48d52cca64df33cbcf7e4c90ae",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
5467428 | pes2o/s2orc | v3-fos-license | Profile: Professor Philippa Garety
Healthy professional one-upmanship is exemplified in Philippa Garety's position as a professor of clinical psychology, a clinical director and a joint leader of a psychosis clinical academic group. Julia Bland sought to discover whether psychiatrists have anything substantial to offer that psychologists cannot.
Instinctively collaborative, she was attracted to the job by the triumvirate power-sharing. She is the clinical lead, with an academic lead, Professor Philip McGuire, and a service lead, Lucy Canning, and they work closely together. The shared leadership model attracted her as a way of getting away from a medical director with clinical responsibility often 'keeping distance from the service director in order to be able to critique the service director'. She genuinely believes that the new collaborative style of leadership is 'more supportive of innovative and excellent clinical practice with a strategic focus'. They are trying to improve delivery of mental health services across the psychosis spectrum, from early onset, first episode, through longer term and recovery. The approach is non-doctrinaire, pragmatic, aiming for excellence in a cash-strapped health environment, not a task for the faint hearted. It has meant convincing staff, managers and finally commissioners that continual cutting of funding has the perverse consequence of increasing costs, by increasing instability and bed usage, and overwhelming staff who in turn become less effective. They have pulled off a piece of magic, successfully persuading local commissioners that investment is more effective at cost containment than disinvestment: boosting community services, along the lines successful with early intervention for psychosis, with a view to reducing bed usage and thus costs in the longer term -'It's early days, but bed usage is going in the right direction'. Strategy needs to be logical: 'I know data can be used in different ways, and that there is no simple identity between data and reality, but over the years we have seen decisions being made which are highly politicised and/or pragmatic, without really looking for the information.' There is 'some evidence' for early intervention 'when case-loads are sensible, there is good leadership, the best range of interventions is available and staff are sufficiently skilled, and supervised to offer them'. She also emphasised the need for 'understanding about the importance of engagement, working with families and carers, excellent medical intervention, vocational specialists'. And how often is that the case on the ground, I wondered?
We talked about the evidence for supportive generic outreach versus more specific and formal structured interventions. She acknowledged the importance of continuity
Julia Bland
Professor Philippa Garety, Clinical Director and Joint Leader, Psychosis Clinical Academic Group in the relationship between service user and service deliverer, whatever they think they are delivering. She also agreed that the sophisticated nurse, psychologist or psychiatrist actually uses a range of therapeutic skills within any one consultation, which are difficult to measure.
Nevertheless, it is clear that she respects scientific research methodology: it may be difficult to ask the right questions, and to deliver robust answers to questions about what is most useful and why, but this does not obviate the need to try to get the most reliable answers we can, and base policy on those rather than on mere hunch.
Over the course of our conversation, I came to the conclusion that her drivers are simple and unfashionable: compassion and a passion to provide the optimum treatment (psychological, social or pharmacological) for people with psychosis, who, as we all know, suffer enough without incompetent service delivery. Philippa Garety is not preachy, and she certainly did not directly state this motivation. But it shone through her lack of vainglory.
She was brought up in Mill Hill, in North London, and was sent to a convent by her Catholic parents. Her Christian roots have moved to the Church of England, and although it felt too intrusive to press her on theological matters, I got a sense that this is solid and important to her. At Cambridge, she studied philosophy for 2 years, with the intention of becoming a lawyer, but decided it would be 'too boring'. She came to the (then) Institute of Psychiatry to become a clinical psychologist in 1979, and perhaps it is not entirely coincidental that this institution has recently been renamed the full mouthful of the Institute of Psychiatry, Psychology and Neuroscience.
I wanted to know how philosophy had informed her thinking as a psychologist. She has described her curiosity being piqued, when, as a young psychologist, a patient stated baldly, 'You put my mother in the washing machine'.
She brings epistemological questions to bear on psychotic experience -'how can we claim to know what we know?' Wittgenstein and linguistic philosophy was the hot area during her time in Cambridge, but she was drawn to the philosophy of science and epistemology; how different models of understanding arise, how meaning is attributed. Personally, she is a flexible thinker, able to see different perspectives, and is allergic to the doctrinaire: 'one can operate within certain frameworks, as long as they seem useful, but they don't have to be seen as representing single notions of reality'. Perhaps this flexibility has enhanced her interprofessional capacities? Working, as she does, constantly, with psychiatrists, described as 'my closest colleagues'.
Interprofessional rivalry: why bother with psychiatrists?
Philippa Garety's sweetly reasonable stance did not make for an exciting discussion of interprofessional rivalry. Being interviewed for a psychiatric journal aside, I failed to elicit any personal animus towards psychiatrists.
She points out that historically clinical leadership has been attached to psychiatry and nursing rather than psychology, and 'there are far more of you'. She almost makes me feel sorry for psychologists who she claims are 'more inclined to lead other psychologists, and psychological services, but not so comfortable with other professionals'. I wonder whether she could be referring to us overconfident medics.
We discussed the unequivocal challenge to psychiatric hegemony from Allan S. Mariner, an American psychologist, who sees medical training as largely irrelevant to mental health practice: 'in the mental health field . . . only three professional activities . . . are firmly operationally connected with medical training: performing physical examinations, prescribing drugs, and giving electroshock treatment . . . the mental health practitioner, whether he be psychiatrist, psychologist, psychiatric social worker, or lay analyst, is basically practising applied psychology [my italics]. A truly relevant curriculum leading to a doctorate in mental health must be developed from the curriculum in clinical psychology, not the medical curriculum'. 1 Neuropsychiatry and liaison psychiatry aside, does Dr Mariner have a point? But Professor Garety is too tactful to make any superior claims for psychology (in spite of the better research training we do agree that psychology trainees receive).
In her peacable world, promotion should be based simply on the best person for the job, profession immaterial. Although this is eminently sensible, it does not describe the reality of mutual suspicion, ignorance, rivalry and disrespect, which I have seen in several dark corners of mental health provision.
Her one criticism of contemporary clinical psychology was the lack of social mix in the intake to training. It is so hard to be accepted for clinical psychology training that the system inevitably favours those whose parents can support them while they do unpaid work to spruce up their CVs. 'It's the internship argument', she says. I suggest that the preponderance of middle-class blonde women becoming clinical psychologists does not make for a profession that mirrors the diversity of the service users they serve, and she agrees that there is a worrying bias in gender and class.
CBT for psychosis
But Professor Garety has not entirely avoided controversy. As a leading researcher in the psychological treatment of psychosis, she was instrumental in the National Institute for Health and Care Excellence guidelines recommending cognitive-behavioural therapy (CBT) and family intervention for psychosis (2003, updated 2014). 2 Later studies have attacked the effectiveness of CBT for psychosis in preventing relapse. She concedes this may be true, but still supports the symptomatic relief that CBT affords. 'It is fairly clear on about 12 meta-analyses that CBT does offer benefit on a range of outcomes with a small-to mediumeffect size . . . Oddly enough it seems that CBT is better for symptom relief but not so effective in preventing relapse, while [family intervention] is the other way round.' So is it still worth trying to roll out CBT? An unequivocal yes: 'But of course it's not a panacea. The next generation of CBT will be more targeted for the individual, addressing command hallucinations, paranoia or depression more specifically, I hope'. She also sees future trials as being more targeted to specific symptoms.
One of her own main research interests is the emotional and cognitive context of delusions, clearly still fuelled by being intrigued by the philosophical as well as clinical questions around derivation of meaning for the individual. She still sees patients one afternoon a week.
Her empathic and sensible view is of circular reinforcement: anxiety increases the negative appraisal of self and others, promoting paranoid thoughts, which in turn increase anxiety. Social isolation and adverse environments clearly play into this cycle.
The clinician's attitude to the person with psychosis should be validating and normalising: the 'anyone who had been through what you've been through' approach, relating to the patient as someone who has experienced adversity and is coping with confusing and distressing experiences. The mental health professional has to be not only noncolluding but also non-confrontative, attempting to develop a collaborative understanding with the patient of the patient's own thinking. Easier said than done.
Her service-based research includes a randomised controlled trial of community-based assertive outreach in first-episode psychosis, with Professor Tom Craig. She has looked at the psychological effect of the deprived urban environment on individuals with persecutory delusions, and at how factors such as increased social support in a low expressed emotion environment can increase the flexibility of delusional belief.
What about life beyond psychology and psychosis? Her happiest weekends are full of family and fresh air: she and her partner (a woman and Anglican priest) sail, walk and spend time with the extended family. Philippa is the youngest of six, several of whom are also keen sailors and have houses in the same village on the Hampshire coast. Generally upbeat, she minimises the difficulties she has experienced around coming out as gay in a Catholic family, but admits that it has got easier as she has got older.
Family life is not all play: the long haul of helping her elderly parents maintain as much independence for as long as possible has taken its toll.
Her unlived life is that of a farmer, but in her world after psychology, she will settle for an orchard, although 'definitely not fiddling with making jam'. From the perspective of her current busy professional life, she does harbour a pull towards a more contemplative, outdoors life, with more room for silence: 'I think I'm quite an introvert'. All the more decent then, of this thoughtful, intelligent and energetic woman, to place her time at the disposal of a hectic organisation which demands constant interaction with people, psychotic and otherwise. | 2016-05-04T20:20:58.661Z | 2016-04-01T00:00:00.000 | {
"year": 2016,
"sha1": "8ed850d61e31b44e9e4f80d6bbf34b868995eb2d",
"oa_license": "CCBY",
"oa_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/E70DEE53358B98360E31351DF30FEAAD/S2056469400001686a.pdf/div-class-title-profile-professor-philippa-garety-div.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "8ed850d61e31b44e9e4f80d6bbf34b868995eb2d",
"s2fieldsofstudy": [
"Psychology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
256082334 | pes2o/s2orc | v3-fos-license | Predicting respiratory failure in patients infected by SARS-CoV-2 by admission sex-specific biomarkers
Several biomarkers have been identified to predict the outcome of COVID-19 severity, but few data are available regarding sex differences in their predictive role. Aim of this study was to identify sex-specific biomarkers of severity and progression of acute respiratory distress syndrome (ARDS) in COVID-19. Plasma levels of sex hormones (testosterone and 17β-estradiol), sex-hormone dependent circulating molecules (ACE2 and Angiotensin1-7) and other known biomarkers for COVID-19 severity were measured in male and female COVID-19 patients at admission to hospital. The association of plasma biomarker levels with ARDS severity at admission and with the occurrence of respiratory deterioration during hospitalization was analysed in aggregated and sex disaggregated form. Our data show that some biomarkers could be predictive both for males and female patients and others only for one sex. Angiotensin1-7 plasma levels and neutrophil count predicted the outcome of ARDS only in females, whereas testosterone plasma levels and lymphocytes counts only in males. Sex is a biological variable affecting the choice of the correct biomarker that might predict worsening of COVID-19 to severe respiratory failure. The definition of sex specific biomarkers can be useful to alert patients to be safely discharged versus those who need respiratory monitoring.
Background Men and women are infected by SARS-CoV-2 at same rate, but men have shown higher risk of developing severe disease and more often undergo complications, including acute respiratory distress syndrome (ARDS), intensive care unit admission and death, compared to women [1][2][3][4][5][6][7][8]. However, factors determining the sex differences in COVID- 19 have not been completely clarified yet. Several markers and risk factors for COVID-19 severity have been identified, such as D-Dimer and ferritin plasma levels, neutrophil and lymphocyte counts and the presence of comorbidities, but their link to sex-specific outcome of disease severity and lung disability has been scarcely investigated [9,10]. Sex hormones, in particular estrogen and testosterone, have been suggested to play a crucial role in determining COVID-19 progression. Regarding testosterone, growing evidence suggests that the levels of this hormone are involved in disease progression [11,12]. Testosterone has an anti-inflammatory effect, decreasing pro-inflammatory cytokines such as, IL-1β, IL-6, and TNF-α which have a central role in the progression of COVID-19 infection [13]. Accordingly, in elderly men the abating of testosterone correlates with the increase of a pro-inflammatory state [14]. Notably, plasma testosterone levels decrease in the presence of obesity and diabetes [15], which are comorbidities frequently associated to COVID-19. Interestingly, in male patients with COVID-19, low testosterone concentrations associated with increased disease severity and inflammation, suggesting a protective role for this hormone [16]. In contrast, other studies supported a pathogenic effect of testosterone, describing that androgen receptor activation increases the expression of the SARS-CoV-2 coreceptor transmembrane protease serine 2 (TMPRSS2) [11,17]. Accordingly, a recent study showed that prostate cancer patients receiving androgen-deprivation therapies are partially protected from SARS-CoV-2 infections [18].
Estrogen, in particular 17β-estradiol (E2), has been observed to have a protective effect in COVID-19 hampering the inflammatory storm, promoting effective immune responses, and promoting the fusion of endosomes and lysosomes inducing the virus' degradation [19]. Moreover, E2 is able to induce the expression of the angiotensin-converting enzyme 2 (ACE2) [20]. In addition to acting as SARS-CoV-2 receptor, ACE2 hydrolyses Angiotensin-II to Angiotensin-(1-7) (Ang1-7) which, by its anti-inflammatory effect, displays a protective role in several pathologies, such as hypertension, cardiovascular diseases, and ARDS that represent a risk of worse prognosis in COVID-19 [21][22][23][24]. According to ACE2 pulmonary protective effect, in animal models of acute lung injury, ACE2-knockout mice show a more severe lung failure compared with wild-type control mice [25]. Further supporting E2 protective effects, a positive association between COVID-19 and menopausal status has been detected [26,27]. In addition, E2 therapy has been shown to reduce the fatality risk by more than 50% in female COVID-19 patients > 50 years [28]. In line with these data, lower rates of hospitalization, less need of respiratory support, and shorter period of hospitalization have been observed in pre-menopausal women compared to post-menopausal women [29]. Several clinical trial are underway testing the effect of E2 and estrogen receptor selective modulators on clinical response and mortality of COVID-19 (https:// clini caltr ials. gov/).
Here, to the aim to identify sex-specific markers of severity and/or progression of COVID-19 we examined plasma levels of testosterone, E2, soluble ACE2 (sACE2) and Ang1-7 together with the known biomarkers for COVID-19 severity (D-Dimer and ferritin plasma levels, neutrophil and lymphocyte number) and the presence of pre-existing comorbidities in male and female patients taking into account the severity of respiratory disease at the time of admission and the occurrence of respiratory deterioration during hospitalization.
Patients and methods
Patients 160 adult patients (80 females and 80 age matched males) admitted to the National Institute for Infectious Diseases Lazzaro Spallanzani Hospital, Rome, Italy between March and September 2020, positive for SARS-CoV-2 by RT-PCR from nasopharyngeal swabs, and able to provide informed consent were enrolled. Intersex and transgender individuals were not represented in this study. Demographic and medical history, the presence of pre-existing comorbidities (diabetes, cancer, cardiological diseases, hypertension, asthma, respiratory diseases, kidney failure, liver failure, neurological disorders, metabolic disorders, obesity), and laboratory biomarkers (D-Dimer and ferritin plasma levels, neutrophil and lymphocyte number), were obtained and recorded for each patient upon admission to hospital. Laboratory biomarker analyses were repeated during hospitalization. Plasma samples were collected at clinical admission.
The arterial oxygen partial pressure (PaO 2 ) to fractional inspired oxygen (FiO 2 ) ratio (P/F), was assessed in each patient at admission and during hospitalization. The P/F cutoff assessment of 200 mmHg was used to distinguish two categories of patients: those with a P/F < 200 mmHg indicating moderate/severe ARDS [30,31] and those with a P/F ≥ 200 mmHg, indicating mild/no ARDS. We divided patients in two further groups: (i) patients with deteriorated disease, when the worst P/F during hospitalization was < 200 mmHg, and was lower than that observed at admission, and (ii) patients with stabilized disease, when the worst P/F during hospitalization was P/F ≥ 200 mmHg, indicating mild/no ARDS, and when the worst P/F was < 200 mmHg, but higher or equal to that observed at admission.
The study was approved by our local ethics committee, and was conducted according to the Declaration of Helsinki. The patients provided their written informed consent to participate in this study.
Statistical analysis
Data were summarized according to groups as medians and interquartile interval (25th-75th percentile), and percentages. Categorical variables, whenever dichotomous or nominal, were reported as frequencies and percentages and analysed through the Chi-square test. Statistical analysis for groups comparison was performed by the Mann-Whitney U test, Spearman's rank correlation or the Chi-Square test or Fisher test as appropriate. A P value ≤ 0.05 was considered statistically significant.
The age, for privacy reason was not collected, only two age classes were defined, up to 45 years and 55 years and more, and the patients were assigned accordingly. No patients between 46 and 54 years were included.
Multivariable linear regression analysis was performed to assess independent relationships of P/F (in mmHg), age and the other biomarkers as predictors in women and men with COVID-19. For repeated measures, such lymphocytes and neutrophils count, the worst result was considered (lower lymphocyte count and higher neutrophils count). To summarize comorbidities was created a categorical variable with the following values: 0 = no comorbidities, 1 = one comorbidity, 2 = two or more comorbidities.
Variables from the univariate analysis that showed a significant correlation with the outcome were included in the multivariable linear regression analysis. The regression was stratified by sex to provide separate models for women and men.
STATA 16 (StataCorp, Texas USA) was used for statistical analysis.
Analysis of circulating parameters in male and female COVID-19 patients divided according to P/F at hospital admission
Lymphocyte and neutrophil number, D-Dimer, ferritin, E2, testosterone, soluble ACE2 and Ang1-7 plasma levels were measured in total sex-aggregated patients and in female and male patients, stratified according to their P/F values at admission: P/F < 200 mmHg indicating moderate/severe ARDS and P/F ≥ 200 mmHg, indicating mild/ no ARDS.
As shown in Fig. 1, lymphocyte number was significantly lower in patients with moderate/severe ARDS in comparison with patients with mild/no ARDS, sex independently. Neutrophil number was significantly higher in total and in female patients with moderate/severe ARDS than in those with mild/no ARDS, but not in male patients.
D-Dimer plasma level was significantly higher in total and male patients with moderate/severe ARDS. Interestingly, D-Dimer level did not vary in female patients according to P/F ratio, suggesting the usefulness of this biomarker only in males. Ferritin plasma level was significantly higher in patients with moderate/severe ARDS in comparison with patients with mild/no ARDS, both in female and male patients.
No significant differences were observed in plasma levels of E2, Ang1-7 and sACE2 in the two groups of patients stratified according to P/F, whereas testosterone level was significantly higher in plasma from male patients with mild/no ARDS than in those with moderate/ severe ARDS.
From the correlation analysis of P/F with the circulating parameters (Fig. 2) a significant positive correlation of lymphocyte number and a significant negative correlation of neutrophil number were found in total patients, as well as in female and male patients. P/F showed a significant negative correlation also with plasma D-dimer and ferritin. Regarding E2 levels, we observed no correlation with P/F in total patients, but, interestingly, we observed a significant positive correlation in female and a significant negative correlation in males. Testosterone levels positively correlated with P/F but only in male patients and Ang1-7 positively correlated only in females, supporting its indirect dependence by estrogens that induce ACE2 expression. No correlation was observed between P/F and sACE2.
Analysis of circulating parameters in male and female COVID-19 patients divided according to deterioration or stabilization of lung disease during hospitalization
The circulating parameters reported above were also analysed in total, sex-aggregated patients, and in female and in male patients, divided in two groups according to the deterioration or stabilization of lung disease as described in "Patients and methods" (Fig. 3). Lymphocyte and neutrophil counts were significantly lower and higher, respectively, in total patients with deteriorated disease than in those with stabilized disease. To note, a significant association with disease deterioration was detected only in men for lymphocyte count and only in women for neutrophil count. These data support a sex-specific value of these biomarkers to predict disease progression. D-Dimer and ferritin plasma levels were significantly higher in total, male and female patients with deteriorated disease than in those with stabilized disease.
No significant differences were observed in E2, and sACE2 plasma levels, whereas testosterone and Ang1-7 were significantly lower in total patients with deteriorated disease than in those with stabilized disease. Disaggregating by sex, testosterone and Ang1-7 were significantly lower in male and in female patients with deteriorated disease, respectively, than in those with stabilized disease.
Pre-existing co-morbidities in male and female COVID-19 patients stratified according to P/F
The percentage of patients with moderate/severe ARDS during hospitalization and with pre-existing diabetes, cancer, cardiological diseases, hypertension, respiratory diseases and metabolic diseases was significantly higher than the percentage of patients with the same comorbidities but with mild/no ARDS (Table 1). Asthma, liver and kidney failure, neurological diseases and obesity did not show any association with ARDS severity in total, sexaggregated patients. Female patients with pre-existing diabetes, tumor, cardiological diseases, hypertension, metabolic diseases and obesity had more frequently moderate/severe ARDS during hospitalization than mild/ no ARDS. In male patients, hypertension, and metabolic diseases were significantly associated with moderate/ severe ARDS at time of admission.
Associations between the analysed parameters in female and in male COVID-19 patients separately by multivariate linear regression analyses
Multivariate linear regression analyses were performed separately by sex ( Table 2). The model incorporating age classes (up to 45 years and 55 years and more) and comorbidities, revealed that in women, being 55 or more years old (β = − 72.061; 95%CI from − 131.901 to − 12.221), having a high number of neutrophils (β = − 8.575; 95%CI from − 12.806 to − 4.344) and high ferritin concentration (β = − 0.032; 95%CI from − 0.060 to − 0.005) was negatively associated with P/F. In women a positive association with P/F was observed for E2 concentration (β = 0.766; 95%CI from 0.017 to 1.514) and lowest lymphocyte count (β = 0.451; 95%CI from 0.097 to 0.805).
Discussion
COVID-19 causes a wide range of disease, from asymptomatic to severe respiratory failure. After SARS-CoV-2 infection some patients develop a rapidly progressive respiratory failure that requires ventilatory support. Several biomarkers have been identified to predict the outcome of COVID-19 severity [32][33][34] but few data are available on sex difference in their predictive value and in risk factors for COVID-19 progression [10,35]. Since sex differences in COVID-19 lethality rate exist, it is conceivable to hypothesize that some biomarkers could predict disease severity in a sex specific manner. Accordingly, experimental and epidemiologic evidences suggest that most of the biomarkers that have been tested in the context of the risk of infection and the severity of COVID-19 are different in healthy male and female [36]. To clarify this point, we analysed plasma levels of sex hormones, and sex-hormone dependent circulating molecules, i.e., sACE2 and Ang1-7 together with the known biomarkers for COVID-19 severity (D-Dimer and ferritin plasma levels) and with the occurrence of respiratory deterioration during hospitalization and we observed that some biomarkers could be useful both in males and in females but others are sex specific (Table 3). We also evaluated whether the presence of pre-existing comorbidities was associated with ARDS severity in a sex-dependent manner.
Considering already known circulating biomarkers, we found that ferritin levels were significantly higher and lymphocyte count significantly lower, at admission, in patients with moderate/severe ARDS in comparison to those with no or mild ARDS both by sex-aggregated and by sex-disaggregated analyses, confirming data present in literature regarding the role for these circulating parameters as biomarkers of COVID-19 severity [10,37,38].
With regard to neutrophil count and D-Dimer level, our data showed that their plasma levels significantly increased in patients with moderate/severe ARDS in comparison to those with no or mild ARDS when sexaggregated analysis was performed, confirming previous reported data [12,[37][38][39]. Disaggregating the data by sex, the increased neutrophil count was observed only in females, whereas the increased D-Dimer level only in males both with moderate/severe ARDS, supporting the crucial value of considering sex differences in the search of clinical biomarkers in COVID-19.
With regard to the role of these circulating parameters as predictive biomarkers, all of them had a significant value in predicting worsening of the disease during hospitalization in total patients. However, sex disaggregated analysis of the data suggested a sex-linked role of the circulating biomarkers in worsening the outcome of COVID-19 in the hospitalized patients. In fact, we found that, whereas D-Dimer and ferritin were useful predictors of deterioration both in male and in female patients, lymphocyte count were significantly lower in men with a deteriorated disease than in those with a stabilized disease. At opposite, neutrophil count increased only in women with a deteriorated disease than in those with a stabilized one.
We also found that plasma levels of testosterone were significantly higher in male patients with mild /no ARDS in comparison to those with moderate/severe ARDS and in male patients with stabilized disease then in those with a deteriorated disease. Accordingly, a randomized controlled trial described a progress in respiratory functionality in men taking testosterone replacement therapy [40]. Moreover, a decline in testosterone has been demonstrated to be predictive of a poor prognosis in men with COVID-19 [41]. Recently, a low level of testosterone was found to be a marker of clinical severity of COVID-19 in men [42] and lower testosterone concentrations during hospitalization were associated with increased disease severity and inflammation in men [16]. Hence, these data supported a protective role of testosterone and the usefulness of measuring its plasma level to predict COVID-19 progression in men. As regard to estrogen, we observed that E2 and pulmonary functionality positively correlated in women and negatively correlated in men. This sex-dependent opposite effect could be explained by the biphasic effect of E2 depending on its concentration. At high levels, e.g., periovulatory to pregnancy levels, E2 has mainly anti-inflammatory effects, by inhibiting the production of pro-inflammatory cytokines, such as TNFα, IL-1β and IL-6, involved in cytokine storm and disease severity [43,44]. At low levels, as those observed in plasma from male patients, E2 stimulates the production of pro-inflammatory cytokines.
To note, E2 is able to upregulate the expression of ACE2 [20]. Furthermore, the gene encoding ACE2 is located on the X-chromosome, in sites commonly escaping the inactivation of one X-chromosome in mammalian XX cells (XCI), supporting the higher expression of this receptor in female than in males [24,45]. During infection, binding of viral surface spike (S) glycoprotein to membrane-bound ACE2 (mACE2) triggers ACE2 shedding. The shed soluble ACE2 (sACE2) retains its catalytic activity, but its precise role in viral entry is still unclear. Several studies have reported the beneficial and preventive role of therapeutic sACE2 in COVID-19 [46][47][48][49]. Surprisingly, clinical data suggest that patients with low mACE2 and high sACE2 faced more disease severity and fatality [49,50]. In this study, in total patients and also disaggregating by sex, in female and in male patients, we did not observe a significant association of sACE2 levels neither with ARDS severity at admission time nor with deterioration of disease during hospitalization, denying a possible use of this molecule as a biomarker of lung failure.
(See figure on next page.) Fig. 3 Analysis of circulating parameters in stabilized and deteriorated COVID-19 patients. Lymphocyte and neutrophil number, D-Dimer, ferritin, E2, testosterone, soluble ACE2 and Ang1-7 plasma levels were analysed in total (A), female (B) and male (C) patients, divided in two groups: (i) patients with deteriorated disease, when the worst P/F during hospitalization was < 200 mmHg and was lower than that observed at admission; (ii) patients with stabilized disease, when the worst P/F during hospitalization was P/F ≥ 200 mmHg, indicating mild/no ARDS and when the worst P/F was < 200 mmHg, but higher or equal to that observed at admission. Data referred to parameters are reported as mean ± SD. Data were analysed using Mann-Whitney U test. *P < 0.05 was considered statistically significant Regarding Ang1-7, this molecule is indirectly modulated by E2 which upregulates ACE2 expression. Accordingly, we observed that Ang1-7 positively correlated with P/F and its decrease predicted disease deterioration in total patients, but disaggregating data by sex showed that its predictive value was valid only for female and not for male patients. With regard to the role of pre-existing comorbidities as sex-specific risk factors of ARDS progression, the preexistence of hypertension and metabolic diseases were associated with a more severe disease in males, while pre-existing diabetes, tumors, cardiological diseases, hypertension, metabolic diseases and obesity appeared to be risk factors for more severe ARDS in females.
To explore the associations between all the analysed parameters in female and in male patients separately, we applied multivariate linear regression models and we observed that, considering age, women, being 55 or more years old, having a high number of neutrophils, and high ferritin concentration were associated with worst lung functionality, whereas E2 concentration was associated with better pulmonary function. In men, the presence of two or more comorbidities, a high number of neutrophils and high ferritin concentration negatively impacted on lung functionality, while high testosterone concentration was associated with better pulmonary function.
A limitation that could have reduced the capacity of our study to better discriminate between the severity of the disease and the sex related conditions was the focus on a single COVID-19 associated parameter, the P/F ratio only, without taking into account the final outcome. Another limitation was the categorization in two age groups only, due to restriction to the individual data for privacy reasons that, although allowed to divide according to the menopause status (mainly for women), did not allow a finer age-related analysis of the correlation.
Perspectives and significance
Our data, although requiring validation through a wider population analysis, highlight that sex should be considered as biological variable influencing the choice of the appropriate biomarker and underline the need to personalize the assistance of patients from the time of admission also basing on sex. | 2023-01-23T14:36:23.704Z | 2021-11-22T00:00:00.000 | {
"year": 2021,
"sha1": "332eeba9013042686d59b570c1528d160dec67de",
"oa_license": "CCBY",
"oa_url": "https://bsd.biomedcentral.com/counter/pdf/10.1186/s13293-021-00407-x",
"oa_status": "GOLD",
"pdf_src": "SpringerNature",
"pdf_hash": "332eeba9013042686d59b570c1528d160dec67de",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
3923236 | pes2o/s2orc | v3-fos-license | Occurrence of lactational mastitis and medical management: A prospective cohort study in Glasgow
Background Lactational mastitis is a painful, debilitating condition that if inappropriately managed, may lead women to discontinue breastfeeding prematurely. The aim of this paper is to report the incidence of mastitis in the first six months postpartum in a Scottish population, its impact on breastfeeding duration and to describe the type and appropriateness of the support and management received by affected women from health professionals. Methods A longitudinal study of 420 breastfeeding women was undertaken in Glasgow in 2004/05. Participants were recruited and completed a baseline questionnaire before discharge from hospital. Cases of mastitis were reported either directly to the researchers or were detected during regular follow-up telephone interviews at weeks 3, 8, 18 and 26. Women experiencing mastitis provided further information of their symptoms and the management and advice they received from health professionals. Results In total, 74 women (18%) experienced at least one episode of mastitis. More than one half of initial episodes (53%) occurred within the first four weeks postpartum. One in ten women (6/57) were inappropriately advised to either stop breastfeeding from the affected breast or to discontinue breastfeeding altogether. Conclusion Approximately one in six women is likely to experience one or more episodes of mastitis whilst breastfeeding. A small but clinically important proportion of women continue to receive inappropriate management advice from health professionals which, if followed, could lead them to unnecessarily deprive their infants prematurely of the known nutritional and immunological benefits of breast milk.
Background
Lactational mastitis is a painful, debilitating condition that can adversely affect mothers in their efforts to breast-feed their babies [1,2]. Despite being a relatively common complication of lactation, surprisingly few studies documenting the incidence of, and risk factors for, the condi-tion have been reported. To our knowledge no large longitudinal study of mastitis has been conducted in the UK in recent times. Observational studies conducted in the USA [3], Finland [4] New Zealand [5] and Australia [6,7] suggest however, that up to 20-25% of breastfeeding women will develop mastitis during the course of lactation and approximately 20-35% of women who develop mastitis will experience recurrent episodes [5][6][7][8].
In some studies, mastitis has been associated with the premature cessation of breastfeeding [1,2]. Sufferers of mastitis may stop breastfeeding because of the pain associated with the condition or because they have been inappropriately advised by a health professional to do so. Health professionals therefore, have a responsibility to provide appropriate advice and support to assist women to manage the condition successfully and to continue breastfeeding.
The purpose of this paper is to report the incidence of mastitis in the first six months postpartum in a Scottish population, its impact on breastfeeding duration and to describe the type and appropriateness of the support and management received by affected women from health professionals.
Methods
Study participants were recruited from women who had given birth at the Princess Royal Maternity Hospital in Glasgow between April 2004 and January 2005. Women who were breastfeeding at the time of recruitment and had delivered a full-term, healthy singleton were eligible to participate. Those who were fully formula feeding at the time of recruitment, lived outside the Greater Glasgow Conurbation, had a multiple delivery, a low birth weight infant (< 2500 g), a premature delivery (< 37 weeks gestation) or who could not read and speak English were excluded from the study.
Women meeting the inclusion criteria were visited on the maternity ward by the research midwife (SM) within 24-48 hours of delivery. They were given a written and verbal description of the study and were informed that they could decline to participate in, or withdraw from, the study without prejudice to their care. The study was approved by the Local Research Ethics Committee of the North Glasgow University Hospitals NHS Trust and written informed consent was obtained from participants.
Participants completed a baseline questionnaire prior to or shortly after discharge from hospital, providing information on demographic characteristics and previous births, breastfeeding experience and episodes of mastitis. Women were followed up by telephone interview at 3, 8, 18 and 26 weeks or up to the time that they discontinued breastfeeding. Follow-up interviews included questions on breastfeeding practices, breast and feeding related problems, breast care, use of breast pumps and maternal and infant health.
Case definition
Mastitis was defined as a red, tender, hot, swollen area of the breast, accompanied by one or more of the following [6]: i) an elevated temperature (either estimated or measured as being ≥ 38°C ) or ii) one or more of the constitutional symptoms of fever (body aches, headaches and chills) or iii) diagnosis of mastitis from a medical practitioner.
Symptoms had to have been present for a minimum duration of 24 hours [7].
Women who experienced an episode of mastitis post-discharge were instructed to telephone the research team to inform them of the event if their symptoms persisted for 24 hours or more. Additional cases of mastitis were identified retrospectively at the time of the telephone followup interviews. Those identified by either method as having mastitis were requested to complete and return a 'Mastitis Case' questionnaire, a copy of which had been given to them as part of an information pack at the time of recruitment (see Additional file). This questionnaire elicited information about events leading up to the episode of mastitis, management advice received and followed, and outcome.
Statistical analysis
Data were analysed using the Statistical Package for the Social Sciences (SPSS for Windows Version 15) [9]. The estimated incidence of mastitis was based on women who had symptoms meeting our criteria who either directly contacted the research team or who were identified at one of the follow-up interviews. The target sample size of 500 had the power to detect an incidence of mastitis of 20% with an accuracy of 3.5%. The achieved analysis population of 420 had the power to detect an incidence of 20% with an accuracy of 3.8%.
Descriptive statistics include means and standard deviations or percentages. Survival analysis and the log-rank test were performed to determine if there was a difference in the duration of breastfeeding between women who developed mastitis and women who did not.
Results
During the study period 1141 breastfeeding women were identified. Of these, 183 were missed or discharged prior to completion of recruitment and 372 did not meet the eligibility criteria. The remaining 586 women were invited to participate and of these 40 women refused to participate, a further 46 agreed to participate but provided incorrect or insufficient follow-up contact details and a further 80 discontinued breastfeeding within two weeks of delivery. In most cases women discontinuing breastfeeding before 2 weeks failed to establish breastfeeding successfully due to feeding difficulties. As an objective of the study was to identify determinants of mastitis (to be reported separately), any woman discontinuing breastfeeding before 2 weeks for reasons other than mastitis was excluded from the analysis population as their inclusion might reduce or mask the association of feeding difficulties and mastitis.
Incidence of mastitis
In total, 420 women were included in the analysis population (response rate 72%) and 74 women (18%, 95%CI: 14%, 21%) were identified as having experienced at least one episode of mastitis in the first 26 weeks postpartum. Mastitis was not significantly more common in women with a prior history of mastitis compared with those with no prior history (Table 1). Fifty-seven women (77% of cases) completed the Mastitis Case questionnaire, 39 of whom (68%) reported only one episode of mastitis but 13 women (23%) reported two episodes and five women (9%) three or more episodes of mastitis. In total 76 episodes of mastitis were reported by the 57 women completing the Mastitis Case Questionnaire. Thirty initial episodes (53%) of mastitis, and 33 (43%) of all episodes, occurred within the first 4 weeks postpartum ( Figure 1). The mean time to first episode of mastitis was 6.3 (SD 6.6) weeks.
Mastitis and breastfeeding duration
Of the 74 women who developed mastitis 28 (38%) stopped breastfeeding before the end of the study (26 weeks). The duration from time of mastitis to stopping breastfeeding was 6.1 (SD 6.6) weeks. Of the other 46 women who developed mastitis, 41 were continuing to breastfeed at 26 weeks while the remaining 5 were lost to follow-up during the study but were breastfeeding at the time of last study contact. The occurrence of mastitis was not negatively associated with breastfeeding duration (Figure 2). In fact, those women who suffered an episode of mastitis were significantly more likely to be breastfeeding at 26 weeks than those who did not suffer mastitis (Log-rank test χ 2 = 8.81, df = 1, p = 0.003).
Advice and treatment
Information on advice and treatment was available for the 57 women who completed the Mastitis Case questionnaire. Of these women, 21 (37%) were able to manage their first episode of mastitis without consulting a health professional. Thirty six women sought advice from a health professional: 21 (37%) from their General Practitioner (GP), 18 (32%) from a Community Midwife and 12 (21%) from a Health Visitor. Twelve women (21%) consulted more than one health professional, possibly indicating that more serious cases were referred on to a woman's GP by her Community Midwife or Health Visitor.
The type of advice given to women by their health professional is listed in Table 2. Six women (10%) were told to stop feeding either from the affected breast or altogether.
Discussion
The incidence of mastitis reported in this study is comparable to the incidence rates of 17.3% and 20% reported in recent Australian studies [1,6,10]. The only other study of Scottish women identified was conducted in the 1940s and reported an incidence of breast abscess of 8.6% [11]. However, as non-suppurative forms of mastitis were not identified, the real incidence of mastitis was likely to be higher. This and other earlier studies [8,12] probably underestimated the true incidence due to limitations in case ascertainment and the short time period that women were followed postpartum. For instance, earlier studies identified only those women with mastitis who sought medical treatment from the hospital where they were delivered and used hospital medical records as their source of data [8,12]. A more recent US study [3] reported a 9.5% incidence of health care provider-diagnosed lactational mastitis. In our study only one woman reported visiting a hospital casualty department with approximately one third of women visiting their GP. The majority of women either self-managed their mastitis or consulted only their community midwife and/or health visitor for management advice. In general, incidence rates for mastitis are below 10% when medical records and women seeking medical advice are used as a source of data, whereas incidence rates of around 20% are seen in studies where diagnosis is based on self-reported symptoms [1].
Duration of breastfeeding, by mastitis (n = 420) Figure 2 Duration of breastfeeding, by mastitis (n = 420). The recommended management of mastitis is usually conservative, the key recommendation being that mothers continue to breastfeed and to feed more frequently or express milk from the affected breast(s), in an effort to clear blocked ducts and engorgement [13][14][15]. Most women received advice consistent with current recommendations, however one in ten affected women were inappropriately advised by a health professional either to stop breastfeeding from the affected breast or to stop breastfeeding altogether. While in relative terms this may seem small, in absolute terms this represents approximately 680 women across Scotland receiving inappropriate advice annually. This estimation is based on the assumption that 70% of Scottish women delivering in 2004 [16] (n = 53957) initiated breastfeeding [17] and that 18% of these women experienced mastitis, of whom 10% received inappropriate advice.
There appear to be differences between countries in the extent to which antibiotics are prescribed to treat mastitis. In our study just over half (53%) of women who reported mastitis were prescribed antibiotics. This is higher than the 38% reported in a Finnish study [4] but lower than the 75% or more in Australian studies [6,18] and 86% in a recent US study [3]. In a recent Swedish study just under 15% of women with mastitis were prescribed antibiotics [19]. Of these, 3.3% of cases were prescribed antibiotics on the basis of their symptoms and the remaining cases (11.4%) were prescribed antibiotics on the basis of culture results.
The bacteriological analysis of breast milk is not routinely practiced in the UK with women usually being prescribed antibiotics on the basis of the severity and duration of their symptoms. Potentially pathogenic bacteria are found in the breast milk of healthy breastfeeding women and because the results from bacterial cultures may be difficult to interpret, it has been suggested that the bacteriological examination of breast milk is not particularly informative in the decision to treat mastitis with antibiotics [19]. However, in light of the fact that community acquired methicillin resistant Staphylococcus aureus (MRSA) is becoming more common, breast milk culture and sensitivity testing is recommended if the condition does not respond to antibiotic therapy within two days or if the mastitis recurs [14,15].
The difference in prescribing rates may be related to the number of women who self-manage their condition or seek advice from a heath professional other than their GP. Scottish women tended to consult their Community Midwife or Health Visitor, some of whom may have organised a prescription for antibiotics, with only just over a third consulting their GP. Whereas in an Australian study the majority of women (73%) had sought treatment and advice from their GP [6] and all of the women in a US study [3] were diagnosed following a medical consultation, thus increasing the likelihood of antibiotics being prescribed.
Staphylococcus aureus is the most common organism responsible for mastitis [20] and recent Clinical Practice Guidelines [15,21] recommend penicillinase-resistant penicillins such as flucloxacillin and dicloxacillin as the drug of first choice, or cephalexin and clindamycin in women who are allergic to penicillin. While the WHO publication on mastitis also recommends amoxicillin and erythromycin [14] more recent guidelines advise against the use of these drugs on the basis that a significant proportion of isolates of Staphylococcus aureus are resistant to these antibiotics [15,21]. Of the women who could recall the antibiotic they were prescribed (20/30) almost half (9/20) were prescribed an antibiotic that was not consistent with current practice guidelines. Kvist et al. recommend that the "imprudent use of antibiotics be avoided because of the spread of MRSA and other multi-resistant pathogens" [19]. Both their and our results suggest that a relatively large proportion of women can conservatively manage their mastitis without resorting to taking antibiotics.
Mastitis has been associated with the premature cessation of breastfeeding [1,2]. However, this was not the case in a recent study of Australian women where no association between mastitis and breastfeeding duration was found [10]. In our study, women who experienced mastitis were significantly more likely to be breastfeeding at 26 weeks than those who did not experience mastitis, which is sim- Percentages add to more than 100 as some women may have received more than one form of advice ilar to the finding of a study of New Zealand mothers [5]. Vogel et al. concluded that mastitis is more likely to occur in mothers with ample milk supply, who may be more at risk of milk stasis if they delay or miss a feed [5].
The strengths of this study are the relatively high response (72%) and the high follow-up rate (95%). In addition, we had frequent and regular contact with women allowing us to pinpoint the timing of onset of mastitis. There are also a number of limitations to this study. Firstly, women identified as cases through the follow-up interviews were only identified if they answered yes to having had mastitis specifically. They were not asked if they had experienced any symptoms suggestive of mastitis. However, the results of our study are strikingly similar to those of Amir et al. who, in order to reduce bias, avoided asking about mastitis directly but collected information about mastitis symptoms [10]. A further limitation of this study is that almost half of participants were continuing to breastfeed at 6 months, compared with the national average of 25% [17], suggesting that our sample was not necessarily representative of all breastfeeding women in Scotland. Despite these limitations, the mastitis incidence rate in this study is reasonably consistent with the incidence rates from studies of women in other Western countries.
Conclusion
The findings of this study suggest that one in six women may develop lactational mastitis. While most women receive appropriate management advice from health professionals, a clinically significant number of women are advised to stop breastfeeding from the affected breast or to stop breastfeeding altogether. If followed, this advice could lead to women unnecessarily depriving their infants prematurely of the known nutritional and immunological benefits of breast milk. | 2014-10-01T00:00:00.000Z | 2008-08-25T00:00:00.000 | {
"year": 2008,
"sha1": "33711cd4e502b3a33487958c67fed74b3ea9f430",
"oa_license": "CCBY",
"oa_url": "https://internationalbreastfeedingjournal.biomedcentral.com/track/pdf/10.1186/1746-4358-3-21",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "33711cd4e502b3a33487958c67fed74b3ea9f430",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
172126050 | pes2o/s2orc | v3-fos-license | Uma narrativa histórica da educação nacional entre mitos: do fundador ao de democracia racial
This paper aims to develop a debate about the characteristics that consolidated an autonomous national project with imposing pretensions, markedly from the turn of the 19th century to the 20th century. It tries to address how the different myths about equality between people have been treated historically by national education, up to the current document of the National Curricular Common Base. Through the literature review, a historical trajectory is recovered in six stages. It is identified that the myth of "racial democracy" serves as an argument to exclude compensatory measures in the name of a meritocracy that treats dissimilar as similar. Therefore, in addition to recovering the historical silencing, it seeks to indicate means for the construction of a democratic pedagogical political project for the national education system.
Introduction
This paper presents the results of a master's thesis, developed in the Post-Graduate Program in Human and Social Sciences, of the Federal University of ABC, Brazil. The main objective is to develop a debate about the characteristics that consolidated an autonomous national project with imposing pretensions, markedly from the turn of the 19th century to the 20th century. An essential period for the creation and consolidation of a supposed homogeneous narrative of "nation", which in turn would be consolidated, according to the interests of the dominant classes, through the incipient urbanization and industrialization of that period. This unitary perspective of a people and territories markedly different from each other disregarded the specificities of our colonial and slave-owning formation, spelling out what Marilena Chauí (2000) called the "founding mythic imaginary." According to the author, this imaginary consists of a fantasy, socially disseminated, that reproduces a national past, in which the different social subjects (black, indigenous, European) lived together in an integrated, harmonious and peaceful way in the construction of the "Brazilian nation". This imagery has endured as an important factor of silencing and non-recognition of ethnic-racial issues, among other subjects, for the understanding of the relation and inheritance present in the discourses about the black and the indigenous, in a country with continental dimensions of enormous inequalities and social, political and economic conflicts such as Brazil.
The purpose of this article is to address how the different myths about equality between people have been treated historically by national education, up to the current document of the National Curricular Common Base, now referred to as BNCC. Moreover, since this document intends to standardize contents and strategies of teaching in a territory marked by differences, it is necessary to analyze it under the historical perspective of constructing the idea of "Brazilian nation". The understanding, therefore, of who the Brazilian is and his representations, becomes a question of identity and relations of power. The curriculum analysis, in turn, longs to interpret how the process of learning and teaching developed in the Brazilian experience, marked by the exclusion of nonwhite actors in the process of building national memory. According to Nascimento (1978), the negative representations of blacks and indigenous peoples find in the diffusion of the myth of "racial democracy" a determinant condition for the questions that involve the struggle for the identity recognition suppressed in this process. However, the attempt to create a unit, through a multicultural territory, becomes central to the understanding of issues involving curriculum, race, and ethnicity. The search for civility or "Brazilian nation", in this perspective, did not consider or contemplated the diversity of indigenous and African peoples, being these, submitted to the colonial yoke under the aegis of the European.
The civilizational model imposed by the Europeans was considered superior in relation to other peoples and models of society. In this case, education and the curricular field are instruments that allow the dissemination and questioning of representations about blacks and indigenous peoples, politically and ideologically submitted to the systems of European settlers. According to Silva (2007), public education in Brazil was born excluder and racist, making it necessary to relate the dispute over the curriculum as a possible unfolding of the struggle for rights and recognition of groups, whose identities were violated violently in the name of a country project that was intended to be white and European. Still according to the author, black and indigenous identities defend the emphasis on the difference vis-a-vis the historically produced segregation in the country.
In this direction, Gomes (1995) explains that identity must be understood as a way of being in the world, expressing practical activities such as languages, celebrations, rituals, eating behaviors and popular traditions that are references and marks of a group. It is not enough, therefore, to contemplate, in the official documents, the questions that deal with ethnic-racial relations. In this sense, attributed by the author, as a curriculum is a dispute (political, ideological, social, etc.), education for ethnic-racial relations intends to enter this contest in an active and protagonist way, becoming the center, not periphery, of the debate.
In order to carry out the discussions proposed in this article, we try to recover a historical trajectory in six stages, beginning with the construction of the Brazilian nation, passing through the (2.) slave nation, (3) by the place of the indigenous peoples, (4.) by the proclamation of the republic and the abolition of slavery, (5.) by the origin of Brazilian racial thought, until the last stop, (6.) which concerns a possible myth of Brazilian racial democracy.
On the construction of the Brazilian nation
The issue of ethnic-racial relations in the construction of the Brazilian BNCC allows us to recover the process of formation in Brazil of a society built, according to Faoro (2000), on a slave-owning and latifundia perspective. This history was based on aristocratic bases, with full control of public and private institutions, as well as political and economic control by the local elites. The slave model, however, is not restricted to the colonial and imperial past, and part of the recent disputes are a recognition by the country's institutions of this process and its social, political, economic and cultural developments. In this context, we seek to analyze the meanings and intentions in the disputes involving the construction of the BNCC, which seeks to unify contents for basic education, making it imperative to follow the historical process of organization and struggle of Afro-Brazilians and indigenous people through recognition and access to rights by the institutions that make up the Brazilian State.
These foundations, which form and sustain a territory marked by the exclusion of social, cultural and economic orders, are still present in contemporary debates and demands; among the exclusions, the ethnic-racial question and the curricular field allow a contextualized rescue of this process. According to Chauí (2000), Brazilian society is the product of a myth, created by the country's intellectual elites, which, in turn, has been able to keep them linked and aligned in the long process of formation of the nation. Therefore, it is necessary to seek, in the history of Afro-Brazilians and indigenous peoples, the recognition and challenges that permeate the incessant struggle of these groups for recognition and access to rights, of which they were historically denied to them in this country.
The history of the black and indigenous in the formation of Brazilian society is now part of the official school curriculum of the nation, through the publication of Federal Law 10.639 of January 2003 which, in turn, changed the LDB 9.394 of 1996. Among the objectives of this curricular insertion, is the search for recovering and valuing the protagonism of these subjects in the formation of the Brazilian nation. This achievement sought to bring into the field of curriculum and education a counterpoint to the Eurocentric vision, historically consolidated in national educational institutions, in order to dispute and re-significate the social function of the school.
In this way, the representation of mixed, gay and plural country gains strength in the popular imagination, after all, the history of Afro-descendants and indigenous people is valued at school. But for Nascimento (1982), this narrative that values mix and diversity also covers the genocide that was imposed by the Brazilian and European ruling classes that occurred in these five centuries of colonial post-domination history. Violence is perpetuated, with a clear ethnic-racial cut, as we can see, for example: in the executions of young people in the outskirts of the big cities, as well as the violence that affects indigenous peoples in conflict with farmers in the land dispute in the Brazilian northeast. According to a report by Pelegrini (2012), for example, data from the international amnesty show that the number of homicides in Brazil in 2012 was 56 thousand. Of these, 30,000 youths were murdered, and most of these lives were wasted on the outskirts of large Brazilian cities. In this genocide of the country's youth, 77% are black youth. These figures reveal the existence of a genocide with a clear profile: young, black and residents of the outskirts of big cities. Another example comes from the state of Maranhão, in Borges and Nossa (2017), in which the journalist recounts the attack of gunslingers suffered by the ethnic group "gamela", in dispute for lands in the State, which ended with about 13 injured indigenous people being some with their hands cut off.
In this way it is also representative of this context of violence and exclusion, the characteristics of the educational system of the country. This is because the system, in turn, historically reproduces Eurocentric discourses and practices that devalue and ignore the knowledge and protagonism of natives and exiles for the history of Brazil, which is why a series of affirmative achievements in different spheres of the educational and social field, should be valued. Malachias, Bento and Silva (2010, p. 143) argue that "Law 10.639 / 2003 [...] crowns a trajectory of struggle that has been waged by the Black Movement in its claims for affirmative educational policies." Therefore, understanding the narrative that seeks to silence the ethnic-racial conflicts that characterize the past and the present in the curricular and educational field requires a historical perspective on the narrative of an integrated and peaceful nation. Faced with this, it becomes urgent to understand the "founding myth" and its impact for the re-signification of this conflicting past. Myth, as an attempt to mitigate conflicts or explain socially disseminated practices, is present in the history of human relations in various contexts. In colonial, agrarian, slave, and postcolonial Brazil, the invention and consolidation of the National States in Europe would arrive as an imposition for an elite that was claimed to be universal and modern. The construction of a grandiose and universalizing narrative for an eminently rural and archaic country has become a difficult problem. It was necessary to rewrite history.
The presence of the founding myth is the starting point for the understanding and centrality that this process carries to problematize and understand the current issues that permeate the territory of the curriculum and the disputes involving the construction of BNCC.
The myth, in this perspective, acts to consolidate the imagination that naturalizes and idealizes the process of formation of the country. This myth, on the other hand, offers an idealized representation of the reality and the national past and that continues being updated and adapted according to the convenience of the economic elites of Brazil. With this capacity of adaptation and updating, the allegory can be repeated indefinitely. In order to build and consolidate the idea of unity and harmony among the different, in a class society, a conciliatory solution was needed. The apparent departure was to reach a point in common, as emphasized by Chauí (2000, p.17), who aimed to incorporate "in a single belief the rival beliefs, that is, the class appeal, the political appeal and the religious appeal did not need to dispute the loyalty of citizens because all these beliefs could express themselves for one another under the common background of nationality". This myth and his symbolic strength continue allowing and constructing new meanings that ensure a sense of unity and communion among the Brazilian population. When this harmony was threatened by the social and economic division generated by the model of capitalist production, in a process of consolidation and global hegemony at the turn of the nineteenth century to the twentieth, the idea of nationality acted as a symbolic force in the maintenance of social inequalities in Brazil. The national question is the thread that leads and organizes the narratives during the imperial period as a way to "naturalize" great asymmetries and the conflicts that threatened the current social order. The regent revolts 3 , hard-pressed by the imperial government, are examples of the resistance of the excluded and the reactions of the local elites, in defense of the privileges they inherited from the colonial period. The threats were fought in the institutional military field, with the precious aid of political and social identification represented by the idea of nation.
The process of independence of Brazil was negotiated with the metropolis and the local elites as a way of maintaining the social and economic unit unchanged. The Constitution of 1824 guaranteed the privileges of the slave elite and provided for the creation of a national We can see that this decree sought to equate the condition of the slave with that of a contagious patient, exposing, in social, political and economic aspects, the denial of rights to blacks, in the incipient public education system. What is implied in this form of suppression is the idea that both, black and sick, would infect the social body, and it was found that access and the right to education in the newly created Brazilian nation would continue to equate blacks with a category of sub according to the hegemonic thinking that prevailed at that time.
The lack of historical approaches and documentary sources about the presence of the black in the institutionalized education of Brazil, during the beginning of the Brazilian "nation", reveals that the memories and the narratives constructed under the idea of a people's unity go through the silencing of non-the formation of the country. Although not massively, layers of the freed black population, during the imperial period and the beginning of the Republic, accessed formal education through networks of solidarity. They created their own schools in quilombos, in partnerships with religious orders directed to blacks, long before the universalization of public education in Brazil.
The struggle of black groups for admission to the official schooling process proves that, even on the verge of citizenship, blacks followed the processes of transformation of the "Brazilian nation" and exerted influence on them. School space for the Negro served as a form of social ascension to the extent that it reinforced discrimination through such exclusion and lack of mass access to the educational system. This right to memory and the need for a look at ethnic-racial issues should be equivalent to the valuation of Europeans and Orientals, as Kabengele Munanga (2008) points out: African and indigenous cultural heritages constitute one of the fundamental matrices of the so-called national culture and should therefore occupy the same position as European, Arab, Jewish, Oriental, and other legacies. Together, these legacies constitute the collective memory of Brazil, a plural, non-mestizo or unitary memory. A memory to be cultivated and preserved through family memories and the educational system, for a people without memory is like a people without history. It is precisely here that the problem arises, since the African and indigenous cultural heritages in Brazil have never occupied a position of equality with the others in the national education system. If this were the case, then the laws 10,639 / 2003 and 11,645 / 2008 promulgated by the current President of the Republic, 115 years after the abolition would not make any sense 4 . (MUNANGA, 2008, p. 49-50) The inflection for the idea of unity around a great people had a sudden repercussion in the intellectual elites, in the transition from the nineteenth to the twentieth century, in Brazil.
The thought produced in this period shows the condemnation to the limbo of Afro-Brazilians and Indians as inferior races, which should be extirpated by the natural process of "laundering" that the new great nation would be bound to attain. This historical process of invention of the Brazilian nation is present from the disputes involving the process of colonial independence in the imperial period and enters the Republic, remaining in the debates in a decisive way in the twentieth century.
The constitution of a nation and its national identity were the great concerns of the The slave system as the economic base of the "nation" The slave system, the economic base of the colony and the Empire, was marked by the naturalization of violence by the owners and state control institutions, as well as by the great resistance of the slaves, which was characterized by small daily insubordination, such as the numerous revolts , to the formation of quilombos, like Palmares (AL). Estimates, according to Nascimento (1978), revolved around the arrival of some 5 million captives from various regions of the African continent. Still according to Nascimento (1978), about a third of the slaves who entered the country were Bantu, coming from Central Africa and Angola. The mix also included groups from Senegal, Congo, the Mine Coast and the Gulf of Benin. This diversity was present, in great quantity, in the heavy workings of the field and in the intimate daily life of the big house. The names of baptism were added by color and place of birth. The slave masters preferred heavy labor to a large number of laborers from various parts of Africa, and preferably newcomers, to avoid rebellion and escape -common forms of resistance and present throughout the country the slave-owning cycle of Brazil.
According to Alencastro (2000), in the attempt to uproot the captive of his African origin, the process of de-socialization and depersonalization acted as important resources in the attempt to commodify and commodify the individual in slave systems under different contexts, such as Brazilians. Nevertheless, the diversity of African ethnic groups allowed for great alliances, loyalty ties and cultural exchanges between different ethnic groups. The strenuous crossing of the Malungos (the way the ship's companions and friends were called) represented a great massacre and ethnic genocide, with diseases, suicides and other violence practiced by the colonial slave-owning elites. On the other hand, there was an intense exchange in this process between the cultures that helped to build Brazil.
Throughout the enslaved period, Schwarcz and Starling (2015) explain, the enslaved sought in the breaches of the system ways to recreate their cultures, to maintain the rituals, to reproduce and adapt their beliefs, to invent desires and to fight incessantly for freedom. The Europeans had known slavery since antiquity, with Greek cities and the Roman Empire being great examples of enslaved societies at that time. Modern slavery was the foundation of the colony. In places such as the Recôncavo, in Bahia, slaves made up more than 75% of the population. The African continent, on the other hand, did not have at the height of the slave trade a political or religious unity, the different forms of social organization occurred by lineage and kinship. It should be noted that the African traders organized the mercantile relations taking care of their own economic interests, so they were not passive in this relation.
In order to understand a grandiose model of commerce involving human lives, a moral justification was therefore required. Colonial, export-oriented units based on slave labor and monoculture latifundia produced one of the greatest ethnic genocides in history. Violence was the engine of control of this system. Public punishment, whipping, rape were common in this asymmetrical relationship, always counting on the consent of religious power, disseminating fear and dread as a constant form of control.
A cruel and inhuman practice such as slavery leaves fundamental social marks for naturalizing the differences between whites and blacks in Brazil. This social marker of difference lies today, to stay in some examples of this Brazilian wound, in the violent peripheries of the country, through the programs that spectacularise and trivialize violence against the black, until the disputes over public policies that recognize the genocide and the specificity of the black cause. The enslaved society considered all physical work, or that demanded humiliating effort. In order to meet the demand for specialized labor, urban slaves were leased by day or week for the exercise of multiple activities and trades, such as carpentry, painting, stowage, blacksmithing, barber shop, shoe store, among others. The women were domestic workers, cooks, housekeepers, laundresses, and other crafts. The spread of slavery was not restricted only to large estates, for it shaped social relations in all spheres, defining social inequalities, and placing race and color as fundamental markers of differences. The slave labor was exploited by merchants, military, civil servants, innkeepers, artisans, and even freedmen. Socially disseminated memory by common sense often implies a passive character in the struggle of the Negro for rights and against slavery. This imaginary ignores the black protagonism in the resistance to the condition in which it was, due to great ties of affection, religious associations and other disguises and negotiations that mark this process, as emphasized by Schwarcz and Starling (2015, p. 96] the enslaved did not behave like "things," having always acted their place and condition, struggled to get their leisure time, maintain their family, recreate their customs in strange lands, worship their gods and practices, preserve their children and care their".
According to the authors, this daily violence and resistance, during the period of institutionalized slavery, caused panic in the local elites, with the threats of rebellions and revolts of the slaves. Organized uprisings, mass fugues, and the assassinations of overseers terrified the elites, always preoccupied with maintaining their own privileges. The escapes gave rise to warlike quilombos, where the power of negotiation and resistance became a concrete alternative to the slave order. This threat was systematically fought by the official The educational model, historically constructed in Brazil, is directly responsible for the silencing of these contributions, as well as for the dissemination of the idea of passivity and conformism. With regard to indigenous peoples, the memory silenced and succumbed by the violence of the colonizing process, continues as a central element for the understanding of the ethnic-racial issues that allow to identify the meanings in the discourses that base the national curricular matrices in the present time. In this sense, it is necessary to identify the social place attributed to indigenous peoples in this process.
The place of indigenous peoples
According to Santos (2006) for example, inhabited where today is the coastal part of the national territory, formed a myriad of distinct groups without, however, constituting themselves as a political unit.
Throughout the process of consolidation of Portuguese colonization, different ethnic groups warred and confronted Europeans, demanding in their turn great efforts to achieve domination. The resistance to imposition and domination imposed by the invader marked the process of conquest and is still present in the Brazilian territory, as Darcy Ribeiro (1995, p. 34) points out in stating that "this is what is happening today, five hundred years later, with the Yanomami of the northern border of Brazil. " The fiction of the unified territory, as a legitimating presupposition of the State, and composed of a single great people, was the modern basis that would justify the greatest atrocities in the name of territorial unity, tradition and the great Brazilian nation. For Brazil, which carries in the history of the massacres of colonization and the slave model as the central base of the economy and social organization, it remains a problem to define the concept of Brazilian people, especially for the dominated and silenced of history. The destruction of ethnic minorities, which threatened national unity, was the object of practical actions by institutional power concerned with the nation-state, while it was ignored for years by scholars of the human sciences, who prioritized the class clipping to the detriment of ethnic minorities, treated as a second-order issue. According to Santos (2006), in view of this context, attention is drawn to the actions taken by the Brazilian power to protect indigenous peoples, who are Empire and then to the Republic, the Indian oscillated between a model of Brazilianness as an example of delay and annoyance to national "development." One possible explanation for such variability lies in the question of labor, after all, indigenous peoples were not central to the consolidation of the colonial production system. At that time, excluding possible imprisonments, the indigenous arm was marginalized in the economic sectors during the colonization. In this context, Oliveira draws attention to the ambiguous role attributed to the Indian in our history, because, faced with the indigenous "problem", the indigenous question should not be dissociated from the structural aspects of the nation that need confrontation, such as: corruption, impunity and other social ills that mark the history of the country. This "problem", on the other hand, must be regarded as a persuasive force because: When speaking of an indigenous problem, the impression of most listeners is that this is only a common expression, without any more precise content, a simple act of pointing, without anticipating any interpretation on the subject. Things, however, do not happen this way: what often appears as innocent creation of common sense, in reality constitutes a powerful instrument of unification of ideas, facts and domains, that otherwise would not be seen as related of necessary or even habitual 5 . (OLIVEIRA, 1995, p. 64) According to the author, in this perspective, the category "Indian" must be understood as the creation of the invader, in order to generalize and "domesticate" the myriad groups that inhabited the territory we now call Brazil. The transition in the country to peoples who recognized themselves as ethnic groups, such as the Xavantes, Bororo, among others, to the uniform identity of "Indians", implies new forms of recognition, resistance and differentiation of the European invaders and subsequent construction of a nationality unified. All these forms must be considered when thinking about the place of the indigenous in the national imagination. In this process, indigenous peoples become understood as a minority that appropriates and builds a new collective identity, of indigenous peoples, who struggle to access rights and political achievements within the State and public policies. This identity resource continues to mark the relationship based on the tutelage of indigenous peoples in an 5 Original text: Quando se fala em problema indígena, a impressão da maioria dos ouvintes é que isso constitui apenas uma expressão corriqueira, sem qualquer conteúdo mais preciso, um simples ato de apontar, sem antecipar qualquer interpretação sobre o tema. As coisas, no entanto, não se passam dessa forma: o que frequentemente aparece como inocente criação do senso comum, na realidade se constitui em um poderoso instrumento de unificação de ideias, fatos e domínios, que de outra forma não seriam vistos como relacionados de modo necessário ou mesmo habitual. elitist and authoritarian way, becoming an element of great tensions and violence in the countryside and in the great Brazilian cities. Historically, guardianship sought to consolidate an official political practice, which acts to avoid, or at least disrupt, the coordination of groups for collective action. This process, in turn, would help to overcome their fragility in the face of power. Thus, the indigenous battle develops in the state arena, benefiting from the political divide, consisting of a technique in which the different groups learn that they are Indians, discover that they have rights and seek to find group ways to consolidate these rights.
The In this context, the colonization process, imposed on the American continent, must be emphasized as a permanent deployment of Latin America in which, according to Galeano, everything became capital for the Europeans. The characteristics of this domain imposed a historical cost to the nations and ethnic groups that inhabited the usurped territory. In this sense, beyond Brazil, domination represented an impact that lasts for much of the American continent as a cost to the consolidation of world capitalism, as highlighted by the author: For those who view history as a dispute, the backwardness and misery of Latin America is the result of its failure. We lost; others won. But it turns out that those who won won thanks to what we lost: the history of the underdevelopment of Latin America integrates, as has already been said, the history of the development of world capitalism. Our defeat was always implicit in the victory of others, our wealth has always generated our poverty to feed the prosperity of others: the empires and their native agents. In colonial and neocolonial alchemy, gold turns into scrap and food becomes poison. Potosí, Zacatecas, and Ouro Preto fell from the top of the splendors of precious metals to the bottom of empty voids, and ruin was the fate of the Chilean pampas of saltpeter and the Amazon jungle of rubber; the northeastern sugarcane plantations in Brazil, the Argentine jungle forests, or some oil-producing settlements in Maracaibo have painful reasons for believing in the mortality of the fortunes that nature bestows and imperialism usurp. The rain that irrigates the centers of imperialist power drowns the vast suburbs of the system. In the same way, and symmetrically, the well-being of our dominant classes-domineering inwards, dominated from without-is the curse of our multitudes, condemned to a life of beasts of burden 6 . (GALEANO, 1992, p. 5-6) As much as the imposition In short, it can be seen that the ethnic-racial question is a determining condition for the understanding of the formation of the Brazilian homeland, as well as the place historically conditioned to the indigenous as part of this debate. In a territory projected as a "great consolidating the ideal of a racist and Eurocentric miscegenation , to think of the newly proclaimed country as a Federative Republic.
The process of abolishing slavery and the transition to the republic
The It is up to this moment to reinforce the active aspect of the Negro and the indigenous in the fight for their own freedom. The multiple revolts, the formation of solidary networks among freedmen, the small sabotages in the productive units, and the other actions of resistance of that period should be emphasized, so as not to run the risk of imputing to the kindness of some the abolition of 1888. Thus, the combination of some external factors, such as the influence of liberal thought on the largely urban elites, the changes in the capitalist production model that reorganized the world of labor and world consumption, the condemnation of the great capitalist powers against the slave trade, helps to understand the spread of abolitionist ideals throughout the Brazilian territory in the late 1880s. There was no plausible justification for maintaining the slave system, or for support of the imperial government. According to Skidmore (1976), this transition sought, once again, to maintain social and political control of the country's economic elites without altering the asymmetries and privileges of the ruling classes.
But Brazil came to the formal abolition of slavery slowly, without a bloody civil war.
Controllers of the policy have taken every precaution to stifle the impending social conflict, using legal parameters. Another factor that contributed to a less traumatic passage for the process, in turn, did not fail to reproduce, in the intellectual elite of the country, a racist view on the inferior condition of the Negro as emphasized by Skidmore (1976): This was the conception that reigned in the elite: Brazil had escaped the prejudice of color. As Nabuco wrote in O abolicionismo: "Slavery, for our happiness, has never sourced the soul of the slave against you, speaking collectively, nor created between the two races the reciprocal hatred that naturally exists between oppressors and oppressed." Moreover, recent experiences have shown that "color in Brazil is not, as in the United States, a social prejudice against whose obstinacy the character, talent, merit of the person who incurs it cannot do so much." In contrast to what occurred in the United States, abolitionists in Brazil were rarely obliged to discuss the question of race itself, because advocates of slavery practically never turned to racial inferiority theories 7 . (SKIDMORE, 1976, p. 62) In this context, it seems clear that many abolitionists defended the theories of European superiority and racial inferiority, which would serve as justification and ideological force for the defense of the laundering of the "Brazilian nation", and the consequent marginalization and determination of a place of inferiority for blacks and in the construction of a new imaginary country. This marginal role, in turn, would mark the consolidation of the myth of Brazilian racial democracy by inserting the idea of miscegenation as a positive, valued element that has maintained, however, the specificity of the Negro and the indigenous in secondary conditions.
The originality of Brazilian racial thinking
The understanding of the role of racial theories in the transition from the nineteenth to the twentieth century in Brazil cannot be reduced to the influence of European and American theories on white racial superiority. This is because racialist thought in this period has originality that lies in the discarding of eugenicist theories as necessary, in the adaptation and inclusion of explanations that would justify the racial question as central to the social explanation of the country. In a country mixed as Brazil, the simple transfer of racialist theories would be impracticable. It was necessary to create something new and to adapt the perspective of superiority and racial inferiority in this context. These scientists believed they carried the responsibility of the destiny of the nation.
The scientific thought of the period did not seek methods of experimentation that proved some theory, according to experimental model prevailing in Europe and the United States. The scientific activity was characterized by debates and discourses valued socially by the national literate elites, with strong influence of the social-Darwinism and the evolutionist model, popularized as theoretical justification of imperialist domination in these regions.
These men of "sciences" 8 did not form a cohesive theoretical unity. These national racialist interpreters argued that African inferiority should not even be compared to "higher" races. As a common point, the authors sought theoretical justifications to prove that lower groups, such as former slaves, were unable to become part of a new civilization. The justification for the national backwardness was, in the current thinking, tied to the presence of the Negro in the formation and history of Brazil.
In spite of their differences, these thinkers had the stamp of the institutions they represented -São Paulo School of Law, Faculty of Medicine in Bahia, lyceums and art museums in Rio de Janeiro, press offices in São Paulo, Rio de Janeiro and Recife -the common objective in the treatment of the great scientific subjects, in the light of the national questions. These self-taught people consumed manuals and theories, such as positivism, social-Darwinism, and evolutionism, often outdated in Europe. These intellectual activities intended to explain Brazil from the racial perspective, in a society marked by miscegenation and miscegenation, constructed in violent ways, according to the characteristics of the slave model that prevailed in the country.
We can briefly divide two great theoretical perspectives that marked Western thought in the eighteenth and nineteenth centuries: the French perspective, which naturalized human equality; and the view of superiority and capacity of certain genetic types as determinants and characteristics of "higher" moral and intellectual standards. The perspective that sought to naturalize the differences prevailed among Brazilian "men of sciences".
The fact that Brazil was a "mestizo nation" by some of these thinkers brought new dilemmas in the attempt to reconcile the ideals of grandeur and national identity with a mixed and diverse people like Brazil. Equality, as advocated in the new Constitution of 1891, found racialist discourse as an important barrier to the social ascension of the excluded, as Schwarcz (1993) observed: The same context that finds in a liberal project the solution to its new political configuration looks for in the deterministic and anthropological theories subsidies to transform social differences in fundamental biological barriers. After slavery and establishment and democracy through the Republic, it takes force a racial discourse, late compared to the liberal model present since 1822 9 . (SCHWARCZ, 1993, p. 240) The abolition followed by the Proclamation of the Republic has awakened the prospect of a great country. The negotiated transition among economic elites, however, had a major problem. What to do with the mass of free workers facing this new context? How to achieve order by principle and progress at last in the face of so many conflicts and social instability? These and other questions aroused great debate, since the foundation that would justify evolution for a large "Brazilian nation" was supported by the racialist and supposedly scientific discourses of eugenics. Part of the intellectual elite was assured that abolition would not threaten the economic and social hegemony of the ruling classes. According to the author, extinguished slavery would systematically open the path to internalization or recognition of the social hierarchy as an inevitable and deterministic phenomenon.
The immigrant turn-of-the-century project posed itself as the sure justification for building a great nation. In opposition to the reformist thinkers, who admitted the inclusion of blacks in the free labor market, the defenders of laundering, through the insertion of the European in the territory, believed that this would be the guarantee of times of progress. On this aspect, Azevedo (1987) emphasizes: This passage, or transition, was conceived as an orderly time of gradual overcoming of serious racial-social problems in which a set of tactics of control and discipline would be applied in order to attain in the future the much-dreamed time of progress. In the meantime, it was hoped that the country could fill a basic need pointed out by almost all authors already seen so far: nationality. For this, it was necessary to forge a population fully identified with the idea of a homeland, of Brazilian society, not only in terms of geographical limits, but especially in the sense of a national ethic. However, the perception of an explosive socio-racial heterogeneity stands out as a considerable hindrance in the thinking of those who sought to transform the newly independent country into a nation 10 . (AZEVEDO, 1987, p. 60). According to Skidmore (1976), the thesis of innate racial differences counted on three great schools: (a) ethnological-biological, of American origin, that defended the supposed inferiority of the blacks and indigenous people in face of the physical differences in relation to the white; (b.) the historical school, which attributed to the successive triumphs of different peoples who understood themselves as superior; and (c.) social Darwinism, used by racist de se atingir no futuro o tão sonhado tempo de progresso. Nesse meio tempo esperava-se que o país pudesse preencher uma carência básica apontada por quase todos os autores já vistos até aqui: a nacionalidade. Para isso, era preciso que se forjasse uma população plenamente identificada com a ideia de pátria, de sociedade brasileira, não só em termos de limites geográficos como, principalmente, no sentido de uma ética nacional. Contudo, a percepção de uma explosiva heterogenia sócio-racial destaca-se como um considerável entrave no pensamento daqueles que almejavam transformar o país recém-independente em nação. discourses such as the naturalization of "stronger" peoples, who would condemn the disappearance of "inferior" peoples in an irreversible process.
RPGE-
In the period in question, racially concerned scholars were strongly influenced by these three racialist currents, condemning the Negro to extinction or subjugation by the "higher races." The mestizos, on the other hand, obtained in the negation of their own blackness or indigenous blood a real possibility of ascension and social acceptance in elitist circles that had few representatives of that mixture at the top of the social and political hierarchy.
The stimulus to free immigrant workers came accompanied by the marginalization of the black in the process of building the Republic. In this context, racial discourses were directly linked to the idea of nation. A paradox present in this discourse was the very mixed composition of the Brazilian population; the reduced academic universe would need to justify theories of racial superiority without, however, denying national miscegenation. The racial argument would play an important role in maintaining social inequalities and in the lack of access to the rights and citizenship of the excluded layers in the process of construction of the country. Another common justification in this context, for the exclusion of blacks and the encouragement of immigrants, was that blacks, as well as Indians, were lazy, vagabond and with tendencies toward social marginalization.
Conditioned on marginalization, black became an ideal criminal type for racist elites of the period. Immigrants, in turn, became the example and the model to be followed, which could reconcile work and freedom, for the money laundering project proposed in the period.
About the myth of Brazilian racial democracy
In the 1930s, racial themes continued to be central to Brazilian social thought. It is important to point out, however, an important change that would displace miscegenation as a problem or obstacle to the construction of a great country for the positive anthropological field of the search for a national identity, centered on the mixture of whites, blacks and Indians. The biological factor of the miscegenation was transplanted to the social field as an example of harmony and coexistence between the Brazilian population diversity. The rootedness of the myth of peaceful coexistence is the fuel for the "great" civilizing process intended during the Vargas era. Whitening as a dominant ideology remained on the fabric of the mixture that acted as the negation of the black identity in that context, as Clóvis Moura (1988) observes: Whitening as the ideology of power elites will be reflected in the behavior of much of the dominated segment that begins to flee from its ethnic matrices, to mask itself with the values created to discriminate against it. With this the Negro (the mulatto, therefore, too) did not articulate at the level of an ethnic identity consciousness capable of creating a neutralizing counter-ideology manipulated by the dominator. On the contrary. There is a process of accommodation to these values, a fact that will determine the emptying of these Negroes at the level of their ethnic consciousness, thus placing them as simple objects of the historical, social and cultural process 11 (MOURA, 1988, p. 69-70) Racial democracy serves as a justification for a new exclusion, which seeks to keep the exploitation of the black hidden in the name of the friendly coexistence that keeps it in the impossibility of social ascension. From this conception, it emerges that class and race have never been established in the same way, after the dismantling of the slave society and the mode of slave production according to Florestan Fernandes (1989). standards. This model of racism is typically Brazilian, that is, it is not confronted, exposed, but hidden, silenced. Acting in multiple spheres of Brazilian social and institutional relations, Brazilian racism is described, according to Nascimento (1978): We must understand "racial democracy" as meaning the perfect metaphor for Brazilian-style racism: not as obvious as United States racism, nor legalized as apartheid South Africa, but effectively institutionalized at official levels of government as well as diffuse in the social, psychological, economic, political and cultural fabric of the country's society. From the crude classification of blacks as savages and inferiors, to the exaltation of the virtues of blood-mixing as an attempt to eradicate the "black stain", from the operability of religious "syncretism"; to the legal abolition of the black issue through the National Security Law and census omission -manipulating all these methods and resources -Brazil's unofficial history records the long and ancient genocide perpetrated against Afro-Brazilians. Monstrous machine ironically designated "racial democracy" that only gives blacks a single "privilege": that of becoming white, inside and out. The password of this imperialism of whiteness, and of capitalism inherent to it, responds to bastard nicknames like assimilation, acculturation, miscegenation; but we know that beneath the theoretical surface remains the belief in the inferiority of the African and his descendants 12 . (NASCIMENTO, 1978, p. 93) Belief in the harmony between blacks, whites and natives acts as a barrier that hinders the antiracist struggle in Brazil. The fact that the concept of race was abandoned by scientists in the mid-twentieth century does not eliminate racism in Brazil's social relations, institutions, politics, and mentality. According to Munanga (1999), the struggle against racial discrimination continues to be hampered during the first half of the twentieth century, through the built division of blacks and mestizos as separate groups devoid of unity, which hindered the identitary recognition of both. The manipulation of the biological by the ideological through the cognitive categories that sought to name and signify society in a great unity ignores and devalues diversity, making difficult the understanding and the historical dimension of this process of national construction that marked this period. The search for 12 Original text: Devemos compreender "democracia racial" como significando a metáfora perfeita para designar o racismo estilo brasileiro: não tão óbvio como o racismo dos Estados Unidos e nem legalizado qual o apartheid da África do Sul, mas eficazmente institucionalizado nos níveis oficiais de governo assim como difuso no tecido social, psicológico, econômico, político e cultural da sociedade do país. Da classificação grosseira dos negros como selvagens e inferiores, ao enaltecimento das virtudes da mistura de sangue como tentativa de erradicação da "mancha negra", da operatividade do "sincretismo" religioso; à abolição legal da questão negra através da Lei de Segurança Nacional e da omissão censitáriamanipulando todos esses métodos e recursosa história não oficial do Brasil registra o longo e antigo genocídio que se vem perpetrando contra o afro-brasileiro. Monstruosa máquina ironicamente designada "democracia racial" que só concede aos negros um único "privilégio": aquele de se tornarem brancos, por dentro e por fora. A palavra-senha desse imperialismo da brancura, e do capitalismo que lhe é inerente, responde a apelidos bastardos como assimilação, aculturação, miscigenação; mas sabemos que embaixo da superfície teórica permanece intocada a crença na inferioridade do africano e seus descendentes. problematizing the brutality of miscegenation, as the author points out: Portuguese does not: for all those happy predispositions of race, of mesology and of culture to which we refer, not only did it manage to overcome the conditions of climate and of soil unfavorable to the establishment of Europeans in the tropics, as to supply the extreme penury of white people for the colonizing task uniting with woman of color. By intercourse with Indian or black women the colonizer multiplied in a vigorous and ductile mestizo population, even more adaptable than he was pure to the tropical climate. The lack of people, who afflicted him, more than any other colonizer, forcing him to the immediate miscegenation -against which they did not have, besides, scruples of race, only religious preconceptions -was to Portuguese advantage in his work of conquest and colonization of the tropics. Advantage for its better adaptation than biological, social 13 . (FREYRE, 2005, p. 74-75) This classic text, written in the 1930s, helped to construct an ideology of Brazilian culture that excluded class conflicts, inherent to the slave model, because it confused the concepts of culture and race, presenting a mythical and idealized version when dealing with relations between opposites, slaves and masters, as something devoid of conflict between different. The analysis of this author did not seek to understand the organization of power in the different performances in which it was exercised, characteristics of the patriarchal society.
This contribution consolidated a harmonious perspective, which would make it possible to sew the desired quest for a national unity, as Munanga points out (1999): Freyre's great contribution is to have shown that blacks, Indians and mestizos made positive contributions in Brazilian culture; profoundly influenced the lifestyle of the lordly class in matters of food, clothing and sex. Miscegenation, which in the thought of Nina and others caused irreparable damage to Brazil, was seen by him as an immense advantage. In other words, by transforming mestizaje into a positive, non-negative value under the aspect of degeneracy, the author of Casa grande e senzala allowed the definitive completion of the contours of an identity that had long been drawn. Freyre consolidates the original myth of the Brazilian society configured in a triangle whose vertices are the black, white and Indian races. This is how the mixtures came about. The three races also brought their cultural heritages parallel to the racial crossings, which gave rise to another mestizaje in the cultural field. From the idea of this double mix, the myth of racial democracy slowly grew; "We are a democracy because the mixture generated a people without a barrier, without prejudice" 14 . (MUNANGA, 1999, p. 79-80) The documents, such as the curricular guidelines of 2004, Law 10.639, among others that deal with ethnic-racial relations in Brazilian education, identify, in this process of silencing of blacks and indigenous people, one of the central arguments for curricular change, historically, reproduces the idea of mixing as a factor of recognition and inclusion of these segments in the rhetoric of contribution to the construction of Brazil. The organized reaction, which sought to denounce this condition, was marked by the foundation of several initiatives for the reconstruction of a new image for the Negro, as in newspapers, in the experimental theater of the Negro and, in 1931, in the "Black Front", being the education its main field of activity. In these movements, there were simultaneously ambiguous aspects of the black struggle, since the educational model of the period was marked by the white and Eurocentric conception of education. The African element was discarded and omitted in search of an acceptance that obeyed criteria consolidated and embedded in the hegemonic white mentality of the period. According to Kabengele Munanga, the denunciation against discrimination, violence and prejudice coexisted with the denial of African cultural identity, seen as inferior by the descendants themselves.
14 Original text: A grande contribuição de Freyre é ter mostrado que negros, índios e mestiços tiveram contribuições positivas na cultura brasileira; influenciaram profundamente o estilo de vida da classe senhorial em matéria de comida, indumentária e sexo. A mestiçagem, que no pensamento de Nina e outros causava dano irreparável ao Brasil, era vista por ele como uma vantagem imensa. Em outras palavras, ao transformar a mestiçagem num valor positivo e não negativo sob o aspecto de degenerescência, o autor de Casa grande e senzala permitiu completar definitivamente os contornos de uma identidade que há muito vinha sendo desenhada. Freyre consolida o mito originário da sociedade brasileira configurada num triângulo cujos vértices são as raças negra, branca e índia. Foi assim que surgiram as misturas. As três raças trouxeram também suas heranças culturais paralelamente aos cruzamentos raciais, o que deu origem a uma outra mestiçagem no campo cultural. Da ideia dessa dupla mistura, brotou lentamente o mito de democracia racial; "somos uma democracia porque a mistura gerou um povo sem barreira, sem preconceito". The recovery of the African elements, the recognition of this heritage and the overcoming of the racial democracy myth are central arguments in the contemporary political struggle that arises in the disputes involving the construction of the BNCC. Identity, understood as a process negotiated and renegotiated by ideological and political disputes, according to the established context and power relations, becomes the driving force of the disputes around ethnic-racial issues and the Brazilian education system. In a context of disputes and conceptions that seek to diverge from the commercial consensus imputed to the national curricular model, the BNCC is the scene of disputes that defend the multifaceted and plural recognition of the Brazilian people and its right to memory and to multiple national identities. In this clipping, we can understand the importance of the founding myth and racial democracy as theoretical foundations that today seek to be deconstructed by the curricular guidelines that deal with ethnic-racial relations in the documents that organized Brazilian education historically and which, today, underpin the dispute and the discussions on the feasibility and relevance of a BNCC for the country.
In order to analyze the symbolic meanings present in the BNCC discourse, it becomes necessary to understand the field about curriculum. In spite of their disputes and disputes, ethnic-racial issues find their point of interest and search for centrality in the post-critical perspective of the curricular field in this proposal that intends to establish a national standard for the organization of Brazilian basic education. Understanding the history of the curriculum as a field of study, identifying the different perspectives, such as technicalist, critical and post-critical visions, is necessary for the understanding of ethnic-racial relations and the disputes for meanings that characterize this analytical gaze for the curricular field.
Final remarks
As I went through the construction of the idea of a homogeneous and grandiose nation in the crossing of the nineteenth century to the XX in Brazil, as well as the trajectory of struggles and resistance on the part of the Brazilian blacks and Indians, I was able to perceive the importance and the presence of this debate that seeks deconstruct the founding myth and the myth of "Brazilian racial democracy" as a central condition for any educational project of national character. Shifting the rhetoric and the discourse of the field of tolerance and respect to the area of confrontation and conflict of narratives and ideas is fundamental to advancing in the construction of a more egalitarian and less violent society. | 2019-06-01T13:16:46.801Z | 2019-01-02T00:00:00.000 | {
"year": 2019,
"sha1": "e6405992bec4321bff9828ff6396729ad119e226",
"oa_license": "CCBYNCSA",
"oa_url": "https://doi.org/10.22633/rpge.v23i1.11506",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "19e2355f702cbf442e8127c30c4b6211240e4b1a",
"s2fieldsofstudy": [
"Education",
"History",
"Political Science"
],
"extfieldsofstudy": [
"Political Science"
]
} |
246863963 | pes2o/s2orc | v3-fos-license | Theoretical Analysis of SIRVVD Model to Provide Insight on the Target Rate of COVID-19/SARS-CoV-2 Vaccination in Japan
The effectiveness of the first and second dose vaccinations are different for COVID-19; therefore, a susceptible-infected-recovered-vaccination1-vaccination2-death (SIRVVD) model that can represent the states of the first and second vaccination doses has been proposed. By the previous study, we can carry out simulating the spread of infectious disease considering the effects of the first and second doses of the vaccination based on the SIRVVD model. However, theoretical analysis of the SIRVVD Model is insufficient. Therefore, we obtained an analytical expression of the infectious number, by treating the numbers of susceptible persons and vaccinated persons as parameters. We used the solution to determine the target rate of the vaccination for decreasing the infection numbers of the COVID-19 Delta variant (B.1.617) in Japan. Further, we investigated the target vaccination rates for cases with strong or weak variants by comparison with the COVID-19 Delta variant (B.1.617). This study contributes to the mathematical development of the SIRVVD model and provides insight into the target rate of the vaccination to decrease the number of infections.
Introduction
As of December 2021, the COVID-19 pandemic continues to be a global concern. Telework [1], lockdown [2], airport quarantine [3], and tracing apps [4] are used as countermeasures for overcoming COVID-19 spread. The most important countermeasure is the COVID-19 vaccine (e.g., BNT162b (Pfizer) [5], mRNA-1273 (Moderna) [6], and ChAdOx1 (AstraZeneca) [7], and others). Mathematical simulation is a desirable approach to gain insight into the effect of vaccination on the infectious spread. COVID-19 country-based simulations (Malaysia [8], Saudi Arabia [9], Spain [10], United States [11], and others) have already been reported. A fast simulation method of differential equations is a susceptible-infected-recovered-vaccination (SIRV) model [12][13][14][15][16]. However, these models can represent only the first dose of the vaccination. For COVID-19, the effects of the first and second doses on infectious prevention are different; for the COVID-19 Delta variant (B.1.617), the effects of the first and second doses on infectious prevention are 35.6% and 88.0%, respectively [17]. In this context, the SIRVVD model that represents the first and second doses of vaccination was proposed [18]. However, theoretical analysis of the SIRVVD Model is insufficient. Therefore, we find an analytical expression of the infectious number based on the SIRVVD model, by treating the numbers of susceptible persons and vaccinated persons as pa-Figure 1: SIRVVD model [18] rameters.
In order to stop the spread of infection, it is important to achieve herd immunity through vaccination. Some previous studies have reported various target values for achieving herd immunity through vaccination [19][20][21][22][23]. Because the target vaccination rate to achieve herd immunity depends on the country's lifestyle (wearing masks and social distancing) [24], it is important to obtain target values of vaccination for each country. Therefore, we calculated the target rate of the vaccinated persons in the scenario of spreading the COVID-19 Delta variant (B.1.617) in Japan based on the solution of infectious numbers of the SIRVVD model. In addition, we investigated target rates in the case of becoming strong or weak variants compared with the COVID-19 Delta variant (B.1.617).
SIRVVD model 2.1 Differential equations
The SIRVVD model [18] for representing the first and second doses of vaccination is defined aṡ where the dot notation represents the time derivative, i.e.,ḟ (t) := df (t)/dt. Further, S(t) represents the number of susceptible persons; I j∈{0,1,2} (t) represents the number of infected persons of j times the dose of the vaccination (j = 0 means no vaccination). V j∈{1,2} (t) represents the number of the vaccinated persons of j times the dose of the vaccination. R(t) represents the number of recovered persons, and D(t) represents the number of dead persons. Moreover, the total infection number I(t) is defined as Then, we describe the coefficient parameters as follows: β(t) represents infectivity, θ −1 j∈{0,1,2} represents the average antibody period by the j times the dose of vaccination (j = 0 implies the natural infection). α j∈{1,2} represents the transition rate to V j (t); σ j∈{1,2} represents the effectiveness of j times the dose of vaccination (decreasing the value of infectious probability). γ −1 denotes the average infection period, and δ j∈{0,1,2} denotes the fatality rate of the persons of j times the dose of vaccination. The states transitions are presented in Figure 1. The infection probability and fatality rate of the vaccinated persons V j∈{1,2} decrease because of the vaccination effects. Then, the total population N is where R + 0 represents the set of plus real numbers that include zero.
Analytical expression of the infection number
It is desirable that the function of the infectious number I j∈{0,1,2} (t) is clear to investigate the infectious spread. However, it is difficult to obtain an analytical solution in the absence of any assumptions be-causeİ j (t) is complex (including various functions). Therefore, we assume that S(t), V 1 (t), V 2 (t), and β(t) are not time-dependent functions but the constants of S, V 1 , V 2 , and β, respectively. Then, the total number of persons that have an infectious possibility N is given by If the infectious number is sufficiently small, N N . I j (t) defined by Equations (4)-(6) can be simplified as where Herein, if we define then, Equation (10) can be simplified intȯ as the first-order separable simultaneous differential equations. Here, we transform it into 1 I(t) dI(t) = A dt, and then, we can represent where U represents the diagonal matrix of A, and these relationships can be represented by where P comprises the eigenvectors of the coefficient matrix A, and λ j∈{0,1,2} represents the eigenvalues. In other words, the solution I(t) is clear by calculating the eigenvectors and eigenvalues of the coefficient matrix A.
To be clear, when λ j∈{0,1,2} , let a 3 × 3 unit matrix be denoted by E; we calculate In addition, the eigenvalues λ are obtained by solving the eigenequation det(A − λE) = 0 as From the eigenvalues, P comprises eigenvectors that can be represented as where The solution I(t) of the differential equationİ(t) can be represented by substituting Equations (12)- (14) into Equation (11). Then, solutions I j∈{0,1,2} are represented as The coefficients are given by where, These equations require constant values of S(t), V 1 (t), and V 2 (t); thus, these equations are not suitable for long-term simulations.
Target rate of the vaccination
γ represents the inverse value of the average infection period, and therefore, γ > 0. Thus, lim t→∞ e −γt = 0.
Further, because the coefficients of Equation (15) are constants, each solution I j∈{0,1,2} (t) can be approximated by S was deleted by Equation (9). Therefore, we get the following theorems: where, j ∈ {0, 1, 2}. The infectious spread stops by creating a scenario as indicated in Equation (17). We collected terms on virus parameters to the right side and vaccination parameters on the left side to clarify this scenario. As a result, Equation (17) can be simplified into Here, N denotes the total number of persons with a probability of infection; σ j∈{1,2} represents the effectivity of the vaccination; and V j∈{1,2} represens the number of vaccinated persons. Thus, the left side represents the total number of people who have the probability of infection considering the vaccination (the left side is small when using highly effective vaccination). The longer the infection period and the higher the infectivity, the smaller is the right term because γ −1 represents the average infection period and β represents the infectivity. In addition, in the case a virus has a long infection period and high infectivity, a highly effective vaccination is required to stop the spread of infection.
Here, let V obj.
2 denote the target number of the second dose of vaccination required to stop the spread of infection. We can obtain V obj.
Further, the target rate of the second dose of vaccination required to stop the infection spread P obj.
Limitation of the vaccination
We consider the condition of not stopping the infection spread even if all persons receive the second dose of vaccination. We obtain the following theorem because this scenario would mean substituting V 2 = N , V 1 = 0 for Eq. (16).
Even if all persons receive the second dose of vaccination, the infection number continues to increase when the effectivity of the second dose of the vaccination σ 2 is insufficient. In contrast, the condition of Equation (20) is not satisfied when the vaccination that has perfect prevention (i.e., σ 2 = 1) because N > 0, β > 0, and γ > 0. In other words, the spread of infection stops when developing a vaccination that has a perfect prevention effect (i.e., σ 2 = 1) and getting a vaccination to all persons (i.e., V 2 = N ). However, it is important to know the condition of Equation (20) because achieving this is difficult.
Conditions under which the vaccine is not required
We consider conditions under which the vaccine is not required; this implies substituting V 2 = 0, V 1 = 0 for Equation (18). We can obtain a simplified equation as 2.6 Validity of the target number V obj.
2
Here, we check whether the infection spread stops by satisfying V obj.
2 } over the initial number of persons who get the second dose of vaccination V 2 to verify the cases that satisfy V obj.
.
The simulation results for cases V 1 = 0 and V 1 = N/2 are shown in Figures 2 and 3, respectively. The horizontal axis represents V 2 ; the vertical axis, I(t end )/I(0). In addition, the black dashed line represents I(t end )/I(0) = 1; the red dashed line, V 2 = V obj.
2
. I(t end )/I(0) > 1 indicates that the infection number increases, and I(t end )/I(0) < 1 implies that it decreases. In all cases, when V 2 = V obj.
Infectivity of the COVID-19 Delta variant (B.1.617) in Japan
Let β delta denote the infectivity of the COVID-19 Delta variant (B.1.617) in Japan; we estimate this value. According to previous research [25], infectivity β can be represented as where r denotes the effective reproduction number. Further, β calculated by Equation (22) [27]. Therefore, we consider that the infectivity of the Delta variant (B.1.617) is underestimated owing the effect of the vaccination when considering the newer cases of infection during this period. Thus, we estimated the infectivity of the Delta variant (B.1.617) in Japan using data from the period when the vaccination was not started (before February 2021). According to the COVID-19 advisory board for the Japanese government [28], the Delta variant (period: August 2021) had a 1.9 times higher infection rate compared to the that of the regular COVID-19 virus (period: December 2020) in Japan because no people were vaccinated in December 2020. Thus, we consider the infectivity of the Delta variant (B.1.617) as the 1.9 times the infectivity of this period. We assume where β delta and β regular represent the infectivity of the Delta variant (B.1.617) and regular COVID-19 virus, respectively. The maximum value of the effective reproduction number in December 2020 in Japan was r = 1.18 [29]. [19] World 82.5% 45 -60 % Gumel [20] US 70% 80 % Gumel [20] US 95% 60 % Kadkhoda [21] -100% 62 -72 % Kadkhoda [21] -95% 63 -76 % Fontanet [22] France -67 % Liu [23] China 90% 83 -92% Ours Japan** 88% 63% *: Prevention effect of vaccine. **: The case of lifestyle on Dec. 2020 in Japan.
Kobayashi et al. estimated 0.17 as the γ parameter of COVID-19 in Japan from 0.13 to 0.17 [30]. We adopted γ = 0.13 because we consider optimism to be undesirable, and this leads to a long infection period. The total population in Japan is approximately S = 1.2×10 8 persons. We obtain β regular 1.28×10 −9 as the infectivity of the COVID-19 regular virus by substituting these parameters in Equation (22). Moreover, the infectivity of the Delta variant (B.1.617) is β delta 2.43 × 10 −9 by substituting β regular for Equation (23) and by substituting β regular .
Target number V obj.
2 and target rate P obj.
in Japan
We consider the target number V obj. 2 and target rate P obj. 2 in Japan for the infectious spread of the Delta variant (B.1.617). We assume that all persons who received the first dose will receive the second dose, i.e., V 1 = 0. Further, the total population in Japan is about 1.2 × 10 8 . V obj. can be represented by substituting these values for Equation (19) as The target rate of 63% is for the case of the lifestyle of December 2020 in Japan. If we obtain the lifestyle values before the appearance of COVID-19 (before December 2019), a target rate of about 63% will appear to be insufficient.
Snehota et al. [31] reported that about 72.5% of the population in various countries have taken the COVID-19 vaccine. We believe that the target vaccination rate can be achieved in Japan because the target rate P obj. 2 = 0.63 is lower than the reported global value of 72.5%. In Japan, on December 27, 2021, 78% of the total population received the second dose of vaccination [27], and the infection number decreased (e.g., the average number of new cases of infection per day from November 1, 2021, to December 31, 2021 was lower than 3 persons per million persons [32].).
Comparison with other studies
The target rates of the vaccinated persons reported by other researchers are listed in Table 1. These are target rates for achieving herd immunity to stop infection spread. We check the various values in this study.
One reason for the different values is that the target rate is dependent on the lifestyle of each country. For example, Shen et al. [24] reported that the target rate of vaccination is dependent on wearing masks and social distancing.
The target rate of vaccinated persons to stop the infection spread is not the same because the lifestyles of each country are not the same. Thus, it is necessary to calculate the target rate of each country. P obj. 2 = 0.63 represents the target rate for the lifestyle in December 2020 in Japan.
The target rates P obj.
2 in the case of perfect prevention (σ 2 = 1) are shown in Figure 4. The vertical axis represents the average infection period γ −1 [day], and the horizontal axis represents the infectivity β (1.0 × β delta represents the Delta variant infectivity). The percentage values are P obj. However, the perfect prevention vaccine is not realistic. Therefore, we calculate the case of σ 2 = 0.88, as shown in Figure 5. Painted black markers represent the results satisfying Equation (20), i.e., even if all persons receive the second dose of vaccination, the infection spread does not stop. This scenario occurs in the cases of long infection periods and high infectivity. For such a scenario, we consider not only vaccinations but also lockdowns, such as stay-at-home orders.
Finally, the case of a new variant that leads to a decrease in the vaccination effect (σ 2 = 0.70) is shown in Figure 6. There are many cases wherein only the vaccination cannot stop the spread of the infection caused by a new strong variant. The results indicate that it is important to develop more high-effect vaccines to overcome the new strong variant.
Conclusion
We found the analytical expressions of I j∈{0,1,2} (t) of the differential equationsİ j∈{0,1,2} (t) of the SIRVVD model. Furthermore, we proposed a method to determine P obj.
2
, which is the target rate of the vaccination required to stop the spread of infection. Assuming the COVID-19 Delta variant (B.1.617) in Japan, we estimated P obj. 2 to be approximately 63%. Further, we calculated the target rate of the vaccination by assuming a new strong or weak variant (Figures 4, 5, and 6). We consider that these values are important to control the infection spread.
We did not consider the validity term of the vaccination. The vaccination effect decreases based on elapsed time [33]. In future studies, we plan to investigate vaccination strategies that consider the validity of the vaccination. | 2022-02-16T06:48:06.955Z | 2022-02-13T00:00:00.000 | {
"year": 2022,
"sha1": "cd33747b360a75f35967aa2101934fe9d1091af1",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "cd33747b360a75f35967aa2101934fe9d1091af1",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Computer Science",
"Biology",
"Mathematics"
]
} |
21134896 | pes2o/s2orc | v3-fos-license | Exploring the relationships among service quality features , perceived value and customer satisfaction
The purpose of this paper is to explore the relationships among service quality features (responsiveness, assurance, and empathy), perceived value and customer satisfaction in the context of Malaysia. The empirical data are drawn from 102 members of an academic staff of a Malaysian public institution of higher learning using a survey questionnaire. The results indicate three important findings: firstly, the interaction between perceived value and responsiveness was not significantly correlated with customer satisfaction. Secondly, the interaction between perceived value and assurance also did not correlate significantly with customer satisfaction. Thirdly, the interaction between perceived value and empathy correlated significantly with customer satisfaction. Thus the results demonstrate that perceived value had increased the effect of empathy on customer satisfaction, but it had not increased the effect of responsiveness and assurance on customer satisfaction. In sum, this study confirms that perceived value act as a partial moderating variable in the service quality models of the organizational sample. In addition, implications and limitations of this study, as well as directions for future research are discussed.
Introduction
Service quality and customer satisfaction are inarguably the two core concepts that are at the crux of the marketing theory and practice (Spreng & Mackoy, 1996).In today's world of intense competition, the key to sustainable competitive advantage lies in delivering high quality services that will in turn result in satisfied customers (Shemwell, Yavas, & Bilgin, 1998).Therefore, there is not even an iota of doubt concerning the importance of service quality as the ultimate goal of service providers throughout the world (Sureshchandar, Rejendran, & Anantharaman, 2002).In an era of intense global competition; many organizations have now shifted the paradigm of service quality to customer's perspective (Parasuraman, Zeithaml, & Berry, 1985).Based on this paradigm, a customer will judge the quality of service accorded and determine whether it met his/her expectations (Grönroos, 1984;Parasuraman et al., 1985Parasuraman et al., , 1988)).Many scholars think that employee satisfaction with the accompanying service features may increase retention and loyalty (Alexandris, Dimitriadis, & Makata, 2002), thus paving the way to increased organizational competitiveness (Shemwell et al., 1998).
Service quality has been defined as a form of attitude -a long-run overall evaluation (Zeithaml, 1988;Parasuraman et al., 1988) while perceived service quality as a general, overall appraisal of service.The global value judgement on the superiority of the overall service could occur at multiple levels in a particular organization (Sureshchandar et al., 2002).Many researchers such as Parasuraman et al. (1988), Juwaheer and Ross (2003) and Walker, Johnson and Leonard (2006) highlight that responsiveness, assurance and empathy as among the three most important service quality features.
Responsiveness has often been defined as the willingness of a service provider to provide the needful services accurately and promptly (Juwaheer & Ross, 2003).
Assurance refers to credibility, competence and security in delivering those services (Juwaheer & Ross, 2003).Empathy is related to care, attention and understanding of the individual customer needs and interests when dispensing the prescriptive core services (Juwaheer & Ross, 2003).
Extant research in this area shows that the proper implementing and dispensing of the core service quality features may justifiably increase customer satisfaction (Gronroos, 1984;Parasuraman et al., 1988;Walker et al., 2006).In a quality and what the customer receives in actual terms (Caruana, Money, & Berthon, 2000;Parasuraman et al., 1988).If the service provided by an organization does meet a customer's needs and expectations, then this may subsequently lead to higher customer satisfaction (Foster, 2004;Parasuraman et al., 1988;Walker et al., 2006).Surprisingly, a thorough investigation of such relationships reveals that the effect of service quality features on customer satisfaction is not consistent if perceived value is present in organizations (Caruana et al., 2000;Varki & Colgate, 2001).
Perceived value is considered as customer recognition and appreciation on the utility of a product that is given by a service provider which may fulfil his/her expectation (Foster, 2004;Heininen, 2004;Walker et al., 2006).In a service management framework, many researchers are in tandem that responsiveness, assurance, empathy, perceived value and customer satisfaction are distinct constructs that are highly interrelated.For example, the ability of an organisation to incorporate the use of responsiveness, assurance and empathy in the delivering services will inadvertently promote an increase in customer perceptions of value; and this in turn will motion a higher level in customer satisfaction (Parasuraman et al., 1988;Sureshchandar, 2000;Sureshchandar et al., 2002).
Even though many studies have been carried out in this area of research, little is known about the moderating effect of perceived value in service quality models.
Most studies have found that the mediating role of perceived value has been less emphasized in previous quality system research.This is because of an over emphasis of a segmented approach in analyzing service quality features, the heavy reliance on a direct effect model testing, and negligence pertaining to the role of human needs and expectations in developing previous service quality models (Bitner, 1990;Caruana et al., 2000).Consequently, the findings from such research approaches may not highlight the influence of perceived value as a moderating variable in service quality literature (Eggert & Ulaga, 2002;Sureschchandar et al., 2002).
The current study has three major research questions: first, which features of service quality may affect customer satisfaction?Second, which features of service quality may affect perceived value?Finally, does perceived value affect the relationship between service quality features and customer satisfaction?Hence, this study is conducted to measure three main objectives: first, the relationship between the service quality features and customer satisfaction.Second, the relationship between the service quality features and perceived value.Lastly, the moderating effect of perceived value in the relationship between service quality features and customer satisfaction that occurs in one selected Malaysian public university in East Malaysia.For confidential reasons, the name of the institution is kept anonymous.
The remaining discussions of this paper are structured as follows: first, to elaborate the related literature review and research hypotheses.Second, to describe the methodology and procedure involved in conducting the study; and finally to discuss the results of hypothesis testing and conclusion.
Literature review and research hypotheses
Several studies about soft quality programs in Western organizational settings show that service quality may both, directly and indirectly, affect customer satisfaction.In terms of direct relationship perspective, Bitner (1990) examined the quality service based on a sample of 145 tourists and found that service quality was an important antecedent of customer satisfaction.In addition, Caruana et al. (2000) conducted a research comprising 80 personal interviews with customers of an audit firm and found that service quality positively correlated with perceived value.Hence, it can be hypothesized that: There is a positive relationship between responsiveness and customer satisfaction H2: There is a positive relationship between assurance and customer satisfaction H3: There is a positive relationship between empathy and customer satisfaction H4: There is a positive relationship between responsiveness and perceived value The service quality research literature is consistent with the notion of perceived quality models.For example, Parasuraman et al. (1985) state that a conceptual service quality model highlights that the match between service quality standards and customers' standards may decrease service performance gap and increase customer perceived value about the quality systems.Consequently, it may lead to higher customer satisfaction.This literature has been used to develop the conceptual framework for this study as shown in Figure 1.Based on the framework, it can be hypothesized that:
Research design
This study used a cross-sectional research design which allowed the researchers to integrate service quality literature, in-depth interviews, a pilot study and the actual survey as the main procedure to gather accurate and less bias data (Davis, 1996;Cresswell, 1998;Sekaran, 2000).This study was conducted in a selected public institution of higher learning in East Malaysia (HIGHINSTITUTION), where a center
Measures
Back translation technique was used to translate the content of the questionnaire in Malay and English in order to increase the validity and reliability of the instrument (Wright, 1996).The survey questionnaire was developed based on the modification of SERVQUAL instrument (Parasuraman et al., 1985).This questionnaire consists of three major sections: firstly, responsiveness was measured using 4 items.The items used to measure this variable are (1) repair teaching and learning facilities, (2) assist academic staff in operating teaching and learning facilities, (3) booking system for lecture halls and/or lecture theatres and (4) ready to assist academic staff if needed.Secondly, assurance was measured using 3 items.The items used to measure this variable are (1) confidence in the service provider, (2) comfortable in dealing with the service provider and (3) efficiency in providing services.Thirdly, empathy was measured using 3 items.
The items used to measure this variable are (1) understandable about academic staffs' needs; (2) put a priority in monitoring central teaching buildings and (3) ability in fulfilling the requests of the academic staff.
Fourthly, perceived value has 3 items that have been modified from service quality related perceived value (Caruana et al., 2000;Foster, 2004).The items used to measure this variable are (1) teaching and learning facilities are useful for teaching and learning, (2) teaching and learning facilities help to improve teaching and learning and (3) teaching and learning spaces are comfortable for teaching activities.Fifthly, customer satisfaction has 4 items that are modified from service quality related customer satisfaction (Bitner, 1990;Eggert & Ulaga, 2002;Walker et al., 2006).The items used to measure this variable are satisfaction or dissatisfaction with (1) the service provider's attitude and behavior, (2) the ability of the service provider's treatment of the academic staff, (3) the ability of the service provider to communicate with the academic staff and (4) the willingness of the service provider to maintain the teaching and learning conditions of the central teaching buildings.All these items are measured using a 7-point Likert scale ranging from 'very strongly disagree' (1) to 'very strongly agree' (7).Demographic variables are used as a controlling variable because this study is focused on employee attitude.
Sample
The unit of analysis for this study is the 800 members of the academic staff who are in employment at the HIGHINSTITUTION in East Malaysia.In the first step of data collection procedure, the researchers met the university's HR department to find out about the rules for distributing the survey questionnaire to the academic staff.Considering the organizational rule, a quota sampling was used to determine the sample size based on the length of the study and the budget constraint involving the 120 academic staff.Further, a convenient sampling technique was used to distribute the 120 questionnaires to the academic staff of all the faculties at the main campus of the university.From that number, a total of 102 survey questionnaire were returned to the researchers yielding an 84 percent response rate.The survey questionnaire were answered by the participants by consent and on a voluntarily basis.
Data analysis
A statistical package for social sciences (SPSS) version 15.0 was used to analyze the data from the questionnaire.Firstly, exploratory factor analysis (EFA) was used to assess the validity and reliability of measurement scales (Hair, Anderson, Tatham, & Black, 1998).Secondly, Pearson correlation analysis (r) and descriptive statistics were conducted to determine the collinearity problem and the usefulness of the data set.Finally, a hierarchical regression analysis, as recommended by Cohen and Cohen (1983), was used to measure the moderating effect of perceived value in the hypothesized model.Moderating effect is an interaction that shows the degree of relationship between the independent variables and dependent variables and that will change if other variables exist in the relationship (Cohen & Cohen, 1983;Jaccard, Turrisi, & Wan, 1990).The results of an interaction are evident when the relationship between interacting terms and the dependent variable is significant.The fact that the significant main effects of predictor variables and moderator variables simultaneously exist in analysis it does not affect the moderator hypothesis and is significant to interpret the interaction term (Baron & Kenny, 1986).
Sample profile
In relation to the sample profile, Table 1 shows the following characteristics: 1) the majority of the respondents are females (58.8%); 2) the respondents are mainly between the ages of 31 and 35 years old (54.9%); 3) many of them possess a masters' degree (80.4%); their length of service range from 4 to 7 years (49.0%), and most of the lecturers have an educational background in social sciences and humanities (58.0%).
General Characteristics Sub Characteristics Percentage
Sex
Validity and reliability analyses for measurement scales
The validity and reliability analyses were conducted based on the procedures established by Hair et al. (1998), and Nunally and Bernstein (1994).A principal component factor analysis with oblique rotation using direct oblimin was used to determine the possible dimensions of the constructs.Furthermore, the Kaiser-Mayer-Olkin Test (KMO) which is a measure of sampling adequacy was conducted for each variable and the results indicate that it was acceptable.The original survey questionnaire has 38 items which are related to five variables: responsibility (7 items), assurance (7 items), empathy (7 items), perceived value (7 items) and customer satisfaction (10 items).The factor analysis was conducted to condense the 38 items to 17 items.Table 4 shows the results of the validity and reliability analyses where (1) all research variables exceeded the minimum standard of Kaiser-Meyer-Olkin's value of 0.6 and are significant in Bartlett's test of sphericity (BTS), (2) all research variables had eigenvalues larger than 1, (3) the items for each variable exceeded factor loadings of 0.40 (Hair et al., 1998), and (4) all variables exceeded the acceptable standard of reliability analysis of 0.70 (Nunally and Bernstein, 1994).
These statistical results support the notion of the perceived value theories (Parasuraman et al., 1985), and empirical studies (Eggert and Ulaga, 2002;Varki and Colgate, 2001), signifying the goodness of data for this study.
Correlation and regression analysis
As depicted in Table 5, the mean value for each variable is between 5.01 and 5.43, indicating the level of responsiveness, assurance and empathy, perceived value and customer satisfaction ranging from a level of high (4) to highest (7).The Pearson correlation coefficients between the independent variables (i.e., responsiveness, assurance and empathy) and moderating variable (i.e., perceived value) and between dependent variable (i.e., customer satisfaction) are less than 0.90, indicating the data are not affected by serious collinearity problem (Hair et al., 1998).These correlations also provide further evidence of validity and reliability for measurement scales used in this research (Hair et al., 1998).
Measures
Table 5 shows the results of testing two direct effect hypotheses: first, the relationship between service quality features and customer satisfaction, and second, the relationship between service quality features and perceived value.The outcome of testing the first type of relationship shows three important findings: 1) responsiveness is positively and significantly related to customer satisfaction (r=.47, p<0.01), therefore H1 is supported.2) assurance is positively and significantly related to customer satisfaction (r=.35, p<0.01), therefore H2 is supported.3) empathy is positively and significantly related to customer satisfaction (r=.53, p<0.01), therefore H3 is supported.These statistical results show that the three service quality characteristics are important predictors of customer satisfaction in the organizational sample.
Besides that, the outcomes of testing the second type of relationship also display three important findings: 1) responsiveness is positively and significantly related to perceived value (r=.47, p<0.01), therefore H4 is supported.2) assurance is positively and significantly related to perceived value (r=.35, p<0.01), therefore H5 is supported.3) empathy is positively and significantly related to perceived value (r=.56, p<0.01), therefore H6 is supported.These statistical results show that the three service quality characteristics are important antecedents of perceived value in the organizational sample.Table 6: "Results for hierarchical regression analysis with perceived value as the moderating variable and customer satisfaction as the dependent variable".
Table 6 shows the outcome of testing the moderating hypotheses in Model 3.
Firstly, the interacting variable (responsiveness x perceived value) is insignificantly and negatively correlated with customer satisfaction (β=-0.22,p>0.05), therefore H1 is rejected.This result demonstrates that the inclusion of perceived value has not increased the effect of responsiveness on customer satisfaction.This indicates that perceived value does not act as a moderating variable in such relationships.
Secondly, the interacting variable (assurance x perceived value) is also insignificantly and negatively correlated with customer satisfaction (β=-1.83,p>0.05), therefore H2 is also rejected.This result demonstrates that the inclusion of perceived value has not increased the effect of assurance on customer satisfaction.This indicates that perceived value does not act as a moderating variable in such relationships.Thirdly, the interacting variable (empathy x perceived value) is significantly and positively correlated with customer satisfaction (β=1.72,p<0.05), therefore H3 is accepted.This result demonstrates that the inclusion of perceived value has increased the effect of empathy on customer satisfaction.This indicates that perceived value does act as a positive moderating variable in such relationships.
Discussion and conclusion
The finding for this research shows that perceived value does act as a partial moderating variable in the relationship between service quality features and customer satisfaction.In the context of HIGH INSTITUTION, Malaysia the service provider (center for teaching and learning) is given a major responsibility to plan, maintain, and monitor sophisticated teaching and learning facilities (i.e., lightings, air-conditioners, computers, multimedia and physical equipments) in the central teaching buildings based on the rules set up by the leadership of the university.In general, a majority of the academic staff perceived that implementation of such service quality practices has improved the delivery of teaching and learning services and this may invoke their perceived value about the quality systems.
Specifically, a majority of the academic staff perceived that their values about the use of empathy in delivering teaching and learning services might be an important predictor of customer satisfaction.Conversely, a majority of the academic staff perceived that their values about the use of responsiveness and assurance in delivering teaching and learning services might not be major antecedents of customer satisfaction.
A thorough review of the in-depth interview reveals that perceived value does not moderate the effect of responsiveness and assurance on customer satisfaction; this may be caused by external factors.Firstly, the teaching and learning facilities in the central teaching buildings are intensively used from morning till night within learning semesters for the purposes of teaching and conducting seminars, short courses and workshops.If these teaching and learning facilities are not properly utilized, they may not be repaired or replaced with other equipment within a short time.The duration of repairing and/or replacing with other equipment will be longer; and this may increase the academic staffs' complaints and criticisms of the service provider.
Secondly, as a new campus, the teaching and learning facilities in the central teaching buildings have been properly installed and managed by the service This study provides significant implications and impacts on three major aspects: theoretical contribution, robustness of research methodology, and practical contribution.In terms of theoretical contribution, the findings of this study provide three important elements: firstly, perceived value does positively moderate the relationship between empathy in teaching and learning services on customer satisfaction.In this sense, the inclusion of perceived value in the analysis has increased the effect of empathy in teaching and learning services on customer satisfaction.This finding is consistent with previous studies by Eggert and Ulaga (2002) and Varki and Colgate (2001).Secondly, perceived value does not moderate the relationship between responsiveness in the delivery of teaching and learning services on customer satisfaction.This result shows that the inclusion of perceived value in the analysis has not increased the effect of responsiveness in teaching and learning services on customer satisfaction.This finding is consistent with studies by Caruana et al. (2000) and Eggert and Ulaga (2002).Thirdly, perceived value also does not moderate the relationship between assurance in the delivery of teaching and learning services on customer satisfaction.This result demonstrates that the inclusion of perceived value in the analysis has not increased the effect of assurance in the teaching and learning services on customer satisfaction.This finding is not consistent with previous study by Eggert and Ulaga (2002).In sum, this study has extended previous research conducted in most Western countries and provided a great potential to understand the notion of Regarding the robustness of the research methodology, the data gathered using service quality literature, the in-depth interviews, pilot study and survey questionnaire have exceeded the minimum standard of validity and reliability analyses and this can lead to the production of accurate and reliable findings.
With respect to practical contributions, the findings of this study can be used as a guideline by the management to improve the design and administration of service quality programs in organisations.Specifically, the design and management of service quality may be improved if managers emphasize on the suggestions: firstly, staff who are involved in providing teaching and learning facilities need to be engaged in proper quality management training programs.Through these training programs, the staff will have the opportunity to acquire new knowledge, skills, abilities, as well as be imbued with good moral values.Consequently, these learning outcomes may be used to increase efficiency when engaged in current and future job undertakings.Secondly, staff members who are involved in providing teaching and learning facilities need to be given better rewards.For example, those who are involved in improving teaching and learning facilities are from the low level positions and who are eligible to do overtime.In this situation, providing extra monetary incentives to them will invoke their satisfaction and this may lead to increased motivation to perform their jobs better.Thirdly, recruitment policy needs to be changed from hiring fresh employees to experienced employees.For example, staff members who are involved in providing teaching and learning facilities are usually dealing with professional employees.If organization hires experienced staff this will encourage them to minimize errors and increase efficiency in installing, maintaining and monitoring sophisticated teaching and learning facilities.As a result, it may lead to an improved customer service and customer satisfaction.
Fourthly, staff performance need to be properly assessed.For example, members who are involved in providing teaching and learning facilities deal with backgrounds clients from varied background.The nature of their work requires different performance appraisal styles.For example, immediate boss and customers views should be used to determine accurate performance ratings.As a result, it may lead to increased staff motivation in the workplace.2001).Although a substantial amount of variance in dependent measure that is explained by the significant predictors is identified, there are still a number of unexplained factors that can be incorporated to identify the causal relationships among variables and their relative explanatory power.Therefore, one should be cautious about generalising the statistical results of this study.Finally, the sample of this study uses only the academic staff from a single university and they are selected by using a convenient sampling technique.The nature of this sample may decrease the ability of generalizing the results of this research to other organisational settings.
The conceptual and methodological limitations of this study need to be considered when designing future research.Firstly, this study sets up a foundation for research on relationships between service qualities, perceived value and customer satisfaction.It has raised many questions as well as confirming initial propositions.
A few research areas can be further explored as a result of this study.Secondly, the organisational and personal characteristics as a potential variable that can influence perceived value about service quality needs to be further explored.Using given more attention in considerable service quality literature (Alexandris et al., 2002;Parasuraman et al., 1988;Walker et al., 2006).The importance of these issues needs to be further explained in future research.
In sum, the findings of this study confirm that perceived value does act as a partial moderating role in the service quality model of the organisation.These results have partially supported and broadened service quality research literature published in most Western countries.Therefore, current research and practice within service quality models needs to consider individuals' perceived value as a critical aspect of service quality.The findings of this study further suggests that perceived value should be seen as a crucial aspect of service quality where perceived value about service quality may strongly induce positive subsequent personal outcomes (e.g., satisfaction, retention, commitment and thus loyalty).Thus, it may lead employees to maintain and increase academic excellence in higher learning institutions.
Figure 1 .
Figure 1."Perceived value as a moderator on the relationships between service quality features and customer satisfaction" for teaching and learning (central teaching building) has been established to develop and manage sophisticated teaching and learning facilities for the educational institution.This center uses teaching and learning quality standards set by the Ministry of Higher Education, Malaysia, to establish and manage lecture theatres and lecture halls in the target educational institution.In order to understand the nature of quality service, in-depth interviews are first conducted involving four experienced officers comprising of an assistant registrar, an information systems manager, an assistant administrative officer and a supporting staff.All of them are currently in active employment at the center.They are selected based on purposive sampling where the employees possess adequate knowledge and experience in the area of designing and administering quality service programs.The information garnered from such employees has helped the researchers to understand the nature of quality service policies and procedures, employees' perceptions of value about service quality and customer satisfaction characteristics, as well as the relationships between such variables in the target institution.After refining, categorizing and comparing the information with relevant theoretical and empirical evidence, the information was used as a guideline to develop the content of survey questionnaires for a pilot study.Next, the pilot study is conducted by discussing survey questionnaires with five experienced academic staff with a sound background knowledge and experience in social sciences, humanities, sciences and technology.The information is then used to verify the content and format of a survey questionnaire for the actual research.doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 236 A. Ismail; M.M.B. Abdullah; S.K. Francis doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 237 A. Ismail; M.M.B. Abdullah; S.K. Francis doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 238 A. Ismail; M.M.B. Abdullah; S.K. Francis 4 Results doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 240 A. Ismail; M.M.B. Abdullah; S.K. Francis doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 243 A. Ismail; M.M.B. Abdullah; S.K. Francis doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 244 A. Ismail; M.M.B. Abdullah; S.K. Francisprovider.For example, the service provider is given a major responsibility to provide teaching and learning aids for specific purposes such as teaching, conducting seminars, short courses and workshops.Although the service provider may provide teaching and learning equipment, it does not have sufficient equipment to meet the expectations of the academic staff.Besides that, the service provider does not have enough sufficient manpower and therefore will not be able monitor and maintain the equipment needed if there are many functions held at several of the lecture theatres and lecture halls in the central teaching buildings.The inability to manage such teaching and learning facilities may decrease the desired comfort, trust and courtesy of the academic staff when teaching and organizing seminars, short courses and workshops in the central teaching buildings.As a result, it may increase feelings of dissatisfaction and misjudgment about the services from among the members of the academic staff.
doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 247 A. Ismail; M.M.B. Abdullah; S.K. Francisthese organisational (e.g., ownership and type) and personal (e.g., gender, age, education and position) characteristics may provide meaningful perspectives for the understanding of how individual similarities and differences affect service quality policies within an organisation.Thirdly, the cross-sectional research design has a number of shortcomings; therefore other research designs such as longitudinal studies could be used as a procedure for collecting data and describing the patterns of change and the direction and magnitude of causal relationships between variables of interest.Fourthly, the findings of this study rely too much on the sample taken from a single organizational sector.To fully understand the effects of service quality on individual attitudes and behaviour via its impact upon perceived value, more organisational sectors (e.g., government linked companies, business organizations, and non-profit organizations) need to be used in future study.Finally, other personal outcomes of perceived value such as retention, loyalty and commitment should be considered in future research because they are
Table 2 .
Table2shows that gender has no significant difference, which means that responsiveness, assurance, empathy, perceived value and customer satisfaction, are not differently perceived by females and males."Independent Samples T-Test for the Differences between Gender".
Table 3
shows that age does not have a significant difference, meaning that responsiveness, assurance, empathy, perceived value and customer satisfaction are not differently perceived by age structures.Thus, post-hoc tests do not show the differences among the age structures.
Table 3 .
"The Results of One-Way ANOVA for the Difference between Age Structures".
Table 4 .
"Results of validity and reliability analyses".
Table 5
. "Pearson correlation coefficients between variables and descriptive statistics".doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 242 A. Ismail; M.M.B. Abdullah; S.K. Francis If organisations seriously consider such suggestions then this will help the academic staff to improve their teaching and learning activities.This in turn will motivate students' motivation to acquire new knowledge, skills, abilities and good moral values.Consequently, it may lead to enhance the effectiveness of teaching and learning in the universities.The conclusions drawn from the results of this study should consider the following limitations.Firstly, this study was a cross-sectional research design where the data was taken at one point of time within the duration of this study.In this sense, this research design did not capture the developmental issues (e.g., intra-individual change and restrictions of making inference to participants) and/or causal connections between variables of interest.Secondly, this study only examined the relationship between latent/unobserved variables (i.e., responsiveness, assurance, empathy, perceived value and customer satisfaction) and the conclusion drawn from this study does not specify the relationship between specific indicators for the independent variables, moderating variable, and dependent variable.Thirdly, the outcomes of multiple regression analysis have focused on the level of performance variation explained by the regression equations and it is also helpful to indicate the amount of dependent variable variation that is not explained(Tabachnick & Fidell, doi:10.3926/jiem.2009.v2n1.p230-250©© JIEM, 2009 -2(1): 230-250 -ISSN: 2013-0953 Exploring the relationships among service quality features, perceived value and customer satisfaction 246 A. Ismail; M.M.B. Abdullah; S.K. Francis | 2017-09-07T07:15:56.637Z | 2009-07-09T00:00:00.000 | {
"year": 2009,
"sha1": "30e4aa83f5238648f5b14768bb59ff89cb28ec17",
"oa_license": "CCBY",
"oa_url": "http://www.jiem.org/index.php/jiem/article/download/48/22",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "30e4aa83f5238648f5b14768bb59ff89cb28ec17",
"s2fieldsofstudy": [
"Business"
],
"extfieldsofstudy": [
"Psychology"
]
} |
210945085 | pes2o/s2orc | v3-fos-license | Double penetration wound: A nail gun injury involving the head and heart
Nail gun injuries usually occur at the extremities due to working accidents. Intracranial or intrathoracic injuries are relatively rare, and cases combined with both injuries are even rarer. Such situations pose challenges for surgeons due to their uniqueness during operation. Radiologic imaging findings in our case were significant in indicating surgical findings. Herein, we report a patient who shot himself using a nail gun, damaging his brain and heart.
Introduction
Nail gun injuries involving vital organs, especially those involving the brain or heart, often result in severe wounds and require surgical interventions immediately. Despite the traumatizing results, nail gun injuries have a better outcome compared with other penetrating injuries, such as those caused by knives or bullets [1] . Imaging techniques, especially computed tomography (CT), are often required either to survey the wound or to plan surgical procedures and sometimes predict the outcome for the patient. We report a case of a double-nail gun penetration injury involving the cranium and the thorax. The patient did not lose consciousness at the scene, and Acknowledgments: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declarations of competing interest: None. * Corresponding author.
his conscious was clear throughout his journey from his home to the hospital. Imaging results of the chest in our case revealed the related positions between the penetrated nail and the coronary artery (which were adjacent but not encountered), predicted surgical findings, and helped arrange treatment methods in our case. Median sternotomy instead of thoracotomy was performed for a better surgical view due to penetration of the left ventricle, which was also shown on the CT. Cardiopulmonary bypass (CPB) machines were not used in our case while repairing the heart due to simultaneously presenting hemorrhagic brain injury, which is quite the opposite in regular surgical heart repairs [2] . By this case, we would like to share our imaging findings and consequent management methods with readers and specialists in associated fields.
Case description
A 58-year-old man had underlying diseases of multiple old cerebral vascular infarctions, coronary 2-vessel-disease post stenting, type 2 diabetes mellitus, hypertension and end stage renal disease (ESRD). Two days after being diagnosed with major depression disorder, he committed suicide by shooting himself twice with a pneumatic nail gun. The first shot was in the chest, and after finding himself still alive, he shot himself again in the head. These nails were stainless, with 50 mm in length and 2 mm in diameter. He was discovered after 6 hours in his bathtub alone and sent to the emergency department (ED), without losing consciousness (Glasgow Coma Scale: E4V5M6) and limb weakness. After being transferred to the ED, he complained of headache and chest pain. Two penetrating wounds were found in the left chest and the temple around the right forehead, without gross external bleeding from the wounds ( Fig. 1 ). Chest x-ray and CT revealed a nail penetrating the left lower anterior chest wall to the heart with significant hemopericardium ( Fig. 2 ). Brain CT revealed intracranial hemorrhage (ICH) in the right frontal lobe and subdural hemorrhage at the right frontotemporal region ( Fig. 3 ). Emergent operation was then arranged. After median sternotomy was performed, a massive bleed was noted after opening the pericardium. The left ventricular (LV) rupture site was adjacent to the left anterior descending (LAD) artery, which was intact at first glance. After removing the nail, we repaired LV with Teflon felt interrupted mattress sutures immediately.
Next, we performed a craniotomy around the nail ( Fig. 4 ). After removing the nail, we cleaned the hemorrhage of subdural hemorrhage and ICH. Hemodynamics were stable during and after operation, and we arranged intensive care unit admission afterwards.
After emergent operation, this patient's hemodynamic status was stable and neither cardiac arrhythmia nor ischemia signs noted. Intracranial pressure monitor data was stable during postoperative period. Postoperative brain CT showed absorption of right frontal region ICH. However, patient did not recover his consciousness after we stopped sedative med-ications. Around 20 days after operation, his family decided to give up and withdrawal life supports.
Penetrating intracardiac injury
Mortality rates of intracardiac penetration injuries vary depending on the penetrating object. Nail gun injuries penetrating the heart have a mortality rate of up to 25% [1] . Mortality rates of stabbing injuries of the heart range from 22% to 62%. Gunshot injuries have a higher mortality rate from 60% to 95% [1] .
CT of the head and chest were done upon arrival at the ED, and in the axial and sagittal views, we could clearly see the left descending branch of the coronary artery, as in Figure 2 B. Comparing the site of the LAD branch and the nail penetrating the heart, it could be seen that though adjacent, the nail did not penetrate the LAD of the coronary artery. By viewing the image results, prediction of the findings during surgery could be made, which was validated in our patient.
The projectile speed of a nail gun can range from 150 ft/s (45.7 m/s) (pneumatic nail gun) to 1400 ft/s (426.7 m/s) (powder-actuated tool). The former and the latter are used for driving nails into wood and concrete/metal, respectively. Even though the velocity can reach as high as a handgun (700 ft/s) (213.4 m/s) [3] , the lateral distribution energy of nail guns is limited and most of the trauma develop along the route of penetration [4] . Unlike nails, a bullet will spin when shot, which deals more lateral distribution damage to the tissue .
When the penetration injury involves the heart, right ventricle (RV) injuries are most reported [5] , which was different from our case (LV). The reason might be the penetration angle being more lateral caused by our patient. The chamber pressure of LV is higher than RV, which might not only cause more severe hemopericardium, but also make it harder to repair if penetrated. The technique used in repairing an RV-penetrating injury often involves purse string suture to prevent bleeding and pledget to minimize tear during nail removal. A purse string suture around the penetrating site can be simultaneously closed as the nail is withdrawn [3] . However, unlike RV, such technique is difficult to apply in repairing LV, since LV cannot be directly seen during median sternotomy. We need to rotate the heart medially to expose LV, remove the nail from LV, and finally perform Teflon felt interrupted mattress sutures immediately to minimize bleeding.
For penetrating cardiac nail gun injuries, the most commonly preferred surgical options are sternotomy or thoracotomy [3] , and the choice is up to the surgeon. In our case, sternotomy was chosen due to better view of LV and bigger space for a heart-lung machine if needed. In order to minimize harm to the heart during nail removal, it is recommended that intravenous adenosine be utilized to reduce the heart rate and continuous transesophageal echocardiography be applied to assess the timing of nail removal, particularly in certain situations when the interventricular septum is involved [6] . In our case, however, the nail did not penetrate through the interventricular septum, so transesophageal echocardiography and adenosine were not mandatory.
After operation, the medical team should be aware of possible postoperative arrhythmia. Panicker et al stated that atrial flutter was identified in their case and the patient was treated with amiodarone [3] . In our case, an amiodarone pump was used due to developing atrial fibrillation 1 day after surgery. Sinus rhythm returned after 4 days of continuous use of the amiodarone pump.
Penetrating brain injury
The velocity of most nail guns is often enough for the nail to penetrate through the skull [7] . However, most pa- tients present clear consciousness when sent to ED [8] , and mortality following these intracranial nail injuries are low [9] . This might be due to the limitation of lateral distribution damage caused by the nail [4] , which was discussed above.
Brain CT is essential for examining nail locations and ICH [10 ,11] . However, significant artifact is frequently present if there are too many nails [8] , which could also be seen in our patient.
Surgical procedures for nail removal include craniotomy or simple blind extraction. Craniotomy is strongly suggested if a nail is totally embedded in the brain [12] ; while simple blind extraction could be considered if the head of the nail is outsides the skull [13][14][15] , but complicated delayed ICH has been reported in some studies [16][17][18] . No matter which surgical option is adopted, a continuous intracranial pressure monitor is suggested if at all possible [19] .
Postoperative complications for penetrating brain injury include infection and epilepsy. The percentage of infectious complications, such as wound infection, cerebral abscess and meningitis, is around 55% and 90% within 3 and 6 weeks, respectively [20] ; therefore, prophylactic tetanus and broad-spectrum antibiotics should be initiated as early as possible [19 ,21] . However, there is still no consensus on how long antibiotics should be administered, with estimates ranging from 7-14 days [22] to 6 weeks [21 ,23] .
Brain protection during cardiac surgery
Most cardiac operations involving heart valves require transient cardiopulmonary arrest, in which CPB was applied [2] . However, the brain, which is highly susceptible to hypoxia, is at risk of thromboembolism, hypoperfusion, and inflammation during CPB [2] . In order to counter these adverse events, anticoagulants, hypothermia, and systemic glucocorticoids are suggested while employing CPB [2] .
In our case, which was complicated with ESRD and ICH, usage of anticoagulants was extremely risky. Patients with ESRD have bleeding tendency due to platelet dysfunction [27] , which makes hemostasis for ICH harder in our case. Since the ventricles were not penetrated in our patient, we decided to repair the heart off-pump, meaning with the heart beating during operation.
Conclusion
We report a case of nail gun injury involving both heart and brain. Chest CT imaging findings assisted surgeons in surveying the coronary arteries relative to the penetration wound and planning surgical procedures. As far as intracardiac penetrating injury is concerned, CPB is often utilized when repairing such injuries; however, brain damage caused by hypoperfusion and thromboembolism during CPB cannot be underestimated, especially when there is already brain injury present (noted in our case). Anticoagulants for thromboembolism prevention were considerably risky in our case due to severe coagulopathy caused by ESRD, which was the reason we elected to repair the heart off-pump. In this research, we would like to share our experience in image findings and adoption of a more favorable strategy if similar cases are to be encountered. | 2020-01-23T09:06:04.520Z | 2020-01-22T00:00:00.000 | {
"year": 2020,
"sha1": "3e3b90c8be851f8f89ac476cf8c3288f926c9fd3",
"oa_license": "CCBYNCND",
"oa_url": "https://doi.org/10.1016/j.radcr.2019.11.021",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "b69de765b4da69c8110fc92f81412570cbd3786c",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
267320240 | pes2o/s2orc | v3-fos-license | Learning Agent-based Modeling with LLM Companions: Experiences of Novices and Experts Using ChatGPT & NetLogo Chat
Large Language Models (LLMs) have the potential to fundamentally change the way people engage in computer programming. Agent-based modeling (ABM) has become ubiquitous in natural and social sciences and education, yet no prior studies have explored the potential of LLMs to assist it. We designed NetLogo Chat to support the learning and practice of NetLogo, a programming language for ABM. To understand how users perceive, use, and need LLM-based interfaces, we interviewed 30 participants from global academia, industry, and graduate schools. Experts reported more perceived benefits than novices and were more inclined to adopt LLMs in their workflow. We found significant differences between experts and novices in their perceptions, behaviors, and needs for human-AI collaboration. We surfaced a knowledge gap between experts and novices as a possible reason for the benefit gap. We identified guidance, personalization, and integration as major needs for LLM-based interfaces to support the programming of ABM.
INTRODUCTION
The advent of coding-capable Large Language Models (LLMs) has the potential to fundamentally change the way people engage in computer programming [25].As LLM-based programming interfaces (e.g.GitHub Copilot; ChatGPT) become increasingly popular [46], some studies started to study their user perceptions [84].However, the research on their potential learning impacts is still limited.Many prior studies only focus on impressions of educators [46] or students [100], with little empirical data on the actual learning usage of these tools.On the other hand, a few studies started to explore how LLM-based interfaces can be designed to facilitate programming education, indicating potential advantages for learners.Notably, these studies suggest that learners with more prior programming experience tend to benefit more [42,57].While a recent study identifies some challenges for novice learners with LLM-based interfaces [101], there is a gap in understanding why experienced programmers seem to gain more learning benefits from these tools.
In this paper, we present the design of a novel LLM-based interface, NetLogo Chat, for the learning and practice of NetLogo.NetLogo is a widely used programming language for agent-based modeling (ABM), which applies simple rules on multiple individual agents to simulate complex systems [94].It is particularly powerful in capturing emergent phenomena, e.g., the spread of viruses or predator-prey systems [93].It is an important methodology in computational modeling across scientific disciplines and education from K-12 to postgraduate levels [88], where scientists and educators are highly in need of LLM-based interfaces [21,59].As an important part of computational modeling, the priorities of ABM differ from general programming [65].A modeler needs to verify that their conceptual design of individual rules matches the real-world patterns (e.g. a predator needs food to survive), the code matches the design (i.e.there are no unexpected or implicit assumptions), and the aggregated outcome matches real-world phenomena (e.g. if all prey die out, predators die too) [28].As most LLM-related studies on computer programming work on general-purpose languages that LLMs perform best (e.g.Python or Javascript), no LLM-related studies have explored ABM or other forms of computational modeling at this point.
NetLogo Chat was designed with constructionist learning principles and incorporated known best practices for ABM and computer programming.Constructionism advocates for the design of learning experiences where learners construct their understanding of the world (e.g.knowledge of ABM) through building personally meaningful artifacts (e.g. an agent-based model around learners' interests) [61].Similar to GitHub Copilot Chat[1], NetLogo Chat was integrated into an integrated development environment (IDE).Different from previous designs, it aims to give users more control over the human-AI collaboration processes, strives to incorporate authoritative sources, and tries to provide more support for troubleshooting.
Using both ChatGPT and NetLogo Chat as a probe [101], we conducted a qualitative study to highlight the different perceptions, behaviors, and needs of experts and novices during open-ended modeling sessions.We interviewed 30 expert and novice participants from academia, industry, and graduate schools around the world.Participants proposed diverse NetLogo tasks from their disciplines and worked toward their modeling goals.We asked interview questions before, during, and after their interaction with each design.We answered the research questions: (1) What perceptions -strengths, weaknesses, and adoption plans -do expert and novice users perceive LLM-driven interfaces to support their NetLogo learning and practice?(2) How do expert and novice users use LLM-driven interfaces to support their NetLogo learning and practice?(3) What are expert and novice users' needs for LLM-based interfaces to support their NetLogo learning and practice?
Learners generally agreed with our design principles and suggested additional features for future designs.As in other studies, experts reported more perceived benefits than novices.Comparing the different interaction patterns between experts and novices, our study reveals a behavioral gap that might explain the gap in benefits.We found that experts collaborated with LLM-based interfaces with more human judgment in all activities than novices, helping them overcome AI hallucinations, while novices struggled with evaluating and debugging AI responses.From there, we identified components of a knowledge gap between novices and experts.We reported experts' and novices' needs in LLM-based interfaces in three key themes: guidance (from LLMs); personalization (of LLMs); and integration (into modeling environments), many of which confirm and develop the design decisions of NetLogo Chat.The contributions of this paper include: (1) The design and implementation of NetLogo Chat, an LLMbased system that supports learning and practice of NetLogo, a widely-used programming language for ABM; (2) An empirical study that contributes to the understanding of how novices and experts perceive, use, and express needs for LLM-based programming interfaces in different ways; (3) A theorization of the knowledge gap between experts and novices that might lead to the behavioral gap, and suggestions of potential design interventions; (4) The design discussion and suggestions for building LLMbased programming interfaces that benefit both experts and novices in agent-based modeling more equitably.
RELATED WORK 2.1 LLMs for Computational Programming and Modeling
Researchers have been exploring natural-language-based interfaces for programming for decades, yet early attempts were mostly exploratory, being limited in capabilities.NaturalJava [64] required users to follow a strict pattern when prompting, while later systems (e.g.NaLIX [47] or Eviza [73]) asked for a specific set of English expressions.This created difficulties for users and system designers, as they felt "a main challenge of NLP interfaces is in communicating to the user what inputs are supported." [73] Without the capability to generate natural languages, those interfaces were also constrained to one-off interactions.
Recently, a new generation of LLMs demonstrated the capability to understand and generate both natural languages and computer languages.GPT-3 was examined in writing code explanations [52], documentation [44], and providing feedback for assignments [3].Soon, educators started to believe that Codex could be used to solve simple programming problems [27,90].Embedded in ChatGPT, GPT-3.5-turbo and GPT-4 demonstrated even stronger capabilities in programming.More and more LLMs have started to gain the capability of coding (e.g.PALM 2; Claude 2; CodeLLaMA 2), ushering in a new era of natural language interfaces for programming.
Even the most powerful LLMs suffer from hallucinations and may misunderstand human intentions.Early users of ChatGPT complained about incorrect responses and struggled to prompt ChatGPT for a desired output [74].While LLMs might outperform average humans in specific, structured tasks [58], the evaluation criteria might have been flawed [50], as LLMs struggled to combine existing solutions for a novel challenge [23].A study suggested that developers should not rely on ChatGPT when dealing with new problems [81].
LLMs are naturally less prepared in low-resource programming languages (LRPL).Here, our working definition for LRPL is similar to that of natural languages: with relatively scarce online resources and have been less studied by the AI field [53].LRPLs are not less important: NetLogo, the most widely used programming language for agent-based modeling (ABM) [80], is used by hundreds of thousands of scientists, educators, and students for computational modeling.Using simple computational rules for individual agents, ABM could simulate complicated emergent phenomena.It has been frequently used in different scientific disciplines [93] and science education [35] for recent decades.With considerably fewer online resources to train on, LLMs are much more prone to errors and/or hallucinations with LRPLs [79].
A few studies attempted to improve LLMs' performance with LRPLs in two directions.First, some studies fine-tuned foundational LLMs with LRPL datasets [12].While this approach demands considerable datasets and computational power, it has not been applied to generative tasks yet [31].Second, some studies used prompt engineering techniques.For example, aiming at simple tasks, a study creates sets of grammar rules for LLMs to fill in [86].Another study leveraged compiler outputs, allowing LLMs to iteratively improve their Rust code, but was only tested in a smaller number of fixed tasks [97].The potential of LLMs in scientific disciplines, including in computational modeling, is rarely explored.At this point, the only study targeted at STEM helps with a very specific engineering task [45].
User Perception and Behaviors with LLM-based Programming Interfaces
Two strands of user perception and behaviors studies informed our design and study: studies of conversational agents (CAs); and of LLM-based programming interfaces.For education, CAs were used to develop learners' writing [85], self-talk [29], and programming skills [95].Many of them are pedagogical conversational agents (PCA) with the aim to adaptively mimic the behaviors of human tutors [96].PCAs could serve in multiple roles, such as tutors [85], motivators [11], peer players [30], or learning companions [29].
Prior research of CAs underscored the importance of understanding user perception and behaviors [30], yet the technical boundaries of the pre-LLM era limited the freedom of designers.Previous studies have explored aspects such as trust, mutual understanding, perceived roles [18], privacy [69], human-likeness [36], utilitarian benefits, and user-related factors [49] to understand users' acceptance and willingness to use CAs.However, many CAs before LLMs had to use pre-programmed responses [87], and simply emulating functional rules from human speech failed to deliver people's high expectations of CAs [19].Without the capability to read or write code, pre-LLM CAs for computing education were largely limited to providing relevant knowledge [95] or supporting conceptual understanding of programming [48].
Recent studies have started to understand user perception and behaviors with LLM-based programming interfaces.In education, early studies focused on instructors' and students' perceptions of LLM-based interfaces for programming.Computer science students self-reported many potential benefits of using ChatGPT and were less inclined to report potential drawbacks [100].On the other hand, computer science instructors were significantly concerned over students' widespread usage of ChatGPT [46].While some instructors went as far as banning ChatGPT altogether, others suggested exposing students to the capabilities and limitations of AI tools, leveraging mistakes in generated code for learning opportunities.Both instructors and students expressed the need to adapt to a new, LLM-era way of teaching and learning [102].
For professionals, challenges and opportunities co-exist with LLM-based programming interfaces.Recent studies found programmers preferred to use Copilot [84] and finished tasks faster with Copilot [62].Yet, Copilot struggled with more complicated problems, providing buggy or non-reproducible solutions [23].Professional programmers faced difficulties in understanding and debugging Copilot-generated code, which hinders their task-solving effectiveness [84].Programmers who trusted AI were prone to write insecure code with AI [63].For conversational interfaces, despite inputs being in natural languages, users felt that they needed to learn LLM's "syntax" [26,37].
Our understanding of user perception and behaviors with LLMbased interfaces during (the learning of) computer programming is still very limited.As the field just started exploring this direction, previous studies mostly focused on general user impressions [102], or conducted behavioral tasks on pre-scripted, close-ended tasks [62].While close-ended settings made it easier to assess objective metrics [7], open-ended contexts open a wider window to understanding users' learning patterns, behaviors, perceptions, and preferences [8].For example, a recent study observed two modes that professional programmers interact in open-ended tasks with Copilot: acceleration, where the programmer already knows what they want to do next; and exploration, where the programmer uses AI to explore their options [4].Another study on professionals' prompt engineering shed light on their struggles, challenges, and potential sources of behaviors [101].
Still, we noticed two gaps in previous studies.First, a majority of studies chose professional programmers or computer science instructors/students as participants, while LLM-based interfaces are also used by millions of people without a CS background for programming tasks.Second, as HCI studies mostly focus on languages that LLMs are known to perform best, e.g.Python or HTML, little is known about user perceptions and behaviors when computational modeling or LRPLs are involved.
LLM-based Interfaces for Learning Programming and Modeling
While LLMs have shown promising potential in supporting human-AI collaboration in programming, most design studies were preliminary, and LLM-based interfaces for computational modeling remained understudied.For example, the Programmer's Assistant integrated a chat window into an IDE [68].Going beyond simple integrations, GitHub Copilot Chat[1] provided in-context support within code editors, yet its user studies were still preliminary [10].
A similar design was done on XCode without a user study [78].
Another study explored the integration between computational notebooks with LLMs and emphasized the role of the domain (in this case, data science) on LLM-based interface design [55].
LLMs have gained much attention among programming educators, but the design study is insufficient.Recent studies tested LLMs on introductory programming tasks and achieved unsurprisingly high scores [16,70].This prospect leads to great concerns among computer science instructors as they observed the widespread usage of ChatGPT among students [46].Yet, only a few LLM-based design studies targeted programming learning.Using a Wizard of Oz prototype, a study underscored the importance of supporting students' varied degrees of prior expertise [67].A design study reported positive short-term performance gains when young, novice programming learners engaged with Codex [42].Another study also found LLMs' benefits for novice programmers [57].Both studies found that more experienced programmers tended to benefit more, yet the reason was still unclear.
In this study, we invoke the learning theory of Constructionism [61] to inform our LLM-based system and empirical study design.While there is no rigid definition for Constructionism, it argues that learning happens most felicitously when learners "consciously engage in constructing a public entity" [61].In the context of computer programming, it means learning happens naturally through programming computers, as it iteratively externalizes learners' internal understanding of the world in code, and then allows learners to improve their understanding through watching how the code runs [60].Moreover, it argues that computer programming is not as abstract or formal as it appears; individual programmers' approaches are often concrete and personal, in pluralistic ways [83].However, the pluralism in thoughts is more difficult to capture by close-ended tasks (such as a problem set) and objective metrics (such as completion rate/time) [8].As such, constructionist learning studies often prefer open-ended tasks (e.g.making games [40], designing instructional software [32], creating agent-based models in NetLogo [8]) and qualitative studies, as they open windows into the nuances of learners' perceptions and behaviors in more natural and realistic settings.
The Logo programming language and its descendants (e.g.Scratch; Alice; NetLogo) succeeded in supporting multiple ways of knowing and thinking in computing education and in scientific research [75], yet to our knowledge, no published studies have explored their synergy with LLMs.Many prominent constructionist design principles could be applied to AI-based interfaces [41] and inspired the design of NetLogo Chat.For instance, "low floor, high ceiling, wide walls" asks learning environments to provide 1) an easy entrance for novices (low floor); 2) the possibility for experts to work on sophisticated projects (high ceiling); 3) the support of a wide range of different explorations (wide walls); 4) the support of many learning paths and styles [66].We also learned from previous design studies that stress the importance of adaptive scaffolding [14,72] and support debugging [9] for novices to learn NetLogo.Hence, we contributed to the field one of the first design studies of LLM-based interfaces for learning programming that follow the constructionist tradition.
NETLOGO CHAT SYSTEM
NetLogo Chat is an LLM-based system for learning and programming with NetLogo.It comprises two main parts: a web-based interface integrated with Turtle Universe (a version of NetLogo) [13] (See 3.1); and an LLM-based workflow that improves the quality of AI responses and powers the interface (See 3.2).We iteratively designed the system by: (1) Based on authors' experiences in teaching NetLogo, we created a design prototype based on the constructionist learning theory (see 2.3), with a focus on supporting users iteratively build up their prompts and smaller code snippets before working on entire models.We developed a proof-of-concept system, using prompt engineering techniques to interact with GPT-3.5-turbo-0314.
(2) We internally evaluated the proof-of-concept with a group of NetLogo experts.During this process, we encountered frequent hallucinations with NetLogo (grammatical or conceptual mistakes; inventing keywords that do not exist; etc).
For the system to provide guidance, we realized that authoritative sources are necessary for LLMs' performance; (3) We incorporated the official NetLogo documentation and code examples into the system using prompt engineering techniques (see 3.2), evaluated other LLMs' potential, and then conducted pilot interviews to evaluate the system with three external NetLogo experts invited from NetLogo's mailing lists.The interviews used a protocol similar to the one we formally used (see 4.2), with more flexibility and openendedness; (4) Based on the external feedback, we identified the need for supporting troubleshooting, leading to the design decision 3.1.3.We upgraded the underlying LLM to GPT-3.5-turbo-0613, fixed many minor usability issues, and finalized the prototype that we used in the empirical study.
Design Overview
Figure 1: NetLogo Chat asking for details about human's needs.
3.1.1Enable users to program the computer, rather than being programmed by the computer.Over-reliance on LLM-based interfaces has become a major concern among both educators and some learners, where students blindly follow the instructions given by LLMs without attempting to construct their representations of knowledge.Such a scenario is antithetical to the constructionist learning tradition, where Seymour Papert's fear of "computers program children" comes back to life again [60].
Inspired by the Logo language, the design of NetLogo Chat aims to give control back to learners: to suppress LLMs' tendency to give a quick response that often assumes too much about the learner's inclination, we force it to ask clarification questions more often.For this feature to work effectively, it is essential to ask questions with quality.To achieve this, we used a few-shot approach and crafted templates for LLMs to follow.We conducted an informal evaluation of LLM's generated questions during our development process and empirical study.Across the board, the LLM we used was able to generate questions with acceptable quality, similar to the one demonstrated in Fig 1 .A future design could embed a larger set of templates and retrieve a few relevant templates when needed.Following previous examples in related tasks [38], we integrated NetLogo's official documentation and model examples to help improve LLMs' and human performance.Different from previous studies, we not only provided related examples to LLMs, but also revealed them to users.By doing so, we seek to improve the transparency of LLM's mechanism, foster trust in the LLM-driven system, and provide authoritative guides and examples for users even when LLMs might fail to provide precise support.
Integrate with the IDE and Enhance
Troubleshooting.We seek to integrate NetLogo Chat into NetLogo's IDE beyond integrating a conversational assistant parallel to the code editor.To facilitate a constructionist learning experience, the code editor needs to be integrated into the conversational interface, where learners can work with smaller snippets of code with more ease.Thus, the design might lower the threshold for learners to tinker with the code, a key learning process advocated by the constructionist literature [61,83].
Fig 3 provides a concrete example, where the embedded editor displays a piece of generated code.Instead of having to copy and paste the piece back into the main editor, the user could first see if any syntax issues exist in the code; run the code within a conversation; and ask follow-up questions or raise additional requests, before putting back a working code snippet into their projects.
To further support the user's troubleshooting, in addition to error messages, NetLogo Chat will display extra debugging options for users.Users could choose to look for an explanation, or ask the LLM to attempt fixing the issue on its own, or with the user's ideas.During the process, the system will attempt to find documentation and related code examples to reduce hallucinations.Building on the literature on error messages' impact on learning [5], we also clarified many messages to provide a better context for humans and both LLM-based systems used in the study.Since OpenAI started to provide fine-tuning on GPT-3.5-turbo(the version also used in ChatGPT Free) only after we concluded the main study in July, NetLogo Chat was implemented with prompt engineering techniques.We built our project on ReAct [99], a promptbased framework that could reduce hallucination, improve human interpretability, and increase the trustworthiness of LLMs.By requiring LLMs to generate an action plan and delegate the action to a third-party conventional agent (e.g.search for documentation, ask clarification questions, conduct a static syntax check, etc.) before composing the final response, the framework provides a promising pathway to integrate external inputs (e.g.human input, official documentation) into LLM workflows.Fig 4 depicts a rough outline of NetLogo Chat's workflow.Imagine a user requests to "create a predation model":
Technical Implementation
(1) The LLM is instructed, in the prompt, to first elaborate on the request (planning): "The user intends to create an agentbased biology model related to predation.However, it is unclear what exactly the user wants.We need to ask followup questions." (2) Next, the LLM is instructed to choose an action from the list: Ask clarification question(s); Search for documentation; Write a response; Say sorry.Here, imagine the LLM chooses "Ask clarification question(s)" based on the planning.(3) Then, the LLM needs to generate some questions based on the request.Because LLMs are trained on real-world data, it is not difficult for them to come up with some ideas.For example, "What species do you want to put in the model?" The LLM is also instructed to provide some examples, e.g."Wolf", "Sheep".( 4) When the user replies to the questions, the loop restarts from step (1).Since there is sufficient information about the request, the LLM decides to search for information, and also generates keywords for the search, e.g."Wolf-sheep predation model in NetLogo".(5) The system conducts a semantic search on a pre-assembled database of NetLogo's official documentation and code examples.The system returns the search result, use it as a new round of input, and restarts from step (1).( 6) With inputs from both the user, who clarified the request; and the database, which supplies the example; the LLM plans again, chooses to write a response, and generates its final response.
In the example, we initiated three requests with the LLM, each with a prompt template that results in a structured response [99] (e.g.any response needs to have a Plan, an Action, and a Parameter).Each request could use a different LLM that works best for the specific request.Using this approach, the system has the potential to balance cost, performance, speed, and privacy.For example, a future iteration of NetLogo Chat could leverage a fine-tuned local LLM to probe the user's intentions and search for documentation.Then, with any personal or sensitive information stripped away, the system could forward the compiled request to a powerful online LLM (e.g.GPT-4).
For the empirical study, we chose GPT-3.5-turbo-0613 as Net-Logo Chat's LLM backend.First, we expect most participants to be using the free version of ChatGPT, driven by the same LLM.In this way, we would have a fair playing field for the empirical study, where both systems will be used.Second, at the time of our study, the response time for GPT-4 was too long to sustain a realtime experience, while we had no access to other NetLogo-capable LLMs' APIs.Although we did observe some remarkable improvement when internally evaluating the system (e.g.ChatGPT has trouble answering questions for lesser-known NetLogo keywords, while NetLogo Chat does not), a more systematic evaluation rubric is needed for future research.Building on the tradition of understanding the difference between experts and novices [17], we separated the participants into experts and novices using self-reported survey data.To mitigate the effect of inaccurate responses, NetLogo experts in the team, who have been core developers and instructors of NetLogo, watched every video and decided if a participant greatly overestimated or underestimated their capabilities.We considered the participant's discussions with the interviewer, the think-aloud process, and the coding behaviors.A vast majority of users' reports correspond with the experts' judgment.Then, to simplify the analysis, we separated participants (Table 2) by their levels into two main categories: experts, who are either experts in NetLogo or programming in general; and novices.In the study, we denote experts by the prefix E (E01-E17) and novices by N (N01-N13).13 experts had previous experience with ChatGPT (76%), including programming (65%, n=11).11 novices (85%) also used ChatGPT before, but much less for programming (38%, n=5).
Interviews
Our study was conducted in 3 phases: (1) We pilot interviewed 3 experts invited from NetLogo's online community.Each was asked to comment on LLMs for NetLogo learning, as well as on ChatGPT and an early prototype of NetLogo Chat.Each semi-structured interview lasted between 60-90 minutes and was video recorded.Prior to each formal interview, participants were asked to come up with a short NetLogo task that they were interested in working on.Almost every participant brought forward a modeling task from their career domain or personal interest, e.g. to model "how honeybees decide to regulate the temperature of the hive", or "the spread of conflicting ideas".Only once, when the task scope was too complicated for the session, did we ask the participant to bring another.During any part of the interview process, interviewers generally followed the protocol, asking followup questions when needed.Specifically: (1) We asked baseline questions, e.g., "What do you think are the potential advantages / disadvantages of using LLMs in supporting your learning and programming of NetLogo?" (in 2 separate questions) (2) We asked the participant to work on their task with the help of ChatGPT.Then, we asked the same baseline questions again, then asked "What do you like or dislike about the interface".Repeat the procedure with NetLogo Chat; (3) If time permitted, we further asked about their preferences for learning and/or programming with NetLogo and asked which feature they wanted to add/remove from either system.Here, the objective was not to strictly compare between the two systems, but to elicit more in-depth discussions over LLM-based interfaces.
Since almost all users have already engaged with ChatGPT, we did not randomize the order of ChatGPT/NetLogo Chat.Also, 3 participants used the paid version (GPT-4) during the task with ChatGPT.While much of the generated data comes from the inevitable comparison between the two systems, we chose not to interpret them as objective comparisons.Instead, the different design principles underpinning the systems presented two objects to think with [60], that our participants drew on during their reflections and discussions of LLM-based programming interfaces.
Data Analysis
Our interviews resulted in around 40 hours of video data.Around half of our data is behavioral in nature, where participants worked on their tasks and were encouraged to think aloud; the other half is more verbal, where participants answered questions.As such, each interview was not only transcribed verbatim, but also watched by a researcher to create observational notes.The two streams were then combined into a single archive for analysis.
Based on our research questions, we iteratively applied the grounded theory approach [22] to analyze our data.During each step, the research team fully discussed the discrepancies between each researcher and iteratively refined the codebook to improve consistency.The analysis reached theoretical saturation at around 50% of interviews, when additional interviews no longer revealed unexpected major insights for our research questions.Then, we finished the rest of qualitative coding with the finalized codebook (Table 3).
(1) Four researchers open-coded 2 interviews, one from a novice and one from an expert, to summarize the topics mentioned by participants.During this process, researchers coded in different tabs to avoid interference.Three broad themes emerged from this phase: participants' approaches to programming; participants' interactions with AI systems; and their comments on AI systems.(2) Taking notes of the emerging themes, the first author created a preliminary codebook that categorizes dozens of codes into themes.Each researcher coded another 2 interviews in different tabs.In this phase, we refined the themes into approaches to programming (which also helps to separate experts and novices); perceptions and observed behaviors related to AI systems; and comments on AI systems' abilities.(3) Based on the coding results, the first author created a formal codebook, with definitions clarified based on the discrepancies between researchers (Table 3).To reduce the unbalanced influence of subjective interpretation, researchers only coded explicit behaviors; or direct comments.To avoid missing insights, researchers were instructed to highlight places where existing codes are insufficient to cover the topics.During the first two weeks, a few codes were created or merged as a result of discussions.We retrospectively revised our coding.
Based on the codebook, the first author iteratively incorporates themes into an outline.To further mitigate individual differences, researchers were asked to include as many codes as possible for each quote or observation.
FINDINGS
5.1 Perception: Before and After Interaction 5.1.1Before Interaction: Positive Expectations.Prior to the tasks, both novices and experts had positive expectations of LLM-based interfaces for NetLogo, with novices holding higher expectations than experts.
Both novices and experts expected LLM-based interfaces to save human time and support human effort, especially compared to other help-seeking activities.With LLMs, human time and energy could be liberated for more high-level tasks ( E12 , N03 ).Educators felt that LLMs could facilitate more efficient teaching, allowing students to "more complicated things with relative ease", spiking "their imagination."( E02 ) LLMs can also bring emotional benefits by reducing the fear of "bothering the teachers or the experts" ( E14 ) or asking "stupid questions" ( N06 ).
Most participants highlighted AI's potential to help them with NetLogo's syntax.For most participants, NetLogo is not the main programming language they used.Before the advent of ChatGPT, N06 felt that she needed to "recite the words (syntax of NetLogo)".Yet, the need was eliminated when "AI can teach you very quickly".Many experts also needed support, as NetLogo "has very strict syntax rules" ( E07 ) which makes writing more difficult.
Novices, in particular, expected that AI could be helpful for troubleshooting.N08 , for instance, felt that LLMs could help him through the troubleshooting process by describing "what I'm trying to do and get a snippet of code that helps get me past that block".For novices without a background in programming, this future looks promising.N12 is interested in the potential to "make programming more approachable to students".
5.1.2Before Interaction: Negative Expectations.Almost every participant expressed concerns or reservations about LLM-based interfaces.Yet, the concerns of novices and experts were conspicuously different.
Experts focused on preserving human judgment.E01 believed that AI should not "replace human judgment and ability".Similarly, E06 insisted that "(human) has to do the main thinking and ideas and all of that."E17 felt that humans cannot let AI "take over the main reasoning and emotions, the emotions intervening in the decisions."Many educators were also "concerned about learning" ( E13 ), fearing the tendency to "default to the AI system to come up with the answers instead of working through it ourselves" ( E12 ).Many experts explicitly explained their rationales.For example, E08 was concerned that "if a model points me to a suboptimal direction, I will have no idea, because I haven't considered alternative structure".E15 feared that relying on AI responses might "make your horizon narrow" because she would miss learning opportunities when browsing through the models library.For computational modeling, AI also might lack "in-depth knowledge in a specific field" to create an entire model ( E05 ).As such, E05 would only trust AI to "finish a specific task".
Novices were more optimistic and more concerned with their capabilities of understanding AI's responses or making AI understand them.For example, while N04 thought "one of the hypothetical drawbacks" to LLMs being "confidently incorrect", they added that "people are like this too".On the other hand, N03 feared that she would waste more time with AI if "it didn't understand me, or if I had difficulty expressing".N02 acknowledged that "there is a limitation to not knowing how to code (on how much AI could help)."Without knowledge of NetLogo, N11 felt difficult to spot LLM-generated mistakes.
After Interactions: Different Impacts of Hallucination.
All participants encountered AI hallucinations throughout the sessions.While some participants rated NetLogo Chat higher than ChatGPT's free version, most participants had similar changes in perceptions: experts, in general, reported more benefits from LLMs than novices.Some participants reported more positively about NetLogo Chat's capabilities.Several experts questioned ChatGPT's training in Net-Logo, yet they trusted more in NetLogo Chat, for it incorporates authoritative sources (see 3.1.2).E16 believed that NetLogo Chat LLMs could save human time and effort, especially for syntax, and provide emotional benefits.
Before, Positive
LLMs could help troubleshooting.LLMs could mislead humans to suboptimal directions.
While LLMs may make mistakes, it is no worse than humans.LLMs could hinder learning processes.
LLMs may not understand human intentions.
Before, Negative
LLMs could only work on smaller tasks.LLMs' responses are difficult to understand.LLMs supported learning or practicing by saving time.
LLMs supported learning or practicing by saving time.
After Interaction
Will continue to use LLMs for learning or practicing NetLogo.
Will seek alternative learning resources before continuing to use LLMs.
"understands your NetLogo syntax" and "the basic aspects of Net-Logo".N02 thought NetLogo Chat still had bugs but was "much more informative and precise than ChatGPT."As NetLogo Chat is designed to support troubleshooting (see 3.1.3),E04 thought NetLogo Chat "was able to kind of do some better troubleshooting to a certain extent, for it clarifies error codes".In both cases, experts understood hallucinations as an inevitable part of human-AI collaboration and reacted with more leniency.When E03 first encountered an incorrect response, he exclaimed: "Very interesting!You're mistaken."E05 felt that LLMs helped him "finish most of the code", though he still needed to "debug and see if the code makes sense logically."As experts did not rely on LLMs to resolve issues but mostly leveraged them as a shortcut, E06 stated that hallucinations were instances "where the programmer needs to use own experience and discretion", as risks would escalate if one extrapolates "what ChatGPT provides you in a wrong manner".
Novices, on the other hand, reported more obstacles and frustration, as they relied more on LLMs for their tasks.N07 emotionally responded to a hallucination that ChatGPT "apparently made that shit up".N01 had difficulties to "fix the bugs that were in it (the generated code)."N08 's session ended up "hitting a dead end", with the frustration leading him to "go consult other resources".
Most novices and experts still thought that LLM-based interfaces supported their learning or practicing by saving time.Even though N03 had "low trust" in ChatGPT, she still felt more confident after collaboration, for it "narrowed down the stuff I have to figure out myself and has made me much faster already."As an educator, N12 felt that LLMs facilitated a constructionist learning experience in which "you're being thrown into the culture and have to learn it on the fly."E13 thought he learned a syntax from ChatGPT that would "save me time in the future" and the learning process was "a lot faster than if I were doing it by hand".
As experts reported more perceived benefits, they predominantly intended to continue using LLM-based interfaces for NetLogo.After the task, E11 felt confident that "I can write anything I want to write".Yet, many novices, driven by their frustration with LLMs, sought alternative learning resources before considering a return.N04 , for instance, had a 180-degree turn: expressing great hope before the tasks, they now inclined to "build more by myself with my own code, without AI."N13 thought that she would prefer to work with "someone who is familiar with the programming language" together with LLMs.
The Behavioral Gap Between Novices and Experts
5.2.1 Behavioral Gap in Planning and Prompting.While experts' and novices' tasks were similar in terms of complexity, we observed differences between how novices and experts plan out their tasks.Since most participants gradually adapted their prompting styles, we focused on participants' first-round prompts.Two initial prompting patterns, one emphasizing modeling the entire system and another focusing on smaller, initial aspects of the task, emerged from our interviews.Most novices adopted the first pattern (11/13, 85%), while many experts adopted the second pattern (9/17, 53%).Below, we introduce one vignette for each pattern: (1) N05 started by asking: "I need to make a model of the bunch of agents who are trying to promote political views to other people (...)".Although he used GPT-4, the returned code still came with several syntax errors.N05 then spent the next 20 minutes trying to ask GPT-4 to fix issues without success.He expected to "put the idea into it and we'll run the code", but in the end "it didn't happen."(2) E07 started by asking ChatGPT to "write code for drawing a rectangle".When GPT-3.5 failed to further divide the rectangle, E07 instantly pivoted to another strategy: "I have the following code that draws a rectangle.I want you to modify it so the rectangle is divided by two".GPT-3.5 still failed, yet it produced working code and did "something close to it".
The second prompting pattern involved remarkable mental efforts to decompose and plan out the task.For example, E07 described his approach as "separate into small, general tasks you want to do." E04 explained that he "just likes to iteratively build (the code)".On the other hand, in the first pattern, many participants attempted to shortcut the efforts by delegating the tasks to AI, as N05 said: "I just want to ask it (ChatGPT) to just directly make a code for this task and that's it." By the end of the task, most participants had realized the importance of breaking tasks into smaller pieces for coding with AI.Naturally, when an LLM-based interface generated code with mistakes, a participant would be (implicitly) guided to ask smaller follow-up questions.Soon, many of them realized the benefits.N01 thought it would be better if one "works through real small problems first, before getting to more complicated problems."N10 would "start with something really basic."Experts using the first pattern had similar ideas.For example, E12 decided to restart "with something simple and just work with it."
Behavioral Gap in Coding and
Debugging.As most participants engaged with an agent-based modeling task that they never worked on, both experts and novices learned some aspects of Net-Logo with the help of AI -although, in different ways.Experts usually took a much more measured, prudent, and critical approach during coding and debugging, while novices mostly followed AI's instructions.
Most novices focused on reading AI's explanations and followed AI's instructions during their coding processes.ChatGPT often gives instructions like "You can copy and paste this code into NetLogo and run it".Even without this hint, almost all novices would copy and paste the generated code without much reading.The tendency worried some novices, but they had no choice: "I feel like I'm waiting for someone to tell me the answer, rather than learning how to solve it."( N11 ) Experts put more emphasis on the code, often ignoring the explanations provided by AI.During their reading, experts evaluated and often criticized the responses, planning their next steps along the way.Only a few experts tried copying and pasting the code to see if they worked out of the box.Other experts selectively copied and pasted parts of the code into their programs, or wrote their programs with generated code on the side.Even when they copied and pasted the code, experts were more cautious.For example, while E04 decided to "just take this and see what this does", he also realized that AI-generated code would override his ideas and manually edited the code.
All participants inevitably had to debug parts of the generated code.Yet, novices sought support from AI more frequently and often struggled with AI responses.For example, N12 would regularly "copy the code that doesn't make sense and go back to AI to see if it can help me."N09 complained that while ChatGPT gave suggestions, "it obviously requires fiddling around with it."As she had little idea about NetLogo, it became a purely trial-and-error experience.Even when AI did solve some errors, it was challenging for novices to learn from the process.For example, N04 commented that while NetLogo Chat provided an automated process, it was still difficult for him to get the lesson, "since I didn't write it myself."
Behind the Behavioral Gap:
The Knowledge Gap.We identified a knowledge gap that may lead to the behavioral gap.When novices realized that they needed to spend more effort decomposing the task or vetting AI responses, they found themselves lacking the necessary knowledge.We summarized, in participants' own words, the four components of a knowledge gap that novices need to overcome when working with AI.
Novices reported the need for conceptual knowledge of modeling.For example, N07 described his experience as "like being adrift on an ocean.Without a compass, and without a map."With only a basic understanding of agent-based modeling, N11 felt compelled to accept ChatGPT's response as "I don't really know how to interpret some of the output from it."Such feelings correspond with novices' tendency to skim through AI responses.Whereas, some novices asked for help from LLMs with different degrees of success.N04 first asked: "(...) Can you tell me what I will need to do before we begin?"With AI's suggestions, N04 had some more success asking follow-up questions.
The unfamiliarity with the basic concepts of NetLogo and/or coding in general further adds to the difficulty in prompting and understanding.After reading a guide suggested by NetLogo Chat, N07 realized that he "probably wouldn't have chosen NetLogo to ever begin with" for his database-related task.Other novices Debug with (more) help from AI.
Debugging "Oh, I didn't ask him to move.That is my problem." "I'm going to ask it the same question, but I'm confused why it said something about patches." were often confused by NetLogo's terms, even when they were mostly in plain English.N03 was confused about "why (ChatGPT) said something about patches" (note: patches are static agents that form NetLogo's modeling world), and that deepened her reliance on ChatGPT.N10 realized that she "only understand 20% of what I am reading, so I can't vet it myself."When the interviewer asked about adding comments into code, N03 replied that while it might be helpful, she was still missing "the high-level understanding of how it comes together."Many novices also lack the experience for debugging, leading to more unsuccessful attempts and more frustrations.Participants, in particular novices, were often confused by error messages from Net-Logo.N01 acknowledged that "without background knowledge, it is hard to figure out what the bugs are, if (LLM) gives you information that is inaccurate."Without experience in debugging, many novices felt frustrated and helpless as previously reported.On the other hand, E12 noted that his students "might not be comfortable with the idea that debugging is a normal part of the process."E01 believed that "the user needs a little practice in debugging their own code" before working with LLM-based interfaces.
Most novices felt a need to learn to interact with LLMs.After repeated failures, N01 felt that he did not "even know what questions to ask to get it to, because it is not doing the right thing."N06 thought AI would help a lot if she could "learn more about how to use AI." N05 realized that he needed to use the correct keywords, for otherwise it "will never generate a good model."This knowledge is relatively easier to acquire though: while N09 felt that "how to ask questions is very important", she believed that "you learn by actually doing it."
Needs for Guidance, Personalization, and Integration
5.3.1 "Good" Responses, "Bad" Responses.Participants generally appreciate and expect less technical, clear instructions.Many of them appreciate NetLogo Chat's design decisions that include authoritative sources in responses (see 3.1.2)and ask back clarification questions (see 3.1.1).However, participants' preferences are also highly personal and situational.For both designs, some participants explicitly went against excessive or unnecessary explanations, particularly when the goal is primarily to accomplish a task at hand.For instance, E09 complained that GPT-4 "talks too much.I want the code, not the explanation yet."E14 complained that while related code samples provided by NetLogo Chat could "contain a lot of good suggestions", she wanted to move them to "another box or an expandable line".Some participants appreciated and hoped that LLMs could stay on topic and give smaller pieces of information at a time.E01 thought NetLogo Chat would be more helpful if it only attempted to solve a bug "one at a time", for users "always overfill their buffer".Novices, in particular, prefer concrete, step-by-step responses, given the focus they put on AI-generated instructions.N04 wanted to "test one by one if (LLM) gave me multiple suggestions."Going beyond text responses, N03 hoped that there could be "a visual to help me better understand, or internalize what different elements of the code are", so her learning could move to a higher-level understanding of the code's intention.
For NetLogo Chat, most participants reacted positively to the reference to authoritative sources (see 3.1.2),the usage of NetLogo's language, and the provision of links to sources.E03 believed that "the possibility to go directly from this AI to the documentation" would be helpful for his students.N10 "automatically like (Net-Logo Chat's response) better" because it used "NetLogo's kind of turtle and patch language."E12 felt "a little bit more confident in the information I was getting because it seemed to be coming from inside of the application."However, sticking too much to authoritative explanations might have a downside.E10 complained that NetLogo Chat gave him "dictionary reference", and "dictionary definitions are not especially helpful." Many participants, in particular experts, reacted positively when NetLogo Chat assumed less about and stuck more to their intentions (e.g.asking questions back, see 3.1.1).For example, E09 commented that ChatGPT (GPT-4) "assumed what I wanted it to do, whereas this one makes you specify your assumptions."He prefers NetLogo Chat's approach, because "it makes you think about the code more."E12 felt that NetLogo Chat's clarification of intention was akin to "progressively guiding me towards a better prompt."As transparency is a key factor in computational modeling, N11 feared that if "anyone can produce an agent-based model, but without actually understanding all the parameters", hidden assumptions introduced by ChatGPT could be detrimental.
Need for Personalization.
In this section, we break down the strong needs of experts and novices for more personalization, besides response styles, into two themes: knowledge levels and help-seeking needs.
Novices, in particular, felt a strong need for LLM-based interfaces to acknowledge their knowledge levels and produce responses accordingly.N07 gave a stringent critique of both systems, feeling both systems were "not useful at all", for both "presumes you know something about NetLogo".N08 felt that "ChatGPT has no idea of how much or how little I know about how to code in NetLogo, or how to code in general."Solving this issue would require more personalized approaches.Coming from an educational background, both N02 and E03 suggested that LLMs should first probe the knowledge level of users before providing answers.
Participants gave a variety of suggestions that were at times conflicting: (1) Some participants prefer a guided walkthrough.N08 hoped that LLMs could walk him through the process and provide starting points.Both E14 and N03 hoped that LLMs could be used alongside video tutorials, where they could first see a successful example of human-AI collaboration and then ask follow-up questions.(2) Some participants prefer contextual recommendations.
N11 hoped that LLMs could show related code examples and provide "two or three other ways that you might look with".E10 suggested that LLMs provide in-context explanations if "you don't remember the definition or explanation of a particular command".(3) Some participants hope that LLMs could support helpseeking from humans.E01 hoped that LLMs could help novices "explain my situation so that I can paste it to the user group", so human experts could intervene more easily when AI fails to unstuck novices.Similarly, E17 suggested that AI could be combined with "peer to peer answers and collaboration".(4) Some participants believed that incorrect responses could become a learning opportunity.E02 was concerned that students might be "exposed to fewer options" with AI, compared with "coding from scratch".E03 feared that a system capable of directly producing solutions might deprive students of the debugging process, where they would have learned." Novices also had similar feelings.After many hallucinated responses, N08 thought that ChatGPT "forces me to learn as opposed to just getting code that's ready to go."To fully transform the moment of mistake into learning opportunities, educators suggest the design not to frame mistakes as failures, but rather "as a learning moment" ( E12 ).
Need for Integration.
Compared with ChatGPT in a separate browser window, most participants appreciated the NetLogo Chat interface being an integrated part of the modeling environment.They are particularly in favor of the deep integration in NetLogo Chat's design that goes beyond placing a CA and an IDE side-byside.We further identified many participants' need for a deeper integration.Many participants appreciated the integration of a sandbox-like code editor in NetLogo Chat, where they can tinker with smaller, AI-generated code chunks and execute them on the fly (see 3.1.3).N12 "definitely liked this feature of being able to go easily between the code and see what was changed and what was added."N04 appreciated that one can "see the code run" in the NetLogo IDE, which ChatGPT could not do.N13 thought while some code generated by ChatGPT was "so comprehensive", NetLogo Chat was able to break it down and make them "more conducive".Participants also expressed further needs for iterative modeling.E13 hoped that NetLogo Chat could help him "modularize all of my commands" by splitting the code into many smaller, more manageable chunks.E12 asked for a comparison feature between versions of code chunks that could help him "iterative changes quickly".
In addition, participants also hoped that LLMs could help them reflect on longer pieces of (existing) code.N02 and E02 wanted AI to support the combination of multiple, smaller code chunks into a single, coherent code.As such, LLM-based interfaces should be able to work with longer pieces of code.Both N06 and E08 hoped that NetLogo Chat could "look at my code and make suggestions based on my code".
Many participants needed adaptive support for modeling more than just coding.Many requested AI support in building model interfaces that could be used to take in inputs or send out outputs.For example, N06 needed NetLogo for her academic paper, hence plotting became "very important".For educators like E13 , while the canvas output was "good for the three-quarters of a project", it hid "the real power of agent-based modeling -tracking the emergent properties of the model, rather than simply making bits run around the screen."During the modeling processes, many interface parts could become necessary or unnecessary depending on situational needs.Integrated LLM-based interfaces need to go beyond a "side chat window" and support various spatial configurations for advanced users to decide on.
DISCUSSIONS
Our study first reported, in detail, how novices and experts perceive and use LLM-based interfaces (ChatGPT & NetLogo Chat) differently to support their learning and practice of computational modeling in an open-ended setting.Most participants appreciated the design direction NetLogo Chat is heading toward.However, they also expressed their needs for improved guidance, personalization, and integration which opens up huge design spaces for future improvement.
6.1 Guidance: Bridging the Novice-Expert Gap For most participants, guidance is what they need most from LLMs in programming.While hallucinations from LLMs constantly present a challenge to everyone, with a higher frequency to evaluate and debug AI responses, experts suffered less negative impact than novices.As a result, experts reported higher levels of perceived gains and more optimistic adoption plans than novices.While novices in our study also attempt to evaluate and debug AI responses, they are ill-equipped for these tasks.Without understanding the knowledge gap between experts and novices, it becomes impossible to design effective guidance.
Based on our empirical findings, we theorize the two types of knowledge novices might need when collaborating with AI in computational modeling (Fig 5).First, the knowledge to effectively decompose and plan modeling tasks.Second, the knowledge to evaluate AI responses and identify potential issues.We further identified four components of knowledge that both novices and experts reported to be essential: conceptual knowledge of modeling; basic concepts of NetLogo and coding; experiences of debugging; and how to interact with LLMs.To mitigate the impact of currently inevitable hallucinations of LLMs, it is essential to help novices get over the knowledge gap.
We propose three learning moments where design intervention might work best.The first moment is when users plan their next steps.While most novices started by delegating the planning process to AI, most of them eventually planned on their own.Here, we follow the constructionist learning theory for a broader understanding of planning that includes both rigid, formal plans and "softer", ad-hoc exploration of problem spaces [83].Both planning styles should be recognized as legitimate in learning and supported by the design [83].With our current design, most novices reported positive feelings when NetLogo Chat attempted to clarify their intentions and produce a plan for their task.Since this phase does not involve any generated code, more support could be provided, as novices may have fewer problems reading and evaluating natural language responses.They may also feel more comfortable asking questions about modeling or programming ideas, relating them to the generated code later, without fearing that they cannot (yet) read or write code.Moreover, LLMs could expand learners' visions by suggesting new ideas, proposing new plans, or taking notes of human ideas.When novices are confused about basic concepts, LLMs could suggest video or textual tutorials and provide Q&A along the way.
The second moment is when users read and evaluate LLMgenerated code.Reading and understanding code is one of the most important aspects of computing education [51].However, novices in our study were neither confident nor equipped for reading code.As a result, they intended to skip the code section.As predicted by the interest development framework [56], the lack of skills (knowledge) and confidence (identity) may mutually enhance each other.Breaking the feedback loop requires designers to scaffold their reading experiences in both directions.By making explanations within code (as comments or tooltips) or visualizing the code structures (e.g.[76]), we might be able to help build novices' connections between code syntax and real-world meanings.To build up learners' confidence, LLMs should deliver code pieces and explanations in adaptive sizes that work for learners.For learners who still could not succeed, the interface should further provide ad-hoc support that helps novices ask follow-up questions, or lead them to appropriate learning resources.
The third moment is when users need to debug their code.Debugging is considered a rich learning opportunity in constructionist learning [39].However, it is often associated with negative feelings that both manifested in prior literature [91], as well as our findings.Unfortunately, cognitive science has found that negative moods may further impede debugging performance [43], enlarging the gap between novices and experts.Following the suggestions of educators in our study, we suggest that LLM-based interfaces could frame bugs in a more positive light, while providing a link to a successful human-AI collaborative debugging process for firsttime learners.Both novices' and LLMs' debugging processes are often stuck in loops [92,97].While such situations are inevitable, some expert participants suggest that LLM-based interfaces could encourage learners to seek help from another human.Help-seeking is recognized as an important part of programming education, yet novices often struggle with it [54].In such cases, LLM-based interfaces should further help them frame questions for human experts.
Personalization: Beyond "Correctness" of LLMs
Personalization has been identified as an essential factor for perceived autonomy when users interact with conversational agents [98], for emotional and relational connections [89], and for various educational benefits [6].Adding to previous literature, we found personalization to be a crucial factor for LLMs to facilitate effective guidance for learning, as participants expect LLMs to recognize their knowledge levels and react accordingly.
While LLMs might have the potential to further the personalization of learning, recent research in LLMs focused on the "objective" capabilities, ignoring the personalized aspect of its evaluation.For example, technical reports of LLMs all reported benchmarks in whether they could produce functionally correct programs (HumanEval) [15]; if they could correctly answer multi-choice questions (MMLU) [33]; or if they could produce the correct answer of grade school mathematical problems (GSM-8K) [20].While working toward such "correctness" benchmarks is certainly crucial for LLMs to reduce hallucination and produce better responses, it becomes problematic when the definition of "helpfulness" or "harmfulness" is measured with a ubiquitous scale without individual differences [2], and such a definition has since been adopted by all major players in LLMs.
At least in learning and practice programming, we argue that helpfulness cannot be a singular metric, but instead varies based on many factors.Corroborating with constructionist design principles [66], we identified some potentially important factors such as knowledge levels and help-seeking preferences, while other factors such as culture, ethnicity, and gender could be as important.To support human learning, the full potential of LLMs could only be achieved through the recognition of epistemological pluralism [83]: humans have different approaches toward learning, and technology needs to be tailored to human needs.
Most participants in our study expected or asked for personalization, in the sense that LLMs recognize their knowledge levels and help-seeking needs, yet today's designs are still far from that.While it is virtually impossible to fine-tune thousands of LLM variants, LLMs' role-play capabilities and novel prompt-based workflows (e.g. the one used by NetLogo Chat, or the concept of GPTs very recently released by OpenAI) have shown promising potential.As personalization requires the inevitable and sometimes controversial collection of user data, we suggest a more upfront approach: only collecting data that directly contributes to a more helpful AI (e.g. the knowledge level), only using data for this purpose, and explaining the benefits, risks, and privacy processes at the beginning.Alternatively, designers could also consider flowing the pathway of cognitive modeling, which deduces learners' knowledge levels from known interactions with the system [77].On the other hand, our understanding of users' perceptions, behaviors, and needs for LLM-based programming interfaces has just begun, and we call on more studies to pursue this direction.
Integration: LLMs for Computational Modeling
For most participants, integration between LLM-based interfaces and modeling environments goes beyond stitching a chat window into the IDE.While most of them appreciated NetLogo Chat's design directions, they put forward many needs that are worth considering in future design.Here, we briefly discuss the two major themes: support for troubleshooting; and support for modeling.For troubleshooting: (1) The capability to work on smaller snippets of code, with the capability to execute, explain, and debug code in context.For both humans and LLMs [34], debugging complicated code is known to be difficult.NetLogo Chat has made the first step in reducing the scope to smaller code chunks.As such, it becomes easier for both humans to debug and LLMs to support their debugging processes.Whereas, more work is needed to bring together the code chunks into coherent full programs.(2) The capability to leverage authoritative NetLogo documentation in generated responses, as well as for the user's own reference.In debugging contexts, LLMs' tendency to hallucinate becomes more frustrating.By providing users and LLMs with authoritative explanations within the debugging context, NetLogo Chat may reduce the effort for users to seek related information, which is also known to be difficult for novices [24].More work is needed to explain in a more personalized way: for example, pure novices may need explanations for every basic term.(3) The capability to automatically send in contextual information (i.e.code and error messages) for LLM to troubleshoot.Users generally appreciated NetLogo Chat's design decision to support troubleshooting.However, the convenience came with a potential price: when using NetLogo Chat, users were more likely to ask LLMs for help, which might lead to fewer human attempts and learning opportunities.Further studies are needed to understand this design balance better.
Many participants also asked for features that specifically support their computational modeling tasks, which are known to have different priorities from programming in general [65].Here, two more capabilities are warranted: (1) The capability to assume less, actively probe, and stick to user intentions.In addition to the potential learning opportunities (see Discussion 1), for participants, hidden assumptions in scientific modeling are particularly harmful.While users appreciate NetLogo Chat's direction in having LLMs ask questions back, future interfaces should be able to facilitate the conversational build-up of plans and steps, further supporting users to program computers piece-bypiece rather than falling to hidden assumptions made by LLMs.
(2) The capability to support modeling practices beyond coding.Building the program is only one step; computational modeling also involves design, data visualization, and validation [88].For LLM-based interfaces to support modeling practices, future interfaces should go beyond coding to support users' efforts throughout the modeling process.
LIMITATIONS AND FUTURE WORK
There are limitations to our study that warrant future work.As a widely used agent-based modeling language, a deeper understanding of user perceptions, behaviors, and needs for LLM-based interfaces around NetLogo may inform us of design choices for other modeling environments.Future work should consider computational modeling or programming environments that might have different priorities.Since the NetLogo language was designed for an audience without a computer science background [82], it becomes more important and meaningful to understand how to design for bridging the novice-expert gap in LLM-based interfaces.However, it is unclear whether our findings and suggestions would sufficiently support novices' and experts' learning and practice of NetLogo.Using a more rigid rubric to distinguish between experts and novices might improve the rigor of our study.A quantitative, controlled study in the future might further (in)validate our findings and suggestions.As such, we plan to work on a new iteration of NetLogo Chat design and empirical study to fully understand the design implications.
Although we aimed to recruit participants representative of Net-Logo's global audience, our participant pool was not as representative as we hoped in two key dimensions.First, our participants were mostly professionals, academics, and graduate students.While K-12 teachers and learners are another major audience for NetLogo and agent-based modeling and may have different priorities and preferences [71], only one K-12 teacher was present in the study.More studies are warranted to further the empirical understanding of LLM-based interfaces in education contexts.Second, the demographics of our participants skewed towards North American and European, highly educated, and male.Such a group of participants, recruited voluntarily, might manifest higher than average acceptability toward novel technology, e.g.most of our participants have already engaged with ChatGPT.For future work, researchers need to recruit a more balanced and diverse group of participants, if the goal is for LLM-based programming interfaces to equitably support novices and experts throughout the world.
CONCLUSION
As Large language models (LLMs) have the potential to fundamentally change how people learn and practice computational modeling and programming in general, it is crucial that we gain a deeper understanding of users' perceptions, behaviors, and needs in a more naturalistic setting.For this purpose, we designed and developed NetLogo Chat, a novel LLM-based system that supports and integrates with a version of NetLogo IDE.We conducted an interview study with 30 adult participants to understand how they perceived, collaborated with, and asked for LLM-based interfaces for learning and practice of NetLogo.Consistent with previous studies, experts reported more perceived benefits than novices.We found remarkable differences between novices and experts in their perceptions, behaviors, and needs.We identified a knowledge gap that might have contributed to the differences.We proposed design recommendations around participants' main needs: guidance, personalization, and integration.Our findings inform future design of LLM-based programming interfaces, especially for computational modeling.
Fig 1 and Fig 2 provide an exemplary comparison between NetLogo Chat and ChatGPT's reaction to a simple modeling request.Here, ChatGPT immediately assumes details of the user's needs and generates an entire model for the user to copy and paste.Whereas,
3.1.2Invoke Authoritative Sources Whenever Possible.Hallucination is another major concern for LLMs, particularly in an LRPL like NetLogo.For example, the code generated by ChatGPT in Fig 2 contains multiple syntax issues and requires human experts to address them.More powerful LLMs suffer from the same symptoms.We submitted similar sample requests to GPT-4, PaLM2, Anthropic Claude 2, and Falcon-180B: none was able to produce syntactically correct code for a classical NetLogo model.
( 2 )
We improved the design of NetLogo Chat based on what we learned from the pilot interviews and revised the interview protocol accordingly.(3)We conducted formal interviews with 27 online participants (30 in total).
Figure 5 :
Figure 5: A preliminary theorization of the novice-expert knowledge gap.
Table 1 :
Overview of Participant Demographics (n=30) ute learning resources of agent-based modeling (ABM).The exact breakdown of participants' demographic data can be seen in Table1.The participant pool largely represented the scientific modeling community in NetLogo's main audience, with a majority of participants coming from STEM disciplines.Many participants were also related to the educator sector.6 participants (20%) were instructors who teach or are interested in teaching NetLogo in classrooms; 4 (13%) were graduate-level students interested in learning NetLogo, making up a third of the population.Participation in the study was voluntary.All participants signed an online consent form on Qualtrics.
Table 3 :
An Overview of the Codebook
Table 4 :
Novices and Experts' Perceptions on LLM-based Interfaces for NetLogo
Table 5 :
Novices and Experts' Behaviors During Human-AI Collaboration | 2024-01-31T06:45:09.490Z | 2024-01-30T00:00:00.000 | {
"year": 2024,
"sha1": "f41e70709f1eaaeb2387ab3f4eb6d8de8ed04c53",
"oa_license": "CCBYNCSA",
"oa_url": "https://dl.acm.org/doi/pdf/10.1145/3613904.3642377",
"oa_status": "HYBRID",
"pdf_src": "ArXiv",
"pdf_hash": "f41e70709f1eaaeb2387ab3f4eb6d8de8ed04c53",
"s2fieldsofstudy": [
"Computer Science",
"Education"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
237322299 | pes2o/s2orc | v3-fos-license | Identification of Riptortus pedestris Salivary Proteins and Their Roles in Inducing Plant Defenses
Simple Summary The bean bug, Riptortus pedestris (Fabricius) is a notorious pest of soybean crops in Asia. During the feeding process, the bug secretes a mixture of salivary components, which play critical roles in the insect–plant interactions. In the present study, a total of 136 salivary proteins were identified by transcriptomic and proteomic approaches. Among them, five proteins (RpSP10.3, RpSP13.4, RpSP13.8, RpSP17.8, and RpSP10.2) were capable of inducing cell death, reactive oxygen species (ROS) burst, and hormone signal changes, indicating the potential roles of these proteins in eliciting plant defenses. Our results provide a good resource for future functional studies of bug salivary effectors and might be useful in pest management. Abstract The bean bug, Riptortus pedestris (Fabricius), is one of the most important soybean pests. It damages soybean leaves and pods with its piercing-sucking mouthparts, causing staygreen-like syndromes in the infested crops. During the feeding process, R. pedestris secretes a mixture of salivary proteins, which play critical roles in the insect–plant interactions and may be responsible for staygreen-like syndromes. The present study aimed to identify the major salivary proteins in R. pedestris saliva by transcriptomic and proteomic approaches, and to screen the proteins that potentially induced plant defense responses. Altogether, 136 salivary proteins were identified, and a majority of them were involved in hydrolase and binding. Additionally, R. pedestris saliva contained abundant bug-specific proteins with unknown function. Transient expression of salivary proteins in Nicotiana benthamiana leaves identified that RpSP10.3, RpSP13.4, RpSP13.8, RpSP17.8, and RpSP10.2 were capable of inducing cell death, reactive oxygen species (ROS) burst, and hormone signal changes, indicating the potential roles of these proteins in eliciting plant defenses. Our results will shed more light on the molecular mechanisms underlying the plant–insect interactions and are useful for pest management.
Introduction
Plants and herbivorous insects have been engaged in a long-term co-evolutionary arms race. Generally, insects, especially those with piercing-sucking mouthparts, inject saliva into plant tissues, while the injected saliva then acts as the biochemical interface that plays critical roles in food processing [1]. Upon secretion, the bioactive saliva has an array of functions from anchoring and lubricating stylets, digesting nutrients and plant cell components to modulating plant defenses [2][3][4][5]. The identification of saliva components is the first step in understanding their functions. In recent decades, with the development of omic technology, salivary proteins from some agriculturally important pests have been reported, including aphids [6][7][8], planthoppers [9,10], leafhoppers [11], whiteflies [12], and plants (strain: Wandou 27) at 26 ± 0.5 • C with humidity of 50 ± 5% under a 16/8 h (light/dark) photoperiod.
Transcriptomic Analysis
The salivary glands (SGs) were collected from adult R. pedestris using a pair of forceps in a phosphate-buffered saline (PBS) solution (137 mM NaCl, 2.68 mM KCl, 8.1 mM Na 2 HPO 4 and 1.47 mM KH 2 PO 4 , pH 7.4). The isolated SGs were immediately transferred to RNAiso plus (TaKaRa, Dalian, China) and the total RNA was extracted according to the manufacturer's recommendations. Altogether two biological replicates were performed, with each containing 20 SGs. Thereafter, the RNA samples with high integrity and quantity proceeded for library construction and Illumina sequencing in Novogene (Beijing, China). Briefly, poly (A) + RNA was enriched by oligo (dT) magnetic beads, and fragmentation was performed with cations at 94 • C for 5 min. Afterwards, the extracted RNA was reverse transcribed using N6 random primers. After end-repairing and adaptor ligation, the sample was subjected to PCR amplification and purification to create a cDNA library with the QIAquick PCR purification kit (Qiagen, Hilden, Germany). Subsequently, the library was sequenced on an Illumina platform, and the output raw data were filtered to remove the low quality reads. Then, the clean reads were aligned to the reference genome using HISAT2 [32]. The transcripts per million (TPM) expression value of each gene was calculated by StringTie [33]. The TPM values of each gene from R. pedestris guts, fat body, muscles, carcass, testes, and ovaries have been reported in our previous work [30]. For the identification of SG-specific genes, the TPM value of each gene in SG was compared with those in the other six tissues, respectively. The gene-relative abundance (ratio) and p-value were calculated based on differential expression analysis embedded in DESeq2 [34]. Only genes with fold changes > 10 and p-value < 0.05 in all comparison groups were considered as SG-specific genes.
Proteomic Analysis
The SGs used for proteomic analysis were collected from adult R. pedestris. The 50 SGs were pooled into one sample, which was later lysed with lysis buffer that contained 100 mM NH 4 HCO 3 , 8 M Urea and 0.2% SDS, followed by 5 min of ultrasonication on ice. The lysate was subsequently centrifuged at 12,000× g for 15 min at 4 • C, and the supernatant was transferred to a clean tube. The sample was then reduced with 10 mM DTT for 1 h at 56 • C and alkylated with sufficient iodoacetamide for 1 h at room temperature in dark. Then, the sample was completely mixed with 4 volumes of precooled acetone under vortexing and incubated at −20 • C for at least 2 h. The sample was then centrifuged and the precipitation was collected. After washing twice with cold acetone, the pellet was dissolved with the dissolution buffer. For trypsin digestion, 3 µL of 1 µg/µL trypsin and 500 µL of 50 mM TEAB buffer were added, and the sample was mixed and digested at 37 • C overnight. Later, formic acid was added into the digested sample, and centrifuged at 12,000× g for 5 min at room temperature. The supernatant was slowly loaded to the C18 desalting column, washed with washing buffer (0.1% formic acid, 3% acetonitrile) thrice, and eluted with the elution buffer (0.1% formic acid, 70% acetonitrile). The eluent of sample was collected and lyophilized.
LC-MS/MS analysis was performed following the methods described below. Briefly, the lyophilized powder was dissolved in 10 µL solution A (100% water, 0.1% formic acid), and centrifuged at 14,000× g for 20 min at 4 • C; later, 1 µg of sample was injected into a home-made C18 Nano-Trap column (2 cm × 75 µm, 3µm). Peptides were separated in a home-made analytical column (15 cm × 150 µm, 1.9 µm) using the linear gradient elution. The separated peptides were analyzed by the Q Exactive HF-X mass spectrometer (Thermo Fisher, Bremen, Germany), with the ion source of Nanospray Flex™ (ESI) (Thermo Fisher, Bremen, Germany) at the spray voltage of 2.3 kV and the ion transport capillary temperature of 320 • C. The top 40 precursors with the highest abundance in the full scan were selected and fragmented by higher energy collisional dissociation (HCD) and analyzed in MS/MS. Afterwards, the resulting spectra from each fraction were searched separately against the R. pedestris database [30] by the search engines: Proteome Discoverer 2.2 (PD 2.2, Thermo Fisher, Bremen, Germany). The search parameters were set as follows: the mass tolerance for precursor ion was 10 ppm and the mass tolerance for product ion was 0.02 Da. Carbamidomethyl was specified as fixed modifications, oxidation of methionine (M) as dynamic modification, and acetylation as N-Terminal modification in PD 2.2. At most 2 missed cleavage sites were allowed. In order to improve the quality of analysis results, the PD 2.2 software was used to further filter the retrieved results.
Bioinformatics Analysis
Blast2GO was utilized to annotate the salivary proteins. BGI WEGO (http://wego. genomics.org.cn/cgi-bin/wego/index.pl accessed on 2 October 2020) was adopted to classify proteins into three main categories (biological process, cellular component and molecular function). TBtool was employed to identify the enriched GO terms using the following formula: where N represents the number of gene with GO annotation, n represents the number of interested genes in N, M stands for the number of genes in each GO term, m indicates the number of tissue-specific genes in each GO term [35]. A value of p < 0.05 was considered significant enrichment. To analyze the proteins potentially secreted, the presence of signal peptides was firstly predicated by the SignalP 4.1 Server (http://www.cbs.dtu.dk/services/SignalP/ accessed on 2 October 2020). Then, proteins containing a signal peptide were searched against the TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/ accessed on 2 October 2020), so as to identify the potential transmembrane domains. Proteins without or with one transmembrane domain included in the predicted signal peptide were considered as the secreted protein.
In Planta Expression of Salivary Proteins
To investigate the roles of salivary proteins in inducing plant defenses, in planta expression of salivary proteins was performed in Nicotiana benthamiana leaves. Briefly, the coding sequences without signal peptides were amplified from cDNA of R. pedestris SG using the primers listed in Table S1. Then, the purified PCR products were cloned into binary vector that containing GFP using the Trelief TM SoSoo Cloning Kit (Tsingke, Nanjing, China). The recombinant plasmids were subsequently introduced into Agrobacterium tumefaciens GV3101. Afterwards, the A. tumefaciens containing target plasmids were harvested by centrifugation and resuspended in the infiltration buffer (10 mM MgCl 2 , 10 mM MES, 150 mM acetosyringone). Infiltration was performed by injecting agrobacterium cell suspensions into N. benthamiana leaves using a needleless syringe. The plants were kept in a climate chamber under the temperature of 23 ± 1 • C, the RH of 70-80%, and the light/dark photoperiod of 16/8 h. At 48 h post-infiltration, the samples were collected for DAB staining and real-time quantitative PCR (qPCR) analysis.
DAB Staining
The ROS level in N. benthamiana was detected by DAB staining. In brief, the tobacco leaf was cut and immersed into 1 mg/mL DAB-HCl (pH 3.8, Sigma, St. Louis, MO, USA) for 6 h. Then, the DAB solution was replaced with 100% ethanol and decolored overnight at 65 • C. The stained sample was then photographed using a Canon EOS 80D camera (Canon Inc., Tokyo, Japan).
qPCR Analysis
The N. benthamiana samples were collected by a punch (8 mm in diameter) and homogenized in RNAiso plus. After extraction, the extracted total RNA was reverse transcribed using the HiScript II Q RT SuperMix Kit (Vazyme, Nanjing, China) to synthesize the first-strand cDNA. In this process, the contaminant genomic DNA was removed by adding gDNA remover in the kit. qPCR was performed using the SYBR Green Supermix Kit (Yeasen, Shanghai, China) with the Roche Light Cycler ® 480 Real-Time PCR System (Roche) by the following reaction program: denaturation for 5 min at 95 • C, followed by 40 cycles at 95 • C for 10 s and 60 • C for 30 s. The qPCR primers, which were designed by Primer Premier 6.0, are listed in Table S1. The housekeeping genes for actin was used as the internal control. The output data were calculated by a relative quantitative method (2 −∆∆Ct ) [49]. Statistical significance between GFP control and each salivary protein was calculated using Student's t test. Three biological replicates were performed.
Overview of the Riptortus pedestris Transcriptome
The salivary proteins of R. pedestris were identified according to the workflow shown in Figure 1. For transcriptomic analysis, the library of SG was constructed and sequenced on the Illumina platform. After removing the low quality data, the clean reads were mapped to the reference genome, and then the expression level of each gene was determined. Altogether 8042 genes were validated to be expressed in SG with TPM > 1, which accounted for 42.3% of the predicated genes (Table S2). Among them, 456 genes were abundantly expressed with TPM > 100. Noteworthily, as many as 71 ribosomal proteins were abundantly expressed in SG, indicating the active protein synthetic process in this secretory tissue. In addition, 181 proteins were predicted to contain a signal peptide, which was indicative of the secretory property. Additionally, genes associated with hydrolase activity, including serine protease, amylase, lipase, and carboxypeptidase, were abundantly expressed in SG ( Figure 2A; Table S2).
Many salivary proteins are specifically expressed in SG [12]. The transcriptome from R. pedestris guts, fat body, muscles, carcass, testes, and ovaries has been reported in our previous study. Here, this work investigated the SG-specific genes based on the transcriptomic data in our previous study [30] and the present study. A total of 226 genes were found to be specifically expressed in SG (Table S3). Gene ontology (GO) analysis demonstrated that genes associated with catalytic activity, hydrolase activity, and metabolism were significantly enriched ( Figure 2B).
Moreover, Venn diagram analysis between SG-abundant genes and SG-specific genes was conducted, which identified 178 overlapping genes in two groups. Among these genes, 140 were predicated to contain a signal peptide, including four possessing at least one transmembrane domain besides the signal peptide. Such result indicated that these proteins might be embedded in the cell membranes of salivary glands. Taken together, there were 136 secretory genes that were abundantly and specifically expressed in R. pedestris SG (Table S4). Overview of workflow used to identify the defense-induced salivary proteins. The salivary glands isolated from adult Riptortus pedestris were subjected to transcriptomic and proteomic analyses. The secretory proteins that were abundantly and specifically expressed in R. pedestris salivary gland were considered as salivary proteins. Then, the salivary proteins were transiently expressed in Nicotiana benthamiana, and the defense-induced salivary proteins that were capable of inducing cell death, reactive oxygen species (ROS) burst, or hormone signal changes were identified.
Figure 1.
Overview of workflow used to identify the defense-induced salivary proteins. The salivary glands isolated from adult Riptortus pedestris were subjected to transcriptomic and proteomic analyses. The secretory proteins that were abundantly and specifically expressed in R. pedestris salivary gland were considered as salivary proteins. Then, the salivary proteins were transiently expressed in Nicotiana benthamiana, and the defense-induced salivary proteins that were capable of inducing cell death, reactive oxygen species (ROS) burst, or hormone signal changes were identified.
Figure 1.
Overview of workflow used to identify the defense-induced salivary proteins. The salivary glands isolated from adult Riptortus pedestris were subjected to transcriptomic and proteomic analyses. The secretory proteins that were abundantly and specifically expressed in R. pedestris salivary gland were considered as salivary proteins. Then, the salivary proteins were transiently expressed in Nicotiana benthamiana, and the defense-induced salivary proteins that were capable of inducing cell death, reactive oxygen species (ROS) burst, or hormone signal changes were identified.
Overview of the Riptortus pedestris Proteome
The presence of salivary proteins in R. pedestris SG was investigated at the protein level by LC-MS/MS. A total of 775 proteins were detected (Table S5). Among them, 377 (48.6%) had binding functions, including ion binding, cofactor binding, protein binding, and lipid binding; 165 (21.3%) had regulatory functions, including metabolic regulation, signaling regulation, and developmental regulation; 266 (34.3%) were related to hydrolase, oxidoreductase, and lyase activity; whereas 39 (5.0%) were associated with transportation and transduction ( Figure S1). Among the 136 secretory genes that were abundantly and specifically expressed in SG, 88 were found to be translated into proteins (Table S4). For the other 53 genes, no protein was detected, which might be caused by the insensitivity of LC-MS/MS analysis, protein degradation, or the failure to translate these genes, and further investigation was needed.
Characteristics of Salivary Proteins
First of all, this study investigated the developmental expression patterns of 136 predicted salivary proteins. The insects (reared on soybean plants) of 37 different development stages were analyzed. It was found that all the salivary proteins were expressed at the lowest level during the egg periods ( Figure 3). This is the typical expression pattern for most salivary proteins, as the insect does not need to feed before hatching [50]. Most salivary proteins were stably expressed in nymph and adult periods. However, yellow-like and four unannotated proteins (Rp.chr2.1812, Rp.chr3.1239, Rp.chr3.1227, Rp.chr1.0390) were highly expressed in the adult period. It will be interesting to investigate their specific functions in the adult period.
Interestingly, it was found in this study that salivary proteins were unevenly distributed in the genome. More than one half of salivary proteins were located on chromosome 1. In contrast, the sex chromosome did not contain any salivary protein. Additionally, 95 salivary proteins contained two or more amino acid tandem repeats (Table S4). For example, the seven serine rich proteins, with a homologous gene only found in Oncopeltus fasciatus, were tandemly arrayed. Similar phenomena were also observed in venom serine protease, carboxypeptidase, and some R. pedestris-specific genes, indicating that the salivary genes underwent expansion during the long-term evolution process. Interestingly, it was found in this study that salivary proteins were unevenly distributed in the genome. More than one half of salivary proteins were located on chromosome 1. In contrast, the sex chromosome did not contain any salivary protein. Additionally, 95 salivary proteins contained two or more amino acid tandem repeats (Table S4). For example, the seven serine rich proteins, with a homologous gene only found in Oncopeltus fas- All of the R. pedestris salivary proteins were compared with those from 24 other arthropod species (Table S6). Altogether, seven gene families were found to be commonly identified between R. pedestris and other arthropod species, including serine protease, lipase, carboxypeptidase B, protein yellow, amylase, inositol polyphosphate phosphatase, and lectin. Besides, up to 86 proteins were specifically identified in R. pedestris or other evolutionarily related bug species (Table S4). These proteins with unknown function have not been investigated in previous studies. However, their specific roles in R. pedestris and host plant interactions remain further investigation.
Function Analysis of Salivary Proteins
Cell death and ROS burst are considered to play critical roles in the plant defense responses induced by salivary components. To identify the candidate salivary proteins capable of eliciting plant defense signals, the salivary proteins were transiently expressed in N. benthamiana by agrobacterium infiltration. Five salivary proteins were found to be capable of inducing cell death, chlorosis, or ROS burst ( Figure 4). As these proteins were bug-specific with unknown functions, we temporarily annotated them as RpSP10.3, RpSP13.4, RpSP13.8, RpSP17.8, and RpSP30.2 (RpSP is abbreviation of Riptortus pedestris salivary protein) based on their molecular weights. showed that both proteins located specifically in the cytoplasm ( Figure 5). Pathogenesisrelated proteins are induced by the plants as a defense response system in stress conditions like microbial and insect infections [51]. Both RpSP13.4 and RpSP13.8 induced significant upregulation of pathogenesis-related protein (PR1) and PR4 ( Figure 6). Plant hormone regulatory networks, especially the jasmonate (JA) and salicylic acid (SA) pathways, sophisticatedly responded to insect attacks [52]. We found that the JA-responsive gene PDF1.2 was significantly repressed in N. benthamiana transiently expressed with RpSP13.4 and RpSP13.8. In contrast, the SA-responsive gene NPR1 was significantly upregulated in RpSP13.4-expressed leaves when compared with that of GFP control ( Figure 6). Chlorosis phenotypes were observed in N. benthamiana transiently expressed with RpSP10.3 and RpSP17.8. These two proteins also induced ROS burst and the expression of PR1 and PR4. Figure 5). Pathogenesis-related proteins are induced by the plants as a defense response system in stress conditions like microbial and insect infections [51]. Both RpSP13.4 and RpSP13.8 induced significant upregulation of pathogenesis-related protein (PR1) and PR4 ( Figure 6). Plant hormone regulatory networks, especially the jasmonate (JA) and salicylic acid (SA) pathways, sophisticatedly responded to insect attacks [52]. We found that the JA-responsive gene PDF1.2 was significantly repressed in N. benthamiana transiently expressed with RpSP13.4 and RpSP13.8. In contrast, the SA-responsive gene NPR1 was significantly upregulated in RpSP13.4-expressed leaves when compared with that of GFP control ( Figure 6). Chlorosis phenotypes were observed in N. benthamiana transiently expressed with RpSP10.3 and RpSP17.8. These two proteins also induced ROS burst and the expression of PR1 and PR4. RpSP10.3 was mainly expressed in both the cytoplasm and the nucleus, while RpSP17.8 was specifically expressed in the cytoplasm ( Figure 5). A weaker, but significant ROS burst was observed in N. benthamiana transiently expressed with RpSP30.2. RpSP30.2 also induced the expression of PR4 and NPR1. In contrast, the expression of PDF1.2 was significantly repressed in RpSP30.2-expressed leaves ( Figure 6).
Discussion
Saliva contains a mixture of bioactive molecules that play critical roles in mediating the insect-plant interactions. In this study, the salivary proteins from R. pedestris were identified by transcriptomic and proteomic approaches. It was found that a majority of salivary proteins were tandemly arrayed in the genome and showed low expression in the egg stages. Except for proteins involved in hydrolase and binding, many salivary proteins were bug-specific or even R. pedestris-specific, and some of them were capable of inducing plant defenses.
Salivary proteins associated with hydrolase were significantly enriched, including serine protease, trypsin, carboxypeptidase, lipase, nuclease, and amylase. The bean bug uses the soybean seeds and plants as its main food sources, and the protein/oil-rich seeds are necessary to complete the insect life cycle [53]. Proteases and lipases are possibly secreted to digest proteins and oils in seeds. Previously, extra-oral digestion is generally presumed to be critical for carnivorous insects, which can secrete trypsin, chymotrypsin, and other digestive enzymes to decompose solid contents into liquid and increase the food extraction efficiency [54]. In recent years, the extensive identification of protease, amylase, and lipases in herbivorous saliva indicates that extra-oral digestion is also applied for herbivorous insects [55]. However, the abundance of hydrolase in insect saliva varies across species. As for planthopper that mainly depends on plant phloem saps, no trypsin is specifically expressed in SG, and the insect secretes less amounts of proteases during feeding [56]. For R. pedestris, abundant trypsins are specifically expressed in SG at high levels, indicating that the secretion of proteases may be more important for bugs during feeding. In rearing R. pedestris, the bug can smoothly pierce and suck dry the soybean seeds. It was presumed in this study that R. pedestris secreted abundant hydrolases to break down the solid proteins, oils, and amyloses into liquid, and then absorbed this liquified nutrition.
Discussion
Saliva contains a mixture of bioactive molecules that play critical roles in mediating the insect-plant interactions. In this study, the salivary proteins from R. pedestris were identified by transcriptomic and proteomic approaches. It was found that a majority of salivary proteins were tandemly arrayed in the genome and showed low expression in the egg stages. Except for proteins involved in hydrolase and binding, many salivary proteins were bug-specific or even R. pedestris-specific, and some of them were capable of inducing plant defenses.
Salivary proteins associated with hydrolase were significantly enriched, including serine protease, trypsin, carboxypeptidase, lipase, nuclease, and amylase. The bean bug uses the soybean seeds and plants as its main food sources, and the protein/oil-rich seeds are necessary to complete the insect life cycle [53]. Proteases and lipases are possibly secreted to digest proteins and oils in seeds. Previously, extra-oral digestion is generally presumed to be critical for carnivorous insects, which can secrete trypsin, chymotrypsin, and other digestive enzymes to decompose solid contents into liquid and increase the food extraction efficiency [54]. In recent years, the extensive identification of protease, amylase, and lipases in herbivorous saliva indicates that extra-oral digestion is also applied for herbivorous insects [55]. However, the abundance of hydrolase in insect saliva varies across species. As for planthopper that mainly depends on plant phloem saps, no trypsin is specifically expressed in SG, and the insect secretes less amounts of proteases during feeding [56]. For R. pedestris, abundant trypsins are specifically expressed in SG at high levels, indicating that the secretion of proteases may be more important for bugs during feeding. In rearing R. pedestris, the bug can smoothly pierce and suck dry the soybean seeds. It was presumed in this study that R. pedestris secreted abundant hydrolases to break down the solid proteins, oils, and amyloses into liquid, and then absorbed this liquified nutrition.
There were 86 proteins specifically identified in R. pedestris or its evolutionarily related bug species. Although these species-specific proteins have not been well characterized yet, their functions in R. pedestris and plant interactions cannot be ignored. During insect feeding, the secreted saliva intimately interacts with host plants, and the compositions can be easily shaped by natural selection [48]. For the time being, many salivary effectors identified are found to be species-specific. In Nilaparvata lugens, three planthopper-specific proteins (NlSHP, salivap3, and NlMul) are found to be indispensable for insect feeding on rice plants [9,18,50]. In Acyrthosiphon pisum, the critical salivary effectors (C002 and ACYPI009881) are limited to aphid species [21,57,58]. The species-specific salivary protein is possibly the result of the co-evolutional arm race between insects and plants. For example, the whitefly-specific salivary effector Bt56 attenuates plant defense responses by targeting plant NTH202 [59]. Further, the aphid-specific salivary effector Mp1 increases insect virulence by interacting with plant VPS52 [60]. The soybean staygreen-like syndromes induced by R. pedestris infection are significantly different from plant syndromes caused by aphids or whiteflies. Staygreen-like syndromes may be correlated with some bug-specific salivary proteins, which deserves further investigations.
Herbivore-associated molecular patterns derived from saliva have been documented to activate an array of plant defenses against herbivores, including ROS burst, mitogenactivated protein kinase (MAPK) cascade activation, and hormone signal transduction [1,31]. This study is the first to screen the bug-specific salivary components that potentially initiate plant defense responses. According to our results, transient in planta expression of five proteins (RpSP10.3, RpSP13.4, RpSP13.8, RpSP17.8, and RpSP30.2) induced ROS burst in N. benthamiana. Generally, rapid and transient accumulation of ROS is a common strategy for plants to cope with biotic and abiotic stresses [61,62]. It is the first line of barrier against subsequent attack by pathogens and herbivores, which acts as the secondary messengers to control defense responses [61,62]. Therefore, cell death and alteration in the hormone pathways of transfected plants may be positively correlated with ROS burst [20]. Although the five proteins are found to induce plant defenses, there may be other salivary components that suppress these detrimental effects, such as macrophage migration inhibitory factor in M. percicae, [25] and Al6 in Apolygus lucorum [27]. Further studies are warranted to unveil the complex mechanisms of R. pedestris saliva in insect-plant interaction.
Conclusions
This study analyzed the salivary components of R. pedestris by transcriptomic and proteomic approaches. Altogether, 136 secretory genes were discovered to be abundantly and specifically expressed in SG. Besides, it was found that salivary proteins associated with hydrolase potentially contributed to extra-oral digestion, while the bug-specific salivary proteins might be correlated with long-term R. pedestris-plant interaction. Among the bugspecific proteins, RpSP10.3, RpSP13.4, RpSP13.8, RpSP17.8, and RpSP30.2 were identified as the potential salivary effectors that induced plant defenses. The salivary secretome of R. pedestris described in this study helps to understand the molecular mechanisms underlying plant-insect interactions and provides valuable resources for further research.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/biology10080753/s1, Figure S1: Function analysis of proteins detected in salivary glands. The proteins detected in salivary glands were classified into cell component (red), molecular function (orange), and biological process (blue), respectively. Table S1: Primers used in this study. Table S2: The Riptortus pedestris genes expressed in the salivary glands. Table S3: Expression patterns of SG-specific genes in different Riptortus pedestris tissues.
Data Availability Statement:
The data presented in this study are available to public. Raw reads generated by transcriptomic sequencing was submitted to the NCBI Sequence Read Archive under accession number: PRJNA671796. The raw data generated by LC-MS/MS analysis was submitted to Zenodo (https://zenodo.org/ accessed on 2 October 2020) with the http://doi.org/10.5281/zenodo. 5112678.
Conflicts of Interest:
The authors have no competing interest to declare. | 2021-08-28T05:19:31.356Z | 2021-08-01T00:00:00.000 | {
"year": 2021,
"sha1": "3725fabc4af591ed9adc1b4378210ec5123c0458",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2079-7737/10/8/753/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "3725fabc4af591ed9adc1b4378210ec5123c0458",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
21659132 | pes2o/s2orc | v3-fos-license | CAM-chem: Description and evaluation of interactive atmospheric chemistry in the Community Earth System Model
We discuss and evaluate the representation of atmospheric chemistry in the global Community Atmosphere Model (CAM) version 4, the atmospheric component of the Community Earth System Model (CESM). We present a variety of configurations for the representation of tropospheric and stratospheric chemistry, wet removal, and online and offline meteorology. Results from simulations illustrating these configurations are compared with surface, aircraft and satellite observations. Major biases include a negative bias in the high-latitude CO distribution, a positive bias in uppertropospheric/lower-stratospheric ozone, and a positive bias in summertime surface ozone (over the United States and Europe). The tropospheric net chemical ozone production varies significantly between configurations, partly related to variations in stratosphere-troposphere exchange. Aerosol optical depth tends to be underestimated over most regions, while comparison with aerosol surface measurements over the United States indicate reasonable results for sulfate , especially in the online simulation. Other aerosol species exhibit significant biases. Overall, the model-data comparison indicates that the offline simulation driven by GEOS5 meteorological analyses provides the best simulation, possibly due in part to the increased vertical resolution (52 levels instead of 26 for online dynamics). The CAM-chem code as described in this paper, along with all the necessary datasets needed to perform the simulations described here, are available for download at www.cesm.ucar.edu . 1 Introduction Atmospheric chemistry plays an integral role in the distribution of the non-CO2 radiatively active gases and aerosols (Forster et al., 2007). In addition, climate and its evolution strongly influences atmospheric chemistry and air quality (Chen et al., 2009; Jacob and Winner, 2009). Because of the nonlinear behavior of chemistry and the importance of regionally-varying emissions, it is critical to represent the interactions between chemistry and climate using a global three-dimensional model. We discuss here the representation of atmospheric chemistry (gas phase and aerosol species) in the global Community Atmosphere Model (CAM version 4, Neale et al., 2011), the atmospheric component of the Community Earth System Model (CESM). Due to its full integration in the CESM, CAM-chem provides the flexibility of using the same code to perform climate simulations and simulations with specified meteorological fields. Therefore, CAM-chem can be used in three separate modes, all embedded within CESM: (1) a fully coupled Earth System model (i.e., with all climate components active, which offers the possibility to connect the chemistry with biogeochemical processes in the land and ocean models), (2) with specified sea-surface and seaice distributions and (3) with specified meteorological fields. Configurations (1) and (2) are usually referred to as online, while configuration (3) is referred to as offline. In configurations (2) and (3), only the atmosphere and land components of CESM are active. When run in a fully coupled mode, the ocean and ice dynamics are also allowed to respond to Published by Copernicus Publications on behalf of the European Geosciences Union. 370 J.-F. Lamarque et al.: CAM-chem description and evaluation Atmosphere
Model
version
4
Introduction
Atmospheric chemistry plays an integral role in the distribution of the non-CO 2 radiatively active gases and aerosols (Forster et al., 2007). In addition, climate and its evolution strongly influences atmospheric chemistry and air quality (Chen et al., 2009;Jacob and Winner, 2009). Because of the nonlinear behavior of chemistry and the importance of regionally-varying emissions, it is critical to represent the interactions between chemistry and climate using a global three-dimensional model.
We discuss here the representation of atmospheric chemistry (gas phase and aerosol species) in the global Community Atmosphere Model (CAM version 4, Neale et al., 2011), the atmospheric component of the Community Earth System Model (CESM). Due to its full integration in the CESM, CAM-chem provides the flexibility of using the same code to perform climate simulations and simulations with specified meteorological fields. Therefore, CAM-chem can be used in three separate modes, all embedded within CESM: (1) a fully coupled Earth System model (i.e., with all climate components active, which offers the possibility to connect the chemistry with biogeochemical processes in the land and ocean models), (2) with specified sea-surface and seaice distributions and (3) with specified meteorological fields. Configurations (1) and (2) are usually referred to as online, while configuration (3) is referred to as offline. In configurations (2) and (3), only the atmosphere and land components of CESM are active. When run in a fully coupled mode, changes in the atmosphere. This allows "slow responses" (for example in ocean currents and ice extent) within the climate system to occur that are prevented in modes 2 and 3. Mode 2 only allows fast atmospheric responses, and land responses to occur, while mode 3 does not allow any meteorological responses.
The availability of a specified dynamics configuration of CAM-chem offers a number of advantages over traditional chemistry-transport models. (1) It allows for consistent simulations between the online and offline versions; (2) it allows the offline version to be run in an Earth System framework so as to fully exploit other model components. In particular it allows an incorporation of the biogeochemical algorithms in the land components; (3) it allows for the radiative algorithms incorporated into CAM4 to be fully exploited in the offline version. This allows a calculation of the instantaneous radiative forcing within the offline model version, including a calculation of the instantaneous radiative forcing for specific events (e.g., forest fires) (Pfister et al., 2008;Randerson et al., 2006); (4) it allows for the strict conservation of tracer mass by accounting for changes in mixing ratio as the water vapor concentration changes within the atmosphere. In addition, CAM-chem uses a chemical preprocessor that provides extensive flexibility in the definition of the chemical mechanism, allowing for ease of update and modification.
It is the purpose of this paper to document all these aspects of CAM-chem, along with results from associated simulations and their evaluation against observations. The paper is therefore organized as follows: in Sect. 2, we provide a short introduction to CAM4. In Sect. 3, we discuss all the chemistry-specific parameterizations of CAM-chem. The implementation of the specified-dynamics version is discussed in Sect. 4. Section 5 presents the various chemical mechanisms used in this study. Model simulation setups, including emissions are discussed in Sect. 6, while the comparison to observations is shown in Sect. 7. Discussion and conclusions are in Sect. 8.
CAM4 description and definition of CAM-chem
CAM-chem refers to the implementation of atmospheric chemistry in the Community Earth System Model (CESM). All the subroutines responsible for the representation of chemistry are included in the build of CESM only when explicitly requested; therefore a user only interested in a climate simulation (for which the atmospheric composition is specified) is not impacted by the additional code and additional cost to simulate chemistry. The chemistry is fully integrated into the Community Atmosphere Model (Fig. 1), meaning that the representation of dynamics (including transport) and physics (radiation, convection and large-scale precipitation, boundary-layer and diffusion) is the same whether the model is using computed (online) or specified (specified dynamics, dotted box in Fig. 1) meteorological fields. CAM-chem is therefore CAM4 with chemistry activated.
In the configurations described in this paper, atmospheric chemistry interacts with the climate only through radiation since no cloud-aerosol interaction is available in CAM4; the impact on temperature is of course overwritten in the case of specified dynamics. Also, the ocean and sea-ice components of the CESM are inactive due to the use of specified seasurface temperatures and sea-ice distribution. On the other hand, the land component is fully active and provides the chemistry with deposition velocities and biogenic emissions. When CAM4 is run without chemistry, it uses prescribed ozone and aerosol fields, usually monthly-averaged threedimensional fields from a previously performed simulation with CAM-chem. The Community Atmosphere Model, version 4 (CAM4, Neale et al., 2011) was released as the atmosphere component of the Community Climate System Model, version 4 (CCSM4) and contains improvements over CAM3 . In particular, the finite volume dynamical core option available in CAM3 is now the default primarily due to its superior tracer transport properties . As in CAM3, deep convection is parameterized using the Zhang-McFarlane approach (1995), but with modifications as discussed below, while shallow convection follows Hack et al. (2006). Additional information on the representation of clouds and precipitation processes can be found in Boville et al. (2006). Finally, processes in the planetary boundary layer are represented using the Holtslag and Boville (1993) parameterization. The vertical coordinate is a hybrid sigmapressure.
Changes made to the representation of deep convection include a dilute plume calculation and the introduction of Convective Momentum Transport (CMT; Richter and Rasch, 2007;Neale et al., 2008). In addition, the cloud fraction has an additional calculation to improve thermodynamic consistency. A freeze-drying modification is further made to the cloud fraction calculation in very dry environments, such as Arctic winter, where cloud fraction and cloud water estimates were somewhat inconsistent in CAM3.
Altogether, only marginal improvements over CAM3 are found in the large-scale aspects of the simulated climate (see Neale et al., 2011 for a complete description of the model performance). Indeed, it is found that the implementation of the finite volume dynamical core leads to a degradation in the excessive trade-wind simulation, but with an accompanying reduction in zonal stresses at higher latitudes. But, CMT reduces much of the excessive trade-wind biases. Plume dilution leads to moister deep tropics alleviating much of the mid-tropospheric dry biases and reduces the persistent precipitation biases over the Arabian peninsula and the southern Indian ocean associated with the Indian Monsoon. Finally, the freeze-drying modification alleviates much of the wintertime excessive cloud bias and improves the associated surface cloud-related energy budget.
Chemistry-specific parameterizations
CAM-chem borrows heavily from MOZART-4 . In particular, many of the parameterizations needed to represent atmospheric chemistry in a global model are adapted or expanded from their equivalents in MOZART-4. However, for completeness, we will include a brief description of those parameterizations (and their updates, whenever applicable).
Dry deposition
Dry deposition is represented following the resistance approach originally described in Wesely (1989); as discussed in E2010, this earlier paper was subsequently updated and we have included all updates (Walcek et al., 1986;Walmsley and Wesely, 1996;Wesely and Hicks, 2000). Following this approach, all deposited chemical species (the specific list of deposited species is defined along with the chemical mechanisms, see Sect. 4) are mapped to a weighted-combination of ozone and sulfur dioxide depositions; this combination represents a definition of the ability of each considered species to oxidize or to be taken up by water. In particular, the latter is dependent on the effective Henry's law coefficient. While this weighting is applicable to many species, we have included specific representations for CO/H 2 (Yonemura et al., 2000;Sanderson et al., 2003), and peroxyacetylnitrate (PAN, Sparks et al., 2003;Turnipseed et al., 2006). Furthermore, it is assumed that the surface resistance for SO 2 can be neglected (Walcek et al., 1986). Finally, following Cooke et al. (1999), the deposition velocities of black and organic carbonaceous aerosols are specified to be 0.1 cm s −1 over all surfaces. Dust and sea-salt are represented following Mahowald et al. (2006a, b).
In particular, the computation of surface resistances in CLM leads to a representation at the level of each plant functional type (Table 1) weighted-mean over all land cover types available at each grid box. This ensures that the impact on deposition velocities from changes in land cover, land use or climate can be taken into account.
In addition, the same dry deposition approach is separately applied to water surfaces i.e., lakes and oceans, including sea-ice. It is then combined with the land-based value, weighted by the ocean/sea-ice and land fractions in each model grid cell.
Biogenic emissions
Similar to the treatment of dry deposition over land, biogenic emissions of volatile organic compounds (isoprene and monoterpenes) are calculated based upon the land cover. These are made available for atmospheric chemistry, unless the user decides to explicitly set those emissions using predefined (i.e., contained in a file) gridded values. Details of this implementation in the CLM3 are discussed in Heald et al. (2008); we provide a brief overview here.
Vegetation in the CLM model is described by 17 plant function types (PFTs, see Table 1). Present-day land surface parameters such as leaf area index are consistent with MODIS land surface data sets (Lawrence and Chase, 2007). Alternate land cover and density can be either specified or interactively simulated with the dynamic vegetation model of the CLM for any time period of interest.
Isoprene emissions follow the MEGAN2 algorithms for a detailed canopy model (CLM). This includes mapped PFT-specific emission factors to account for species divergent emissions of isoprene. These standard emission factors are modulated by activity factors accounting for the effect of temperature, radiation, leaf age, vegetation density (identified by the leaf-area index) and soil moisture. The annual totals are in the range of 500-600 Tg yr −1 , depending on model configuration and associated climate conditions (see Table 9).
Total monoterpene emissions follow the earlier work of Guenther et al. (1995) as implemented in the CLM by Levis et al. (2003). Baseline emission factors are specified for each plant function type and are scaled by an exponential function of leaf temperature.
Wet deposition
Wet removal of soluble gas-phase species is the combination of two processes: in-cloud, or nucleation scavenging (rainout), which is the local uptake of soluble gases and aerosols by the formation of initial cloud droplets and their conversion to precipitation, and below-cloud, or impaction scavenging (washout), which is the collection of soluble species from the interstitial air by falling droplets or from the liquid phase via accretion processes (e.g., Rotstayn and Lohmann, 2002). Removal is modeled as a simple first-order loss process X iscav =X i × F × (1 − exp(−λ t)). In this formula, X iscav is the species mass (in kg) of X i scavenged in time step t, F is the fraction of the grid box from which tracer is being removed, and λ is the loss rate. In-cloud scavenging is proportional to the amount of condensate converted to precipitation, and the loss rate depends on the amount of cloud water, the rate of precipitation formation, and the rate of tracer uptake by the liquid phase. Below-cloud scavenging is proportional to the precipitation flux in each layer and the loss rate depends on the precipitation rate and either the rate of tracer uptake by the liquid phase (for accretion processes), the mass-transfer rate (for highly soluble gases and small aerosols), or the collision rate (for larger aerosols). In CAMchem two separate parameterizations are available: Horowitz et al. (2003) from MOZART-2 and Neu and Prather (2011). The distinguishing features of the Neu and Prather scheme are related to three aspects of the parameterization: (1) the partitioning between in-cloud and below cloud scavenging, (2) the treatment of soluble gas uptake by ice and (3) accounting for the spatial distribution of clouds in a column and the overlap of condensate and precipitation. Given a cloud fraction and precipitation rate in each layer, the scheme determines the fraction of the gridbox exposed to precipitation from above and that exposed to new precipitation formation under the assumption of maximum overlap of the precipitating fraction. Each model level is partitioned into as many as four sections, each with a gridbox fraction, precipitation rate, and precipitation diameter: (1) cloudy with precipitation falling through from above; (2) cloudy with no precipitation falling through from above; (3) clear sky with precipitation falling through from above; (4) clear sky with no precipitation falling from above. Any new precipitation formation is spread evenly between the cloudy fractions (1 and 2). In region 3, we assume a constant rate of evaporation that reduces both the precipitation area and amount so that the rain rate remains constant. Between levels, we average the properties of the precipitation and retain only two categories, precipitation falling into cloud and precipitation falling into ambient air, at the top boundary of each level. If the precipitation rate drops to zero, we assume full evaporation and random overlap with any precipitating levels below. Our partitioning of each level and overlap assumptions are in many ways similar to those used for the moist physics in the ECMWF model (Jakob and Klein, 2000).
The transfer of soluble gases into liquid condensate is calculated using Henry's Law, assuming equilibrium between the gas and liquid phase. Nucleation scavenging by ice, however, is treated as a burial process in which trace gas species deposit on the surface along with water vapor and are buried as the ice crystal grows. Kärcher and Voigt (2006) have found that the burial model successfully reproduces the molar ratio of HNO 3 to H 2 O on ice crystals as a function of temperature for a large number of aircraft campaigns spanning a wide variety of meteorological conditions. We use the empirical relationship between the HNO 3 :H 2 O molar ratio and temperature given by Kärcher and Voigt (2006) to determine in-cloud scavenging during ice particle formation, which is applied to nitric acid only. Below-cloud scavenging by ice is calculated using a rough representation of the riming process modeled as a collision-limited first order loss process. Neu and Prather (2011) provide a full description of the scavenging algorithm. On the other hand, the Horowitz approach uses the rain generation diagnostics from the large-scale and convection precipitation parameterizations in CAM; equilibrium between gas-phase and liquid phase is then assumed based on the effective Henry's law. Below-cloud removal is applied to nitric acid and hydrogen peroxide only.
While using the same information on rain formation and precipitation, the wet removal of aerosols is handled separately, using the parameterization described in Barth et al. (2000).
Lightning
The emissions of NO from lightning are included as in E2010, i.e., using the Price parameterization (Price and Rind, 1992;Price et al., 1997), scaled to provide a global annual emission of 3-5 Tg(N) yr −1 (see Table 9), slightly lower than the global estimate of Hudman et al. (2007). This range is due to interannnual variability in convective activity. The vertical distribution follows DeCaria et al. (2006) as in E2010. In addition, the strength of intra-cloud (IC) lightning strikes is assumed to be equal to cloud-to-ground strikes, as recommended by Ridley et al. (2005).
Polar stratospheric clouds and associated ozone depletion
The representation of polar stratospheric clouds is an update over the version used in all the CCMval-2 analysis papers (e.g., Austin et al., 2010) in which it was shown that CAMchem (identified in those studies as CAM3.5) was underestimating the extent and depth of Antarctic ozone hole depletion. In particular, we are now using a strict enforcement of the conservation of total (organic and inorganic) chlorine and total bromine under advection. Indeed, it has been identified that the existence of strong gradients in the stratosphere led to non-conservation issues of the total bromine and chlorine, as computed from the sum of their components, related to inaccuracies in transport algorithms. We are therefore forcing the conservation through the addition of two additional tracers: TCly and TBry. These tracers are specified at the lower boundary and reflect the total amount of Br and Cl atoms (organic and inorganic) in the atmosphere following the observed concentrations of all considered halogen species in the model. To ensure mass conservation, at each grid point the total mass after advection of the summed Cl-containing species is scaled to be the same as the mass of TCly. Uniform scaling is applied to each component. The overall impact is to increase the amount of reactive bromine and chlorine in the polar stratosphere and, consequently, of ozone loss.
In addition, we have updated the heterogeneous chemistry module to reflect that the model was underestimating the supercooled ternary solution (STS) surface area density (SAD). Heterogeneous processes on liquid sulfate aerosols and polar stratospheric clouds are included following the approach of Considine et al. (2000). This approach represents the surface area density, effective radius, and composition of liquid binary sulfate (LBS), supercooled ternary solution (STS), nitric acid tri-hydrate (NAT), and water-ice. There are six heterogeneous reactions on liquid sulfate aerosol (LBS or STS), five reactions on solid NAT aerosol, and six reactions on solid water-ice aerosol. The process of denitrification is derived in the chemistry module; the process of dehydration is derived in the prognostic H 2 O approach used in CAM. Details of the heterogeneous module are discussed in Kinnison et al., 2007. This previous version (used in the CCMval-2 simulations) allowed the available HNO 3 to first form nitric acid trihydrate (NAT); then the resulting gas-phase HNO 3 was available to form STS SAD. In the new version the approach is reversed, with the available HNO 3 to first form STS aerosol. This enhances the STS SAD that is used to derive heterogeneous conversion of reservoir species to more active, odd-oxygen Table 3. List of species in the considered chemical mechanisms. In addition, we list the chemical solver (Explicit or Implicit), the potential use of emissions of lower-boundary conditions and of deposition processes (wet and dry). Observational studies have shown that STS is the main PSC for odd-oxygen loss and therefore this is a better representation of stratospheric heterogeneous processes (e.g., Lowe and MacKenzie, 2008). After formation of STS aerosol, there is still enough gas-phase HNO 3 available to form NAT. The effective radius of NAT is then used to settle the condensed phase HNO 3 and eventual irreversible denitrification occurs. These updates have led to a considerable improvement in the representation of the polar stratospheric ozone loss (see Sect. 7.4).
Photolysis
In CAM-chem, for wavelengths longer than 200 nm (up to 750 nm), the lookup table approach from MOZART-3 (Kinnison et al., 2007) is the only method available at this time.
We have also included the online calculation of photolysis rates for wavelengths shorter than 200 nm (121-200 nm) from MOZART-3; this was shown to be important for ozone chemistry in the tropical upper troposphere (Prather, 2009). Therefore, the combined (online-lookup table) approach is used in all model configurations. In addition, because the standard configuration of CAM only extends into the lower stratosphere (model top is ≈ 40 km), we have included an additional layer of ozone and oxygen above the model top to provide a very accurate representation of photolysis rates in the upper portion of the model (Fig. 2) as compared to the equivalent calculation using a fully-resolved stratospheric distribution. The fully resolved stratospheric module was evaluated in Chapter 6 of the SPARC CCMVal report on the Evaluation of Chemistry-Climate Models. This photolysis module was shown to be one of the more accurate modules used in CCMs (see SPARC, 2010 for details). Table 4. List of reactions. Temperature (T ) is expressed in K, pressure (P ) in Pa, air density (M) in molec cm −3 , ki and ko in cm 3 molec −1 s −1 .
The impact of clouds on photolysis rates is parameterized following Madronich (1987). However, because we use a lookup table approach, the impact of aerosols (tropospheric or stratospheric) on photolysis rates cannot be represented.
Offline meteorology and transport
CAM-chem has the capability to perform simulations using specified dynamics, where offline meteorological fields are input into the model instead of calculated online. This procedure can allow for more precise comparisons between measurements of atmospheric composition and model output. To use input meteorological fields we follow the same procedure defined originally in the Model of Atmospheric Transport and Chemistry (MATCH) (Rasch et al., 1997) and subsequently applied in all versions of MOZART (E2010; Kinnison et al., 2007Brasseur et al., 1998). In this procedure only the horizontal wind components, air temperature, surface temperature, surface pressure, sensible and latent heat flux and wind stress (see Table 2) are read from the input meteorological dataset; in all cases discussed here, the input datasets are available every 6 h. For timesteps between the reading times, all fields are linearly interpolated to avoid jumps. These fields are subsequently used to internally generate (using the existing CAM4 parameterizations) the variables necessary for (1) calculating subgrid scale transport including boundary layer transport and convective transport; (2) the variables necessary for specifying the hydrological cycle, including cloud and water vapor distributions and rainfall (see Rasch et al., 2007 for more details).
CAM4 (and therefore CAM-chem) uses a sub-stepping procedure to solve the advection equations for the mass flux, temperature and velocity fields over each timestep. Regardless of the configuration (online or specified dynamics), the meteorological fields are allowed to evolve over each substep. When using specified dynamics, we have found that this sub-stepping dampens some of the inconsistencies between the inserted and model-computed velocity and mass fields subsequently used for tracer transport. The mass flux (atmospheric mass and tracer mass) at each sub-step is accumulated to produce the net mass flux over the entire time step. This allows transport to be performed using a longer time step than the dynamics computations. A graphical explanation of the sub-stepping is given in Lauritzen et al. (2011).
In addition, an atmospheric mass fixer algorithm is necessary as the mass flux computed from the offline meteorological winds input into CAM4 will not in general produce a mass distribution consistent with the offline surface pressure field. If uncorrected this may lead to spurious changes in tracer mass, concentration or surface pressure (Rotman et al., 2004). The mass fixer algorithm ensures that the calculated surface pressure closely matches the surface pressure in the offline meteorological dataset (see discussion in Rotman et al., 2004). The mass fixer algorithm makes the appropriate adjustments to the horizontal mass fluxes to produce a resulting mass distribution consistent with the evolution of surface pressure in the input meteorological dataset. The procedure follows the algorithm given in Rotman et al. (2004): first it uses an efficient algorithm to find the correction to the vertically integrated mass flux, then the corrected mass flux is distributed in the vertical in proportion to the dependence of each model level on the surface pressure in a hybrid coordinate system. The edge pressures of the Lagrangian mass surfaces are consistently adjusted to allow for the vertical remapping of the transported fields to the fixed hybrid pressure coordinate system. Following the corrections in mass flux and the edge pressures the constituent tracers are transported by the large-scale wind fields.
Currently, we recommend using the NASA Goddard Global Modeling and Assimilation Office (GMAO) GEOS5 generated meteorology. The meteorological fields were generated using the operational forecast model and datasets (labeled below GEOS5) or under the Modern Era Retrospective-Analysis For Research And Applications (MERRA) setup (Rienecker et al., 2011). These differ in their assimilation methods and, to a lesser extent, assimilated datasets; see Rienecker et al. (2011) for more details. Using either of these meteorological datasets and the formulation of offline CAM-chem as described above, multi-year simulations (see Sect. 5 and Fig. 7) do not seem to require the use of limiters of stratosphere-troposphere exchange such as SYNOZ (McLinden et al., 2000). All GEOS5/MERRA meteorological datasets used in this study are made available at the standard CAM resolution of 1.9 • × 2.5 • on the Earth System Grid (http://www.earthsystemgrid.org/home. htm). These files were generated from the original resolution (1/2 • × 2/3 • ) by using a conservative regridding procedure based on the same 1-D operators as used in the transport scheme of the finite-volume dynamical core used in GEOS5/MERRA and CAM (S.-J. Lin, personal communication, 2009). Note that because of a difference in the sign convention of the surface wind stress (TAUX and TAUY) between CESM and GEOS5/MERRA, these fields in the interpolated datasets have been reversed from the original files supplied by GMAO. In addition, it is important for users to recognize the importance of specifying the correct surface geopotential height (PHIS) to ensure consistency with the input dynamical fields, which is important to prevent unrealistic vertical mixing. Over the range between the surface and 4 hPa, the number of vertical levels in the GEOS5/MERRA fields is 56, instead of 26 for the online dynamics (see Fig. 3). In particular, between 800 hPa and the surface, GEOS5/MERRA has 13 levels while the online version has 4. The choice of 26 levels is dictated by the use of CAM4 in climate simulations. No adjustment is made to the CAM4 physics parameterizations for this increase in the number of levels.
Chemical mechanisms
As mentioned in the Introduction, CAM-chem uses the same chemical preprocessor as MOZART-4. This preprocessor generates Fortran code specific to each chemical mechanism, allowing for an easy update and modification of existing chemical mechanisms. In particular, the generated code provides two chemical solvers, one explicit and one semiimplicit, which the user specifies based on the chemical lifetime of each species. Because the semi-implicit solver is quite efficient, it is recommended to preferentially use it unless the chemical species has a long lifetime everywhere.
We describe in this paper two chemical mechanisms, (1) extensive tropospheric chemistry, and (2) extensive tropospheric and stratospheric chemistry. All species and reactions are listed in Tables 3 and 4, respectively. In both mechanisms described in this paper, CAM-chem uses the bulk aerosol model discussed in Lamarque et al. (2005) and E2010. This model has a representation of aerosols based on the work by Tie et al. (2001Tie et al. ( , 2005, i.e., sulfate aerosol is formed by the oxidation of SO 2 in the gas phase (by reaction with the hydroxyl radical) and in the aqueous phase (by reaction with ozone and hydrogen peroxide). Furthermore, the model includes a representation of ammonium nitrate that is dependent on the amount of sulfate present in the air mass following the parameterization of gas/aerosol partitioning by Metzger et al. (2002). Because only the bulk mass is calculated, a lognormal distribution (Table 5, also see E2010) is assumed for all aerosols, with different mean radius and geometric standard deviation (Liao et al., 2003) for each aerosol type. The conversion of carbonaceous aerosols (organic and black) from hydrophobic to hydrophilic is assumed to occur within a fixed 1.6 days . Natural aerosols (desert dust and sea salt) are implemented following Mahowald et al. (2006a, b), and the sources of these aerosols are derived based on the model calculated wind speed and surface conditions. Size-dependent gravitational settling is included for dust and sea-salt. In addition, secondary-organic aerosols (SOA) are linked to the gasphase chemistry through the oxidation of atmospheric nonmethane hydrocarbons (NMHCs) as in Lack et al. (2004). Note that, because of the representation of SOA formation through a 2-product method, there is no simple way to represent that in our chemical mechanism (and therefore Table 4); this calculation is therefore performed in a separate portion of the code, using information from the chemical mechanism (rate of VOC oxidation).
The extensive tropospheric chemistry scheme represents a minor update to the MOZART-4 mechanism, fully described in E2010. In particular, we have included chemical reactions for C 2 H 2 , HCOOH, HCN and CH 3 CN and minor changes to the isoprene oxidation scheme, including an increase in the production of glyoxal. Reaction rates have been updated to JPL-2006(Sander et al., 2006. This mechanism is mainly of relevance in the troposphere and is intended for simulations for which variability in the stratospheric composition is not crucial. Therefore, in this configuration, the stratospheric distributions of long-lived species (see discussion below) are specified from previously performed WACCM simulations (Garcia et al., 2007;see Sect. 6.3).
On the other hand, in the case where changes in stratospheric composition are important, for which the dynamics is calculated online, we have added a stratospheric portion to the tropospheric chemistry mechanism described above. This addition (of species and reactions, see Tables 3 and 4) is taken from the WACCM mechanism as this has been shown to perform very well in the recent CCMval-2 analysis (SPARC, 2010). The overall description of this chemistry is discussed in Kinnison et al. (2011).
Emissions
Available with the distribution of the CAM-chem are emissions for tropospheric chemistry that are an extension of the datasets discussed in E2010, covering 1992-2010 tively (http://www.mnp.nl/edgar/). For Asia, these inventories have been replaced by the Regional Emission inventory for Asia (REAS) with the corresponding annual inventory for each year simulated (Ohara et al., 2007 (Kloster et al., 2006). SO 2 emissions from continuously outgassing volcanoes are from the GEIAv1 inventory (Andres and Kasgnoc, 1998). Totals for each year and emitted species are listed in Table 7. All emissions but volcanoes are released in the model bottom layer and implemented as a flux boundary condition for the vertical diffusion. Note that while the emissions are provided at the model resolution, any emissions resolution can be used and the model automatically interpolates to the model resolution. At this point, this interpolation is a simple bilinear interpolation and therefore does not ensure exact conservation of emissions between resolutions. Errors are usually small and limited to areas of strong gradients.
Lower boundary conditions
For all long-lived species (see Table 3), the surface concentrations are specified using the historical reconstruction from Meinshausen et al. (2011). In addition, for CO 2 and CH 4 , an observationally-based seasonal cycle and latitudinal gradient are imposed on the annual average values provided by Meinshausen et al. (2011). These values are used in the model by overwriting at each time step the corresponding model mixing ratio in the lowest model level with the time (and latitude, if applicable) interpolated specified mixing ratio.
Specified stratospheric distributions
In the case where no stratospheric chemistry is explicitly represented in the model, it is necessary to ensure a proper distribution of some chemically-active stratospheric (namely O 3 , NO, NO 2 , HNO 3 , CO, CH 4 , N 2 O, and N 2 O 5 ) species, as is the case for MOZART-4. This monthly-mean climatological distribution is obtained from WACCM simulations covering 1950covering -2005covering (Garcia et al., 2007. Because of the vast changes that occur over that time period, our data distribution provides files for three separate periods : 1950-1959, 1980-1989 and 1996-2005. This ensures that users can perform simulations with a stratospheric climatology representative of the pre-CFC era, as well as during the high CFC and post-Pinatubo era. Note that additional datasets can easily be constructed if necessary. While transport and chemistry are applied to all species in the stratosphere, the concentration of the species listed above are explicitly overwritten from the model top to 50 hPa. Between that level and two model levels above the tropopause (computed from the temperature profile), a 10-day relaxation is applied to force the model concentrations towards the observations.
Comparison with observations
The purpose of this section is to document the model chemistry performance against observations for the model setups described in Sect. 6 (see Table 6 for a summary). Model performance in simulating climate and meteorological features can be found in Lamarque et al. (2008), Lamarque and Solomon (2010) and in Neale et al. (2011). Here, we contrast the fields generated by CAM4 in the simulations listed in Table 6.
The zonal mean distribution (Fig. 4a) of relative humidity is very similar between GEOS5 (considered here as the reference since it is the operational forecast product and assimilates the most observations) and MERRA, except in the Northern Hemisphere polar stratosphere and in the lower troposphere, where MERRA is slightly drier than GEOS5 (not shown). On the other hand, it is clear that the online distribution is wetter in the tropical lower troposphere, and drier in the tropical upper troposphere.
The zonal mean wind is essentially the same in the GEOS5 and MERRA simulations, but the online polar jets are stronger in both hemispheres, leading to a more isolated polar stratosphere. Other features, such as the position and strength of the mid-latitude jet, are very similar between the simulations.
In terms of temperature, the online configuration is characterized by a slightly higher-altitude tropical minimum, as well as a colder Southern Hemisphere polar stratosphere, related to the stronger jet. We also find that the online simulation tends to be colder in the Northern polar lower stratosphere. The online produces a shift of the tropical tropopause, with a warming in the upper-troposphere and cooling in the lower stratosphere, similar to the findings of Collins et al. (2011) when using interactive ozone. In the lower troposphere (700 hPa, Fig. 4b), there is no clear indication of a bias, except for slightly warmer land areas of the Northern Hemisphere (1-2 K, not shown) in online compared to GEOS5/MERRA. Many of the precipitation patterns are similar between the configurations (Fig. 4b), but the difference with respect to GEOS5 indicates that MERRA has slightly different tropical structure (stronger precipitation) while the online configuration has an overall stronger precipitation in both the tropical and midlatitude ocean regions. Land masses tend to be drier in the online version, except for Central Africa, China and the Himalayas.
The chemical composition evaluation below makes use of a variety of measurements: surface, airborne and satellite. In the case of the online stratosphere-troposphere version, the comparison will include evaluation of modeled total ozone column. It is important to note that, because the online model is only driven by the observed sea-surface temperatures, there is no expectation that a single-year in the model simulation will be directly comparable with observations; instead, the most meaningful comparison is at the climatological level.
Comparison with ozone: sondes and surface
Due to its central role in tropospheric chemistry and the availability of numerous ozone sonde measurements dating several decades (Logan, 1994), we focus our first evaluation on tropospheric ozone using ozone sonde measurements, both at specific locations and averaged over representative regions (supplement Fig. S1; Tilmes et al., 2011). For a variety of sites spanning the whole globe (especially in the meridional direction), the data coverage allows the comparison of profiles (Fig. S2), seasonal cycles (Fig. S3) and long-term changes (Fig. 7).
In order to provide a more concise description of the model performance under various configurations, we first display the annual bias at specific pressure levels (250 hPa, 500 hPa and 900 hPa) to span the troposphere and lower stratosphere (Fig. 5). In particular, at high latitudes, the 250 hPa surface will be located in the stratosphere during a fraction of the year.
We find that the model tends to underestimate the ozone concentration at 250 hPa in the high latitudes. On the other hand, most of the mid-latitude sites indicate a positive bias. This is an indication that the model seems to provide a position of the chemical tropopause that is lower than observed, i.e., an overestimate of the ozone mixing ratio in the lower stratosphere (i.e., 200-300 hPa, see Fig. 5). This is confirmed by the Taylor diagrams (generated using regional averages of ozone sondes and equivalent model results from monthly output), which indicate that all versions are quite similar at 250 hPa (Fig. 6). In particular, the seasonal cycle (quantified by the correlation coefficient, computed using monthlyaveraged data and model output) ranges between 0.7 and 0.9, indicating a reasonable representation of ozone variations at 250 hPa.
At 500 hPa, GEOS5 is clearly providing the best performance, with many stations with a bias smaller in absolute value than 5 ppbv. It is interesting to note that MERRA behaves quite differently than GEOS5. This is most likely due to differences in the assimilation method and datasets used. In particular, as discussed below, the MERRA meteorology leads to a much stronger stratosphere-troposphere flux of ozone, likely leading to the 500 hPa biases. In terms of bias, the online simulation performs better than MERRA, especially in the Southern Hemisphere. However, the correlation is much worse (0.5-0.6) for many of the Northern Hemisphere stations. The seasonal cycle in both MERRA and GEOS5 seems equally good. In the lower troposphere (900 hPa) the various configurations tend to exhibit similar regional annual biases, especially in the Northern Hemisphere. Correlation of the annual cycle is, for most regions, in the range 0.6-0.9, with some clear misrepresentations, especially the MERRA simulation over North America. Surface ozone is discussed below.
The analysis of long-term changes (Fig. 7) indicates that observed meteorology is important in representing interannual variability, especially in the upper troposphere, as the correlation for GEOS5/MERRA tends to be higher than in the online configuration (Table 8; stations were selected for the availability of fairly continuous long-term records). Nevertheless, from Table 8, the bias at 250 hPa seems to be better captured by the online simulation for the analysis stations equatorward of 40 • , possibly related to consistently representing transport and chemistry. Overall, Table 8 confirms the previous analysis that, in the free troposphere (500 hPa), GEOS5 provides the best representation of ozone. That configuration also tends to provide a better simulation at 800 hPa, with a lower mean bias and usually high (>0.7) correlation coefficient. The seasonal cycle at the tropical station San Cristobal (Galapagos) is clearly misrepresented at that altitude in all configurations.
The annual budget for tropospheric ozone is summarized in Table 9; note that these are averaged numbers, with an interannual variability on the order of 10 %. We find that the online and offline versions have similar tropospheric (defined here as the region of the atmosphere where the ozone mixing ratio is lower than 100 ppbv) burdens and depositions, but with an overall smaller net chemical ozone production in the case of the offline meteorology. The corollary to this comparison is that the diagnosed stratosphere-troposphere flux of ozone (computed as the difference between the deposition and net chemical production) ranges from 410-420 Tg yr −1 (online and GEOS5) to 675 Tg yr −1 (MERRA), within the range of published modelderived estimates (e.g., 515-550 Tg yr −1 , Hsu et al., 2005; 556 ± 154 Tg yr −1 , Stevenson et al., 2005). In the case of the online simulation, this leads to an ozone lifetime of ≈ 27 days, in very good agreement with Stevenson et al. (2005).
To discuss surface ozone, we present in Fig. 8a and b the comparison of summertime (June-August) daily 8-h maximum (usually afternoon) for the United States (from the CASTNET network) and Europe (from the EMEP network). Because of the very different chemical regimes between Western and Eastern United States, we have separated the stations using longitude 100 • W as the line of demarcation. We find that all model configurations tend to reproduce the Western sites quite well, with a propensity for the online configuration to be slightly higher. Over the Eastern United States, all configurations are biased high, with the MERRA configuration leading to the highest biases (40-60 ppbv). It is unclear why ozone is biased high over those regions and is likely to be a combination of incorrect emissions, coarse resolution and misrepresentation of physical processes (Lin et al. 2008). It is less likely to be meteorology-driven since online and specified dynamics behave similarly. Over Europe (Fig. 8b), the model also tends to overestimate surface ozone. It is however clear that the online meteorology provides a much more biased ozone distribution (as can be seen in Fig. 6, region #5) than the specified dynamics. It is interesting to note that, in the model, the concentrations saturate at 80-90 ppbv, depending on the configuration, unlike the United States sites. Further analysis is necessary to understand this behavior.
Comparison with aircraft observations
As a standard benchmark evaluation, we have performed comparison with the aircraft observations in the Emmons et al. (2000) climatology (Fig. S4). In this section, we summarize the content of those figures (Fig. 9) by focusing on the regional averages for each campaign in the Emmons et al. (2000) climatology. We further concentrate our analysis on a specific range of altitude, 2-6 km in this case. This is chosen as to be representative of the free troposphere and be less directly influenced in possible emission shortcomings. Because we focus on the regional averages on the climatology, we include all simulations. In each case, we have used the monthly-averaged model results for the period 2006-2008, although the use of a longer period does not alter the conclusions (not shown).
As expected from the ozone analysis presented in Sect. 7.1, the GEOS5 simulation performs better than the other ones in its ozone distribution. Also, MERRA strongly overestimates ozone during the TOPSE (Northern Hemisphere high latitudes) campaigns, reinforcing the view of too strong mixing from the stratosphere. Nitrogen oxides (NO x ) tends to be highest in the online configuration (Fig. 9a), although all versions seem to exhibit similar biases; in particular, there is clear low NO x bias in the case of TOPSE. Similarly, PAN and nitric acid (HNO 3 ) are consistently simulated across configurations. In most cases, the lowest CO distribution is associated with the GEOS5 simulation. This is discussed in more details in Sect. 7.3.
Except for the TOPSE campaign, the non-methane hydrocarbons ( Fig. 9b) are quite well represented in the model over most of the campaigns, TOPSE being again the most biased with a strong underestimate in C 2 H 6 . There is a larger spread between configurations on methylhydroperoxide (CH 3 OOH), where the online simulation is consistently the lowest and on hydrogen peroxide (H 2 O 2 ) for which MERRA tends to be the lowest.
In order to identify the role of the wet removal parameterization (see Sect. 3.3), we perform the same analysis as above using the MERRA simulations (with the Horowitz and Neu and Prather schemes, see Table 6). Over the analysis regions (Fig. 10), we find little impact on ozone, NO x or PAN. On the other hand, CO is consistently smaller (and therefore worse) when the Neu and Prather scheme is used. Furthermore, the nitric acid and hydrogen peroxide distributions over TOPSE are more accurately represented with the Horowitz scheme. However, in most cases, little difference can be found between those simulations.
Comparison with surface carbon monoxide
Surface mixing ratio of carbon monoxide represents one of longest records of tropospheric composition; for this comparison we use all years available from the National Oceanic and Atmospheric Administration data (available at ftp://ftp.cmdl. noaa.gov/ccg/co/flask/event/, Novelli and Masarie, 2010). Furthermore, because of its strong link to the hydroxyl radical OH, the overall concentration and seasonal cycle of CO (which in turn depends on emissions of CO and its precursors) is an important indicator of the representation of the tropospheric oxidative capacity (Lawrence et al., 2001). For that purpose, we compare the model results (in this case the online and MERRA simulations, in order to have a sufficiently long record) in terms of the latitudinal distribution of the annual mean and seasonal cycle ( Fig. 11a and b, respectively). Overall, the model accurately represents the latitudinal distribution of CO (strongly driven by gradients in emissions); it also represents the Southern Hemisphere annual mean, but underestimates the high-latitude Northern Hemisphere values, similar to the multi-model results in Shindell et al. (2006). The much higher than observed value close to the Equator is related to the Bukit Koto Tabang (West Sumatra) station, which being an elevated site (865 m) makes it challenging for coarse-grid models, especially for narrow islands. The bias could also be related to emission errors, although the Guam station (at 13 • N) does not indicate a strong bias. Statistical evaluation is listed in Table 10. We find that, while the biases (annual mean and seasonal cycle) are lower in MERRA, the root-mean square difference is larger in that case. The higher correlation however indicates a better representation of the surface CO distribution in MERRA. Similarly, the seasonal cycle is quite well represented over all the latitudes, except in the Polar Northern Hemisphere. This is further confirmed by the comparison to the retrieved CO 500 hPa concentrations by the Measurements of Pollution in The Troposphere (MOPITT v4, Deeter et al., 2010, Fig. 12). We find that all versions tend to overestimate the CO over Africa in December-February. During summer, the online version tends to reproduce the African maximum better.
The estimated CO tropospheric lifetime (with respect to OH loss, ozone <100 ppbv) is approximately 1.5 months in the MERRA and GEOS5 simulations and 1.75 in the online simulation, similar to the results from Horowitz et al. (2003) and Shindell et al. (2006). We have also compared (see Supplement, Fig. S5) our tropospheric OH distribution with the Spivakovsky et al. (2000) climatology using the Lawrence et al. (2001) diagnostic approach. In that case, we find that our OH distribution is in quite good agreement with that climatology. It is however smaller (20-30%) in the tropical mid-troposphere, especially in the Southern Hemisphere; it is also larger in the Northern mid-latitudes, except in the midtroposphere. Based on this analysis, the closest OH distribution to Spivakosky's is provided by the GEOS5 simulation.
The tropospheric methane lifetime (reaction with OH only, computed as total burden divided by tropospheric loss, with the troposphere defined as the region with ozone <100 ppbv) ranges between 8.7 and 9.8 yr (Table 9), similar to Shindell et al. (2006) and Horowitz et al. (2003), but lower than MOZART-4 (10.5 yr, E2010, albeit computed slightly differently) and therefore consistent with the above analysis of OH. This is also consistent with a drier tropical troposphere in the online simulation (Fig. 4a) and higher isoprene emissions (Table 6).
Comparison with total ozone column
The availability of stratospheric chemistry in the online simulation warrants the comparison with the satellite observed total ozone column. In particular, we use here the gridded EP-TOMS and OMI (both available at http://toms.gsfc.nasa. gov). The data are zonally-and monthly-averaged before comparison with the model field (Fig. 13). The overall features (latitudinal distribution and seasonal cycle, including the Antarctic ozone hole) and interannual variability are well reproduced. In particular, even though the model has a limited vertical extent and resolution in the stratosphere, the tropical ozone column is quite well reproduced.
To further compare with observed values, we focus (Fig. 14) on comparing the long-term variability in highlatitude spring ozone (March in the Northern Hemisphere and October in the Southern Hemisphere). Because the online model is only driven by the observed sea-surface temperatures, there is no expectation that a single-year in the model simulation will be directly comparable with observations; instead, only the mean and standard deviation are relevant in this case. Figure 14 shows that, unlike the version used in the CCMVal-2 simulations (Austin et al., 2010), this updated version has a good representation of the ozone hole (mean and interannual variability), with a limited underestimate (mean bias is −5.0 DU not considering the highly unusual 2002 conditions) of the mean October Antarctic ozone hole. Similarly the mean Northern Hemisphere March ozone distribution is slightly negatively biased (mean bias is −7.5 DU). These negative biases are likely due to the cold bias over the polar regions in the online configuration (see Fig. 4a). Note however that the model is not quite able to reproduce the Northern Hemisphere dynamical interannual variability due to its limited representation of the stratosphere .
Comparison with aerosol observations
At the regional scale and over land, the modeled aerosol optical depth (Fig. 15) is generally lower than the satellite observations (from the Moderate Resolution Imaging Spectroradiometer and the Multi-angle Imaging SpectroRadiometer, all years 2001-2010), except over China. As discussed in Lamarque et al. (2011b), none of these simulations include the impact of water uptake on black carbon optical properties, lowering the optical depth associated with this compound and partially explaining the negative bias, particularly over South America where biomass burning is the major source of optical depth. It is likely that the more important tropical precipitation in the online simulation significantly contributes to the underestimation in optical depth.
Owing to the availability of a large set of surface observations (time and speciation), we focus our analysis on the United States Interagency Monitoring of Protected Visual Environments (IMPROVE, Malm et al., 2004) dataset (available for download at http://vista.cira.colostate. edu/IMPROVE/Data/data.htm). We perform a comparison of sulfate, elemental carbon, organic carbon (including secondary organic aerosols, SOA) and ammonium nitrate.
We first present correlations (Fig. 16) between long-term mean (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009) observations and model results (interpolated to the location of the observing stations); note that the IMPROVE sites are located in remote locations (such as National Parks) and are therefore representative of the rural environment, not urban. We find that, as discussed in Lamarque et al. (2011b), sulfate is quite well represented by the online configuration, however slightly positively biased. On the other hand, both elemental and organic carbon aerosols are underestimated (and with large scatter) in all configurations.
The linear fit parameters (Table 11) indicate that all configurations behave similarly, except in the case of sulfate. In that case, the slope for the online configuration is closest to the observations, while GEOS5 and MERRA overestimate the surface values by a factor of 2. It is possible that such disparity would be reduced if the SO 2 emissions were released Surface layer concentration (microg/m3) Fig. 17. Probability density function of observed (IMPROVE sites, black line) and modeled (interpolated to observation sites) aerosol surface concentration (sulfate, top row; elemental carbon, second row; organic carbon, third row; ammonium nitrate, bottom row). The simulation results shown here are for the online stratosphere-troposphere (red) and the GEOS5 simulation (green). Analysis is shown for annual (left column), winter (December-January-February, center column) and summer (June-July-August, right column).
at a specific height instead of the bottom model layer, which is much thinner in the specified dynamics set up (Fig. 3).
To further document the behavior of the bulk-aerosol scheme over the United States, we present in Fig. 17 the probability density function of the mean annual and seasonal (summer and winter) observed and modeled surface concentrations. Using this diagnostic, we find that the modeled sulfate cannot capture the lowest observed values and instead peaks at higher values and has a broader distribution; in addition, the annual mean seems to exhibit a longer tail in the distribution than the observations. Furthermore, the distribution function for elemental carbon is quite well reproduced, except for a higher tendency for small values. On the other hand it is clear that the model simulations of organic carbon cannot capture the mid-range values, especially in the summertime, most likely a representation of the lack of significant SOA production with the current scheme (Lack et al., 2004). Somewhat surprisingly, owing to the inherent difficulties in representing ammonium nitrate in a model and to the fairly simple representation of its formation in the model, we find that ammonium nitrate is quite reasonable, with a slightly smaller proportion for the high concentrations (possibly due to the model coarse resolution) and higher proportion for the low concentrations.
Discussion and conclusions
Using a variety of diagnostics and evaluation datasets (including satellite, aircraft, surface measurements and surface data), we have demonstrated the capability of CAM-chem in reasonably representing tropospheric and stratospheric chemistry.
Based on the simulations discussed here, we have found that the CO and CH 4 lifetimes are in good agreement with previously published estimates; similarly, our OH distribution is in reasonable agreement with the Spivakovsky et al. (2000) climatology. However, our CO distribution in the high Northern latitudes is underestimated when compared to surface, aircraft and satellite observations, indicating an overestimate of the CO loss by OH or underestimate of its emissions or chemical production.
Ozone in the troposphere is simulated reasonably well, with some overestimation of ozone in the upper troposphere/lower stratosphere region in high northern latitudes in comparison to ozone sondes. Even though the model top is limited to 40 km, stratospheric composition is acceptable and the polar ozone depletion is reasonably well reproduced in the online configuration. All configurations of CAM-chem suffer from a significant overestimate of summertime surface ozone over the Eastern United States and Europe. On other hand, Western United States sites are quite accurately represented.
Aerosol optical depth tends to be underestimated over most regions when compared to satellite retrievals. Additional comparison over the United States indicates an overestimate of sulfate in the case of MERRA and GEOS5, and an underestimate of elemental and organic carbon in all configurations. Analysis of the seasonal (summer/winter) and annual probability density functions indicates strong similarities between model configurations.
Using a variety of statistical measures and especially Taylor diagrams, we have found that the CAM-chem configuration with GEOS5 meteorology provides the best representation of tropospheric chemistry. This is particularly clear for the 500 hPa (best bias and correlation) and 900 hPa (best bias) regional ozone distribution. The increased vertical resolution in GEOS5 and MERRA (56 levels instead of 26 for the online meteorology) might play a role in this better performance, although this hypothesis has not been explicitly tested. Based on the ozone budget analysis (and supported by ozone analysis in the high-latitudes), we have found that the MERRA meteorology leads to a stronger ozone flux from the stratosphere. This is likely associated with the different assimilation procedures used in MERRA than in GEOS5 (Rienecke et al., 2011).
When compared against regional climatologies from field campaigns, the inclusion of the Neu and Prather (2011) parameterization for wet removal of gas-phase species shows little impact on 2-6 km ozone. On the other hand, CO is consistently smaller (by 5-10 ppb). Consistent with this, the methane lifetime is slightly shorter with the Neu and Prather scheme. Nitric acid and hydrogen peroxide are increased over the high northern latitudes (TOPSE campaign). Overall, our analysis suggests only small differences from this new parameterization.
All necessary inputs (model code and datasets) to perform the simulations described here on a wide variety of computing platforms and compilers can be found at http://www.cesm.ucar.edu/models/cesm1.0/. | 2018-05-16T17:12:30.647Z | 2012-03-27T00:00:00.000 | {
"year": 2012,
"sha1": "7678554254f76540f747a6c4e04ab7be965cac2c",
"oa_license": "CCBY",
"oa_url": "https://gmd.copernicus.org/articles/5/369/2012/gmd-5-369-2012.pdf",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "b9a45c8423904366a88ed965b94d52e9cff26b88",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": []
} |
214310145 | pes2o/s2orc | v3-fos-license | IS GUARANTEE OF ORIGIN REALLY AN EFFECTIVE ENERGY POLICY TOOL IN EUROPE? A CRITICAL APPROACH
Guarantees of origin are tradeable energy certifi cates defi ned by directives 2009/28/EC and 2018/2001/EU of the European Union. They serve the aim of informing fi nal consumers on energy sources used for their electricity supply. They are also expected to encourage new investments in renewable electricity generation. This paper investigates how the use of guarantees of origin meets these expectations. A literature review, an analysis of related regulations and an evaluation of empirical data shows that there are regulatory failures both at national and the European Union levels. Furthermore, due to a contradiction between certain rules in European Union level regulation, consumers receive unreliable information on their electricity consumption mix. Therefore, although national rules should be improved, the problem of reliability cannot be resolved until the Union level framework is modifi ed. Furthermore, the present framework does not incentivise investments in renewable energy technologies either. Accordingly, recommendations are formulated for policy makers to ensure reliable and suffi cient operation of the certifi cate system. in renewable energy generation. This how the usage of GOs meets these goals. This topic is The ‘Clean of A new to The structure of this paper can be outlined as follows. Section 2 provides an overview on the evolution of regulation and market of GOs. Section 3 summarizes the challenges regarding GOs based on a review of the literature and also highlights a contradiction between EU level regulations. The review is based on articles published in peer reviewed scholarly journals, as these papers provide grounded, methodologically proper results and conclusions. Beside these papers, however, policy documents were also used for the review, since the regulatory frameworks for GOs in European countries were constructed and have been con-tinuously improved in the last two decades and such policy documents provided important background and analysis for these rules. Databases and library serv-ices were also used for the search (www.sciencedirect.com, www.emeraldinsight. com, www.ebsco.com). Keywords for the literature search were ‘guarantee(s) of origin’, ‘renewable/green energy/electricity certificate(s)’, ‘energy/electricity consumption/consumer(s)’, ‘electricity disclosure’. The literature review is based on papers published after 2001, when GOs were first introduced in EU level regulation. Since the regulation was changed in 2009, a special focus was set on papers published after 2009. Section 4 provides results from an empirical analysis on the international flows of electricity and GOs. Section 5 includes some policy recommendations. Finally,
INTRODUCTION
As the environmental challenges of human activities became increasingly distressing in the last decades of the 20 th century, a significant consumer demand emerged for sustainable products (Pogutz -Micale 2011). Several market actors have been striving to satisfy this demand, even in the energy sector which is one of the most polluting industries (Stern 2007). Besides energy saving, utilizing renewable energy sources (RES) is one of the most important steps towards a sustainable energy sector. However, in the case of energy it is impossible to track energy flows through the grid, therefore it has been challenging to improve green products for consumers. First, market actors attempted to use unique solutions to prove that their energy supplied is of renewable origin. Later, more and more countries introduced regulations on reporting the origin of energy; most of them refer to some kind of tradeable certificate. Such tradeable certificates for tracking energy attributes exist in Europe (Gkarakis -Dagoumas 2016), in the United States (Gillenwater 2008), and also in Asian countries (Kumar -Agarwala 2013;Chuang et al. 2018). Certificates of this kind are suitable to support renewable or green energy products of suppliers. This paper focuses on guarantees of origin (GOs), a tool for tracking electricity attributes used in the member states of the European Union (EU) and some other European countries.
GOs were first introduced in the legislation of EU by directive 2001/77/EC. The first and still effective definition of GOs was set in Article 2 of directive 2009/28/EC (the Renewable Energy Directive): "an electronic document which has the sole function of providing proof to a final customer that a given share or quantity of energy was produced from renewable sources". The definition of GOs also refers to an obligation set in Article 3(9) of directive 2009/72/EC (the Electricity Market Directive) that forces electricity suppliers "to inform their consumers on the contribution of each energy source to the overall fuel mix of the supplier over the preceding year". This obligation is known as the 'disclosure' obligation and usually GOs are used to verify the content of disclosure. Beside this, another aim of using GOs is formulated in the Renewable Energy Directive: "it is appropriate to allow the emerging consumer market for electricity from renewable energy sources to contribute to the construction of new installations for energy from renewable sources" (22). Furthermore, the European Commission declares that an EU wide standardized system of GOs can "help Member States develop their renewable energy resources in the most cost effective manner possible" (EC 2008); and also that a functioning GO market "could help supplement or possibly in the longer term supersede public support for renewable energy" (EC 2016a: 152). Therefore, GOs should serve two policy aims: (i) informing final consumers; and (ii) drive new investments in renewable energy generation. This paper aims to examine how the usage of GOs meets these goals.
This topic is related to the latest legislator outcomes in the EU. The 'Clean Energy for all Europeans' package issued by the European Commission -among many others -covers the above mentioned orders of the Renewable Energy Directive and the Electricity Market Directive as well (EC 2016b;EC 2016c). A new directive on the promotion of the use of energy from renewable sources already entered into force in December 2018. Furthermore, amendments to the Electricity Market Directive and other regulations of the package were adopted by the European Parliament on the 18 th of January 2019. However, the Renewable Energy Directive and the Electricity Market Directive will remain in effect for a few years.
The structure of this paper can be outlined as follows. Section 2 provides an overview on the evolution of regulation and market of GOs. Section 3 summarizes the challenges regarding GOs based on a review of the literature and also highlights a contradiction between EU level regulations. The review is based on articles published in peer reviewed scholarly journals, as these papers provide grounded, methodologically proper results and conclusions. Beside these papers, however, policy documents were also used for the review, since the regulatory frameworks for GOs in European countries were constructed and have been continuously improved in the last two decades and such policy documents provided important background and analysis for these rules. Databases and library services were also used for the search (www.sciencedirect.com, www.emeraldinsight. com, www.ebsco.com). Keywords for the literature search were 'guarantee(s) of origin', 'renewable/green energy/electricity certificate(s)', 'energy/electricity consumption/consumer(s)', 'electricity disclosure'. The literature review is based on papers published after 2001, when GOs were first introduced in EU level regulation. Since the regulation was changed in 2009, a special focus was set on papers published after 2009. Section 4 provides results from an empirical analysis on the international flows of electricity and GOs. Section 5 includes some policy recommendations. Finally, Section 6 presents conclusions.
EVOLUTION OF GREEN ELECTRICITY MARKETS
Verifying the renewable origin of supplied electricity with tradeable certificates is a well-known practice today. In most European countries, GOs are used for this purpose. A number of papers have already described the approach and methodology of the GO system (e.g. Raadal et al. 2009;Bröckl et al. 2011;Gkarakis -Dagoumas 2016). Briefly, all GO systems throughout Europe are designed as follows: one GO refers to one MWh of electricity generated; GOs are issued in an electronic registry; market participants as account holders are able to transfer GOs in the registry separately from physical or business flows of electricity; GOs should be cancelled from the registry in order to prove the origin of electricity supplied to a consumer.
However, at the time when consumer demand for renewable electricity appeared in Europe, this tool had not been created yet. The first renewable power products appeared at the end of the 80s and during the 90s in European countries, such as the United Kingdom, Switzerland, Germany, the Netherlands, Sweden or Finland (Wüstenhagen et al. 2000;Lipp 2001;Salmela -Varho 2005;Markard -Truffer 2006;Wüstenhagen -Bilharz 2006). These early products were usually introduced through market initiatives and this activity was not regulated in the beginning. The first system based on tradeable certificates was established only in 1998 in the Netherlands (Bertoldi -Huld 2006). The EU also did not have any regulation on renewable electricity products in the 90s. Later, recognizing the emerging consumer demand for renewable electricity (EC 2000), directive 2001/77/EC introduced the concept of GO, and forced member states to ensure that a GO is issued upon request. According to directive 2001/77/EC, GOs were not necessarily tradeable certificates. Nevertheless, the approach of tradeable certificates became so widespread that the Association of Issuing Bodies (hereafter AIB) was established as a cooperation of several national bodies in 2002. AIB aimed for cooperation and harmonization among its members and for this purpose introduced the European Energy Certificate Scheme (EECS) that contains common rules, methodology and standard for GOs (AIB 2018).
In the meantime, the progress of market liberalization and the first steps towards a common European energy market began. Several papers have already discussed this evolution (e.g. Jamasb -Pollitt 2005;Joskow 2008). As an important part of liberalization, consumers became able to freely choose among offers of electricity suppliers. Directive 2003/54/EC aimed to promote competition and help consumer choices, as it obligated electricity suppliers to disclose their energy mix to consumers. Directive 2003/54/EC still did not contain reference to GOs regarding the disclosure obligation, however, some member states recognized that this obligation could be best managed with the help of GOs.
2009 was an important milestone in the evolution of the EU legislation regarding the renewable energy sector and moving toward the common European energy market as well. Both 2001/77/EC and 2003/54/EC directives were repealed and the still effective Renewable Energy Directive and Electricity Market Directive entered into force. Regulation on disclosure obligation remained unchanged in the Electricity Market Directive. Contrary, the Renewable Energy Directive brought several new and more detailed rules for GOs. This directive introduced the definition and the purposes of GOs that were presented in Section 1. From this point in time, GOs became tradeable electronic certificates by definition with the standard size of 1 MWh and containing specific information set in Article 15 of the directive. This specific information refers to the attributes of energy production (e.g. energy source, production period, information on the installation where the energy was produced). Article 15(9) of the Renewable Energy Directive also declares that member states "shall recognise guarantees of origin issued by other Member States". (It should be remarked that directives 2004/8/EC and 2012/27/EU introduced GOs for electricity generated with high efficient co-generation technology. In this article, however, high efficient co-generation GOs are disregarded.) Meanwhile, the consumer market for renewable electricity products has gone through a significant increase. GOs served as a basis for renewable electricity products. Consumers have had the opportunity to choose among several such products in a number of countries (van der Linden et al. 2004;Lise et al. 2007;Bröckl et al. 2011;Kaenzig et al. 2013;Hast et al. 2015;Mulder -Zomer 2016). Not only households, but also corporate consumers show interest for GOs as they can use them for carbon accounting purposes (Csutora -Harangozó 2017) or for marketing communication. According to data from the AIB website, the amount of electricity verified by GOs has increased a hundredfold between 2002 and 2018. In the years between 2015 and 2018, the most GOs were issued in Norway, Italy and Switzerland. The most GOs were cancelled for consumers in Germany during this period. Figures 1 and 2 provide an overview on this market. Several papers tried to identify the existing market price premiums or the willingness-to- pay for green electricity products (eg. Zoric -Hrovatin 2012;Hast et al. 2015;Mulder -Zomer 2016;Dagoumas -Koltsakis 2017). Although they found that the market premium for such tariffs is very low, it is a fact that the total value of the European GO market is estimated to be around 120 million Euro in year 2016 (Jansen 2017).
The new directive 2018/2001/EU on the promotion of the use of energy from renewable sources that should be implemented by member states by the 30 th of June 2021 contains more sophisticated regulations on GOs, however, the basic conceptual framework did not change.
Informing fi nal consumers
The Renewable Energy Directive prescribes that the sole function of GOs is to inform final consumers about the origin of electricity and also defines some fundamental criteria national frameworks should be based on. These fundamental criteria are the following: (i) specific rules on the data content of GOs; (ii) trade of GOs independently from physical flows should be enabled; (iii) international trade of GOs should be enabled. These fundamental rules will remain, according to directive 2018/2001. According to (iii), the EU is a common market for GOs where no limitations may occur which hinder international trade. Some non-EU countries, Iceland, Norway, Switzerland, also adopted EU regulations on GOs, therefore market participants of these countries are also able to act on this market.
It is obvious that the reliability and accuracy of disclosure information verified by GOs is crucial. An important issue identified by researchers is to avoid the double counting of any electricity amount (Raadal et al. 2009;Klimscheffskij et al. 2015). This problem may occur because while GOs are not issued for every MWh of electricity produced, disclosure obligation of suppliers refers to all electricity consumed. Therefore, disclosure statements are partly based on statistical energy generation mix data beside GOs. If a supplier uses GOs to verify the origin of an electricity amount, this amount should be subtracted from energy statistics. Otherwise one amount of RES electricity would be taken into account twice: once with GOs and once in the statistics. This modified energy mix not containing those electricity amounts GOs were issued for, is called the residual mix. The residual mix is an important tool: for consumers who do not choose green electricity covered by GOs, this residual mix should be presented as disclosure information. EU funded projects E-TRACK and RE-DISS formulated recommendations and protocols for a proper methodology to avoid double counting. Details are assessed by Draeck et al. (2009) andRE-DISS (2015). An important point in this regard is the fact that if a portion of GOs are used for a certain consumer, the share of RES in the residual mix will be less because the GOs used should be extracted from the energy mix. Also, if GOs are exported to another country, the domestic residual mix will become less green.
Some papers, however, drew attention to methodological or regulatory failures that cause double counting or endanger the reliability of disclosure information in other ways. Draeck et al. (2009) analysed the regulation of 29 European countries (EU27, Norway, Switzerland) and some problems were identified in all cases without exception. The RE-DISS project still found a number of discrepancies in its final report (2015) published six years later. Namely, among the 31 European countries that were subject to the report, 11 still had not implemented the disclosure regulation of the Electricity Market Directive and 20 had not implemented any rules of the Renewable Energy Directive on GOs. Other papers focus only on one or a few countries. Boardman and Palmer (2007) found a number of peculiarities in the United Kingdom that hamper consumer choice and also result in unreliable information. Eight years later Hast et al. (2015) stated that double counting of electricity still happens in the United Kingdom due to defects of legislation. This issue, together with confusing information techniques of the suppliers, results in a lack of trust that blocks the rise of demand for renewable electricity tariffs. Trust, however, is crucial for marketing renewable electricity products (Hanimann et al. 2015). Bröckl et al. (2011) focus on Nordic countries (Denmark, Finland, Iceland, Norway, Sweden) and conclude that other unregulated tools used to track electricity beside GOs also make it impossible to publish reliable information and to avoid double counting. Winther and Ericson (2013) point out that Norwegian disclosure information are incomprehensive and therefore often unreliable. Categories "unknown" or "import" also appear among the energy sources in the disclosure statement, due to inadequate methodology that cannot identify the share of all energy sources.
The issue of reliability becomes more complex considering Article 15(9) of the Renewable Energy Directive that obliges member states to recognize GOs from other member states. According to this paragraph, the international trade of GOs on the common European market should be enabled. Considering the issue of residual mixes, in such a framework of international GO trade, double counting could be avoided and reliable information could be provided only if national rules were harmonized. This goal is served by the activity of AIB, whose member organizations have already harmonized their regulations on GOs according to the common EECS rules. The E-TRACK and RE-DISS projects have recommended a wide harmonization among states concerned (Draeck et al. 2009;RE-DISS 2015). Several scholarly papers also declare that harmonized national regulations are needed to ensure reliable information for consumers and to avoid double counting and other discrepancies (Lise et al. 2007;Raadal et al. 2009;Bröckl et al. 2011;Gkarakis -Dagoumas 2016). Stakeholder organizations suggest harmonization too (BEUC 2015; Jansen 2017).
However, full harmonisation among EU countries has still not been realized. AIB has member organizations only from 21 countries (Austria, Belgium, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Iceland, Ireland, Italy, Lithuania (since June 2018), Luxembourg, Netherlands, Norway, Slovenia, Spain, Sweden, Switzerland), of which 17 are members of the EU. Thus, there are no member organizations in the AIB for 10 member states (Bulgaria, Greece, Hungary, Latvia, Malta, Poland, Portugal, Romania, Slovakia, United Kingdom). Therefore, particular rules in national legislation still differ and different methodologies result in double counting and other discrepancies (Klimscheffskij et al. 2015;Jansen 2017).
The issues discussed above (discrepancies, incomprehensive information, double counting, lack of harmonization) could be handled with the improvement of regulation and methodology. In some cases, the necessary improvements might have happened since the publication of the cited papers. However, some papers identify quite different concerns regarding the reliability of GO information. Aasen et al. (2010) focus on Norwegian corporate consumers of renewable electricity tariffs and conclude that the consumers do not really trust the GOs and dis-closure information. Their mistrust is not a result of an insufficient methodology (e.g. double counting, lack of harmonization), but the fact that physical electricity cannot be tracked. Winther and Ericson (2013: 378) analysed the attitude of Norwegian household consumers and points out that "people's conceptualisation of electricity in terms of its physical characteristics and the issue of tracing were important barriers in making them understand and believe in the information". Both papers draw attention to a peculiar speciality of Norway in this regard. Namely, Norwegian consumers are aware of the fact that the entire electricity production of Norway is renewable. Despite this well-known fact, many consumers are informed through disclosure that they consume electricity generated from fossil sources. This can occur because a large amount of Norwegian GOs are exported to other countries and therefore the local consumption is covered with a residual mix. This residual mix of the Norwegian consumption contains fossil fuels as well, coming from other countries. (The methodology used in this case is assessed by the RE-DISS (2015) project.) The same problem was also referred to by Bröckl et al. (2011) for Iceland. The case of Iceland is especially remarkable, because this land has no physical interconnectors to other countries.
Research papers identify three issues regarding information based on GOs: -inadequate national regulations and methodologies that endanger reliability and accuracy (e.g. through double counting); -lack of harmonization between countries that also results in unreliable information for consumers in the case of international GO trade; -incomplete consumer understanding and acceptance of information based on GOs due to their knowledge on physical electricity. However, none of these papers draw attention to one serious inconsistency in the EU level regulation. According to the Renewable Energy Directive, all member states should reach a predefined binding target by 2020 regarding the share of RES in the energy supply. GOs enable the trade of electricity attributes between member states, and consumers are informed about their electricity consumption based on GOs. However, Article 15(2) of the Renewable Energy Directive declares that GOs traded internationally "shall have no function in terms of a Member State's compliance with" the 2020 goals. Similarly, Regulation No. 1099/2008 of the European Parliament and of the Council on energy statistics also does not specify any opportunity to include GOs in official energy statistics. Therefore, a consumer choice for renewable electricity does not appear in any official statistic nor have any effect on official national energy mixes. Consumers get disclosure information from the supplier but officially another statement is generated. If any amount of GOs become traded internationally, the two statements will surely contain different information. On one hand, it is reasonable that the EU legislation does not allow national achievements in RES development financed by costly support schemes or investment subsidies to vanish through exporting GOs by market participants. If a member state makes a lot of effort to reach a binding target, it would not be acceptable that unregulated voluntary demand for certificates defeats the outcome. However, it is clear that this inconsistency within EU rules endangers the trustworthiness of disclosure information. Accordingly, the following questions might be raised: does disclosure provide only fictive information; are the consumers who choose RES electricity and pay a price premium aware of this issue? This is not a question of proper national level rules or methodology but it affects the fundamental approach of the EU level legislation. Further research may need to focus on this issue.
To sum up, international trade is a crucial issue for several reasons: -in the case of international trade of GOs, lack of harmonization among countries leads to disclosure problems; -the consumer awareness of the physics of electricity generation lead to scepticism regarding disclosure information; -in the case of international trade of GOs, contradictory rules in EU level regulation lead to different statements on the same phenomena.
Driving new RES development
Several papers analysed the question whether GOs and green electricity products can drive new investments in the generating capacity. Theoretically, it is obvious that green electricity products based on GOs may contribute to new RES capacities. Selling GOs provides RES generators with additional income, therefore it is an advantage to them in market competition against conventional power plants. Based on this concept, the European Commission (2016) states that an improved market of GOs "should help supplement or possibly in the longer term supersede public support for renewable energy". Raadal et al. (2012) also conclude that green electricity products might contribute to the development of RES electricity generation. Furthermore, Wüstenhagen and Bilharz (2006) add that green electricity markets can better drive cost reduction in RES generation than feed-in tariff support. Markard and Truffer (2006: 318) state that green electricity market "is more compatible with a liberalized market environment" than support schemes.
However, a number of papers conclude that the concept mentioned above does not work in practice. Markard and Truffer (2006) accomplished an analysis of five European markets (Germany, Netherlands, Sweden, Switzerland, United Kingdom). They realized that price premiums on these free markets of green electric-ity products are so low that their incentive effect on building new RES generating capacities is little. Boardman and Palmer (2007) focused only on the market of the United Kingdom and concluded that disclosure does not drive any change on the production side. According to the latest results, no such market development has occurred that could trigger any substantial change in this regard. Raadal et al. (2012) compared the Swedish and Norwegian green electricity markets and disclosure system with the tradable green certificate (TGC) support scheme. While TGC could bring a significant change in generating capacity, GOs and disclosure have had no effect on the production side. Hast et al. (2015) made an assessment on the green electricity markets of Finland, Germany, and the United Kingdom. According to their conclusions, green electricity products based on GOs cannot bring new RES capacities, because the large amount of GOs from already existing RES power plants creates oversupply and reduces prices. Usually large Norwegian, Swiss and Austrian hydropower generators, that were commissioned decades ago and their costs had been returned beforehand, are responsible for the oversupply. The price premium calculated by the authors is therefore rather low, around 0-5%. Similarly, low premium was calculated by Mulder and Zomer (2016), who focused on the Dutch GO market. They also state that the effect of GOs on new RES development is weak and the reason for this is the high share of imported Norwegian GOs with low marginal costs. GOs are used rather just as a marketing instrument by suppliers. Dagoumas and Koltsakis (2017) conducted an econometric analysis on Greece and confirmed that prices of GOs are by far not enough to provide incentive signals for investors. A panel data analysis on a sample of 30 European countries also does not find a significant relationship between the consumer market for green electricity products supported by GOs and RES development (Hamburger -Harangozó 2018).
So, according to the consensual judgement of the literature, green electricity products based on GOs could not bring any incentives for new RES development. The reasons behind this defect are oversupply on the markets and therefore low prices. Oversupply emerges due to the large amount of import from old hydro power plants. This issue might be connected to the high share of Norwegian GOs on the market mentioned in the previous subsection. Another reason for the failure in driving new RES capacities might be that support schemes provide a more reliable incentive than voluntary GO demand (Markard -Truffer 2006;Raadal et al. 2012). Either way, if GOs and disclosure cannot excite a mechanism that brings new RES capacities, the legitimacy of the legal framework will be weakened (Aasen et al. 2010). Again, enabling limitless international trade is a barrier for promoting RES generation.
CONTRAST BETWEEN FLOWS OF ELECTRICITY AND ATTRIBUTES
According to the literature reviewed in the previous section, the usage of GOs faces heavy challenges with regards to informing final consumers and driving new RES development. An important issue in both fields is the international trade of GOs. Regarding consumer information, lack of harmonization causes discrepancies in case of international trade but also consumer's approach to electricity in its physical form results in mistrust if GOs are traded between countries. Additionally, in case of international trade of GOs, a contradiction between EU level rules on GOs and energy statistics causes that the disclosure and statistical statements produce different information on electricity consumption. Regarding RES development, a large share of GOs from a few countries with good natural endowments bring prices down and hinder price signals towards investors. Based on these conclusions from the previous section, an investigation with the scope of international trade of GOs follows.
An empirical analysis was conducted on physical electricity flows and the international trade of GOs among AIB member countries. Although electricity market liberalization has been fulfilled in all countries and international trade of electricity became possible (see Section 2), this still does not mean that electricity can be delivered from one country to another without any limit. The business flows of electricity are limited to the available physical capacities of the interconnector networks. Traders should participate on auctions for these capacities (Van den Bergh 2015). These capacities allocated to market participants might be even curtailed by the network operator in the case of emergency situations, according to Regulation 714/2009 of the European Parliament and of the Council. Contrary, the international trade of GOs should not be limited by any physical or legal restrictions according to the Renewable Energy Directive. Consequently, international trade of GOs may differ from physical or business electricity flows.
Since several papers pointed out that consumers realize the difference between physical electricity flows and GO trade and this results in mistrust, an analysis was conducted with empirical data to investigate to what extent the international trade of GOs differ from the physical flows of electricity. Based on empirical data, the aim of this analysis is to investigate to what extent the mistrust generated by the difference between physical and certificate flows is grounded. It also aims to examine whether a remarkable gap between the two flows can be explored. This analysis makes a significant contribution to the papers cited in the previous section, because it uses empirical data for a number of countries. The data of the European Network of Transmission System Operators (hereafter ENTSO-E) on physical flows of electricity (ENTSO-E 2019) and the data of AIB on international trade of GOs (AIB 2019) were used for this purpose. The sample contained 19 AIB member countries. 1 Monthly data were taken for the analysis from the years 2016 to 2018, leading to a total of 684 observations.
For each country and for each month the amount of physical electricity export flows in the direction to other AIB member countries were summarized. Similarly, the amount of exported GOs to other AIB countries was ascertained for each observation. This was followed by the assessment of whether the amount of exported GOs exceeds the amount of physical electricity export flow in any of the cases. According to the analysis, in 277 cases (40.50% of the observations), the amount of exported electricity attributes was higher than the physical flow of electricity. This is already a remarkable share, however, taking the electricity values instead of the number of observations, a much higher share appears. From the total 1 212 193 262 MWh GO exports, 875 957 448 MWh were realized over the values of physical export flows. Accordingly, 72.26% of exported GOs were traded over electricity flows. The same calculation was made for import data. Import of GOs exceeded physical import flows in 209 cases, 30.56% of all observations. From a total 1 164 561 208 MWh of GO imports, 730 389 679 MWh were realized over the physical import flows. This means that 62.72% percent of imported GOs was over the physical import flows of electricity. (The country of destination in the case of exported GOs and the origin in the case of imported GOs was not known, therefore the database does not include such data. However, it is sure that if these data were available, the share of GOs traded over the electricity flows would be even higher.) Based on these high shares of surplus, the firm statement can be formulated that the flows of international GO trade are highly different from physical electricity flows. Figure 3 indicates how many GOs were exported to other AIB countries and how much the share of GO surplus is over physical flows. Figure 4 indicates the same information for import.
Looking at Figures 3 and 4, the following country-specific information can be drawn: -In the case of export 11, in the case of import 8 countries have more than 30% surplus of GO trade over physical flows. -The case of Norway is remarkable, since the amounts are the largest in export and import as well and the share of GO surplus is very high for both directions too. 90.08% of exported and 94.86% of imported GOs were over 1 Two AIB member countries were dropped from the sample. Belgium was dropped because the regional organizations of Belgium are members of the AIB independently and data on GOs are accessible separately for the regions. ENTSO-E data are given for countries, therefore, GO flows from and to Belgian regions cannot be compared to the physical electricity flows. Lithuania was also dropped, since it was only a member of AIB for six months of the period under analysis.
the real physical electricity flows. Taking the absolute values, 50.03% of all exported GOs to 19 countries were from Norway. This information supports the conclusions (Hast et al. 2015;Mulder -Zomer, 2016) saying that Norwegian GOs overflow the markets. Total GO import from other AIB countries (MWh) Share of GO surplus in import from other AIB countries (%) flows are possible from or to these countries. However, Iceland is a significant exporter of GOs among these countries. -In the case of three countries (Cyprus, Czech Republic, Slovenia), GO flows never exceeded the physical flows. In the case of three countries (Croatia, Germany , Switzerland), the GO export flows never exceeded the physical export. Regarding the physical flows and the international trade of GOs, the following conclusions can be drawn: -It is clear that the international flow of GOs is the multiple of international physical flows. -Considering that the interconnector capacities are highly exploited (Spiridonova 2016), the quantities of traded electricity attributes are much higher than what could ever be realized in physical flows with the present interconnector capacities. -In several cases, GO trade even occurs from or to countries lacking any interconnector capacities.
GOs are also traded outside of AIB countries. According to data from the website of AIB, in the years under analysis, a number of GOs were exported from AIB member countries to several other European countries, but also to such far away lands as Australia, Brazil, Chile, China, India, Japan, Qatar, Thailand, Saudi Arabia, the United Arab Emirates, and the United States. Market actors who trade GOs are obviously ready to utilize such opportunities to penetrate these markets. If GOs are exported to a third country, the domestic residual mix will contain a lower RES share. But in reality, it might be incomprehensible to imagine how a consumer in the Far East paying for RES GOs could affect the residual mix of a European consumer. It is doubtful that this practice fits with the legislative aims of GOs set in the Renewable Energy Directive. Consumers get information on their energy mix that is biased by consumer demand in distant countries. Although the share of GOs exported outside AIB member countries is currently low, this practice includes a fundamental distortion. It is also questionable if European consumers are aware of this practice.
Based on the regulations, the literature review, and the empirical data, several defects were identified regarding GOs that set back the realization of the aims that were formulated in the Renewable Energy Directive. These defects result in unreliable disclosure information and hinder incentivising investments. Until EU level rules remain the same, the improvement of national level regulations cannot help solve the fundamental problems. The next section summarizes recommendations of other papers and aims to articulate certain personal recommendations regarding the framework of GOs.
DISCUSSION AND POLICY RECOMMENDATION
The international trade of GOs enabled by EU level regulations fits the approach of a common European market. The rules on international GO trade, however, do not refer to the costs and barriers of transferring energy. Furthermore, the disclosure information based on GOs is also contradictory to the official statistics. Based on these rules, a virtual trade of GOs emerged in the past years which is unable to be a basis for informing consumers, although that would be the aim of using GOs. Several papers identified that even those consumers who pay for RES electricity products do not understand or do not believe the disclosure information (Aasen et al. 2010;Winther -Ericson, 2013;Klimscheffskij et al. 2015;Jansen 2017). These rules may drive away consumers from purchasing RES electricity products. Although several papers and organizations (Bröckl et al. 2011;BEUC 2016;Jansen et al. 2016) suggest merely the improvement of communication and supporting better consumer understanding of the system, this cannot be a solution. A practice that is so distant from reality should not be communicated as real, instead, the rules should be enclosed to physical reality. Some of the papers referred to in Section 3 identify one or more issues regarding the disclosure information based on GOs, and make suggestions on regulatory framework. Klimscheffskij et al. (2015) suggest improvements on methodologies. Other papers (Lise et al. 2007;Raadal et al. 2009;Bröckl et al. 2011;Gkarakis -Dagoumas 2016) recommend stronger harmonization among states. These recommendations aim at national level regulations to provide more precise disclosure information while respecting the EU level framework. However, this EU level framework caused the flows of GOs to become rolled away from physical reality, as shown in Section 4. Even if all national methodologies and rules were precise and a full harmonization was fulfilled among European countries, the difference between physical reality and the trade of GOs would remain unchanged. And if this remains unchanged, nothing prevents the framework against consumer mistrust.
The usage of GOs also does not succeed in incentivizing new investments. The cause behind this failure is not mistrust, but low prices on the market due to a large amount of GOs from Norway and other countries with good natural endowments for RES electricity generation (Hast et al. 2015;Mulder -Zomer 2016). Data introduced in this paper verifies this opinion. Several studies recognized that GOs cannot bring more investments; therefore a different type of amendment might be needed. Hast et el. (2015Hast et el. ( : 1381 recognized that GOs cannot incentivize investments and in order to handle this problem suggested that "more attention should be paid on ensuring that price premiums paid by voluntary consumers are efficiently used to finance environmental benefits that would not otherwise occur". Mulder and Zomer (2016) have more concrete suggestions. First, they recommend to allow trade of GOs only in EU countries. This way, the Norwegian GOs would be locked out of the market. This suggestion tries to solve an actual problem -namely the dump of Norwegian GOs -, however, does not address the fundamental issues of the framework. Second, they suggest that only new power plants could get GOs for their electricity generation. This suggestion aims to boost the investment motivating effect of GOs. But GO is only a tool of informing consumers and if only new power plants got GOs, it would not fit the definition of GOs. This might be a sufficient solution but then the present concept of GOs would not be valid anymore and policy makers would need to rethink it.
Based on the findings of the previous section, this paper also formulates recommendations. Of course, harmonization of national regulations, as others also have suggested, is necessary to handle reliability issues in the case of international trade of GOs. But it would not be enough, because the failures of both legislative aims result from the limitless international trade of GOs. Consumers may not believe or do not understand disclosure statements based on GOs, especially if they are aware of the physical characteristics of electricity and of the contradictions in the EU level regulation. Also, limitless international trade allows Norwegian GOs to dump the market and push down prices. These findings challenge the current EU level framework on GOs.
Physical barriers of the transportation of electricity should be taken into account. This is the case in electricity trade as well where market participants have to compete for interconnector net capacities. Electricity trade is separated from the physical flows, but physical flows and infrastructure remain barriers in the international trade of electricity and market participants have to take part in auctions for interconnector capacities. Similarly, a methodology should be elaborated that ensures that GO trade also considers physical reality. The basis of such a methodology could be the metered electricity flows on the interconnectors in a given time period, e.g. a month. Market players willing to transfer GOs from one country to another should acquire the empowerment for the transfer through a tendering process. The maximum limit for the amount of GOs enabled to transfer through an interconnector in a time period should fit the amount of realized electricity flow on that interconnector during the given period. This restriction should be applied for borders within Europe or the AIB member states and also in the direction of third countries. Export to distant countries would not be possible. The elaboration of the detailed rules might be the subject of another research. This amendment would only allow a physically realistic amount of GOs to be traded between countries and therefore consumers would get more reliable disclosure information. The GO dump from Norway would also be cut down to a lower level; therefore, market prices might be higher and might be able to incentivize investors. Of course, the recommended framework would be in conflict with the interest of traders, but could help to realize policy goals.
Another issue is how the problem of contradicting disclosure information and official statistics can be resolved. Even if the allowed amount of international GO trade would be adjusted to physical flows of electricity, a difference would remain between official statistics and GO information. Taking consumer understanding into consideration, it is extremely important to avoid such inconsistencies. As Wiser (1998) and Markard and Truffer (2006) also point out, education is a crucial point in the functioning of green energy markets. Markard and Truffer (2006) add that one important outcome of such markets is an eco-learning process of consumers and market participants. Nevertheless, it is clear that contradictory information hinders learning. Policy makers and further research should focus on this issue and ensure that inconsistency between the two kinds of information could be avoided.
CONCLUSION
The lack of harmonization, inaccurate national regulations, and also the present EU level regulatory framework on GOs and disclosure lead to consumers not receiving reliable information on the energy mix of their electricity consumption. Enabling the limitless trade of GOs and the contradiction between EU regulations on GOs and statistics result in unreliable energy mix disclosure information for consumers. We suggest that the possible amount of international GO trade should be determined and adjusted to physical electricity flows. This would help to provide reliable information for consumers. However, a major concern is that the new directive 2018/2001/EU on the promotion of the use of energy from renewable sources did not change the framework in this regard.
This paper identifies some problems that might be in the scope of researchers and policy makers in the future. They may determine the details of the framework and proper methodology of tendering for international trade of GOs that was proposed in the previous section. Further research should focus on the contradiction between the rules on GOs and official statistics. | 2020-01-02T21:45:11.632Z | 2019-12-01T00:00:00.000 | {
"year": 2019,
"sha1": "73c72cdd74b4cb1e777e98808b3a038721c74742",
"oa_license": "CCBYNC",
"oa_url": "https://akjournals.com/downloadpdf/journals/204/41/4/article-p487.pdf",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "71f8fa24001fd051a755d30e348060814793b264",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Business"
]
} |
141166161 | pes2o/s2orc | v3-fos-license | Experimental Study on the Small-strain Stiffness and the Damping Properties of Strong Structural Clay
The small strain parameters of soil are important indexes for correct evaluation of the soil deformation. Based on the Stokoe resonant column system, the small strain dynamic characteristics of Zhanjiang strong structured clay under different consolidation stresses are investigated, including the dynamic shear modulus and damping ratio. Results show that under the same stress level, the maximum dynamic shear modulus of undisturbed soil is slightly larger than that of the remolded, and both of them decrease with the increase of shear strain. And the higher stress level is, the higher the maximum dynamic shear modulus would be, and the more significant the attenuation would be. As the shear strain increases to a certain extent, the damping ratio of the undisturbed soil increases sharply, but that of the remolded increases slowly. The maximum dynamic shear modulus of undisturbed soil shows a singular feature that it increases at the beginning and then decreases with the increase of effective confining pressure; and the inflection point corresponding to the effective confining pressure is larger than the structural yield stress of soil. The maximum dynamic shear modulus of the undisturbed soil is not only affected by the positive effect of the consolidation pressure and void ratio, but also affected by the negative effect of the structural damage induced by the stress level.
Introduction
The small-strain parameters of soil are important metrics for evaluating their deformations. Burland noted that soil strains surrounding underground structures (e.g., foundation pits, tunnels and bases) range from 0.01% to 0.1% in most areas except for a few that contain plastic zones [1]. Atkinson et al. classified the soil strain into three types, namely, the micro-strain (≤0.001%), the small strain (0.001%-0.1%) and the large strain (>0.1%) [2]. Extensive engineering field measurements demonstrate that a considerable number of soils surrounding geotechnical engineering structures are in the small-strain condition under working loads. In recent years, extensive studies have been conducted to examine the small-strain behaviour of soil through high-precision laboratory and in situ tests, and the stiffness of soil has been found to be highly nonlinear under the condition of the small strain [3][4][5][6][7].
Generally, study on the small-strain properties of soil is focused on two areas: the initial modulus or maximum dynamic shear modulus G max at the small strain and the variation trend of soil stiffness within the small-strain range. Mandy factors would affect the G max of soil, such as the strain amplitude, stress level and void ratio. The Hardin equation can be used to calculate the G max of soil at the small strain and reflect the effect of the over-consolidation ratio, void ratio and effective stress on G max [8]. In engineering practice, G max of a soil at the small strain is generally estimated based on its physical parameters and stress state using an empirical equation. The decrease of stiffness with increasing 4th International Conference on Energy Equipment Science and Engineering IOP Conf. Series: Earth and Environmental Science 242 (2019) 062082 IOP Publishing doi: 10.1088/1755-1315/242/6/062082 2 strain is described based on the dynamic shear moduli G corresponding to various shear strains γ. A critical shear strain γ 0.7 is often used to quantify the extent of decrease in the small-strain stiffness of soil. The Hardin-Drnevich equation can depict the whole curve of the decrease of G with γ [9].
Structured clays are widely distributed, and their strength and deformation are both restricted by their strong structure properties. The structure of a soil is closely related to its stress state [10][11]. To date, effects of the structure on G at various stress levels have been rarely investigated. Park statistically analyzed the existing results and summarized the initial G-confining pressure σ relationship for various soils as a power function relationship [12]. For most sandy, silty and soft soils, G increases with increasing consolidation pressure (i.e., σ). However, the mechanical properties of structured soils differ from those of normal soils, and their mechanical behaviours differ significantly when they are subjected to a lower or higher pressure than the structural yield stress. Whether the variation trend of G with σ remains unchanged after the point of structure yielding stress deserves further in-depth investigation. In addition, structural damages in the structured soil evolve with the increase of average stress. In characterizing the G max of a soil at the small strain, the Hardin equation only takes into consideration the increase of G max caused by the increase in effective stress but neglects the effect of the structural damage. Hence, the applicability of the Hardin equation to calculate G max for structured soils at the small strain remains unclear.
In view of the above mentioned limitations, in this study, resonant column tests are performed on undisturbed and remoulded specimens of strong structured clay at various σ levels to investigate the variation law of its dynamic parameters at the small strain. Through comparison, the G and damping ratio D of the structured clay at various σ levels are analyzed. In addition, the correlation between the shear strength of the structured clay and variation of its stiffness is discussed.
Test soil sample
A soil sample was collected from a sedimentation layer at a depth ranging from 7.0 m to 10.0 m in a coastal zone of Zhanjiang, in China. The soil layer contained horizontal thin beddings with intermediate visible sand grains. Preliminary mechanical tests revealed that the undisturbed soil sample had an unconfined compressive strength of 150 kPa, a structural yield strength as high as 400-600 kPa and a sensitivity of 5-7, which belongs to a typical high-sensitivity and strong-structured soil.
Test equipment
The resonant column test can measure the dynamic deformation properties of soil within a strain range of 10 -6 -10 -3 and is considered the most reliable laboratory method for measuring the soil small-strain parameters. In this study, a fixed-free Stokoe resonant column apparatus (RCA) (GDS Instruments, UK) was used. The GDS RCA consists of a drive system, a monitoring system, a drainage system and a pressure chamber, as shown in Figure 1. The GDS RCA applies vibration excitation to the top of a cylindrical solid soil specimen, generates a torsional or longitudinal excitation via the electromagnetic drive system, determines the transmitted wave velocity and the damping of the soil specimen by measuring its motion at its free end, and calculates G based on the measured shear velocity and specimen density. Figure 1. The GDS RCA test system. Figure 2 shows the drive system of the Stokoe RCA. The drive system consists of a four-arm rotor and a support column. There is a permanent magnet at the bottom of each arm of the four-arm rotor. The support column is used to fix four pairs of coils. A sinusoidal voltage is first applied to the electromagnetic coils, which then generates a driving force on the driving head and thereby applies a torsional force on the specimen. The frequency and amplitude of the excitation applied to the specimen can be varied by adjusting the frequency and amplitude of the current applied to the electromagnetic coils. When the driving frequency equals to the natural (resonant) frequency of the specimen, the specimen reaches its maximum vibration amplitude, which can be monitored by the acceleration sensor. The frequency corresponding to the maximum vibration amplitude of the specimen is the natural (resonant) frequency, based on which the G of the specimen can be calculated. where I is the rotational inertia of the soil specimen, and I 0 is the rotational inertia of the drive system of the resonant column. However, due to the complex geometric shape of the drive system, an accurate value of I 0 cannot be mathematically determined. Generally, I 0 is determined through an empirical value table.
The shear wave velocity V s can be calculated using the following equation: G can be determined based on V s : where ρ is the mass density of the soil specimen, G is the dynamic shear modulus of the soil specimen, f is the resonant frequency, H is the height of the soil specimen, and β is the eigenvalue of the torsional vibration frequency equation. The viscous damping ratio D of soil specimen measured by the resonant column is obtained based on a free vibration damping curve, which is measured by the acceleration sensor installed on the drive plate of the resonant column. After the torsional (flexural) resonance test, the resonant frequency of the specimen is obtained. Subsequently, a damping test is conducted, and the resonant frequency automatically provides the damping test parameter. A sinusoidal wave is applied to the soil specimen; then, the excitation is terminated, and the free vibration of the soil specimen is measured. The decrement δ in the logarithmic expression of the damping curve can be calculated based on the continuous cyclic vibration amplitude ratio: δ is calculated by plotting a curve depicting the relationship between the peak vibration amplitude and number of cycles. Theoretically, this curve is expected to be a straight line with a slope of δ.
The viscous damping ratio D of the soil specimen is then calculated based on δ:
Test method
Undisturbed and remoulded specimens of Zhanjiang strong-structure clay (ZSSC) were each examined by the resonant column test. Each specimen was subjected to a multi-level consolidation (isotropic consolidation) test. After each specimen was drained and consolidated at a certain pressure level, the amount of consolidation settlement and amount of drainage were measured. Subsequently, the vibration stress was increased incrementally, and the vibration frequency of the specimen and the corresponding shear stress were measured. The power was then cut off, the vibration damping curve of the specimen was measured, and the volumetric change and void ratio of the specimen were calculated. In addition, G for the specimen was calculated based on equation (3). The specimen was then subjected to the next level of consolidation until it had been tested all the σ levels.
A high-quality undisturbed ZSSC sample was collected using a fixed piston thin-walled sampler. A remoulded specimen was prepared by kneading. The structure of a piece of undisturbed ZSSC sample was destructively converted into a sludge form, and then was remoulded and shaped. Each specimen had a diameter of 50 mm and a height of 100 mm. Each specimen was saturated using two techniques successively (namely, vacuum saturation and back-pressure saturation) to allow its degree of saturation to reach above 98%, after which the specimen was tested. The following effective σ levels were used in the test: 100, 200, 400, 600 and 800 kPa. Figure 3 shows the G-γ curves of ZSSC at various σ levels. G gradually decreases as γ increases. When γ is relatively small, G decreases slowly as γ increases. However, after γ increases to a certain . G-γ curves of ZSSC. extent, G starts to decrease rapidly as γ increases. A comparison of the test curves of the undisturbed and remoulded specimens at various σ levels shows that under the same excitation voltage, a higher σ corresponds to a smaller measured minimum γ, and more significant decreases in G with the increase of γ. In addition, an inconsistency of the G variation can also be observed between the undisturbed and remoulded specimens with changing σ. For the remoulded specimen, G-γ curve gradually goes up as σ increases, and it is varied monotonically. For the undisturbed specimen, G-γ curve does not change monotonically but instead first goes up and then down with the increase of σ. The notable phenomenon can be observed in Figure 4 Here, the Hardin-Drnevich [9] hyperbolic model is used to describe the relationship between the measured dynamic shear stress τ and γ amplitude:
The variation of dynamic parameters at the small strain
where a and b are test parameters. Furthermore, G for the soil can be expressed as follows: Based on equation (6), a straight-line relationship exists between 1/G and γ. By fitting the curves, as shown in Figure 4, G max values at various σ levels can be obtained (i.e., G max = 1/a).
The test data of G for the undisturbed specimen were normalized relative to G max , as shown in Figure 5(a). In addition, the normalized results for the G/G max -γ curves of soils obtained from other [13][14][15]) are plotted in Figure 5(a) for comparison to elucidate the differences in dynamic properties between ZSSC and other clays. Similarly, the relation curves between the damping ratio D and the shear strain γ for the undisturbed specimen were conducted and shown in Figure 5(b). In Figure 5 i.e., all the values are significantly higher than those from the earthquake resistant code. Thus, it is too conservative an approach to select dynamic parameters based on the specification.
In Figure 5(b), the test values of D of ZSSC at a strain smaller than 10 -4 are mainly concentrated near the values of Chen et al. and Yuan et al., and start to become dispersed when the strain exceeds 10 -4 . The test values of D are lower than those from the earthquake resistant code. And the results for D are more dispersed than those for G/G max , which is consistent with the current understanding of damping. Figure 6 shows the D-γ curves of the undisturbed and remoulded specimens at various σ levels. As demonstrated in Figure 6(a), for the undisturbed specimen, D increases as γ increases at each σ level and increases sharply after γ increases to a certain extent. Compared with Figure 3(a), it is found that the stage of G decreasing rapidly corresponds to the stage of D increasing rapidly. By contrast, the Dγ curves of the remoulded specimen increase relatively smooth with the increase of γ (Figure 6(b)).
(a) D-γ curve of the undisturbed specimen (b) D-γ curve of the remoulded specimen
Description of G and D
Based on the Mohr-Coulomb criterion and the Hardin-Drnevich hyperbolic nonlinear stress-strain model, the following G equation was proposed [9]: where γ ref is the reference shear strain. For the equation (7), all of parameters have their definite physical meanings. However, it is not fit for different soil types due to the relatively small number of fitting parameters. Several researchers have revised the Hardin-Drnevich model by using a three-parameter Davidenkov model to correct the f(γ) [16]: where γ 0 , A and B are fitting parameters related to the soil type (γ 0 is no longer a reference shear strain with a well-defined physical meaning). The Davidenkov stress-strain relationship model can be expressed as follows: Based on G, Hardin & Drnevich proposed the following empirical equation for calculating D: where D max is the maximum damping ratio (D = D max when G = 0). In engineering practice, because the fit of D to the test results is not ideal, another empirical equation is often used to describe the D-γ relationship [16]: where D min is the minimum damping ratio, which is the basic damping ratio of the soil and is related to the properties and consolidation state of the soil, and n is a fitting parameter related to soil type.
(a) The undisturbed specimen (b) The remoulded specimen Figure 7 shows the distribution of the test values of the undisturbed and remoulded specimens at each σ level in addition to the model-fitted curves. The test values of the undisturbed and remoulded specimens are relatively insignificantly dispersed and distributed in a relatively narrow zone. The Hartin-Davidenkov model and the empirical D-γ equation (equation (12)) can satisfactorily fit the variation of G/G max and D with the change of γ. Figure 8 shows the G max -σ curves of ZSSC obtained from the tests. G max of the remoulded specimen increases with the increasing of σ, which shows a feature of monotonicity. Through comparison, a peculiar phenomenon is observed -G max of the undisturbed specimen first increases and then decreases as σ increases. The value of σ corresponding to the inflection point is greater than the structural yield stress of soil, suggesting the occurrence of hysteresis. A comparison of the undisturbed and remoulded specimens under the same consolidation conditions shows that the G of the undisturbed specimen is slightly higher than that of the remoulded specimen at relatively low σ levels, but becomes lower than that of the remoulded specimen as σ increases. As σ reaches 600 kPa, the G max of the undisturbed specimen starts to decrease, its skeleton gradually becomes fractured due to compression, and its stiffness gradually decreases. At the post-structural yield stage, σ increases until the structure of specimen collapses, at which time the specimen completely loses its structure. Figure 8. G max -σ curves of ZSSC. Figure 9. S u -σ curves of ZSSC. Figure 9 shows the undrained shear strength S u -σ curves of the undisturbed and remoulded specimens under the static load. The S u -σ curve of the undisturbed specimen has a notable inflection point, whereas the S u -σ curve of the remoulded specimen is a straight line. For the undisturbed specimen, when σ is lower than the structural yield stress of the soil, its structural strength plays the dominant role, and the effects of consolidation and compaction are insignificant; consequently, its deformation is insignificant. When σ exceeds the structural yield stress, its structural strength gradually decreases to zero, and the cementation between particles decreases; consequently, the structure is significantly compacted, which is reflected by the inflection point of the S u -σ curve.
Test Correlation between the stiffness and strength of ZSSC
The resonant column and triaxial undrained consolidation test results both demonstrate that under low effective confining pressures, the G max and peak Su of the undisturbed specimen are both higher than those of the remoulded specimen; however, after the pressure exceeds the structural yield stress of the soil, due to consolidation, the remoulded specimen has a relatively high G and τ, which is reflected by its relatively high stiffness and strength. Figure 10. Variation in the G max /S u ratio of ZSSC with σ. Figure 10 shows the variation in the G max /S u ratio of ZSSC with the change of σ. As demonstrated in Figure 10, the G max /S u ratio displays notable multi-stage characteristics as σ increases. The clay sustains relatively small structural damage at low σ levels. As σ increases, both the strength and stiffness of the clay increase. The G max /S u ratio of the undisturbed specimen also increases as σ increases. An inflection point appears on the G max /S u -σ curve of the undisturbed specimen at σ value of 400 kPa. After σ reaches 400 kPa, the G of the undisturbed specimen still increases as σ increases, but at a much slower rate than S u , resulting in a decrease in the G max /S u ratio. As σ further increases, the undisturbed specimen sustains progressive structural damage, and its strength gradually approaches that of the remoulded specimen. In addition, as σ increases, the S u of the undisturbed specimen starts to increase at an increasing rate, whereas the G max /S u ratio decreases significantly. This result suggests that for the undisturbed strong -structured soil, the degree of attenuation of its stiffness caused by the structural damage is greater than that of its shear strength.
Discussions
A peculiar phenomenon is observed from the G max -σ curve of the undisturbed ZSSC specimen-G max first increases and then decreases with the increasing σ. The σ corresponding to inflection point is higher than the structural yield stress of soil. Only the stress corresponding to the inflection point on G max /S u -σ curve or G max -σ curve normalized by the void ratio function is almost equal to the structural yield stress of soil. If the initial consolidation pressure falls within the range between the structural yield stress and the pressure corresponding to the inflection point of the G max -σ curve, G increases to a smaller extent than S u as σ increases from the beginning.
Compressive hardening of soil is one of its important features during compression, i.e, soil modulus increases with its increasing density. For the structured clay, this property would be significant difference. Based on the influence mechanism of compressive hardening and structure on the soil strength and stiffness, it is concluded that G max of the undisturbed ZSSC specimen is not only positively affected by σ and the void ratio function, but also negatively affected by the structural damage induced by σ. When σ is lower than the structural yield stress of soil, the positive effects play the dominant role; when σ is higher than the structural yield stress, the positive effects are less significant than the negative weakening effects of the structural damage of the undisturbed specimen on G. In addition, the structural damage of the undisturbed specimen caused by increasing σ is progressive and irrecoverable, and affects G and the shear strength Su to varying degrees at different stages.
Conclusions
This study investigated the small-strain parameters-G and D of undisturbed and remoulded ZSSC specimens with the resonant column tests. The conclusions derived from this study are summarized as follows: 10 1) At the same σ level, the G max of the undisturbed specimen was slightly higher than that of the remoulded specimen. The G of the undisturbed and remoulded specimens both decreased as γ increased. A higher σ corresponded to a higher G max , and a more significant decrease in G max with the increase of γ.
2) The D of both the undisturbed and remoulded specimens increased as γ increased. However, after γ increased to a certain extent, the D of the undisturbed specimen started to increase sharply, whereas the D of the remoulded specimen increased smoothly.
3) A peculiar phenomenon was observed for the G max of the undisturbed specimen-the value first increased and then decreased as σ increased. In addition, σ corresponding to the inflection point of the G max -σ curve of the undisturbed specimen was higher than the structural yield stress. The G max of the undisturbed specimen was positively affected by σ and the void ratio, and negatively affected by σinduced structural damage.
4) The structural damage of the clay caused by σ resulted in a more significant decrease in G max than in S u . | 2019-05-01T13:04:16.986Z | 2019-03-30T00:00:00.000 | {
"year": 2019,
"sha1": "2a64aea58b7809ded819d537f6035564b1fbaf58",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1755-1315/242/6/062082",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "c00a3342ef891328b071fbd5074eb6710a28d7f1",
"s2fieldsofstudy": [
"Engineering"
],
"extfieldsofstudy": [
"Materials Science"
]
} |
253526914 | pes2o/s2orc | v3-fos-license | On a path to becoming more self-regulated: Reflective journals’ impact on Chinese English as a foreign language students’ self-regulated writing strategy use
A number of studies have confirmed the positive effect of writing reflective journals on L2 learning. However, the relationship between writing reflective journals and the use of self-regulated writing strategies remains unclear. To redress this knowledge gap, we assigned 38 Chinese English as a foreign language (EFL) students three journal-writing tasks in which they reflected on their writing processes and explored (1) the types of self-regulated writing strategies and changes to those strategies that the students’ reflective journals documented; (2) how students with varied writing-proficiency levels differed in their use of self-regulated writing strategies; and (3) the effects of reflective-journal writing on students’ self-perceived use of self-regulated writing strategies in particular, and on their L2 writing in general. Among the 19 kinds of strategies identified in 112 reflective-journal entries, only five (i.e., handling feedback, resource management, text processing, emotion regulation, and idea planning) were demonstrated relatively frequently. The use of seven strategies (i.e., self-monitoring and evaluation, idea planning, perspective change, emotional control, effort regulation, peer learning, and resource management) exhibited significant increases over time, especially during the second-half of the focal semester. In addition, our journal data highlighted individual variation in proficiency levels: with high-proficiency students significantly more likely than others to apply idea planning, feedback handling, and resource management and low-proficiency ones significantly more likely than others to engage in goal-setting. The qualitative results suggest that the practice of journaling raised students’ awareness and may have contributed to an increase in their use of self-regulated writing strategies. In particular, the findings reveal how students internalized and reconstructed the various SRL processes taking place via writing reflective journals. For L2 educators using or considering using reflective journals, these findings contain fresh insights that could help them not only to increase their students’ SRL levels, but also to provide more individualized SRL guidance.
Introduction
Writing is arguably the most essential, yet the most challenging, second-language (L2) learning skill (Zhang and Guo, 2012). This is not only due to it requiring sufficient L2 language knowledge during the writing process (Manchón, 2011), but also because it is a highly structured process intertwined with individual and environmental factors such as motivation, working memory, cognitive and metacognitive processing, and the task environment (Flower and Hayes, 1981;Hayes, 1996;Winne and Hadwin, 1998). According to Zimmerman and Risemberg (1997), becoming an adept writer depends on high level use of selfregulated learning (SRL) strategies as writing is a goal-driven, selfinitiated, and self-maintained activity.
A second gap in the existing literature is that few studies have explored how students writing reflective journals demonstrate and develop their awareness of different types of SRL-strategy use, or how individual learner differences influence self-regulated writing processes (Farahian et al., 2021). With regard to the latter, L2 proficiency level is an important individual-difference factor, yet findings about its influence are relatively limited (Yabukoshi, 2020). More studies of how students with varied writingproficiency levels reflect on their writing processes via reflective journals are therefore warranted.
Accordingly, the present work assigned Chinese students of English as a foreign language (EFL) three journal-writing tasks, each consisting of reflection on their writing processes, and examined the resulting data for (1) the types of self-regulated writing strategies they used and changes to those strategies over time; (2) how students with varied writing-proficiency levels differed in their use of self-regulated writing strategies; and (3) the effects of reflective-journal writing on students' self-perceived selfregulated writing strategy use in particular and on their selfperceived L2 writing in general. It is hoped that its findings will equip L2 educators who are considering using reflective journals with fresh insights that could help them increase their students' SRL levels and provide more individualized SRL guidance.
Literature review
The use of self-regulated learning strategies in L2 writing SRL in first-language (L1) writing has been studied widely (e.g., Zimmerman and Risemberg, 1997;Zimmerman and Kitsantas, 2002), but the equivalent body of work on SRL in L2 writing started relatively late and remains small (Bai and Guo, 2018). This is surprising, given that writing is an inherently complex activity that requires learners to exhibit strong cognitive as well as linguistic abilities (Hayes, 1996;Kellogg, 1996), and that L2 writers generally strive to obtain language knowledge during the writing process (Manchón, 2011). L2 writing is also timeconsuming, requiring learners' motivation and regulation (Kormos, 2012). Therefore, in theory, SRL could play a key role in determining how students initiate, sustain, regulate, and monitor their L2 writing processes.
Unlike classifications of L2 writing strategy from a cognitive perspective that classified writing strategies in terms of three cognitive phases (i.e., pre-writing, composing, and revising; e.g., Cumming, 1989;Wenden, 1991), some studies have sought to establish which strategies or other factors are key to L2 writing from a socio-cognitive angle (e.g., Kormos, 2012;Csizér and Tankó, 2017;Sun and Wang, 2020). This line of research integrated social and motivational aspect into writing strategy use by highlighting the important role of SRL in writing (e.g., Cumming et al., 2002). Such research was thus regarded as more comprehensive as it acknowledged "the multidimensional nature of the writing process" (Teng and Zhang, 2016a, p. 5). For example, following the framework created by Zimmerman and Risemberg (1997), Sun and Wang (2020) used a questionnaire covering three dimensions, i.e., environmental, behavioral, and personal factors, to measure English writers' SRL strategies. In another study, Teng and Zhang (2016a) validated a questionnaire designed to measure SRL strategy use in EFL writing. They regarded self-regulation as a higher-order construct over nine lower-level writing strategies grouped according to whether they involved cognitive, metacognitive, social-behavioral or motivational regulation. In Frontiers in Psychology 03 frontiersin.org general, to achieve sustainable L2 writing, a person must actively employ four distinct types of self-regulated writing strategies: cognitive, metacognitive, behavioral, and motivational (e.g., Teng andZhang, 2016a, 2018;Csizér and Tankó, 2017;Hu and Gao, 2018). Learners use cognitive strategies to learn to accomplish tasks using cognitive abilities such as knowledge processing, constructing, rehearsing, and transforming (Zimmerman, 2000); use metacognitive strategies to plan, monitor, control, and evaluate the entire learning process (Pintrich, 2000;Zimmerman, 2000); use behavioral strategies to select or create an adaptive learning environment (Pintrich, 2000); and finally, use motivational strategies to actively manage their motivation or motivational processing during learning (Zimmerman and Schunk, 2008). In general, the quantity of using self-regulated writing strategies by L2 writers is moderate (e.g., Abadikhah et al., 2018;Bai et al., 2022). This also holds true of its subcomponents including cognitive strategies (Sethuraman and Radhakrishnan, 2020), planning Guo, 2018, 2021;Bai et al., 2020), selfmonitoring , help-seeking, and motivational regulation (Mbato and Cendra, 2019). On the other hand, some studies have reported that students did not fully adopt selfregulated writing strategies (e.g., Mallahi, 2020;Akhmedjanova and Moeyaert, 2022), with three strategies being especially poorly represented. They were goal-setting (Abadikhah et al., 2018;Sun and Wang, 2020), reviewing of records (Sun and Wang, 2020), and self-consequence (Abadikhah et al., 2018). Meanwhile, Xu (2021) found that Chinese university EFL students adopted fewer socialbehavior strategies than other self-regulated writing strategies.
Studies of the associations between L2 learners' writing proficiency and various aspects of self-regulated writing strategies have, in general, reported them to be positive (e.g., Teng and Zhang, 2016a;Hu and Gao, 2018;Bai et al., 2020;Jackson and Park, 2020;Sun and Wang, 2020;Bai and Guo, 2021;Guo and Bai, 2022;Shen and Bai, 2022). In particular, such findings have involved the strategies of noticing (Hu and Gao, 2018), selecting (Hu and Gao, 2018), text processing (Alanazi, 2020), memorizing (Hu and Gao, 2018;Alanazi, 2020), planning (Chien, 2012;Bai et al., 2020Bai et al., , 2022Guo and Bai, 2022), text-generation (Chien, 2012;Bai et al., 2020Bai et al., , 2022, self-monitoring/evaluation (Hu and Gao, 2018;Bai et al., 2020Bai et al., , 2022Sun and Wang, 2020;Guo and Bai, 2022), reviewing of records (Sun and Wang, 2020), seeking opportunities (Sun and Wang, 2020), revising (Chien, 2012;Bai et al., 2020;Guo and Bai, 2022), and motivational regulation (Teng and Zhang, 2018;Teng et al., 2020;Shen and Bai, 2022). However, some other studies have reported non-significant correlations between writing proficiency and the use of SRL strategies (e.g., Mallahi, 2020). For example, Csizér and Tankó (2017) found that even students with relatively high language proficiency might not be aware of the importance of SRL strategy use. Such results imply that levels of using self-regulated writing strategies and the relationship between strategy use and writing proficiency are both subject to marked individual and environmental variations (see also Inan-Karagul and Seker, 2021). Regarding change over time in self-regulated writing strategies, SRL theory itself emphasizes the central role of time and ordered SRL sequencing during the writing process (Zimmerman, 2008;Zimmerman and Schunk, 2011), and that students' journey to adopting such strategies may be quite lengthy (Akhmedjanova and Moeyaert, 2022). Therefore, it is important to trace developmental patterns in students' use of them. Some studies have reported fluctuating use of self-regulated writing strategies over time: with Jackson and Park (2020), for example, capturing three participants' dynamic views of SRL at nine time-points over a semester. In general, their results suggested that these individuals' SRL did not increase much over the semester, and that the nature of its fluctuations varied from person to person. Similarly, Wilby (2020) reported that although students' writing motivation and self-efficacy increased significantly over a semester, their metacognitive self-regulation skills (i.e., planning, monitoring, self-control, and self-reflection) did not exhibit significant change. Wilby speculated that this could have been due to their lack of opportunities to reflect on their work -a process that, according to Pintrich (2000), plays a key role in SRL. This further highlights the need to build reflection-related activity into L2 writing instruction. In a more recent study, Saqr et al. (2021) showed how SRL behavior deteriorated over time in an academic-writing class. Specifically, both writing arguments and reflecting were prominent during the students' completion of their first assignment, but their incidence dropped during the second and the third assignments.
However, some other studies have suggested that self-regulated writing strategy use can improve over time. For example, Han and Hiver (2018) identified three profiles of learners based on their SRL, self-efficacy, and anxiety about English writing. The authors found that, throughout the focal semester, the students who fit each of these profiles either developed or consolidated their SRL. The same authors suggested that such improvement may have involved the instruction the students received, which scaffolded their attention to task requirements, use of linguistic resources, and choices as writers. In the same year, Sasaki et al. (2018) reported the results of their three-and-a-half-year observation of how Japanese EFL learners used three writing strategies: i.e., global planning, local planning, and L1-to-L2 translation. Their participants tended to use both global planning and local planning strategies more frequently as time went on, and their L2 writing proficiency also increased. The same authors also pointed out that their participants' developmental pattern was not linear; that between-subjects variation overshadowed its within-subjects counterpart; and that this overshadowing became more marked with the passage of time. In sum, the change of selfregulated writing strategy use over time appears nearly impossible to generalize. Instead, it should be examined on a case-by-case basis, especially when considering individual difference and the effects of particular pedagogical practices, such as reflection.
The use of reflective journals in L2 learning
Reflective journals are "written documents that students create as they think about various concepts, events, or interactions over a Frontiers in Psychology 04 frontiersin.org period of time for the purposes of gaining insights into selfawareness and learning" (Thorpe, 2004, p. 328). From a sociocultural perspective, keeping a reflective journal is a dialogic activity that triggers an interaction between a learner and him-or her-self (Rassaei, 2015). At the core of reflective-journal writing is one's ability to reflect on one's own learning process and progress and thus obtain new knowledge that can guide future learning actions (Moon, 2006;Chang and Lin, 2014;Rassaei, 2015;Hussein, 2018;Baek, 2019;Hussein et al., 2020;Sudirman et al., 2021). In this sense, reflectivejournal writing would appear to be closely linked to the final selfreflection phase of Zimmerman's (2000) cyclical modal of SRL. Various advantages of reflective journals in L2 contexts have been documented, including their potential use as indices of improvement in writing effectiveness (Wu and Lin, 2015;Hussein et al., 2020;Lv et al., 2021;Sudirman et al., 2021), reading effectiveness (Chang and Lin, 2014), recast effectiveness (Rassaei, 2015), and a growth mindset (Hussein, 2018). Some studies have reported that the use of reflective writing can promote students' SRL in a general way (Jafarigohar and Mortazavi, 2013;Chang et al., 2015), and others, in specific ways: notably, by boosting organizational skills (Chang and Lin, 2014), critical reflection (Sudirman et al., 2021), and critical thinking (Ahmed, 2020). Alongside these broadly positive findings about reflective-journal writing, however, a recent study by Akhmedjanova and Moeyaert (2022) reported that a sample of eight Southeast Asian students did not value the use of SRL journals in the process of learning and writing English, and that at least five of them considered the SRL journal assignments to be annoying and boring. This further highlights the importance of taking account of contextual variation when examining learners' self-regulated writing strategy uptake (Inan-Karagul and Seker, 2021).
As briefly noted above, individual variation, notably in proficiency levels, also needs to be considered. For example, in a study by Farahian et al. (2021), an experimental group that used reflection sheets did not exhibit a significantly higher level of critical reflection than a control group. The authors speculated that learners' L2 proficiency levels could have played a role in limiting their use of higher levels of reflection. Similarly, Yabukoshi (2020) asked four EFL learners to document their SRL processes in reflective journals and found that while the two high achievers' writings demonstrated high self-regulation skills, good selfmotivation, and adaptive decisions about future plans, those of the two low achievers were less likely to include reflection on their learning experiences or to identify their problems. Subsequently, Akhmedjanova and Moeyaert (2022) reported that their only student participant who understood key aspects of SRL-related journal writing or its value also scored highest on the baseline essay. This finding provided further evidence of a potential relationship between pre-existing proficiency and self-regulated writing strategy use while writing reflective journals. However, given the abovecited studies' very small sample sizes, more research should be conducted to shed light on this possible relationship.
Another under-explored issue arises from the fact that most previous studies have treated reflective-journal writing simply as a means to an end, rather than investigating how students demonstrate and develop their awareness of different types of SRL strategies while engaged in it. The few exceptions notably include Yabukoshi (2020), who used reflective journals as evidence of L2 students' individual variation in metacognitive awareness. However, that study's analysis of SRL focused disproportionately on its metacognitive side, and on phases rather than on strategies. Another exception, Baek (2019), showed that affective learning strategies were used most frequently, and memorization strategies, least frequently, in students' reflective journals. However, Baek's context was EFL reading, not writing, and the coding scheme did not fully represent the SRL framework. More recently, Akhmedjanova and Moeyaert (2022) coded SRL in students' reflective journals to show how different aspects of it (i.e., goal-setting, task management, progress monitoring, and reflection) differed before and after an SRL intervention; but again, a full picture of self-regulated writing strategy use cannot be provided by just those four areas. Lastly, Teng (2022) utilized reflective journals as a source of data for exploring EFL learners' perceptions of a formative assessment in which an SRL intervention was embedded. The results suggested that EFL students reported increased levels of self-regulated behaviors. In particular, Teng adopted a combination of bottom-up and top-down coding approaches to explore how students engaged with the SRL intervention in terms of the relative value they placed on goalsetting, peer learning, feedback-handling, self-assessment, and reflection on their performance. But in any case, studies investigating how students demonstrate and develop their use of self-regulated writing strategies in reflective journals have been quite limited in terms of both amount and perspectives, suggesting the need for more exploration toward a comprehensive picture.
Based on the above review of the existing literature, we asked the following research questions: RQ1. When they write reflective journals, how do students demonstrate their use of self-regulated writing strategies, and does such strategy use change over time?
RQ2. How do students with varied writing-proficiency levels differ from one another in terms of their use of self-regulated writing strategies when writing reflective journals?
RQ3. Does the use of reflective journals affect students' perceptions of their own self-regulated writing strategies and/or L2 writing?
Methodology Participants and context
This study took place in an undergraduate course at a top Chinese university during the fall semester of 2021. A total of 38 English majors participated in this research with informed consent. Among them, there were 22 females (57.9%) and 16 males (42.1%), ranging in age between 17 and 21 (M = 18.16, SD = 0.89). Because most of them did not have standardized writing-test scores (e.g., TOEFL or IELTS) for reference, they were asked to rate their own Frontiers in Psychology 05 frontiersin.org English-writing proficiency on a five-point scale, ranging from "1" = "very poor" to "5" = "very good. " Their average self-rated proficiency was 3.03 (SD = 0.75). Before participating in this study, none of them had received any self-regulated strategy instruction. The course, titled English Writing and Critical Thinking, lasted 15 weeks. The class met twice a week for 90 min each time. It was aimed at equipping its students with the basic skills of critical thinking and enabling them to use such skills in English academic writing. Over the semester, the students were required to submit five writing assignments. In the first assignment, they were expected to write a one-to two-page introduction to an essay. The second was an annotated bibliography of three to five references they planned to use in the same essay. The third assignment was a detailed outline of the essay, showcasing the structure of each of its parts as well as the claim(s), subclaim(s), and evidence that the writer intended to use. The fourth was a six-to eight-page draft of the essay developed on the basis of the outline. The final course assignment was the eight-to 10-page essay itself. In addition, there were three peerreview sessions held after the submission of the second, third, and fourth assignments, respectively. In them, students were divided into groups according to their essay topics and offered feedback to their fellow group members on conventions, content, and composition.
Procedures
Our data-collection procedures are presented in Table 1. Data were collected from questionnaires and reflective journals. At the beginning of the semester, students filled out a background information questionnaire and a self-regulated writing strategy use questionnaire. In Week 7, they filled out the self-regulated writing strategy use questionnaire again and wrote the first reflective journal. In Week 11, they submitted the second reflective journal, after finishing the second and the third assignments and experiencing two peer review sessions. In Week 15, the students completed the self-regulated writing strategy use questionnaire for a third time. Two weeks later, they submitted their final reflective journal after completing all of their coursework.
Instruments Questionnaire
The questionnaire used in this study had two parts. The first part comprised five items covering the respondent's name, ID number, age, gender, and self-rated English-writing proficiency. These items were aimed at offering an overview of the participants' background information.
The second part consisted of 35 items on self-regulated writing strategy use, all responded to on the same five-point Likert scale ranging from 1 = "strongly disagree" to 5 = "strongly agree" (Appendix I). The 35 items collectively covered four dimensions: cognition, metacognition, social behavior, and motivational regulation. The 21 items that measured the first three of those dimensions were adapted from Teng and Zhang (2016a) Writing Strategies for Self-regulated Learning Questionnaire (WSSRLQ). It should be noted that course memory, one of the two cognitive subcategories in the original questionnaire, was not included, because it was irrelevant to the design of the current study's focal course. Motivational regulation was measured using the 14-item L2 Writing Strategies for Motivational Regulation Questionnaire (L2WSMRQ), adapted from Teng and Zhang (2016b). In this case, environmental structuring, the last subcategory in the original questionnaire, was not pertinent to the focal course's design and was therefore removed.
As the questionnaire was filled out by the participants three times to obtain process information about their self-regulated writing-strategy development, the internal reliability of each subcategory was also tested at those three time points (T1 = Week 1, T2 = Week 7, and T3 = Week 15). The results of those tests are Week 1 The background information questionnaire and the self-regulated writing strategy use questionnaire Week 6 Assignment 1 (introduction) Week 7 The self-regulated writing strategy use questionnaire The first reflective journal Week 8 Assignment 2 (planned references) Week 9 First peer review Week 10 Assignment 3 (outline) Week 11 Second peer review The second reflective journal Week 13 Assignment 4 (essay draft) Week 14 Third peer review Week 15 The self-regulated writing strategy use questionnaire Week 17 Assignment 5 (final essay) The third reflective journal Frontiers in Psychology 06 frontiersin.org presented in Table 2. It should be noted that although most of the scales at different time points reached a satisfactory level (i.e., larger than 0.70; Tavakol and Dennick, 2011), some scales showed low reliability. For example, the reliability of emotional control was low at three time points, and mastery self-talk was low at T1. We postulate that this may be due to the low number of items (i.e., three items) of the measured scale, which, as Pallant (2020) pointed out, could easily lead to low Cronbach values.
Reflective journals
The students received general guiding questions (Appendix II) before each reflective journal writing session to help ensure that their journal content would be relevant but not restricted. The first two reflective journals, hereafter J1 and J2, were based on the first and the third assignments, respectively. Therefore, their guiding questions focused on changes between the first drafts and the submitted versions of these assignments; the strengths and weaknesses of their work; and their plans for future improvement. The last reflective journal, J3, was a reflection not only on the final essay but also on the learning process throughout the semester, and the guiding questions were tailored to both purposes. To better capture their experiences and feelings, the students were encouraged to write in their native language, with no limitation on word count. Those non-English entries were translated into English for report by the second, third, and fourth authors, and the first author checked all translation for accuracy. Students were assured that their journals' content would not be rated or influence their course grades in any way, and only the researchers have access to their work. To avoid potential researcher bias and for reference in this research, the journal data were anonymized by replacing name information with an ID number ranging between 1 and 38. Because two participants failed to submit their final reflective journal entries, 112 were collected: i.e., 38 in the first round, 38 in the second, and 36 in the third.
Writing scores
The final scores students received on their essays were used to determine their writing proficiency levels for further analysis. The scores were given by the course instructor according to students' performance in five dimensions: disciplinary content understanding, quality of argument, use of sources, responsiveness to the question, and clarity/focus of the writing. The theoretical maximum score for each essay was 100 and the minimum was 0.
Data analysis
To answer our first research question, the second and third authors familiarized themselves with the reflective journal entries by reading them several times prior to coding. The entries were then coded using a combination of top-down and bottom-up approaches and continuously checked against the theoretical framework developed by Teng and Zhang (2016a) and Zhang andZou (2022a, 2022b). Every time an act of strategy use was identified in a journal, it would be counted as one time of application for the corresponding strategy (or strategies, as sometimes one act might reflect multiple strategies). The total number of use for every strategy would be recorded for further analysis. If the same act was mentioned repeatedly, it would be recorded only once, so it was a count of "acts" rather than "accounts, " unaffected by how many words or details used for description. To ensure the credibility of coding, the same two researchers first coded the journal entries independently and then discussed their decisions with each other and with the first author, until all disagreements were resolved. The final coding scheme can be found in Appendix III. Since the data were not normally distributed, Friedman tests were conducted to explore whether the students' use of self-regulated writing strategies changed over the semester. To further pin down when the differences actually occurred, the data collected from the three reflective journals were compared pairwise using sign tests. A Bonferroni adjustment was made to the probability criterion of statistical significance (p < 0.05/3 ≈ 0.017).
To answer our second research question, the students were divided into high-, medium-, and low-proficiency groups based on the final scores they received on their essays. The top 25% (10 students) and bottom 25% (10 students), respectively, made up the high-proficiency and low-proficiency groups (as per Sun et al., Frontiers in Psychology 07 frontiersin.org 2021). The remaining 50% belonged to the medium-proficiency group (18 students). The overall use of self-regulated writing strategies exhibited in reflective journals by the high-and low-proficiency groups were then compared using Mann-Whitney U tests, due to the non-normal data distribution.
To answer our final research question, quantitative and qualitative analyses were conducted. The quantitative part used one-way repeated-measures analysis of variance (ANOVA) on the questionnaire data to compare the differences in self-perceived self-regulated writing strategies at three time points. The data met the assumptions of normal distribution.
The results of Mauchly's test suggested that all variables except for peer learning (Mauchly's W = 0.78, p = 0.01, Greenhouse-Geisser = 0.82, Huynh-Feldt = 0.85) violated the assumption of sphericity. Thus, for the peer-learning variable, the Huynh-Feldt correction instead of F-ratio was used in the relevant one-way repeated-measures ANOVA (Field, 2009). To further determine when significant differences occurred in the measured SRL strategy variables, post-hoc tests with Bonferroni adjustment (p < 0.05/3 ≈ 0.017) were conducted.
Our qualitative analysis chiefly addressed students' perceptions of the use of reflective journals according to their statements in the final reflective journal. The journal entries were read carefully by the third and the fourth authors. Then, initial open codes were allocated to the recurrent expressions of a given idea. After identifying the open codes, the researchers used axial coding and grouped the codes into the following 14 topical categories: summary, review, self-talk, self-recognition, reminder, record, improvement, inspiration, envisioning, feeling of accomplishment and satisfaction, willingness to keep on writing and make progress, strengthening writing techniques, instructor connection, and negative comments. In the final stage, the researchers used selective coding to merge the recognized categories into six distinct themes (see Appendix IV for the coding scheme and Appendix V for samples of students' reflective journals).
Use and change of self-regulated writing strategies demonstrated in reflective journals
A total of 963 descriptions of self-regulated writing strategy use were identified in the 112 reflective journal entries we collected. Among the 19 kinds of strategies, feedback handling (22.43%), resource management (12.98%), text processing (9.45%), emotional control (9.45%), and idea planning (9.03%) were the top five most frequently seen, accounting for almost two-thirds of all occurrences of self-regulated writing strategies. A full description of the distribution can be found in Table 3. Table 4 presents the results of Friedman tests, which show statistically significant changes in the use of eight strategies: selfmonitoring and evaluation (X 2 = 9.08, p < 0.05), idea planning (X 2 = 9.12, p < 0.05), text processing (X 2 = 18.12, p < 0.001), perspective change (X 2 = 7.58, p < 0.05), emotional control (X 2 = 28.83, p < 0.001), effort regulation (X 2 = 13.02, p < 0.01), peer learning (X 2 = 29.96, p < 0.001), and resource management (X 2 = 11.52, p < 0.01). For these eight strategies, the results of post-hoc tests ( Table 5) further indicated that: (1) between J1 and J2, text processing decreased at the p < 0.001 level; (2) between J2 and J3, self-monitoring and evaluation increased at the p < 0.05 level, resource management increased at the p < 0.01 level, and both emotional control and peer learning increased at the p < 0.001 level; and (3) between J1 and J3, both idea planning and effort regulation increased at the p < 0.01 level, and both emotional control and peer learning increased at the p < 0.001 level. Differences in the use of self-regulated writing strategies between the high-and low-proficiency groups As shown in Table 6, Mann-Whitney U tests revealed significant differences between the high-and low-proficiency groups in their use of four self-regulated writing strategies as demonstrated in their reflective journals. These were: idea planning (Z = −2.46, p < 0.05), goal-setting (Z = −2.01, p < 0.05), feedback handling (Z = −2.32, p < 0.05), and resource management (Z = −4.16, p < 0.001). Specifically, the high-proficiency group applied idea planning, feedback handling, and resource management significantly more often than the low-proficiency group did, with mean rank differences of 10.14, 10.14, and 17.74, respectively. The low-proficiency group, in contrast, tended to use goal-setting significantly more frequently, with a mean rank difference of 5.04.
The relationship between writing reflective journals on self-regulated writing strategies and L2 writing Table 7 presents the descriptive statistics and results of one-way repeated-measures ANOVA of change over time in the participants' self-perceived self-regulated writing strategy use reflected in the questionnaire data. The use of reflective journals was associated with statistically significant increases in the mean scores assigned to the following four strategies: idea planning (p < 0.001), goal-oriented monitoring (p < 0.05), peer learning (p < 0.01), and interest enhancement (p < 0.001). Results of post-hoc tests (Table 8) indicated that, for idea planning, the statistically significant changes occurred between T1 and T3, with a mean difference of 0.36 (p < 0.001), and between T2 and T3, with a mean difference of 0.22 (p < 0.05). For peer learning, the statistically significant change occurred between T2 and T3, with a mean difference of 0.34 (p < 0.05); and for interest enhancement, such changes occurred between T1 and T3, with a mean difference of 0.30 (p < 0.01), and between T2 and T3, with a mean difference of 0.36 (p < 0.001). However, for goal-oriented monitoring, no statistically significant change can be said to have occurred. In Frontiers in Psychology 08 frontiersin.org other words, as compared with the first half of the semester, there was an accelerated increase in the students' use of idea planning, peer learning and interest enhancement in the second half. Our qualitative analysis of the participants' perceptions of reflective-journal writing revealed that in general, they held a positive attitude toward it. First, they perceived such writing as a way of promoting self-monitoring and evaluation. To be more specific, they perceived it as a means of: (1) summarizing, for example, "Reflective writing helps me summarize what I have learnt in the process of writing and revising, which would benefit my future writing" (Participant 24); (2) reviewing, for example, "Reflective writing enables me to look back on what I have learnt and record my personal growth in a timely manner" (Participant 8); (3) engaging in self-talk, for example, "Reflective writing is like a dialog with myself" (Participant 30); (4) engaging in self-recognition, for example, "Reflective writing gives me a chance to evaluate strengths and weaknesses in my writing" (Participant 31); (5) reminding, for example, "Reflective writing reminds me to focus on my problems in writing and push me to think how I can do better next time" (Participant 16); and (6) recording, for example, "In reflective writing, I would write down my thoughts before, during, and after writing my essays" (Participant 20).
Second, these students perceived reflective writing as a process of goal-setting and future idea planning. This included: (1) writing improvement, for example, "I always review my reflective writings before starting to write in order to make improvements and avoid past problems" (Participant 16); (2) inspiration, for example, "I am greatly inspired by reflective writing, from which I learn (4) extension, for example, "Reflective writing not only helps me in writing, but in every other course I am taking" (Participant 10). Third, the participants perceived reflective writing as a way of stimulating their motivation. Through writing reflective journals, they said, they gained (1) a sense of accomplishment and satisfaction, for example, "Reflective writing helps me polish my essays, which gives me a sense of accomplishment" (Participant 20) and (2) willingness to keep on writing and make progress, for example, "Through reflective writing, I realize that many problems are not as difficult to solve as I expected. I am willing to find out the root cause and solve it" (Participant 36).
Fourth, they perceived reflective writing as a way of strengthening their writing techniques. One examples of this perception was "Through reflective writing, I pay close attention to my language, logic, and supporting facts. My writing greatly improves after my reflection" (Participant 27). The improvement in writing techniques was "systematic" (Participant 10), as Participant 27 put it: "The awareness of constantly paying attention to logical coherence and using facts to support points of view has been deeply engrained in every piece of writing of mine." Finally, the participants perceived reflective writing as a way of establishing connections with and gaining support from their teachers, for example, "Reflective writing enables me to create a close connection with my teacher. I have learnt a lot from her feedback on my reflective writing. She also encourages me a lot" (Participant 7). Reflective writing was especially useful when they were challenged with negative feelings. "I told my teacher about my anxiety and confusion through reflective writing and felt very lucky to get encouragement and suggestions from her," wrote Participant 25.
Although the majority of the sampled students expressed positive thoughts and feelings about reflective-journal writing, Frontiers in Psychology 10 frontiersin.org some responses characterized it as "time-consuming" (Participant 17), "causing anxieties" (Participant 11), and even "miserable" (Participant 18). Some felt that this activity had barely any effect on the quality of their writing, for example, "I still cannot guarantee I'll get a higher score next time" (Participant 35), and that "Even if I am aware of my writing problems, I may make mistakes again in my next writing assignment. Reflective writing is not an easy way to completely correct my mistakes" (Participant 33).
Discussion
Use and change of self-regulated writing strategies as demonstrated in reflective journals Some previous questionnaire-based studies have suggested a moderate-to-high level of self-regulated writing strategy use (e.g., Abadikhah et al., 2018;Mbato and Cendra, 2019;Bai and Guo, 2021). The present study found more variations: among the 19 kinds of strategies identified in reflective-journal writing, only five (i.e., handling feedback, resource management, text processing, emotion regulation, and idea planning) were demonstrated relatively frequently, accounting for almost two thirds of all appearances of self-regulated writing strategies. On one hand, this indicates the special link between reflective-journal writing and SRL, as evidenced by the fact that all four types of self-regulated writing strategies (i.e., cognitive, metacognitive, behavioral, and motivational) are reflected (e.g., Teng andZhang, 2016a, 2018;Csizér and Tankó, 2017;Hu and Gao, 2018). On the other, it suggests a disproportionate prioritization of some types of strategies (e.g., handling feedback) over others.
Interestingly, the above results contradict Xu's (2021) finding that Chinese EFL writers adopted fewer social-behavior strategies than other self-regulated writing strategies. A possible explanation for this discrepancy is that the choice of social-behavior strategies is closely linked to specific guidance provided by English teachers (Bai et al., 2020). In our study, the teacher placed a high value on article revision, and the course's second and subsequent writing assignments depended heavily on previous ones. These distinctive course features both led to the students making strong efforts to address how they handled feedback from the teacher and their peers. In addition, they used reflective-journal writing as a major channel for conversing with the teacher as well as with themselves, implying the dialogic nature of such writing from a sociocultural perspective (Rassaei, 2015). On the whole, we believe this is a positive sign about reflective writing, i.e., that it pushes students both to think critically, and to deal carefully with the feedback they receive (Thorpe, 2004;Ahmed, 2020;Sudirman et al., 2021) even if they do not fully understand it (Zhang andZou, 2022a, 2022b). At the same time, however, we argue thatgiven the overall key role of SRL in determining how students initiate, sustain, regulate and monitor their L2 learning (Dörnyei, 2005;Tseng et al., 2006), teachers designing guidance for reflective-journal writing should pay special attention to those strategies that were infrequently used by our participants, that is, strategies which learners may be unfamiliar with and unable to use. Such explicit strategy instruction would be beneficial as regulation may benefit from shifting from being self-regulation to co/other-regulation, so that students are prepared to be successful self-regulated learners (Thomas and Rose, 2019;Thomas et al., 2021). Regarding change over time in the use of self-regulated writing strategies, our journal data firstly suggest that not all such strategies grew significantly. This echoes previous findings (Jackson and Park, 2020;Wilby, 2020) and further confirms the complex nature of the use of self-regulated writing strategies in the real world (Oxford, 2017). Secondly, these data show that seven strategies (self-monitoring and evaluation, idea planning, perspective change, emotional control, effort regulation, peer learning, and resource management) did significantly increase over time, and most such changes occurred during the second half of the semester, adding empirical evidence to the existing literature regarding such improvement (Han and Hiver, 2018;Sasaki et al., 2018). Importantly, these seven strategies reflected all four aspects of selfregulated writing (i.e., cognitive, metacognitive, behavioral, and motivational), lending support to the idea that reflective writing can promote SRL-related strategies (e.g., Jafarigohar and Mortazavi, 2013;Chang et al., 2015). And thirdly, we found that text processing decreased significantly from J1 to J2. In this study, text processing was defined as using linguistic knowledge (e.g., grammar, spelling, punctuation, clear expression) to revise or improve writing texts, so our finding could be explained by the fact that novice L2 writers usually pay more attention to "mechanics rather than content" (Mu and Carrington, 2007, p. 10) and are relatively lacking in knowledge of the generic, discourse, and rhetorical aspects of academic writing (e.g., Swales, 2004;Lillis and Curry, 2010). As the course progressed, the students received an increasing amount of overt instruction in generic, discourse, and rhetorical moves (e.g., Cargill et al., 2018;Li and Flowerdew, 2020), and this could explain why less text processing was found in the latter half of the course. Differences in the use of self-regulated writing strategies between the high-and low-proficiency groups Some previous studies on the use of reflective writing have tended to hypothesize a potential influence of individual proficiency levels on observable SRL-related behaviors (Farahian et al., 2021), while others argued that such a link exists while treating SRL strategy use as a holistic construct (Yabukoshi, 2020;Akhmedjanova and Moeyaert, 2022). Our study has not only revealed the influence of individual-proficiency variation on the use of self-regulated writing strategies when writing reflective journals, but also paints a clear picture of where such variation lies. First, we found that our highproficiency group applied idea planning, feedback handling, and resource management significantly more often than the low-proficiency group did. This clearly reflects the critical role of the three major processes of L2 writing, i.e., planning, composing, and reviewing (Flower and Hayes, 1981;Hayes, 1996), and especially how "before writing, " "during writing, " and "after writing" strategies may all contribute to writing. This finding extends, to a new context, previous ones regarding how expert writers outperform novice ones in planning (e.g., Chien, 2012;Bai et al., 2020Bai et al., , 2022Guo and Bai, 2022). However, our finding regarding the roles of feedback handling TP, text processing; IP, idea planning; GM, goal-oriented monitoring; PL, peer learning; FH, feedback handling; IE, interest enhancement; MS, mastery self-talk; EC, emotional control; and PS, performance self-talk.
Frontiers in Psychology 12 frontiersin.org and resource management in differentiating students do not appear to have been reported in any previous study. We believe this finding could add weight to arguments in favor of the use of reflective journals to promote SRL strategy-use behavior, over and above this practice's known cognitive and metacognitive advantages (Thorpe, 2004;Ahmed, 2020;Sudirman et al., 2021). The same finding may also have important pedagogical implications for L2 teachers who design, assign, and assess reflective journals, as their support could be important for students on their way toward successful selfregulated learners (Thomas and Rose, 2019;Thomas et al., 2021). In particular, for those learners who have difficulties in writing these journals in particular, additional scaffolding for idea planning, feedback handling, and resource management should be put in place. Additionally, we found that while high-proficiency students reported significantly more idea planning strategies in their reflectivejournal entries than their low-proficiency counterparts did, the former demonstrated significantly less goal-setting than the latter. A clarification of these two constructs could help explain this apparent contradiction. Unlike Sun and Wang (2020), who combined them under the label goal-setting and idea planning, we found them to be clearly distinguishable from each other in our data. That is, goalsetting involved setting more abstract goals and sub-goals of learning, for example, "Although I am still unable to have perfect logic in my article, this should be a goal for me to work on, and I will continue to improve my writing ability by applying this academic writing method" (Participant 30). Idea planning, on the other hand, was more specific, executable and task-oriented, for example, "I planned to study the changes in female images in Chinese TV series in chronological order and argue that such changes did not alter people' s stereotypes of females" (Participant 3). In this context, it should also be noted that idea planning and goal-setting as captured in our study were consistent with how those two constructs were conceptualized in some previous research (e.g., Zhang, 2010). For example, in Teng and Zhang's (2016a) WSSRLQ, one example given for idea planning is "Before writing, I read related articles to help me plan, " whereas an example of goal-oriented monitoring and evaluating is "When learning to write, I set up goals for myself in order to direct my learning activities" (p. 700). It may also be important that, although the arrangement of personal goals is vital in SRL Schunk, 2008, 2011), only goals that incorporate specific performance standards, are proximal, and are of a reasonable difficulty level are widely considered 'good' goals (Locke et al., 1981;Bandura, 1988). Such goals promote performance by focusing on the amount of effort required for success and self-anticipation (Schunk, 1990). Given that most of the goals we identified in our journal data were rather broad and vague, it is perhaps unsurprising that low-proficiency students demonstrated significantly more use of goal-setting than high-proficiency students did.
The use of reflective-journal writing
The qualitative results suggest that the practice of journaling raised students' awareness and may have contributed to an increase in their self-regulated writing strategy use. This finding is consistent with a number of previous studies regarding the positive role of reflective writing in L2 writing instruction (e.g., Chang and Lin, 2014;Chang et al., 2015;Rassaei, 2015;Wu and Lin, 2015;Hussein, 2018;Baek, 2019;Hussein et al., 2020;Farahian et al., 2021;Lv et al., 2021;Sudirman et al., 2021). In addition, our participants reported that reflective journaling not only helped to strengthen their writing techniques, but also made them feel more connected to their teacher. From a sociocultural point of view, this finding illustrates a co-regulated language learning mode to be explored (Gao and Hu, 2020;Thomas et al., 2021) and highlights the importance of encouraging students to have conversations with themselves and with the teacher (Rassaei, 2015) as a means of gaining new knowledge (Moon, 2006;Chang and Lin, 2014;Hussein, 2018;Baek, 2019;Hussein et al., 2020;Sudirman et al., 2021). Our findings augment the literature by showing how reflective journals may create a space for (1) developing one's goalsetting and idea-planning ability; (2) enhancing one's interest in writing subsequent assignments; and (3) fostering one's ability to learn from peers. These findings highlight how crucial reflective journaling could be to SRL's self-reflection phase (Zimmerman, 2000) and all four aspects of L2 writing (e.g., Teng andZhang, 2016a, 2018;Csizér and Tankó, 2017;Hu and Gao, 2018). In addition, our findings suggest that reflective writing is a sociocultural activity (Rassaei, 2015), which (1) is capable of facilitating the formation of writing communities (Graham, 2018), (2) treats writing as a process rather than a product (Murray, 1972), and (3) deepens teacher-student relationships (Lee and Schallert, 2008). All of these factors seem likely to contribute positively to various aspects of students' self-regulated writing strategy growth.
On the other hand, the sporadic negative comments we received on reflective writing echoed Akhmedjanova and Moeyaert's (2022) findings of students failing to recognize the value of writing SRL journal, despite the relatively lower burden of our reflective-writing assignments. A possible explanation could be that the students who complained were relatively lacking in declarative knowledge about SRL strategies (Mallahi, 2020;Sun and Wang, 2020;Akhmedjanova and Moeyaert, 2022), and perhaps especially about the power of reflection in writing (e.g., Hussein, 2018;Baek, 2019;Farahian et al., 2021;Sudirman et al., 2021). If so, they would have been less likely to access these mental resources while composing both reflective writing and writing assignments. In addition, students' individual-difference factors such as their attitudes toward and experience of reflective writing (Atkinson, 2016) could also have led to negative perceptions of our journalwriting assignments. However, the existence of these few negative comments should not be taken as detracting from the participants' overall positive perceptions of reflective-journal writing.
Conclusion
To maximize the benefits of self-reflection, many L2 educators require their students to write reflective journals (e.g., Takeuchi, 2019;Sudirman et al., 2021), but the relationship between this practice and self-regulated writing strategy use has long remained unclear. To help clarify this issue, the present study began by tracing how students demonstrated and developed their awareness of different types of SRL strategies when writing reflective journals. It found such demonstration and its development to be uneven. Specifically, handling feedback was exhibited the most in students' reflective journals, and seven strategies (self-monitoring and evaluation, idea planning, perspective change, emotional control, effort regulation, peer learning, and resource management) showed significant increase over time, especially during the second half of the semester. In addition, we identified considerable variation in SRL use by writing-proficiency levels, with highly proficient students applying idea planning, feedback handling, and resource management and low-proficiency students applying goal-setting, significantly more often than the other group did. Furthermore, the qualitative results of the current study provide evidence about the potential effectiveness of using reflective-journal writing to promote self-regulated writing strategies.
Pedagogically, this work's most important implication is that teachers should consider using reflective journaling as a powerful tool for SRL development. They should purposefully design structures for reflective writing that cover the cognitive, metacognitive, motivational and behavioral aspects of their students' learning. Specifically, given the uneven distribution of SRL-strategy adoption, both across students and across contexts, it is necessary for teachers to insert appropriate guiding questions to scaffold students' reflection, rather than merely asking them to reflect. Second, reflective journaling should be clearly defined as a process rather than a product. That is, assigning journal writing at multiple time-points could help develop students' awareness of self-regulated writing strategies gradually and steadily. Also, these tasks should be assigned in reasonable quantities and at reasonable intervals, to avoid students becoming anxious about them, annoyed, or bored. And fourth, teachers should use reflective-journal writing as an opportunity to facilitate more individualized SRL guidance. For instance, when guiding low-proficiency students in this activity, they could offer explicit instruction in how any goals expressed should be specific, executable, and task-oriented instead of broad and generic.
This study has the following limitations. First, it only recruited English-major undergraduates from one top Chinese university, and the number of participants in the high-and low-proficiency groups was rather small; so, studying a larger number of participants from more diverse academic and ethnolinguistic backgrounds could yield different insights. Second, in the case of certain strategies such as peer learning and feedback handling, reflective journaling may not have been the only influential factor-especially due to the existence of peer-review sessions. Future studies could refine their research designs to allow clearer discernment of the reflective journals' effects. Third, in this study we compared the differences in the use of self-regulated writing strategies between high-and low-proficiency groups as betweensubjects variations. Future studies could consider testing betweensubjects variation at an individual level in order to better generalize strategy use across participants. Fourth, although the qualitative findings seem to suggest that journal writing may have contributed to an increase in self-regulated writing strategy use, a causal relationship cannot be established without the inclusion of a control group (students who did not participate in reflective journaling). Fifth, our three time-points for data collection may have been few and close together to reflect the full fluctuating nature of SRL strategy use over time that previous researchers have reported. Thus, data collection at a larger number of time points across a longer overall period could yield more accurate results regarding such change.
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Ethics statement
The studies involving human participants were reviewed and approved by Tsinghua University. The patients/participants provided their written informed consent to participate in this study.
Author contributions
YZ: conception, design, literature review, data analysis, and writing of the manuscript. ML: literature search, data collection and analysis. YC: literature search, data collection and analysis. SH: literature search and data analysis. All authors contributed to the article and approved the submitted version. | 2022-11-16T15:38:40.715Z | 2022-11-16T00:00:00.000 | {
"year": 2022,
"sha1": "8bb8fb05d89ab8605324235fe83f72c5db27ca8a",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Frontier",
"pdf_hash": "8bb8fb05d89ab8605324235fe83f72c5db27ca8a",
"s2fieldsofstudy": [
"Education"
],
"extfieldsofstudy": []
} |
238825656 | pes2o/s2orc | v3-fos-license | HEDYCHIUM SPICATUM BUCH-HAM. (KUCHRI), A TREASURE HOUSE OF ESSENTIAL OILS
Medicinal plants have a very significant role in the health care system. They are served as the primary source of modern drugs. One of such important medicinal plant is Hedychium spicatum Buch-ham. which belongs to the Zingiberaceae family (ginger family). The plant is commonly known as the spiked ginger lily in English and Kuchri in Hindi and Shati in Sanskrit. It is a commercially valuable plant due to its rhizomes. This rhizomatous plant holds a significant place in Ayurveda due to its extraordinary disease-curing properties. It is mentioned as Shwasahara mahakashaya dravya in Ayurveda. It is used in many folk cultures around the world as a remedy against many diseases like diarrhoea, liver-related problems, pain, vomiting, stomachache, inflammation, nausea, headache, fever etc. It is a therapeutically important plant due to the presence of numerous important essential oils as major phytochemical constituents like 1,8- Cineole, camphene, sabinene, β -p inene, myrcene, α -phellandrene, etc. The main therapeutic properties of the plant are anti-inflammatory, anti-microbial, hepatoprotective, tranquilizer, antipyretic, anti-diabetic, pediculicidal, anti-helminthic etc. The aim of the present review is to provide information related to phytochemistry, therapeutic properties, traditional uses of Hedychium spicatum in Ayurveda and folk medicinal system .
INTRODUCTION
Medicinal plants are the major segment of native traditional systems of medicine for years. Herbal products are used in almost each and every corner of the world (Intentional as well as domestic markets) under various categories such as herbal drugs, botanical drugs, botanicals, phytomedicines, traditional medicines, herbal medicines, traditional Chinese medicines (TCMs), traditional herbal medicinal products, natural health products, or plant food supplements, because of their remarkable multi-target therapeutic actions, safety and easily availability [1][2][3][4]. Medicinal herbs play a critical role in the manufacturing of new drugs because they contain phytochemicals of diverse nature which are medically important [5,6]. In the modern era, half of the drugs which are used clinically, have been developed primarily from plant sources [7]. There are numerous plants exist in mother nature which are extremely valuable as they are the rich source of anti-microbial, anti-inflammatory, anti-oxidant drugs. Hedychium spicatum Buch-ham. ( fig. 1) is one of such important medicinal plants which belongs to the zinger family Zingiberaceae. The plant is commonly known as spiked ginger lily, garland flower in English. Whereas it got its trade name Kapurakachari due to its commercially valuable rhizomes [8]. Locally the plant is known as ban haldi or kuchri. Zingiberaceae family is commonly known for its extraordinary therapeutic properties. This family is found growing wild in the tropical areas of the world like Southeast Asia. Gingers are served as an important ingredient in food, spices, medicines, dyes and perfume etc. There are approximately 53 genera having 1200 different variety of species present in Zingiberaceae family. 20 genera and almost 200 different species of Zingiberaceae family have been reported from various regions of India [9][10][11]. The genus name Hedychium is derived from 'hedys' a greek word which means sweet and 'chion' means snow [12]. It is a well-recognized genus for its specific aroma, beautiful foliage and attractive, diverse and showy nature of flowers and comprises of almost 50 different species [13]. Almost each species of this genus is native to central and Southeastern Asia and more specifically found in southern China and some Himalayan regions except H. peregrinum, which is a native plant of Madagascar [14]. Hedychium spicatum has a very rich history of its utilization in traditional medication systems and is commonly known as an anti-diabetic plant [15]. It has a significant position in World's most traditional medicine system i.e. Ayurveda. It is used in many Ayurvedic herbal formulations to treat a variety of ailments. The use of Hedychium spicatum rhizome has been mentioned in Ayurvedic literature for treating hair loss, digestion and problems related to the respiratory system, joint pain, and hiccups and to maintain cardiac health. In Ayurveda the plant rhizomes are usually consumed in powder, syrup or tablet form [16]. Hedychium spicatum is a rich source of various kind of essential oils which are responsible for its extraordinary therapeutic properties like anti-microbial, tranquilizing, anti-oxidant, anti-inflammatory, pediculicidal, hepatoprotective, antidiabetic and anti-helminthic etc. The essential oils are used in manufacturing of soap, hair oil, face powder and incense. Hedychium spicatum is utilized by many different cultures and tribes of the world as remedy to treat many diseases like asthma, piles, diarrhoea, liver-related problems, constipation, dysentery, stomachache, bronchitis, cough, headache, pain, inflammation and as an anti-venom against snakebites, skin disorders, dropsy and hair fall. Rhizomes are used as an appetizer, carminative and stimulant. Rhizomes are used for making well-known tonic as well as food supplement called Chayawanprash. The roots and leaves of this plant are used in Tibetan medicines. Apart from its clinical and therapeutic use, it is famous for its sweet-scented flowers which are used for ornamental purposes. The powdered form of dried rhizomes is used as herbal holi color (an Indian festival of colors). The rhizomes are used to provide aroma to tobacco. Due to the insecticidal properties of its foliage, it is being used to make floor mats [17][18][19][20][21][22][23][24][25][26]. As per the International Union for Conservation of Nature and Natural resources (IUCN) report, H. spicatum has now listed under vulnerable species and near threatened plant species containing essential oils. The main reason behind it overexploitation and habitat degradation of the plant [27][28][29]. Vernacular names and taxonomic classification of Hedychium spicatum is given in table 1 and 2 respectively. International Journal of Current Pharmaceutical Research reddish brown layer is present over the one edge of each piece having numerous scars and circular rings and rudiments of rootlets [32][33][34].
Geographical distribution of Hedychium spicatum
H. spicatum is a native plant of south-eastern Asian countries. It is found more commonly in the subtropical Himalayas. In India it is primarily found in Assam, Arunachal Pradesh and Uttarakhand at an elevation of 1800-3000 m. Whereas it is also found in Jammu and Kashmir, Thiruvananthapuram hills of Kerala. It is also found widely growing in the Malysia, Nepal, Bhutan, Myanmar, Northern Thailand and China [35,36]. The plant impressively tolerates wide range of climatic conditions of forest margins from 1500 to 2800m asl. H. spicatum grows well in light (sandy), medium (loamy) and heavy (clay) soils. It prefers acid, neutral and basic (alkaline) soils [37].
Ayurvedic view of Hedychium spicatum
The origin of Ayurveda is almost 2500 and 500 BC old. The word Ayurveda depicts the meaning "science of life". Its treats the diseases by balancing the three body components/doshas of the body i.e. kapha (water and earth), pitta (fire) and vata (space and air) [49][50][51]. Hedychium spicatum is a popular medicinal plant Ayurveda science of life. It is commonly known as Shati in Ayurveda. It consists of Kaphavataghna properties i.e. it balances the kapha and vata Doshas. In Charak Samhita, Hedychium spicatum is mentioned under Shwasahara mahakashaya dravya i.e. drugs having the potential to treat cough and other problems related to the respiratory system. It is mainly used to treat cough, wound ulcer, fever, respiratory problems and hiccough, respiratory tract issues, dermatological diseases, and tropical pulmonary eosinophilia [52][53][54][55]. Rasapanchak of Hedychium spicatum is given in table 3.
Ayurvedic formulations of Hedychium spicatum [60, 61]
The most famous Ayurvedic formulation of of Hedychium spicatum is "Bharangyadi" is well known polyherbal Ayurvedic formulation which is made up of Clerodendrum serratum, Hedychium spicatum and Inula racemosa. In Ayurveda, it is used for treating allergic rhinitis and allergic asthma. In vitro analysis of this polyherbal formulation done by Kajaria D et al., suggested its use as a significant immunomodulatory as well as an anti-microbial and anti-inflammatory drug.
Folk view of Hedychium spicatum
Some plants are used in many medicinal folk practices which represent their rich ethnobotanical significance [62]. Ethnobotany not only promotes the value of cultural beliefs but also plays a vital role in the health care system and drug development [63].
Hedychium spicatum is associated with many ethnobotanical uses. For instance, the root powder of this plant is used traditionally as a remedy to asthma [64]. Kurumba tribal people of Chemmankarai, Nilgiri use leaves to get relief from headache. They use preparation made up of rhizome powder and goat milk as a cure to asthma [65]. In Western Mizoram, rhizome is traditionally used for many medicinal purposes like, liver-related problems, pain, vomiting, stomachache, inflammation and snakebite. The rhizome is used as an expectorant, tonic, carminative and stimulant [66]. Some local communities of Rudraprayag District, Uttarakhand apply the root paste topically in swelling [67]. In the Jaunsar-Bawar Hills of Uttra Pradesh, people treat diarrhoea with the paste made up of Hedychium spicatum rhizome [68]. In Darjeeling and Sikkim, rhizome is used against stomach and liver-related issues and vomiting whereas in some other regions it is used to cure asthma, cough, diarrhoea, vomiting and headache [69,70]. People from Dhwaj sacred grove from the Central region of Indian Himalayas traditionally treat cold, cough, tonic, asthma, gastric problem, liver diseases, fever with the Hedychium spicatum rhizome [71]. Traditional healers of Bhubaneshwar, Odisha use rhizome powder of H. spicatum to treat bronchial asthma, also used as an appetite enhancer. They also treat local inflammation, nausea, hiccups, halitosis and vomiting with the plant rhizome [72]. Tribes of Koraput Odisha, use tuber paste externally as a remedy for rheumatism and loose motion [73]. People use rhizome of this plant in folk medicinal practices to treat asthma, piles, bronchitis, snake bite and nausea in the Bageshwar valley of Uttarakhand [74]. In Marginal hill community, Uttrarakhand, people use Hedychium spicatum rhizome in gastrointestinal (problems related to the intestine and used as a purgative, laxative and carminative) respiratory (it is used to treat cough), dermatological problems (cosmetics and used against lice) [75]. In Kumaun, Himalaya, people use root against problems related to liver, fever, vomiting, diarrhoea, inflammation, pains [76]. Traditional healers of Nanda Devi Biosphere Reserve, Uttarakhand use this plant against asthma, piles and liver related problems [77]. People in Siwalik region of Uttarakhand treats dysentery traditionally with the plant rhizome [78]. Women from Garhwali region use rhizome as a blood purifier and in treating rheumatic pain [79]. In some areas of Uttarakhand root powder is used to treat neuromuscular disorders and body pain, asthma and also used as an anti-cancer and antimicrobial [80,81]. The Migratory Shepherds in Summer Hill of District Shimla use the plant rhizome against asthma and cough [82]. In some areas of kullu, rhizomes are used against asthma, bronchitis and as a blood purifier [83]. The Kanwar tribe of Chhattisgarh, use Hedychium spicatum tuber in sexual enforcement [84]. In many districts of Nagaland, plant is used to treat diseases like fever, headache, vomiting, diarrhoea, inflammation [85]. Some indigenous groups of Northern Chin State, Myanmar use rhizome of the plant to get relief from menstrual bleeding, wound bleeding, asthma, and kidney problems. They also use it as a tonic [86]. Some ethnic groups of Parbat district, Western Nepal use the rhizome decoction against indigestion and high fever [87]. A Nepalese tribe named Raji, use plant rhizome to enhance digestion, loss of appetite and to treat constipation, stomachache [88]. Some native groups of Lore Lindu National Park, Central Sulawesi, Indonesia, use the rhizome in cosmetics and as a spice [89]. The rhizome paste of this plant is used around many parts of the world as a traditional remedy for abscesses (dermatological disorder) [90,91].
Modern view of Hedychium spicatum
Herbal medicinal products are preferred more because these have negligible adverse impacts and toxicity associated with them. But concerns regarding the safety of these products are raising due to the practice of adulteration and contamination [92-96,]. Adulteration always leads to the degradation of herbal products which may cause severe health risks [97]. Common adulterants of herbal products are orthodox drugs and fake or inferior plant materials and foreign materials. Species misidentification in the Global market of herbal products is a very common type of adulteration. Misidentification of plant species induces undesirable, unrelated species which may have the potential to cause severe impacts on consumer health. This can be intentional or unintentional. Intentional adulteration is carried out to derive maximum profit by cleverly increasing the weight or quantity of the herbal product [98,99]. For instance, Kaempferia galanga Linn is a plant species whose rhizomes are sold out in the market as the same name as that of Hedychium spicatum i. e Shati. This may happen due to morphological similarity [100]. Improper packaging and storage are also associated with contamination [101]. The quality of herbal products can only be assured if proper standardization tools and techniques are set up from the beginning of the process to the ending i.e. from the collection of raw material to the production of end product [102]. For the detection of adulteration, the most preferred technique which is used nowadays is DNA barcoding. Its molecular-based technique which has a great potential to identify the species and detect out the adulterant and contaminants in the herbal products [103].
Therapeutic properties of Hedychium Spicatum
Hedychium spicatum has extraordinary therapeutic properties due to its wide range of phytochemical constituents. Some of the therapeutic properties of Hedychium spicatum are discussed below.
Anti-inflammatory activity
Chachad et al., conducted a comparative study to evaluate the antiinflammatory potential of Hedychium spicatum Buch-Ham ex Smith, Kaempferia galanga Linn. and Curcuma zedoaria Rosc. The study was conducted out on carrageenan-induced rat paw edema models. The ethanolic extract of Hedychium spicatum rhizome showed the maximum anti-inflammatory behavior. The study confirms that the plant is associated with anti-inflammatory activity [104].
Hepatoprotective activity
Hepatoprotective activity of Hedychium spicatum was studied by Joshi Uttara in an in vitro study. The study was conducted on primary rat hepatocytes in which hepatotoxicity was induced by paracetamol. Dietrpene isolated from methanol extracts significantly worked on hepatocytes viability and other associated parameters like glutamic transaminase, glutamic pyruvic transaminase and total protein by restoring them. The study suggested that Hedychium spicatum has potent hepatoprotective activity [108].
Anti-diabetic
Kaur et al., conducted an in vivo study to evaluate the anti-diabetic potential of Hedychium spicatum in rat models in which diabetes was induced artificially by intraperitoneal injecting alloxan monohydrate solution. The treatment of models by orally administering rhizome essential oil at 0.3 Ml dose for 14 d, resulted in reduced levels of blood glucose and urea. It was also observed that essential oil helped Islets of Langerhans in regaining the original shape [109].
Tranquilizing activity
Chopra et al., evaluated the tranquilizing activity of Hedychium spicatum in rat models and the results showed that rhizome essential oil is associated with the tranquilizing activity. It significantly worked on avoidance response caused by the depressed condition, performance of rota-rod and hypnosis induced by phenobarbitone and analgesia induced by morphine in rats [110].
Ameliorative efficacy
Chaudhary et al., conducted an in vivo study on eight weeks old, white leghorn male chicks which were intoxicated with indoxacarb to evaluate the ameliorative actions of Hedychium spicatum. Root powder of the plant showed its ameliorative actions by normalizing the levels of total erythrocyte count (TEC), total leukocyte count (TLC), haemoglobin (Hb), packed cell volume (PCV), absolute lymphocyte count (ALC), absolute heterophill count (AHC) and lymphocyte and by reducing the value of differential leucocyte count (DLC %) to normal. The study concluded that the plant has ameliorative actions [111].
Pediculicidal
Jhadav et al., studied the pediculicidal activity of Hedychium spicatum in an in vitro study. The study was conducted out on Pediculus humanus capitis (head louse). It was observed that essential oil extracted from rhizomes of the plant at the concentration of 5%, 2% and 1% caused 95-100% mortality rate in the models after the interval of 2 h of treatment with the extract. The study confirms that the essential oil showed better results than the 1% permethrin-based products available in the market [112].
Anti-helminthic
Sravani et al., analyzed the anti-helminthic potential of Hedychium spicatum in in vitro using adult Indian earthworms named Pheretima posthuma and in silico study. It was found from the results of in vitro study that beta-sitosterol extracted from the rhizomes showed potent inhibitory actions than the standard drug called Piperazine citrate. The time taken by beta-sitosterol for paralysis and death was less than that of Piperazine citrate. Whereas no significant findings were obtained from the in silico study [113].
Anti-pyretic
Arora et al., investigated the antipyretic activity of Hedychium spicatum in an in vivo study on Wister strain albino rat models. The models were induced with pyrexia artificially by Brewer's yeast. Methanolic extract of the plant was found to be an effective measure against pyrexia as it significantly elevated the body temperature of models at the dosage of 100 mg/kg and 200 mg/kg. Thus, the study suggested the use of Hedychium spicatum as an antipyretic agent [114].
CONCLUSION
The present study is an attempt to provide detailed information about an important medicinal plant species Hedychium spicatum. Modern pharmacological studies indicated that this plant has extraordinary biological potential. It is strongly believed that the data presented in this review on the utilization of Hedychium spicatum plant in Ayurveda and folk cultures might draw the attention of researchers to use this plant in modern medicines. The diverse kind of phytochemicals present in this plant can be the promising source of anti-microbial, hepatoprotective, immunomodulatory, pediculicidal drugs. | 2021-09-28T16:31:51.268Z | 2021-07-15T00:00:00.000 | {
"year": 2021,
"sha1": "226e405897afb2180cca2dd988916f53309717ed",
"oa_license": "CCBYNC",
"oa_url": "https://innovareacademics.in/journals/index.php/ijcpr/article/download/42738/25135",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "a12112ed0545bb251b07470b609327e0b707ceb4",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Art"
]
} |
238407964 | pes2o/s2orc | v3-fos-license | Complexity of Traveling Tournament Problem with Trip Length More Than Three
The Traveling Tournament Problem is a sports-scheduling problem where the goal is to minimize the total travel distance of teams playing a double round-robin tournament. The constraint 'k' is an imposed upper bound on the number of consecutive home or away matches. It is known that TTP is NP-Hard for k=3 as well as k=infinity. In this work, the general case has been settled by proving that TTP-k is NP-Complete for any fixed k>3.
Introduction
Sports tournaments are very popular events all over the world. A huge amount of money is involved in organizing these tournaments and also a lots of revenue is generated by selling the tickets and broadcasting rights. Scheduling the matches is a very important aspect of these tournaments. Scheduling specifies the sequence of matches played by each participating team along with the venues. In a Double Round-robin Tournament, every pair of participating teams play exactly two matches between them once in both of their home venues. The Traveling Tournament Problem (TTP) asks for a double round robin schedule minimizing the total travel distance of the participating teams. The fairness condition imposes an upper bound k to the maximum number of consecutive home or away matches played by any team. TTP was first introduced by Easton, Nemhauser, and Trick [1]. The problem bears some similarity with Traveling Salesman Problem (TSP). In fact, a reduction of the unconstrained version of TTP (or k = ∞) from TSP has been shown by Bhattacharyya [2] proving the basic problem to be NP hard. When the maximum permissible length consecutive home or away matches is set to 3, TTP is proven to be NP-Hard as well by Thielen and Westphal [3]. In this work, the natural question has been asked, Is TTP NP-Hard for any fixed k > 3?
Problem Definition
Let T be the set of teams with |T | = n. Let ∆ be a square matrix of dimension n × n whose element at i th row and j th column corresponds to the distance between the home venues of i th and j th team in T for all i, j ≤ |T |. ∆ is symmetric with diagonal terms 0 and all the distances in ∆ satisfy triangle inequality. Definition 1.1 Decision Version of TTP-k: For a fixed a natural number k, an even integer n, a given set of teams T with |T | = n, mutual distances matrix (∆) and a rational number δ, is it possible to schedule a double round-robin tournament such that the total travel distance of the tournament is less or equal to δ, where no team can have more than k consecutive away matches or consecutive home matches and no team plays its consecutive matches with same opponent.
Previous Work
The Traveling Tournament Problem was first introduced by Easton, Nemhauser, and Trick [1]. Since then most of works focused on finding approximation algorithms or heuristics when k = 3 [4,5,6,7,8]. Thielen and Westphal [3] proved NP hardness for TTP-3. In a series of two papers [9, 10] Imahori, Matsui, and Miyahiro showed approximation algorithms for the unconstrained TTP. Bhattacharyya [2] complemented this by proving the NP-hardness of unconstrained TTP. For other values of k, only upper bound results are known. Approximation for TTP-2 is done by Thielen and Westphal [11] and improved by Xiao and Kou [12]. For TTP-k with k > 3, approximation algorithms are given by Yamaguchi et al. [13] and Westphal and Noparlik [14].
Many different problems related to sports scheduling have been thoroughly analyzed in the past. For detail surveys and study on round-robin tournament scheduling, the readers are referred to [15,16,17]. Graph Theoretic approach for solving scheduling problems can be found in [18,19].
Approach towards the Problem
In this work, generalization of the approach by Thielen and Westphal [3] has been done, who showed the NP Hardness of TTP-3. Like them, a reduction from the satisfiability problem has been shown. While [3] showed reduction from 3-SAT, a reduction from k-SAT is shown here. However the reduction shown here is different in few crucial aspects. Firstly, the construction of the reduced TTP instance graph is different from that in [3]. In order to accommodate trips of length k, a new graph in terms of vertices and edge weights is required. In addition, the trip structures and their synchronous assembly to get the tournament schedule is different from that of TTP-3. The reconstruction of k-SAT with specific properties of clauses require different technique.
Result
Here the main theorem is the following. Theorem 1.2 TTP-k for a fixed k > 3 is strongly NP-Complete.
Proof of Theorem 1.2
This reduction requires that the input instance of the satisfiability problem satisfies certain properties. The first step is to show that any input instance can be transformed into an instance with properties required for the reduction. Following notations are used throughout the paper. If x ∈ {0, 1} is a boolean variable,x denotes its complement,x = 1 ⊕ x. Lemma 2.1 For a k-SAT instance F k with t variables and p clauses there exists another k-SAT instance F k with t variable and p clauses such that a satisfying assignment of F k implies a satisfying assignment of the variables of F k and vice-versa. Moreover, the number of occurrence of all the t variables and there compliments are equal in F k , t ≤ t + k+1 2 and k | p with k, t, t , p, p ∈ N.
Proof.
Let x i andx i (i ∈ {1, . . . , t}) be the variables in F k . Suppose n i andn i are the number of occurrence of x i andx i in F k respectively. Without loss of generality it can be assume that n i ≤n i . To make n i =n i , few clauses has been added to F k depending on the value of k is even or odd.
2. When k is even, then using Handshaking Lemma it can be said that there exist even number of indices j ∈ {1, . . . , t} such that (n j +n j ) is odd. So, n j −n j is odd. Without loss of generality, it can be assumed thatn j > n j . By identifying two indices of this kind, namely m and n, clauses of the form (x m ∨ x n ∨ x t+1 ∨x t+1 ∨ · · · ∨ x t+ k 2 −1 ∨x t+ k 2 −1 ) are added until n i −n i is even ∀i ∈ {1, . . . , t} by introducing (k/2 − 1) new variables. Now ∀j ∈ {1, . . . , t} such that n j =n j , pair of clauses of the forms ( ) are added until n j =n j ∀j ∈ {1, . . . , t}. Then by adding at most , the number of clauses can be made divisible by k Let the resulting k-SAT problem expression be F k with t variables and p clauses where, k|p . In both the cases explained above, all the additional clauses always give truth values. So, F k will have a truth assignment of variables x i for all i ∈ {1, . . . , t } if and only if F k have a truth assignment of variables x i ∀i ∈ {1, . . . , t}. This proves the lemma.
TTP-k is NP-Complete
The first step is to show that TTP-k is indeed in NP.
Lemma 2.2 TTP-k is in NP.
Proof. In a decision version of TTP-k with given set T of n teams, and the constraint k it is verifiable in O(n 2 ) whether a schedule is valid and gives a total travel distance less than δ or not. This ensures the membership of TTP-k in NP.
Next, a reduction from k-SAT to TTP-k has been shown. For this, a special weighted graph G = (V, E) has been constructed with one or more teams situated at each vertex of G and a predefined value δ of total travel distance such that there is a satisfying assignment of variables for k-SAT if and only if there is a TTPk schedule between the teams in G with total travel distance less than δ. First the input k-SAT problem is modified using Lemma 2.1 such that the resulting formula has the following properties. (c) Number of clauses p is divisible by k, k|p.
The Construction
We start with the construction of the reduced instance graph G. Recall that k is the upper bound of number of consecutive home or away matches, and n i is the number of occurrence of the variable x i . The main part of the graph is the union of t many sub-graphs G 1 , G 2 , · · · , G t where t is the number of variables in the (modified) input SAT instance. Each G i consists of (k + 1)n i vertices.
(a) n i many vertices are denoted by x i,j and n i many vertices are denoted byx i,j where j ∈ {1, . . . , n i }.
In addition, there are (k − 1)p + 1 many vertices in the graph where p is the number of clauses of the (modified) input SAT instance. There is a central vertex v. Then there are p vertices, C m and (k − 2)p vertices, z l m ∀l ∈ {1, . . . , k − 2} and ∀m ∈ {1, . . . , p}. For ease of explanation, we summarize the important parameters related to G. The total number of vertices in G is where each of these consecutive vertices in this serial connection is connected with each other with an edge of weight 2. Thenx i,j is connected to w 1 i,q with an edge of weight 2 where q = j(mod n i ) + 1 and also y i,j is connected to both x i,j andx i,j with edges of weight 2 each as described in Figure 1.
For the connection between the remaining vertices in G, first x i,j orx i,j are connected to c m with an edge of weight 2 if x i orx i,j has its j th occurrence in the m th clause of the modified k-SAT. Then c m is connected with z k−2 m through k − 3 vertices serially namely z 1 m , z 2 m , . . . , z k−3 m where each of these consecutive vertices in this serial connection is connected with each other with an edge of weight 2. At last z k−2 m and c m is connected to v by edges of weights (M − 2k + 6) and (M + 2) respectively as described in Figure 2.
Formally, weights of all the edges of G are listed as follows: • Weight(x i,j , c m )=2, if j th occurrence of x i is present in m th clause of the given k-SAT expression.
• Weight(x i,j , c m )=2, if j th occurrence ofx i is present in m th clause of the given k-SAT expression.
All other edges in the complete graph G is given maximum possible weights without violating triangle inequality. CREATING TTP-k INSTANCE. Now the teams are placed on the vertices of G to construct the reduced instance.
• Total number of teams is equal to a 3 + a.
• At each vertex of G except v, only one team is placed. This set of vertices or teams is denoted as U .
• a 3 teams are situated at v and distance between them is 0 and call this set of vertices or teams T .
We fix, Weights of the edges of G preserve the triangle Inequality.
Proof. For a tuple (v, x i,j , y i,j ) or (v,x i,j , y i,j ) for all i ∈ {1, . . . , t} and j ∈ {1, . . . , n i }, the triangle inequality is preserved as, For a tuple (v, x i,j , w 1 i,j ) or (v,x i,j , w 1 i,j ) for all i ∈ {1, . . . , t} and j ∈ {1, . . . , n i }, the triangle inequality is preserved as, For a tuple (v, x i,j , c m ) or (v,x i,j , c m ) where j th occurrence of x i orx i is present in m th clause of the given k-SAT expression, the triangle inequality is preserved as, The triangle inequality for all other tuple of three vertices in G is followed from these above three cases as the weights are given maximum possible values without violating triangle inequality while assigned.
The Reduction
As a desired value of δ has been got, now the only remaining part is the reduction. First, it has been shown that a given satisfying assignment of variables in a k-SAT ⇒ a TTP-k schedule of total travel distance less than δ. First, the tours for the a 3 vertices situated at v are constructed and also showed that, these tours are so cheap in terms of travel distance that tours of similar structure must be there in a tournament where the total travel distance is desired to be as low as possible. So, a k node disjoint tours are constructed for a team at v where, all the vertices x i,j ,x i,j , y i,j , w r i,j , c m , z r m are visited. First, observe that as there is a satisfying assignment of variables in the k-SAT. Let us define two conditions denoting with the value of a variable b m in the following manner, to z k−2 m and return to v, it travels the same distance, 2(M + 2), and also visits k vertices in a single trip which is a desired situation here. Afterwards in another trip to a vertexx i,j , in spite of directly traveling to and fro tox i,j only, i.e. 2M , if it travels to w k−2 i,q first then through all w s i,q for q = j(mod n i ) + 1 and s ∈ {1, . . . , k − 3} tox i,j and y i,j and returns back to v, the travel distance is same as 2M and also k vertices will be visited in a single trip. Multiple trip to these extra (k − 1) vertices would cost much more in comparison. Similar tours are taken when b m = 0 only interchanging x i,j and y i,j withx i,q and y i,q respectively. This leaves (k − 2)p/2 number of x i,j type vertices to visit. As k|p, this can be done in p(k − 2)/2k trips each of length k and travel distance less than 2(M + k(k − 1)). So the total distance traveled by a team situated at v is upper bounded by δ 1 . Where, With tours involving distance M is minimized by covering exactly k vertices in each of the tours. Now these tours can be numbered from 1 to a k and vertices in each tour can be numbered from 1 to k. So, all the a vertices of U are named as u i,j for all i ∈ {1, . . . , k} and j ∈ {1, . . . , a k } such that u i,j is the i th visited team of j th tour. Also the vertices in T are partitioned in a 2 disjoint sets T 1 , T 2 , . . . , T a 2 each of size a. Moreover T q = {t r,q such that r ∈ {1, . . . , a}}. The tours by the teams in T are now designed in such a way that, t 1,1 will take tour number 1, i.e. travel through u 1,1 , u 2,1 , . . . , u k,1 . Then u 1,1 , u 2,1 , . . . , u k,1 visit t 1,1 in the same order. Similarly, for all i ∈ {2, . . . , k}, t i,1 follows the same tour and visited by the same teams as t 1,1 but with a time delay of (i − 1). This way all the teams in T 1 first complete visits to all the teams in tour 1 and then get visited by them. Then T 1 starts tour 2. This way teams in T 1 plays with the teams in tour 1 to tour a 2k in such a way that the teams in T 1 visit first and then they get visited. Similarly, teams in T q for all q ∈ {1, . . . , a 2 2 } follow a similar travel like T 1 . The matches between teams in T q for all q ∈ { a 2 2 + 1, . . . , a 2 } and the teams in U that are in tour j, for all j ∈ { a 2k + 1, . . . , a k } are also played in a similar fashion with the change that the teams in U visit the teams in T first and then they get visited by the teams in T . This way the sets T and U both are divided in two parts according to the teams they completed playing. Formally, Afterwards by changing the roles of T a and T b , matches between teams in T a and U b are arranged and similarly between teams in T b and U a . So, the main remaining part is schedule of matches between the teams in U and that of the teams in T . For scheduling matches between teams in U , the teams in U are categorized in k categories depending on their occurrence in the tours, i.e., for all i ∈ {1, . . . , k}, U i = {u i,j such that j ∈ {1, . . . , a k }}. For all i, the teams in U i play against the teams in T at the same slots but half of them play at home and the other half play away. More specifically the teams in U i play with the teams in T at exactly on the slots (2i − 1) to (2a 3 + 2i − 2). Keeping these busy slots in mind, schedule of a single round-robin tournament of the teams in U is designed using canonical tournament introduced by de Werra [19]. This is done by assigning a vertex to each of the teams in a U i in a special graph as done in the canonical tournament design. This tournament structure gives assurance that each team plays every other team exactly once and no team has a long sequence of home and away matches. Here a match between i and j signifies a match between a team in U i and a team in U j . At the end the same tournament is repeated with changed match venues, i.e. nullification of home field advantage. Now the only remaining part is scheduling of the matches between the teams in T . Let for all t ∈ T, d(t) denote the first slot on which team t plays a team in U and let T i = {t ∈ T such that ((d(t) − 1)mod k) + 1 = i} for all i ∈ {1, . . . , k}. Then every T i is partitioned into 2a 2 groups of cardinality a 2k such that d(t 1 ) = d(t 2 ) for every two members t 1 , t 2 of the same group. For every T i , the matches between teams in different groups in T i are scheduled. Among the teams in T i , a k will always be busy playing with some teams of U . To handle this fact, two dummy groups U 1 and U 2 , as defined before, are introduced that play with these busy a k teams of T i . Then each group is treated as a team and again canonical tournament structure is applied only skipping the day at which the two dummy groups meet.
For scheduling the matches between the members of two groups g and h of T i in l rounds, where The following is done. The i th round contains the matches between h j and g ((i+j) mod l)+1 for all i, j ∈ {1, 2, . . . , l} with game taking place at the home of h j . This restricts the lengths of away trips and home stands for all the teams from being long. Then the same schedule is repeated with altered venues. Matches between the teams of some group of T i with the teams of a dummy group has already been taken care when the matches between U and T were designed. Now, for the scenario where the two dummy groups meet, two kind of matches are there which differ in length. The encounter between two groups consists of a k slots, while the encounter between two dummy groups, i.e. U 1 an U 2 consists of a slots. The encounters between the groups of T i are scheduled in usual way using a k slots and the extra (k−1)a k slots are used to schedule matches between the teams in different T i 's.
To schedule matches between the teams in different T i 's, first each T i is partitioned in two equal size, namely T i,1 and T i,2 . Now, considering each T i,j as a single team for all i ∈ {1, 2, . . . , k} and all j ∈ {1, 2}, a canonical tournament is again applied on these teams skipping the day at which T i,1 encounters T i,2 for all i. These scenario can be achieved by properly initializing the canonical tournament. Then the same schedule is repeated with altered venues to nullify home advantage as done in all earlier canonical tournament structures.
For a team situated in one of the vertex in E \ v to visit the teams in v, it has to travel at most 2(M +2)a 3 k using a 3 k trips of length k. So, the total travel distance of all the teams in E \ {v} to the teams in v can be bounded by δ 2 . Where, δ 2 = 2(M + 2)a 4 k As all the distances between the vertices in E \ {v} in G is less or equal to 2k, the travel between all the teams in E \ {v} can be bounded above by δ 3 . Where, The teams in T visit each other at zero cost as they are situated at the same point. So, the total travel distance of the tournament is bounded by, In another way, This completes the first direction of the proof.
Proof of the other direction
For the other direction of the proof, it has to be shown that k-SAT is not satisfiable ⇒ a TTP-k schedule of total travel distance less than δ is not possible. Here it is assumed that there is no satisfying assignment of variables of the k-SAT. In the forward direction of the proof explained above, it is shown that the proposed schedule is compact and gives an optimized value of travel distance for the tours of the teams in T to the teams in U . But the travels are designed depending on a truth assignment of x variables. The travel to c m vertex goes through a x i,j orx i,j which is there in the m th clause of the k-SAT and assigned with value 1. Also the other variable among these two covers y i,j and w s i,j s together in another tour. Let assume that, there exist optimum tours similar to the forward direction of the proof although there is no satisfying assignment of variables of the k-SAT. So, there is an optimum path through each c m for all m ∈ {1, . . . , p} that includes a vertex x i,j orx i,j which is present in the m th clause of the k-SAT expression. Now if value 1 is assigned to each of these variables then that must end in a wrong assignment of variables as it contradicts the assumption otherwise. This imply that x i,j =x i,j = 1 has been assigned for some i ∈ {1, . . . , t} and j ∈ {1, . . . , n i }. That means both x i,j andx i,j are included in optimized tours from v to c i and c j , where i, j ∈ {1, . . . , p}. Now, it will not be possible to design a trips from v, that includes y i,j , w 1 i,j , . . . , w k−2 i,j for all i ∈ {1, . . . , k} and j ∈ {1, . . . , a k } using the optimized trips. So, it is not possible to cover all the c m , y i,j , w s i,j s along with all x i,j s for all i, j using these optimized tours. To cover all the vertices in U each of the a 3 vertices of T has to tour at least once more to the vertices of U . This extra tours will increase the total travel distance by at least 2 · M · a 3 . As the total travel distance by the teams in U for matches among them is O(a 2 ), so the tours of the teams of T to those of U dictates the total travel distance of the tournament. So, increase in this part will increase the total distance significantly and for M being θ(a 5 ), the total travel distance will be more than δ. This completes the other direction of the proof and the reduction. | 2021-10-07T01:16:04.228Z | 2021-10-05T00:00:00.000 | {
"year": 2021,
"sha1": "d64776a2fb81c6d3b8704c1e997ab44a4295c58f",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "d64776a2fb81c6d3b8704c1e997ab44a4295c58f",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
4562976 | pes2o/s2orc | v3-fos-license | Mesenchymal Stem Cells As A Therapeutic Option For Patients With ALS
Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease with poorly understood pathophysiology. Death of upper and lower motor neuron results in weakness and paralysis of skeletal muscles including respiratory musculature. Disease typically begins with muscle weakness and atrophy or progressive bulbar palsy and leads to respiratory distress and death about 5 years from the diagnosis. About 1-2 new cases per 100000 per year are diagnosed, about 10% is familial and connected with superoxide dismutase 1 (SOD1) mutation.
Background
Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease with poorly understood pathophysiology. Death of upper and lower motor neuron results in weakness and paralysis of skeletal muscles including respiratory musculature. Disease typically begins with muscle weakness and atrophy or progressive bulbar palsy and leads to respiratory distress and death about 5 years from the diagnosis. About 1-2 new cases per 100000 per year are diagnosed, about 10% is familial and connected with superoxide dismutase 1 (SOD1) mutation.
Despite scientific development, there is no successful treatment for ALS, riluzol only prolongs survival for about 10-20% [1]. Stem cell treatment might be a disease-modifying strategy for ALS patients. Mesenchymal stem cells (MSC) have been defined by International Society for Cellular Therapy, they are known to be weakly or nonimmunogenic and thus applicable in an allogeneic setting [2]. Wharton's jelly-derived MSC (WJ-MSC) seem to be a preferable source because of easiness and safety of the harvesting procedure as well as a rich number of cells contained in the umbilical cord. WJ-MSC have high immunomodulatory capabilities as well as good proliferation and differentiation potential. Their characteristics close to embryo-derived stem cells but with lower risk of potential tumorigenesis [3].
Mechanism of action
The mechanism behind ALS are not entirely known. Studies on animal models have been undertaken to deepen understanding of the disease and explore potential mechanisms of treatment with MSC. The hypothesis include indirect effects such as the delivery of growth factors in situ, modulation of glial cells secretome, up-regulating regulatory T-cell activity. Preclinical studies showed that MSC express 12 neural genes and 11 transcription factors, and are able to differentiate into neural and glial cells in an appropriate media [4].
Sun et al.
[5] study on ALS animal model proved that MSC plays a role in neuroprotection of motor neurons, astrocytes, microglia, and modulates their response to apoptosis. Modulation of response to inflammation by down-regulating expression of TNFα and IL-6 and iNOS was observed as well. According to Uccelli et al.
[6] study, paracrine mechanisms are more probable than local production of growth factors since MSC intravenous infusion into mouse ALS model lead to amelioration of motor neuron function, slowdown of weight loss and survival improvement, despite low engraftment rates. The results of co-culture of MSC with peripheral blood mononuclear cells Abstract It has been postulated that mesenchymal stem cells could play a role in treatment of patients with amyotrophic lateral sclerosis. Variety of mechanism behind MSC action has been proposed including immunomodulation and delivery of trophic factors. The scientific rationale in preclinical studies is being proved along with first clinical uses.
Preclinical studies
It has been confirmed during ex vivo experiment that WJ-MSC maintain their immunomodulatory properties upon a long-term culture and preserve genetic stability through expansion up to 15 passages without chromosomal changes and malignant transformation both in vitro and after MSC injection into nude mice with follow-up of 4 weeks Another study administered autologous mesenchymal stem cells intravenously and intrathecally. Stem cells infused through lumbar puncture were committed to neuronal differentiation. Disease progression was observed in study and control group, however its dynamics was significantly slower in patients who received stem cell therapy [19].
Oh et al. [20] conducted an open label phase I clinical trial with intrathecal injections of autologous BM-MSC. Seven out of 8 patients received 2 doses with time interval of about 1 month. No serious adverse effects were noted within 12 months of follow-up. The course of the disease seemed to be slower after therapy in comparison to previous decline. A retrospective analysis of 57 ALS patients, including 20 controls and 37 patients treated with autologous bone marrow mononuclear cells, showed survival benefit of stem cell therapy [21]. Another approach was demonstrated by Deda et al. [22] and Blanquer et al. [23]. The first above-mentioned study delivered bone marrowderived hematopoietic progenitor cells were to 13 patients via total laminectomy and under general anesthesia [22]. The second one administered neurosurgically BM-MNC to patients with bulbar-onset ALS and noted slower disease progression [23].
Summary
Application of stem cells as an actual therapy for ALS patients is still in debate. It is a promising method, however current results are difficult to evaluate due to vast heterogeneity of patients and therapeutic schemes as well as lack of long-term follow-up. A growing number of clinical trials undertake this issue to assess answers about MSC efficacy | 2018-05-08T18:09:27.565Z | 2016-04-27T00:00:00.000 | {
"year": 2016,
"sha1": "6e33b51abbfe2780a7614cf18b49053d4fb19b72",
"oa_license": "CCBY",
"oa_url": "https://www.omicsonline.org/open-access/mesenchymal-stem-cells-as-a-therapeutic-option-for-patients-with-als-2327-5146-1000235.pdf",
"oa_status": "HYBRID",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "6e33b51abbfe2780a7614cf18b49053d4fb19b72",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
6515972 | pes2o/s2orc | v3-fos-license | Akt Deficiency Attenuates Muscle Size and Function but Not the Response to ActRIIB Inhibition
Background Akt is a critical mediator of developmental skeletal muscle growth. Treatment with a soluble ActRIIB fusion protein (ActRIIB-mFc) increases skeletal muscle mass and strength by inhibiting myostatin and related peptides. Recent in vitro studies have suggested that Akt signaling is necessary for the ability of ActRIIB inhibition to induce muscle hypertrophy. Thus, we hypothesized that mice deficient in either Akt1 or Akt2 would not respond to in vivo inhibition of ActRIIB with ActRIIB-mFc treatment. Methodology and Principal Findings We analyzed body composition and muscle parameters in wild-type C57BL/6J and Akt1 and Akt2 knockout mice, and compared the responses to blockade of ActRIIB signaling via ActRIIB-mFc treatment. Mice lacking Akt1 or Akt2 had reduced muscle mass, grip strength and contractile force. However, deficiency of Akt1 or Akt2 did not prevent the ability of ActRIIB-mFc treatment to induce muscle hypertrophy, or increase grip strength and contractile force. Akt1 and Akt2 deficient mice responded similarly as wild type mice to ActRIIB-mFc treatment by increasing fiber size. Conclusions and Significance Akt1 and Akt2 are important for the regulation of skeletal muscle mass and function. However, these Akt isoforms are not essential for the ability of ActRIIB inhibition to regulate muscle size, fiber type, strength or contractile force.
Introduction
Skeletal muscle adapts to environmental stimuli and alters its mass accordingly. New myofilaments are added to existing myofibrils through an increase in protein synthesis under the control of mechanical load, nutrients and hormones. Signaling through the phosphatidylinositol 3-kinase (PI3K)/Akt pathway increases protein synthesis [1,2]. Akt is sufficient to induce muscle hypertrophy as demonstrated by muscle-specific expression of a constitutively active form of Akt [3]. Three Akt isoforms (Akt1/2/ 3) have been identified in mice and humans, and implicated in the regulation of growth and metabolism [4,5]. Germline ablation of Akt1 results in a proportional reduction in body size [6,7,8]. In contrast, conditional Akt1 overexpression in muscle resulted in muscle hypertrophy and increase in strength [9,10]. Disruption of Akt2 in mice resulted in insulin resistance, hyperglycemia, hyperinsulinemia, and glucose intolerance [11,12]. As with Akt1, mice lacking Akt2 develop a mild growth deficiency [11,12].
Deficiency of both Akt1 and Akt2 resulted in multiple developmental defects including muscle atrophy [13].
Another key mediator of muscle size is the activin receptor type IIB (ActRIIB) [14]. ActRIIB is a type II transforming growth factor (TGF)-b superfamily receptor that is emerging as a key player in the regulation of muscle size and strength [15,16]. Ligands, including myostatin and GDF-11, bind to the ActRIIB leading to phosphorylation and nuclear translocation of Smad2/3, which mediates muscle atrophy [17]. Inhibition of ActRIIB signaling can be achieved by genetically overexpressing regulatory proteins (such as follistatin), which binds and inhibits endogenous TGFb superfamily ligands [18,19]. Other methods for inhibiting ActRIIB include expression of a dominant-negative form of ActRIIB [19] or postnatal injection of a decoy ActRIIB receptor [15,16]. These methods result in a dramatic increase in muscle mass, more than what is seen in myostatin deficiency alone, indicating that more than one ActRIIB ligand is important for the control of muscle size [16,19,20].
Studies done in cell lines and through electroporation have demonstrated a dominant role of Akt signaling in the regulation of myocellular hypertrophy resulting from inhibition of ActRIIB [10,21,22,23,24,25]. It is unknown, however, whether the interaction between Akt and ActRIIB signaling is responsible for the resulting hypertrophy in vivo. We have previously shown that treatment with a decoy ActRIIB soluble receptor (ActRIIB-mFc) increased muscle size and contraction force [15,26]. We hypothesized that deficiency of Akt1 or Akt2 would prevent the in vivo effects of ActRIIB inhibition on muscle.
Effects of Akt deficiency and ActRIIB blockade on body composition
We first performed immunoblotting on lysates from gastrocnemius muscles to assess the effects of ActRIIB-mFc on ActRIIB signaling on the levels of total Akt and phospho-Akt (Ser473). ActRIIB-mFc treatment increased total Akt significantly ( Fig. 1). The level of phospho-Akt increased slightly in response to ActRIIB-mFc, but this was not significant (Figure 1).
We compared body composition among wild type (WT) and Akt1 knockout (ko) and Akt2ko mice, and also examined the effects of ActRIIB-mFc after 10 weeks. There was a significant effect on body weight with respect to genotype (Figure 2A, p = 0.0008). Akt1ko mice weighed significantly less than WT (14.1%, p = 0.0265) and Akt2ko mice tended to weigh less than WT (13.8%, p = 0.0751). There was a significant effect on lean mass ( Figure 2B, p,0.01), and fat mass ( Figure 2C, p,0.01) with respect to genotype. Lean mass was 15% lower in Akt1ko mice as compared to WT ( Figure 2B, p,0.01) whereas Akt2ko mice were not significantly different. Fat mass was 44% decreased in Akt2ko mice as compared to WT mice ( Figure 2C, p,0.05), whereas it was unchanged in Akt1ko mice. The proportions of lean and fat tissue were unchanged from WT in Akt1ko mice (WT lean: 78.361.7%, WT fat: 18.461.6%; Akt1ko lean: 76.961.5%, Akt1ko fat: 20.561.9%). In contrast, Akt2ko mice had a significant increase in percentage lean tissue and a decrease in percentage fat (Akt2ko lean: 83.760.3%, fat: 12.160.4%, p,0.01 vs. WT for both) when compared to WT. ActRIIB-mFc treatment had a significant effect on body weight (120% in WT, 115% in Akt1ko, and 124% in Akt2ko. Figure 2A, p,0.0001). In accordance with changes in body weight, lean mass was significantly affected by treatment ( Figure 2B, p,0.0001). Fat mass tended to decrease with ActRIIB-mFc treatment in WT and Akt1ko mice cohorts but increased in Akt2ko mice ( Figure 2C, p,0.05). With ActRIIB-mFc treatment, the proportions of lean tissue significantly increased in WT
Effects of Akt deficiency and ActRIIB blockade on muscle structure
We examined the effects of Akt deficiency on various skeletal muscles ( Figures 3A-E). The weights of extensor digitorum longus (EDL), gastrocnemius, anterior quadriceps, tibialis anterior, and soleus were significantly affected by genotype (p,0.01 for all). The mass of EDL, gastrocnemius, and anterior quadriceps was significantly decreased in Akt1ko compared to WT ( Figure 3A-C). Akt2ko mice showed a decrease in EDL and gastrocnemius mass ( Figure 3A-B) and a 21% increase in soleus mass as compared to WT ( Figure 3E, p,0.01). Heart mass was significantly affected by genotype ( Figure 3F, p = 0.0012). Akt1ko heart mass was 20% lower than WT (p,0.05) but unaffected in Akt2ko. ActRIIB-mFc treatment increased the weights of all skeletal muscles measured ( Figures 3A-E). Heart mass was decreased in Akt1ko mice with treatment however not in WT or Akt2ko mice ( Figure 3F, p,0.05).
Next, we compared the effects of Akt deficiency and ActRIIB blockade on EDL and soleus muscles as examples of glycolytic versus oxidative muscles, respectively. The mean cross sectional area (CSA) of EDL was significantly altered with genotype (WT: 1.7560.14 mm 2 , Akt1ko: 1.2260.07 mm 2 , Akt2ko: 1.2560.13 mm 2 , p,0.0001). There was a shift to the left in the distribution of EDL fiber CSA in Akt1ko and Akt2ko mice ( Figure 4A) as represented by a decrease in mean EDL fiber CSA ( Figure 4B, p,0.0001 for both Akt1ko and Akt2ko vs. WT). The number of EDL fibers was affected by genotype ( Figure 4C, p,0.01) however direct comparisons to WT were non-significant. ActRIIB-mFc treatment increased the mean EDL fiber CSA and resulted in a right shift in the distribution of fibers in WT, Akt1ko and Akt2ko mice ( Figure 4A). There was no significant change in the total number of EDL muscle fibers following ActRIIB-mFc treatment ( Figure 4C). These data confirmed that ActRIIB-mFc induced EDL muscle hypertrophy and not hyperplasia [15].
In contrast to EDL, Akt deficiency did not change the soleus CSA (WT: 1 of Akt1ko mice as compared to WT ( Figure 5A-B, p = 0.0582), but the number of type I fibers was not significantly affected (Figures 5C). Type II fibers were smaller in the soleus muscle of Akt1ko mice ( Figure 5D-E, p,0.05) and the number of type II fibers was reduced (Figures 5F, p,0.05). ActRIIB-mFc increased soleus CSA (WT: 1.8860.18, Akt1: 1.7560.15, Akt2: 1.946 0.17 mm 2 , p,0.05) but did not change the total number of fibers (WT: 828.8629.2, Akt1: 854.3674.7, Akt2: 860.5642.2). Ac-tRIIB-mFc did not alter the fiber composition of soleus muscles in WT, Akt1ko and Akt2ko mice (Percent Type II in WT: 61.562.4%, Akt1ko: 54.763.3%, Akt2ko: 65.661.7%). The mean CSA of type I and type II fibers was significantly altered with ActRIIB-mFc treatment ( Figure 5B, 5E p,0.01 for both) and the distribution shifted to the right in all genotypes following ActRIIB-mFc treatment ( Figure 5A, 5D). However, the total numbers of type I and type II fibers were not affected by ActRIIB-mFc ( Figure 5C, 5F).
Effects of Akt deficiency and ActRIIB blockade on muscle function
Forelimb grip strength was reduced in Akt1ko and Akt2ko compared to WT ( Figure 6A, p,0.01). We subjected WT, Akt1ko and Akt2ko mice to an endurance treadmill exercise test until fatigue. There were no apparent differences in the distance run ( Figure 6B), or work done during exercise among the genotypes ( Figure 6C). ActRIIB-mFc increased grip strength across genotypes ( Figure 6A, p = 0.0024). Despite a clear trend we observed a non-significant decrease in distance run ( Figure 6B) and work done ( Figure 6C) during endurance testing in WT and Akt1ko mice with treatment.
Muscle contraction was measured in EDL muscles ex vivo. Maximum twitch force was decreased in Akt1ko and Akt2ko as compared to WT ( Figure 7A, p,0.01 for both). However, there was no significant difference in the twitch dynamics among WT, Akt1ko or Akt2ko EDL muscles ( Figure 7B). The time to peak contraction (CT) was similar across genotypes (WT: 35.063.3, Akt1ko: 32.062.5, Akt2ko: 28.862.3 msec). The time to half maximum force (KRT) was not affected by genotype (WT: 32.561.6, Akt1ko: 31.062.3, Akt2ko: 27.561.6 msec). The maximum tetanic force was decreased in Akt1ko and Akt2ko ( Figure 7C, p,0.01), but tetanus dynamics were not different ( Figure 7D).
Effect of Akt deficiency and ActRIIB blockade on expression of atrophy-related genes
We detected no significant effect of Akt1 or Akt2 deficiency on the expression of atrophy-related E3-ubiquitin ligases, MAFbx/ atrogin-1 and MuRF1 ( Figure S1). Inhibition of ActRIIB signaling resulted in non-significant decrease in expression of MAFbx and MuRF1 in WT mice, and significant decrease in Akt2ko mice ( Figure S1, p,0.01 and p,0.05 respectively).
Effects of Akt deficiency and ActRIIB blockade on Akt signaling
WT, Akt1ko, and Akt2ko gastrocnemius lysates were blotted for expression of Akt signaling proteins ( Figure 8A). Total Akt1 and Akt2 levels were lower in Akt1ko and Akt2ko lysates compared to WT whereas phospho-Akt levels were similar across genotypes. Total glycogen synthase kinase-3 (GSK-3) levels in Akt mutant lysates were comparable to WT, but phosphorylated GSK-3 levels were reduced in Akt1ko and Akt2ko. Mammalian target of rapamycin (mTOR) levels were similar across genotypes, while phosphorylated S6 kinase (S6K) tended to be variable among WT, Akt1ko and Akt2ko ( Figure 8A). Levels of phosphorylated S6 ribosomal protein were undetectable in both Akt1ko and Akt2ko, and total S6 levels were decreased. Glycogen content was significantly lower in Akt1ko and Akt2ko compared to WT ( Figure 8B, p,0.0001 in Akt1ko and Akt2ko versus WT). ActRIIB inhibition increased Akt1 and Akt2 in WT, and Akt2 in Akt1ko mice. We did not detect changes in phosphorylated Akt, phosphorylated GSK-3, total GSK-3, mTOR, and phosphorylated S6K in response to ActRIIB inhibition. Levels of phosphorylated S6 were increased in response to ActRIIB inhibition in WT, whereas total S6 levels were not altered substantially. Glycogen content was significantly increased by ActRIIB inhibition in WT (p,0.05), and tended to increase in Akt1ko and Akt2ko mice ( Figure 8B).
Discussion
Akt plays a pivotal role in mediating the responses to insulin and growth factors [1,27]. Akt1 and Akt2 are highly expressed in insulin-sensitive tissues. Constitutive expression of Akt1 in L6 myoblasts promotes glucose and amino acid uptake and protein synthesis [28]. Overexpression of Akt2, but not Akt1, stimulates C2C12 muscle cell differentiation, suggesting that Akt1 and Akt2 isoforms are necessary for optimal muscle growth [29]. Mice lacking Akt1 are smaller and have been reported to show a decrease in muscle size compared with wild-type mice [6,7,8]. Akt2 deficient mice exhibit a diabetic phenotype, associated with reduced glucose uptake in the EDL but not the soleus muscle, suggesting a fiber type specific insulin resistance (15). Mice with Akt1/Akt2 deficiency develop multiple developmental defects and severe muscle atrophy, indicating that both Akt isoforms contribute to the determination of muscle size. However, detailed analyses of the effects of Akt1 and Akt2 deficiency on body composition, muscle structure and function are lacking.
Our studies revealed differential effects of Akt1 and Akt2 deficiency on body composition and skeletal muscle. In agreement with previous reports, Akt1ko mice were smaller, and displayed proportional decreases in body and lean mass compared to wildtype mice [6,7,8]. In contrast, Akt2ko mice weighed less than wildtype mice, an effect explained mainly by a decrease in body fat. Akt1ko mice had significant decreases in the weights of gastrocnemius, EDL, and quadriceps, i.e. muscles composed mainly of type II fibers [30]. By examining the histology of EDL, we found the Akt1 deficiency was associated with a significant decrease in the cross-sectional area of fibers. In contrast to EDL, Akt1 deficiency did not change the weight of soleus muscle, which is composed of both type I and type II fibers. Nonetheless, Akt1 deficiency decreased the number of type II fibers in soleus muscle. Thus, Akt1 is essential for the determination of the size and number of type II fibers. Our results are in agreement with Easton et al. who demonstrated that cardiomyocytes from Akt1ko mice have smaller CSA [8]. Akt2 deficiency resulted in significant reductions the weights of EDL and gastrocnemius, whereas the weight of soleus muscle was increased relative to wild-type mice. The reduction in EDL weight in Akt2ko was explained, at least partly, by a reduction in fiber cross-sectional area. The reasons for the increase in soleus muscle weight in Akt2ko mice are unknown. There were no apparent changes in the fiber cross-sectional area or number of fibers counted at the mid-section of the soleus muscle. However, it is possible that an increase in soleus muscle triglyceride or extracellular matrix that were not measured in our study may have played a role in the increase in soleus weight observed in Akt2ko mice.
Overexpression of constitutively active Akt1 or Akt2 in muscle results in increased levels of phosphorylated S6K and glycogen content [31]. In agreement with these findings, we found that Akt1 and Akt2 deficient animals have decreased phosphorylated S6 levels, a marker of S6K activity, and decreased levels of phosphorylated GSK-3 and glycogen content.
While Akt is important for the development of muscle size and function, we found that the hypertrophy resulting from ActRIIB inhibition with ActRIIB-mFc was not dependent on either Akt1 or Akt2 isoforms. Similar to wild-type mice, ActRIIB-mFc treatment increased the mean EDL fiber cross-sectional area and shifted the fiber distribution to the right in Akt1ko and Akt2ko mice, consistent with type II fiber hypertrophy [15,32,33]. ActRIIB-mFc treatment induced type I and type II fiber hypertrophy in the soleus muscle. This result is in agreement with a recent study in which ACE031 (the humanized analogue of ActRIIB-mFc) treatment increased the mean fiber cross-sectional area in plantaris, a predominantly type II muscle, and as well as type I and type II fibers in the soleus muscle [26].
Our results suggest that the increase in muscle mass resulting from ActRIIB-mFc treatment is specialized for strength. Forelimb grip strength and EDL twitch and tetanic contractions were increased by ActRIIB-mFc in wild-type as well as Akt1ko and Akt2ko mice. However, ActRIIB-mFc treatment did not improve the distance run or endurance during treadmill exercise. It is unclear whether a longer treatment would reveal distinct effects of Akt deficiency on exercise.
In vitro treatment with the ActRIIB ligand, myostatin, has been shown to increase the expression of atrophy-mediating E3ubiquitin ligases, MAFbx/Atrogin-1 and MuRF1 [21,34]. Additionally, overexpression of Akt in cultured myotubes inhibits glucocorticoid-induced expression of MAFbx [35]. Consistent with these observations, in vivo ActRIIB inhibition increased Akt protein levels and decreased mRNA expression of MAFbx and MuRF1 in gastrocnemius, suggesting inhibition of muscle atrophy.
In summary, our results demonstrate that Akt1 and Akt2 are important regulators of muscle size and function; however, neither Akt1 nor Akt2 isoforms are essential for the response to ActRIIB inhibition. In support of our findings, Sartori et al. found that ActRIIB inhibition in muscles where Akt was constitutively active (c.a.Akt) led to even greater hypertrophy than c.a.Akt alone suggesting that ActRIIB and Akt promote muscle hypertrophy through distinct but connected pathways [10]. We speculate that S6K plays a key role in ActRIIB-mFc induced hypertrophy independent of PI3K/Akt as was the case in the response to mechanical stretch in cultured myotubes [36]. Future studies in mFc -treated Akt2 KO (Akt2 KO Rx, blue dashed) on soleus type I fiber distribution; n$393 fibers per group. (B) Type I single fiber cross sectional area (CSA) per genotype (Veh, white bar) and treatment (Rx, black bar); (C) Number of type I fibers in the soleus; (D) Soleus type II fiber distribution; n$458 fibers per group. (E) Type II single fiber CSA per genotype and treatment; (F) Number of type II fibers in the soleus; Data are mean +/2 SEM, n = 5; +P,0.05 vs. WT treated with vehicle; *P,0.05 vs. same genotype treated with vehicle. doi:10.1371/journal.pone.0012707.g005 muscle-specific Akt-deficient mice will provide insights into potential interactions between Akt and ActRIIB, and the implications for growth and function of muscle.
Animals and treatment
All animal work was reviewed and approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania School of Medicine (Protocol # 701656). Eight-week-old male wild-type (WT) C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME) and Akt1ko [7] and Akt2ko mice [11] were housed (n = 5 per cage) under a 12: 12-h light-dark cycle (light on at 0700) and an ambient temperature of 22uC, and allowed free access to water and chow diet. ActRIIB-mFc (also termed RAP031, 10 mg/kg) provided by Acceleron Pharmaceuticals, Cambridge, MA) was injected intraperitoneally twice weekly for 10 weeks [15]. The vehicle was phosphate-buffered saline (PBS). Mice were fed regular chow diet (LabDiet, Richmond, IN, # 5001, containing 4.5% fat, 49.9% carbohydrate, 23.4% protein; 4 kcal/g). Food intake was measured weekly, and body composition was assessed prior to treatment and 10 weeks later with nuclear magnetic resonance (NMR) (Echo Medical Systems, Houston, TX) [15,37].
Treadmill and grip strength
ActRIIB-mFc or vehicle-treated mice were acclimatized to a modular treadmill connected to an open circuit indirect calorimeter (CLAMS, Columbus Instruments, Columbus, OH). After 10 weeks of treatment, the mice were deprived of food for 5 hours in the morning, and underwent forced exercise on the treadmill (angle 10u, 15 min at 10 meters/min and then where m = mass, g = 9.81 m/s 2 , m k = 0.987, h = angle of treadmill incline, dx = horizontal distance, dy = vertical distance. The muscle strength in the forelimbs was measured with a grip meter (TSE; Bad Hamburg, Germany) as previously described [15]. Briefly, mice were trained to grasp a horizontal metal bar while being pulled by their tail and the force was detected by a sensor. Ten measurements were determined for each mouse and averaged.
EDL contraction
The mice were euthanized with carbon dioxide, muscles were dissected and weighed, and contraction of the extensor digitorum longus (EDL) muscle was analyzed ex vivo. Physiological measure-ments, including contraction time, half-relaxation time, peak isometric twitch force, and peak isometric tetanic force were analyzed in freshly dissected EDL muscles as previously described [38]. Briefly, muscles were stimulated in Ringer's solution composed of 100 NaCl mM, 4.7 KCl mM, 3.4 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 25 mM HEPES, and 5.5 mM D-glucose. Muscle length was adjusted to obtain the maximal twitch response, and this length was measured and recorded as optimal length. Peak twitch and tetanus forces were obtained from three maximal contractions (120-Hz stimulation frequency, 500-ms stimulation duration).
Histology
At the end of the physiological studies, EDL and soleus muscles were flash-frozen in isopentane cooled in liquid nitrogen and stored at 280uC prior to sectioning. Serial frozen sections (12 mm) were cut at mid-belly of the EDL muscle using a cryostat at 221uC and placed onto glass slides (Superfrost/Plus, Fisher Scientific). Single-fiber cross-sectional area (CSA) and distribution were determined in images from tissue sections immunostained for laminin [39]. Photomicrographs were taken using an Olympus BX51 microscope equipped with a Magnafire camera. Morphometric measurements were made with Photoshop CS3 (Adobe). Fiber data was used to calculate an average single-fiber area and plot it as a histogram. For ease of interpretation, histograms are presented as smoothened approximations by constructing a spline function (GraphPad Prism), whose average value over each bar Figure 8. Effects of ActRIIB inhibition on Akt signaling and glycogen content. (A) Immunoblots of vehicle (Veh) and ActRIIB-mFc treated (Rx) wild-type (WT), Akt1 knockout (Akt1ko), and Akt2 knockout (Akt2ko) gastrocnemius lysates. P: phospho-specific antibody (B) Effects of genotype (Veh, white bar) and ActRIIB-mFc treatment (Rx, black bar) on quadriceps glycogen content (mg glycogen per g muscle). +++P,0.0001 vs. WT Veh, *p,0.05 vs. vehicle. doi:10.1371/journal.pone.0012707.g008 interval equals the height of that bar. The total number of fibers was counted per muscle section. Type I and type II fibers were distinguished through metachromatic ATPase method [40,41].
Glycogen content
Quadriceps muscles were dissected, weighed, and flash frozen in liquid nitrogen. Samples were homogenized in 0.03 N HCl. The homogenate (100 ml) was mixed with 400 ml of 1.25 N HCl and heated for 1 h at 100uC. Samples were centrifuged at 14,000 rpm, and 10 ml of supernatant was mixed with 1 ml of glucose oxidase reagent (Sigma). After a 10-min incubation at 37uC, the absorbance was read at 505 nm. A standard curve using glycogen type III obtained from rabbit liver (Sigma) was also simultaneously analyzed to determine the final liver glycogen concentrations. | 2014-10-01T00:00:00.000Z | 2010-09-15T00:00:00.000 | {
"year": 2010,
"sha1": "aad169227e769d220b16e1e5611c693b3d990cdf",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0012707&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "aad169227e769d220b16e1e5611c693b3d990cdf",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
113907061 | pes2o/s2orc | v3-fos-license | Possibility of Thermal Storage System Use with Different Accumulating Material in SPbSTU
In this article the possibility of heat accumulator use in one of the buildings of SPBSTU was analyzed. The use of this accumulator could lower the hot water demand. To complete the goal it was assumed that there is an additional café with defined working hours. Different kinds of heat collecting materials were studied: water, red brick, concrete and granite. Depending on charging time of the heat accumulator, the results of the heat gaining abilities of different materials were made. As a result it was calculated that single-phased accumulating materials should not be used as heat collectors in public buildings.
Introduction
Nowadays, energy efficiency is a leading topic in civil engineering.It saves money, resources so in the long run it helps economics and ecology.
To achieve high energy efficiency the heating system distribution should be modernized.One of the variants is to install heat accumulators for hot water and heating systems.In Russia it is popular in small houses outside cities.
In this article one of SPBSTU campuses is studied when it is assumed that there is an additional heating source along with higher hot water demand.
Literature review
Energy efficiency has become world tendency and now is going to be the first thing to consider during the designing process of all constructional elements [2][3][4].In many countries there is legislation for energy efficiency, in Russian Federation it is Federal Law No. 261-FZ of November 23, 2009 "On Energy Saving and Increase of Energy Efficiency and Introduction of Changes into Separate Legislative Acts of the Russian Federation" [5].
During the design development designers in the first place try to apply energy saving space-planning concepts -construction of wide-bodied houses [6,7] allows to reduce specific area of enclosing structures.They also use energy saving construction systems like heat insulation of exterior walls and hinged ventilated facades [8,9].Projects of "Green houses", partly or fully isolated from energy sources, appear [10][11][12].As practice shows, energy saving can be highly increased by implementing new building services systems or by renovating them, for example, by installing automated heating unit [13][14][15].Lately the appliance of heat accumulators became a very popular approach.They can store some amount of heat energy to use it during day periods when charges are higher or during the hot water consumption peaks to reduce the load on the heating unit [16][17].
Purposes and tasks
The purpose of this article and our research was to choose the most suitable type of heat accumulator for help to smooth the peak of hot water consumption or to define that under existing conditions it is inappropriately.
Following problems were solved: x Calculation of hot water demand after the reconstruction of café; x Calculation of possible amount of accumulated heat for different types of singlephased accumulating material (SPAM) in the accumulator; x Analysis of the results and conclusion about possibility of application of different types of heat accumulators for smoothing the hot water consumption peak in considered problem.
Research description
Nowadays, there is a serious problem with cafes in SPBSTU because there is not enough place for all the people.Let us assume that to solve the task an additional café was working during peak loads on the existing café.Changes of the number of water-supply points can be seen in Table 1.
Table 1.
Comparison of number of water-supply points in existing café and in designing café.
Existing café
Additional There are two ways to lower the maximal hot water use: • Recalculation of the substation and an expensive modernization afterwards; • Installation of the heat collector as an additional equipment.The heat collector will be used during when occupation of the café is maximal.From 12.00 to 16.00 there is the highest use of hot water.Outside this period an additional café is not needed.These changes will highly increase the load on hot water supply system and the existing heat unit will not be able to bear it.Nominal capacity of the heat unit is 200 W and maximum hot water flow is 0.96 l/s.In order to smooth the consumption peak it was decided to install heat accumulator in series, when the accumulator is connected into power circuit after the energy source.
Main element of every heat accumulator is accumulating material.Hot water in hot water supply system transfers the heat to the SPAM, which, due to its high heat capacity, can accumulate a huge amount of heat during the charging time.Due to good quality of thermal insulation the heat energy may be stored for a long time and be used when neededin our case during the consumption peak since 12 a.m. to 4 p.m.
In order to find the most profitable accumulator we considered several types of relatively cheap and non-toxical SPAMs (Fig. 1): Fig. 1.Types of heat accumulating materials.
In this article we considered heat accumulators with single-phased SPAM.Let's have a closer look to the working process of heat accumulator.
Heat accumulator charge
Single-phased SPAM (Fig. 2a) is located around the pipe in which heat-transfer medium flows.Tci -input medium temperature, Tco -output medium temperature.Passing through the channel heat-tranfer medium is getting colder (Tci < Tco), transferring heat to the SPAM, thus the temperature of accumulating material Tn is increasing.
Heat accumulator discharge
The most common system is used: cold heat-transfer medium with input temperature Tdi flows through the channel and warms up to output temperature Tdo (Fig. 2b).By changing charging time we can choose the most profitable variant of using the heat accumulator for every type of SPAM.
Calculations
Firstly, we calculate the number of conditional dishes consumed per hour and per one working shift.Café is designed for 50 seats.
Number of dishes consumed per hour can be calculated with the formula (1): Where n -number of seats, n=50; m -number of different customers seating at one seat per hour, m=3 (for students' café); u ୭ -number of conditional dishes consumed by one customer (u ୭ = 2.2).
Heat accumulating material
Single Where u -calculated number of conditional dishes consumed per hour; T -working time of café; K -hourly non-uniformity coefficient (K=1.5).Therefore, for the calculated number of conditional dishes hot water flows were calculated (using SNiP 2.04.01-85 «Domestic water supply and plumbing system"): q h = 0.65 l/s -maximal design flow; q hhr = 1.28 m3/h -maximal hourly flow.Heat flow during the hour of maximal water consumption Q hr = 76800(ccal/h) = 321 546.240 (kJ/h) = 89.3kW.
So, we can calculate how much heat is needed to be stored in the heat accumulator, if we multiply heat flow per hour by working time.
Then, amount of required heat is Q=1 286 184.96 kJ.Physical characteristics of materials, calculated on the base of parameters found, are presented on the Table 2.As a result of modeling of heat accumulators with single-phased SPAM in series we have following functional connections: For the temperature of SPAM (3),(4): Charge: T o -initial temperature of SPAM (To = 283 К); T ci -input temperature of heat-transfer media entering the accumulator during charge (T ci = 338 К).Discharge: T di -input temperature of heat-transfer media entering the accumulator during discharge (T di = 283 К -in summertime, T di = 278 К -in wintertime) For the temperature of heat-transfer media in channel as it leaves heat accumulator (5), ( 6): Charge: As it can be seen from the graph, even during the maximal charging time -20 hours -SPAM can't accumulate the required amount of heat (1286 MJ).
Conclusions
1. Hot water flow in "Hydrotechnical building-2" of campus of SPbSTU during the working hours of designing café has increased almost in two times and became 1.61 l/s instead of 0.96 l/s 2. Applying of heat accumulator is unpractical for full and autonomous supply of café for 50 seats.3. Heat accumulator with SPAM-water can be used as additional power source.
Fig. 3 .
Fig.3.Dependence of heat amount on the charging time.
Café is open for 4 hours in a day.Daily quantity of dishes consumed can be calculated with the formula (2):
Table 2 .
Physical characteristics of different types of SPAM
Table 4 .
) Dependence of heat amount on the charging time Dependence of heat amount on the charging time is shown on the fig.3.
[MJ] The charging time tc, [h]
Heat accumulators with SPAM-granite/brick/concrete don't accumulate a lot of heat when charging time is increased maximally 5. Recalculation of heat unit with nominal capacity of 289.2 kW and hot water flow of 1.61 l/s is needed. | 2018-12-01T04:51:30.521Z | 2016-01-01T00:00:00.000 | {
"year": 2016,
"sha1": "1f45f967d8b2bf9de3b5ba38e334433471306cb8",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1051/matecconf/20167302010",
"oa_status": "GOLD",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "1f45f967d8b2bf9de3b5ba38e334433471306cb8",
"s2fieldsofstudy": [
"Engineering",
"Environmental Science"
],
"extfieldsofstudy": [
"Engineering"
]
} |
254499121 | pes2o/s2orc | v3-fos-license | Emergency Door Capacity: Influence of Door Width, Population Composition and Stress Level
For the new version of the Dutch design guidelines for buildings, a threshold value for the capacity of emergency doors needs to be set. Innovative large-scale laboratory experiments have been performed to investigate the capacity of emergency doors during evacuation conditions. This paper focuses in particular on the relation between capacity and the independent variables doorway width, population composition, stress level and presence of an open door. It turned out that only the experiment with the widest doorway (275 cm) resulted in a capacity lower than the capacity from the current design guidelines (2.25 P/m/s). The average observed capacities are for all widths lowest for the lowest stress level and highest for the highest stress level. The population with a greater part of children has the highest capacity (on average 3.31 P/m/s). This is mainly due to the smaller physical size of children compared to adults and elderly, which makes it possible that more children can pass a door at the same time than adults. The lowest capacity (on average 2.02 P/m/s) is found for the experiment with 5% disabled participants. The presence of a door opened in the escape direction in an angle of 90° for a doorway of 85 cm results in a 20% capacity reduction compared to the reference experiment. The open door does not physically narrow the doorway, but it leads to interactions between participants reducing their speed and the corresponding outflow.
Introduction
Since 1992, the Dutch national building code (''Building decree'') sets requirements to the width of emergency doors. Since 2003, these requirements depend on the number of persons that rely on an emergency door. According to the Building decree a door width of 1 m is sufficient to let 135 persons pass during the period available for safe escape (1 min). This value corresponds to research of Peschl [1], being similar to 2.25 P/m/s. The threshold of 135 persons per meter width during a safe escape time of 1 min has been discussed for years between the Ministry for Housing, Regional Development and the Environment and the fire brigades that are used to allow a maximum of 90 persons per meter width during a safe escape time of 1 min (1.5 P/m/s).
A literature research has been performed to find other research related to similar bottlenecks. In 2002, the department Transport and Planning of the Delft University of Technology performed laboratory experiments with a narrow bottleneck. This resulted in a capacity of 1.77 P/m/s [2], which is closer to the capacity recommended by the fire brigades than in the design guidelines. However, the narrow bottleneck spans a short hallway of five meters long, whereas a doorway usually has a length of 10-40 cm. This will most likely lead to a higher capacity for doorways, since pedestrians may accept short headways for a short period of time.
Kretz et al. [3] performed bottleneck experiments as well. In these experiments, the bottleneck was a thick wall of 40 cm with an opening the pedestrians had to pass. Different widths for the opening have been considered (40 cm, 50 cm, 60 cm, 70 cm, 80 cm, 90 cm, 100 cm, 120 cm, 140 cm, and 160 cm). The participants consisted of healthy students; the experimental conditions were normal. A linear decrease of the capacity is shown with increasing bottleneck width as long as only one person at a time can pass (from 2.2 P/m/s for 40 cm to 1.78 P/m/s for 70 cm width). A constant value of the capacity of around 1.8 P/m/s is shown for larger bottleneck widths (70 cm, 80 cm, 100 cm and 120 cm). Only the very narrow bottleneck thus shows a capacity slightly lower than the capacity indicated in the design guidelines, the other bottlenecks width result in lower capacities.
Experimental research by Mu¨ller [4] and Nagai et al. [5] indicated much higher capacities for bottleneck widths varying between 80 cm and 160 cm, namely between 2.29 P/m/s and 3.23 P/m/s. These high values can be explained by the very high densities at the start of the experiments. Also, the configuration of the bottleneck is slightly different, which affects the measured capacities [6]. These values correspond to the threshold indicated in the design guidelines.
Many observations have been performed in corridors and in areas with many pedestrians present (e.g., stations, inner cities, and stadiums), for overviews see [7,8]. The capacities found vary between 1.03 P/m/s and 1.67 P/m/s, thus much lower than the design guidelines. However, these capacities are found in normal conditions, which most likely will lead lower capacities than in evacuation conditions.
The abovementioned overview indicates a wide variety of capacities found, while most of the capacities are lower than the capacities in the design guidelines. The aim of the research described in this paper is therefore to perform experimental research to collect new information on the capacity of doors during emergency conditions. The next section describes the set up of the experiments in more detail. It also gives a short impression of the day of the experiments. In the third section the methodology is described to calculate the capacities, the results of which are shown in the next section. We end with conclusions and recommendations for future research.
Experimental Set Up
The capacity of an emergency door depends on several aspects, among which the composition of the population using the door, the conditions under which the door is used and the door width. Before describing these experimental variables in more detail, some boundary conditions are set.
In the experiments, an opening represents the emergency door: subjects pass a free passage of a certain width. In this opening, no doorstep is present, to reduce hindrance and prevent possibly dangerous situations for participants. In addition, the pedestrian flow is one-directional, implying that no counter flows are present caused by fire fighters and people from emergency services. In reality, these people will rarely enter a building when the evacuation process is still going on.
The experiments performed by Peschl [1] have been based on a student population. However, in practice, the population will not consist of persons being in good shape, but the persons will have different physical conditions. In this research, we will use age as an indication for a person's physical condition. Here, we distinguish four categories: children (under 18 years of age), adults (between 18 and 65 years of age), elderly (over 65 years of age) and disabled persons. With these age categories, we are able to compose populations corresponding to a variety of situations, see Table 1. The disabled people are represented by three persons in wheelchairs and three blindfolded persons.
The conditions under which an emergency door is used may vary considerably. In the experiments, both the stress level of the participants and the sight are varied. Not much is known on how to introduce stress in an experiment. In the past two methods have been considered favorable: enforcing participants to hurry e.g., by rewarding participants according to their performance and exposing participants to noise. Here, we have chosen to use for the latter option by sounding the slow-whoop signal. In addition, the stress level of the participants is raised by a combination of the slow-whoop signal and stroboscope light. In total, participants have been exposed to three stress levels: none, a slow-whoop signal and a combination of a slow-whoop signal and stroboscope light.
The sight is reduced by reducing illumination to a low level. Two alternative light situations are considered: full lighting (200 lux) and dimmed (1 lux, corresponding to emergency lighting). In the experiments, the opening width is varied between 50 cm (the minimal free passageway of an escape route in the Building decree for existing buildings) and 275 cm. In addition to an opening of 85 cm wide (minimal free passageway of an escape route in the Building decree for new estates) openings are a multiple of 55 cm. Furthermore, an opening of 100 cm is tested to see the correspondence with the normative capacity expressed as the number of persons passing an opening of 1 m wide in 1 min.
The final experimental variable relates to whether or not the outflow of pedestrians after passing the doorway is free. In reality, doors cannot always open 180°, but may be restricted. To restrict the outflow of pedestrians after passing the doorway, we used a door being fixed at an opening of 90°. The total doorway width is not affected.
Ideally, all combinations of experimental variables should be investigated. Since this is not feasible due to time restrictions (the experiments should not last longer than a single day), for each experiment one variable is changed, while for the other variables the default value is maintained. By interpolation of the results of the various experiments, pronouncements can be made on the not performed experiments. The stress levels are varied for all experiments.
Each experiment will be performed multiple times to guarantee the reliability of the observations. To determine the number of repetitions, a total time of congestion of 3 min should be achieved. Since the time of congestion for wide doors is shorter than for narrow doors, more repetitions are performed for the wide doors.
An overview of the experiments is shown in Table 2. A digital video camera and an infrared camera are used to observe the experiments. The infrared camera observes LEDs, attached on top of the caps of the In total 75 children of 11 years old (blue caps), 90 adults (red caps) and 50 elderly persons (yellow caps) have participated in the experiments. This leads to populations of between 90 and 150 persons, which are large enough to cause congestion upstream of the door to observe capacities.
To represent an emergency door, a wall has been built in the middle of a large hallway, perpendicular to the sidewall. In this wall, an opening is made, whose width is easy to vary. At the side of the wall, some space is left to walk from one side of the wall to the other without using the opening. Above the centre of the opening an emergency exit sign has been hung up. An overview of the experimental site is shown in Fig. 1. To use the doorway more efficiently the participants use it in two directions: in the first experiment, they walk from one side of the wall to the other and in the next experiment they walk back again.
Methodology to Calculate Emergency Door Capacity
The images of the digital video camera form the basis to calculate the capacity of an emergency door. The movie of each repetition of an experiment is split into separate images with a frequency of 25 images per minute. Figure 2a shows such an image of the reference experiment with a doorway of 85 cm wide, an average population, 200 lux, no open door and no stress.
To calculate the capacity of the emergency door, the moments that the subjects pass a cross-section directly downstream of the door need to be known. Figure 2b shows the considered cross-sections: the dotted line indicates the cross-section to determine the capacity of the flow from right to left, while the striped line corresponds to the cross-section to determine the capacity of the flow from left to right.
To determine the cumulative curves on the specified cross-sections, the picture lines of the cross-section are placed next to each other for the duration of a repe-
Emergency Door Capacity 59
tition of an experiment using MATLAB Ò . This gives a similar effect as a finish photo, see Fig. 3a. Figure 3b shows a zoom of the picture lines in Fig. 3a. This figure clearly shows that the first participants have passed the cross-section at high speed: their image is very narrow, while the images of the following persons are wider. A similar approach has been used to identify vehicle trajectories from video images made from a helicopter [9]. The next step is to recognize the caps to automatically derive the passage moments of the individuals using the cap colors. In the 'photo finish' image areas are looked for within a pre-specified color range. By not only looking for the colors bright red, bright blue and bright yellow, but by allowing a color range, the shadow effects, the cap edges and the stroboscope effects are compensated for. The detected pixels are then combined to caps. For this, two built-in methods in Matlab Ò have been used: a method to identify areas within a specific color range and clustering methods. More information on this cap recognition methodology can be found in [10]. Figure 3c shows the tracked participants for a repetition of the reference experiment.
For each cluster (participant) its heart is calculated. The x-pixel of this heart indicates the time moment when the participant passes the cross-section, while the y-pixel indicates the lateral position of this event. A cumulative curve can be derived based on these passage moments, where the number of persons having passed the cross-section is plot against the passing moment, see Fig. 3d.
Assuming that the capacity of the door does not change during a repetition of the experiment, a straight line is fit through the cumulative curve. The derivative of this line corresponds to the average capacity of this door during this repetition. The average capacity of the experiment is then the average of the capacities of all repetitions. If the average capacity of each experiment is known, the relations between the capacity and the various experimental variables (door width, population, stress level, etc.) can be determined.
Doorway Width
During the experiments the opening width has been varied between 50 cm and 275 cm. All these experiments have been performed with an average population (with a slight change for the widest doorway), a normal light intensity (200 lux) and without the presence of an open door. Figure 4 shows the results of these experiments.
For each experiment, the observed capacity is shown in the figure. The type of marker indicates the stress level, while the green star represents the average capacity per experiment over all stress level. In addition, the current threshold capacity from the Building decree has been indicated (C = 2.25 P/m/s = 135 P/min).
62
Fire Technology 2012 The figure shows that only the experiment with the widest doorway results in an average capacity lower than the threshold value from the Building decree. Furthermore, the high capacity of the doorway of 220 cm is remarkable, as well as the large difference between the repetitions in this experiment. Figure 4 indicates that the repetitions without stress or with a low stress level result in the highest capacity for a doorway of 220 cm, while the experiment for a doorway of 100 cm contradicts this finding. Figure 5 gives more insight into the influence of the various stress levels. Figure 5 shows that the average observed capacities over all doorway widths are lowest for the lowest stress level and highest for the experiments with slowwhoop and stroboscope considered as the highest stress level. For all cases the average observed capacities are much higher than the value included in the Building decree. As well as in Fig. 4a, the figure shows some outliers for the experiments without stress and with only a slow-whoop signal for a doorway of 220 cm. An explanation can be found in the time of the day this experiment has taken place (see Fig. 4b). Figure 4b indicates that both experiments with the largest variance in capacity have taken place at the beginning of the morning. Moreover, the figure shows that the stress level has an opposite effect for both experiments: for the first experiment the capacity is lowest for the lowest stress level, while in the second experiment the highest capacities occur at the lowest stress level. not show such a clear effect of the various stress level. This leads to the conclusion that the difference is not structural and can be attributed to the conditions (enthusiasm) during the first experiments. Although the number and the distribution of the participants over the three groups (children, adults, elderly) are equal for all experiments, other persons participated. While the participants of the first experiment did not know what to deal with, the participants of the second experiment could wait and see what was happening. Especially at the start of the experiment, these participants were very motivated and were in full focus to pass the door. In the first repetitions (without stress and with slow-whoop, respectively) this led to pushy behavior; this was clearly visible in the video images.
Although the behavior of the participants appears to influence the capacity, it is impossible to motivate the participants just as much for each experiment during the total day. Despite this fact, the capacity of almost all repetitions is higher than the capacity prescribed in the Building decree. Only most repetitions of the experiment with the widest opening are below the capacity threshold from the Building decree. Since the first experiments showed that the capacities appeared to be higher than the planned capacities all adults and elderly have joined this experiment. This led to a slightly different population with relatively less children than the average population, which has a negative effect on the capacity as will be shown in the following. This slightly adapted average population only participated in the experiment with the widest door; in all other experiments the regular average population has participated.
Population Composition
During the experiments also the population has been varied. These experiments have been performed with a doorway of 85 cm wide, a normal light intensity (200 lux) and without an open door. Figure 6 shows the results of these experiments.
The figure above shows that five out of six experiments result in a capacity higher than the capacity threshold indicated in the Building decree. Only the population with 5% disabled persons (three blindfolded participants and three participants in wheelchairs) results in a slightly lower capacity (2.0 P/m/s vs. 2.25 P/m/s). The population with mainly children has the highest capacity. This is mainly caused by the physical fact that children are smaller than adults, which makes it possible for more children to pass a door at the same time. The populations representing a retirement home, a meeting and a shopping centre do not differ much. Conversely, the capacity of the population 'station' varies considerably from the population 'meeting'. The first population consists only of adults, while the second population consists of 90% adults, completed with 5% children and 5% elderly. However, the difference between both capacities is somewhat more than 8%. Also the population 'shopping centre' and 'average' have a substantially different capacity (15%), while the first population has only 5% more children, 5% more adults and 10% less adults. These differences might be explained by the moment of the day the experiment has been performed (see Fig. 6b).
For both situations mentioned above the performance moment of the experiment has a clear but opposite effect. The experiment with the average population 64 Fire Technology 2012 was the fourth experiment of the day just before a short break, while the experiment with the shopping centre population occurred halfway the afternoon. At that moment the fatigue had increased considerably and the enthusiasm decreased, which lead to a lower capacity than the capacity of a comparable average population. Exactly the opposite causes the difference in capacity between the meeting population and the station population. The experiment with the meeting population occurred by one o'clock, when the participants were clearly in need of a lunch break, while the experiment with the station population occurred at the end of the day. To motivate the participants of the station population (consisting of only adults) extra, the challenge was raised to improve the highest capacity until then, set by the school population (''let's see if you can evacuate faster than the children''). This lead to a very strong motivation, resulting in a much higher capacity than the one of a similar population.
The variation in capacity is highest for the school population, which can be attributed to the fact that children strongly react to each other: if the first person passes the doorway very fast, the others will follow very fast as well, whereas if the first person passes the doorway very slow, the others will also take it easy. However, the variation between the experiments with the stroboscope was very small, probably because this unusual external condition makes the children focus more on the aim of the experiments (less distraction).
Presence of an Open Door
The last experimental variable discussed in this paper is the presence of an open door. Figure 7 shows the results for this experimental variable.
From the figure it can be concluded that the capacity decreases up to 80% of the capacity threshold indicated in the Building decree when an open door is present in the doorway. This door does not physically narrow the doorway, but it reduces the outflow of the participants. At the location of the door some kind of narrow corridor exists, while on the other side participants are only hindered in their lateral movement at the moment of passing the wall. In the situation without a door participants fan out in all directions immediately after passing the door, which is not possible when a door is present. This is clearly visible in the trajectories for both situations (see Fig. 8) and an overview of the use of the area downstream of the door (see Fig. 9, the darker the areas, the more frequently this area is used).
When an open door is present, not only less surface directly downstream of the doorway is used, but this surface is also used more intensely (darker areas in Fig. 9). The fact that the doorway is not fully used is not caused by pedestrians maintaining distance to the door, but because pedestrians cannot directly swerve to the right downwards of the wall. This is slightly compensated by pedestrians passing the opening at the left hand side swerving more to the left to give space to pedestrians passing the opening at the right hand side. The angle between the used surface on the left hand side of the wall and the wall is therefore smaller than in the situation that no door is present. Figure 9 also shows that a door not necessarily has to open 180°, since the surface directly behind the wall is not used. A maximum opening angle of 150°appears to be sufficient for a free outflow.
The intense use of the surface immediately downstream of the door results in many interactions between pedestrians, which leads to lower speeds and thus to a lower outflow (and thus capacity) at the doorway.
Multivariate Regression
In the previous section, the influence of each independent variable on the capacity has been discussed separately. In this section we will show the influence of all variables together, performing a multivariate regression. Here, we take into account stress-level S (0 = no stress; 1 = slow whoop signal; 2 = slow whoop signal combined with stroboscope), door width W in meter, population (percentage of children P C , percentage of elderly P E and percentage of disabled persons percentage of adults P A , since this is redundant with the other percentages (P A = 1 -P C -P E -P D ). The capacity C can then be calculated as follows: C ¼ 2:6685 À 0:0065S À 0:1153W þ 1:0612P C À 0:2077P E À 2:1310P D À 0:1789D þ 0:0895L À 0:0850T ð1Þ shows that the reference capacity is 2.6685 P/m/s. The stress level hardly has an effect on the capacity. The door width also has a negative effect, implying that wider doors are less efficiently used. Obviously, the door width for which Eq. 1 can be applied is limited, otherwise the capacity becomes negative.
Since the widest door handled in the experiments had a width of 2.75 m, we suggest to take 3 m as boundary width. With respect to the population, we can see that the children have the largest, positive, effect on capacity: the more children are part of the population, the higher the capacity. A higher percentage of elderly also has a slightly positive effect, whereas more disabled persons in the population have a strong negative effect on the capacity. The presence of an open door also has a negative effect on the capacity, whereas the lighting conditions have a positive effect (more light indicates a higher capacity). Finally, as we saw in earlier analyses, the time moment that the experiment took place has a slightly negative effect on capacity.
68
Fire Technology 2012 The next question is whether these parameters are significant. For this, we applied the t-test for each parameter. Table 3 shows that only the parameter for the stress level is not significant.
Finally, we checked the correlation between the parameters. Table 4 shows that the variables are not correlated.
Conclusions and Recommendations
The main conclusion that can be drawn is that the capacity of thirteen out of sixteen experiments is higher than 2.25 P/m/s. The experiments with a lower capacity have a population with disabled persons, a very wide doorway (275 cm) with less children in the population and an open door.
Another conclusion is that in the performed experiments more pushing does not lead to the 'faster-is-slower' effect. In the experiments a higher urgency (higher stress level) leads to higher speeds and to a higher capacity.
During the experiments, the conditions were optimal. This implies among other things that the inflow of pedestrians towards the door is unhindered, which is also the case for the outflow (except for the experiment with the presence of an open Table 4 Correlation Between Variables . Also, factors such as people wearing winter clothing or carrying bags have not been taken into account. In order to determine the design value of the capacity of an evacuation door, users of the building code and designers of buildings should realize themselves in which conditions their doors will be used, and adapt the capacity value accordingly. Many differences between the observed capacities can be explained by the different experimental variables. The images of the experiments indicate that an explanation can also be found in the individual behavior of the participants. When this microscopic behavior can be predicted, also the capacities can be predicted for a larger variety of conditions. This will be subject of future research. The research described here has explicitly been focused on the capacity of emergency doors. This is only part of the total evacuation process. The previous process (pre-evacuation, route choice, walking towards the exit) has a direct influence on the arrival pattern of pedestrians at the emergency door, and thus whether or not capacity of the door will be reached. This is also subject of future research. | 2022-12-11T14:20:00.361Z | 2010-12-16T00:00:00.000 | {
"year": 2010,
"sha1": "6baf0398702b9f01a845383c2a3ea0f9f8871202",
"oa_license": "CCBYNC",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s10694-010-0202-9.pdf",
"oa_status": "HYBRID",
"pdf_src": "SpringerNature",
"pdf_hash": "6baf0398702b9f01a845383c2a3ea0f9f8871202",
"s2fieldsofstudy": [
"Engineering",
"Environmental Science"
],
"extfieldsofstudy": []
} |
199145471 | pes2o/s2orc | v3-fos-license | Malay Language Learning Innovation in Enhancing Essay Writing Performance
This study aimed to get information about essay writing using the Frog Virtual Learning Environment (VLE Frog). A total of 60 samples of Form Four students in Kuala Lumpur Technical School were selected and divided into experimental group and control group. Quasi-experimental study was conducted to see the performance in essay writing, grasping the main idea and amplifying the contents. Results of t-test analysis in the post-test indicated that the ratio of the total marks for the experimental group and the control group showed significant differences in the t (58) = 3.219, p <0.05. Meanwhile, finding analysis in grasping the main idea in post-test did not show a significant different which is t (58) = 1.287, p<0.05. Results of t-test analysis essay content amplification between the two groups have shown that there was a significant difference in treatment between the two groups of contents at t (58) = 4.264, p <0.05 In conclusion, the use of VLE Frog in teaching essay writing has proven to be more effective than teaching by using traditional teaching methods even though grasping the essay content has less impact on the students’ achievement. Hence, the findings of this study may encourage the teachers to adapt teaching VLE Frog in essay writing and student will be more interested in writing essays.
Introduction
The use of computers in teaching and learning (TnL) has been widely accessed and accelerated in the classroom. This is confirming to the goal of the Ministry of Education Malaysia (MOE) which aims for Malaysia world-class education. Internet technology has changed the lives of community and society around the world in various fields such as education, economy, politics, social and so on. The Internet usage has been growing from time to time (Noraini, 2010). One of the fields that is growing in the internet era is virtual learning.
The virtual learning environment is an internet-based that builds to connect the community members and enable them to talk and collaborate in a task or project. All community members that have been involved in the virtual learning environment need to have a computer and can access online from anywhere. Community members can actively engage by exchanging views on future developments, sharing views on the implementation of certain policies or discussing the progress of their tasks.
Frog's Virtual Learning Environment (Frog VLE) is a web-based learning system that mimics the real learning by integrating the conventional education concepts with virtual methods. For example, teachers can give assignment and test, and check the assignments virtually, meanwhile, students can submit the assignments and check their marks through Frog VLE. Besides, parents can communicate with the school board. While, school administrators can manage school calendars and post the school notice via the internet.
Through Frog VLE, education is no longer limited to the traditional classrooms. Learning can be done anywhere and anytime. This easy-to-use Frog VLE system allows students and teachers to find almost everything in the internet (videos, photos, educational materials) and develops it into an attractive site without a need for any technical expertise. Frog VLE is also a gateway to abundant teaching resources and quality apps across the web, in which all of it can be access in a safe and filtered environment.
Consistent with the Malaysian Education Development Plan (MEDP) 2013 to 2025, Ministry of Education Malaysia (MOE) introduced the Frog Virtual Learning Environment or Frog VLE and has expanded the technology to increase the level of education and prepare the students with some necessary skills. Besides that, all teachers receive 4GB of Internet access together with highly effective gadgets. It is a world-class learning platform and accessing to the resources and technology is in line with the growth and development of electronic media and mass communication.
Problem Statement
Essay writing is a difficult part for students to get high score in Sijil Pelajaran Malaysia (SPM) examination especially for poor students. According to Awang, (2004) write-up work is rather complex and it is a professional work. While, according to Mahzan, (2008) writing activity is considered to be a difficult and challenging activity for a student's ability. Student's lack of interest in writing essay is due to their inability to highlight the content and elaborate the contents.
Writing activity is also a complex high-level thinking activity that reflects the feeling imagination, perception, mind level and cognitive level of the author and is regarded as one of the most difficult language skills. Weakness in writing may as result of students do not have thinking skills, lack of existing knowledge and less effective learning management. According to Mohd. Khir & Marzukhi (2009), majority of the students cannot write the essays well and accurate, and meet the requirements of the questions within the prescribed time frame. In addition, most of the students are not able to display meaningful content, do not have a broad knowledge on the essay topic and not able describe the content clearly and accurately. They are also not able to provide accurate or clear evidence and examples for each content. Perhaps, this phenomenon happens due to lack of reading high-quality materials among students and they get less exposure to methods of writing paragraphs for a quality essay. The situation shows that if the students are weak to master the writing skills, they will not able to write a good essay and present the ideas effectively. Therefore, by using Frog VLE, teachers can upload the materials or issues that to be written in a writing essay. Besides that, students also will have the knowledge before writing an essay. The outcome will absolutely coincide with the requirements of the requested questions and themes. The use of computers especially Frog VLE in TnL writing essay can help the students producing essays that meet the requirements of SPM exam.
The recommended teaching strategies should be emphasized, so that students able to interact and master the learning skills through their own experiences. The reality is that senior teachers often ignore the use of teaching aids as compared to the teachers who have just been completed teaching practice (Mahzan, 2008). Consequently, this creativity can help the teachers to make the TnL process more fun, cheerful and effective (Abdul Rasid, Martini, & Azhar, 2017).
Research Objectives
Generally, this study is aimed to examine the effectiveness of learning using Frog VLE to improve the writing performance. While, the specific objectives of this study are to: 1) Identify the difference between the pre-test scores and the post-test score of essay writing performance between the experimental group (EG) and the control group (CG). 2) Determine that there is a significant difference in the pre-test and post-test of grasping the main idea in essay writing as a whole between EG students and CG students. 3) Determine that there is a significant difference in the pre-test and post-test of essay content amplification level between EG students and CG students.
Research Hypothesis
Alternative hypotheses were selected for this study to analysing students' performance in essay writing by using Frog VLE. The study was conducted based on six alternative hypotheses as follows: Ha1: There is a significant difference between the mean of the EG pre-test score and the mean of the CG's pre-test score.
Ha2: There is a significant difference between the mean score of EG post-test and the mean of CG post-test score.
Ha3: There is a significant difference between the mean of EG pre-test and CG pre-test in grasping the main idea.
Ha4: There is a significant difference between the mean of EG post-test and CG post-test in grasping the main idea.
Ha5: There is a significant difference between the mean of EG pre-test and CG pre-test in amplifying the essay content Ha6: There is a significant difference between the mean of EG post-test and CG post-test in amplifying the essay content
Research Methodology
This study was conducted to examine the effectiveness of Frog VLE on students' essay writing. The study utilizes a quasi-experimental design in order to determine the effect of dependent variables on independent variables. Quasi-experimental design does not involve random distribution (Noraini, 2010). According to Chua (2016), quasi-experimental design is used in a condition of the experimental design is conducted when the respondent group are not randomly assigned, whereby, these two groups of respondents are unbalanced and have no identical characteristics. Therefore, this study uses the quasi-experimental design because the research samples are based on the existing class without random assignments (Zamri & Abdull Sukor, 2009). This situation occurs as the researcher takes the respondent from the existing class in facilitating the TnL.
The application of pre-test and post-test are to determine the effectiveness of Frog VLE's usage in students' essay writing. Experimental method identifies as the most suitable method to obtain information of independent variable impact on dependent variables. The pre-test is piloted before conducting the study while the post-test is held after several TnL sessions by using Frog VLE and traditional methods. The study sample was chosen by the researcher because the quasi-experimental design is about a non-random group (Chua, 2016).
The study was conducted on form four students at a secondary school in Cheras, Kuala Lumpur. Respondents are among male and female students. It involves 60 students who are among good and average achievers based on the preliminary examination who got B, C and D in Malay language subjects. The respondents were divided into two groups; 30 EG and 30 CG.
The researchers have developed a research instrument to be used known as the pre-test and posttest question as well as essay test questions. The selected test title corresponds to the Malay language syllabus. The essay topics for each teaching session are equal for EG and CG. The researchers have conducted four times teaching sessions for each group. Each teaching session took place within 2 periods of Malay language lesson. The data obtained in the pre-test and post-test were analysing by using SPSS version 21.0.
Respondent Demographics
The researchers have gathered some information regarding respondent's background as references to make a finding justification. As such, the researchers have concluded the findings based on the tables below. The experimental group (EG) consists of 14 (47%) male students and 16 (53%) female students. The control group (CG) consists of 17 (57%) male students and 13 (43%) female students. Ha1: There is a significant difference between the mean of the EG pre-test score and the mean of the CG's pre-test score. Table 1, shows the difference mean score for overall writing essay of EG and CG pre-test. The mean score for EG is 53.90 and SD = 9.33 (N = 30). While KK mean score is 57.23 and SD = 6.30 (N = 30). The comparison of mean between EG and CG shows no significant difference with t value (58) = 1.621, p <0.05. Result shows that a significant value of .111 is greater than the significant level .05. Therefore, there was no significant difference between the mean of the EG pre-test score and the mean of the CG's pre-test score. Thus, the alternative hypothesis was rejected.
Ha2: There is a significant difference between the mean score of EG post-test and the mean of CG post-test score. Table 2 shows the mean score for the overall writing essay of EG and CG in post-test level. Based on the table, the finding shows that mean score of EG is 64.33 and SD = 6.08 (N = 30). While CG mean score is 59.13 and SD = 6.42 (N = 30). The comparison of two mean between EG and CG shows a significant difference with t value (58) = 3.219, p <0.05. Result shows a significant value of .002 is smaller than the significant level .05. Thus, there was a significant difference between the mean score of the EG and CG post-test. As such, the alternative hypothesis was accepted.
The finding clearly showed that the use of traditional methods did not help to increase student scores. Therefore, teachers need to use more appropriate teaching approach besides the traditional teaching to ensure that essay writing skill is improving. It is evident that the teaching using Frog VLE is an effective and suitable method for students. Table 3 above shows the comparing of mean for EG and CG main ideas at the pre-test level. Mean for grasping the main idea for EG is 2.167, meanwhile, the mean of grasping the main idea for CG is 2.133. There is only a small difference in grasping the main ideas for both groups. Table 3 shows that there is no significant difference in grasping of main idea aspects for EG and CG at pre-test level when t (58) = .356, p <0.05. As a result, there was no significant difference between the mean of EG pre-test and CG pre-test in grasping the main idea. Therefore, the alternative hypothesis was rejected.
Ha3: There is a significant difference between the mean of EG pre-test and CG pre-test in grasping the main idea.
Ha4: There is a significant difference between the mean of EG post-test and CG post-test in grasping the main idea. Based on Table 4 above, it shows the comparing of mean for EG and CG main ideas at the post-test level. Mean for grasping the main idea for EG is 2.267, meanwhile, the mean of grasping the main idea for CG is 2.133. There is only a small difference in grasping the main ideas for both groups. Table 4 shows that there is no significant difference in grasping the main idea aspects for EG and CG at posttest level when t (58) = 1.287, p <0.05. So, there was no significant difference between the mean of EG post-test and CG post-test in the main idea. Therefore, the alternative hypothesis was rejected.
The conclusion suggested that the traditional teaching or the use of Frog VLE is just the same as because of using both methods enable the students to grasp the main idea aspects. The sample of this study was good and average students, therefore, the sample did not have the problem of expressing the main idea in Malay language essay writing.
Ha5: There is a significant difference between the mean of EG pre-test and CG pre-test in amplifying the essay content Table 5 above shows the comparison mean analysis of EG and CG in amplifying the content at pretest level. The analysis has shown that EG respondents obtained mean 2.667, while CG respondents obtained mean 3.000 and mean difference was -3333. The t-test analysis indicates significant levels p = <0.05, t (58) = -1.980, p <0.05. The finding has shown that the significant value of mastery aspect of content amplification is .053 greater than the significant level of 0.05. So it is clear that there was no significant difference between the mean of EG pre-test and CG pre-test in amplifying the essay content. Therefore, the alternative hypothesis was rejected. Table 6 above shows the comparison mean analysis of EG and CG in amplifying the content at posttest level. The analysis has shown that EG respondents obtained mean of 3,500, while CG respondents obtained mean of 2.933 and a mean difference of .567. The t-test analysis result showed significant level p = <0.05, t value (58) = 4.264, p <0.05. This finding shows that the significant value of amplification profiency aspects of .000 is smaller than the significant level of 0.05. So it is clear that there was a significant difference in content amplification profiency between EG and CG in post-test. As such, an alternative hypothesis was accepted.
Ha6: There is a significant difference between the mean of EG post-test and CG post-test in amplifying the essay content
The finding showed that the respondents of the experimental group were able to make interesting amplification when using Frog VLE as because the students be able to search various materials from google. The material obtained can stimulate the mind of the students to remember the material with the existing knowledge.
Discussion
Based on the finding of the study, the use of Frog VLE in teaching essay writing part B has a positive effect on the students specifically to the writing performance as a whole. After students were given teaching treatment using Frog VLE in four TnL sessions, there was a significant difference in overall writing performance for post-test. It is evident that TnL by Frog VLE enables students to produce better essay after practicing some exercises. The finding showed that the use of Frog VLE does not really help students in mastering the main idea. It was showing that that there is no significant difference in both pre-test and post-test for both groups. Students are still weak to grasp the main idea. It is hard for student to point out the main idea that coincides with the question's need. Whereas, the finding related to interesting amplification requires the students to have the knowledge and available experiences in essay writing. Not all students are able to produce interesting essays (Gagne, Wager, Golas, Keller, & Russell, 2005). The finding of this study showed that the use of Frog VLE can help students in producing interesting essay amplification. Even though the pre-test did not show any significant different, but post-test has shown that there is a significant different. The students who were using Frog VLE are able to amplify the essay content properly.
This matter has been explained in the study of Abdul Rasid, Shamsudin, & Norzila, (2011) after the teaching of KOMSAS text that carried out over the experimental group, students able to generate ideas for producing good essays, whereas, previously the students not able to amplify the content by expanding and developing the content accordingly.
Conclusion
As a conclusion, essay writing is a very important part in TnL of Malay Language at school. Writing skills can be improved by mastering the essay writing. The use of Frog VLE that has been launched on March 2012 and allowed to access by stages can help the teachers to improve students' proficiency in TnL at school. Teachers need to be creative in adapting the use of Frog VLE in TnL essay writing. It's used has helped the students be competent in amplifying essay content and finding the main idea of essay writing. Indirectly, it can motivate students to produce a quality essay.
Overall, the use of Frog VLE is able to upgrade the performance of students' Malay language essay writing. However, the use of Frog VLE is still not fully utilized due to time constraints and technical problems in their respective schools. The school administrators also seem to give less emphasis to Frog VLE usage among teachers. It is recommended that the use of Frog VLE should be fully implemented by teachers to ensure students' achievement in the future with the strong support from the Ministry of Education Malaysia. | 2019-08-02T22:40:12.268Z | 2018-11-30T00:00:00.000 | {
"year": 2018,
"sha1": "14e4edb356457a0ada3f1cb0905ac96044e52442",
"oa_license": "CCBY",
"oa_url": "https://hrmars.com/papers_submitted/4944/Malay_Language_Learning_Innovation_in_Enhancing_Essay_Writing_Performance.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "fee17cb6a745970b9b7bb0073c7518d3210f7f35",
"s2fieldsofstudy": [
"Education"
],
"extfieldsofstudy": [
"Sociology"
]
} |
35634707 | pes2o/s2orc | v3-fos-license | Mathematics of Domains
Two groups of naturally arising questions in the mathematical theory of domains for denotational semantics are addressed. Domains are equipped with Scott topology and represent data types. Scott continuous functions represent computable functions and form the most popular continuous model of computations. Covariant Logic of Domains: Domains are represented as sets of theories, and Scott continuous functions are represented as input-output inference engines. The questions addressed are: A. What constitutes a subdomain? Do subdomains of a given domain $A$ form a domain? B. Which retractions are finitary? C. What is the essence of generalizations of information systems based on non-reflexive logics? Are these generalizations restricted to continuous domains? Analysis on Domains: D. How to describe Scott topologies via generalized distance functions satisfying the requirement of Scott continuity ("abstract computability")? The answer is that the axiom $\rho (x, x) = 0$ is incompatible with Scott continuity of distance functions. The resulting \bf relaxed metrics are studied. E. Is it possible to obtain Scott continuous relaxed metrics via measures of domain subsets representing positive and negative information about domain elements? The positive answer is obtained via the discovery of the novel class of co-continuous valuations on the systems of Scott open sets. Some of these natural questions were studied earlier. However, in each case a novel approach is presented, and the answers are supplied with much more compelling and clear justifications, than were known before.
Preface
This text represents the results of my research in mathematics of approximation domains equipped with Scott topology. I would be the first to classify this research as belonging to the field of pure mathematics, yet it is being presented as a dissertation in computer science. I think, this demands some explanation.
Accepted Practice
There is a widely recognized body of computer science applications of the mathematical theory of domains, reflected in dozens of monographs, and hundreds, if not thousands, of research papers (some of the related references are in the text). Some of the most prominent contributors to this theory work at departments of computer science, rather than at departments of mathematics. Due to these reasons, it is quite often that purely mathematical research in this field is presented as doctoral dissertations in computer science, even in the leading schools.
For example, a famous work by Kim Wagner [57] defended as a PhD Thesis in the School of Computer Science of Carnegie Mellon University, clearly belongs to this class of research. As my main achievements to this moment are in mathematical theory, and not in its applications, I believe it appropriate to follow this practice. v
Recongnized Applications
The main recognized application of domains is to serve as denotational models of programming languages. This application is due to Christofer Strachey, who introduced the idea of denotational semantics, and Dana Scott, who invented approximation domains equipped with what is now called Scott topology and constructed extensional models of lambda-calculus by solving reflexive domain equations like D ∼ = [D → D]. The classical textbook describing this class of applications is [53], and there are many successors. It is widely argued that denotational descriptions of programming languages using approximation domains should be considered canonical definitions of those languages, while other types of descriptions should be considered as derivative. This viewpoint obviously has its share of opponents too.
Among more hands-on applications of this approach are verifications of compiler correctness, uses of abstract interpretation to perform some static analysis of programs, attempts to produce logical frameworks based on denotational semantics, etc.
Another important application, which emerged in the nineties, is to use domains to produce computational models for classical mathematical structures and thus to provide the framework for computations in such structures. This direction of research was originated by Abbas Edalat, who used domains to generalize the notion of Riemann integral to fractal spaces and obtained more efficient methods to compute integrals on such spaces (see [19] and references therein). The research group at Imperial College headed by Edalat is now working towards constructing systematic models of this kind for a wide class of mathematical structures (see [20] for the review of these efforts).
Potential Applications
Many researchers in this field probably have their own lists of prospective applications, which often do not reduce to the two classes of applications described before. My own list includes the following.
When denotational semantics was first introduced, it was suggested that the vi canonical approach would be not to produce denotational description of existing languages, but to design languages fitting the denotational models of interest. While this approach has not received sufficient development, I believe that the potential for such development exists. Thus Problem A is quite crucial for the further applicability of this field.
Simplicity of Presentation and Prerequsites
I make an emphasis on simplicity and intuitive clarity of presentation. The techniques here are deliberately elementary and non-categorical, except for those rare cases when categorical methods are unavoidable.
The Introduction can be used as a brief tutorial for the readers, who are not familiar with the field. The Introduction sacrifices the completeness of presentation of the standard information on domains for the accessibility.
Continuity in Computations and Scott Domains
Here we present evidence of the continuous nature of computations [10].
Continuity in Software Engineering
It seems appropriate to start with a real life observation before dealing with mathematical models. Virtually all software products used today contain a number of bugs, which lead to refusal to work, crashes, and incorrect processing of data. Yet for a usable software product these unpleasant effects are observed relatively infrequently, as most of the time the product works satisfactory. What we should say here is that the real product is sufficiently close to the ideal product, close enough to be used in lieu of the unavailable ideal product. This is a phenomenon of continuity. This continuity results from the process of testing and amending the software and is crucial for our ability to use any software at all. Moreover, it is quite possible that the ideal product cannot exist at all (e.g. if the specifications are for an undecidable problem), yet we still can consider models which contain representatives for such ideal products and for their reallife approximations.
The author so far has not encountered studies dealing with this phenomenon mathematically. Software reliability models do not seem to qualify, because they always abstract from the nature of the particular software involved and just study the dynamics of error rates. We would like to see the difference between the ideal software and its real-life approximation expressed semantically, error classification and rates being built on top of such semantical structure. We think that such approach to software design is potentially more promising than the concept of fully verified software.
Continuity in Constructive Mathematics
Another evidence of continuity of computations comes from effective versions of mathematical analysis (constructive, recursive, or intuitionistic analysis -see, for example [6,25]). All constructive functions from constructive real numbers to constructive real numbers must be continuous. It is impossible to properly treat computations at the points of discontinuity. If one has to deal with non-continuous functions, they have to be partially defined -points of discontinuity cannot belong to their domain (although see [18]).
Continuity in Denotational Semantics of Programming Languages
The most important and mathematically rich example of a continuous approach to computation is denotational semantics (see, for example, [44,46,48,53,59,2] We assume that the reader is familiar with the notions of partially ordered set (partial order), topological space, and continuous function. In some rare cases we will use simple categorical notions, but the reader is able to skip them without excessive harm to the overall comprehension.
Notation
We
Directed Sets
A partially ordered set (poset), (S, ⊑), is directed if ∀x, y ∈ S. ∃z ∈ S. x ⊑ z, y ⊑ z. In particular, empty sets are directed.
The notion of directed set should be considered as a generalization of the notion of increasing sequence, s 1 ⊑ s 2 ⊑ . . ..
Directed Complete Partial Orders (DCPO's)
A poset, (A, ⊑), is a directed complete partial order or dcpo if for any directed S ⊆ A, the least upper bound ⊔S of S exists in A.
Note that, since the empty set is directed, any dcpo A must have the least element, which we denote as ⊥ or ⊥ A .
Because domains without least elements are also considered lately, differences in terminology occur at this point. Sometimes what we call directed complete partial orders is called complete partial orders, and a larger class of partial orders, where only non-empty directed sets must have least upper bounds and, thus, the least element does not have to exist, is called directed complete partial orders.
3.2.3
The "Way Below" Relation and Compact (Finite) Elements Consider a dcpo (A, ⊑) and x, y ∈ A. We say that x ≪ y (x is way below y) if for any directed set S ⊆ A, y ⊑ ⊔S ⇒ ∃s ∈ S. x ⊑ s.
An element x, such that x ≪ x, is called compact or finite.
It is easy to see, that x ≪ y implies x ⊑ y, however the inverse relationship is, in general, much more complicated. For example, it is possible that x ≪ x, but it is also possible that x ❁ y and x ≪ y at the same time. A kind of transitivity,
Continuous and Algebraic DCPO's
Consider a dcpo (A, ⊑) and K ⊆ A. We say that a dcpo A is a continuous dcpo with basis K, if for any a ∈ A, the set K a = {k ∈ K | k ≪ a} is directed and a = ⊔K a . We call elements of K basic elements.
If a continuous dcpo has a basis consisting only of compact elements, such a dcpo is called algebraic.
Bounded Completeness and Domains
We say If an algebraic dcpo is bounded complete, it is called an algebraic Scott domain or, simply, a Scott domain.
If a continuous dcpo is bounded complete, it is called a continuous Scott domain [28].
To be consistent with these definitions, if a dcpo is bounded complete, we call it a complete Scott domain.
Lattices
If each subset of a partially ordered set A has the least upper bound in A, A is called a complete lattice.
A complete lattice which is a continuous dcpo is called a continuous lattice.
A complete lattice which is an algebraic dcpo is called an algebraic lattice.
Note that complete lattices are precisely those complete Scott domains, which have the top element, ⊤. Also note that any subset of a complete Scott domain A also has the greatest lower bound in A.
Aleksandrov and Scott Topologies
It is easy to see that Aleksandrov open sets and Scott open sets form topologies which are called, respectively, the Aleksandrov topology and the Scott topology.
Continuous functions
Consider dcpo's (A, ⊑ A ) and (B, ⊑ B ) with the respective Aleksandrov topologies. It it easy to see that a function f :
Functional Spaces
Consider dcpo's A and B. We define the functional space [A → B] as the set of all Scott It is easy to check that [A → B] is a dcpo.
Retractions, Projections, And Their Pairs
Consider dcpo A and its Scott continuous transformation, f : If r is also a projection, then the terms projection onto, and embedding-projection pair (or simply, projection pair) are used in the corresponding situations.
Interval Numbers
In this chapter we will look at several important domains, including interval numbers, which form the basis for correct approximate computations with real numbers. with the same order.
Vertical Segments and Rays of Real Line
Note that the last two spaces must contain +∞ for directed completeness, and that the last space must also contain −∞, because all domains have the least element.
We will call these domains the domains of lower estimates for reasons, which will become apparent in the next section. We denote these domains as
The "Way Below" Relation
One can easily see that It is also easy to see that given This allows one to understand, why these domains are continuous but not algebraic.
Any subset K, dense in [A, B] in the traditional sense, can be taken as a basis.
Interval Numbers
In this section we define the domain of interval numbers belonging to [A, B], where the cases of A = −∞ and/or B = +∞ are included. This domain will consist of segments [a, b], where A ≤ a ≤ b ≤ B. We denote this domain as R I or R I [A, B] . Informally speaking, the segment [a, b] is interpreted as a partially defined number x, about which it is known that We can consider a product domain R + × R − with coordinate-wise order. Then R I can be defined as a subset of R + × R − , obtained by by elimination of pairs a, b , gives more information about a number, than [a, b]. In such a case, we say that [a, b] approximates [c, d], and write [a, b] We flip this triangle so that the best defined interval numbers, like [a, a], are on the top, and the least defined one, [0, B], is on the bottom.
The "Way Below" Relation
One can easily see that It also easy to see that [a, b] It is easy to see that R I is a continuous Scott domain. Consider any set Q, dense in [A, B] in the traditional sense, and take elements [A, b], [a, B], and [a, b], such that a < b and a, b ∈ Q, as a basis K ⊆ R I .
Scott topology
We consider the basis K built above and build the base of Scott topology. For
Algebraic Information Systems and Domains
This chapter presents the approach of information systems in the case of algebraic Scott domains as can be found, e.g. in [48,35,36].
The algebraic information systems and domains described in this chapter reflect the standard notions of inference and theories in traditional formal theories.
Information Systems and Domains
The approach of information systems describes an approximation domain as a set of theories in a logical calculus. There are two ways to follow this approach. One could consider a given approximation domain, which obeys certain specific axioms, and then try to build a logical calculus such that its theories form an isomorphic domain. This approach is very useful, when one needs to investigate which class of domains is covered by a specific variant of information systems, or when one starts with a given domain to begin with.
However, for didactic purposes, a different approach has proven much more valuable. This approach was used in the first key paper on information systems by Dana Scott [48] and works as follows. We presume that there is some approximation domain on the background, but we do not specify it precisely. Then, having this hypothetical domain in mind, we reason about the desired properties of our logical systems and impose appropriate axioms describing the behavior of these systems. Then we define domains as sets of theories and establish their properties as theorems rather than postulating them as axioms. The reason for this distinction is that we view a domain element, x, as a dynamic object, about which only some of the information is known at any given time of the computational process, but which later can be supplemented with more information, and thus replaced with y, x ⊑ y. The information, which was unknown about x,
UNKNOWN as a Truth Value
can become true about y. This dynamic process is, however, viewed as monotonic with respect to time -once a piece of information becomes true about an element, it remains this way further on.
It is possible to talk about false pieces of information about x -these are such pieces of information which are not true about any y, such that x ⊑ y, that is, the pieces which cannot become true about x during its arbitrary monotonic evolution. The false truth values will be treated via consistency mechanism, however they will always remain auxiliary, and usually can be easily eliminated from the scene if necessary, while true and unknown truth values are essential.
Example: Formal Theories
The most natural example one should keep in mind is any traditional formal theory, where elementary statements are all formulas, and the domain in question is the domain of all theories ordered by ordinary set-theoretic inclusion.
If a statement belongs to a theory, we will say that it is true in (or, if you wish, "about") this theory, otherwise it is unknown in this theory. If the theory cannot be refined to a larger non-contradictory theory to include a specific statement, we might wish to say that this statement is false in ("about") this theory.
In the traditional formal theories any deduction is thought of as a formal text of finite length. Hence any entailment of a statement or a contradiction from a set of statements is made on the basis of a finite subset of this set. This finitarity property will be reflected in the definitions below.
Consistency and Entailment
Let us denote the set of elementary statements as D, and the set of all its finite subsets as P fin (D). We interpret a finite subset, u ⊆ D, u = {d 1 , · · · , d n }, as a finite conjuction of statements d 1 , · · · , d n .
Keeping the hypothetical approximation domain at the background, we would like to call a finite conjuction, u, of elementary statements consistent, if there is a domain element, x, about which all statements of u are true.
It is traditional to assume that ∅ is consistent, that is, that the domain in question is non-empty [48]. We do follow this tradition here. It is also traditional to assume that any single elementary statement is true about some domain element, that is, {d} is consistent for any elementary statement d. This means that one considers contradictory elementary statements to be "junk" and wants to exclude them from an information system. We would like to depart from this convention, as it seems to gain nothing, and makes it more difficult to talk about some natural examples, like the one considered in the previous subsection, and also makes it impossible to talk about effective structures on domains in full generality (the point of view, advocated, in particular, by V.Yu.Sazonov; our style of information system is, in fact, intermediate between the traditional one and the style of Sazonov [43] and seems to be the most convenient).
We would like to say, that a finite conjuction, u, of elementary statements entails an elementary statement, d, if whenever u is true about a domain element, x, statement d is also true about x.
Again, it is traditional to consider only consistent conjuctions, u, in the context of entailment. However, it is much more convenient to assume that a conjuction, which is not consistent, entails everything. It is technically convenient to introduce a special false statement, ∇, which entails everything and follows from all inconsistent conjunctions.
Taking all this into account, the following definition becomes quite natural.
The equivalence between this and the classical definition of information system will be shown in Subsection 5.1.7.
Theories as Domain Elements
Definition 5.1.2 Domain |A| associated with an algebraic information system A is the set of theories, Informally, we will say that if d ∈ x, then d is true about x, otherwise d is unknown about x.
Properties of Domains
Assume that algebraic information system A is given. Domain |A| is partially ordered by set-theoretical inclusion, ⊑ A =⊆ DA . We study the properties of this order. Proof. This simple proof is still quite instructive, so we go through it here.
Deductive Closure and the Least Element
First of all, we want to show that x is deductively closed, that is if u ⊆ x and u ⊢ A v Formula u ⊢ A v implies finiteness of u and v, hence consider u = {d 1 , . . . , d n }.
Consider u ′ = u 1 ∪ . . . ∪ u n . Then, combining the rule of conjuction elimination and the rule of transitivity, we obtain u ′ ⊢ A {d 1 }, . . . , u ′ ⊢ A {d n }, applying the same combination of rules in their conjuction introduction and reflexivity incarnations once again, we obtain u ′ ⊢ A u, applying transitivity once again we obtain u ′ ⊢ A v, and from this and u ′ ⊆ x one concludes that v ⊆ x.
Consistency follows straightforwardly from the two previous definitions. ✷ Assume that S = {s i | i ∈ I} is a non-empty directed set of elements of domain |A|. Let us establish that i∈I s i is an element of |A|. From this it would be easy to see that ⊔ i∈I s i = i∈I s i .
First, establish the deductive closeness of i∈I s i . Assume that u = {d 1 , . . . , d n } ⊆ i∈I s i . Then d 1 ∈ s i1 , . . . , d n ∈ s in . By directness of (s i , i ∈ I), there is such i ∈ I, that s i1 ⊆ s i , . . . , s in ⊆ s i . Then u ⊆ s i , and by deductive closeness of Consistency is trivial. Proof. Assume, that x = {d 1 , . . . , d n } and that x ⊑ ⊔S, where S is a directed set. By the proof of Lemma 5.1.4, ⊔S = s∈S s, hence, since d 1 , . . . , d n ∈ x, there are s 1 , ..., s n ∈ S, such that d 1 ∈ s 1 , . . . , d n ∈ s n . Hence, by directness of S, there is s ∈ S, such that d 1 ∈ s, . . . , d n ∈ s. Hence, by the deductive closeness of s, x ⊆ s, and x ⊑ s.
Conversely, assume that x cannot be represented as a deductive closure of its finite subset. Then consider set U = {u | u ⊆ x, u finite}. It is easy to see, that S = {u | u ∈ U } is a directed set, and that x = ⊔S, but for no u ∈ U , x ⊑ u, hence x is not compact. ✷ The second part of the proof of the previous Lemma implies that the set of compact elements characterized there possesses the properties of a basis of dcpo and, hence, the following Lemma holds.
Lemma 5.1.6 Domain |A| associated with algebraic information system A is an algebraic dcpo.
Bounded Completeness
Lemma 5.1.7 Domain |A| associated with algebraic information system A is bounded complete.
Proof. Consider Y ⊆ |A|, such that there is x ∈ |A|, such that for all y ∈ Y , y ⊑ x. Then y∈Y y is a consistent set, and y∈Y y is the desired least upper bound of Y . ✷
Algebraic Scott Domains
The discource of this section can be summarized by the following Theorem.
Representing Algebraic Scott Domains
Here we show that algebraic information systems describe exactly algebraic Scott domains.
Theorem 5.1.2 For any algebraic Scott domain (X, ⊑ X ), there is an algebraic information system A, such that the partial order (|A|, ⊑ A ) is isomorphic to partial order (X, ⊑ X ).
Proof. Denote the set of compact elements of X as K. Take where ∇ A is a new token.
For finite subsets u, v ⊆ D A , we say that 2. or u has an upper bound in X, and for any k ∈ v, k ⊑ X ⊔ X u.
First, we must prove that A is an algebraic information system. It is a simple proof, we only show transitivity axiom here, which is only non-trivial if u has an upper bound in X. Then assumption, that u ⊢ A v 1 , . . . , u ⊢ A v n , means that any element k of any of these v i 's is less or equal to ⊔ X u. In turn, v 1 ∪ . . . ∪ v n ⊢ A w, implies that all elements of w are under the least upper bound of all those k's. But ⊔ X u is (some) upper bound of those k's, hence all elements of w are under ⊔ X u.
Now we have to establish an order-preserving isomorphism between X and |A|.
With every x ∈ X we associate K x = {k ∈ K | k ⊑ x}. Now we are going to prove that all K x are elements of |A|, and every element of |A| equals to K x for some x ∈ X. Then, due to the fact that x ⊑ y ⇔ K x ⊆ K y , the relation x ↔ K x would yield the necessary isomorphism.
It is easy to establish that K x is an element, by using the directness of K x .
In the opposite direction, consider S ⊆ K, which is an element of |A|. First of all, let us show that S is directed. By consistency of S any finite u ⊆ S is bounded.
By bounded completeness of X, there is ⊔ X u. Then, from the directness of K ⊔X u , we immediately obtain the existence of compact k ∈ K ⊔X u , which is an upper bound of u (of course, this immediately implies k = ⊔ X u, i.e. the least upper bound of a finite bounded set of compact elements is compact). By definition of ⊢ A , u ⊢ A {k}. Hence, the deductive closeness of S implies that S is directed. Now, using the fact that X is directed complete, consider x ∈ X, such that x = ⊔ X S. Let us prove that S = K x . The only non-trivial direction is to prove K x ⊆ S.
Consider compact element k ∈ K x . The definition of "way-below" relation and of element x, together with the facts k ≪ k and k ⊑ x, imply that there is k ′ ∈ S, such that k ⊑ k ′ .
Then deductive closeness of S implies k ∈ S. ✷
Equivalence between Classical Information Systems and
Our Definition , is called a classical information system if (i) ∅ ∈ Con ′ A (non-degeneracy; empty domains are barred); The domain |A| associated with a classical information system A is defined as a set of x ∈ D ′ A satisfying the consistency condition, ∀u ⊆ x. u is finite ⇒ u ∈ Con ′ A , and the condition of deductive closeness, ∀u ⊆ Of course, ⊑ |A| is defined to equal ⊆.
All such domains are algebraic Scott domains, and for any algebraic Scott domain X it is possible to build a classical information system A, such that |A| is isomorphic to X. So everything is very similar to our version of algebraic information systems. Now we are going to establish an even stronger equivalence between these two versions. We are going to build two translations as follows. Given an algebraic informa- and given a classical information system, we will build an algebraic information system A = S(B), such that the following properties hold.
obtained by restricting ⊢ A on P fin (D Ac ) × P fin (D Ac ). Basically, the only disturbance that S • C cannot avoid is that other "junk" tokens equivalent to ∇ A die.
The translations are defined by the following formulas.
It is easy to check that if A is an algebraic information system and B is a classical information system, then C(A) is a classical information system, S(B) is an algebraic information system, and our claims above hold.
Approximable Mappings and Continuous Functions
Information systems allow to think about domain elements as theories. Likewise, this approach allows to think about Scott continuous functions as special inference engines, which infer information about output, f (x), from information about input, x.
Unfortunately, these input-output inference engines are called approximable mappings for their property to approximate one another. While this is an important property, Scott continuous functions thought of as graphs, or, in fact, any functions to domains approximate one another. Moreover, the "mappings" in question are not even functions.
However, at this point of the development of the field one does not have much of a choice, but to follow the accepted terminology.
The following definition takes into account the intuition, that consistent information about input should produce consistent information about output, and that "native" inference relations of input and output domains can be used by an input-output inference engine.
Definition 5.2.1 Given two information systems, A and B, relation f ⊆ P fin (D A ) × P fin (D B ) is called an input-output inference relation or an approximable mapping between A and B, if the following axioms hold: inconsistency about an input allows to infer anything about the corresponding output);
Scott Continuous Functions
An approximable mapping, f , between information systems A and B naturally gives rise to a function, |f | : |A| → |B|, where |f |(x) is computed by inferring all information from x using f .
We will see that |f | is always Scott continuous, and that for any Scott continuous function g : |A| → |B| one can find an approximable mapping g between A and B, such that | g| = g, |f | = f . Together with the results about identity maps and composition this yields an equivalence between the category of algebraic information systems and approximable mappings and the category of algebraic Scott domains and Scott continuous functions.
Definition 5.2.2
Given an approximable mapping f between A and B, define the as- Proof. First, of all, to prove correctness we need to show, that |f |(x) ∈ |B|.
We need to show deductive closeness and consistency of |f |(x). Assume, that v = Then, by axioms of information systems and approximable mappings, uf v, where u = Then there are s 1 , . . . , s n ∈ S, such that d 1 ∈ s 1 , . . . , d n ∈ s n . Then, by directness of S, we can find such s ∈ S, that v ⊆ s. Hence d ∈ |f |(s) and belongs to the union in question. ✷ Definition 5.2.3 Given a Scott continuous function g : |A| → |B|, define the associated It is easy to check that g is an approximable mapping.
This and the Scott continuity of g allow to establish easily, that | g| = g, |f | = f .
Identity and Composition
It is easy to see that One can think of g • f as an input-output inference engine obtained by hooking the input of engine g to the output of engine f : It is easy to check that |g • f | = |g| • |f |.
Functional Spaces
Consider algebraic information systems A and B. We are going to define an information | would consist precisely of all approximable mappings between A and B.
Since there is an isomorphism between approximable mappings and Scott continuous functions, and since the set-theoretical inclusion of approximable mappings corresponds to the partial order on the domain of Scott continuous functions, [|A| → |B|], we will use the information system [A → B] to represent this functional space. Take We leave the necessary correctness checks to the reader.
Effective Domains and Computations
We say that an algebraic information system A is effective, if D A and ⊢ A are recursively enumerable. The corresponding domain |A| is called effective too.
An important example is produced by the effective domain of arithmetic theories in any of the usual systems of arithmetic. This example shows why our degree of generality is the right one.
Usually people give a more restrictive definition, which is equivalent to D A and ⊢ A being recursive. One of the reasons for this is the inconvenience of the definition of a classical information system. Indeed, if A is an effective algebraic information system, then Con ′ C(A) is co-recursively enumerable, and ⊢ ′ C(A) has to be defined effectively on a co-recursively enumerable domain of definition, which is not a trivial undertaking, and, in any case, the result would be awkward.
An element x ∈ |A| is called computable if it is a recursively enumerable set.
Note that this condition is equivalent to the recursive enumerability of K x under our construction from Section 5.1.6. Non-reflexive Logics for
Non-algebraic Domains
In the previous chapter we saw that information systems based on the ordinary logic correspond to algebraic Scott domains. Hence, in order to generalize the logical approach to larger classes of dcpo's, one has to modify the logic of inference and/or the notion of theory.
In this chapter we concern ourselves with bounded complete dcpo's. Some progress for some classes of non-bounded complete spaces was achieved by Abramsky [1].
Non-reflexive Logic
The seminal paper [28] by R.Hoofman generalized the logical approach to non-algebraic continuous domains. Hoofman replaced the ordinary relexivity rule, A ⊢ A, with a weaker property of inverse modus ponens: His paper is a well-written one, but at the same time the actual and simple reasons for his construction to work are hidden in its later sections. Thus, the average reader gets an impression that it is a miracle construction, and the real intuition behind this paper is lost.
In particular, another important feature, which allows to exploit non-relexivity, is underemphasized. This feature is the change in the notion of a theory (i.e. domain element). A theory, x, is traditionally a consistent, deductively closed set of statements.
Hoofman adds an additional requirement of inverse deductive closeness, which is trivial We will soon see, that this feature is essential for his approach, while the inverse modus ponens rule is important only in order to maintain continuity of the resulting domain. In fact, we generalize the results of [28] to arbitrary spaces of fixed points of Scott continuous transformations of algebraic Scott domains, by omitting both the inverse and the standard rule of modus ponens.
Functions and Inference Rules
Scott continuous functions are equivalent to input-output inference engines (approximable mappings). In particular, given an algebraic information system A, Scott continuous transformations |f | of the corresponding domain |A| are equivalent to the generalized inferences f ⊆ P fin (D A ) × P fin (D A ), which are transitive with the standard ⊢ A and respect consistency and inconsistency.
Transitivity
It is easy to see, that ∀u, v, w ∈ P fin (D A ). uf v, vf w ⇒ uf w is equivalent to the condition |f | • |f | ⊑ [A→A] |f |.
Inverse Transitivity
It is easy to see, that the Hoofman rule of inverse transitivity, ∀u, w ∈ P fin (D A ). uf w ⇒ ∃v ∈ P fin (D A ). uf v, vf w is equivalent to |f | ⊑ [A→A] |f | • |f |.
Retractions
The previous two paragraphs imply that f is transitive and inversely transtive if and only if |f | is a retraction: |f | = |f | • |f |.
Inverse Reflexivity
The rule ∀x ∈ |A|. |f |(x) ⊑ A x is equivalent to the following rule of "inverse reflexivity": uf v ⇒ u ⊢ v, which will be transformed into uf v ⇒ v ⊆ u in some simple cases.
Closures and Projections
It is easy to see, that reflexivity implies inverse transitivity, and that inverse reflexivity implies transitivity.
Hence, the combination of reflexivity and transitivity yields precisely closures (such retractions |f |, that x ⊑ A |f |(x)), and the combination of inverse reflexivity and inverse transitivity yields precisely projections (such retractions |f |, that |f |(x) ⊑ A x).
Fixed Points as Theories
Consider the definition of |f |: |f |(x) = {d | ∃u ∈ x. uf {d}}. So |f |(x) consists of tokens which are inferrable from x via inference engine f .
Hence, the deductive closeness of x with respect to inference f , ∀u, v ∈ P fin (D A ).
Similarly, the inverse deductive closeness of x with respect to inference f , namely Hence, together the deductive closeness and the inverse deductive closeness of x with respect to inference f is equivalent to x being a fixed point of |f |: |f |(x) = x.
Domains of Fixed Points
The previous paragraph suggests the following procedure. Consider an algebraic information system A and replace its entailment relation with an approximable mapping f between A and A.
Then replace the notion of theory with the consistent subset x ⊆ D A , such that x is deductively closed and inversely deductively closed with respect to f . The result is We can call |A f | a fixed-point subdomain of |A|.
However, we want to introduce a more general notion of a fixed-point subdomain.
Adding the Top Element
Consider an algebraic information system A = (D A , ∇ A , ⊢ A ) and the corresponding domain |A|. Consider an element ∇ A ⊤ ∈ D A and define the algebraic information system A ⊤ as follows. Then
General Notion of Fixed-Point Subdomain
In this chapter we consider approximable mappings as generalized entailment relations.
Sometimes we would like an approximable mapping f to entail the contradiction from some of finite conjuctions of statements from D A , noncontradictory under ⊢ A . In order to formalize such a situation we have to consider approximable mappings f from A to The general notion of a fixed-point subdomain of A is the set of fixed points of If, for the sake of tradition, one wants to impose the requirement that the subdomains are non-empty, one has to add the requirement |f |(⊥) = ⊤.
Application: Removing the Compact Top Element
Assume that an algebraic Scott domain |B| has the compact top element Then The compactness of ⊤ B is important, since a Scott continuous function on an algebraic Scott domain is completly defined by its values on compact elements.
Powersets and Qualitative Domains
Consider set D, such that ∇ A ∈ D, and define If we consider W ⊂ P fin (D A ), such that ∅ ∈ W , {∇ A } ∈ W , and modify the previous construction so that u ⊢ A v iff v ⊆ u or ∃w ∈ W. w ⊆ u, then we obtain a qualitative domain (see [28]). We can modify the construction of cutting the compact top element above to obtain any qualitative domain from the powerset domain.
General Notion of Finitary Information System And Finitary Scott Domain
Consider an algebraic information system A = (D A , ∇ A , ⊢ A ) and an approximable mapping f from A to A ⊤ . We agreed that f describes a fixed-point subdomain of |A|. Observe that f also describes a fixed-point subdomain of |A m |. Moreover, the sets of fixed points of corresponding Scott continuous functions |f | : |A| → |A ⊤ | and |f | : This leads us to a general definition of what we call a finitary information system and the definition of the corresponding finitary Scott domain. We will respect the rule |f |(⊥) = ⊤ meaning the non-emptyness of the resulting domains. These definitions will be respectively a streamlined equivalent of A m together with an approximable mapping f between A m and A m ⊤ and a description of the set of fixed points of the corresponding function |f |.
a calculus must admit at least one non-contradictory theory); In order to check the correctness of our discourse, one has to consider information systems A m = (D A , ∇ A , ⊢ m A ) and the corresponding system A m ⊤ . Given a finitary infor-
Domains of Fixed-Point Subdomains
In the subsections 6.1.5 and 6.1.8 we defined a finitary information system and the corresponding finitary Scott domain of fixed points of |f | for any approximable mapping f , such that ¬(∅f {∇ A ⊤ }), between any algebraic information system A = (D A , ∇ A , ⊢ A ) and the corresponding system A ⊤ .
The approximable mapping f , such that ∅f {∇ A ⊤ }, corresponding to Scott continuous function |f |, such that |f |(⊥) = ⊤ A ⊤ , is the top compact element of the domain [A → A ⊤ ], and, hence, it can be removed by the technique described above, yielding the Because these mappings f are in one-to-one correspondence with fixed-point subdomains of |A|, we can take the domain |[A → A ⊤ ] ⊤ | as representing fixed-point subdomains of |A|. Of course, different f might have the same set of fixed points, and hence the corresponding fixed-point domains will coincide as sets, but their underlying entailment relations will differ, so they should be considered different subdomains.
Hence we call |[A → A ⊤ ] ⊤ | the domain of fixed-point subdomains of |A|.
Discussion
One should observe that f is not an arbitrary new entailment relation, but is closely related to the original ⊢ A , namely f is transitive with respect to ⊢ A due to the axioms of approximable mappings. This property is responsible for the fact that all elements of the resulting fixed-point subdomain belong to the original domain.
We can think about a fixed-point subdomain as the result of some elements of the original domain being destroyed. There are three mechanisms of such a destruction.
An element can lose its deductive closeness or consistency. When these two mechanisms are involved, we still remain within the realm of reflexive logic and algebraic subdomains result. This important case is covered in details in Chapter 7. differ. Hence this alternative definition is unsatisfactory.
Results for Continuous Scott Domains
Now the results on continuous information systems by Hoofman can be easily explained.
He retained transitivity of f and replaced its reflexivity by inverse transitivity. This resulted in |f | being a retraction.
It is well known that retractions of continuous lattices are continuous lattices, and that all continuous lattices can be obtained as retractions of powersets, and these results can be generalized for continuous Scott
Infinitary Logic
In [43] Sazonov and Sviridenko introduced another generalization of logic. They kept reflexivity and transitivity, but removed the finitarity requirement that an infinite set of statements x only infers those statements, which the finite subsets u ⊆ x infer.
Their approach describes exactly all bounded complete partial orders. Directed completeness is equivalent to the following requirement of partial finitarity: a set of statements x cannot infer the contradiction, unless one of its finite subsets u ⊆ x infers the contradiction.
Hence partially finitary systems of Sazonov and Sviridenko exactly correspond to complete Scott domains.
Sazonov and Sviridenko also gave characterizations of some subclasses of complete Scott domains in their system, in particular, they described a class of their logic corresponding to continuous Scott domains.
The Class of Finitary Scott Domains
What is the class of domains described as fixed-point domains is an important open question. It is well known that the set of fixed points of a Scott continuous transformation of a complete lattice is a complete lattice. It was traditionally thought that the converse is also true, namely that any complete lattice can be obtained as a set of fixed points of a Scott continuous transformation of a sufficiently large powerset. For example, Exercise 18.4.3(ii)(1) on page 491 of the famous textbook on lambda calculus by Barendregt [5] asks to establish this for the case of countable bases of the respective Scott topologies.
Hence we expected that our approach would allow us to obtain all complete Scott domains. We thus expected to follow the terminology of continuous information systems by Hoofman and call the fixed-point information systems developed in this chapter complete information systems.
However, our analysis of literature and numerous conversations with experts in the field convinced us that the problem stated in the textbook by Barendregt
Important Subclasses
What subclasses of finitary domains will be obtained, if we require f to be transitive or, more strongly, inversely reflexive? The conjecture is that we still obtain all finitary Scott domains as sets of fixed points of the corresponding functions |f |.
Transitivity is, of course, a very desirable property of a logical system, so it is of interest to check whether it restricts generality. These questions come in at least two flavors: for f defining a continuous transformation of |A m | and for f defining a continuous transformation of an arbitrary algebraic Scott domain |A|.
Approximable Mappings And Other Issues for Finitary Information Systems
Hoofman successfully introduced the notion of continuous approximable mapping for continuous information systems based on the fact, that given spaces A and B and their Unfortunately, no such simple solution seems possible for finitary information systems. Yet, a satisfactory notion of finitary approximable mapping seems to be necessary for further successful studies of finitary information systems.
There is a possibly simpler variant of this problem if f is required to be transitive.
Possible Translation between Nonreflexive and Nonfinitary Logics
In [43] Sazonov and Sviridenko gave a translation between their logic and the logic of Hoofman for continuous Scott domains.
Naturally there is a question whether such a translation is possible for a larger class of domains.
It also might be of interest to combine the features of nonreflexive and nonfinitary logic in one system.
Subdomains for the Algebraic Case
This chapter represents our results obtained in 1986-1988 and presented in [7]. With the introduction of our new version of the definition of algebraic information system and our ideas in nonreflexive logic the presentation is greatly simplified.
For the duration of this chapter "domain" means algebraic Scott domain, "subdomain" means algebraic Scott subdomain, and "information system" means algebraic information system.
The Brief History of the Question
In this chapter, we study the question "Which subsets of a domain should be considered subdomains?" for the domains described by algebraic information systems. One criterion for a good answer is a requirement that all subdomains of domain |A| should themselves form a domain, |Sub(A)|. We would also like to be able to solve domain equations such as |X| ∼ = |Sub(X)| + . . .. It will be shown in the next few pages that the framework of domains as abstract cpo's does not provide sufficient hints for the solution, but information systems do.
The systematic use of the operation of taking a certain subset of a domain to form another domain and the systematic consideration of domains, whose elements are domains, first appear in the mid-'70's [47,49]. It is notable that nobody has tried to consider arbitrary subsets of domain |A| that would satisfy a particular version of domain axioms, although at the first glance this would seem to be the most intuitive and general version of a subdomain definition in the framework of domains as abstract cpo's. The reason why nobody has considered such a definition is that such arbitrary subsets might satisfy the domain axioms for random causes, and their collection would be totally unmanageable; in particular, one cannot hope to form a domain, |Sub(A)|, with these subsets as elements.
The next idea, which originated by Dana Scott in [47], is to define a subdomain as a set of fixed points of a retraction belonging to a certain class. There are two problems here: a) What are the reasons for using retractions? b) What is the appropriate class of retractions? None of these problems receives a clear answer in the framework of domains as abstract cpo's. On the use of retractions Scott writes: "it seems almost an accident that the idea [to use the retractions] can be applied" (see [47], p.540).
For the algebraic case, Scott suggests using closure operations, that is, retractions |r|, such that |r| ⊒ id. This suggestion can be motivated by the theorem (see [47], In the paper entitled "Data Types as Objects" [49], Shamir and Wadge suggest that to describe the systems of polymorphic types, one should consider data types that incorporate their own subtypes as elements, and allow an element of a data type to belong simultaneously to different subtypes of this data type. Let us briefly present the main features of subtypes in [49]. The subtype |Y | = {x ∈ |D| | x ⊑ y} is associated with each element y of a data type |D|. We can view such subtype as the set of fixed points of the (finitary, in the algebraic case) projection x → x ⊓ y. Subtypes are ordered simply by inclusion. Thus, this simple approach yields the result that |D| ∼ = |Sub(D)| for every domain |D|. Each element represents the subtype of its approximations, and an element x belongs to all subtypes |Y | of |D| that are represented by y's, such that x ⊑ y.
At the end of the Introduction to [7], we discussed the rather adverse relations between retraction-based approaches to subtyping ( [47,49], and this chapter are based on retractions) and the dominant form of modern theories of typing that deals with types of intensional objects (the type polymorphism in ML is, probably, the most widely known example).
Algebraic Subdomains Correspond to Closures
When we considered this problem in [7], we were using classical information systems.
We wrote: "Let us consider informally, what should we do to an information system A to obtain a subdomain |B| ⊆ |A|. We would like to destroy some elements of |A|. The elements are consistent and deductively closed sets of assertions in A. Therefore, some elements of |A| must lose their consistency or deductive closure to be destroyed. Also notice, that the loss of deductive closure means intensification of entailment." Then we figured out, that inconsistency can be treated as entailment of contradiction in A ⊤ , and hence the loss of consistency can be also treated as strengthening of entailment. The need to keep considering separately entailment and consistency in a number of technical situations, nevertheless, complicated our whole discourse.
With the definition of algebraic information system introduced in this Thesis, the matters become much simpler.
It is easy to see that the case when D A = D B does not lead to new domains |B|, so we do not consider this case further.
Since every algebraic information system is a finitary information system, the discourse from the previous chapter applies. Namely, we can see that if ⊢ A ⊆⊢ B , the approximable mapping f from A to A ⊤ , which corresponds to ⊢ B , is reflexive.
Since B is an algebraic information system, the restriction of f to A → A approximable mapping (or its extension to A ⊤ → A ⊤ ) is also transitive, so the corresponding restriction or extension of |f | would be a closure. Also the non-triviality Definition 7.2.2 An approximable mapping f from an algebraic information system A to the algebraic information system A ⊤ is called a generalized non-trivial closure, if Proof. Consider the algebraic information system F = [A → A ⊤ ]. We need to define the relation R, which would axiomatize properties of generalized non-trivial closures.
is the embedding λx.x). ⊢ A is also a theory with respect to R. Hence, the corollary from the previous section implies, that domain |F R ′ | consists of those elements of the domains |F |, which are theories closed with respect to R. It is easy to see that these elements are precisely all generalized non-trivial closures.
Hence, the information system F R ′ is exactly the information system Sub(A) we have been looking for. ✷ This differs favorably from [49] and from the other approaches to subdomains as sets of fixed points of projections, where subdomains are ordered by inclusion. Indeed, the smaller a subdomain is as a set, the more information we have about each of its elements, and hence, the larger this subdomain should be informationally (integers ⊒ reals is desirable).
Relationship between Subdomains and Subtypes
In [7] we expressed the opinion that compile-time and inheritance-oriented approaches to types, like subclasses in object-oriented programming and variants of strongly normalizing typed lambda-calculi, are incompatible with retraction-based subtyping in domain theory. This opinion was further confirmed by our reading of Section 10.4.4 on partial equivalence relations as types for systems with subtyping in the textbook by Mitchell [40].
general, be stronger, than entailment for the subtype. This and next chapters show that this is impossible for retractions yielding algebraic domains, at least as long as we stay with the reflexive transitive finitary logic.
Thus the issue of developing practical programming language with types conforming to "data types as objects" paradigm [49], where types are run-time first-class objects, and any compile-time type checking and inference is considered to be an op- For the duration of this chapter "domain" means algebraic Scott domain, "subdomain" means algebraic Scott subdomain, and "information system" means algebraic information system. Proof. Checking the axioms of algebraic information system for A r is straightforward. Observe that reflexivity of ⊢ r A follows from our selection of only such elements u ∈ P fin (D A ) as members of D r A , that uru. Now we use the intermediate reflexive property of r. For any x ∈ Fix(|r|), the set {u ∈ D r A | u ⊆ x} belongs to |A r |. For any y ∈ |A r |, the set {u | u ∈ y} is a fixed point of |r|. These are monotonic injective maps. This establishes the desired isomorphism.
✷ Lemma 8.1.3 Given an algebraic information system A and an approximable mapping r from A to A defining a retraction |r|, the following conditions are equivalent: 1. |r| is finitary.
2 ⇒ 1 follows from the previous lemma.
Here we prove 1 ⇒ 2. Consider an algebraic information system B and approximable mappings i and j from the Lemma 8.1.1. Consider u, w ∈ P fin (D A ), such that , such that v ′ iv and vjv ′ . Then urv, vrv, and vrw. ✷
Finitary Projections
Lemma 8.1.4 Given an algebraic information system A and an approximable mapping p from A to A, the following conditions are equivalent: 1. |p| is a finitary projection.
Proof. 1 ⇒ 2 follows from the previous lemma and the fact that for a projection To prove 2 ⇒ 1, observe that in the condition 2 upv and vpw, so |p| is a finitary retraction, and u ⊢ A w, so |p| is a projection.
Conjuctive Completeness
We say that algebraic information system A is (finitely) conjunctively complete, if for The step D r A = {u ∈ P fin (D A ) | uru} in Lemma 8.1.2 provides for conjuctive completion. Actually, A ⊢A A yield precisely the conjuctive completion of algebraic information system A, and |A ⊢A A | ∼ = |A|. It is actually enough to add conjuctions only for such u ∈ P fin (D A ), that uru and there is no such d ∈ D A , that {d} ⊢ A u and u ⊢ {d}, so if we start from a conjuctively complete system to begin with, the conjuctive completion step can be omitted.
Domains of Fixed Points of Finitary Projections
When p defines a finitary projection, the construction of Lemma 8.1.2 is rewritten as follows.
An algebraic information system
A . As before, |A p | ∼ = Fix(|p|). If the information system A was conjuctively complete to begin with, the construction above could be modified as follows: . Then for any x ∈ Fix(|p|), the set x D p A would belong to |A p |. For any y ∈ |A p |, the deductive closure of y in A is a fixed point of |p|. This establishes the desired isomorphism. In this case, for any x ∈ |A|, |p|(x) is the deductive closure of x D p A in A.
Domains of Fixed Points of Generalized Nontrivial Finitary Retractions
In order to fully incorporate the discourse of the previous two chapters into the framework of finitary retractions, one has to consider generalized nontrivial retractions |r| : |A| → |A ⊤ |, such that |r|(⊥) = ⊤ A ⊤ and |r|(|r|(x)) = |r|(x), when |r|(x) = ⊤ A ⊤ .
In order to use the setup of Lemma 8.1.1, one has to consider retraction |r ′ | : Notice that |B| has the compact top element, and that Fix(|r|) ∼ = |B ⊤ |.
Other Criteria of Finitarity
The following criterion of finitarity for projections was known before. A projection |p| : This criterion can be obtained from Lemma 1(ii) of [29], which in effect says that the set The consideration of information systems easily produces a much nicer criterion based on the intermediate reflexive property (Lemma 8.1.4). It does not require consideration of arbitrary non-finite elements x and allows to select finitary projections from the space of all continuous functions rather than from the space of projections. It can be literally rewritten in term of abstract cpo's: ∀x 0 , z 0 ∈ A 0 . z 0 ⊑ |p|(x 0 ) ⇒ ∃y 0 ∈ A 0 . x 0 ⊒ y 0 ⊒ z 0 , y 0 = |p|(y 0 ). Yet, one is unlikely to chose this criterion if he does not consider information systems because it involves inequality z 0 ⊑ |p|(x 0 ). Also notice that in this case we have to consider |p|(x 0 ), which is generally not finite.
For retractions, the criterion based on the intermediate reflexive property is given by Lemma 8.1.3. In terms of abstract cpo's our criterion can be rewritten as follows: The finite elements of Fix(|r|) are obtained as images of the finite elements of |A| under |r|. Hence, in the style close to [29], another criterion can be written as follows:
Domain of Finitary Projections
The set of all finitary projections |A| → |A| can be obtained as the set of fixed points of the finitary projection |P r| of |A → A|. Specifically, if g ∈ |A → A| define f = |P r|(g) It is easy to check that f is a finitary projection, and that if g is a finitary projection then g = |P r|(g). To prove that |P r| itself is a continuous function and a finitary projection, one should notice that {(u 1 , w 1 ), . . . , Hence this set is an algebraic Scott domain.
Omitted Issues
The details of construction of |Ret| is omitted in the present text. This construction is the result of mating the construction of |P r| with the construction of the domain of subdomains.
Limits of Projective Sequences
It is possible to generalize the construction of the limits of some projective sequences of domains defined by algebraic information systems, { |A n |, |A n+1 |, |i n |, |j n | | |i n | : |A n | → |A n+1 |; |j n | : |A n+1 | → |A n |; |i n |, |j n | is an embedding-projection pair; n = 1, . . .}, from [35,36] to arbitrary sequences of this kind. We do not give the details of the generalized construction here.
Issues Related to the Domain of Finitary Retractions
I do not know, whether |Ret| is unique.
The question might be somewhat related to the properties of the space of all retractions, say, of a powerset. This space is a complete lattice, but it is not algebraic and not even continuous.
Other Subclasses of Finitary Retractions
Here we look at two special subclasses of the class of finitary retractions, which may deserve special attention. [19]; the fundamental text in the theory of valuations on Scott opens sets is [31]).
CC-valuations
Remark: If a valuation µ is bounded and µ(X) = 0, then it is always easy to replace it with a normalized valuation µ ′ (U ) = µ(U )/µ(X).
Definition 1.4. A valuation µ is called continuous when for any directed system
of open sets µ( i∈I U i ) = sup i∈I µ(U i ).
We introduce two new properties of valuations.
This is, obviously, a very strong requirement, and we will see later that it might be reasonable to look for weaker non-degeneracy conditions.
Consider the interior of the intersection of these sets. It is easy to see that for a valuation µ
Definition 1.7.
A continuous, normalized, strongly non-degenerate, co-continuous valuation µ is called a CC-valuation. We should also notice that since our valuations are bounded, they can be extended onto closed sets via formula µ(C) = µ(A) − µ(A \ C), and all definitions of this section can be expressed in the dual form. This gives us considerable benefits in applications of these valuations and also suggests that the generalization of these applications to the case of unbounded valuations is a non-trivial undertaking.
A bounded valuation µ can be uniquely extended to an additive measure defined on the ring of sets generated from the open sets by operations ∩, ∪, \ [42]. We feel that it is useful to draft a possible construction here. We will denote the resulting additive measure also as µ. The issues of σ-additivity are not in the scope of this text (interested readers are referred to [31,4]). We deal with the specific infinite systems of sets we need, and mainly focus on quite orthogonal conditions given to us by co-continuity of µ.
Valuations Based on Weights of Basic Elements
Consider a continuous dcpo A with a countable basis K. Assign the converging system of weights to basic elements: w(k) > 0, k∈K w(k) = 1. Define µ(U ) = k∈U w(k). It is easy to see that µ is a continuous, normalized, strongly non-degenerate valuation.
However, µ is co-continuous if and only if all basic elements are compact (which is possible only if A is algebraic). To see why this is so, we undertake a small discourse, which will be useful for us later in this text.
First, observe that intersection Proof. By the definition of basis, x = ⊔K x , where K x = {k ∈ K | k ≪ x} is a directed set. For any k, consider U k = {y | y ≫ k}. Because K x is directed, it is easy to see that
Proof. (of co-continuity) Consider a filtered system of open sets
Let us show that x ∈ Int(B). If x ∈ Int(B) then there is y ∈ B such that y ≪ x by the Border Lemma. Then y ⊑ k for some k ∈ K x , and since B is upwardly closed, k ∈ B. This means that ∀k ′ ∈ K x . k ′ ≪ k yielding x ⊑ k and k ≪ x. Therefore k = x and x is compact, which is a contradiction. Thus, Since k∈B w(k) = inf k∈Kx µ(U k ) the Lemma is proved. ✷
Constructing CC-valuations
In this section we build a CC-valuation for all continuous dcpo's with countable bases.
The construction generalizes the one of Subsection 9.2.1. We are still going to assign weights, w(k) > 0, to compact elements. For non-compact basic elements we proceed as follows. We focus our attention on the pairs of non-compact basic elements, (k ′ , k ′′ ), which do not have any compact elements between them, and call such elements continuously connected. We observe, that for every such pair we can construct a special kind of vertical chain, which "behaves like the vertical segment [0, 1] of real line". We call such chain a stick. We assign weights, v(k ′ , k ′′ ) > 0, to sticks as well, in such a way that the sum of all w(k) and all v(k ′ , k ′′ ) is 1.
As in Subsection 9.2.1, compact elements k contribute w(k) to µ(U ), if k ∈ U .
An intersection of the stick, associated with a continuously connected pair (k ′ , k ′′ ), with an open set U "behaves as either (q, 1] or [q, 1]", where q ∈ [0, 1]. Such stick contributes The resulting µ is the desired CC-valuation.
It is possible to associate a complete lattice homomorphism from the lattice of Scott open sets to [0, 1] with every compact element and with every stick defined by basic continuously connected elements, k ′ and k ′′ . Then, as suggested by Keimel [32], all these homomorphisms together can be thought of as an injective complete lattice homomorphism to [0, 1] J . From this point of view, our construction of µ is the same as in [32].
Thus the discourse in this section yields the proof of the following: Remark: This implies that x and y are not compact.
Proof. We use the well-known theorem on intermediate values that x ≪ y ⇒ ∃z ∈
A. x ≪ z ≪ y (see [28]). Applying this theorem again and again we build a countable system of elements between x and y as follows, using rational numbers as indices for intermediate elements: x ≪ a 1/2 ≪ y, x ≪ a 1/4 ≪ a 1/2 ≪ a 3/4 ≪ y, . . .
All these elements are non-compact and hence non-equal. Now consider a directed set
where r is a real number, 0 < r < 1. Introduce b r = ⊔{a i | i ≤ r}. We prove that if r < s then b r ≪ b s , and also that x ≪ b r ≪ b s ≪ y, thus obtaining the required cardinality. Indeed it is easy to find such n and numbers q 1 , q 2 , q 3 , q 4 , that We call the set of continuum different non-compact elements {a r | r ∈ (0, 1)} between continuously connected x ≪ y, built in the proof above, such that x ≪ a r ≪ a q ≪ z ⇔ r < q a (vertical) stick.
Proof of Theorem 9.3.1
Consider a continuous dcpo A with a countable basis K. As discussed earlier, with every compact k ∈ K we associate weight w(k) > 0, and with every continuously connected Since K is countable, we can require w(k) + v(k ′ , k ′′ ) = 1.
Whenever we have an upwardly closed (i.e. Alexandrov open) set U , for any stick It is easy to show that µ is a normalized valuation. The rest follows from the following Lemmas. Proof. Consider a directed system of open sets, {U i , i ∈ I}, and U = i∈I U i . We need to show that for any ǫ > 0, there is such Take enough (a finite number of) compact elements, k 1 , . . . , k n , and continuously connected pairs of basic elements, (k ′ 1 , k ′′ 1 ), . . . , (k ′ m , k ′′ m ), so that w(k 1 ) + . . .
An upper bound of these U ij and U i ′ j is the desired U i . ✷ Lemma 9.3.3 µ is strongly non-degenerate.
Proof. Let U and V be Scott open subsets of A and U ⊂ V . Let us prove that V \ U contains either a compact element or a stick between basic elements. Take x ∈ V \ U . If x is compact, then we are fine. Assume that x is not compact. We know that x = ⊔K x , Otherwise, since any basis includes all compact elements, k and x are continuously connected.
Now, as in the theorem of intermediate values
If k ∈ V \ U and k is compact, then µ(V ) − µ(U ) ≥ w(k) > 0. If the stick formed . The further development is essentially dual to the proof of Lemma 9.3.2. We need to show that for any ǫ > 0, there is such U i , i ∈ I, that µ(U i ) − µ(Int(B)) < ǫ.
Take enough (a finite number) of compact elements, k 1 , . . . , k n , and continuously connected pairs of basic elements, (k ′ . A lower bound of these U ij and U i ′ j is the desired U i . ✷ It should be noted that Bob Flagg suggested and Klaus Keimel showed that Lemma 5.3 of [23] can be adapted to obtain a dual proof of existence of CC-valuations (see [22] for one presentation of this). Klaus Keimel also noted that one can consider all pairs k, k ′ of basic elements, such that k ≪ k ′ , instead of considering just continuously connected pairs and compact elements.
Open Problems
The key open problems are related to fast computation of valuations and integrals and to canonical valuations on functional spaces and reflexive domains.
Since the similar problems exist for relaxed metrics we will return to them at the end of the next chapter.
Relaxed and Partial Metrics
In this chapter we present our joint results with Joshua Scott [11,12,13]. We presume that the methods of denotational semantics allow us to obtain adequate descriptions of program behavior (e.g., see [53]). The term domain in this chapter denotes a directed complete partial order (dcpo) equipped with the Scott topology.
Axiom ρ(x, x) = 0 Cannot Hold
Assume that there is an element 0 ∈ D representing the ordinary numerical 0. Let us show that ∀x. ρ(x, x) = 0 cannot be true under reasonable assumptions. We will see later that all other properties of ordinary metrics can be preserved at least for all continuous dcpo's with countable bases (see the next chapter).
It seems reasonable to assume that any reasonable construction of T [ρ] for any generalized distance function ρ : A × A → D, should satisfy the following axiom, regardless of whether the distance space D is a domain, or whether ρ is continuous: Lemma 10.1.1 Assume that there are at least two elements x, y ∈ A, such that x ❁ A y.
Proof. The continuity of ρ implies its monotonicity with respect to the both of its arguments. Then x ❁ A y implies d = ρ(x, x) ⊑ D ρ(x, y) ⊑ D ρ(y, y) = d. This yields ρ(x, y) = d, and, similarly, ρ(y, x) = d. ✷ Then we can obtain the following simple, but important result. The topologies used in domain theory are usually T 0 ; in particular, the Scott topology is T 0 . This justifies studying continuous generalized metrics ρ, such that ρ(x, x) = 0 is false for some x, more closely.
Intuition behind ρ(x, x) = 0
There are compelling intuitive reasons not to expect ρ(x, x) = 0, when x is not a maximal element of A. The computational intuition behind ρ(x, y) is that the elements in question are actually x ′ and y ′ , x ⊑ A x ′ , y ⊑ A y ′ , but not all information is usually known about them. The correctness condition ρ(x, y) ⊑ D ρ(x ′ , y ′ ) is provided by the monotonicity of ρ.
In particular, even if x = y, this only means that we know the same information about x ′ and y ′ , but this does not mean that x ′ = y ′ . Consider x ′ = y ′ , such that x ❁ A x ′ and x ❁ A y ′ . Then ρ(x, x) ⊑ D ρ(x ′ , y ′ ) and ρ(x, x) ⊑ D ρ(y ′ , x ′ ), and at least one of ρ(x ′ , y ′ ) and ρ(y ′ , x ′ ) is non-zero, if we want ρ to yield a T 0 topology (we do not assume symmetry yet).
Partial Metrics
Historically, partial metrics are the first generalized distances on domains for which the axiom ρ(x, x) = 0 does not hold. They were introduced by Matthews [39,38] and further investigated by Vickers [54] and O'Neill [41].
Partial metrics satisfy a number of additional axioms in lieu of ρ(x, x) = 0 (see Section 10.5). Matthews and Vickers state that ρ(x, x) = 0 is caused by the fact, that x expresses a partially defined object. The most essential component of the central construction in this and the next chapter is a partial metric (Section 10.6).
Our Contribution
We build partial metrics yielding Scott topologies for a wider class of domains that was known before (see the next chapter). This class -all continuous dcpo's with countable bases -is sufficiently big to solve interesting domain equations and to define denotational semantics of at least sequential deterministic programming languages [53].
Actually, a recent careful analysis of papers [34,39] together had shown that an even more general result followed easily from this papers taken together, however this fact remained unwritten folklore known only to a few people.
We introduce the notion of relaxed metric (Section 10.3), which maintains the intuitively clear requirement to reject axiom ρ(x, x) = 0, but does not impose the specific axioms of partial metrics. We believe that the applicability of these specific axioms is more limited (Section 10.9.1).
We introduce the idea that a space of distances should be thought of as a data type in the context of denotational semantics and, thus, should be represented by a domain. We also introduce the requirement that distance functions should be computable and, thus, Scott continuous (the use of continuous valuations in [41] should be considered as a step in this direction).
These considerations lead to an understanding that relaxed metrics should map pairs of partial elements to upper estimates of some "ideal" distances, where the distance domain of upper estimates, R − , is equipped with a dual informational order: ⊑ R − =≥.
We also consider lower estimates of "ideal" distances, thus, introducing the distance domain of interval numbers, R I . Continuous lower estimates are useful during actual computations of distances (Section 10.8) and for defining and computing an induced metric structure on the space of total elements (Theorem 10.6.2).
There is a comparison in [39] between partial metrics and alternative generalized distance structures such as quasi-metrics and weighted metrics [34]. We provide what we believe to be the strongest argument in favor of partial metrics so far -among all those alternatives only partial metrics can be thought of as Scott continuous, computable functions.
Relaxed Metrics
Consider distance domains in greater detail. It is conventional to think about distances as non-negative real numbers. When it comes to considering approximate information about reals, it is conventional to use some kind of interval numbers.
We follow both conventions in this text. The distance domain consists of pairs a, b (also denoted as [a, b]) of non-negative reals (+∞ included), such that a ≤ b.
Recall from the Chapter 4 that we denote this domain as R I and that [a, b] Also recall that we can think about R I as a subset of R + × R − , where ⊑ R + =≤, ⊑ R − =≥, and both R + and R − consist of non-negative reals and +∞. We call R + a domain of lower bounds, and R − a domain of upper bounds. Thus a distance function ρ : A × A → R I can be thought of as a pair of distance functions l, u , l : We think about l(x, y) and u(x, y) as, respectively, lower and upper bounds of some "ideal" distance σ(x, y). We do not try to formalize the "ideal" distances, but we refer to them to motivate our axioms. There are good reasons to impose the triangle inequality, u(x, z) ≤ u(x, y) + u(y, z). Assume that for our "ideal" distance, the triangle inequality, σ(x, z) ≤ σ(x, y) + σ(y, z), holds. If u(x, z) > u(x, y) + u(y, z), then u(x, y) + u(y, z) gives a better upper estimate for σ(x, z) than u(x, z). This means that unless u(x, z) ≤ u(x, y) + u(y, z), u could be easily improved and, hence, would be very imperfect.
This kind of reasoning is not valid for l(x, z). In fact, there are reasonable situations, when l(x, z) = 0, but l(x, y) = l(y, z) = 0. E.g., consider Example 10. Also only u plays a role in the subsequent definition of the relaxed metric topology, and the most important results remain true even if we take l(x, y) = 0. In the last case we sometimes take D = R − instead of D = R I making the distance domain look more like ordinary numbers (it is important to remember, that ⊑ R − =≥ and, hence, 0 is the largest element of R − ).
We also impose the symmetry axiom on the function ρ. The motivation here is that we presume our "ideal" distance to be symmetric, hence, we should be able to make symmetric upper and lower estimates.
We state a definition summarizing the discourse above: when it satisfies the triangle inequality. A symmetric function ρ : A × A → R I is called a relaxed metric when its upper part u is a relaxed metric.
Relaxed Metric Topology
An open ball with a center x ∈ A and a real radius ǫ is defined as B x,ǫ = {y ∈ A | u(x, y) < ǫ}. Notice that only upper bounds are used in this definition -the ball only includes those points y, about which we are sure that they are not too far from x.
We should formulate the notion of a relaxed metric open set more carefully than for ordinary metrics, because it is now possible to have a ball of a non-zero positive radius, which does not contain its own center.
Partial Metrics
The distances p with p(x, x) = 0 were first introduced by Matthews [39,38]. They are known as partial metrics and obey the following axioms: Whenever partial metrics are used to describe a partially ordered domain, a stronger form of the first two axioms is used: If x ⊑ y then p(x, x) = p(x, y), otherwise p(x, x) < p(x, y). We include the stronger form in the definition of partial metrics for the purposes of this work.
Central Construction
Here we construct continuous relaxed metrics yielding the Scott topology for all continuous Scott domains with countable bases. Our construction closely resembles one by O'Neill [41]. We also use valuations, but we consider continuous valuations on the If, in addition, we would like the next theorem to hold, we have to consider a different version of ρ with u(x, y) = 1 − W (K x ∩ K y ) − W (I x ∩ I y ) and l(x, y) = W (K x ∩ I y ) + W (K y ∩ I x ). The previous theorem still holds.
We introduce the notion of a regular basis. A set of maximal elements in A is denoted as Total (A) . We say that the basis K is regular if ∀k ∈ K, x ∈ Total (A). k ⊑ x ⇒ k ≪ x. In particular, if K consists of compact elements, thus making A an algebraic Scott domain, K is regular. Here we prove relatively difficult parts of these theorems for the case when u(x, y) = 1 − W (K x ∩ K y ) − W (I x ∩ I y ) and l(x, y) = W (K x ∩ I y ) + W (K y ∩ I x ). Lemma 10.7.2 is needed for Theorem 10.6.2, and other lemmas are needed for Theorem 10.6.1.
Proof. Using K x ∩I x = ∅ we can rewrite u and l. u(x, y) We want to show that V ⊆ U , for which it is enough to show that V ∩ U = ∅.
We show that (K x ∩ I y ) ∩ U = ∅. Then by symmetry the same will be true for K y ∩ I x , and hence for V .
(K x ∩I y )∩U = (K x ∩I y ∩K x ∩K y )∪(K x ∩I y ∩I x ∩I y ). But K x ∩I y ∩K x ∩K y ⊆ I y ∩ K y = ∅. Similarly, K x ∩ I y ∩ I x ∩ I y = ∅. ✷ Lemma 10.7.2 If K is a regular basis and x, y ∈ Total (A), then l(x, y) = u(x, y).
Proof. Using the notations of the previous proof we want to show that U ⊆ V .
Let us show first that K x ∪ I x = K y ∪ I y = K. Consider k ∈ K. Since x ∈ Total (A), if k ∈ I x , then k ⊑ x. Now from the regularity of K we obtain k ≪ x and k ∈ K x . Same for y. Now, if k ∈ U , then k ∈ K x or k ∈ K y . Because of the symmetry it is enough to consider k ∈ K x . Then k ∈ I x . Then, using k ∈ U once again, k ∈ I y . Then k ∈ K y .
and the similar formula holds for I's.
Then the result follows from the following simple facts: Proof. Monotonicity of ρ is trivial.
Consider a directed set B ⊆ A and some z ∈ A. We have to show that ρ(z, ⊔B) = Rewriting this, we want to show that W (K z ∩K ⊔B )+W (I z ∩I ⊔B ) = sup{W (K z ∩ Monotonicity considerations trivially yield both "≥" inequalities, so it is enough to show "≤" inequalities. In fact, we will show that for any sets C ⊆ A It is easy to show that K ⊔B = ∪{K x | x ∈ B} by showing first that the set ∪{K x | x ∈ B} is directed and ⊔B = ⊔(∪{K x | x ∈ B}). Let us prove that I ⊔B = ∪{I x | x ∈ B}. "⊇" is trivial. Let us prove "⊆". Assume that k ∈ ∪{I x | x ∈ B}, i.e. ∀x ∈ B. k ⊔ x exists. It is easy to see that because B is a directed set, {k ⊔ x | x ∈ B} is also directed. Then k ⊑ ⊔{k ⊔ x | x ∈ B} ⊒ ⊔B, implying existence of k ⊔ (⊔B) and, hence, k ∈ I ⊔B . Now consider enumerations of the countable or finite sets K ⊔B and I ⊔B : k 1 , ..., k n , ... and k ′ 1 , ..., k ′ n , ..., respectively. Define tail sums S n = w(k n ) + w(k n+1 ) +... and Observe that (S n ) and (S ′ n ) converge to 0.
Pick for every k n some x n ∈ B such that k n ∈ K xn . Pick for every k ′ n some x ′ n ∈ B such that k ′ n ∈ I x ′ n . Then using the directness of B, we can for any n pick such y n ∈ B, that x 1 , ..., x n , x ′ 1 , ..., x ′ n ⊑ y n . Then k 1 , ..., k n ∈ K yn and k ′ 1 , ..., k ′ n ∈ I yn .
It is easy to see that (W ( . Therefore u(x, y) ≥ u(x, x) + w(k) and thus y ∈ B x,ǫ . ✷
Computability and Continuity
Consider an effective algebraic Scott domain A and its computable element x. We recall that this means that K x is recursively enumerable (Section 5.4). It is easy to show that in this case I x must also be recursively enumerable. However, as we discussed in Section 5.4, one almost never should expect them to be recursive.
The actual computation of ρ(x, y) goes as follows. Start with [0, 1] and go along the recursive enumerations of K x , K y , I x , and I y . Whenever we discover that some k occurs in both K x and K y , or in both I x and I y , subtract w(k) from the upper boundary.
Whenever we discover that some k occurs in both K x and I y , or in both I x and K y , add w(k) to the lower boundary. If this process continues long enough, [l(x, y), u(x, y)] is approximated as well as desired.
However, there is no general way to compute a better lower estimate for u(x, y) than l(x, y), or to compute a better upper estimate for l(x, y) than u(x, y). Consequently, there is no general way to determine how close is the convergence process to the actual values of l(x, y) and u(x, y), except that we know that u(x, y) is not less than the currently computed lower bound, and l(x, y) is not greater than the currently computed upper bound. Of course, for large x and y this knowledge might provide a lot of information, and if the basis of our domain is regular, for total elements x and y this knowledge provides us with precise estimates -i.e. if the basis is regular, then the resulting classical metric on Total (A) can be nicely computed.
The computational situation is very different with regard to quasi-metrics. Con- . This is a quasi-metric in the style of [50,33], and it yields a Scott topology [12,13]. However, as discussed in [12,13], typically K x \ K y is not recursive. Moreover, one should not expect K x \ K y or its complement to be recursively enumerable. This precludes us from building a generally applicable method computing d(x, y) and illustrates that it is computationally incorrect to subtract one upper bound from another. Whenever the upper part u(x, y) of a relaxed metric is based on common information shared by x and y yielding a negative contribution to the distance (we subtract the weight of common information from the universal distance 1 in this chapter), both u(x, z) ≤ u(x, y) + u(y, z) − u(y, y) and u(x, x) ≤ u(x, y) should hold. As we shall see in full generality in the next chapter, when u(x, y) = 1 − µ(Common information between x and y), axioms of partial metrics hold.
However, to specify function u in Example 10.1.1, we use information about x and y, which cannot be thought of as common information shared by x and y. In such case we still expect a relaxed metric ρ, but its upper part u does not have to satisfy the axioms of partial metrics.
Functoriality
One of the open issues in the field remains to develop a functorial approach to relaxed metrics. This approach should facilitate our abilities to both define canonical distances on domains and to compute these distances at a reasonable cost.
We also hope to be able to extend this approach to reflexive domains and to obtain some invariance properties of ρ in the process.
In all cases we would like to ensure that relaxed metric topology and Scott topology coincide. In particular, in order for Scott open sets to be relaxed metric open it is necessary (not sufficient) that u(x, x) ≥ u(x, y) ⇒ x ⊑ y.
However, for general relaxed metrics it is easy to come up with examples, when , and the similar situation takes place for A → B. However, the strong form of the axiom of small self- u(x, y)) seems to eliminate these problems.
Applications and Open Problems
Let us briefly state where we stand with regard to the applications to programs. We are able to introduce relaxed metrics on a class of domains sufficiently large for practical applications in the spirit of [53]. and u(P, P ) = inf u(P N , P N ). Of course, none of these distances has to be zero.
However, we do not know yet how to build relaxed distances so that not only nice convergence properties are true, but also that distances between particular pairs of programs "look right" -a notion, which is more difficult to formalize, than convergence.
Also, we compute these distances via recursive enumeration now, and a more efficient scheme is needed. It should be noted here that since we are going to distinguish functions topologically by integrating them over A, we need these measures to be based on strongly non-degenerate valuations.
Obtaining Generalized
Distances from Valuations
In this chapter we continue to present our joint results with Svetlana Shorina [15,16].
In the previous chapter we introduced a construction of partial metrics based on the mechanism of common information between elements x and y bringing negative contribution to u(x, y). This construction was based on assigning finite weights to basic elements and gave topologically meaningful relaxed metrics for all continuous Scott domains with countable bases. For such domains with regular bases (which included all algebraic Scott domains) this construction gave relaxed metrics which "behaved well" on the total elements of the domain.
Here we introduce a general method of defining partial and relaxed metrics via information about elements for all dcpo's via mechanism of µInfo-structures.
Then we show how to obtain such a µInfo-structure from a CC-valuation for any continuous Scott domain. Since we know, how to construct a CC-valuation for any continuous dcpo with countable basis, this method of constructing partial and relaxed metrics via CC-valuations and the resulting µInfo-structures works for all continuous Scott domains with countable bases.
(SCOTT CONTINUITY OF THE INDUCED RELAXED METRIC)
if B is a directed subset of A and y ∈ A, then (a) µ(Info(⊔B) ∩ Info(y)) = sup x∈B (µ(Info(x) ∩ Info(y)),
(SCOTT OPEN SETS ARE RELAXED METRIC OPEN)
for any (basic) Scott open set U ⊆ A and x ∈ U , there is an ǫ > 0, such that We will also consider deficient µInfo-structures, when the strong respect for totality axiom is not imposed.
In terms of lattice theory, µ is a (normalized) valuation on a lattice M(I). The consideration of unbounded measures is beyond the scope of this work, and µ(I) = 1 is assumed for convenience. Axioms relating ⊑ and Info are in the spirit of information systems approach [48], although we are not considering any inference structure over I in this chapter.
The requirements for negative information are relatively weak, because it is quite natural to have ∀x ∈ A. Neginfo(x) = ∅ if A has the top element.
The axiom that for x ∈ Total (A), Info(x) ∪ Neginfo(x) = I, is desirable because indeed, if some i ∈ Info(x) does not belong to Info(x) and x can not be further increased, then by our intuition behind Neginfo(x), i should belong to Neginfo(x). However, this axiom might be too strong and will be further discussed later.
The last two axioms are not quite satisfactory -they almost immediately imply the properties, after which they are named, but they are complicated and might be difficult to establish. We hope, that these axioms will be replaced by something more tractable in the future. One of the obstacles seems to be the fact in some valuable approaches (in particular, in this chapter) it is not correct that x 1 ⊑ x 2 ⊑ · · · implies that Info(⊔ i∈N x i ) = i∈N Info(x i ).
The nature of these set-theoretical representations, I, of domains may vary: one can consider sets of tokens of information systems, powersets of domains bases, or powersets of domains themselves, custom-made sets for specific domains, etc. The approach via powersets of domain bases of the previous chapter can be thought of as a partial case of the approach via powersets of domains themselves adopted in this chapter.
Partial and Relaxed Metrics via µInfo-structures
Define the (upper estimate of the) distance between x and y from A as u : I.e. the more information x and y have in common the smaller is the distance between them. However a partially defined element might not have too much information at all, It is possible to find information which will never made it into Info(x) ∩ Info(y) or Neginfo(x) ∩ Neginfo(y) even when x and y are arbitrarily increased. In particular, Info(x) ∩ Neginfo(y) and Info(y) ∩ Neginfo(x) represent such information. Then we can introduce the lower estimate of the distance l : A × A → R + : l(x, y) = µ(Info(x) ∩ Neginfo(y)) + µ(Info(y) ∩ Neginfo(x)).
The proof of Lemma 10.7.3 goes through.
The symmetry u(x, y) = u(y, x) and l(x, y) = l(y, x) is obvious. Another helpful fact is We are using the axiom x ⊑ y ⇒ Info(x) ⊆ Info(y), which implies Info(x) ∩ Info(y) = Info(x), and the same for Neginfo.
This follows from the following fact:
Any relaxed metric open set is Scott open.
This follows at once from the following property: 7. The induced relaxed metric is a Scott continuous function.
This also follows immediately from the corresponding axiom of µInfo-structure.
Thus, the following theorem is proved (cf. Theorem 10.6.1). Observe, that the strong respect for totality axiom is NOT used in the proof. However, it might still be fruitful to consider replacing the axiom ∀x ∈ Total (A).
A Previously Known Construction
Here we recall a construction from the previous chapter based on a generally non-cocontinuous valuation of Subsection 9.2.1. We will reformulate it in our terms of µInfo-structures. In the previous chapter it was natural to think that I = K. Here we reformulate that construction in terms of I = A, thus abandoning the condition x ∈ Total (A) ⇒ Info(x) ∪ Neginfo(x) = I altogether.
Define Info(x) = P x , Neginfo(x) = I x . Consider a valuation µ of Subsection 9.2.1: for any S ⊂ I = A, µ(S) = k∈S∩K w(k). µ is a continuous strongly non-degenerate valuation, but it is not co-continuous unless K consists only of compact elements.
Because of this we cannot replace an inconvenient definition of Info(x) = P x by non-compact basic element, leading to the non-continuity of the partial metric u(x, y).
Also the reliance on countable system of finite weights excludes such natural partial metrics as metric u : 1] is the set [0, 1] equipped with the dual partial order ⊑ = ≥, and u(x, y) = max(x, y). We rectify all these problems in the next Subsection. It is easy to see that valuation, Info, and Neginfo axioms of µInfo-structure hold. We
Partial and Relaxed Metrics via CC-valuations for Continuous Scott Domains
These axioms are established by the Lemmas below.
You will also see that for such bare-bones partial metrics, as u(x, y) = 1−µ(C x ∩ C y ), which are nevertheless quite sufficient for topological purposes and for domains with ⊤, only co-continuity of valuations matters, continuity is not important.
Observe also that since the construction in Section 9.2.1 does form a CCvaluation for algebraic Scott domains with bases of compact elements, the construction in the previous chapter can be considered as a partial case of our current construction if the basis does not contain non-compact elements. Remark: Note that continuity of µ is not required here.
Proof. Part A: Let Q be a closed subset of A. Then (C x ∩ Q, x ∈ B) form a directed set of closed sets. We need to show that ( x∈B C x ) ∩ Q = C ⊔B ∩ Q, then the result will follow by co-continuity. "⊆" is trivial. Let us proof "⊇". Consider However, for ∀y ∈ P x , since y ≪ x and x ⊑ ⊔B, ∃z ∈ B. y ⊑ z, hence y ∈ C z ∩ Q. Hence This follows from µ(C ⊔B ) = sup x∈B µ(C x ) by Part A and µ(C ⊔B \Q) ≥ sup x∈B µ(C x \Q) by monotonicity of µ. ✷ Lemma 11.1.2 Assume that µ is a continuous valuation and B is a directed subset of A. Then µ(I ⊔B ∩ Q) = sup x∈B (µ(I x ∩ Q)), were Q is an open or closed subset of A.
Remark: Co-continuity is not needed here.
Proof. During the proof of Lemma 10.7.4 it was established that I ⊔B = x∈B I x (think about k ∈ A and observe that that proof goes through). It is our current feeling, that instead of trying to formalize weaker non-degeneracy conditions, it is often more fruitful to change a µInfo-structure. In particular, one can build a µInfo-structure based on I = [0, 1] × [0, 1] in the situation described above.
Negative Information and Tolerances
In this chapter we present our later joint results with Svetlana Shorina [17,14]. In the previous chapter we showed how to obtain such a µInfo-structure from a CCvaluation for any continuous Scott domain. Certain pathologies in the behavior of I x = {y ∈ A | {x, y} is unbounded} precluded us from extending this method beyond bounded complete domains.
In this chapter, we obtain meaningful Scott continuous relaxed metrics for continuous dcpo's by replacing I x with Neginfo(x) = Int(J x ), where J x = {y ∈ A | x ∈ Int(I y )}. This result can be understood in terms of interplay between negation duality and Stone duality. Since we know how to construct a CC-valuation for any continuous dcpo with countable basis, this method of constructing partial and relaxed metrics via CC-valuations and the resulting µInfo-structures works for all continuous dcpo's with countable bases.
Escardo [21,52] defined a topological space A to be weakly Hausdorff, if its consistency relation is closed. The consistency relation is given by formula x ↑ y = { x, y | ∃z ∈ A. x ⊑ z&y ⊑ z}, where ⊑ is the specialization order of A. Since the Lawson condition is equivalent to the formula ∀x ∈ Total (A).
which is a weakening of I = J, which is, in turn, equivalent to A being weakly Hausdorff, we call the spaces satisfying the Lawson condition very weakly Hausdorff.
Tolerances and a Smyth Conjecture
Recently Mike Smyth [52] and Julian Webster [58] advanced the approach in which tolerance is considered not as an alternative to standard topology, but as a structure complementary to topology. In particular, it seems to be fruitful to equip Scott domains with tolerances. Sections 12.6 and 12.7 make a small contribution to this emerging theory.
Smyth [52] defines a tolerance as a reflexive symmetric relation following Poincare and Zeeman. He defines a topological tolerance space as a topological space equipped with a tolerance relation closed in the product topology.
For a weakly Hausdorff space, ↑ is the least closed tolerance.
The set I x = {y ∈ A | {x, y} is unbounded} is an observable continuous representation of negative information about x ∈ A for a weakly Hausdorff continuous dcpo A with the Scott topology. We will see in Section 12.4 that, when A is not weakly Hausdorff the largest continuous approximation of I x is represented by J x = {y ∈ A | x ∈ Int(I y )}, and the largest observable continuous representation of I x is represented by J ′ x = Int(J x ).
Smyth conjectured, that J or J ′ is closely related to the least symmetric closed tolerance on A. In this chapter we establish that, indeed, { x, y | y ∈ J ′ x } is the complement of this tolerance.
We also establish a relationship between this tolerance and lower bounds of relaxed metrics on A. 12.
Proof. ⇒. Consider y ∈ I ⊔B . y ∈ I ⊔B ⇔ ⊔B ∈ I y . If I y is Scott open, ∃b ∈ B. b ∈ I y , and b ∈ I y ⇔ y ∈ I b .
⇐. Consider directed B, such that ⊔B ∈ I y . ⊔B ∈ I y ⇔ y ∈ I ⊔B . Condition I ⊔B = b∈B I b implies that ∃b ∈ B. y ∈ I b , and y ∈ I b ⇔ b ∈ I y . ✷
Negation Duality and Stone Duality
We consider equality I x = {y ∈ A | x ∈ I y }, which is just another way to state the negation duality. Then, instead of taking Int(I x ), we consider J x = {y ∈ A | x ∈ Int(I y )}.
Then the continuity property for the resulting relaxed metric will be restored. It is going to be technically convenient to replace I x not with J x , but with Int(J x ), but we do not think that this feature is principal.
We will see that the reasons for J x to work in this situation can be best under- . We should emphasize here, that what is going on in this chapter is a rather subtle interplay of two different dualities -negation duality and Stone duality -none of which seems to be reducible to another.
We will also see, that the function x → J x is the largest "negative" Scott contin-
Lawson Condition
In general we only get deficient µInfo-structures via the use of Neginfo(x) = J x or Neginfo(x) = Int(J x ). Namely, the totality property does not hold in general, and thus we still do not get an induced metric on Total (A). However, in this situation it helps to impose the Lawson condition, that relative Scott and Lawson topologies on Total (A) are equal.
The Lawson condition was introduced in [37], and is widely used lately. As Lawson writes in the Introduction to [37], "this turns out to be a very fruitful notion that permit great generality, but at the same time permits the derivation of many important structure results". This condition is now a standard part of the notion of a computational model for a topological space [22].
In this chapter, we present two equivalent formulations of the Lawson condition:
Polish Spaces
Lawson has shown in [37], that for every continuous dcpo A with countable basis, Lawson In Section 12.4 we analyze the pathologies of behavior of I x on a specific example.
We then study the properties of J x , which serves as a replacement for I x , and explain those properties from the viewpoint of Stone duality.
In Section 12.5 we find equivalent formulations of the Lawson condition and use these formulations to establish the totality property of the resulting µInfo-structures with its ramifications for the induced metrics on Total (A).
In Section 12.6 we talk about tolerances and prove the Smyth Conjecture.
In Section 12.7 we establish that for the relaxed metrics defined above, x ∼ y if and only if l(x, y) = 0, and build a continuous family of tolerances.
When I x Behaves Well
In this section we study cases, when for all x ∈ A, I x is Scott open, or, equivalently, when for all directed B ⊆ A, I ⊔B = x∈B I x . We already know, that this situation takes place for continuous Scott domains.
Another class of domains, for which these properties can be established, is the class of coherently continuous dcpo's with countable basis. The term "coherence" here is understood in the weak sense of [1] (weaker, than bounded completeness), and not in the strong sense of [26] (stronger, than bounded completeness). Proof. Consider y ∈ I x . Since the space has a countable basis, we only need to show, that if there is a sequence of basic elements, k 1 ⊑ k 2 ⊑ . . ., such that y = ⊔k i , then some k i belongs to I x .
By contradiction, assume that this is not the case. Then for all i, k i and x have an upper bound. Using the presence of a countable basis again, approximate x with a sequence of basic elements as well: l 1 ⊑ l 2 ⊑ . . ., x = ⊔l i . Then for all i, k i and l i have an upper bound.
Now we are going to build the sequence of (not necessarily basic) elements, . ., such that for all i, k i ⊑ u i and l i ⊑ u i . Then ⊔u i would be an upper bound of x and y, yielding the desired contradiction.
Consider an element v ∈ MUB(k i , l i ). Define the height of v as maximal j, such that there is w ∈ MUB(k j , l j ), such that v ⊑ w. If there is no such maximal natural number, we say that v is of infinite height. Using coherence condition, it is easy to see, that there is an element u 1 ∈ MUB(k 1 , l 1 ) of infinite height. Now consider only elements v ∈ MUB(k 2 , l 2 ), such that u 1 ⊑ v. Using coherence condition, it is easy to see once again, that there is an element u 2 ∈ MUB(k 2 , l 2 ) of infinite height, such that u 1 ⊑ u 2 .
Continuing this process, we obtain the desired sequence. ✷ Therefore, one can use N x = I x in order to obtain all results of the previous section not only for continuous Scott domains, but also for coherently continuous dcpo's with countable bases. Now we will see how negation duality works in this example. In our previous Proof. The potentially non-trivial direction is to prove Int(J) ⊆ m∈M Int(J m ). Consider y ∈ Int(J). By the Border Lemma (Lemma 9.2.1) there is x ∈ J, such that x ≪ y.
When
Then, because of the condition of the Lemma we are currently proving, there is m ∈ M , such that x ∈ J m . Then applying the Border Lemma again, we obtain y ∈ Int(J m ). ✷ Hence, N x = Int(J x ) enables us to satisfy all the requirements of the setup of Section 11.1.4, except for the requirement that for all x ∈ Total (A), C x ∪ N x = A, which does not hold in general. E.g. consider our example domain E, and observe that e ∞ ∈ Total (A), but 0 E ∈ C x ∪ N x . (Observe, also that changing N x to J x does not fix this.) Thus the Theorem 11.1.1 holds, but the Theorem 11.1.2 about the equality of l(x, y) and u(x, y) does not have to hold, and the resulting induced metric on Total (A) cannot in general be obtained. E.g., in our example domain E, we have that e ∞ ∈ Total (E), but u(e ∞ , e ∞ ) = µ(I e∞ ) − µ(J e∞ ) = µ({0 E }), which is, in general, not zero, since 0 E is compact.
Stone Duality
The first Lemma in the previous subsection holds for the reasons, which are not related to such specific features of J x as the use of Int(I y ) (any open set can be used instead) and the fact, that x and y belong to the same set A.
We analyze this situation in the spirit of Stone duality [30,55], which is a contravariant equivalence between categories of spatial frames (of open sets) and sober topological spaces.
For the purpose of this subsection only, assume that there is a continuous dcpo A (Scott topologies of continuous dcpo's are sober [30]) and a set D, and that we are
J x Is the Largest Continuous Approximation of I x
Both I x and J x can be considered as functions from A to the powerset of A, P(A).
However, in general, only J x is Scott continuous. The following theorem shows that, in some sense, J x is the best we can do. Proof. Assume that such Scott continuous function f is given, and for some x, there is y ∈ f (x), such that y ∈ J x , i.e. x ∈ Int(I y ). However, y ∈ I x means x ∈ I y . Now consider a directed set B = K x . We have that x = ⊔B and that all b ∈ B are way below x. Then, taking into account x ∈ Int(I y ) and applying the Border Lemma, we obtain that ∀b ∈ B. b ∈ I y .
However, the assumption of continuity of f means, that f (⊔B) = b∈B f (b).
Hence, since y ∈ f (⊔B), there is some b ∈ B, such that y ∈ f (b), hence y ∈ I b , hence b ∈ I y , contradicting the last formula of the previous paragraph.
The Use of the Lawson Condition
We start with the equivalent formulation of the Lawson condition. What is going on here is described by the following Theorem. Proof. Assume that the Lawson condition holds and x ∈ Total (A). Assume, that y ⊑ x, i.e. y ∈ C x and y ∈ I x , using the totality of x. Thus, by duality, x ∈ I y . Because x ∈ Total (A) and because due to the Lawson condition I y ∩Total (A) = Int(I y )∩Total (A), we obtain x ∈ Int(I y ), hence y ∈ J x .
Conversely, assume ∀x ∈ Total (A). J x = I x . Let us prove I y ∩ Total (A) = Int(I y ) ∩ Total (A), thus proving the Lawson condition. Take x ∈ I y ∩ Total (A). By negation duality, y ∈ I x , then, by totality of x and our assumptions, y ∈ J x , which, by definition of J x , means that x ∈ Int(I y ). 12.6 Tolerances and Negative Information 12.6.1 Tolerances, Distinguishability, and Observability Smyth [52] requires that a tolerance relation is closed in the product topology. Here are informal reasons for this.
The typical meaning of two points being in the relation of tolerance, x ∼ y, is that x cannot be distinguished from y, i.e. there is no way to establish, that x and y differ.
The natural way to interpret the statement, that x and y can be distinguished, is to give some "effective" procedure for making such a distinction. Thus, the property of being distinguishable is observable [51]. Correspondingly, the property x ∼ y is refutable, hence ∼ should be closed.
The fact that the least closed tolerance for a weakly Hausdorff continuous dcpo is ↑ also is quite natural in this framework. Indeed, domain elements are thought of as being only partially known and dynamically increasing in the course of their lives.
Hence the fact that x ↑ y, that is ∃z. x ⊑ z, y ⊑ z, precisely means that x and y may approximate the same "genuine" element z, hence we cannot distinguish between them.
Since ↑ is closed in the weakly Hausdorff case, its complement is open, hence observable.
That means that when x ↑ y does not hold, there is some "finite" way to distinguish between x and y.
J ′ and the Least Closed Tolerance (a Smyth Conjecture)
Consider a continuous dcpo A. In this subsection x, y, v, w ∈ A. Recall that we defined Lemma 12.6.1 x ∼ y = { x, y | ∃ v, w . v ≪ x, w ≪ y, {v, w} is unbounded}.
Proof. Using the Border Lemma we get x ∈ Int(J y ) iff ∃v ∈ J y . v ≪ x. By the definition of J y , v ∈ J y iff y ∈ Int(I v ) i.e. ∃w ∈ I v . w ≪ y. Finally recall that w ∈ I v iff {v, w} is unbounded . ✷ It is an immediate corollary that ∼ is symmetric. x, y ∈ W \ ∼.
Since W is open, we can choose two open sets U ⊆ X and V ⊆ X, such that x, y ∈ U × V ⊆ W . Consider a pair p, r , p ∈ U, r ∈ V.p ≪ x, r ≪ y. The pair p, r is bounded, otherwise x, y ∈ ∼, so we can take z such that p ⊑ z, r ⊑ z, therefore z ∈ U , z ∈ V , so z, z ∈ U × V and z, z ∈ W . So the complement of W is not a tolerance, because it is not reflexive. ✷
Examples
In our example domain E, the pair e ∞ , 0 E is unbounded, but belongs to the least closed tolerance, since these elements cannot be distinguished by looking at the approximation pairs, e 1 , 0 E , e 2 , 0 E , . . ..
Tolerances and Lower Bounds of Relaxed Metrics
We are going to prove the following statement.
⇐. l(x, y) > 0 means C x ∩ J ′ y = ∅ or C y ∩ J ′ x = ∅. It is enough to consider C x ∩ J ′ y = ∅. Since x is the largest element of C x , we obtain x ∈ J ′ y , hence x ∼ y.
Open Problems
It might be useful to extend the Stone duality analysis to Int(J x ) and to be able to speak about the intuition behind the Lawson condition in the spirit of [51].
Another open question is as follows. If Lawson condition does not hold, can we obtain some negative results about the existence of µInfo-structures with totality property, or, more generally, about the existence of relaxed metrics with the property ∀x, y ∈ Total (A). l(x, y) = u(x, y)? Obviously, this question allows a number of variations, e.g. we know now, that when this question is restricted to the case of Info(x) = C x , such negative results can indeed be obtained.
It seems that tolerances will play an increasingly important role in domain theory. One particularly promising direction of development is to use tolerances and especially their asymmetric generalizations instead of transitivity of logical inference to formally express and study the ideas of A.S.Esenin-Vol'pin and P.Vopenka, that large numbers should be considered infinite, and long proofs and computations should be considered meaningless [56].
Conclusion
In the preface we identified the key unsolved Problem A, namely to learn how to find reasonable approximations for a sufficiently wide class of definitions of domain elements and Scott continuous functions while spending realistic amount of resources, as the main obstacle on the path of wider practical applicability of domains with Scott topology in software engineering.
In our work we achieved considerable progress in the development of analysis on such domains. In conclusion, we would like to stress the extreme importance of finding Failing the solution of Problem A, it is possible that certain mathematical constructions in domain theory might give us some hints on how to construct novel algorithms. | 2015-12-12T03:36:23.000Z | 2015-12-12T00:00:00.000 | {
"year": 2015,
"sha1": "db59502380befacc46461cb024254b573f97facd",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "792384a28faf1dc73d08edbebdc2feaf43de6884",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science",
"Mathematics"
]
} |
15045381 | pes2o/s2orc | v3-fos-license | Polyunsaturated Fatty Acid Intake and Risk of Lung Cancer: A Meta-Analysis of Prospective Studies
Background Studies have reported inconsistent results for the existence of an association between polyunsaturated fatty acid (PUFA) intake and risk of lung cancer. The purpose of this study is to summarize the evidence regarding this relationship using a dose response meta-analytic approach. Methodology and Principal Findings We searched the PubMed, EmBase, and Cochrane Library electronic databases for related articles published through July 2013. Only prospective studies that reported effect estimates with 95% confidence intervals (CIs) of lung cancer incidence for greater than 2 categories of PUFA intake were included. We did random-effects meta-analyses of study-specific incremental estimates to determine the risk of lung cancer associated with a 5 g per day increase in PUFA intake. Overall, we included 8 prospective cohort studies reporting data on 1,268,442 individuals. High PUFA intake had little or no effect on lung cancer risk (risk ratio [RR], 0.91; 95% CI, 0.78–1.06; P = 0.230). Furthermore, the dose-response meta-analysis also suggested that a 5 g per day increase in PUFA has no significant effect on the risk of lung cancer (RR, 0.98; 95%CI: 0.96–1.01; P = 0.142). Finally, the findings of dose response curve suggested that PUFA intake of up to 15 g/d seemed to increase the risk of lung cancer. Furthermore, PUFA intake greater than 15 g/d was associated with a small beneficial effect and borderline statistical significance. Subgroup analyses for 5 g per day increment in PUFA indicated that the protective effect of PUFA was more evident in women (RR, 0.94; 95% CI, 0.87–1.01; P = 0.095) than in men (RR, 1.00; 95% CI, 0.98–1.02; P = 0.784). Conclusion/Significance Our study indicated that PUFA intake had little or no effect on lung cancer risk. PUFA intake might play an important role in lung cancer prevention in women.
Introduction
Lung cancer is the leading cause of cancer-related death worldwide for both men and women, and around 1.5 million new cases are diagnosed each year [1][2]. For the past few decades, studies have shown that eicosanoids derived from polyunsaturated fatty acid (PUFA) consumption influence many physiological processes, including calcium transport across cell membranes, angiogenesis, apoptosis, cell proliferation, and immune cell function [3][4][5]. Epidemiologic studies have suggested that a healthy diet and lifestyle are critical for prevention of lung cancer [6][7][8]. Dietary fat has been closely related to lung cancer risk. Among subtypes of dietary fat, PUFA is the most promising for inhibiting carcinogenesis and reducing lung cancer risk. However, data on the effect of PUFA intake on subsequent lung cancer morbidity are limited and inconclusive.
The results of a previous prospective study indicated that high PUFA intake was associated with lower lung cancer risk [9]. In contrast, another important study showed that high PUFA intake was associated with greater risk of lung cancer morbidity [10]. Clarifying the optimal daily intake of PUFA is particularly important in the general population, as it has not been definitively determined. Here, we attempted a large-scale examination of the available prospective studies to determine the association between PUFA intake and lung cancer morbidity. We also performed a dose response meta-analysis to quantify the risk of lung cancer with an incremental increase in PUFA intake for the general population.
Data Sources, Search Strategy, and Selection Criteria
This review was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Statement issued in 2009 (Checklist S1) [11]. Any prospective study that examined the relationship between PUFA intake and lung cancer morbidity was eligible for inclusion in our study, and no restrictions were placed on language or publication status (published, in press, or in progress). We searched the PubMed, EmBase, and Cochrane Library electronic databases for articles published through July 2013 using the following search terms: (''fat'' OR ''fatty acid'' OR ''docosahexaenoic acid'' OR ''eicosapentaenoic acid'' OR ''docosapentaenoic acid'' OR ''alpha-linolenic acid'' OR ''polyunsaturated fatty acid'' OR ''omega-3 fatty acid'' OR ''n-3 fatty acid'' OR ''fish'' OR ''fish oil'' OR ''seafood'' OR ''PUFA'') AND (''lung cancer'' OR ''lung neoplasm'' OR ''lung carcinoma'') AND (''cohort'' OR ''cohort studies'' OR ''nest case-control studies'') AND (''human''). We also conducted manual searches of reference lists from all relevant original and review articles to identify additional eligible studies. The medical subject heading, methods, population, study design, exposure, and outcome variables of these articles were used to identify the relevant studies.
The literature search was independently undertaken by 2 authors (YFZ and HFG) using a standardized approach. Any inconsistencies between these 2 authors were settled by the primary author (YHZ) until a consensus was reached. A study was eligible for inclusion if the following criteria were met: (1) the study had a prospective design (prospective cohort or nest prospective case-control study); (2) the study investigated the association between PUFA intake and risk of lung cancer; and (3) the authors reported effect estimates [risk ratio (RR), hazard ratio (HR), or odds ratio (OR)] and 95% confidence intervals (CIs) for comparisons of high and low PUFA intake (with more than 2 categories of PUFA intake). We excluded all case-control studies because various confounding factors could bias the results.
Data Collection and Quality Assessment
The following data elements were collected: name of the first author or study group, publication year, country, study design, assessment of PUFA exposure, sample size, age at baseline, percentage of male patients, follow-up duration, effect estimate, endpoints reported, comparison categories, and covariates in the fully adjusted model. We also extracted the numbers of cases per person or per person-year, effect of the different exposure categories, and 95% CIs. For studies that reported several multivariable adjusted RRs, we selected the effect estimate that was maximally adjusted for potential confounders.
The Newcastle-Ottawa Scale (NOS) was used to evaluate methodological quality. The NOS is a comprehensive tool that has been partially validated for evaluating quality of observational studies in meta-analyses [12][13]. The NOS is based on the following 3 subscales: selection (4 items), comparability (1 item), and outcome (3 items). A ''star system'' (range 0-9) has been developed for assessment (Table S1). The data extraction and quality assessment were conducted independently by 2 authors (YFZ and HFG). Information was examined and adjudicated independently by an additional author (YHZ), who referred to the original studies.
Statistical Analysis
We examined the relationship between PUFA intake and risk of lung cancer on the basis of the effect estimate (RR or HR) and its 95% CI published in each study. We first used the random-effects model [14,15] to calculate summary RRs and 95% CIs for high PUFA intake compared to low PUFA intake. We subsequently transformed category-specific risk estimates into estimates of the risk ratio (RR) associated with every 5 g per day increase in PUFA intake by use of the method of generalized least-squares for trend estimation [16]. These estimates were calculated from the assumption of a linear relation between the natural logarithm of risk ratio and increasing PUFA intake. We converted fish and fish oil into PUFA, defined 100 g of fish as 2.5 g PUFA and 100 g of fish oil as 30 g PUFA. The value assigned to each PUFA category was the mid-point for closed categories, and the median for open categories (assuming a normal distribution for PUFA intake). We conbined the risk ratios for each 5 g per day increase in PUFA intake by use of random-effect meta-analysis [14]. Unless otherwise stated, we used the most adjusted risk estimate from each study as stated above. We finally conducted a dose response random-effects meta-analysis from the correlated natural log of RRs or HRs across the PUFA intake categories [16,17]. To derive the dose response curve, we modeled PUFA by using restricted cubic splines with 3 knots at fixed percentiles of 10%, 50%, and 90% of the distribution [16]. This method requires the effect measure with its variance estimate for at least 3 known categories of exposure. Heterogeneity between studies was investigated by using the Q statistic, and we considered P values,0.10 as indicative of significant heterogeneity [18,19]. Subgroup analyses were conducted for lung cancer on the basis of country, sex, assessment of exposure, and duration of follow-up. We also performed a sensitivity analysis by removing each individual study from the meta-analysis. Several methods were used to check for potential publication bias. Visual inspections of funnel plots for lung cancer were conducted. The Egger and Begg tests were also used to statistically assess publication bias for lung cancer [20,21]. All reported P values were 2-sided, and P values ,0.05 were considered statistically significant for all included studies. Statistical analyses were performed using STATA software (version 12.0; Stata Corporation, College Station, TX, USA).
Results
The results of the study selection process are shown in Figure 1. We identified 1,137 articles in our initial electronic search; 79 remained after exclusion of duplicates and irrelevant studies. After detailed evaluation, 8 prospective studies were selected for the final meta-analysis [9,10,[22][23][24][25][26][27]. A manual search of the reference lists of these studies did not yield any new eligible studies. General characteristics of the included studies are presented in Table 1.
All 8 included studies were prospective cohort studies (for a total of 1,268,442 individuals). Between 4,538 and 492,186 individuals were included in each study, and follow-up periods ranged from 7.0 to 24.8 years. Two studies were conducted in the United States [22,23], 4 in Europe [10,24,26,27], and the remaining 2 were conducted in Japan [9,25]. Study quality was assessed using the NOS (Table S1) [12]. Here we considered a study with a score $7 as being of high quality. Overall, 3 studies had a score of 9 [9,10,24], 3 studies had a score of 8 [22,26,27], and 2 studies had a score of 7 [23,25].
After pooling included studies, the summary RR showed that a high PUFA intake was not associated with lung cancer (RR, 0.91; 95% CI, 0.78-1.06; P = 0.230; Figure 2A), but potential evidence of significant heterogeneity was seen (I 2 = 67.7%; P = 0.001). The findings of dose-response meta-analysis also suggested that no association with risk of lung cancer per 5 g/day increment of PUFA intake (RR, 0.98; 95%CI: 0.96-1.01; P = 0.142; Figure 2B), heterogeneity between studies was high for lung cancer (I 2 = 69.5%; P,0.001). As a result, a sensitivity analysis was conducted, and after each study was sequentially excluded from the pooled analysis, the conclusion was not affected by exclusion of any specific study.
All studies were included in the dose response curve between PUFA intake and incidence of lung cancer. As shown in Figure 3, PUFA intake of 3.6-15.0 g/d seemed to increase the risk of lung cancer. This harmful effect was observed for PUFA intake #15 g/ d; however, PUFA intake greater than 15 g/d was associated with a small beneficial effect and borderline statistical significance. However, as shown by the P value of nonlinearity (P = 0.233), there was no evidence of a potential nonlinear relationship ( Figure 3).
Owing to a P value of ,0.10 for heterogeneity testing, we conducted subgroup analyses to minimize heterogeneity among the included studies. Overall, we noted that a high PUFA intake was associated with a reduction in lung cancer risk if the follow-up period was greater than 10 years, when excluding Veierod et al. 's study [10] (RR, 0.61; 95% CI, 0.37-0.98; P = 0.041, Figure 4A). Furthermore, we found that 5 g of PUFA intake increment per day may be a protective factor for lung cancer in women (RR, 0.94; 95% CI, 0.87-1.01; P = 0.095, Figure 4B). No other significant differences in effect were identified between PUFA intake and the risk of lung cancer.
A review of funnel plots could not rule out the potential for publication bias for lung cancer ( Figure 5). However, the Egger [20] and Begg test [21] results showed no evidence of publication bias for lung cancer (Egger: P = 0.186 for high versus low PUFA intake, and P = 0.135 for per 5 g per day increment in PUFA intake; Begg: P = 0.213 for high versus low PUFA intake, and P = 0.276 for per 5 g per day increment in PUFA intake).
Discussion
Previous observational studies of the association between PUFA intake and lung cancer risk have been inconclusive. Several casecontrol studies found a decreased risk with PUFA consumption [28][29][30]. Two other case-control studies failed to find a significant association, although the observed odd ratios were below unity [31,32]. A positive association was found in 1 study in China, but the subjects consumed fish relatively infrequently and the authors suggested the influence of a potential residual bias [33]. However, various confounding factors in case-control studies could bias the results. Furthermore, the cutoff value for optimal PUFA intake categories differed between studies. Therefore, we conducted a dose response meta-analysis of prospective studies to determine the risk of lung cacner with an incremental increase in PUFA intake.
Our current study was based on prospective studies and explored all possible correlations between PUFA intake and risk of lung cancer. This large quantitative study included 1,268,442 individuals from 8 prospective cohort studies with a broad range of populations. The findings from our current meta-analysis suggest that increased PUFA intake had little or no effect on the incidence of lung cancer. To the best of our knowledge, this is the first time a meta-analysis has evaluated, systematically and quantitatively, the association between intake of PUFA and risk of lung cancer.
Of the 8 studies examined, the majority indicated no association between intake of PUFA and lung cancer incidence, but 2 prospective studies reported a beneficial effect of PUFA in some special populations [9,26]. In a prospective study of 5,885 individuals, Takezaki et al. found that participants who consumed fish 3 or more times per week had an 81% lower risk of lung cancer than those who consumed fish less than 1 time per week [9]. Laake et al. also suggested that high PUFA intake significantly reduced the risk of lung cancer in women [26]. The pooled results of our meta-analysis were consistent with most studies analyzedi.e., no evidence of an association between PUFA intake and lung cancer risk was noted. Moreover, we discovered that the pooled RR estimate points were ,1 and had a potential trend to deviate to the left. Our dose response curve showed nonsignificant nonlinear relationships between PUFA intake and lung cancer: a low PUFA intake (3.6-15.0 g/d) seemed to increase risk of lung cancer; high PUFA intake (.15 g/d) seemed to have a small beneficial effect on risk of lung cancer. Hence, we suggest there might be a potential protective effect of high PUFA intake on lung cancer incidence; however, this protective effect may not be clinically significant and should be validated by further research.
In our current study, there was no significant difference between increased PUFA intake and the risk of lung cancer. The degree of association may be too low to detect an expected protective effect. Two possible explanations are: (1) different cooking method may moderate the effect of PUFA intake, and (2) different histological or cell types of lung cancer might provide a biased view of the study question. Previous studies showed that consumption of salty or dried fish increased risk of lung cancer [32,34]. In addition, fish oil rapidly degrades owing to oxidation and other chemical changes when exposed to air, light, heat, and processing, thereby decreasing the concentration of PUFA [35]. In regards to the type of lung cancer, previous studies suggested that high PUFA intake was associated with reduced the risk of adenocarcinomas but not squamous cell or small cell carcinomas [36].
Subgroup analyses indicated that the protective effect of increased PUFA intake was more evident in women than in men. One possible explanation for this could be higher smoking rates among men than among women (i.e., the protective effect of PUFAs on lung cancer risk may be negated by smoking among men). The proportion of current smokers was highest in the lowest category of PUFA intake [26]; it is possible that adjustment for smoking was insufficient in these studies. In never-smokers, Laake et al. found a significant positive association between PUFA intake and lung cancer risk in men, and a positive nonsignificant association in women [26]. Therefore, PUFA intake in men might be too low to detect a protective effect. In addition, we observed that the protective effect of PUFA was more evident in studies with follow-up periods greater than 10 years as compared to those with shorter follow-up periods. The reason for this difference could be that studies with shorter follow-up periods (less than 10 years) did not reach statistical significance owing to low incidence of lung cancer. In addition, only 3 prospective studies had follow-up periods of more than 10 years [9,24,26]. This conclusion may be unreliable since smaller cohorts were included in such subset. Therefore, we generated a relative result by comparing high PUFA intake with low PUFA intake, and provided a comprehensive review of the model.
Four strengths of our study should be highlighted. First, only prospective studies were included, which should eliminate selection and recall bias. Second, the large sample size allowed us to quantitatively assess the association of PUFA intake with risk of lung cancer, thus making it more powerful than any individual study. Third, the dose response analysis included a wide range of PUFA intake, which allowed for accurate assessment of the dose relationship between PUFA intake and lung cancer risk. Fourth, no meta-analysis had been performed to date to provide a comprehensive view of the association between PUFA and lung cancer risk.
The limitations of our study are as follows: (1) publication bias is very possible in meta-analyses of published studies; (2) data on histological or cell types of lung cancer and non-smoking subjects were not available so we could not differentiate effects of PUFA intake by histological or cell types of lung cancer; and (3) the analysis used pooled data (individual data were not available), which restricted us from performing a more detailed relevant analysis and obtaining more comprehensive results.
The findings of this study suggest that increased PUFA intake had no significant effect on lung cancer risk. Subgroup analyses suggested that increased PUFA intake might play an important role in lung cancer prevention in women. According to dose response curve, a PUFA intake of more than 15 g/d seemed to lower the risk of lung cancer, but the difference in risk was not statistically significant. Future studies should: (1) focus on specific populations to evaluate strategies for primary prevention of lung cancer; (2) ascertain the specific histological or cell types of lung cancer and analyze effects by type; and (3) take cooking method into consideration when assessing PUFA intake and its effects on clinical outcomes.
Supporting Information
Table S1 Quality scores of prospective cohort studies using Newcastle-Ottawa Scale.
Author Contributions
Conceived and designed the experiments: YHZ. Performed the experiments: YHZ HFG YFZ. Analyzed the data: YHZ JL FFY YFZ. Contributed reagents/materials/analysis tools: YHZ. Wrote the paper: YHZ. | 2017-04-19T18:01:06.920Z | 2014-06-12T00:00:00.000 | {
"year": 2014,
"sha1": "feaf07f2f0ee108b248f4df51622ebc7df8d4051",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0099637&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "feaf07f2f0ee108b248f4df51622ebc7df8d4051",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
237733105 | pes2o/s2orc | v3-fos-license | Implementation of Permaculture Design to Achieve Food Self- Sufficiency in Mood Forest Farm, Hualien County, Eastern Taiwan
CONTACT elok.sukmarani@gmail.com
INTRODUCTION
The Covid-19 (coronavirus disease of 2019) pandemic is a global health crisis. The COVID-19 is an infectious disease caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) [1]. The virus has affected the lives of many people in the world. Wuhan, the capital of China's Hubei province, was the location of the first detected cases of this coronavirus outbreak. At present, this pandemic hit more than 100 countries, specifically 220 countries, and territories, including Taiwan. As of May 29, 2021, a total of 169,118,995 confirmed cases with 3,519,175 deaths globally were reported by WHO [2]. All countries affected by Covid-19 are facing a difficult time overcoming the outbreak. The agricultural sector is one of the sectors that has experienced the impact of the Covid-19 pandemic. For example, limited distribution, reduced workforce, increased prices for certain commodities, and disruption of supply chains. Food security is being threatened as a result of this pandemic because it affects the supply and demand for food in society [3]. One of the strategies and efforts that can solve it is create plant diversity, design cropping patterns, optimize land, and design multifunctionality systems on the farm with the aim of practitioners being able to produce and harvest their own plants as their food supply.
Currently, permaculture is a design that is a good concept in agricultural sector. A harmonious system to integrate of landscape and people providing their food, energy, shelter, and other material and non-material needs in a sustainable way [4]. Permaculture applies a more holistic design to both farming crops and animal stock which have multifunctional to each other. For instance, the practitioners are designing food forest in ther land with the function to harvest a lot product of them such as fruit, vegetable, and animal product in one area system. For instance, the practitioners are designing food forests in their land with the function to harvest a lot of products of them such as fruit, vegetable, herbs, and other product. Besides it, permaculture has ethics and principles ecologically as a basic to implemented in surrounding us. Recently, permaculture practice has actually been spread rapidly internationally and many countries have adopted this concept [5]. Taiwan has not been able to achieve food self-sufficiency according to the country's target. Especially, with the impact of the pandemic that is currently attacking the local population of this country. While in the process to achieve food self-sufficiency, there are also many challenges. Intensive farming in this country is still high, such as the use of agricultural chemicals to achieve maximum production. Although Intensive farming practices can achieve maximum productivity, they also can have a negative impact on both human health and natural ecosystems. The farmers use highly intensive inputs and energy to achieve it, such as the use of chemical pesticides and fertilizer [6]. Therefore, an effort is needed to improve the environment, the quality, and the safety of agriculture products through ecological balance. Through the concept of permaculture, practitioners can be more productive with their environment and able to provide their food in ecology and sustainable way, without damage nature surrounding them. The practitioners concern food safety of their families through the use of organic material in their farming system. They can make problems to solutions such as rotated plants to improve the soil to make it healthier and fertile, also integrating ecological ponds around the rice fields that have multifunctions to each other.
Eastern Taiwan includes two counties, Hualien and Taitung. Eastern Taiwan has many natural resources, agricultural culture, and diverse local communities. This wealth is a great opportunity for farmers, especially permaculture practitioners in this area. The study aims to analyze the implementation of permaculture design for food self-sufficiency on farm in eastern Taiwan. One example of its implementation is carried out by a family who is a permaculture practitioner on their land called Mood Forest (牧于森林) farm, in Hualien County.
METHODS
This study was used a qualitative method with a case study approach. This study is located in one of the permaculture farms, namely Mood Forest in Pinghe Village, Shoufeng Townships, Hualien County, Eastern Taiwan. The study was used qualitative methods through interviews and participant observation, field notes, and documentation to collect data. Interviews were conducted directly with two people as permaculture practitioners on this farm. They are Hsu (Taiwanese, male, farmer) and Alinggo (Taiwanese, female, half-farmer/X-farmer, teacher) with unique and different backgrounds. Because the interview was also conducted in the participant observation activity, this interview has matched the reality on-site. In addition, the data is also supported by my field note and documentation. Participant observation is a way to collect data in naturalistic settings by who observe the common and uncommon activities of people being studied [7]. In this data collection session, we were involved in several activities on the farm location. The results of data collection through interviews, participant observations, field notes, and documentation were analyzed and obtained three main themes. There are permaculture concepts, achieve food self-sufficiency, also problem and challenge faced. Specifically, the result of the study is explained in the next section.
RESULTS AND DISCUSSIONS
Based on the data result with various methods described in the previous session, we interpret them, then compare them with several previous studies and related literature. There are three main results with their discussion as described below.
Permaculture Concept
Based on the results of interviews that have been conducted with key person (Hsu) that permaculture is one of the ways farming in agriculture. He is who has been a permaculture practitioner since 2014 in his farm, said: "Permaculture just way to achieve natural farming" "Natural farming is an idea in agriculture" The concept of permaculture that she and her family integrate into this land is the right choice and decision for their family life. Alinggo feels happy about her family's decision. The practitioners have various designs of permaculture implementation. The practitioners have crop patterns design based on microclimatology. Also, they have ecology ponds with a lot of fish, water plants, and other elements. Other designs are pest control design, herbs garden, chicken shelter with a fruit tree on this land. The location of the land has a beautiful view, then Alinggo said when she first found out about this area because under Liyu mountain and has a lot of natural resources.
"Wow here is beautiful. If I could, I hope I can have land in here" with the feeling of being happy to have owned this land and managed it together with her family, they proved it through changes in this area. They developed a small land that had nothing to a large area with various elements that they currently have present. At present, their land area is around 3000 ping (9917.3 square meters). Moreover, they also prove it by getting several certificates for their land, such as certificates for eco-friendly farming and organic farming. Environmentally-friendly farming certificates are proof of their land joining with a nongovernment association in Hualien related to the field of permaculture and environmentally friendly farming. This purpose of the association is one of the mission's government that related the policies to realize environmentallyfriendly agriculture. While organic certificates, they got in 2016 and 2018.
However, their journey while managing the land for many years, they do not know about PDC or Permaculture Design Course. Alinggo said: "PDC? I am not sure talk about this" This activity is an important activity that is usually followed by practitioners to learn permaculture both for beginners or expert people. They will get a certificate after taking this course. The activities are usually carried out such as attending seminars but also practice on the field. Alinggo family is only active in participating in the activities of permaculture associations or environmentally friendly and organic agriculture. For example, presentations related to their land if there is a visit event, permaculture volunteer activities, and sometimes participating in other members' farms.
Uniquely, Hsu has a background in the journalist field on previous his job, he never had any experience in agriculture before, it did not make him hesitate in pursuing agriculture for these 7 years. At present, he has the ability to make bake, as a saxophonist, be a runner, and a farmer, especially a permaculture practitioner. In the beginning, He learned from reading various books and websites, attending agricultural seminars, experimenting, researching by himself, and various experiences from the important person who expert in the agriculture sector. Unlike Alinggo, whose background has in the agricultural sector. Alinggo had agriculture experiences because of her childhood with her family in the agriculture field. Alinggo said with her enthusiastic expression: "He didn't have a childhood like me. But he learned it. He used a website, just like to bake" "He learned how to bake, learned how to be a farmer, how to be a saxophone player" In addition, when they face a problem in their land, they never give up on solving it and try to make various designs as a challenge for them such as choose the kind of plant, new pest management, and other designs. They adopt the mindset of continuing to try when they fail to do something on their soil, as Hsu said that: "Mistake and try it again" and Alinggo also said that: "If we applying permaculture design, we must work with seriousness and patience because we are working with nature, and we hope that nature will do well and can cooperate with us" Therefore, they never give up on their land, always try and learn. Not only do they both play a role in managing the land, but the involvement of their children is also very helpful and proves that their togetherness can make the family very harmonious and happy. Their children have drawing, read a book, and photographer hobbies. It no wonders, if this family has a principle is the learning through art for their permaculture farm.
Achieve Food Self-Sufficiency
By the integration of permaculture, Hsu can design various cropping patterns based on the microclimate, soil type, plant kind, and other functions. Then, his family can get a lot of benefits from every design function that he integrates and implements on his land. For example, his food forest design. He integrated several fruit trees such as guava, lemon, passion fruit, herb plants, ginger, and raising hen chicken, also rabbit in one area (Figure 1). Through the various functions of elements in one area, they can harvest fruit in a long or short period of time, and they can harvest eggs from their hens every month. a b c These elements also support to each other. For example, chicken manure in this area is a source an organic fertilizer for the plants to grow and develop very well. Also, the existing fruit trees are a shelter for the life of chickens, and the chickens often sleeping on the branches of the tree. Based on my observation and data analysis that the fruit tree's existence not only works for chickens but also for the life of rabbits and the small plants underneath. The fallen dry leaves also of course also function as organic matter to fertilize the surrounding soil. According to Alinggo (Hsu wife), they have many reasons for every element of their design. She said: "I grow some types of ginger in under greenhouse and I also plant passion fruit on the top greenhouse with the function to reduce the sunlight". it correlates with literature which statemented that in Sub-tropical regions, ginger is often interplant in agroforestry systems, because ginger is moderately tolerant of shade is well adapted to agroforestry systems [8].
Even when they harvest fruit or vegetables, they always share them with friends, neighbors, or visitors who come. This is because they often feel self-sufficient and share their land products as a form of caring for others. Their actions are one of permaculture ethics implementation, namely care of people because they care to each other, itself, community, and society. They have also applied other ethics such as caring for the earth such as not produce waste, and even utilizing natural resources in balanced and sustainable principles [4], [9].
In addition, their paddy fields are an important element in this land because it has economic value for their lives. Every harvest time, they always have customers who will order the rice, even though some of them are sold but they can still set some of their harvests to provide food for their own families. They have a rice planting season twice a year, specifically in wintertime and summertime. The practitioners have a rule for rice customers. The rice customers must commit order their rice three months before harvest. Alinggo and Hsu provide a social media group such as LINE and a Facebook page related to information on products from their land, especially rice yields. They manage their rice fields independently, although sometimes they use rice plowing field machines, they still go to the field to manage it. Alingo said: "
Even we used the machine on the big land, we still need to have some opportunity to do something like this (use by hand) because the big machine only can go in a straight line and still have this space"
The practitioner (Hsu) manages the rice paddy field from land management, soil and plant nutrients management, harvest till post-harvest management with his family. Uniquely, their packaging of rice products has the name of the farm and drawings their daughter made. It is also for other products, such as rosella jam.
In these two years, they have only one period of planting rice once. However, they can produce abundant yields, even more than before when they planted rice paddy twice per year. They prefer to rest their land by planting soybeans to improve soil nutrition. When the next planting period, their land will produce better and more abundant rice production than before. Moreover, the yield of this soybean crop is also for the availability of vegetables for their food. This management correlated with the literature regarding the role of soybean plants. The soybean is one of the legumes plants whose roots can reserves nutrients through a nitrogen fixation process by rhizobacteria. So, they can provide nitrogen nutrients to make fertile soils and then affect paddy plants in the next planting period [10]. The practitioners wish to get high yields in the next period through concern to make their fertility soil first. Based on consideration in every decision in land management here, they are never worried about the lack of stock of rice, vegetables, even fruit or eggs from their livestock as a food supply for this family and they are even able to provide their own food or independently. This condition is helpful when they faced the impact of the COVID-19 pandemic, especially through the integration of permaculture design on their land
Problem and Challenges Faced
In mood forest farm, the main problem is pests on the rice paddy fields. Based on the interview result, the pests attack their rice paddy is a small bird. Their rice paddy fields are the main target of little birds may be because of the strategic location of their land. Result of our observations, it obvious that there were no paddy fields like Mood Forest field. There are only vegetable and fruit farm, annual plant, reservoirs, wild grasses, and empty land surrounding Mood Forest location. Therefore, it is not wondrous that little birds like come to eat paddy grains. Supported by Hsu's statements that the grains of rice paddy were damaged and even empty of being eaten by little birds. We tried to identify this little bird in question. Based on its anatomical features, this bird is a sparrow (Passer spp.) which includes the Family Ploceidae. This bird is regular pests of ripening grain crops, especially rice paddy [11]. Hsu used natural enemies or predators of this bird and other creative designs to control the pest here. Recently, Hsu and his family designed a tool for controlling this bird. The use of giant bird kites is one of the creative ideas for controlling rice pests since the last 5 months by practitioners here (Figure 1a). They put up 3 strong bamboo poles in the edges of the field and each bamboo was connected with a rope which functions as a place for the kites to move when pushed by the wind movement. This design is a simulation of the presence of large birds upcoming and circling paddy fields that can become enemies for birds to prevent them from coming. They also implemented other creative solutions, it as bamboo high poles for eagle place in the middle of their rice fields (Figure 1b).
Figure 2. The Giant Bird Kite and Eagle Bamboo Pole
A pole made of bamboo with a very tall size was deliberately placed in the middle of the paddy field as a place for the eagle to stop. It is known that the eagle is one of the natural enemies of these birds which often become a pest of Mood Forest rice paddy field. One of the most famous predators in the bird world is the eagle [11].
Alinggo also explained that they not only used the eagle stand and bird kites faced the little bird. They also used the mechanic method for pest management, which is the use of firecrackers (figure 2). However, they won't use this tool every time because it costs a lot of money to buy new firecrackers. They will set on firecrackers when the birds come on a large scale, such as more than 15 birds. But, after a few minutes burned the firecrackers, the birds come back to the fields. Therefore, it is still a challenge for them to control bird pests. Moreover, the problem and challenges faced by practitioners depend on the character and condition of the land. Relate to previous research that practitioners in permaculture must understand the available elements in detail surrounding them that they can find and decide which design patterns are appropriate to land conditions [12]. The location of Mood Forest in the countryside and under the mountain specifically, but they were can to optimize the land with various permaculture designs without damaging nature surrounding them. Conversely, the problem and challenges face practitioners in the urban area is different, such as the limitation area of farming. But through permaculture design, practitioners might use their rooftop, yard or create the design on a small scale. Thus, they can still be productive and harvest from the elements they have. Then, they also get harvest products as a food supply for their own and their families.
CONCLUSIONS
The concept of permaculture has been proven to increase productivity, harmony in human life with their nature, and sustainability of the environment through using natural principles in an ecological way. Furthermore, this concept is one of the best ways to be food self-sufficient in the Covid-19 pandemic situation through integrated several designs on the land. The practitioner family always tries to make optimum use of their yard and land. Especially, those elements that have economic value as a result of food that can support their daily lives as food supply and also create income for them through rice sales for customers. In addition, they also have created income from other product sales such as chicken eggs, processed foods, herb products, and many other products. Besides rice paddy, they have crops that include various beans, chili, lettuce, banana, guava, papaya, lemon, passion fruit, pineapple, loquat, also other fruits and vegetables. For livestock in mood forest including poultry such as various chicken types and some geese and fish in their pond. At the beginning of getting this land, practitioners explained an emotional journey until finally now they get what they want easily on their land, without having to buy it in the market, except for other products that they do not have. Even though they buy products in the market, they still prioritize organic products even at expensive prices because their mindset is based on permaculture principles and ethics. | 2021-09-27T20:55:33.837Z | 2021-07-25T00:00:00.000 | {
"year": 2021,
"sha1": "feacbc6d86fa30e5e6dadc07a8d7b2656b6d73c7",
"oa_license": "CCBY",
"oa_url": "http://ejournal.ppi.id/index.php/taiwanproceeding/article/download/117/76",
"oa_status": "HYBRID",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "6dfd4764af13f6487117bf8400bf59a2ae1eb4b5",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Geography"
]
} |
117207165 | pes2o/s2orc | v3-fos-license | Borsuk-Ulam type theorems for metric spaces
In this paper we study the problems of the following kind: For a pair of topological spaces $X$ and $Y$ find sufficient conditions that under every continuous map $f : X\to Y$ a pair of sufficiently distant points is mapped to a single point.
Introduction
In this paper we are going to give new proofs, using the recent ideas of M. Gromov, to the classical Borsuk-Ulam and Hopf theorems and their generalizations, and study some their consequences, as well as separate results, in the spirit of Urysohn width and Gromov waist.
Recall the famous Borsuk-Ulam theorem [3]: Borsuk, S. Ulam, 1933). Under any continuous map f : S n → R n some two opposite points are mapped to a single point.
A deep generalization of this result is the Hopf theorem [7]: Hopf, 1944). Let X be a compact Riemannian manifold of dimension n and f : X → R n be a continuous map. For any prescribed δ > 0, there exists a pair x, y ∈ X such that f (x) = f (y) and the points x and y are connected by a geodesic of length δ.
After these results, different definitions of width of a map were introduced by Kolmogorov, Alexandrov, and Urysohn; see the definitions in the book [14] (and the book [13] in English). Generally, the width is some guaranteed size (in some sense) of a preimage of a point y ∈ Y under any continuous (or linear in some definitions) map f : X → Y , see Section 6 for more precise definitions. Even for the simplest domain spaces like the Euclidean unit ball B n and the round sphere (sphere with the standard Riemannian metric) S n not all values of such widths are known.
In the papers [5,12] another particular kind of width (called waist there) 1 was studied, based on the volume of the preimage of a point, and the waist of the round sphere was found to be the volume its equatorial subsphere of appropriate dimension. In Section 7 we give some results on this kind of width/waist too, though our present technique is limited to the simple case of codimension 1 map.
Our presentation is greatly inspired by the results of [6], where the estimates for the size of the preimage of a point were proved using the technique of "contracting in the space of (co)cycles". One of the questions addressed in this paper is how this technique can be applied to the Borsuk-Ulam and Hopf theorems. Such an application turns out to be possible and these old theorems are generalized (see Theorems 2.3 and Theorem 4.1).
Acknowledgment. The authors thank Evgeniy Shchepin and Vladimir Dol'nikov for useful discussions and remarks.
A Borsuk-Ulam type theorem for metric spaces
We are going to utilize the ideas of M. Gromov [6] to give a coincidence theorem. Let us make a few definitions.
Definition 2.1. Let X be a topological space. Denote P X the space of paths, i.e. the continuous maps c : [0, 1] → X. This space has a natural Z 2 -action by the change of parameter t → 1 − t, and a natural Z 2 -equivariant map π : P X → X × X, c → (c(0), c(1)).
Such short path maps may be given by assigning a shortest path to a pair of points in a metric space. If X is a compact Riemannian manifold then such short path maps do exist. Now we are ready to state: . Suppose X is a compact manifold of dimension n, Y is another manifold of dimension n, and f : X → Y is a continuous map of even degree. Then for any short path map s : X × X → P X there exists a pair (x, y) ∈ D(s) such that f (x) = f (y).
The classical Borsuk-Ulam theorem [3] follows from this theorem by considering X = S n and s to be the shortest path map in the standard metric. Theorem 1.2 (of Hopf) does not follow from this theorem because here we may only obtain an inequality on dist(x, y). An advantage of Theorem 2.3 is that the codomain Y may be arbitrary.
Space of cycles and the proof of Theorem 2.3
We start from explaining the main ideas underlying what Gromov calls "contraction in the space of cycles" [6] in a particular case. Denote cl 0 (X; F 2 ) the space of 0-cycles mod 2 in X, that is the space of formal finite combinations x∈X a x x with a x ∈ F 2 with an appropriate topology.
A more tangible description of cl 0 (X; F 2 ) is the union over k ≥ 0 of spaces of unordered 2k-tuples B(X, 2k) ⊂ X ×2k /Σ 2k . Informally, the topology in cl 0 (X; F 2 ) = k≥0 B(X, 2k) is such that when two points of a set c ∈ B(X, 2k) tend to a single point then they "annihilate" giving a configuration in B(X, 2k − 2) in an obvious way, and conversely a pair of points may be "created" from a single point giving a configuration in B(X, 2k + 2).
In the case when X is an n-dimensional manifold we define the canonical class ξ in H n (cl 0 (X; F 2 ); F 2 ) as follows. Any n-dimensional homology class of cl 0 (X; F 2 ) can be represented by a chain c, which is given by a map of an n-dimensional mod 2 pseudomanifold K to cl 0 (X; F 2 ). Considering any element of cl 0 (X; F 2 ) as a subset of X we may consider c as a set valued map from K to X. Its graph Γ c is a subset of K × X, which is again a mod 2 pseudomanifold. Hence the degree mod 2 of the natural projection Γ c → X is well defined. This degree will be the value ξ(c) by definition. Another informal way to define ξ would be to count how many times a generic point x 0 ∈ X participates in the 0-cycles from the chain c.
Now we return to the proof of the theorem. From the compactness considerations it is sufficient to prove the theorem for smooth generic maps f . In this case we may define the natural map f c : Y → cl 0 (X; F 2 ), which maps any y ∈ Y to the F 2 -cycle where c(x) is the mod 2 multiplicity of the map f at x. This map is well-defined because the degree of f is even by the hypothesis. The image of f c represents an n-dimensional mod 2 homology class in cl 0 (X; F 2 ) and by the definition of the fundamental class ξ it is obvious that ξ(f c (Y )) = 1. Therefore the map f c is homotopically nontrivial.
But we are going to deform the map f c to the constant map by a homotopy h t , using the short path map s. Put We have to check whether this map is continuous in y and t. If the preimage f −1 (y) changes by "annihilating" a pair points or "creating" a pair of points, the components of h t (y) are also "annihilated" or "created" pairwise (here we use the Z 2 -equivariance of the short path map and its behavior over the diagonal).
If the parameter t approaches 0 then h t approaches f c , because in every f −1 (y) we have an even number of points with even multiplicities c x , so in the expression of h t we approach every point x ∈ f −1 (y) (such that c x = 1) odd number of times. If t approaches 1 the points s(x 1 , x 2 )(t/2) and s(x 2 , x 1 )(t/2) tend to "annihilate" (and do "annihilate" at t = 1), and therefore h 1 maps the whole Y to zero cycle. Thus the proof is complete.
Remark 3.1. In [9] a simplified version of the reasoning in [6], in the particular case of the problem of probability of covering by a simplex, was presented, which avoids an explicit use of the space of (co)cycles. In the above proof a similar trick is also possible in the following way: Assume the map f to be generic (or even piece-wise linear). Then for any y ∈ Y consider the finite set f −1 (y) and the full graph (1-dimensional complex) G y on the vertices f −1 (y). Denote the union of these full graphs over y ∈ Y by G f . With some natural topology it can be interpreted as an abstract chain (in piece-wise linear case it can be made rigorous) with boundary ∂G f = X = y∈Y f −1 (y) (here we use the even degree of f ). If any edge of G y ⊂ G f can be realized in X (for example, using a short path map) then G f is mapped to X continuously. So G f becomes an (n + 1)-dimensional chain in C n+1 (X; F 2 ) with boundary ∂G f = [X] mod 2. But the fundamental class of X mod 2 cannot vanish, which is a contradiction.
Hopf type results
The method of contraction in the space of cycles allows to prove the following generalization of the Hopf theorem: is an open manifold of dimension n, and f : X → Y is a continuous map. Assume that X has a metric with injectivity radius ρ and 0 < δ ≤ ρ. Then there exist a pair of points x, y ∈ X at distance δ such that f (x) = f (y).
Remark 4.2. Compared to the Hopf theorem, in this theorem we assume additionally that δ is at most the injectivity radius, but we allow arbitrary open manifold in place of R n as the codomain.
Proof. We mostly follow the proof of Theorem 2.3. Assume that f is generic and consider the preimages of a regular value y ∈ Y . Since Y is open the degree of f is even and f −1 (y) consists of even number of points.
Assuming that no two points in f −1 (y) are at distance δ, make a graph G y on vertices f −1 (y) and edges corresponding to pairs x, y at distance less than δ. By the assumption on the injectivity radius this graph can be drawn by shortest paths on X and depends continuously on y while y does not cross special values of f .
From general position considerations, for any two regular values y ′ and y ′′ , it is possible to connect them by a path so that the graph G y while passing from y ′ to y ′′ is modified by a series of "simple events" (corresponding to folds of the map f ), when a pair of connected vertices x 1 and x 2 is added to G y and they get connected to the same set of For generic f , let S 1 ⊂ Y be the set of special values of f ; it has codimension at least 1. Let the set S 2 ⊂ S 1 correspond to "not simple events", that is S 1 \ S 2 is a fold of f . For generic f , the set S 2 has codimension 2 in Y and its preimage f −1 (S 2 ) has codimension 2 in X. We will ignore S 2 in the reasoning with the fundamental class of X or Y , because the homology is not affected by codimension 2 changes. Now we want to repeat the part of the proof of Theorem 2.3 using the homotopy in the space of cycles along the edges of G y : Like in Remark 3.1, this homotopy may be interpreted as an (n + 1)-dimensional chain in X. But unlike the proof of Theorem 2.3, the mod 2 boundary of this chain is not the fundamental class [X], but is the set of those points x ∈ X \ f −1 (S 1 ) that come with odd degree in their corresponding graph G f (x) . Fortunately, we will show that actually all the points of X \ f −1 (S 1 ) have odd degrees in their G f (x) and the proof can be finished similar to the proof of Theorem 2.3.
Without loss of generality assume that X is connected and move a point has codimension at least 2 and does not spoil the connectedness). During such a move there may be two possible modifications of the graph 2) the vertex x collides with another vertex x ′ in G f (x) and they exchange. Because their sets of neighbors are the same, N(x) \ x ′ = N(x ′ ) \ x, then the degree of x does not change at this event.
Therefore for any x ∈ X \ f −1 (S 1 ) the degree of x in G f (x) is the same mod 2. Now remember that Y is open and X is closed, then for some y ∈ Y the graph G y must be empty and while moving to a nonempty graph it will first generate a pair of points connected by an edge. Hence for some point x ∈ X \ f −1 (S 1 ) its degree in G f (x) must be odd and therefore it must be odd for every x ∈ X \ f −1 (S 1 ). So the image of h t (y) is a chain in C n+1 (X; F 2 ) (see Remark 3.1) with boundary [X] mod 2, which is a contradiction, because X is closed.
Another approach to Hopf type results is possible, following [16]. Informally, we may increase the dimension of Y , drop the compactness assumption on X, but require an assumption on its Stiefel-Whitney classes: Theorem 4.3. Let f : X → Y be a continuous map between manifolds that induce a zero map on cohomology modulo 2 in positive dimensions. Supposew k (T X) = 0 (the dual Stiefel-Whitney class), dim X + k − 1 ≥ dim Y , X is a Riemannian manifold, and δ is a prescribed real number. Then there exists a pair x, y ∈ X such that f (x) = f (y) and the points x and y are connected by a geodesic of length δ.
Proof. Consider the space S X of pairs (x, v), where x is an arbitrary point in X and v is a unit tangent vector at x. This space has a natural For Z 2 -spaces the following invariant is well-known. The natural Z 2 -equivariant map to the one-point space π S X → pt induces the map of the equivariant cohomology is a polynomial ring with onedimensional generator t. The maximal power of t that is mapped nontrivially to the equivariant cohomology of S X is called the homological index of S X and denoted ind S X . In [4] the following is proved: take the maximal k so that the dual Stiefel-Whitney class w(T X) is nonzero, then ind S X = dim X + k − 1, under the assumption of this theorem ind S X ≥ dim Y . Now consider the map h : S X → X defined as follows: start a geodesic from x with tangent v and consider its point h(x, v) at distance δ/2 from x. Now the composition f • h maps S X to Y and induces a zero map on the mod 2 cohomology of positive dimension. By the main result from [16] we see that some two pairs (x, v) and (x, −v) should be mapped to the same point, which gives the required pair connected by a geodesic of length δ.
A possible approach to the multiplicity of generic smooth maps
In the spirit of the proof of Theorem 2.3 it makes sense to consider the following construction. Let X be a closed n-dimensional smooth manifold. Filter the space cl 0 (X; Z) (of zero dimensional cycles with integer coefficients) as follows: assign to the cycle x∈X a x x the norm x∈X |a x | and denote cl (k) 0 (X; Z) = {c ∈ cl 0 (X; Z) : |c| ≤ k}. Consider another n-dimensional closed smooth manifold Y . For a generic smooth map f : X → Y of degree 0 we construct the cycle map f c : Y → cl 0 (X; Z) as above. As in the of proof of Theorem 2.3, it is known that the canonical class ξ ∈ H n (cl 0 (X; Z); Z) does not vanish on f c (Y ).
An important question is estimating from below the multiplicity of f (see [6,8]), which is equal to the least number k such that cl (k) 0 (X; Z) ⊇ f c (Y ). One possible way to show that the multiplicity of any smooth generic f : X → Y is at least k is to show that the restriction ξ| cl (k−1) 0 (X;Z) cannot be detected on a continuous image of Y . A particular case Y = R n ⊆ S n would follow from proving that the restriction ξ| cl (k−1) 0 (X;Z) is aspheric (not detectable on π n (cl (k−1) 0 (X; Z))).
With the space of cycles cl 0 (X; F 2 ) with mod 2 coefficients the above construction becomes even more explicit. Since cl 0 (X; F 2 ) = j≥0 B(X, 2j), we describe the filtering as cl j=0 . Here the unordered configuration spaces B(X, 2j) resemble the classical approach to the multiplicity using configuration spaces (see [8]).
A remarkable fact is the Dold-Thom-Almgren theorem [1] that describes the homotopy groups of this cycle space: π * (cl 0 (X; Z)) =H * (X; Z). In [1] it was proved for homology with integer coefficients, but it seems that the version for mod 2 coefficients also works. But it seems that there is no satisfactory description of the homotopy groups of cl (k) 0 (X; Z) or cl (k) 0 (X; F 2 ), which prevents us from developing this approach at the moment.
Results on the Urysohn type width
Let us give the definition of the Urysohn type width: where inf is taken over all continuous maps f : X → Y .
Informally, the width guarantees that some two far enough points are mapped to a single point under every continuous map f : X → Y . In the book [14] the value inf dim Y =k w(X, Y ) (the infimum over all k-dimensional polyhedra) is called the Urysohn width and denoted u k (X).
Note that if X is a Riemannian manifold and ρ(X) is its injectivity radius, then for every two points x, y ∈ X at distance < ρ(X) there exists a unique shortest path between x and y depending continuously on the pair (x, y), see [11] for example.
The Borsuk-Ulam theorem asserts that w(S n , R n ) = π (we measure lengths on the sphere by geodesics), in [16] it is proved that R n can be replaced by any open n-dimensional manifold. A general results is: Corollary 6.2. The inequality w(X, Y ) ≥ ρ(X) holds for a compact Riemannian manifold X and an open manifold Y such that dim X = dim Y . Remark 6.3. The particular case Y = R n (with n = dim X) of this theorem follows from the Hopf theorem.
Proof. Apply Theorem 2.3 to the shortest path map, which is well defined for pairs x, y ∈ X closer than ρ(X). The even degree of a map f : Similarly, Theorem 4.3 implies: Corollary 6.4. The inequality w(X, R N ) ≥ ρ(X) holds for a Riemannian manifold X and N ≤ dim X + k − 1, where k is the maximum k withw k (T X) = 0. Now we are going to discuss the case when the target space is not a manifold, which is important in the definition of the Urysohn width. Definition 6.5. Suppose X is a compact Riemannian manifold. Let κ(X) be the maximum number such that for any 0 < δ < κ(X) any ball in X of radius δ is strictly convex. Call κ(X) the convexity radius. Remark 6.6. Obviously ρ(X) ≥ 2κ(X), because touching strictly convex balls can intersect at one point only. It is also known that κ(X) > 0 for compact Riemannian manifolds.
Following the reasoning in [14] we prove: Theorem 6.7. Let X be a compact n-dimensional Riemannian manifold. Suppose f : X → Y is a continuous map to a polyhedron. For dim X ≥ n, we also require that f is surjective and f * : Before presenting the proof we discuss some consequences of this theorem. Corollary 6.8. We have the following estimate for the Urysohn width of an n-dimensional closed Riemannian manifold: Remark 6.9. For a sphere S n this theorem gives u n−1 (S n ) ≥ π/2. In fact, it is known [14, pp. 84-85, 268] that for a unit Euclidean ball u n−1 (B n ) ≥ 2n+2 n (the edge length of a regular simplex inscribed into B n ), and using a similar method for the half-sphere S n + of the round sphere S n it can be proved that u n−1 (S n ) ≥ u n−1 (S n + ) ≥ arccos − 1 n . Thus Theorem 6.7 is asymptotically sharp.
Remark 6.10. In a private communication E.V. Shchepin told that u n−1 (B n ) = 2n+2 n . In order to prove the equality we have to give an example of a map. Let us take the orthogonal projection f to the inscribed simplex ∆ ⊆ B n , and then take the PL-map g of the barycentric subdivision of ∆ to the cone over the skeleton 0 * ∆ n−2 . The map g is the identity over the skeleton ∆ n−2 , and maps any center of a facet to 0. It can be checked by hand that any preimage of a point under g • f is a cone over a regular simplex of dimension ≤ n, and its diameter is at most 2n+2 n . The values u k (B n ) are not known precisely for n/2 < k < n − 1.
Since every k-dimensional polyhedron can be embedded into some 2k-dimensional manifold [17, Lemma 7.1], Theorem 2.3 implies: Corollary 6.11. We have the following estimate for the Urysohn width of an n-dimensional closed Riemannian manifold for k ≤ n 2 : u k (X) ≥ ρ(X).
Proof of Theorem 6.7. Assume the contrary. Take fine enough subdivision of the polyhedron Y so that for every vertex v ∈ Y the preimage of st v has diameter ≤ δ < κ(X). This is possible from the standard compactness reasoning. For every vertex v ∈ Y consider the set X v = f −1 (st v), it is nonempty by the assumption. Since diam X v ≤ δ we can select a point φ(v) ∈ X v and note that X v ⊆ B δ (φ(v)). Now let us extend the map φ from the vertex set of Y to the map φ : Y → X. If σ is a simplex in Y with vertices v 1 , . . . , v k , then the sets X v 1 , . . . , X v k have a common point p(σ) (any point in the preimage f −1 (σ)). Therefore the set {φ(v 1 ), . . . , φ(v k )} is contained in B δ (p(σ)) and it makes sense to consider its convex hull C(σ) (because δ < κ(X)). Note that such convex hulls are homotopy trivial subsets of X, and therefore we can extend φ by induction so that for every simplex σ ∈ Y the image φ(σ) is contained in C(σ).
Since δ < κ(X) < ρ(x) the map h is homotopy equivalent to the identity map of X. Hence h * : H n (X; F 2 ) → H n (X; F 2 ) is the identity map, which contradicts the equality h * = f * • φ * and the cohomology condition on f .
Relation to the Gromov waist
In [5] (see also [12]) another notion similar to width was introduced: where inf is taken over all piecewise smooth generic maps f : It is shown in [5,12] that γ(S n , R m ) (here S n is the round sphere) equals the (n − m)dimensional volume of the equatorial subsphere S n−m ⊂ S n . In fact, a stronger result in the spirit of concentration phenomena was proved, but we only discuss the result about volumes here.
In [10] is is shown that R m on the right hand side can be replaced with arbitrary m-dimensional manifold. The open question (see [5,Question of page 194]) is: Can one one bound from below γ(S n , K) (defined as (n − m)-dimensional volume) for any m-dimensional polyhedron K?
In this section we investigate this question and some its generalizations in the simplest case, when the map f has codimension −1 (i.e. drops dimension by 1). In this case the preimage of a point is generally something 1-dimensional. Theorem 7.3. Let Y be an (n − 1)-dimensional polyhedron, then γ(S n , Y ) ≥ π.
Proof. Let f : S n → Y be a generic smooth map. Put X y = π 0 (f −1 (y)), that is the set of all connected components of f −1 (y). On the unioñ Y = y∈Y X y we define the metric as the minimal distance between disjoint compacta. The map f is a composition of continuous maps f = g •f so that Note that all the preimages g −1 (y) are zero-dimensional, soỸ has dimension at most n − 1.
Using the result of K. Sitnikov, cited in [14, p. 268], we note that for some pointỹ ∈Ỹ the preimage W =f −1 (ỹ) is not contained in an open half-sphere of S n . By the definition ofỸ , the set W is connected, hence it intersects any equatorial subsphere S n−1 ⊂ S n . By subdividing W into smaller segments it is easy to show that the probability of intersection between W and some equatorial S n−1 is at most vol 1 W π , so vol 1 W ≥ π.
Remark 7.4. Note that the estimate of [5,12] is better: vol 1 W ≥ 2π. This can also be explained in the above terms: if Y is a manifold, then for a generic f the set W is a manifold, i.e. topologically a circle. Hence it intersects a generic equatorial subsphere in at least two points! Now it makes sense to state the following conjecture, which would answer Question 7.2: Conjecture 7.5. Let Y be an m-dimensional polyhedron, n > m. For every generic map f : S n → Y there is a connected subset W ⊆ S n such that f (W ) is a point and W intersects any equatorial subsphere S m ⊂ S n .
We also extend Theorem 7.3 to any Riemannian manifold as the domain space: Theorem 7.6. Let X be a compact n-dimensional Riemannian manifold and Y be an (n − 1)-dimensional polyhedron, then γ(X, Y ) ≥ 2κ(X).
Proof. First pass to the mapf : X →Ỹ with connected preimages, as in the proof of Theorem 7.3. Now assume the contrary: for anyỹ ∈Ỹ the preimage W =f −1 (ỹ) has length < 2κ. From the compactness reasoning we may assume that vol 1 W < 2δ < 2κ for all W . Since W is connected it has a unique ball B(W ) of radius ≤ δ that contains W and is the least ball containing W . The rest of the proof follows the proof of Theorem 6.7 by constructing a map h : X → X, which has (n − 1)-dimensional image and is homotopic to the identity because of the condition dist(h(x), x) ≤ δ.
We can improve Theorem 7.6 in the case, when Y is a manifold, thus extending the main result of [5,12,10] (in the particular case of codimension −1 maps) to any CAT(1) metric space as the domain: Theorem 7.7. Let X be a compact n-dimensional manifold with CAT(1) metric and Y be an (n − 1)-dimensional manifold, then γ(X, Y ) ≥ 2π.
The proof will follow from several lemmas, all the spheres are assumed to be the standard Euclidean unit spheres. Lemma 7.8. Suppose C is a closed path on the unit sphere S 2 with length not greater than 2π. Then C lies in a hemisphere.
Proof. We need the Kirszbraun theorem for the sphere [15]. It asserts that for any U ⊂ S 2 and any 1-Lipschitz map f : U → S 2 there exists 1-Lipschitz extension f : S 2 → S 2 (i.e f | U = f ). Now consider a geodesic circle σ on S 2 and let z be one of its two centers. There obviously exists a 1-Lipschitz map f : σ → C. This map can be extended to a 1-Lipschitz map f : S 2 → S 2 . Thus the distance from the point f (z) to any point in C = f (σ) is at most π/2. This means that the hemisphere with center f (x) contains C. Proof. Suppose a ′ is a reflection of a with respect to m. The perimeter of the quadrangle aba ′ c is less than 2π. By Lemma 7.8 the quadrangle aba ′ c is contained in a hemisphere, and therefore d(a, a ′ ) = d(a, m) + d(m, a ′ ) = 2d(a, m).
Lemma 7.10. Let C be a closed curve in X (a CAT(1) metric space) with length ℓ(C) < 2π. Then C can be covered by a ball with radius ≤ ℓ(C)/4 < π/2.
Proof. Denote ℓ(C) simply by ℓ. Let a and b be two points on C that correspond to parameters 0 and ℓ/2. Suppose m is a midpoint of the geodesic [ab]. Let us show that the ball with center at m and radius ℓ/4 covers the whole C. Let x be any point on C with parameter α. Without loss of generality we assume that α < ℓ/2. Then d(x, a) ≤ α and d(x, b) ≤ ℓ/2 − α.
By the definition of a CAT(1) metric space, the distance d(m, x) is less or equal to the corresponding median in the model triangle on S 2 . In other words, if we consider a triangle △a ′ b ′ x ′ with the same as △abx side lengths on S 2 ; this is possible because the perimeter of △abx is < 2π. Then the length of the median [x ′ m ′ ] will be ≥ d(x, m). Applying lemma 7.9 we obtain: d(x ′ , m ′ ) ≤ ℓ/4, and therefore d(x, m) ≤ ℓ/4. This is exactly what we need.
Proof of Theorem 7.7. Note that for any generic f : X → Y and a generic y ∈ Y the preimage f −1 (y) is a union of closed curves. Then we argue as in the proof of Theorem 7.6 and use Lemmas 7.10 and 7.11. | 2012-09-06T10:37:26.000Z | 2012-09-06T00:00:00.000 | {
"year": 2012,
"sha1": "c0af91a9d57a2b8ea623e3ef0f35b00bcb0ecb20",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "c0af91a9d57a2b8ea623e3ef0f35b00bcb0ecb20",
"s2fieldsofstudy": [
"Mathematics"
],
"extfieldsofstudy": [
"Mathematics"
]
} |
13225399 | pes2o/s2orc | v3-fos-license | Acquired bloodstream infection in the intensive care unit: incidence and attributable mortality
Introduction To estimate the incidence of intensive care unit (ICU)-acquired bloodstream infection (BSI) and its independent effect on hospital mortality. Methods We retrospectively studied acquisition of BSI during admissions of >72 hours to adult ICUs from two university-affiliated hospitals. We obtained demographics, illness severity and co-morbidity data from ICU databases and microbiological diagnoses from departmental electronic records. We assessed survival at hospital discharge or at 90 days if still hospitalized. Results We identified 6339 ICU admissions, 330 of which were complicated by BSI (5.2%). Median time to first positive culture was 7 days (IQR 5-12). Overall mortality was 23.5%, 41.2% in patients with BSI and 22.5% in those without. Patients who developed BSI had higher illness severity at ICU admission (median APACHE III score: 79 vs. 68, P < 0.001). After controlling for illness severity and baseline demographics by Cox proportional-hazard model, BSI remained independently associated with risk of death (hazard ratio from diagnosis 2.89; 95% confidence interval 2.41-3.46; P < 0.001). However, only 5% of the deaths in this model could be attributed to acquired-BSI, equivalent to an absolute decrease in survival of 1% of the total population. When analyzed by microbiological classification, Candida, Staphylococcus aureus and gram-negative bacilli infections were independently associated with increased risk of death. In a sub-group analysis intravascular catheter associated BSI remained associated with significant risk of death (hazard ratio 2.64; 95% confidence interval 1.44-4.83; P = 0.002). Conclusions ICU-acquired BSI is associated with greater in-hospital mortality, but complicates only 5% of ICU admissions and its absolute effect on population mortality is limited. These findings have implications for the design and interpretation of clinical trials.
Introduction
Nosocomial bloodstream infection (BSI) is a serious and potentially preventable complication of hospitalization and has been estimated to be the eighth leading cause of death in the USA [1]. Critically ill patients are particularly vulnerable to hospital-acquired infections [2,3], which are two to seven times more common in the ICU [4][5][6][7] and can account for approximately half of all hospital-acquired BSI [8].
BSI is more common in patients who have surgery, are immunocompromised, develop multiorgan dysfunction, require mechanical ventilation or renal replacement therapy, and have greater illness severity on ICU admission [3,20,21]. Some critically ill patients may be genetically predisposed to both developing BSI and dying in hospital [22]. Thus BSI may be a marker of illness severity and pre-morbid condition as well as a direct contributor to adverse outcome. As a consequence, our ability to demonstrate the survival benefit of any intervention to prevent BSI will be dependent on the baseline incidence of BSI, the mortality rate of patients who develop it, and, crucially, on its true independent influence on outcome once correction has been made for other markers of illness severity.
Accordingly, we performed an observational study in a large cohort of critically ill patients and sought to estimate the incidence of BSI, the mortality rate of patients who acquire BSI, and its independent influence of mortality.
Study population and data sources
We performed a retrospective observational analysis of the incidence of BSI acquired during ICU admission at two university-affiliated hospitals in Melbourne, Australia. Data were obtained from prospectively collected electronic databases of ICU admissions and hospital microbiology records of positive blood cultures. Standard protocols for the collection, analysis, and reporting of blood cultures were employed. Complete data were available for 11 years (Jan 1998 to Feb 2009) in one centre (Austin Hospital) and six years (Jan 2003 to Dec 2008) in the other (Monash Medical Centre). Local ethics committee approval was obtained for re-analysis of routinely collected data, waiving requirement for specific patient consent. Data were available on tip cultures from intravascular devices in the Austin Hospital allowing a sub-group analysis of proven catheter-associated BSIs in this cohort of patients (72% of the total population).
Definitions
We used Center for Disease Control (CDC) definitions of ICU-acquired BSI (see Additional file 1) [23,24]; we considered both primary and secondary BSIs in our analysis. Nosocomial BSI in the ICU was defined as blood cultures taken in the presence of clinical evidence of infection for a bacterium or fungus obtained more than 72 hours after admission to the ICU. Thus, we included only those blood cultures taken after the third calendar day of ICU stay as reported by previous investigators [14,15,17,25].
Routine drawing of blood cultures was not ICU practice in the participating hospitals; we thus regarded all positive blood cultures obtained in the ICU as indicative of suspected infection.
In accord with CDC guidelines [24], we did not include cultures of coagulase-negative staphylococci or other common commensal skin organisms unless two cultures separately isolated the same species of microorganism. The first positive culture after the third ICU day was used to define the occurrence of ICU-acquired BSI. To allow study of a population at risk, we excluded all ICU admissions of less than 72 hours' duration.
BSI was considered to be catheter-associated if there was a positive tip culture from an intravascular device removed in the two days before or after the positive blood culture and the microbiological isolates from tip and blood were likely to represent the same infection (same species or compatible mixed growth).
Aim
We sought to document the incidence of acquired BSI in our ICU populations and to obtain an estimate of its effect on subsequent survival. We hypothesized that although BSI is likely to be associated with greater risk of death in an individual its relative frequency in the total ICU population might limit its impact on overall mortality.
Data analysis
We merged ICU admission and microbiology result databases and positive blood cultures paired with relevant ICU admission data by date and unique patient identifiers. Data were available on patient demographics, admitting specialty, duration of ICU admission, Acute Physiology and Chronic Health Evaluation III (APACHE III) physiology score on admission, APACHE III chronic health categories, need for mechanical ventilation or renal replacement therapy during ICU stay, and death during hospital admission.
To facilitate statistical analysis, APACHE III chronic health categories were used to define patient groups that might be at increased risk of BSI, namely: disseminated malignancy (metastatic cancer, lymphoma, leukemia, or myeloma), immunodeficiency (immunosuppression by illness or disease including HIV/AIDS), liver disease (hepatic failure and cirrhosis), chronic kidney disease, chronic pulmonary disease, and type 1 diabetes mellitus. Admission type was defined as surgical or non-surgical based on hospital admitting unit. Survival was defined as survival to hospital discharge or 90 days after ICU admission if the patient was still in hospital.
We performed univariate comparisons using Graph-Pad Prism version 5.0a for Mac OS (GraphPad Software, San Diego, California, USA [26]) and multivariate analysis and survival plots using R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria [27]) utilizing the packages survival [28] and Design [29]. Categorical data were reported as percentages, and compared using the chi squared test with Yates' correction. Continuous data were reported as median with inter-quartile range (IQR) and compared using the Mann-Whitney U test with Gaussian approximation. For comparisons, statistical significance was denoted by two-sided P values of less than 0.05.
Baseline risks for ICU-acquired BSI were examined in a multivariate logistic regression analysis with backward elimination of non-significant predictor variables. Significance was assessed against the null model by chi squared test of residual deviance, goodness-of-fit by unweighted sum of squares test, and predictive ability by calculation of the c-statistic. Independent predictors of survival were modeled using a Cox proportionalhazard analysis. As ICU-acquired BSI was not present at baseline it was incorporated into the model as a time-dependent co-variate [30], other factors were either present at ICU admission (admission illness severity, demographics, and comorbidities) or, in the vast majority of cases, were initiated in the first 72 hours of ICU stay (mechanical ventilation and renal replacement therapy) and were treated as time-independent co-variates. We avoided the need to consider hospital discharge as a competing endpoint by deeming all patients discharged alive from hospital to have survived to day 90 for the purposes of the survival analysis rather than censoring them at time of discharge. The proportional-hazard assumption was assessed by inspection of Schoenfeld residual plots. A Cox proportionalhazard analysis was repeated to separately assess the independent effect of the five most common microbiological diagnoses, grouping similar organisms to preserve statistical power. Similarly, in the cohort of patients from the Austin hospital, we modeled the relative effect on survival of catheter-associated and noncatheter-associated BSI.
Results
We studied 6,339 ICU admissions of more than 72 hours' duration. ICU-acquired BSI complicated 5.2% of these admissions (Table 1) and 9.5 new BSIs were acquired in the ICU per 1,000 patient days at risk. Median time to first positive blood culture in those acquiring BSI was seven days (IQR 5-12; Figure 1). Microbiological classification of ICU-acquired BSI is shown in Table 2.
Univariate analysis
Univariate analysis is presented in Table 1. BSI was associated with an 18.7% increase in crude hospital mortality from 22.5% to 41.2%. However, as BSI was infrequent, crude mortality in the total population was 23.5%, only 1% greater than in patients who did not acquire BSI (22.5%). This difference in mortality represents an unadjusted population attributable risk percentage [31] of 4.3% -that is, before adjusting for confounding variables, 4.3% of all deaths could be attributed to excess mortality after acquired-BSI.
During each full year of the study, rates of BSI varied from 4.4% to 8.1%, overall mortality from 21.1% to 26.5% and crude mortality in patients with ICU acquired BSI from 25% to 66%. However, there were no trends toward a systematic alteration in these frequencies over time ( Figure 2).
Patients who developed acquired-BSI had greater ICU length of stay than those who did not (median 15 days vs. 5 days; P < 0.001). Patients acquiring BSI were also significantly sicker at ICU admission and had more comorbidities (Table 1). Overall, 49 patients who developed BSI in the ICU were alive and in hospital 90 days after ICU admission and were treated as survivors in our analysis. Of these, only three subsequently died late during their hospital stay, a rate similar to the overall population.
Prediction of BSI
We examined risks for ICU-acquired BSI by developing a logistic-regression model for its prediction (Table 3). In this model, only higher APACHE III scores, the need for renal replacement therapy, liver disease, and surgical admission were risk factors for acquisition of BSI, whereas older age lessened the odds of a diagnosis of BSI. The model was significantly better than a null model (P << 0.001); however, its predictive ability was poor with a c-statistic of 0.63, implying that factors beyond the baseline predictors examined had a large influence on the development of BSI in the ICU.
Survival analysis
Controlling for baseline difference in a Cox proportional hazard model, we confirmed that BSI was associated with increased risk of death, with a hazard ratio for death from the time of acquisition of BSI of 2.89 (Table 4). Acquired BSI infection was modeled as a timedependent covariate and the effect on actual survival in the model was thus dependent on the time of acquisition of BSI ( Figure 3). This model is dependent on the validity of the proportional hazard assumption for acquired BSI and inspection of residual plots confirmed this was reasonable over the timescale in question (Figure 4). Accordingly for an individual, BSI occurring at day seven (median time of acquisition) was associated with an approximate 20% absolute increase in hospital mortality compared with absence of BSI, when all other baseline hazards were held at population means.
In Figure 5, we modeled the survival effect of all BSI, occurring at the rate and times of acquisition observed in the whole population, comparing against a group not acquiring BSI, with all other baseline hazards held at population means. In this model, the population attributable risk of death at or before day 90 was 4.95% and excess mortality in the entire study population, associated with the observed occurrence of acquired-BSI and independent of the other baseline hazards, was 1%.
Effect of microbiological diagnosis
We separately repeated our survival analysis to model the effect of the five most common classes of BSI on outcome (Table 5). In this analysis enterococci and coagulase-negative staphylococci infections were not significantly associated with survival while Staphylococcus aureus and Gram negative infections both approximately doubled the risk of death and candidemia was associated with an over four-fold risk of dying in hospital.
Catheter-associated BSI
We assessed the possibility of catheter associated bloodstream infection (CABSI) in 167 of 330 cases of BSI, in 127 of these an intravascular device tip was sent for culture in the two days either before or after a positive blood culture. In 34 cases (20.4% of BSI), a positive tip culture with a compatible microorganism identified likely CABSI. CABSI microbiological isolates and types of intravascular devices involved are shown in Tables 6 and 7, respectively. Univariate comparison between catheterassociated and non-catheter-associated BSI demonstrated a non-significant trend toward lower hospital mortality with CABSI (32.4% vs. 44.4%; P = 0.25; Table 8); however, in a Cox proportional-hazard analysis of all Austin Hospital patients, the hazard ratio for in-hospital death before 90 days linked with CABSI was 2.64 (95% confidence interval (CI) = 1.44-4.83; P = 0.002 vs. no BSI;
Time distribution of ICU-acquired BSI
Calendar days elapsed in ICU prior to positive blood culture Table 8).
Statement of main findings
We studied 6,339 admissions of greater than 72 hours in two university-affiliated ICUs. We found that ICUacquired BSI complicated approximately 1 in 20 of these admissions and that increased illness-severity, surgery, immunological compromise, liver disease, mechanical ventilation, and renal replacement therapy predicted its occurrence. We further found that BSI was independently associated with a close to three-fold increased risk of death from the time of positive blood culture and that the proportional hazard assumption was robust. This implies that although the proportional Incidence of ICU-Acquired BSI effect of BSI on mortality is constant over time, if BSI occurs early in ICU admission, when the baseline rate of death is high, the absolute effect on chance of survival is greater than if it occurs later in the course of critical illness. We note, however, that residual cofounders are likely to exist outside of our statistical analysis, a suspicion supported by the weak ability of the logistic-regression model to predict the development of ICU-acquired BSI based on the presence of the baseline predictors available. This suggests that our assessment of the impact of acquired-BSI on survival should be regarded as an upper-estimate of any effect.
We also found that the collective effect of BSI on survival was statistically related to infections with candida, S. aureus, and Gram-negative bacilli, while infections with coagulase-negative staphylococci and enterococci were not significantly associated with increased risk of death in our dataset. In the cohort of patients from the Austin Hospital, BSI identified as likely catheter-associated remained a significant hazard for non-survival, and risk of death was not significantly lower than that related to non-catheter-associated BSIs.
Together these findings imply, that, because of the low incidence and the estimated independent contribution to mortality, the totality of BSI accounted for, at most, an additional 1% excess mortality in the entire ICU population. This adjusted effect of BSI is very close to the unadjusted effect of BSI on survival. This may be because, although patients with acquired BSI were sicker and had more co-morbidities, they had to survive a certain length of time in order to be able to be diagnosed and categorized with BSI. In our model these competing effects appear to offset each other leading to the similarity of the adjusted and unadjusted survival.
Our finding the infection with coagulase-negative staphylococci did not confer significant additional risk of death is in accord with findings that catheter-related BSIs are less strongly associated with risk of death [16,42,43]; however, this was not borne out in our analysis of BSI with evidence of catheter infection. This suggests that the virulence of the microorganism rather than the source of infection may be more important in determining outcome.
Finding an association between illness severity and incidence of BSI in the ICU is in keeping with previous reports [9,[11][12][13][14]. Similarly need for renal replacement therapy in the ICU is a documented risk factor for BSI [44]. The negative association between increasing age and decreased risk of developing BSI has also been observed previously [45]. Older age may be associated with a less pronounced inflammatory response to infection [46], consequently, the likelihood of blood culture sampling may be lower, reducing observed incidence of BSI in older patients.
Significance of study findings
Our study expands current knowledge of the incidence and independent impact of ICU-acquired BSI on survival. First, it suggests that the likely typical incidence of acquired BSI in our ICU population is approximately 5%. This observation is a helpful comparator for control groups in interventional trials aimed at reducing BSI. This finding also suggests that conclusions from studies where the "control" incidence of acquired BSI is higher than this may not be directly transferable to all ICUs. Second, it confirms that some ICU-acquired BSI likely contributes independently to an increased risk of death and that prevention of these infections (predominantly S. aureus, Gram negative and Candida) is an important therapeutic goal. Third, it suggests that catheter-associated BSIs are clinically significant and their prevention is also of importance. Finally, it identifies several important factors that are associated with increased risk of developing BSI. However, our study also indicates that our ability to predict its development using baseline characteristics and major interventions (surgery, renal replacement therapy, and mechanical ventilation) remains limited. Thus, from the data routinely available early on in ICU admission, identification of a sub-group of higher risk patients for BSI-preventive intervention appears difficult. This observation is important, because it suggests that more research is required to identify patient characteristics, beyond those conventionally collected near ICU admission, to allow better prediction of risk of nosocomial infection. Better predictive models might allow appropriate targeting of cumbersome or No significant difference in distribution of microbiological isolates between groups, chi-squared test: P = 0.14. costly preventative interventions and enhance the power of clinical studies examining such interventions. The finding that BSI was independently associated with a three-fold increased risk of death from the time of positive blood culture implies that although the proportional effect of BSI on mortality is constant over time, if BSI occurs early in ICU admission, when the baseline rate of death is high, the absolute effect on chance of survival is greater than if it occurs later in the course of critical illness. We note, however, that residual cofounders are likely to exist outside of our statistical analysis, a suspicion supported by the weak ability of the logistic-regression model to predict the development of ICU-acquired BSI based on the presence of the baseline predictors available. This suggests that our assessment of the impact of acquired-BSI on survival should be regarded as an upper-estimate of any effect.
Finally, because of the relatively low incidence of ICUacquired BSI in our population, ICU-acquired BSI would account for only 5% of deaths, that is 1% excess mortality in the total population. This observation has several implications for randomized controlled trials of interventions aimed at decreasing mortality by preventing ICU-acquired BSI. For example, a putative untargeted intervention capable of preventing 50% of ICUacquired BSI would be expected to reduce the overall mortality of ICU patients staying for longer than 72 hours by only 0.5%. Accordingly, even a very effective intervention would have to be administered to 200 patients to save one life. This effect is impossible to test in any feasibly sized randomized controlled trial (in excess of 100,000 patients would be required for adequate power). Interventions to prevent BSI could have a greater impact on survival by impacting sub-clinical, undiagnosed, or localized infection. However, the strength of such effects would need to be quantified if investigators wish to be assured that interventional studies were adequately sized. Our data do suggest that use of formally diagnosed BSI as a surrogate or secondary endpoint in untargeted interventional studies may not be feasible.
Study strengths and weaknesses
This study has several strengths. We used a large sample of patients. Data were collected by dedicated data collectors and electronically stored and were thus not amenable to manipulation or bias. Similarly, microbiological data were collected as part of patient care. We assessed the independent contribution of BSI to patient outcome, providing useful information for trial design and for the assessment of the relevant interventional literature.
On the other hand, our study also has some weaknesses. Its findings may not be directly generalizable to differing microbiological environments worldwide. However, its results are comparable with those in similar studies conducted in the USA and Europe suggesting a degree of external validity. During the 11-year study period changes in case-mix, clinical workload, and clinical practice could have affected incidence and outcome of ICU-acquired BSI. However, no trend was evident on inspection of the yearly data (see Additional file 1). By examining this time-span we were able to include data from over 6,000 ICU admissions making this one of the largest studies of BSI in intensive care.
We did not have detailed clinical information including exact trigger for drawing blood cultures, antimicrobial therapy, response to treatment, and cause of death. Nor could we determine whether individual episodes of BSI represented true infections. However, we excluded commensal skin organisms isolated in single blood culture bottles, making it more likely that our isolates represented true BSI. The association of such BSI with illness severity, invasive interventions, and mortality all support this notion. Furthermore, although exclusion of a small number of non-clinically significant infections might increase the attributable-mortality of BSI, this would also decrease the observed incidence of BSI and would thus be unlikely to substantially alter our estimate of the effect of BSI on overall survival.
Our analysis of catheter-associated BSI was confined to only one study centre. We required a positive tip culture to confirm a likely catheter source. Thus, we may have missed some catheter-related infection although Table 4.
frequency of catheter-associated infection in our cohort (about 20%) was similar to that reported by some previous investigators [9,14]. Conversely, in a few patients, the catheter might have been secondarily infected by blood-borne infection. However, by identifying a group of patients with likely catheter-associated infection, we were able to demonstrate that increased risk of death remained significant in patients where a catheter source of infection was very likely. We did not compare patients developing BSI with a matched control population. However, such retrospective matching of controls is always approximate and susceptible to unmeasured effects. In our study, without dayto-day clinical data on patients, we were limited to adjusting for baseline factors and major interventions (such as mechanical ventilation) usually commenced early in ICU admission. We also did not assess the effect of BSI on duration of ICU stay. However, the direction of causality is very difficult to determine, because greater length of exposure to risk will tend to increase the incidence of BSI, while, at the same time, occurrence of BSI will tend to delay ICU discharge. Given the inability to assess direction of causality, we did not attempt to incorporate ICU length of stay into our statistical models. We note that infections with coagulase-negative staphylococci, which would be expected to be more common with greater length and complexity of ICU stay, were not significantly associated with risk of death. This suggests that the associations seen with mortality for other microorganisms are likely to be causative. However, because of concerns about unmeasured confounders, our estimate of attributable-mortality from ICU-acquired BSI should be regarded as an upper estimate of any effect. Significantly, however, even using this high estimate of attributablemortality, BSI had little impact on the overall survival of the total population, contributing to, at most, an absolute 1% increase in hospital mortality.
Conclusions
In a study of over 6,000 ICU admissions lasting longer than 72 hours, ICU-acquired BSI was associated with a doubling in risk of death in hospital to approximately 40%. This correlation remained even after adjustment for baseline illness severity, demographics, and co-morbidities in a Cox proportional hazard model and was almost entirely attributable to BSI with significant pathogens. However, ICU-acquired BSI was uncommon thus, although of great clinical impact to those individuals affected by it, its attributable excess mortality could be, at most, 1% of the total population. This effect implies that a) the survival benefit of untargeted interventions aimed at reducing the rate of proven ICUacquired BSI would be undetectable in any practically sized controlled trial; b) claims of improved survival from interventions aimed at reducing acquisition of BSI in the ICU should be treated with caution. | 2014-10-01T00:00:00.000Z | 2011-03-21T00:00:00.000 | {
"year": 2011,
"sha1": "f501fe539ffca032ab42ca38c1e783353ef5b755",
"oa_license": "CCBY",
"oa_url": "https://ccforum.biomedcentral.com/track/pdf/10.1186/cc10114",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "f501fe539ffca032ab42ca38c1e783353ef5b755",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
255813612 | pes2o/s2orc | v3-fos-license | Antibiotic-resistant bacteria in the guts of insects feeding on plants: prospects for discovering plant-derived antibiotics
Although plants produce many secondary metabolites, currently none of these are commercial antibiotics. Insects feeding on specific plants can harbour bacterial strains resistant to known antibiotics suggesting that compounds in the plant have stimulated resistance development. We sought to determine whether the occurrence of antibiotic-resistant bacteria in insect guts was a widespread phenomenon, and whether this could be used as a part of a strategy to identify antibacterial compounds from plants. Six insect/plant pairs were selected and the insect gut bacteria were identified and assessed for antibiotic susceptibilities compared with type strains from culture collections. We found that the gut strains could be more or less susceptible to antibiotics than the type strains, or show no differences. Evidence of antibacterial activity was found in the plant extracts from five of the six plants, and, in one case Catharanthus roseus (Madagascar Periwinkle), compounds with antibacterial activity were identified. Bacterial strains isolated from insect guts show a range of susceptibilities to antibiotics suggesting a complex interplay between species in the insect gut microbiome. Extracts from selected plants can show antibacterial activity but it is not easy to isolate and identify the active components. We found that vindoline, present in Madagascar Periwinkle extracts, possessed moderate antibacterial activity. We suggest that plant-derived antibiotics are a realistic possibility given the advances in genomic and metabolomic methodologies.
gut bacteria confers antibiotic-resistance phenotypes and can readily spread to the environments inhabited by the insects [7,8]. More importantly, the antibiotic-resistance elements present in the insect guts in the absence of clinically-relevant levels of antibiotics suggests that the genes can be maintained in response to other environmental stimuli. We hypothesise that the stimuli responsible for the antibiotic resistance are components in the insect food with antibacterial activity, i.e. these observations suggest that antibiotic compounds are present in the host plant.
A number of insects feed on plants and their products, including plants with medicinal properties. It is not uncommon for insect gut bacteria to be resistant to antimicrobial or toxic components of their foods or to contribute to the detoxification of such compounds. Royal jelly is a honey-bee secretion used for nutrition of queens and larvae and has a potent antimicrobial effect [9]. While core honey bee gut bacteria are resistant to the antibacterial activity of royal jelly, the corresponding environmental strains, non-core species and control strains not associated with bee colonies, are susceptible [10]. In a disc-diffusion assay, six out of eight core isolates were fully resistant to royal jelly, the other two isolates (Neisseriaceae sp. and Lactobacillus sp. A) were mildly susceptible and the corresponding environmental isolates were fully susceptible. For example, Lactobacillus kunkeei gut-isolated strain was fully resistant while the flower-isolated strain was susceptible. It is not clear if such resistance differences in the insect gut microbiota are common; we aimed to explore this question.
To achieve that we selected a number of plants with known medicinal properties, such as eucalyptus, lavender and Madagascar periwinkle, toxic plants, such as ragwort and potato, and a plant with no known medicinal or toxic properties, i.e. cabbage. Each plant was matched with an insect species that could be reared feeding exclusively on that plant. The resulting plant/insect pairs are: Giant Lime Green stick insect (Diapherodes gigantea) feeding on Eucalyptus (Eucalyptus dalrympleana), Diamondback moth (Plutella xylostella) feeding on Chinese Cabbage (Brassica rapa), Cinnabar moth (Tyria jacobaeae) feeding on Ragwort (Jacobaea vulgaris), Rosemary beetle (Chrysolina americana) feeding on Lavender (Lavendula angustifolia), Death's-head Hawkmoth (Acherontia atropos) feeding on Potato leaves (Solanum tuberosum), and Beet Armyworm (Spodoptera exigua) feeding on Madagascar Periwinkle (Catharanthus roseus). We chose insects feeding exclusively on the selected plant species; to ensure this we either chose monophagous insect species or insects reared in captivity that were feeding exclusively on the selected plant.
The aims of this work were essentially two-fold. Firstly, to see whether the phenomenon of unexpected resistance to antibiotics found in gypsy moth gut bacteria (discussed above) could be found in other insects, i.e. is this a general phenomenon? Secondly, to test whether insect gut bacteria can be utilised in assayguided fractionation of plant extracts leading to the identification of plant fractions with antibacterial activities. To do this we compared antibiotic-susceptibility profiles and susceptibility to plant extracts of gutisolated bacterial strains and the corresponding type strains, and used any differences in susceptibilities to identify plant extracts with the most promising antibiotic activity.
Giant lime green stick insect feeding on eucalyptus
Three stick insects were analysed in this study and were treated separately, in part to examine the diversity of species from different individuals. From individual 1, three bacterial species were identified: Bacillus amyloliquefaciens, Microbacterium oxydans and Sphingobacterium multivorum (Table 1). Ba. amyloliquefaciens and M. oxydans are environmental strains, first isolated from soil and air respectively. Sp. mulitvorum, on the other hand, was first isolated from the human spleen. As only a small proportion of bacteria can be cultured under laboratory conditions, we decided to assess what proportion of bacterial strains is recovered from the insect gut by culture-dependent methods. To do this, the other two stick insects were dissected and bacteria from their guts were cultivated and identified. Additionally, bacterial genomic DNA (gDNA) was isolated from the samples using a kit optimised for use with soil bacteria and the samples were submitted for 16S sequencing.
The two samples, originating from two separate stick insects, contained 9450 and 8450 valid reads, yielding 97 and 53 operational taxonomic units (OTUs) respectively. The rarefaction curve reached a plateau indicating that the bacterial diversity in the guts has been sufficiently sampled (Additional file 1: Figure S1). Out of the sample with 97 different OTUs, only 12 were present at above 1% abundance; the most abundant strain was Serratia marcescens, which constituted 33% bacterial reads retrieved from the D. gigantea guts (Additional file 2: Figure S2). The other gut sample contained 53 OTUs; five of these were present above 1% abundance. Again, Se. marcescens was the dominating strain in the gut community with 91% abundance. Se. marcescens is a common pathogen of insects but it can be also found in healthy insects. The only bacterial order found in the guts was Enterobacteriales. Other strains shared by the two gut communities are: Pantoea beijingensis, Lellottia amnigena, and Enterobacter asburiae (Additional file 2: Figure S2).
Because the metagenomic analysis was performed on a new set of insect guts, we cultured and identified strains from these samples as described before. A set of strains was identified: Serratia marcescens, Microbacterium paraoxydans, Kocuria rhizophila and Pectobacterium cypripedi. The older samples, used for the assessment of antibiotic resistance in the insect gut, contained a different set of bacterial strains (see above). It is perhaps surprising that no strains are shared between the two sets of D. gigantea guts; the only closely related strains are the two Microbacterium species. This may reflect intrinsic differences between the gut microbiota between individuals and/or experimental variations such that different bacterial species were dominant in the plated samples. Notwithstanding this, our comparison of culture-dependent and -independent methods suggest that we can culture <5% strains from the D. gigantea guts.
For the antibiotic susceptibility trials, the 3 bacterial species isolated from individual 1 were analysed. In most cases there was no difference in antibiotic susceptibility between the type strains and gut-isolated strains ( Table 2). Ba. amyloliquefaciens gut-isolated strain was more resistant to ciprofloxacin and tetracycline than the type strain. M. oxydans from the stick insect gut was more resistant to ciprofloxacin, but less resistant to tetracycline when compared to the type strain. Gutisolated Sp. multivorum was more susceptible to tetracycline than the type strain, perhaps unsurprisingly as the type strain is a clinical isolate.
We assessed the extract from eucalyptus leaves for its ability to inhibit the growth of the three bacterial species isolated from the gut of individual number 1 (Table 3). No differences in plant extract susceptibility were observed for Ba. amyloliquefaciens; both the gut-isolated strain and type strain were fully resistant to the eucalyptus extract. The gut-isolated Sp. multivorum strain was more resistant to the eucalyptus extract than the type strain. Surprisingly, the opposite was observed for M. oxydans; the gut-isolated strain was more susceptible to eucalyptus extract than the type strain. M. oxydans has shown unexpected ciprofloxacin resistance in the antibiotic susceptibility testing and it is possible that the strain is generally more resistant to several compounds or conditions.
After we established that the eucalyptus extract has antibacterial activity, we attempted to purify an active fraction of the extract (see Methods). After initial purification, HPLC fractions were tested for antibacterial activity and fractions with antibacterial activity was subjected to another round of HPLC, but further separation of the compounds was not achieved. Liquid chromatography coupled with mass spectrometry (LC-MS) analysis of the HPLC-purified fraction revealed the presence of a complex mixture of compounds. Considering only the most abundant ions, positive electrospray MS revealed between 20 and 25 different molecular ions, ranging between m/z 264.8250 and 630.9556. Further isolation of individual components from this mixture and elucidation of their chemical structures were not attempted.
Diamondback moth feeding on Chinese cabbage
The strains identified from Plutella xylostella guts were Sanguibacter keddieii and Raoultella terrigena (Table 1). Sa. keddieii is a rare bacterial species, previously isolated from the blood of otherwise healthy cows [11]. Ra. terrigena was previously classified as Klebsiella terrigena and can be isolated from soil and water [12]. In antibiotic susceptibility trials (Table 2), the Ra. terrigena type strain was less resistant to ampicillin than the gut- isolated strain and more resistant to chloramphenicol, kanamycin and rifampicin. The Sa. keddieii type strain was more resistant to ampicillin, chloramphenicol, ciprofloxacin and kanamycin than the gut-isolated strain. The differences in antibiotic resistance levels of Sa. keddieii were unusual and higher than the other strains tested. We found that the cabbage extract did not inhibit the growth of either of the P. xylostella gut or type strains (Table 3). Since only two strains were isolated from the guts of Diamondback moth larvae, we tested additional strains: an indicator strain Escherichia coli ATCC 25922, and strains isolated from the stick insect gut: Ba. amyloliquefaciens, M. oxydans and Sp. multivorum, with their corresponding type strains. No antibacterial activity was observed against any of these strains (results not shown). No further work was carried out on the cabbage leaf extract.
Cinnabar moth feeding on ragwort
Five bacterial strains were isolated from the guts of cinnabar moth larvae feeding on ragwort leaves (Table 1): Burkholderia fungorum, Bacillus licheniformis, Kocuria rhizophila, Staphylococcus epidermidis and Staphylococcus warneri. Staphylococcus species are common members of human skin microbiota [13], but it is not uncommon for insect gut microbiota to include such isolates. It is unlikely these species are a contamination introduced during gut dissections as the procedure was carried out in sterile conditions following appropriate dissection procedures. St. epidermidis and St. warneri are the only isolates in the guts of cinnabar moth larvae that are normally associated with human microbiota. All the remaining bacterial isolates are environmental strains: Bu. fungorum was first isolated from a fungus and both Ba. licheniformis and K. rhizophila are soil bacteria. In antibiotic susceptibility trials (Table 2), the type strains were often more resistant to antibiotics. The Bu. fungorum type strain was more resistant than the gut-isolated strain to ciprofloxacin and kanamycin. The Ba. licheniformis type strain was more resistant to three antibiotics than the gut-isolated Ba. licheniformis: ampicillin, chloramphenicol and oxytetracycline. There were no differences in antibiotic susceptibility between K. rhizophila gut-isolated strain and the type strain. The St. epidermidis gut-isolated strain was more resistant to ampicillin and less resistant to kanamycin than the type strain and the St. warneri gut-isolated strain was more Table 2 Antibiotic susceptibility profiles of the bacteria isolated from the insect guts. The antibiotic susceptibility of the gut-isolated strains and the type strains was assessed in a broth microdilution assay (green indicates gut strain less susceptible than type strain; red indicates gut strain more susceptible than type strain) resistant to ciprofloxacin than the type strain. Additionally, the St. warneri type strain was more resistant to rifampicin and tetracycline than the gut strain. The large number of cases where the type strain was more antibioticresistant than the gut-isolated strain was unexpected. There were no differences in ragwort extract susceptibility between the gut-isolated strain and type strain of Ba. licheniformis (Table 3). Four gut-isolated strains were more resistant to the ragwort extract than the corresponding type strains: Bu. fungorum, St. epidermidis and St. warnerii. K. rhizophila gut-isolated strain was more susceptible to the ragwort extract than the type strain. Surprisingly, the strains were equally antibioticsusceptible, making it difficult to hypothesise what caused increased resistance to the plant extract.
Even though three out of five type strains tested were more resistant to the plant extract than their corresponding gut strain, the ragwort extract exhibited a high level of antibiotic activity and the extract was investigated further. The partitioned ragwort extract was fractionated on a weak anion exchange SPE column followed by reverse-phase HPLC. Fractions exhibiting antibacterial activity were further analysed by LC-MS. Known pyrrolizidine alkaloids and their N-oxides, which are the compounds responsible for the toxicity of ragwort to cattle [14], were not detected. These alkaloids include: jacobine (m/z 351), jaconine (m/z 387), jacozine (m/z 349), otosenine (m/z 381), retrorsine (m/z 351), seneciphilline (m/z 333), senecionine (m/z 335), and senkirkine (m/z 366), acetylerucifoline (m/z 392), erucifoline (m/z 350), integerrimine (m/z 335), jacoline (m/z 369), riddelline (m/z 350), senecivernine (m/z 335), spartioidine (m/z 333), and usaramine (m/z 351). The Noxides have m/z ratios 16 mass units higher than the parent alkaloids. The absence of these metabolites in our extracts suggests that the pyrrolizidine alkaloids were either not extracted from the plant material or not selected for in the assay-guided HPLC fractionation, and are not responsible for the antibacterial activity of the ragwort extract. However, the generally high level of activity of the ragwort extract suggests that this warrants future investigation. The values were obtained in a disc diffusion assay; the discs carried 10 μg dried extract, the assay was repeated three times
Rosemary beetle feeding on lavender
We were not able to dissect the guts of rosemary beetles and so instead we isolated total bacteria from surfacesterilized beetles. Five different bacterial strains were isolated: Microbacterium foliorum, Microbacterium gubbeenense, Pantoea agglomerans, Rhodococcus erythropolis and Staphylococcus epidermidis. The Rosemary beetle microbiota is probably dominated by the gut bacteria; three of the isolated strains were of environmental origin: M. foliorum was first isolated from the surface of leaves, M. gubbeenense from cheese and Rho. erythrypolis from soil. The two remaining bacterial strains were human-associated: both Pa. agglomerans and St. epidemidis can be found on human skin. In comparison with other insect gut microbiomes there were few differences in antibiotic susceptibilities between the strains isolated from the Rosemary beetles and the matching type strains ( Table 2). There were no differences in antibiotic susceptibility to ampicillin, chloramphenicol, ciprofloxacin, kanamycin, rifampicin or oxytetracycline between the type strains and insectisolated strains of M. foliorum, M. gubbeenense, Pa. agglomerans and Rho. erythropolis. The St. epidermidis type strain was more susceptible to ampicillin, ciprofloxacin, and kanamycin than the gut strain. Even though not many differences in the antibiotic susceptibility between the strains have been detected, in three cases when this was the case the insect-isolated strain was more resistant than the type strain, indicating it was under selective pressure.
Rosemary beetles feed on lavender leaves, but the plants harvested for the preparation of the lavender extract also had flowers. For the purposes of comparison, leaf and flower lavender extracts were prepared and assayed for antibiotic properties. The leaf extract was antibacterial against two of the bacterial strains used and the flower extract was active against all five strains. Both extracts exhibited only a low level of activity. In most cases there were no differences in the susceptibility to lavender extracts between the type and the 'gut' strains. The St. epidermidis insect strain was more resistant to lavender flower extract than the type strain and the M. foliorum type strain was more resistant to lavender leaf extract than the insect-isolated strain. Generally, only a low level of antibacterial activity was detected in lavender and the flower extract tended to be more active than the leaf extract. Only the flower extract was investigated further due to its higher antibacterial activity. The lavender flower extract was subjected to purification with SPE and HPLC. Low-level antibacterial activity was detected in five HPLC fractions ranging between 17 and 50% methanol. The low level of activity hindered the assay-guided fractionation as separating the peaks by HPLC lowered the individual antibiotic activity of each peak below detection level. Semi-preparative HPLC separation did not provide sufficient amounts of material to allow identification of the most active components of the extract.
Death's-head hawkmoth feeding on potato leaves Three different bacterial strains were isolated from the guts of Death's-head Hawkmoths: Enterobacter asburiae, Pseudomonas putida and Raoultella terrigena. En. asburiae is a species normally isolated from clinical specimens while the two other isolates, P. putida and Ra. terrigena, are environmental bacteria commonly isolated from soil and water.
For Death's-head Hawkmoth bacteria there were differences in antibiotic susceptibility in the majority of tests performed and most of the type strains were more antibiotic-resistant than their matching gut-isolated strain ( Table 2). The En. asburiae type strain was more resistant than the gut-isolated strain to all antibiotics tested: ampicillin, chloramphenicol, ciprofloxacin, kanamycin, rifampicin and tetracycline. The Ps. putida type strain was more resistant to chloramphenicol, kanamycin, rifampicin and tetracycline. Ra. terrigena type strain was more resistant than the gut strain to chloramphenicol and more susceptible than the gut strain to kanamycin and tetracycline.
There were no differences in potato leaf extract susceptibility between the gut isolates and type strains of En. asburiae and Ps. putida ( Table 3). The Ra. terrigena gutisolated strain was more susceptible to the potato extract than the type strain, contrary to what was expected. The strains had some unexpected antibiotic resistance differences. The type strain was more chloramphenicolresistant than the gut-isolated strain, opposite to the predicted resistance profile. It is possible that both the chloramphenicol resistance and the increased potato extract resistance are due to the same mechanism. As the antibacterial activity of the extract fractions were relatively low this was not further pursued.
Beet armyworm feeding on Madagascar periwinkle
Four bacterial strains were isolated from Beet Armyworm guts: Bacillus aquimaris, Bacillus vietnamensis, Microbacterium paraoxydans and Rhizobium pusense. Ba. aquamaris, Ba. vietnamensis and Rhi. pusense are environmental isolates, and M. paraoxydans is a clinical isolate. As discussed before it is not unusual for an insect gut to contain both environmental strains and isolates of clinical importance.
Differences in antibiotic susceptibility were detected in over half of the tests performed. The Ba. aquimaris type strain was more susceptible to chloramphenicol and kanamycin, but more resistant to rifampicin than the gut-isolated strain. The Ba. vietnamensis type strain was more resistant to five out of six antibiotics tested: ampicillin, chloramphenicol, ciprofloxacin, kanamycin and rifampicin. The M. paraoxydans type strain was more susceptible to ampicillin and more resistant to ciprofloxacin than the gut-isolated strain. The Rhi. pusense gutisolated strain was more resistant than the type strain to ciprofloxacin, kanamycin and tetracycline.
Both leaf and root extracts were prepared from the Madagascar Periwinkle (C. roseus) plants. There were no differences in Madagascar Periwinkle leaf extract susceptibility between the gut strains and type strains of the two Bacillus isolates: Ba. aquamaris and Ba. vietnamensis. The Rhi. pusense gut isolate was more resistant than the type strain to both the leaf and root extract. Contrary to what was expected, in several cases the type strain was more resistant to the C. roseus extract than the gut-isolated strain. The M. paraoxydans type strain was more resistant to both the leaf and root extract than the gut strain and both Ba. aquamaris and Ba. vietnamensis type strains were fully resistant to the C. roseus extract while the gut-isolated strains were susceptible.
All three type strains that were unusually resistant to the plant extract, also showed unexpected resistance to antibiotics ( Table 2).
The Madagascar Periwinkle root extract was purified and fractionated. The active fraction eluted at the beginning of the methanol gradient with fractions containing less than 15% methanol, indicating high polarity of the active components. The HPLC fraction with antibiotic activity was analysed by LC-MS to determine which metabolites were present. The most abundant ions were identified by running their m/z ratios against the annotated metabolites in the Medicinal Plants Consortium database (http://metnetdb.org/mpmr_public/).The most abundant ion (m/z 349.1700) detected was serpentine, a known terpenoid indole alkaloid previously isolated from roots of the Madagascar Periwinkle [15]. Present in the root extract were also loganic acid and catharanthine. The structures of the metabolites with known structures are shown in Fig. 1. The other abundant ions were matched to metabolites with unknown structures. The identifiers of these metabolites are shown in Table 4, but a b c d Fig. 1 Chemical structures of the identified metabolites from C. roseus extract with antibacterial activity. The active fraction from the root extract contained serpentine (a), loganic acid (b) and catharanthine (c); the leaf sample contained vindoline (d) as the most abundant metabolite, followed by loganic acid (b) and serpentine (a). For the experiments in Table 6, ajmalicine (a) and loganin (b) were used as these are commercially available. Glc = glucose the database does not provide any structural information about these compounds. The most abundant ion identified in the Madagascar Periwinkle leaf extract was vindoline, followed by loganic acid, and serpentine ( Table 5). As before, several ions were matched to metabolites with an unknown structure. Some of the metabolites were shared between the leaf and root extracts. Both serpentine and loganic acid were found in the root extract and leaf extract.
Neither of the fractions contained a single metabolite and it is difficult to hypothesise which one of the metabolites identified was responsible for the antibiotic activity. Instead we established minimal inhibitory concentrations (MICs) for commercially-available compounds detected in the fractions (Table 6). In some cases, we used chemically similar compounds that were commercially available; the differences between the metabolites from the periwinkle extracts and tested compounds are shown in Fig. 1 (a, b). Vindoline had the lowest MIC (150-600 μg/mL); previous data also showed vindoline to be the most active terpenoid indole alkaloid, but the MIC was found to be 1000 μg/mL [16]. The other compounds tested, ajmalicine, catharanthine and loganin, had MICs between 300 and 1250 μg/mL. The MIC values found were relatively high; the lowest MIC values were obtained for permeable Es. coli strain, consistent with more effective entry into the bacteria.
Diversity of bacterial strains present in the insect guts
We isolated a variety of bacterial strains from the guts of the insects investigated. Most of the bacterial isolates were common soil and water strains, but we also identified several strains with clinical significance in human health. Such composition of insect gut microbiomes is not unusual and has been described before (for example Staphylococcus species in termites [17] and Enterococcus species in Galleria mellonella (greater wax moths) [18]). As insects are sometimes vectors of human disease, the presence of pathogens in the healthy insect microbiota is an important factor to bear in mind.
We did not identify any obligatorily anaerobic bacteria from the insect guts; however, this may well be a The ions were matched with known metabolites from the C. roseus metabolomics database based on their m/z ratios. For ions without a good match in the database, the closest match was listed The ions were matched with some known metabolites from the C. roseus metabolomics database based on their m/z ratios. For ions without a good match in the database, the closest match was listed consequence of our methods of isolation. Oxygen levels in herbivorous insects' guts are low [19], especially in the midgut and hindgut. The oxygen level depends partially on the insect food and increases when the insects feed on artificial food. The presence of oxygen in the guts of herbivorous insects can lead to the formation of reactive oxygen species from the ingested plant material. Additionally, the autoxidation of plant phenolics is greatly reduced by the lack of oxygen, even in the highly alkaline guts of some insects. The depletion of oxygen from the insect guts can be abolished by boiling, suggesting either endogenous insect enzymes or microbial activity are responsible. It is possible that both processes are involved and both the host insect and the native microbiota together sustain the gut as a suitable niche. The metagenomic analysis of the Giant Lime Green Stick insect guts suggested that we are only able to identify a minority of bacterial species (<5%) by culturedependent methods. This is not surprising given previous findings [20,21], but assessing the complete gut microbiome of insects was not one of the aims of this study. The metagenomic analysis did reveal some anaerobic species, but our methods may have biased the results in favour of aerobes.
Comparison of the gut bacteria isolated in our study with other studies
For the giant stick insects there were few bacterial species shared between our study and previous findings [22,23]. Diamondback moth and Beet Armyworm microbiomes shared little similarity between our study and previous metagenomic surveys [24,25]. For other insects there were no studies that we could compare our results to, but the bacterial species we isolated were not uncommon in other insect guts.
A study focusing on the characterisation of order Phasmatodea (stick insects) gut anatomy and microbiology implemented culture-dependent methods to find four microbial species in the gut of D. gigantea feeding on Eucalyptus citriodora [22]. In the crop there was one bacterial and one fungal species: Serratia marcescens and Cryptococcus ramirezgomezianus, the midgut was colonised exclusively by Enterobacter cloacea, and the hindgut was the most species-rich with three bacterial strains: En. cloacea, Erwinia persicina, Se. marcescens. Another study identified Se. marcescens and Spiroplasma species in D. gigantea guts via targeted cultureindependent methods [23]. The authors did not speculate about the role of Se. marcescens, but Spiroplasma species are a common commensal endosymbiont, maternally transmitted, and in some insects responsible for impaired reproduction by inducing the selective elimination of male progeny. Our metagenomic analysis of the stick insect gut community confirms the abundance of Se. marcescens in the digestive tract. However from genomic data alone it is not possible to propose a role for Se. marcescens, which is both a human pathogen, common in hospital-acquired infections [26], and an insect pathogen sometimes utilized as a model species in immunity studies [27].
On the other hand there was little similarity between the bacterial species isolated from the Diamondback moth larvae and Beet Armyworm larvae in our experiments and previously reported studies [24,25]. Only two strains have been isolated from the diamondback moth larvae by culture-dependent methods: Sa. keddieii and Ra. terrigena in our experiments. A previous study identified on average 208 operational taxonomic units from seven P. xyllostella larvae midguts [25]. Surprisingly only 25 operational taxonomic units had frequency higher than 0.05%, demonstrating that the majority of isolates present in the guts of Diamondback moth larvae were extremely rare. The authors do not discuss the composition of the P. xylostella microbiomes on the species level, but their dataset can be retrieved through Sequence Read Archive (http://www.ncbi.nlm.nih.gov/Traces/sra/).An analysis of the dataset revealed no sequences with high degree of similarity to either Sa. keddieii or Ra. terrigena, but the rarefaction curves of the dataset did not reach a plateau, indicating that the sequencing depth was insufficient to discover all sequences present in each sample. Another possible explanation is that the strains are not present at all, since the Diamondback moth microbiota may be variable or transient.
Previous studies identified Ochrobactrum sp. and Myroides odoratus as the main culturable strains from Beet Armyworm guts [24]. Rhizobium pusense, isolated in our experiments, is related to Ochrobactrum strains. It is possible that these strains are consistently present in the beet armyworm guts. The Beet Armyworm microbiota was shown to be involved in the production of Nacylaminoacids [24], which can be isolated from insects' oral secretions and have been indicated in eliciting plant defences. One of such N-acylamino acids, microbiallyproduced volicitin (N-(17-hydroxylinolenoyl)-L-glutamine) induces damaged leaves to produce volatile compounds attracting parasitic wasps that infect the armyworm larvae [28].
For cinnabar moth larvae, rosemary beetle, and Death's-head Hawkmoth larvae there are no studies available for comparison. Out of five species of bacteria isolated from cinnabar moth guts, two belonged to genus Staphylococcus: St. epidermidis and St. warneri. Both of these isolates are typical members of human skin microbiota. It is not uncommon for bacteria considered typically human to be found in insects. Staphylococcus species have been described as members of the gut microbiota of Termitidae family of termites and the presence of Staphylococcus species was specific to that group of Australian higher termites [17]; the role of the Staphylococcus species in the termite gut is probably cellulose degradation [29].
The rosemary beetle microbiome was another example of an insect gut community with members typically associated with human microbiota: St. epidermidis and Pa. agglomerans. As discussed before it is not uncommon for insects to harbour such bacteria in their guts. Pa. agglomerans is a prominent member of the desert locust Schistocerca gregaria [30] and produces compounds that lead to the swarming behaviour of the locusts and are necessary for the synthesis of an aggregation/cohesion pheromone.
The Death's-head Hawkmoth gut community was a mixture of environmental strains and strains associated with human microbiota, similar to other insect species described in this chapter. Surprisingly, the Death's-head Hawkmoth larvae harbored strains that were most different from the type strains when the antibiotic susceptibility profiles were compared. The En. asburiae type strain is a clinical isolate and it was the only type strain tested that was more resistant than the gut strain to every antibiotic in the panel. This result suggests the Death's-head Hawkmoths' gut is not a niche with an extremely selective environment, as the gut isolated En. asburiae was not as antibiotic-resistant as the type strain.
Antibiotic-resistant bacteria can be present in the guts of insects feeding on plants
When the antibiotic-resistance profiles of the gutisolated strains and the type strains are compared, there are three possible outcomes. Firstly, no differences between the antibiotic resistance levels can be detected. Secondly, the gut-isolated strain can be more resistant than the type strain, and finally, the type strain can be more antibiotic-resistant than the gut strain. Initially it was hypothesised that no differences and higher resistance of the gut-isolated strains would be the most common outcomes of the antibiotic susceptibility testing, i.e. we expected the gut strains to have developed antibiotic resistance in response to selective pressures in the insect gut. We found that equally often as there being no difference in antibiotic resistance or the gut-isolated strain being more antibiotic-resistant, the type strain had a higher level of resistance to the antibiotics tested ( Table 2). There are several possible explanations for this phenomenon. For example, it is possible that the elevated antibiotic resistance of some type strains is due to their origin as clinical isolates. Bacteria first cultured in hospital settings are likely to have been exposed to antibiotics and to have elevated levels of antibiotic resistance or tolerance [31]. In some cases, a type strain was more resistant than the gut-isolated strain to one antibiotic and less resistant to another, which can be explained by the fitness cost of maintaining multiple resistance genes [32]. Mutations resulting in antibiotic resistance are commonly associated with a reduction in growth rate, although that is not always the case [33].
Susceptibility of bacterial strains to plant extracts
Medicinal and toxic plants were chosen for assessment of antibacterial activity in this study, based on anecdotal evidence and, where available, scientific literature. Interestingly, there is little in the way of systematic studies to assess whether plants used in traditional medicine are more likely to contain valuable bioactive compounds [34]. The use of certain plants in traditional medicine is not indicative of their efficacy; such data can only be provided when standardised and chemically-characterised plant preparations are tested in vitro and in animal models, and their efficacy is confirmed in clinical studies [35].
It was hypothesised that the type strains would be more susceptible to the plant extracts than the strains from the insect gut that have been exposed to the constituents of the extracts. Surprisingly this was not always true. Many of the gut-isolated strains that exhibited unexpected plant extract susceptibility also had some unusual differences in antibiotic resistance levels. For example, Burkholderia fungorum identified in the guts of the Cinnabar moth larvae feeding on ragwort was more susceptible to ciprofloxacin and kanamycin than the type strain. Apart from the elevated antibiotic resistance, the Bu. fungorum type strain was also more resistant to ragwort leaf extract. Similarly, the Ba. vietnamensis type strain was more resistant than the Ba. vietnamensis strain isolated from the guts of Beet Armyworms to ampicillin, chloramphenicol, ciprofloxacin, kanamycin and rifampicin, as well as periwinkle root extract. It is possible that the unexpected antibiotic resistance and plant extract susceptibility are correlated or linked, but it is difficult to rationalise the nature of this phenomenon, but it suggests that the interplay of bacterial strains in the insect gut is likely to be quite complex. For example, it is possible for toxic plant metabolites to be broken down in the gut before they are exposed to gut bacteria [36,37].
Isolation of natural products from plant extracts
A significant issue with this work was the difficulties experienced in isolating natural products from plant extracts. Initially crude extracts were fractionated using normal-phase HPLC, but the performance was unsatisfactory as the waxy components were binding too tightly to the silica gel stationary phase, requiring long wash cycles. To improve the HPLC efficiency, the plant extracts were partitioned between organic solvents and treated with SPE before fractionation. SPE is not as efficient in separating compounds as HPLC, but the variety of matrices available is larger. Two-or three-dimensional fractionation is possible, for example using ion-exchange matrices, size exclusion and silica-or polymer-based matrices that separate compounds based on their hydrophobicity. In the case of the plant extracts tested, sizeexclusion chromatography produced peaks that were too broad and ion-exchange matrices introduced formic acid and ammonium hydroxide, which change sample pH and prevent the growth of bacteria. All experiments including ion exchange SPE included pH controls to avoid false positive results.
The HPLC fractionation was performed on a C18 matrix, which is normally used to separate mixtures of mostly hydrophilic to moderately hydrophobic compounds. Many plant extract components with antibacterial activity were not sufficiently retained by the C18 stationary phase, suggesting that hydrophilic interaction chromatography, ion-pairing reverse-phase or even ion chromatography may be required to separate these highly polar compounds. Additionally, multiple injections had to be made to obtain enough fractionated plant extract for antibacterial activity testing. The efficiency of HPLC decreases when more extract is loaded onto the column. Multiple small injections are favoured if the efficiency of the fractionation is required, making preparative HPLC a slow technique.
Our exploration of the insect/plant pairs suggests that it is straightforward to identify antibiotic resistance in the insect gut associated bacteria and to identify plant extracts with antibiotic activity. Isolation of bioactives from complex natural product extracts in quantities sufficient for structure elucidation, however, remains a challenging task.
Antibacterial activity of the plant extracts
We have identified antibacterial activity in all plant extracts investigated, apart from the cabbage extract. As discussed before, it is not known which plants are more likely to have medicinal properties, making it difficult to hypothesise about the significance of this finding.
We only identified a mild antibacterial activity in the eucalyptus leaf extract. Antibiotic activity in eucalyptus extracts is normally described in the context of the eucalyptus essential oil. Eucalyptus oil (from E. globulus) is known to have an antibacterial activity, but it is relatively low compared to other essential oils [38]. Essential oils are notorious for their high content of reactive and unstable compounds. However, our plant extract processing methodology focused on and selected for hydrophilic compounds, as hydrophobic oily substances interfered with preliminary HPLC fractionations.
The cabbage extract was the only plant extract tested that had no antibiotic activity when tested against the bacteria from the Diamondback moth larvae guts and corresponding type strains. Very few studies describe antibacterial activity in cabbage, and those that do only find weak activity [39], suggesting that it is rooted only in anecdotal evidence of traditional medicine.
Ragwort contains high concentrations of pyrrolizidine alkaloids, which are responsible for its toxicity to cattle and horses. Because of the toxicity of pyrrolizidine alkaloids to generalist herbivores, the compounds are thought to be a feeding deterrent, however it was also suggested they protect the plants against microbial attack [40]. In our experiments ragwort extract was one of the most antibacterial plant extracts tested against insect gut bacteria, but the pyrrolizidine alkaloids were not detected in the plant extract. There are no data about antibacterial activities of pyrrolizidine alkaloids, but they have a moderate antifungal activity against plant pathogens [41]. These data suggest that pyrrolizidine alkaloids are not the compounds responsible for the antibiotic activity in the ragwort extract.
Unusually many HPLC fractions of the lavender extract were active against the bacteria tested. Individually each fraction was weakly active and together they contributed to the mild antibiotic activity of the plant extract. Weakly active lavender extract is consistent with mild evolutionary pressure on the rosemary beetle gut bacteria to develop resistance mechanisms. These data are not consistent with the antimicrobial activity of lavender essential oil. Lavender essential oil is a potent antimicrobial, active against both bacteria and fungi, but there is no consensus on which component of the essential oil is responsible for the activity [42]. Even though the scientific evidence indicates lavender has antibacterial properties, problems around standardization of the oil and the purification of the active components hamper the development of lavender preparations into a therapeutically useful agent.
The potato leaf extract was weakly antibiotic. Potato leaves, similarly to ragwort, are toxic to many herbivores, but they lack the strong antibacterial activity of ragwort extract. Potato leaves have not been previously investigated for antibacterial activity, as other potato waste products have been. Potato tubers unsuitable for sale have been investigated as a source of antibacterial activity [43], and had a mild level of antibiotic properties which have been attributed to phenolic compounds in the extracts. In summary, potato waste products are an attractive source to investigate potential bioactivity. However we have shown the level of antibiotic activity is relatively low and we have not pursued the purification of the active fraction of the extract.
Antibiotic activity of the Madagascar periwinkle metabolites
Madagascar Periwinkle is a widely-studied plant in terms of its secondary metabolites. It has been found to produce terpenoid indole alkaloids that are used in cancer therapy [44], but its antibacterial activity is less welldocumented. Extracts of leaves, stems, flowers and roots have been demonstrated to have antibacterial activity [45] and some of the main alkaloids present in Madagascar Periwinkle have been investigated for antibiotic activity [16]. The most active alkaloid found previously from this plant species was vindoline, but the MIC established for it was much higher than in our study. Additionally, Madagascar Periwinkle seeds contain antibacterial proteins CRCI and CRCII (C. roseus cystatin I and II) [46]. These proteins are thiol protease inhibitors and exhibit antibacterial activity against Es. coli and Staphylococcus aureus, but were ineffective against Ba. subtilis. At 25 μg/mL the zones of inhibition were 11-14 mm in diameter. The authors hypothesise that the cystatins are at least partially responsible for the medicinal properties of Madagascar Periwinkle in traditional preparations.
Our investigation is the first attempt to link the antibacterial properties of the Madagascar Periwinkle extract to the metabolites present in the plant. We have discovered that the main indole alkaloids have mild antibiotic activity against a wide range of bacteria. Vindoline had the highest antibacterial activity, especially when tested against a strain with a permeable cell membrane (Table 6). In general the compounds had MIC values an order of magnitude higher than most commonly used antibiotics, but they would be worth investigating further to determine their targets and modes of action. Indole alkaloids from Madagascar Periwinkle are already wellstudied and that body of knowledge makes them attractive potential leads in antibiotic discovery.
Conclusions
Using six exemplar insect/plant pairs we have shown that insect guts contain both environmental bacterial strains and strains typically associated with humans. In general, we found only a small number of culturable bacterial strains (2-6) from the insect guts, but a metagenomic analysis of the Giant Lime Green Stick insect gut showed that a larger number of species are likely to be present (at least 50-100). There has not been extensive metagenomic analyses of insect gut microbiota, but it appears that insect guts generally contain relatively few microbial species compared with mammalian guts [47]. We demonstrated that bacteria that could be cultured from the insect guts can be antibiotic-resistant (compared to their type strains from culture collections), as had been predicted from literature observations [6].
However, we also found that some species showed no difference in susceptibility to antibiotics between the gut-isolated and the type strains; this was not unexpected. More surprisingly we found several instances where the type strain was more resistant to antibiotics than the gut-isolated strain. In some cases, this can be rationalised by the fact that the origin of the type strain was a clinical isolate, but in other cases such rationalisation was not possible. Taken together these data suggest that comparing antibiotic susceptibilities in gut-isolated and type strain is not necessarily a reliable comparison and that the differences in the antibiotic resistance between gut-isolated strains and type strains cannot be used as a reliable indication of antibacterial activity in the food plant.
We confirmed the extracts of several of the plant species used have antibiotic activity, but that some plants, such as cabbage, lack obvious antibacterial activity. The susceptibility of the bacterial strains to the plant extracts can be used as a guide for identifying plant extract fractions with antibiotic activity. It was possible to fractionate the plant extracts and to identify fractions with antibacterial activity. These fractions were shown to be multi-component mixtures and further separation was not attempted. However, in the case of Madagascar Periwinkle, we could identify compounds in both the root and leaf extract with antibiotic activity and we established the minimal inhibitory concentrations of these and related compounds. Even though these metabolites only exhibit a relatively modest antibacterial activity, they are potentially interesting candidate compounds to follow.
Workflow
For each plant-insect pair the same work flow was followed. First the insect gut bacteria were identified and their antibiotic susceptibility profile was assessed. Then we tested the gut bacteria for susceptibility to the plant extract. Both the antibiotic susceptibility and plant extract susceptibility of the gut bacteria were compared between strains isolated from the insect guts and corresponding type strains obtained from culture collections. Our interest was in culturable strains, so that we could test them for susceptibility. However, in one case (Diapherodes gigantea) we used a culture-independent method to assess the variety of bacterial strains present in the gut microbiota (see below).
Insects
Three adult Giant Lime Green Stick insects (Diapherodes gigantea) were provided by the John Innes Centre (JIC) insectary after they had died of natural causes. The insects were immediately frozen and stored at −20°C until dissection. Five Diamondback moth (Plutella xylostella) larvae, ten Death's-head Hawkmoth (Acherontia atropos) larvae and ten Beet Armyworm (Spodoptera exigua) larvae were obtained from the JIC Insectary. Eleven Cinnabar moth (Tyria jacobaeae) larvae were collected from Ragwort that was growing wild on the playing fields of University of East Anglia (Norwich, UK). Ten Rosemary beetles (Chrysolina americana) were collected from lavender plants on the JIC site. Where possible the insects were starved for two hours before dissection to enrich the gut contents in bacteria. The insects were flash frozen in liquid nitrogen and surface sterilized in 70% ethanol with a subsequent rinse in distilled water.
Insect dissection Rosemary beetle (Chrysolina americana) gut sample preparation
The rigid exoskeleton of the beetles prevented accurate dissection of the gut contents. Instead whole insects were homogenized with sterile micropestles in 1.5 mL Eppendorf tubes with 200 μL PBS buffer (8.0 g/L NaCl, 0.2 g/L KCl, 1.44 g/L Na 2 HPO 4 and 0.24 g/L KH 2 PO 4 ).
Other insect dissections
All steps of the dissection procedure were carried out in a biological safety cabinet. A sterile Petri dish was used as a dissecting surface. Using a sterile razor and stabilising the insect using sterile forceps the head of the insect was removed. Still stabilizing the insect body, an incision was made on the ventral side starting at the previous cut and running down the length of the body. The gut was identified as a large tubular structure running along the body. Where possible the gut tissue was cut open and the frozen gut contents were collected. In other cases the entire gut was removed. Gut contents were suspended in 200 μL PBS buffer and diluted 10, 100 and 1000 times. 50 μL of each dilution was plated on LB (LMM0202, Formedium; final concentrations: 1% (w/v) tryptone, 0.05% (w/v) yeast extract, 1% (w/v) NaCl), pH adjusted to 7.0 with HCl), LBG (LB medium plus glucose at 20 g/L (G8270, Sigma Chemicals)) and TSA (17.0 g/L peptone from casein, 3.0 g/L peptone from soymeal, 2.5 g/L glucose, 5.0 g/L NaCl, 2.5 g/L K 2 HPO 4 , pH adjusted to 7.3 with HCl), media (1% agar (AGA03, Formedium)). Plates were incubated for one to three days at 30°C in aerobic conditions. Plates was also incubated using oxygen-absorbing pouches and sealed bags using Anaerocult (Merck) to generate anaerobic conditions. Isolates were streaked out on fresh agar plates and incubated as before. This ensured the purity of the bacterial isolates before identification.
Identification of insect gut bacteria
Each collected insect was dissected and had its gut bacteria cultivated. We mainly used culture-dependent techniques for characterising the insect gut microbiota, in order to obtain strains that could subsequently be used in antibacterial activity testing of the plant extracts. However, in one case (Diapherodes gigantea) we used a culture-independent method to assess the variety of bacterial strains present in the gut microbiota (see below). For a summary of all isolated bacterial species, see Table 1. Each isolate was identified using 16S PCR with alkaline PEG reagent using 63f and 1389r primers (5′-CAGGCCTAACACATGCAAGTG-3′ and 5′-ACGG GCGGTGTGTACAAG-3′) and Taq DNA polymerase (28,104,Qiagen). Single colonies were picked from agar plates and resuspended in 500 μl distilled water and centrifuged at 6000 rpm for 4 min. Without disturbing the pellets, 490 μl sample was removed and 100 μl alkaline PEG reagent was added. The samples were mixed well by pipetting and incubated for 15 min at room temperature. 1 μl was added to the Taq PCR mix, prepared according to the manufacturer's instructions. Reactions were carried out in a PTC-200 Thermo Cycler (MJ Research). The initial denaturation was carried out for 10 min at 95°C, followed by 30 cycles of denaturation (95°C for 1 min), annealing (57°C for 1 min), and extension (72°C for 2 min). The final extension was carried out at 72°C for 10 min. The PCR products were soaked at 10°C until further use. PCR products were separated on 1% agarose TAE (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8.0) gels and purified using QIAquick PCR Purification Kit (28,104,Qiagen). DNA was sequenced using a BigDye v3.1 kit (Applied Bioscience) in a 10 μL reaction volume. Reactions contained BigDye 3.1 mix, 1× reaction buffer, 50-100 ng DNA template and 20 μM sequencing primer. Reactions were carried out in a PTC-200 Thermo Cycler (MJ Research). The initial denaturation was carried out for 1 min at 95°C, followed by 30 cycles of denaturation (95°C for 30 s), annealing (45°C for 15 s), and extension (60°C for 4 min). The final extension was carried out at 72°C for 10 min. PCR products were soaked at 10°C until further use. Samples were sent to The Genome Analysis Centre (Norwich, UK) for processing. Sequencing data was returned in the form of .txt and.abi chromatogram trace files. The sequences were trimmed to remove poorly recognized bases and run through the blastn algorithm (http://blast.ncbi.nlm.nih.gov/Blast.cgi) against "Nucleotide collection (nr/nt)" database.Bacteria were identified if the sequence was ≥97% similar to 16S RNA gene in the database and had an e-value close or equal to 0.
Metagenomics gDNA isolation
The genomic DNA (gDNA) was isolated from the lysed larval gut samples using FastDNA SPIN Kit for Soil (MP Biomedicals). 50 μL larval gut sample was added to the lysing matrix tube with 978 μL PBS buffer and 122 μL MT buffer (MP Biomedicals). The samples were homogenized in FastPrep instrument for 3 min at setting 6. The tubes were then centrifuged at 13,000 rpm for 15 min to pellet the cell wall debris. The supernatant was transferred to clean 2.0 mL microcentrifuge tubes and 250 μL Protein Precipitation Solution was added. The solutions were mixed by shaking the tube by hand 10 times. The samples were centrifuged at 13,000 rpm for 10 min to pellet the precipitated proteins. The supernatant was transferred to 15 mL tubes and mixed with 1 mL resuspended Binding Matrix. The tubes were inverted by hand for 2 min to allow binding of DNA and then placed in a rack for 3 min to allow settling of the silica matrix. 500 μL of supernatant was discarded without disturbing the settled Binding Matrix. The settled Binding Matrix was resuspended in the remaining supernatant and transferred to the SPIN™ Filter tubes. The SPIN™ Filter tubes were centrifuged at 13,000 rpm for 1 min. 500 μL SEWS-M buffer was added to the filter tubes and the Binding Matrix was resuspended gently, before centrifugation at 13,000 rpm for 1 min. The centrifugation was repeated to dry the filters of residual wash solution. The spin filters were dried for 10 min at room temperature and for 5 min at 37°C. The Binding Matrix was resuspended in 100 μL DNase/ Pyrogen-Free Water and centrifuged at 13,000 rpm for 1 min to elute the DNA. The samples were separated by electrophoresis on 1% agarose TAE gel to confirm the presence of large gDNA fragments.
Sequencing
The gDNA isolated from the stick insect guts was sequenced at ChunLab (Seoul, South Korea) using the MiSeq Nano platform. The filtered sequences were assembled into contigs, which were then classified into operational taxonomic units based on sequence similarity between them. Taxonomic classification was assigned to each operational taxonomic unit at the species level using the ChunLab's EzTaxon-e database and blastn algorithm [48]. Chimeric sequences, which are contaminants originating from two separate DNA sequences, were filtered out using UCHIME program [49]. The sequencing reads were supplied in .clc format. The sequencing results analysed and visualised with CLcommunity software supplied by ChunLab. All sequence data were uploaded to EMBL-EBI.
Type strains
For each identified gut-isolated strain a corresponding type strain was obtained from one of the following culture collections: Centre de Resources Biologiques de l'Institut Pasteur (Paris), Health Protection Agency (Salisbury) or Deutsche Sammlung von Mikroorganismen und Zellkulturen ( Table S1).
Antibiotic susceptibility testing Assay
The minimal inhibitory concentrations (MIC) for ampicillin, chloramphenicol, ciprofloxacin, kanamycin, rifampicin and tetracycline were determined by broth microdilution [50] for each gut-isolated strain and type strain pair. This panel of antibiotics was chosen based on their varied mechanisms of action [51]. Ampicillin inhibits the final stage of cell wall synthesis leading to cell lysis. Chloramphenicol prevents protein synthesis by inhibiting the peptidyl transferase activity of the ribosome. Ciprofloxacin is a gyrase inhibitor and kills cells by creating breaks in DNA. Kanamycin inhibits protein synthesis by binding to the 30S ribosomal subunit. Rifampicin disrupts RNA synthesis by inhibiting RNA polymerase. Tetracycline prevents protein synthesis by blocking the attachment of the aminoacyl-tRNAs to the ribosome.
To assess the levels of antibiotic resistance in the insect guts, we compared the antibiotic susceptibility profiles of the insect-gut isolated strains to type strains available from culture collections. Type strains were obtained from the CRBIP. HPA and DSMZ culture collections (Additional file 3: Table S1); a summary of the antibiotic susceptibilities is shown in Table 2. Briefly, 96well plates with serial dilutions of antibiotics were inoculated with bacterial suspension and incubated for 24 h at 30°C. The MIC was assigned when instead of a suspension of bacterial growth, a well with a clear broth was present. To confirm lack of bacterial growth, the OD was measured at 600 nm in a CLARIOstar plate reader (BMG Labtech). The control organism was Es. coli ATCC 25922. The MICs for vindoline, loganin, loganic acid, ajmalicine and catharanthine were also determined by broth microdilution for the gut-isolated strains from the beet armyworm, matching type strains, Es. coli (ATCC 25922), Mycobacterium smegmatis (ATCC 700084), Ps. aeruginosa (ATCC 15692) and St. aureus (ATCC 29213). 96-well plates with two-fold dilutions of the compounds were inoculated with bacterial suspension at OD 600 = 0.08-0.11 and incubated for 24 h at 30°C. The MICs were assigned as described above.
Extract preparation
Plant extracts were prepared by homogenizing dried leaves of eucalyptus (Eucalyptus dalrympleana), cabbage leaves (Brassica rapa), lavender leaves and flowers (Lavendula angustifolia), ragwort leaves (Jacobaea vulgaris), potato leaves (Solanum tuberosum), and Madagascar periwinkle leaves and roots (Catharanthus roseus) with methanol. 100 g dried plant material was ground to fine powder either with a pestle and mortar or using an electric coffee grinder (Andrew James Stainless Steel Wet and Dry Coffee, Nut and Spice Grinder). The plant powders were soaked overnight in 300 ml methanol, filtered and soaked again twice in 300 ml methanol. The plant extracts initially had a deep green, nearly opaque colour and were soaked until they were no longer green. The three methanolic extracts were filtered, pooled and concentrated in an EZ-2 Elite evaporator (GeneVac). The extracts were de-fatted by liquid-liquid fractionation with petroleum ether in a 1:1 ratio. Only the methanol fraction was used in subsequent purifications. The crude eucalyptus extract was concentrated and de-fatted with petroleum ether. The aqueous partition was further purified using weak anion exchange solid phase extraction cartridges. The active methanol fraction was further fractionated by HPLC using a C18 reverse-phase column.
Antibacterial activity testing
To assess the antibacterial properties of the plant extracts, we tested the susceptibility of gut-isolated strains and type strains to these extracts. The plant extracts were prepared by drying the plant material in an evaporator (DNA SpeedVac, Savant), grinding it to powder and infusing in methanol at room temperature. The extracts were then filtered and crudely purified by solventsolvent partitioning with petroleum ether.
It should be noted that to compare the antibacterial activity of extracts from different plants, the extracts were standardized. Each extract was dried until a dry pellet was obtained, which was resuspended in methanol (100 μg/mL). (Sometimes waxy material was present in the samples and the pellets would not appear dry. In such cases the pellet was evaporated until no more reduction in volume occurred and then weighed and resuspended in methanol at 100 μg/mL.) Resuspended extracts were used to infuse paper discs before assaying on bacteria from the insect guts and the corresponding type strains. This procedure was carried out to standardise the concentrations of plant extracts and compare their antibacterial properties.
Lawns of bacteria were prepared by overlaying TY agar plates with a mixture of 3 mL overnight bacterial culture at OD 600 0.08-0.11 and 3 mL molten and cooled TY agar. Paper discs (Whatman, 5 mm) were infiltrated ten times with 10 μL aliquots of the extract and placed on the bacterial lawns. After 24 h incubation at 30°C the clear zones were measured.
Solid-phase extraction
Plant extracts with activity against bacteria were purified on solid-phase extraction (SPE) weak anion exchange columns (Oasis WAX 6 cm 3 cartridge, Waters) according to the manufacturer's instructions. Briefly, each column was primed with 6 mL methanol and calibrated with 6 mL 2% formic acid in MilliQ water (Merck) before loading no more than 5 mg sample resuspended in methanol. The columns were then washed with 6 mL 2% formic acid, 6 mL methanol and 6 mL 5% ammonium hydroxide in methanol. The flow through was collected, concentrated and assayed for antibiotic activity as described before.
High-pressure liquid chromatography
The active SPE fractions of plant extracts were subjected to reverse-phase high-pressure liquid chromatography (HPLC). Non-concentrated samples were used for the fractionation to avoid loss of separation when too much sample is loaded. The samples were centrifuged for 10 min. If required the supernatants were serially diluted at 1:10, 1:100 and 1:1000 ratios before injection. The samples were analysed on an HPLC instrument (Shimadzu Prominence/Nexera UHPLC). Separation was performed on an analytical 2 × 100 mm 3 μm Luna C18(2) column (Phenomenex) and semi-preparative 10 × 250 mm 5 μm Luna C18(2) column (Phenomenex), run at 0.3 mL/min for the analytical columns and 3 mL/ min for the semi-preparative columns. All separations were run at 40°C. A gradient of methanol against 0.1% formic acid in MilliQ water (Merck) was used to elute compounds. A general HPLC method consisted of a 10-100% methanol gradient over six column volumes followed by a 100% methanol wash over two column | 2023-01-15T14:46:48.691Z | 2017-12-01T00:00:00.000 | {
"year": 2017,
"sha1": "d18f55219f0a3571180d50288f5a851e8dd625f0",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1186/s12866-017-1133-0",
"oa_status": "GOLD",
"pdf_src": "SpringerNature",
"pdf_hash": "d18f55219f0a3571180d50288f5a851e8dd625f0",
"s2fieldsofstudy": [
"Environmental Science",
"Biology"
],
"extfieldsofstudy": []
} |
8607357 | pes2o/s2orc | v3-fos-license | Discrepancies in assessing home care workers’ working conditions in a Norwegian home care service: differing views of stakeholders at three organizational levels
Background The present study is a follow-up study of factors contributing to an undesirable quality of work environment and sick leave rate in the home care services in a Norwegian municipality. The underlying assumption is that organizational discrepancies in the perceptions and appraisals of significant factors and processes in an organization have detrimental effects on the management of the organization and on work environment conditions. Thus, the study aim is to explore potential organizational discrepancies in the appraisals of factors relating to home care workers’ working conditions. Methods The study, using a mixed-methods design, comprised six home care units. It included survey responses of home care workers (80 respondents, response rate 54 %) and qualitative descriptions of stakeholders’ appraisals of organizational issues gathered through semi-structured interviews (33 interviews with stakeholders at three organizational levels). Results Employees at different organizational levels in the home care services expressed divergent appraisals of factors related to the working conditions of home care workers, including impact of organizational measures (i.e. time pressure, work tasks, a new work program, organizational changes, budget model, budget allocation and coping strategies). Survey responses supported interview descriptions by home care workers. Results suggest that organizational discrepancy serve as an important barrier to a sustainable, well-functioning organization in general and to quality-enhancing changes to work procedures in particular. Conclusions It is recommended to improve communication channels and facilitate the exchange of information across levels to ensure a common understanding of matters significant to the organization of the home care services and to the work environment of home care workers. The prevalence and impact of organizational discrepancy should be included in organization research, particularly when exploring explanatory factors of an unhealthy organization.
In 2003 the municipality in this study received orders from the Norwegian Labor Inspectorate due to a high level of unhealthy time pressure and high sick leave in the home care services (HCS). The municipality responded promptly by giving work environment interventions and other alleviating actions high priority. Findings, based on data from 2009, indicated that the work environment interventions in general were perceived by the HCWs to have a positive effect by improving targeted areas identified to cause work strain. However, concurrent changes induced through production system rationalization, such as unit mergers, restructuring, changes to middle-and executive management, and introduction of new technology and new work programs, resulted in negative exposure effects that negated the positive intervention effects, causing an overall deteriorated work situation [11]. Production system rationalization involves a continuous effort to reduce costs and improve quality of an organization's output (e.g. services offered to patients), and has a predominant negative impact on worker health [10]. The majority of HCWs in the present case reported an increase in perceived time pressure from 2004 to 2009. They reported exposure to several occupational risk factors and a high prevalence of musculoskeletal health complaints [6]. Sick leave rate remained high the next two years; 19.1 % in 2011 vs. 18.1 % in 2009, compared to 15.6 % in 2004. Consequently, the HCS seems to be struggling with persistent challenges that are difficult to overcome.
As prolonged high sick leave represents high costs as well as human suffering, it is a puzzle that HCS administrators have not been able to deal with the problem in the aftermath of the unsuccessful initial interventions. A possible explanatory factor is that decision-makers at municipal level have difficulties in recognizing or accepting the working conditions of workers at frontline. Organizational distance, insufficient inter-level exchange of information, and lack of insight in the interruptive influences of rationalizations have been suggested as key factors in this regard [11]. Stakeholders at different organizational levels may have divergent and potential conflicting perceptions and appraisals of organizational quality and productivity measures (i.e. organizational discrepancy); or holding incompatible "mental models" on organizational matters [14]. Literature on the prevalence and impact of organizational discrepancy is rather scarce, but existing research points to discrepancies in perceptions across organizational levels as potentially causing several detrimental consequences [15][16][17][18], such as obstructing effective implementation and leading to diversity in intervention experiences and outcomes [14].
The present study aims to explore potential discrepancies at three levels of an organization in assessing factors relating to HCWs' working conditions. The underlying assumption is that organizational discrepancies in the perceptions and appraisals of significant factors and processes in the organization have detrimental effects on the management of the HCS and work environment conditions of HCWs. To explore this, interviews with stakeholders at all organizational levels (municipal representatives, home care unit leaders and HCWs) and survey responses of HCWs are included in the present study. The introduction of a new work program in the period 2009-2011 was of special interest in the survey, as findings presented in a previous publication indicated that work programs could have a significant impact on the HCWs' workday [11], and informal conversations indicated inter-level discrepancies in this regard.
Method
The study has a mixed-method design by using a combination of quantitative estimations (of perceived change in time pressure the last 2 years and perceived impact of the new work program introduced to the HCWs), and qualitative descriptions of the work situation in the HCS perceived by stakeholders at three organizational levels (municipal representatives, home care unit leaders and HCWs).
Case description
In Norway, the municipalities are responsible for social welfare and health care of residents. The present HCS is organized by a two-level system. The municipal level constitutes the superior 'administrative' authority responsible for ensuring overall good quality of the services. The home care units are separated geographically, and each unit leader is responsible for managing the unit budget and organizing the daily operations. The HCWs are the frontline workers performing the services.
There has been a marked increase in home care services given to patients of all ages with a multiplicity of diagnoses and disabilities. An individual assessment executed by the municipal health and welfare office determines what medical help patients are entitled to receive. Each patient is given an ADL score ("Activities of Daily Living") based on functional level, which constitutes the basis for money and time allocation to the units. An individual resolution contains specified activities for the HCWs to carry out in the patient's home (labeled 'direct time' activities), such as putting on support stockings, giving medications, assisting with morning routines, measuring insulin level, wound treatment and so on. A specified amount of time is allocated to each task on the work list, typically varying from 10 to 45 min. Time estimation per assignment is decided at unit level and may be altered according to changes in patients' functional level. However, the 'norm' from which time is allocated to the units is based on centrally determined ADL scores and the HCS' total time production over a certain period of time.
A normal work day is a 7.5 h shift and may consist of 10 to 27 visits. Additionally, a considerable part of the workday involves transferring between patients and administrative tasks such as documentation and report writing (all labeled 'indirect time' activities), yet these activities are not explicitly specified in the work list. A percentage distribution of 70 ('direct time') -30 ('indirect time') has generally been required to prevent deviation from unit budgets. A new work program introduced to the HCS in the period 2009-2011 has had consequences on the HCWs' workday. The program involves specification and systematization of HCWs' work duties and responsibilities at patients' discharge from hospital as well as follow-up of patients' overall functioning in their homes. The work tasks related to the program involve observations, assessments and documentation (i.e. indirect tasks), formalized in checklists that serve as tools for quality assurance through standardization of services.
Participants
At study start in 2009, 6 (of 11) units signed up for participation (here entitled unit A-F). In the follow-up survey of 2011, one unit (unit E) was omitted due to low response rate. HCWs employed in a ≥ 50 % position were invited to fill in a questionnaire. Interview informants were selected through purposive sampling based on seniority (minimum 7 years) and employment fraction (≥50 %). Table 1 shows detailed information concerning data collection and participation during the study period. The present study is based on the follow-up questionnaire in 2011 and all interviews.
The key informant is a HCW affiliated with a unit not part of the study. Through self-initiated time studies he gained comprehensive knowledge concerning actual time distribution to various work tasks during a typical work day, and was thereby considered to be an important informant in the present study.
Procedure
Initial conversations were carried out with an inspector of the Norwegian Labor Inspectorate, three representatives of the municipality and all six unit leaders to gain insight in aspects of the HCS such as the organization of work duties, organization-specific work demands and significant changes and events relevant for the composition of interview guide and questionnaire. Prior to the data collection, one of the researchers participated on staff meetings at each unit to present the study and give practical information about participation. Questionnaires and informed consent forms were placed in individual envelopes and sent to the units where the unit leaders handed them out to the HCWs. The HCWs individually returned the filledin questionnaire and informed consent form to one of the researchers by mail, and were remunerated with NOK 200 (24€). The data collection was carried out between May 31 2011 and December 8 2011, and was finally closed on February 22 2012. A written inquiry concerning interview participation was placed in selected personal shelves. Participating HCWs filled-in an informed consent form and were remunerated with NOK 300 (36€). The study was approved by the municipal executive, the Regional Committees for Medical and Health Research Ethics (REC) (no. 4.2009.19) and Norwegian Social Science Data Services (NSD) (no. 21036).
Questionnaire
The questionnaire comprised altogether 52 items. The present study utilizes a self-formulated item regarding perceived changes in time pressure the last 2 years. Response categories ranged from 1 (considerable less) to 5 (considerable more) with a neutral mid-point, recoded to a threepoint response scale ("less", "no change" and "more"). The question "If change, why?" enabled open comments. Also, respondents were asked to evaluate the impact of the new work program on their work situation. Response categories ranged from 1 (no impact) to 5 (considerable impact) recoded to a three-point response scale ("minor impact", "some impact" and "considerable impact"). The question "In what ways?" enabled open comments. IBM Statistical Package for Social Sciences (SPSS) version 19 was used to compute frequency distributions. The open comments were briefly formulated by the respondents, thus categorization and enumeration of responses was straight forward.
Interviews
The interviews with HCWs and unit leaders lasted approximately 1 h, and slightly longer for municipal representatives. All interviews were carried out individually, audio-recorded and later transcribed. The interview Material also analyzed in a previous publication [11]. Data guides covered topics concerning work environment of HCWs, work tasks, budget system, and perceived changes in such the last years. The paramount questions were "How do you perceive your work situation today?" and "How do you perceive changes in your work situation to affect you and your work?" Main questions were followed by probe questions such as "Can you give an example of this?", "What do you think caused this?" and so on to stimulate rich descriptions. The interview guides for all levels covered the same topics in order to enable inter-level comparisons. Thus, unit leaders and municipal representatives were asked the same questions as the HCWs yet from their own perspective with regards to the HCWs' work situation and the organization of the HCS. The interview with the representative from the economy department, however, was less focused on working conditions at frontline, but rather on the allocation model (with regards to time). At two occasions after the interviews were carried out, a preliminary understanding of the topics in question, together with new questions where information was lacking, was emailed to unit leaders and municipal staff in order to clear-up possible misunderstanding (mainly involving financial and organization issues). Clarification was also requested when needed during the interviews with HCWs.
Qualitative analysis
The interviews of HCWs from 2009 were reanalyzed and supplemented with interviews of unit leaders and municipal representatives with the purpose of exploring potential organizational discrepancies. Interview topics were similar across the different times of data collection. Descriptions presented by informants interviewed twice (i.e. unit leaders and municipal representatives) did not differ with regard to time of interviewing. The interview data was analyzed by Template analysis as this analyzing technique "works particularly well when the aim is to compare the perspectives of different groups of staff within a specific context" ( [19], p.257). Hence, the data material was separated by organizational level during the analyzing process in order to explore potential inter-level similarities and discrepancies. The software QSR NVivo 9 (http:// www.qsrinternational.com/products_nvivo.aspx) was utilized to aid in organizing and examining the data. The interview guide and findings presented in a previous study [11] served as basis for the initial template consisting of two higher-order themes: 1) "Strenuous work situation for HCWs" and 2) "Economy"; themes initially indicated to be characterized by inter-level discrepancies. The analyzing process was carried out by identifying higher-order themes and further scrutinizing the contents of these themes to identify and differentiate lower-order themes, resulting in a final template presented in the results section. This final template served as basis for interpretation and illumination of the data, in line with recommendations by King [19]. Table 2 shows survey responses on perceived changes in time pressure (A) and impact of the new work program (B). For time pressure, the left column shows the distribution of scale responses regarding changes in time pressure (1 missing). The majority of HCWs perceived increased time pressure. The categorization of open comments concerning what caused the perceived change in time pressure is presented in the right column, with the large majority of factors related to increased work demands (i.e., due to number and characteristics of patients).
Quantitative results
For the new work program, about a third of the HCWs (24) perceived the new work program to have considerable impact on the work situation, while 20 and 23 HCWs perceived some or minor impact, respectively (13 missing). Efficiency demands (14) Increased patient group/ heavier cases (12) Increased workload (10) New work program (4) No comments (8) Decreased (2) Extra personnel (1) Improved work organization (1) No change (32) B. New work program c Negative impact (27) Time-consuming (11) Additional work tasks (10) More stress (6) Positive impact (9) Improved work situation (4) Improved overview (3) Better understanding of patients' situation (2) Negative/positive impact (4) Additional work tasks, but improved quality (4) No comment on quality of impact (27) Table 3 summarizes interview results by showing the final template, differentiated by organizational level. Quotes illustrating inter-level discrepancies or similarities are representative for typical descriptions at each level. The final template was somewhat changed from the initial template by including an additional higher-order theme, resulting in three high-order themes; 'strenuous work situation for HCWs' , 'economy' and 'coping strategies'. The lower-order themes developed within each of them are presented next.
Strenuous work situation for HCWs
Time pressure emerged as a main theme in the interviews, across all organizational levels. All HCWs spontaneously described their work day as busy and stressful, characterized by a constant fight against time. Several HCWs pointed at a negative trend towards increasingly hectic work situation. Increased demands of efficiency were perceived to be prioritized at the expense of compassionate care, which had to be downgraded due to lack of time. These descriptions were in accordance with accounts by unit leaders who in unison referred to employees expressing a feeling of being in a constant hurry and having too much to do. However, a few unit leaders qualified their accounts by pointing at behavior indicating a work situation not so stressful after all (e.g. HCWs coming in for lunch on time), claiming that a 'stop watch-principle' is a natural part of the job, and that the perception of time pressure is 'highly subjective'. All unit leaders' descriptions pointed at the same trend as the HCWs of increasing time pressure, and that work tasks were now more directed towards necessary medical treatment than traditional 'caring' activities. However, some descriptions were characterized by a more pragmatic flavor pointing at the necessity of change; indicating a more productive work organization whereby only medical, documented services were to be carried out ("We are not hired to be social companions"). Representatives at municipal level extended on the notion of time pressure being a subjective matter by pointing at a perceived destructive organizational culture worshiping negativity. HCWs' complaints about lack of time were described as merely 'phrases' uttered with no basis in reality. Accordingly, all representatives at municipal level expressed doubt of an objective increase in time pressure for the HCWs, as "no reasons for such an increase exist". Both work environment interventions and objective figures were referred to as reasons to expect the contrary. This was further supported by a representative from the economy department who also expressed a lack of understanding of HCWs' perceptions of increased time pressure.
Increased indirect task demands emerged as a theme as some HCWs described a tendency towards a more challenging work situation characterized by increased demands of indirect time activities (i.e. documentation, writing reports, and transferring over a large geographical area). These activities were described as increasing sources of time pressure and stealing time from other (direct) work tasks, considered to be more important. Similar descriptions were presented by the unit leaders claiming that the services had become more comprehensive, involving additional areas of responsibility. At municipal level, however, descriptions regarding (direct and) indirect time were somewhat different, focusing more on developments in inspection plans and the controlling body as opposed to actual changes in work duties or workload. Documentation requirements were described as being the same as in previous years, yet with a more strict internal control, to achieve improved quality and thereby a positive change to the HCS. The representative from the economy department described a registered increase in use of indirect time, but could not say whether this was due to an increase in indirect work tasks (as stated by the HCWs) or improved registration (as required by the municipal level).
For the HCWs, the new work program was an example of a requirement directed from municipal level that entailed additional duties (indirect work tasks), resulting in increased workload and time pressure. Unit leaders, being responsible for managing the work organization and budget, expressed frustration regarding the introduction of the work program without supplementary resources. One unit leader illustrated this by referring to a consequent reduction in direct time from 63 to 49 % during implementation, implying that the new indirect tasks steal time from direct patient-related activities that constitute the basis for money allocation. One unit had to discontinue the implementation of the work program in order to secure "the minimum requirement of producing services" because of the additional demands placed on the HCWs. Another unit leader, who had assisted in composing the checklists, admitted that the program involved new, time-consuming work tasks, but she also pointed at benefits such as improved quality. This description was in accordance with representatives at municipal level who focused on quality assurance when describing the work program. Although agreeing that the program could be perceived as involving additional indirect tasks with no financial allocation, they emphasized the UL: Main stressor, Increasing "Some say there is more time pressure and stress now compared to how it used to be. That is probably correct, too". "The time pressure is at the sacrifice of care. You calm and make people feel safe by holding their hand, so we (HCS) are probably heading in the wrong direction…" ML: Improvement expected, cultural problem "'We have never run so fast before!' -this is a phrase, like many other expressions that are often incorrect". "My experience is that a culture revolves around the focus on time pressure, a so-called 'enjoying the misery'-culture".
ED:
No objective reason "The HCWs complaining over more time pressure… It makes no sense. It is part of 'the game'".
Indirect task demands
HCWs: Increasing amount "Well, we have all these requirements directed at us, a lot more now than it used to be, about documentation and all kinds of stuff we must register andwe're not able to do half of what we are supposed to…"; "It is at the expense of the patients, we need to take the time from somewhere, and then we steal it from them".
UL:
Increasing amount "There is an increased demand of indirect tasks. It is very time-consuming, and we need more time (to accomplish these tasks)". "We are required to spend 72 % of our workday on direct patient-related activities, but we can't manage to do that. With all these indirect tasks, there is no chance".
ML: Increased internal control
"There are no reasons for an increased need of time to indirect tasks (…) The requirements of documentation today are the same as previous years, but the internal control is stricter. The follow-up is tighter now. And will be getting even tighter".
ED:
Increased indirect time "We see a registered increase in use of indirect time, but we don't know whether this is due to an increase in indirect work tasks or as a consequence of improved registration". "'The program' † † was not supplied with any resources; it is eating up our timetremendously. This is an example of a contradictory pressure between demands and results, in which we are decreed a lot of tasks, and not even a penny comes with it". "This steals a lot of time and resources. Compulsory attendance here and therethis goes off our budget, but we have no choice".
ML: Improved quality "If they follow the checklists they will feel confident in doing their job. This is quality assurance". "By doing it correct the first time they won't have to do things twice". ED: (neutral) "When they (ML) implement work programs, they must think that there is a gain to it. Maybe they (HCWs) think it is additional work".
Organizational changes
HCWs: Source of work strain "There have been some organizational changes, you know, and it tears on us -When they start with all that, I'm just like: AGAIN!?".
UL:
Increased sick leave "We had high sick leave for a while after the merger, and I definitely believe it was related to the merger and the subsequent effects; these things affect people; new routines, everything must be changed. It caused a lot of commotion, and it took some time before we were back on track; Even after 6 years it is still 'them and us', they are sitting on each side of the table".
ML:
Change is inevitable "The nature of the home care services is to deal with changes and adjustments. That is home care". "There are some myths out there; that organizational changes are terrifying. I believe it is crucial to adjust the organization". Transferring is not taken into consideration on our work lists. We have protested against that many times, we think it is HIGHLY unfair!" UL: Incorrect model (money) "The budget model is not right. This district deals with a lot of psychiatry and since they are in good health physically, they score a very low ADL. They generate no money. It is wrong!". "The real world is different from what the budget model tells us". "A lot of the anticipated positive outcomes of the program such as systematization of work tasks and improved order and efficiency. However, when responding to a question on how the HCWs were to incorporate these additional tasks in practical terms, one representative stated that the work lists should be "squeezed together" and made tighter with "extra assignments per work list in order to relieve others to enable time for documentation", implying increased efficiency demands for the HCWs.
Organizational changes, in particular unit mergers, were described by HCWs in all units affected as causing a more strenuous work situation. Consequences such as culture clashes, unit sizes too large, chaos, larger geographical distances, new localities, establishing and mastering new roles, and new ways of cooperating were cited as stressful factors wearing on them and being sources of increased time pressure. Unit leaders also described undesirable consequences of mergers, mainly in terms of increased sick leave and impaired work environment, with prolonged effects. One unit leader pointed at a severe budget overrun due to poor supervision and thereby loss of control as an example of a negative consequence of organizational changes. In contrast, representatives at municipal level were more visionary in their descriptions of organizational changes, focusing on enhanced quality, financial motives and an inherent need for change in the HCS. Though they did express awareness of adverse consequences for the HCWs and unit leaders ("… there are expenses related to the employees and financial management"), these consequences were brushed aside by focusing on the necessity for change and development.
Economy
Two lower-order themes related to economy emerged in the interviews; budget model and budget allocation. The theme Budget model emerged as informants at all levels presented descriptions concerning the rationale for allocating time to work assignments and money to the units. The HCWs expressed a lot of frustration concerning indirect tasks not being included in the allocation of time. Only direct tasks were specified on their work lists and time to be spent on documentation, transferring, reporting etc. was not included in the time estimates. This was perceived as highly unfair considering the growing amount of such tasks that had to be carried out at the expense of direct-patient time (as described in two preceding themes). The key informant, who has done informal research on HCWs' time distribution during their workday, supported the HCWs' perceptions. Most unit leaders' descriptions of the budget Table 3 Final coding template of interview data, differentiated by organizational level, with quotes to illustrate organizational discrepancies (Continued) new work tasks are of indirect character not taken into consideration as they don't generate any money".
ML:
Improved model "We have been working very well for a long time on quality assurance and budget allocation based on patient weight; they get more resources if patient cases are heavy.
We have a GOOD allocation of the resources available".
ED:
Misunderstandings on lower levels "I can imagine this being a matter of discussion between the HCW and the unit leader. It seems to me that there are some misunderstandings of the model out there…" 2.2 Budget allocation HCWs: Poor economy, less money allocated "You are told that the budget situation is getting worse and worse. And you are told that you have to do more and more in less time. It affects you, you feel; Ok, fine, there is a limit foryeahfor what you can handle". "Will there be directed even more requirements on us? I cannot run any faster and do my job any faster just because the council is short on money. It will make me ill, so I cannot do that".
UL:
Tight budget impossible to comply "We notice a tightening, and we are under more supervision now. Now we have to hand over monthly reports about the financial condition". "It is an invariable requirement that we stick with our budget. But we haven't managed that. That is serious business, oh my, I think it's horrible".
ML:
Increased allocation "I believe many employees would say things have gotten worse, but that is just nonsense if you look at the budget increments". "The budget allocations up till now have actually increased. We have objective figures that tell us that".
ED: N/A N/A
3. Coping strategies HCWs: Stress-reducing and time-managing "We steal time from the patients (…) Or else we wouldn't make it".
UL:
Reducing expenses "My main goal has been to reduce excess spending".
ML: Improving quality "There is an increase in patient load. Patients return home sooner, in a worse condition. I believe that if they (HCWs) focus on following the checklists in the work program † †, they are safe".
ED: N/A N/A
HCWs Home Care Workers; UL Unit Leaders; ML Municipal Level; ED Economy Department †Open survey responses, † † Anonymity model were in accordance with the HCWs' , but their main focus was on money rather than time. Money was also a concern for some unit leaders when they described how the budget model was based on a perceived incorrect coupling between ADL and financial disbursement. Some patient groups were perceived to demand a lot of resources despite low ADL score, and unit leaders managing units in geographical areas with poor socio-demographic status regarded this to be highly unfair. This perception was in great contrast to the descriptions presented at municipal level focusing on the improved and now well-functioning budget model. One representative confirmed unit leaders' perceptions of unfairness, but referred to objective figures stating otherwise. This representative, as well as the economy representative, hinted at a lack of understanding of the budget model among HCWs and unit leaders as the reason for frustrations. Budget allocation involves descriptions concerning resource allocation in the HCS. Informants at all levels described an increased focus on financial issues. The HCWs described poor economy and budgetary constraints as the antecedents of impaired working conditions due to restrictions to hiring temporary workers, vacant posts and forced mergers. This was perceived to cause increased workload and sick leave, and the situation was considered to get worse every year. All units had struggled with budget overrun, and unit leaders described frustrations concerning a tight budget with financial demands that were perceived impossible to comply with. In turn, overspending involved a reduction of 3 % in next year's budget allocation, making this a vicious circle. Reasons for budget overruns were mainly related to hiring of temporary staff ("Last year this unit spent € 1 mill on temporary staff. That is a sick amount of money!"), but also on car repairs, and compulsory projects and work programs, initiated by the municipal level, that required time and resources off the budget. Some unit leaders, like the HCWs, described a trend towards tighter budgets. One unit leader described a stable budget allocation suffering from increased efficiency demands, and another one to stricter control of reporting directed from the municipal level. In contrast, all municipal representatives described an increase in budget allocations, and one representative expressed awareness of an inter-level discrepancy in this regard. However, a couple of representatives described stricter economic control directed at the units, and tightening actions such as a reduction in time to carry out administrative (indirect) tasks, due to a more challenging financial situation for the municipality.
Coping strategies
Informants at all organizational levels described making use of strategies or actions to deal with a strenuous work situation and a poor financial condition. The HCWs mainly described strategies utilized to cope with a strenuous work situation, involving both stress-reducing strategies and actions employed in order to get through the work list on restricted time. Stress-reducing strategies varied among the HCWs. Some described 'active' strategies ("I call for assistance when I need it"), other described 'experience-based' strategies ("I am right here, right now, I cannot worry about the next patient on the list"), and strategies of a 'distancing' character ("I do what I'm supposed to do without engaging myself too much or else I get drained"). Also, 'task-minimizing' coping strategies such as downgrading certain work tasks, skipping lunch, and cutting down time on assignments were described as essential actions in order to get through the work list on time. This was supported by the key informant who described employees distributing time where it is most needed, by either transferring time from one patient to another in more need, or from direct-time to indirect-time activities. He referred to an internal report revealing that HCWs shorten their assignments with 20-25 % on average. Unit leaders, on the other hand, mainly described actions to avoid excess spending, such as removing a car, reducing use of temporary workers, and avoiding employment of new workers in vacant positions. A few unit leaders also mentioned approaches to alleviate worker stress by actions such as hiding extra assignments by adding these to the work lists before the HCWs come to work in the morning, thereby making them unaware of the additional tasks. Municipal representatives described actions mainly focusing on quality assurance and improvement of the services by initiating different work programs. For example, in response to a potential increase in worker strain due to a new national reform, one municipal representative stated that checklists introduced in the new work program would ensure that the HCWs felt safe in their professional role when facing increased challenges in their work.
Discussion
The results of this study support the initial assumption by showing that employees at different organizational levels in the HCS express divergent appraisals of factors related to the working conditions of HCWs, including impact of organizational measures. Follow-up survey responses supported the interview descriptions of HCWs, pointing at a further increase in time pressure due to rationalizationrelated measures. Some of the survey responses pointed at quality enhancing consequences of the new work program, as described by municipal representatives, yet the majority of survey responses revealed adverse strenuous side effects. We posit that organizational discrepancies as identified in the present study serve as an important barrier to a sustainable, well-functioning HCS in general and to qualityenhancing changes to work procedures in particular. Figure 1 presents a graphic summary of the study results and inferences: Organizational level discrepancies in priorities and beliefs concerning allocation of resources and the execution of care are identified. The top-down demands and bottom-up responses are incongruent, indicating a sub-optimal system involving stated demands exceeding perceived capacity. Primary communication across the organizational levels is illustrated by arrows, in which the top-down communication of demands is the strongest (demonstrated by bold arrows). The divergent appraisals of priorities and beliefs bring about further discrepancies in the perception of HCWs' work situation, particularly with regards to strain from time pressure. Discrepancies in the appraisal of timeallocation, indirect tasks, and impact of a new work program and organizational changes translate into a straining work situation for the HCWs as the organizationand allocation-system is determined by decision-makers at a higher organizational level holding a divergent appraisal of significant matters. Finally, Fig. 1 illustrates how organizational discrepancy may result in poor organizational and employee health and functioning. In the following, the themes identified as areas of discrepancies will be discussed.
Time pressure
A high level of time pressure has previously been identified in the HCS [4,5,20,21], and particularly in the present case [11,22]. Informants at all organizational levels agreed that time pressure was the most prevalent straining factor for the HCWs. However, the HCWs' perceptions of their work situation as hectic and stressful were somewhat brushed aside by the management levels as being exaggerations and symptoms of an unfortunate culture, jargon and bad attitudes. These kinds of explanations may be held due to the considerable effort invested to alleviate precursors of time pressure, and hence improvement should be expected. Thus, HCWs' perceptions of impaired working conditions are difficult to comprehend at municipal level for other reasons than being mainly a cultural problem. DeJoy [23] claims that the bias of 'observers' (i.e. managers; as opposed to employees being 'actors') toward internal attribution and underestimating external factors is likely to be exaggerated with increasing organizational distance. Accordingly, the Fig. 1 Graphic summary of results and inferences. Note: At each of the three organizational levels, upper and lower bullet points represent priorities and beliefs, respectively (left side). Primary communication across organizational levels is illustrated by vertical arrows. Bold arrows indicate dominant processes. Organizational discrepancies in priorities and beliefs result in further discrepancies in perceived work situation for HCWs (middle) and bring along adverse consequences for the organization and the individual worker (right side), treated in separate publications [6,11] municipal representatives elaborated on the interventions implemented (to argue for expected improvement) while the HCWs pointed at increased demands (to argue for perceived impairment). Unit leaders, being more closely connected to the daily work duties, supported the HCWs' perception of increased time pressure and more 'instrumental' care. Although this was regarded as an unfortunate trend and a cause of distress for the HCWs, two unit leaders and all municipal representatives described this trend as necessary for improved quality of the Services in terms of being better organized and more efficient. The municipality changed the health division's name from "Health and Care Services" to "Health and Welfare Services", further signalling the downgrading of caring aspects. Erosion of the emotional aspects of home care work due to rationalization has been described in previous research (e.g. [8,24]). The apparent organizational discrepancy regarding "quality of care" has been discussed as 'instrumental' versus 'relational' valued care provision [9,25,26] or 'economical rationality' (i.e. care based on resources available) versus 'caring rationality' (i.e. care based on patients' needs) [27]. This discrepancy has implications for HCWs' perceptions of quality on their job performance. Aronson and Neysmith [28] argue that the standardized, rational way of organizing home care work is difficult to translate into practice, and results in a discrepancy between HCWs' official job descriptions and their own accounts of their work.
Increased indirect task demands
Informants at all organizational levels agreed that there have been changes related to indirect tasks the last years. However, there is an organizational discrepancy concerning what this change involves in terms of added workload. HCWs and unit leaders described an increase in the amount of indirect tasks, resulting in additional time pressure for the HCWs. On the contrary, municipal representatives considered that the impact of changes was primarily stricter control by tighter follow-up of employees (more standardization in rationalization terms). However, this is likely to involve an actual increase in workload for the HCWs, as tighter control results in reduced flexibility; tasks are actually being carried out instead of postponed or given low priority due to time pressure. In this sense, increased control translates into increased requirements and workload for the HCWs, whereas the municipal level representatives can deny requiring additional work tasks, and thereby justify the decision of not allotting more resources. The lack of common understanding likely implies an added source of strain for HCWs. Worrall and Cooper [15] claim that top-level management, being removed from daily operations, are out of touch with the reality of the organization as perceived by the staff. It can be difficult to recognize practical implications in frontline work. Warren et al. [16] refer to striking organizational level differences in the recognition and reporting of occupational risk factors, in which employers seem to have a particularly difficult time identifying psychosocial risks.
New work program
Different work programs have repeatedly been introduced in the Health and Welfare Services [11]. The results reveal organizational discrepancy in both focus and perceived impact of the new work program. The HCWs and the unit leaders presented similar descriptions, though the HCWs focused on increased workload and the unit leaders primarily focused on negative consequences related to work organization (in terms of releasing time to the increasing amount of indirect tasks) and budget. The municipal representatives acknowledged these changes, yet their focus was rather on quality assurance and increased efficiency. However, the responsibility of putting the work program into practice lies with the unit leaders, and the combination of increased workload, tighter work lists and no added financial resources makes it a tough priority for both unit leaders and HCWs. Due to stricter internal control of indirect tasks, the program results in increased strain for the HCWs and a source of frustration for the unit leaders. Decision-makers may be blinded by positive intentions and fail to realize the practical impact of such programs at frontline.
Organizational changes
Similarly, a high degree of organizational discrepancy is identified in focus and perceived impact of organizational changes (e.g. unit mergers). The HCWs described how organizational changes directly and indirectly resulted in adverse consequences for their daily operations. Unit leaders focused more on subsequent consequences such as increased sick leave and budget overrun, as lie within their areas of responsibility and constitute objective indicators of their capabilities of managing the units. These consequences have been communicated to the municipal level, yet descriptions presented by the municipal representatives indicate that these reactions are not given much consideration. The discrepancies also indicate divergent perceptions of impact on service quality; while municipal representatives pointed to enhanced quality of the Services, the HCWs and unit leaders described negative consequences causing impaired quality within areas of their concern. This is in accordance with previous research showing organizational discrepancy in perceived impact of organizational change [15,29,30].
Budget model
There is also a high degree of organizational discrepancy concerning the rationale for allocation of resources. The HCWs were most concerned with an unfair allocation of time, while the unit leaders focused on the allocation of money based on a perceived incorrect model. This is highly discrepant from the descriptions presented by municipal representatives. HCWs and unit leaders referred to "heavy" patient cases (e.g. psychiatric diagnoses, drug abuse) in terms of requiring time and money exceeding what is expected according to the ADL. Conversely, the municipal representatives referred to level of ADL as an indicator of functional level. As the budget model is based on the ADL, an organizational discrepancy in the understanding of what constitutes a "heavy" patient load may have impact on the appraisal of accuracy of the budget model.
Budget allocation
Appraisals concerning budget allocation are to a large extent characterized by incompatible descriptions. The HCWs pointed at a reduction in budget allocation while municipal representatives claimed the opposite. Again, their descriptions are characterized by dissimilar focus, related to their areas of responsibilities: The HCWs focused on how poor economy had adversely affected their working conditions, the unit leaders expressed concern related to budget overrun, while municipal representatives referred to budget increases and optimistic objective figures. However, both unit leaders and municipal representatives described processes that might have resulted in an actual impaired economic situation for the units. For example, increased efficiency demands, several tightening actions and various compulsory work projects (e.g. projects on dental hygiene, protective footwear etc.) are concurrent requirements directed from the municipal level that may override resource increments. Accordingly, the increased focus on financial issues, restrictions to hiring temporary workers, stricter reporting control and budget overrun signalize an impaired economic situation to employees at lower organizational levels.
Coping strategies
Organizational discrepancies are also identified concerning actions employed to cope with work-related challenges, such as types of strategies applied and for what purpose. HCWs focused on dealing with time pressure and a strenuous work situation, unit leaders on economic issues, and the municipal representatives on quality assurance. Again, this reflects their areas of responsibilities; the patients, the budget and the HCS, respectively. For the HCWs and unit leaders, it was the scarcity of resources (i.e. time and money) that needed to be managed in the most constructive way to be able to carry out their job satisfactorily; for HCWs sufficient time is a prerequisite in order to do 'a good job' , for the unit leaders the ability to stick with their budget is an indication of doing 'a good job' , while the municipality is responsible for ensuring 'good quality' of the Services. However, despite well-intended strategies of quality assurance (i.e. introducing work programs; stricter control), this has had adverse effects on HCWs and unit leaders by requiring more time and more money. Semmer [31] claim that negative intervention effects in terms of increase in time pressure and work load occur with some regularity.
Suggestions for practitioners
The present results reveal several areas as possible targets for improvement in order to reduce organizational discrepancy and associated adverse consequences. Corrective measures should be introduced with an overall intention of developing intra-organizational 'shared mental models' of priorities and beliefs concerning matters of significance to the management of HCS (cf. Fig. 1). Improved communication channels across organizational levels likely involve increased sharing of knowledge and understanding of status quo at each level. This is particularly crucial for the HCWs' responses as these seem under-communicated upwards in the system. The top level should be regularly connected to what is actually happening at frontline [15]. Warren et al. [16] suggest removing institutional barriers to the flow of knowledge in an organization characterized by discrepancy. In the present case, unit leaders are intermediaries and the strengthening of their position as messengers can facilitate the exchange of information across levels. In fact, the present HCS has arranged for weekly meetings of unit leaders and municipal representatives. However, New Public Management policies such as unit leaders being measured on budget figures and 'unit performance' may serve as obstacles to report on real unit conditions. Municipal representatives did express some awareness of HCWs' perceptions, but didn't seem to fully understand or acknowledge them as true. A more holistic view of how the organization is managed is recommended, especially during times of change, with increased focus on the relational, as opposed to technical aspects of management [15,32]. Finally, the discrepancy between 'objective job requirements' and 'personal job expectations' likely involves emotional strain for HCWs [11]. Thus, clarification of care and quality is recommended to ensure common understanding and expectations with regard to work tasks and job performance (see also [27]).
These points are antidotes for this particular case, but should nevertheless be important to consider for organizations in general. Generalizability of the present results is limited due to the nature of case studies. However, general principles regarding the disruptive effect of organizational discrepancy to the management of organizations appear valid for most organizations and can probably be transferred to sectors outside health care and in other countries. The portrayals provided by the stakeholders at three organizational levels were elaborative and clearly discrepant though Norwegian municipalities (including the HCS) are characterized by a flat hierarchical structure [33,34].
The disruptive effect of organizational discrepancy is likely even more significant in countries where organizations are characterized by greater organizational distance between decision-makers at top-level and workers at front line.
No considerable differences were detected amongst the units studied in this present study. Nevertheless, results of a previous study showed unit variations in effect of work environment interventions, in exposure to organizational changes, in sick leave rates and in excess spending and budget deficits. Other between-unit factors include geographical extension (impacting on transferring distances between patients' homes) and patient characteristics (psychiatric diagnosis and drug addiction more densely populated in certain areas) [11]. Some of these elements have been treated in the qualitative themes 'organizational changes' and 'budget model'. However, as the aim of the study was to explore the phenomenon of organizational discrepancy at three levels of an organization, the data was analyzed with regard to organizational levels irrespective of unit affiliation.
Conclusions
The study shows that divergent priorities and beliefs of stakeholders at different organizational levels concerning factors significant to the management of the HCS and the performance of care bring about discrepancies in the appraisals of HCWs' working conditions. Further, organizational discrepancy was identified to serve as an important barrier to a well-functioning organization and to quality-enhancing changes to work procedures. It is recommended to improve communication channels across organizational levels, and to facilitate the exchange of information regarding top-down messages (concerning 'the system') and bottom-up messages (concerning HCWs' responses). Also, clarification of expectations to ensure a common understanding of care and quality is recommended. The results and inferences presented in this present study further suggest that prevalence and impact of organizational discrepancy should be included in future research on organizational issues. | 2018-04-03T04:21:37.229Z | 2015-07-25T00:00:00.000 | {
"year": 2015,
"sha1": "a365b79ed8aa1fc56015f77be5b68f7f660b8b6e",
"oa_license": "CCBY",
"oa_url": "https://bmchealthservres.biomedcentral.com/track/pdf/10.1186/s12913-015-0945-6",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "a9fc630150613de73d76243b26d87c19a4155d4d",
"s2fieldsofstudy": [
"Medicine",
"Psychology",
"Business"
],
"extfieldsofstudy": [
"Medicine"
]
} |
31328915 | pes2o/s2orc | v3-fos-license | Hypercalcemia of Malignancy.
Hypercalcemia of malignancy (HCM) is a common concern in patients being treated for cancer, affecting over a quarter of this population. There are multiple causes of HCM, including humoral HCM, osteolytic HCM, ectopic hyperparathyroidism, and vitamin D-secreting lymphomas. Common signs and symptoms of HCM can range from mild gastrointestinal disturbances and fatigue to seizures, coma, or even cardiac arrest depending on the severity of the laboratory abnormality. Treatment has evolved in recent years and varies based on the underlying cause of the HCM. Management options include aggressive hydration, bisphosphonates, denosumab, calcitonin, and corticosteroids. It is imperative that advanced practitioners understand the pathophysiology behind the HCM so that proper treatment can be chosen. Early and appropriate treatment is key to successful outcomes. It is also important for continuous monitoring to occur alongside treatment for HCM to prevent potential adverse effects. Finally, the ultimate resolution of HCM comes only from the treatment of the underlying malignancy; therefore, all previously undiagnosed patients should be referred to an oncologist early after HCM is recognized.
H ypercalcemia of malignancy (HCM) has long been recognized as a common paraneoplastic syndrome associated with poor prognosis in cancer patients. It is estimated that HCM a ects roughly 30% of patients with cancer, with many studies reporting an even higher incidence in those with advanced stages of cancer (Endres, 2012;Mirrakhimov, 2015). Common malignancies associated with HCM in the United States include breast, lung, multiple myeloma, squamous cell carcinoma of the head and neck, renal, ovarian, and certain lymphomas (Gastanaga et al., 2016). Recent research has indicated that overall HCM prevalence seems to be declining in recent years, likely due to the increased use of prophylactic intravenous (IV) bisphosphonates in high-risk malignancies, particularly in cases of bone metastasis and multiple myeloma (Gastanaga et al., 2016;Terpos et al., 2013). However, HCM is still a frequent oncologic emergency that requires appropriate and prompt intervention to improve overall patient outcomes. The purpose of this paper is to review and summarize current literature and guidelines related to the diagnosis and management of HCM.
PATHOPHYSIOLOGY
Although the primary goal of this article is not to discuss the pathophysiology of HCM at length, at least a brief review is necessary to understand how HCM is diagnosed and managed, as treatment can vary at times depending on the underlying cause. Hypercalcemia of malignancy can be divided generally into four types.
The first and most common cause of HCM is humoral HCM. Approximately 80% of HCM cases are thought to be related to this type. Humoral HCM is caused by malignant tumors secreting parathyroid hormone-related protein (PTHrP), which acts similarly to parathyroid hormone (PTH) in the body. Effects of increased PTHrP secretion result in increased osteoclast activity and bone resorption with subsequent release of calcium into the bloodstream. This occurs secondary to the mechanism in which PTHrP acts on osteoblasts, which enhances production of the receptor activator of nuclear factor-κB (RANK) and its associated ligand (RANK-L). Humoral HCM is typically associated with renal, ovarian, breast, and squamous cell cancers as well as human T-cell lymphotropic virus (HTLV)-associated lymphoma (Dellay & Groth, 2016;Mirrakhimov, 2015;Wijaya, Oehadian, & Sumantri, 2014).
The second most common cause of HCM is known as osteolytic HCM. It accounts for 20% of HCM cases and is a result of excessive calcium release from bone secondary to malignant cells in the bone marrow. Local inflammatory response from these malignant cells stimulates osteoclast activity. This type of HCM is frequently seen in malignancies with bone metastasis as well as breast cancer and multiple myeloma (Dellay & Groth, 2016;Mirrakhimov, 2015;Wijaya et al., 2014).
The last two types of HCM are both relatively rare and account for less than 1% of HCM cases. Vitamin D-secreting lymphomas can cause hypercalcemia by increasing intestinal absorption of calcium and decreasing renal calcium excretion. Ectopic hyperparathyroidism causes increased secretion of PTH and acts nearly the same as PTHrP in the pathway discussed above to cause increased serum calcium. Small cell lung cancer and adenocarcinomas are most commonly associated with ectopic hyperparathyroidism, although a handful of documented cases have shown increased PTH secretion due to a parathyroid carcinoma (Cheng, Kuzhively, & Baim, 2017;Dellay & Groth, 2016;Mirrakhimov, 2015;Wijaya et al., 2014). Finally, it is important to note that primary hyperparathyroidism is possible in cancer patients and needs to be distinguished from HCM (Mirrakhimov, 2015).
CLINICAL MANIFESTATIONS
Signs and symptoms of hypercalcemia can vary significantly and depend on the level of hypercalcemia as well as how rapidly the calcium level is rising. Hypercalcemia of malignancy typically occurs more acutely compared to other causes of hypercalcemia, often making the clinical manifestations appear more severe in these cases (Dellay & Groth, 2016;Malangone & Campen, 2015;Mirrakhimov, 2015;Wijaya et al., 2014).
Hypercalcemia is often divided into categories of mild, moderate, severe, and life-threatening depending on the degree of serum calcium elevation. The generally accepted upper limit of normal (ULN) for corrected serum calcium is 10.5 mg/dL. [Note: Corrected serum calcium = serum calcium + 0.8 × (4 -serum albumin)]. It is vital to calculate the corrected serum calcium based on a patient's serum albumin level if it is below normal; otherwise, many cases of hypercalcemia may be missed. Mild hypercalcemia ranges from the ULN to a corrected serum calcium of 11.9 mg/dL; moderate is defined as a corrected serum calcium from 12.0 mg/dL to 13.9 mg/dL; severe is defined as a corrected serum calcium greater than 14.0 mg/dL. Life-threatening hypercalcemia is often defined as a corrected serum calcium that falls in the severe range accompanied by critical clinical manifestations such as coma or cardiac arrest (see Table 1; Gastanaga et al., 2016;Malangone & Campen, 2015).
Hypercalcemia of malignancy can also affect the gastrointestinal system. Common signs and symptoms of mild hypercalcemia are anorexia and constipation. Higher degrees of hypercalcemia may also cause nausea, vomiting, weight loss (if FELDENZER and SARNO REVIEW chronic), and even pancreatitis or peptic ulcer disease (Dellay & Groth, 2016;Malangone & Campen, 2015;Mirrakhimov, 2015;Wijaya et al., 2014).
Clinical manifestations related to the neurologic and musculoskeletal systems are often much more pronounced at moderate to severe levels of hypercalcemia and include anxiety, depression, cognitive dysfunction, lethargy, weakness, fatigue, hyporeflexia, confusion, stupor, and in the most severe cases, coma. Hypercalcemia of malignancy specifically has been associated with posterior reversible leukoencephalopathy syndrome, which presents with headaches, seizures, and subcortical edema on imaging (Dellay & Groth, 2016;Malangone & Campen, 2015;Mirrakhimov, 2015;Wijaya et al., 2014).
Although the systems listed above are the most commonly affected, other reported signs and symptoms of hypercalcemia include pruritus, generalized abdominal pain, and bone pain (Dellay & Groth, 2016;Malangone & Campen, 2015;Mirrakhimov, 2015;Wijaya et al., 2014). It is especially important to recognize this constellation of symptoms in patients not previously diagnosed with a malignancy, as the presentation of hypercalcemia may be an important diagnostic clue to finding an underlying cancer in some patients (Mirrakhimov, 2015).
DIAGNOSTIC APPROACH
Once hypercalcemia is confirmed using clinical presentation and corrected serum calcium level, the next step is to determine which pathophysiologic process discussed above is responsible for causing it. In many cases, a clear history of cancer diagnosis will accompany the patient, so a thorough history and physical exam is the obvious first task. However, even with confirmed malignancy, there are varying HCM mechanisms. Therefore, further investigation should be done. This is an important step, as it can have an impact on which treatment will best manage the hypercalcemia (Dellay & Groth, 2016;Malangone & Campen, 2015;Mirrakhimov, 2015;Rosner & Dalkin, 2012;). First, the serum PTH level should be measured. This assists in determining whether or not primary hyperparathyroidism is the sole cause of hypercalcemia (which, as noted above, can still be the case in some cancer patients), or if HCM is likely present. If the serum PTH level is elevated Note. Summary of common signs and symptoms associated with hypercalcemia, based on severity of laboratory abnormality. Information from Malangone & Campen (2015); Wijaya et al. (2014). or high normal, it is likely that primary hyperparathyroidism is present and the patient should likely be referred to endocrinology for appropriate management (Malangone & Campen, 2015;Mirrakhimov, 2015). In some cases, chronic hypercalcemia with corrected serum calcium levels less than 12.0 mg/dL is also suggestive of primary hyperparathyroidism and should be considered in the differential (Dellay & Groth, 2016). If the serum PTH level is decreased or low normal, HCM is the likely diagnosis and checking a serum PTHrP level should be considered for further evaluation (Malangone & Campen, 2015;Mirrakhimov, 2015). In many institutions, this is a send-out laboratory test that may take up to a week to obtain results. Therefore, assessment of PTHrP level should not delay immediate treatment for hypercalcemia, but it can be an important value for selecting continued management options in HCM (Dellay & Groth, 2016).
If the PTHrP level returns elevated, then humoral HCM is likely the cause. If the PTHrP level is normal or low, the next step is to consider evaluating for bony metastatic disease (if this is not already confirmed) as well as checking the vitamin D level (1,25-dihydroxyvitamin D; Dellay & Groth, 2016;Malangone & Campen, 2015;Mirrakhimov, 2015;Rosner & Dalkin, 2012). Elevated vitamin D levels should point advanced practitioners toward considering lymphoma as the underlying cause (Malangone & Campen, 2015).
Finally, if laboratory evaluation reveals normal values of PTH, PTHrP, and vitamin D, other causes of hypercalcemia must be considered. These include use of certain medications such as thiazide diuretics, corticosteroids, lithium, and vitamin A or D oversupplementation, as well as renal failure (Malangone & Campen, 2015;Mirrakhimov, 2015).
MANAGEMENT OPTIONS Aggressive Hydration
Intravenous fluids are widely accepted as a firstline treatment for hypercalcemia, including HCM. Patients with HCM typically present with dehydration secondary to multiple underlying factors, including anorexia, nausea, vomiting, and nephrogenic diabetes insipidus as discussed previously in clinical manifestations of HCM. Dehydration in turn causes a further decrease in kidney function and a subsequent decrease in calcium excretion (Wijaya et al., 2014).
Aggressive rehydration, generally with 0.9% sodium chloride solution (also known as normal saline), should be initiated rapidly once hypercalcemia is confirmed with laboratory values (see Table 2; Dellay & Groth, 2016;Malangone & Campen, 2015;Wijaya et al., 2014). Practitioners often begin with 1 to 2 liter boluses followed by maintenance IV fluids at a rate of 200 to 500 mL/hr, titrating to a urine output of 100 to 150 mL/hr (Dellay & Groth, 2016;Malangone & Campen, 2015;Wijaya et al., 2014). The goal here is to increase the glo- FELDENZER and SARNO REVIEW merular filtration rate and increase calcium excretion as well as to dilute serum calcium. Intravenous fluids are an accepted treatment in patients with preexisting renal dysfunction and congestive heart failure; however, careful monitoring of fluid status is mandatory (Sternlicht & Glezerman, 2015).
Historically, loop diuretics such as furosemide (Lasix) have been used in conjunction with aggressive IV hydration in the treatment of HCM. However, loop diuretics have been shown to carry a high risk of causing volume depletion and further electrolyte imbalances (Sternlicht & Glezerman, 2015). Evidence currently supports restricting the use of loop diuretics to those patients who are experiencing volume overload secondary to extensive IV fluid administration, and even then, close monitoring of ongoing fluid status is required. Routine use of loop diuretics is no longer recommended (Mirrakhimov, 2015;Rosner & Dalkin, 2012;Sternlicht & Glezerman, 2015).
Aggressive hydration usually elicits a clinical response quickly in HCM. Within a few hours, serum calcium levels decrease by approximately 2 mg/dL (Sternlicht & Glezerman, 2015); however, the effect is often transient and further management or treatment of the underlying malignancy is required (Rosner & Dalkin, 2012).
Bisphosphonates
After or often in conjunction with aggressive hydration, IV bisphosphonate treatment should be started to treat HCM. Bisphosphonates are very well studied in this population of patients and are considered to be quite safe and effective (Sternlicht & Glezerman, 2015;Wijaya et al., 2014).
Bisphosphonates act by inhibiting bone resorption. They bind to the bone and reduce or prevent osteoclast activity, which as discussed above is a major factor in the development of HCM (Dellay & Groth, 2016). Oral absorption of bisphosphonates is extremely low (1%-2%); therefore, IV administration is generally the route seen in the clinical setting (Sternlicht & Glezerman, 2015). First-generation bisphosphonates include etidronate (Didronel) and clodronate (Bonefos), neither of which are commonly used in practice anymore. Second-generation bisphosphonates, or those that contain nitrogen, are much more potent and are frequently used today (Mirrakhimov, 2015;Wijaya et al., 2014). Examples of this class include zoledronate (zoledronic acid or Zometa), pamidronate (Aredia), ibandronate (Boniva), and alendronate (Fosamax). Zoledronic acid and pamidronate are the two bisphosphonates that are currently approved by the US Food and Drug Administration (FDA) for the treatment of HCM in the United States (Sternlicht & Glezerman, 2015;Wijaya et al., 2014). Bisphosphonates can also be used for osteoporosis and to prevent skeletal-related events in metastatic bone disease; however, the dosing and timing is often different in these cases (Mirrakhimov, 2015).
Bisphosphonates should be started early in the treatment of HCM, and certainly within the first 48 hours. It generally takes 2 to 4 days to see a clinical response, with the lowest calcium level generally being 7 to 10 days after bisphosphonate administration (Sternlicht & Glezerman, 2015;Wijaya et al., 2014). Effects last roughly a few weeks, averaging 1 to 3 weeks, with some studies reporting up to 4 weeks in duration. Zoledronic acid has been shown to be superior to pamidronate in both clinical response and amount of time until hypercalcemia relapse. Zoledronic acid has shown a nearly 83% response rate in patients at 7 days after administration compared to approximately 64% with pamidronate. In the same study, zoledronic acid averaged 30 days until relapse vs. 17 days with pamidronate (Dellay & Groth, 2016). Due to this finding, zoledronic acid is often the first choice for bisphosphonate treatment (Sternlicht & Glezerman, 2015;Wijaya et al., 2014). It is important to note that due to their mechanism of action, bisphosphonates are usually most effective in HCM due to metastatic bone disease. In humoral HCM, higher levels of PTHrP are associated with a lower response rate, which needs to be taken into consideration when using bisphosphonates (Rosner & Dalkin, 2012).
There are several side effects that accompany bisphosphonate use. The most common are infusion-related fever and flu-like symptoms, including mild bone aches and pains (Mirrakhimov, 2015;Rosner & Dalkin, 2012;Sternlicht & Glezerman, 2015). There is also a risk of nephrotoxic acute tubular necrosis, particularly in patients who have advanced cancer, have been previously treated with bisphosphonates, and with concur-rent use of nonsteroidal anti-inflammatory drugs (Mirrakhimov, 2015;Rosner & Dalkin, 2012). Due to this risk, experts recommend careful use of bisphosphonates if patients have preexisting kidney injuries or a decrease in renal function (Dellay & Groth, 2016).
Zoledronic acid is generally contraindicated in patients with a creatinine clearance (CrCl) of less than 30 mL/min. Dose reductions, or even holding a dose altogether, of both zoledronic acid and pamidronate should occur in patients with lower-than-normal CrCl (Rosner & Dalkin, 2012). Ibandronate, although not FDA-approved for the treatment of HCM, is believed to be safer for use with patients who have reduced renal function; however, more research is needed in this area (Mirrakhimov, 2015).
One other rare but serious side effect of bisphosphonate use is osteonecrosis of the jaw (ONJ). The risk of ONJ is believed to be less than 1 in 10,000 and is most common in patients diagnosed with multiple myeloma or metastasis to the bone (Mirrakhimov, 2015;Rosner & Dalkin, 2012). Other risk factors leading to this complication include recent dental extractions, poor dental health, a history of smoking, and long-term, regular use of bisphosphonates (Rosner & Dalkin, 2012). A comprehensive oral exam should be completed for each patient prior to starting bisphosphonate treatment (or as soon as possible after starting treatment). Regular dental care is necessary during and after bisphosphonate treatment as well, and dental procedures should be avoided within one month of bisphosphonate use to decrease the risk of complications (Kapoor, Sikka, Arora, & Chaudhary, 2013).
Bisphosphonate cost varies significantly; one dose of zoledronic acid averages $713, while one dose of pamidronate averages $104. Practitioners' choice of bisphosphonate should consider the degree of HCM, potential side effect risks, and cost, especially that associated with long-term use for patients (Dellay & Groth, 2016).
Denosumab
Denosumab (Xgeva) was approved by the FDA in 2010 for the prevention of skeletal-related events in patients with solid tumor malignancies as well as osteoporosis (Dietzek, Connelly, Cotugno, Bar-tel, & McDonnell, 2015;Rosner & Dalkin, 2012). In 2018, the FDA expanded approval to include the prevention of skeletal-related events among patients with multiple myeloma who have developed bone metastases (Varga, Laubach, Anderson, & Richardson, 2018). Denosumab is a fully human monoclonal antibody that binds to RANK-L. This in turn prevents RANK-L from binding to its normal receptor, which as discussed earlier causes a significant decrease in osteoclast activity and subsequent bone resorption (Dellay & Groth, 2016;Dietzek et al., 2015). Although not FDAapproved for the indication of HCM, multiple case reports support denosumab as an option, particularly in the treatment of HCM refractory to bisphosphonates (Ashihara et al., 2016;Dietzek et al., 2015).
Few studies exist on denosumab in the treatment of HCM; however, Hu and colleagues (2014) reported a 64% response rate by 10 days after administration in patients who received denosumab after not having a clinical response to bisphosphonates for 30 days. This study also reported a median duration of response of 104 days, which is significantly greater than what has been reported with bisphosphonate treatment (Hu et al., 2014). Patients who received denosumab for the treatment of HCM refractory to bisphosphonates also demonstrated a decrease in clinical manifestations of HCM, specifically cognitive signs and symptoms such as altered mental status (Hu et al., 2014).
Furthermore, a handful of case studies have depicted safe and successful use of denosumab in patients with HCM and coexisting renal failure (Dellay & Groth, 2016). Denosumab is not eliminated by the kidneys, so dose adjustments are not required in patients with altered renal function (Rosner & Dalkin, 2012); however, more research is needed in this area before routine use in this population can be recommended (Dellay & Groth, 2016).
Denosumab is generally well tolerated, with the most common side effects being arthralgia and mild dyspnea (Hu et al., 2014;Mirrakhimov, 2015;Rosner & Dalkin, 2012). There is also a small risk of ONJ, which is relatively equivalent to the risk associated with zoledronic acid use (Mirrakhimov, 2015). Due to denosumab being a monoclonal antibody, increased infection risk following administration is often a concern; however, current evidence does not support any significant risk of infection associated with its use at this time (Mirrakhimov, 2015).
Denosumab is not commonly used in the treatment of HCM; less than 1% of patients hospitalized for this condition received denosumab as a part of their treatment regimen (Wright et al., 2015). However, with a growing number of studies supporting its safe and effective use, it should be strongly considered as a management option. This is particularly true in patients who are refractory to bisphosphonate use or who have concurrent renal concerns (Dellay & Groth, 2016;Dietzek et al., 2015;Hu et al., 2014;Rosner & Dalkin, 2012). Other factors to consider include ease of administration (subcutaneous vs. IV administration of bisphosphonates), and its fairly high cost as a monoclonal antibody (Ashihara et al., 2016;Dietzek et al., 2015;Hu et al., 2014).
Other Treatment Choices
A few other management options exist in the treatment of HCM, many of which are indicated in specific clinical scenarios. Hemodialysis in the setting of HCM has decreased significantly since bisphosphonates have entered the market; however, it is still appropriate at times (Wijaya et al., 2014). With patients who present with HCM as well as acute or chronic renal failure or a history of heart failure, where aggressive hydration is truly not indicated or is dangerous to the patient, hemodialysis with a very low or no calcium dialysate is feasible and acceptable (Mirrakhimov, 2015;Rosner & Dalkin, 2012;Wijaya et al., 2014).
Another choice in the management of HCM is the administration of calcitonin (Miacalcin, Fortical). Calcitonin acts to inhibit all osteoclast activity as well as to prevent renal absorption of calcium (Dawson, Todd, & Walton, 2014;Dellay & Groth, 2016). This medication exhibits a very rapid, albeit transient, hypocalcemic effect. A decrease in serum calcium level is generally seen about 2 hours following administration; however, calcitonin must be administered every 6 to 8 hours to maintain this effect (Dellay & Groth, 2016). It carries with it great risk for hypocalce-mia, rebound hypercalcemia, and tachyphylaxis that usually develops after 2 or 3 days of treatment (Rosner & Dalkin, 2012). Because of this, calcitonin is generally only used in the acute phase of hypercalcemia, and it is often used as a bridging treatment until effects of bisphosphonates administered are seen (Dellay & Groth, 2016). One case study does report successful chronic use of calcitonin in the treatment of HCM (Dawson et al., 2014); however, this is not routinely seen in practice. Side effects most commonly include nausea, vomiting, and pain or irritation at the injection site (Rosner & Dalkin, 2012).
Corticosteroids are considered a management option in the setting of vitamin D-secreting tumors or lymphomas (Dellay & Groth, 2016). They generally work by decreasing gastric absorption of calcium (Mirrakhimov, 2015). It takes approximately 4 days to see a response as evidenced by a decrease in serum calcium level; in cases where no clinical response is achieved by the tenth day of treatment, corticosteroids should be discontinued and another option should be utilized (Dellay & Groth, 2016;Malangone & Campen, 2015). Treatment is often started with IV hydrocortisone for 3 to 5 days, and if there is a clinical response, patients can be transitioned to oral prednisone for an additional 5 to 7 days of treatment (Rosner & Dalkin, 2012). Significant side effects occur with corticosteroid use, including hyperglycemia, hypertension, steroid-induced psychosis, peptic ulcer disease, and an increased risk for tumor lysis syndrome, which contraindicate long-term use (Malangone & Campen, 2015).
In the past, gallium nitrate (Ganite) was also used to treat HCM. The exact mechanism is unknown, but it appeared to inhibit osteoclast activity and bone resorption (Rosner & Dalkin, 2012). This medication was known to carry a risk of many significant side effects (Rosner & Dalkin, 2012;Sternlicht & Glezerman, 2015). Gallium nitrate was withdrawn from the market in the United States in 1995 and discontinued by the manufacturer in 2012, although while in use it did show effectiveness in treating HCM (Mirrakhimov, 2015;Sternlicht & Glezerman, 2015).
There are also new and developing treatment options for HCM. Cinacalcet (Sensipar), a calcimimetic, works by decreasing PTH production. This has been used successfully to treat HCM secondary to parathyroid carcinoma (Mirrakhimov, 2015;Sternlicht & Glezerman, 2015). Minimal side effects include nausea, vomiting, and headaches. Further research is needed to determine if cinacalcet is effective in treating HCM due to other causes (Doyle & Malcolm, 2014;Sternlicht & Glezerman, 2015). Finally, anti-PTHrP antibodies are currently being tested in animal models. This approach offers future hope for additional treatment options in HCM that is refractory to or not appropriate for bisphosphonate use (Mohammad & Guise, 2016;Sternlicht & Glezerman, 2015).
Lastly, it needs to be understood that all the treatment options discussed here are truly only meant for stabilizing the patient until an oncologist can appropriately treat the underlying malignancy, as this is the only true cure for HCM.
PATIENT MONITORING AND OUTCOMES
Careful monitoring of clinical response is vital for all patients during and after treatment for HCM. Many of the management options discussed here predispose the patient to a risk of hypocalcemia, which carries with it its own dangerous clinical manifestations (Body, Niepel, & Tonini, 2017). This risk is highest in patients who are not monitored closely or who are exposed to long-term bisphosphonate use (Narechania, Thiruchelvam, Lokhande, Kistangari, & Daw, 2015;Wijaya et al., 2014). Continued monitoring of the serum calcium level is necessary, and in the event that hypocalcemia occurs, it will likely need to be treated with vitamin D supplementation and calcium repletion (oral or IV, depending on the severity of hypocalcemia), as well as other appropriate supportive measures (Body et al., 2017).
Other steps can also be taken to improve patient outcomes and reduce the risk for redeveloping hypercalcemia. Advanced practitioners should ensure that any calcium supplementation is discontinued, including that in the form of oral replacement therapy or parenteral feeds (Wijaya et al., 2014). It is also important to discontinue vitamin D supplementation, calcitriol, lithium, thiazide diuretics, and any other medications that may independently contribute to hypercalcemia (Sternlicht & Glezerman, 2015;Wijaya et al., 2014). Finally, weight-bearing exercise as tolerated should be encouraged unless contraindicated (Wijaya et al., 2014).
IMPLICATIONS FOR ADVANCED PRACTITIONERS
Hypercalcemia is a serious neoplastic syndrome that is highly prevalent in cancer patients. Prognosis is generally poor, with as many as 50% of those diagnosed with any degree of HCM not surviving for greater than 30 days (Wright et al., 2015). Prompt recognition and treatment of this condition is key. Practitioners need to be aware of how management options and suggested treatments have evolved over the past 5 to 10 years.
A recent study by Wright and colleagues (2015) revealed that many hospitalized patients with HCM did not receive appropriate treatment. Nearly 33% of patients did not receive bisphosphonate treatment, which is widely considered the first-line therapy in the absence of obvious contraindications. Patients frequently received steroid therapy (27% of patients), despite the fact that this is only truly indicated in the setting of vitamin D-secreting lymphomas, which account for only 1% of HCM cases. Additionally, it was common for contraindicated medications such as thiazide diuretics and lithium to be continued while patients were being treated for HCM. Also of note is that treatment was less likely to follow guidelines when provided by hospitalists or family practice providers (Wright et al., 2015). Advanced practitioners are often in front-line positions to quickly recognize evolving diagnoses in patients, as well as to modify treatment plans accordingly. In clinical scenarios where time is of the essence, such as in HCM, advanced practitioners need to be adequately prepared to effectively and efficiently manage their patients. It is imperative that advanced practitioners utilize up-to-date and evidence-based treatment to improve patient outcomes, particularly in the treatment of HCM. l
Disclosure
The authors have no conflicts of interest to disclose. | 2020-04-08T20:24:45.244Z | 2018-07-01T00:00:00.000 | {
"year": 2018,
"sha1": "0a1d9c52640dfa80fe6b168c8f80de1822554fcb",
"oa_license": "CCBYNC",
"oa_url": "https://link.springer.com/content/pdf/bfm:978-3-642-85073-8/1?pdf=chapter%20toc",
"oa_status": "GREEN",
"pdf_src": "PubMedCentral",
"pdf_hash": "0a1d9c52640dfa80fe6b168c8f80de1822554fcb",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
46929665 | pes2o/s2orc | v3-fos-license | Investigation of protein quaternary structure via stoichiometry and symmetry ınformation
The Protein Data Bank (PDB) is the single worldwide archive of experimentally-determined three-dimensional (3D) structures of proteins and nucleic acids. As of January 2017, the PDB housed more than 125,000 structures and was growing by more than 11,000 structures annually. Since the 3D structure of a protein is vital to understand the mechanisms of biological processes, diseases, and drug design, correct oligomeric assembly information is of critical importance. Unfortunately, the biologically relevant oligomeric form of a 3D structure is not directly obtainable by X-ray crystallography, whilst in solution methods (NMR or single particle EM) it is known from the experiment. Instead, this information may be provided by the PDB Depositor as metadata coming from additional experiments, be inferred by sequence-sequence comparisons with similar proteins of known oligomeric state, or predicted using software, such as PISA (Proteins, Interfaces, Structures and Assemblies) or EPPIC (Evolutionary Protein Protein Interface Classifier). Despite significant efforts by professional PDB Biocurators during data deposition, there remain a number of structures in the archive with incorrect quaternary structure descriptions (or annotations). Further investigation is, therefore, needed to evaluate the correctness of quaternary structure annotations. In this study, we aim to identify the most probable oligomeric states for proteins represented in the PDB. Our approach evaluated the performance of four independent prediction methods, including text mining of primary publications, inference from homologous protein structures, and two computational methods (PISA and EPPIC). Aggregating predictions to give consensus results outperformed all four of the independent prediction methods, yielding 83% correct, 9% wrong, and 8% inconclusive predictions, when tested with a well-curated benchmark dataset. We have developed a freely-available web-based tool to make this approach accessible to researchers and PDB Biocurators (http://quatstruct.rcsb.org/).
Introduction
The Protein Data Bank (PDB, pdb.org) [1] provides detailed information about the threedimensional (3D) structures of biological macromolecules, including proteins and nucleic acids. The PDB was established in 1971 with only 7 X-ray crystal structures of proteins and now contains more than 125,000 structures (as of January 2017). Today, the PDB archive is managed by the international Worldwide Protein Data Bank (wwPDB, wwpdb.org) partnership [2], which includes the RCSB Protein Data Bank (RCSB PDB, rcsb.org) [1], the Protein Data Bank in Europe (PDBe, pdbe.org), Protein Data Bank Japan (PDBj, pdbj.org), and Bio-MagResBank (BMRB, bmrb.org). The majority (~90%) PDB structures were determined by Xray crystallography. This experimental method yields 3D atomic level structures of the socalled asymmetric unit (Fig 1A), which is the repeating unit that makes up the crystal (Fig 1B). Knowledge of the 3D structure of the asymmetric unit and intermolecular interactions among asymmetric units does not provide sufficient information to reveal conclusively the oligomeric structures of protein assemblies, because is often not possible to distinguish biologically relevant intermolecular contacts from contacts that merely stabilize the crystal lattice.
Many proteins form structurally well-characterized thermodynamically stable multimeric complexes, which are important for biological function [e.g., hemoglobin occurs in nature as a heterotetramer (A2B2) with a cyclic (C2) symmetry and dihedral (D2) pseudo-symmetry] [3]. Experimental methods, such as size exclusion chromatography or analytical ultracentrifugation are sometimes required to ascertain the correct oligomerization state for a protein structure determined by X-ray crystallography. Alternatively, correct oligomeric state information may be inferred by comparison with better characterized homologous proteins or be provided by the PDB Depositor as metadata. It can also be predicted using computational methods, such as PISA (Proteins, Interfaces, Structures, and Assemblies) [4] or EPPIC (Evolutionary Protein-Protein Interface Classifier) [5]. Since the PDB was established in 1971, oligomeric state information has been obtained from Depositors or predicted by PQS [6] and more recently with PISA. Although experimental evidence for the oligomeric state was not a mandatory data item in legacy PDB deposition systems, collection of experimental evidence has been improved in the new wwPDB OneDep global deposition, biocuration, and validation system [7].
Quaternary structures of protein homomers and heteromers can be characterized by two main descriptors: stoichiometry and symmetry. Stoichiometry describes the composition of the assembly in terms of subunit number and composition. There are several widely used methods for determining the stoichiometry of protein complexes, including size exclusion chromatography [8], analytical ultracentrifugation [9], and gel-electrophoresis [10,11]. Protein assembly stoichiometry is described using a composition formula. Typically, an uppercase letter, such as A, B, C, etc., represents each type of different protein subunit in alphabetical order. (N.B.: These letters are not the same as the chain identifiers found in PDB archival entries.) The number of equivalent subunits is added as a coefficient next to each letter. For example the stoichiometry of the two-component human hemoglobin heterotetramer is represented as A2B2 (a dimer of heterodimers composed to two distinct polypeptide chains). Symmetry is another important feature of protein tertiary and quaternary structure [12] and plays a key role in understanding protein evolution and structure/function relationships [3], [12], [13], [14], [15], [16]. At the quaternary structure level, we characterize symmetry by the point group, a set of symmetry elements, whose symmetry axes go through a single point [12]. Most oligomeric protein structures (either homomeric or heteromeric) are symmetric macromolecules, which is probably a simple consequence of how subunits associate in solution without aggregating indefinitely [17] and can be classified using closed symmetry groups [18]. They are typically described as cyclic (i.e., C2, C3, C4, . . .), dihedral (i.e., D2, D3, D4, . . .), or cubic (tetrahedral, octahedral, icosahedral). Dihedral and cyclic symmetries are geometrically related: a structure with Dn symmetry can be constructed from n dimers with C2 or from two n-mers with Cn symmetry [19]. Additionally, helical symmetry is also a common open-symmetry encountered in protein structures. In a few instances, stoichiometry and symmetry are insufficient to uniquely characterize the quaternary structure of a protein assembly, where subunits form via different interfaces, while maintaining the same stoichiometry and symmetry. PDB structure 1JYE [20] is an example, where two assemblies have A2 stoichiometry and C2 symmetry, but only one of them is biologically relevant. Our investigation does not cover these special cases, however, detailed results for alternative interfaces are available from EPPIC.
The PDB archive grows by more than 11,000 structures annually. However, because of incomplete data and errors made during data entry, the oligomeric state annotations provided by the PDB are not always correct and reliable [21]. In spite of great efforts to improve the quality of the PDB archive, it has been reported that there are a significant number of PDB entries with incorrect quaternary structure annotations (Fig 2). Levy put the the error rate at 14% [22], while more recently Baskaran et al. [23] reported a lower bound for the error rate of~7%. Development of methods for accurate detection of incorrect annotations and assignment of most probable oligomeric state is, therefore, a matter of some urgency. This study has two main objectives: (i) to enable identification of incorrect protein quaternary structure annotations in the PDB archive, and (ii) to enable assignment the most probable quaternary structure for such cases. To accomplish these goals, we evaluated four different methods for assessing quaternary structure annotations in the PDB. First, we took an evolutionarily approach by clustering proteins related by amino acid sequence and attributed to each member of a given cluster the oligomerization state found to be most prevalent with the cluster. Second, we took a text mining approach by searching through the primary citations of indivdiual PDB entries and extracting information about oligomeric state and experimental evidence thereof. Third and fourth, we took two independent computational approaches using PISA and EPPIC, respectively, to predict oligomeric states. We aggregated results from these methods to generate a consensus prediction for the most probable oligomeric state for each protein structure in the PDB. We tested this combined approach using a well-curated benchmark dataset. During the course of this effort, we developed an efficient approach to evaluate oligomeric states of protein structures in the PDB, which we have made freely available to both PDB Biocurators and researcher as a web-based tool.
Sequence clustering
Since many protein chains in the PDB are similar at the level of sequence, we use this information to cluster polypeptide chains on the basis of amino acid sequence identity and assign a representative oligomeric state for each cluster based on a consistency score. For this purpose, we first constructed sequence clusters at various identity thresholds, including 95%, 90%, 70% and 40%. Clusters were calculated using the BLASTClust algorithm [27], which detects pairwise matches with the blastp algorithm [28] and then places each sequence in a certain cluster if the sequence matches that of at least one cluster member (Fig 4). A cluster is defined as a set of protein chains that are at least k% sequence identical to each other over 90% of the same length, and is associated with two discrete random variables: • S t represents stoichiometry and consists of a list (a,b,c,. . .,m) giving the number of copies of each unique molecule (e.g., A: monomer, A2: homodimer, A2B2: heterotetramer).
For a certain sequence identity threshold k, a consistency score for a given stoichiometry t and symmetry y can be estimated by the joint probability of these two events: After consistency score calculation, we applied a binary decision rule based on the majority probability to predict a representative quaternary structure for a certain cluster. The maximum consistency score in a cluster must be greater than 0.5 to satisfy the majority rule and to predict a representative quaternary structure, otherwise the result is deemed inconclusive.
For the consistency score to be statistically meaningful, there must be minimum number of members in a cluster. We used our benchmark dataset to determine the minimum number of cluster members for different sequence cluster identity thresholds, including 40%, 70%, 90% and 95%. For this purpose, we selected different number of cluster sizes from (n = 1,. . .,50) and predicted the most representative oligomeric state for each cluster using the consistency score described above. Then, we calculated the percentage of correct, incorrect, and inconclusive predictions for each minimum number of cluster members and for each sequence identity threshold. According to these results, each increment in the minimum number of cluster members reduces the number of incorrect results, but increases the number of inconclusive results. Hence, there is a trade-off between errors and cases where the predictions are not available due to an insufficient number of homologs in a cluster. In Fig 5, we plotted predictions for 70% sequence identity against cluster size. Plots for other sequence identities can be found in S1-S3 Figs. Lower sequence identities require more cluster members, while higher sequence identities require fewer. Based on our work with the benchmark dataset results, minimum cluster size should be 5 for 40% sequence identity threshold, yielding a 6% error rate and and 24% inconclusive rate. For 70%, 90%, and 95% sequence identities we found that the cluster size should be at least 3, yielding a 4% error rate and a 26% inconclusive rate. This approach was applied to both homomeric and heteromeric structures in the same way.
Text mining
Nearly 82% of the protein structures archived in the PDB have an associated primary article as of January 2017. Using CrossRef TDM (text and data mining) services (http://tdmsupport. crossref.org/), we were able to extract information regarding oligomeric state and supporting experimental evidence from 8,600 primary publications, describing nearly 32,000 PDB entries.
First, we split the full-text article into sentences. Then, we identified sentences containing oligomeric state information using a keyword list (monomer, dimer, trimer, etc., see S1 Table for the full keyword list). Some of these sentences proved misleading or irrelevant. For example, some sentences describe the asymmetric unit not quaternary structure, and some sentences refer to protein structures other than the one of interest. We, therefore, used a machinelearning approach to eliminate non-relevant sentences by classifying sentences as quaternary structure relevant (positive) or irrelevant (negative). To train and test our machine learning algorithms for the text mining approach, we created a dataset using PDB primary papers. For this task, first, we extracted sentences from the papers using the keyword list in the S1 Table. Then, a list with experimental evidence keywords (S2 Table) was used to split these sentences into two classes. The sentences that contain experimental evidence, were used to create the positive dataset of 5500 positive sentences and the remaining sentences were used to create the negative dataset of 5500 negative sentences. Traditional machine learning algorithms require numerical inputs. Therefore, each sentence is tokenized into words. Then, we converted each word to numerical values using the term frequency-inverse document frequency (tf-idf) method [29] to create a numerical data matrix. This method reflects how important a word is to a document in a corpus using the following formula: were tf(n,s) represents the frequency of word n in sentence s, df(n) represents the number of sentences containing word n, and N is the total number of sentences. To improve the effectiveness of the tf-idf method, all words in each sentence were converted to lower case, with extra spaces and internal punctuation marks removed. Finally, we calculated the tf-idf score for each word in each sentence and created a data matrix for the training procedure (wherein each row represented a single sentence and each column represented a unique word). To avoid the high-dimensional data matrix, we mapped the each column (i.e. features) to a hash-table by using a hash function. In this study, we used the murmurhash3 hash function, proposed by Weinberger et al. [30]. After applying the hashing function, we used two machine-learning algorithms, support vector machines (SVM) [31] and boosted logistic regression (BLR) [32][33][34][35], to classify each sentence in a paper as positive and negative. Next, the dataset was split as 80% training and 20% test set. A grid search was used with 10-fold cross-validation to select optimal parameters in the training set. Two parameters optimized for the SVM algorithm were sigma = 0.013 and cost = 4, and the optimal number of boosting iterations found as 101 for BLR algorithm. Identical parameters were used for the test set to confirm that both datasets were on the same scale and homoscedastic relative to each other. Finally, we tested model performance on the test set (Table 1). SVM performed better than BLR in terms of accuracy, kappa, area under the ROC curve (AUC), sensitivity, negative predictive value, F1 score, and Matthews correlation coefficient. Conversely, BLR showed better specificity and positive predictive value results, suggesting that it predicts positive sentences slightly better than SVM. A similar approach was used to search sentences for experimental evidence of oligomeric state. An experimental evidence keyword list was used for this task (see S2 Table for the full list). After oligomeric state keyword filtering, a second filtering was applied based on these keywords. Finally, after two filtering procedures, the stoichiometry prediction for each PDB entry was made based on the majority probability of remaining oligomeric state keywords. The text mining result was deeemed inconclusive for a particular PDB entry, if machine learning algorithms failed to detect any quaternary structure related sentences and supporting experimental evidence. This approach is only used to predict stoichiometry of a protein structure. Since symmetry information is difficult to retrieve and leads to ambiguous results, we did not provide symmetry prediction using the text mining approach. A general workflow of our text mining approach can be found in Fig 6.
PISA prediction
Following successful crystallographic structure determination efforts are made to identify biologically relevant intermolecular interactions within the crystal [36], and distinguish them from intermolcular contacts that simply stabilize the crystal lattice. The PISA program, developed by Krissinel and Henrick [4], uses a quantitative approach to address this problem [36]. General workflow of the text mining approach. After a publication upload or download, the term frequency-inverse document frequency (tf-idf) method is used to create a numerical data matrix from words, a hash table is created using the hash function, extracted sentences are classified using machine learning alorithms, remaining sentences are searched for experimental evidence and oligomeric state information is determined by using a majority rule. https://doi.org/10.1371/journal.pone.0197176.g006 Investigation of protein quaternary structure via stoichiometry and symmetry information The stability of an oligomeric structure is a function of free energy formation, solvation energy gain, interface area, hydrogen bonds, salt-bridges across the interface, and hydrophobic specificity [4]. PISA uses these properties to analyze protein structures and predict possible stable oligomeric states. Following successful evaluation (i.e., 90% accuracy [21]) using the Ponstingl et al. [24] benchmark data in 2007, PISA was deployed as a web server at the European Bioinformatics Institute (EBI) [36]. Soon thereafter, it was adopted as a quaternary structure validation and annotation tool for PDB archival depositions.
PISA can be accessed through a web service (http://www.ebi.ac.uk/) or as a standalone program (http://www.ccp4.ac.uk/pisa/) from Collaborative Computational Project No. 4 or CCP4 (http://www.ccp4.ac.uk/). The software provides broad information about assemblies, interfaces, and monomers, and gives information regarding possible oligomeric states, such as stoichiometry, solvent accessible surface area (ASA), buried surface area (BSA), and Gibbs free energy of dissociation score, which represents the free energy difference between the associated and dissociated states [37]. Possible oligomeric states yielding positive and negative Gibbs free energy of dissociation score are considered to be chemically stable and unstable quaternary structures, respectively [38]. Those with borderline values are deemed indeterminate.
In this study, we used the command line version of PISA to generate XML files for each possible oligomeric state, which include rotation/translation operators for each of the chains together with ASA, BSA, Gibbs free energy of dissociation score, entropy, internal energy, and macromolecular size. The PISA XML files were then parsed to extract rotation/translation operators and chains with which to build up the atomic coordinates of each possible oligomeric state. Finally, we used BioJava [39] to characterize each of the possible oligomeric states in terms of stoichiometry and symmetry (Fig 7).
EPPIC prediction
To distinguish intermolecular contacts stabilizing biologically relevant oligomeric states from simple crystal contacts, an evolutionary-based classifier (EPPIC; http://www.eppic-web.org) was developed by Duarte et al. [5]. This method uses a geometric measure, number of interfacial residues, and evolutionary features to classify interfaces as biological versus crystal. Unlike other evolutionary-based methods, EPPIC uses only close homologs with >60% sequence identity to ensure homologs share high degrees of quaternary structure similarity. This approach achieved~89% accuracy with the Ponstingl et al. [24] benchmark data. The new version of EPPIC (3.0.1) uses the pairwise interface classifications from version 2 to predict quaternary structure [40]. The crystal lattice is represented as a periodic graph from which the different valid assemblies (those with point group symmetry) can be enumerated. Based on the pairwise scores the software can then decide which of the viable assemblies is the most likely quaternary structure in solution. We obtained early access to the results from the authors. Output results for the whole PDB archive were provided in XML format, containing quaternary structure predictions for each entry in the PDB. We parsed these XML files to extract the respective stoichiometry and symmetry information for PDB entries.
Consensus result approach
Predictions from sequence clustering, text mining, PISA, and EPPIC, were combined and a majority vote rule was applied to create a 4-method consensus result. In the following cases, predictions from individual methods were excluded: • Sequence clustering: insufficient homologous proteins comprising the cluster (at least 3 structures for 70% sequence identity).
• Text mining: no full-text publication available or no clear information regarding quaternary structure therein.
• PISA: method does not produce any assembly result.
Moreover, in the following cases, predictions from individual methods were classified as inconclusive: • Sequence clustering: adequate number of homologous proteins comprising the cluster but the maximum consistency score in a cluster is less than 0.5.
• Text mining: machine learning algorithm could not detect any quaternary structure related sentences which supports the experimental evidence.
• If one of the measures, i.e. stoichiometry or symmetry, is correct and the other one is wrong.
Web-tool development
To make this approach accessible, we developed a user-friendly, easy-to-use web-based tool using R, JavaScript, jQuery, CSS, and HTML (Fig 8). The tool was predominantly constructed using R software [41]. Publications were downloaded either in PDF or XML format, and XML [42], Rcurl [43], tm [44], NLP [45], openNLP [46], and stringr [47] packages were used to Investigation of protein quaternary structure via stoichiometry and symmetry information convert PDF files to plain text files, to parse XML files, to extract sentences, and to split words. The FeatureHashing package [48] was used to map features to the hash table [30]. Machine learning algorithms were trained and tested using the caret package [49]. Data tables were built using the DT package [50], and the shiny package [51] was used to create an interactive web-based application.
The new web tool offers a wide range of methods for evaluating oligomeric states in the PDB, including i. Determination of a representative oligomeric state for a given sequence identity threshold using the sequence clustering approach with consistency scoring.
ii. Generate a PISA oligomeric state prediction for a given structure, rebuild the quaternary structure, and assign stoichiometry and symmetry using BioJava.
iii. Generate a EPPIC oligomeric state prediction for a given structure, including stoichiometry and symmetry. Predictions for multiple interfaces are included in the EPPIC predictions.
iv. Extract oligomeric state information with experimental evidence from any publication describing a crystal structure of a protein.
The tool has a simple user interface, requiring only four character PDB IDs as input and a single mouse click to launch the calculation. Note, in some instances, the full articles used for text mining in the benchmark may not be available in the online tool, which relies on open access PMC articles and PubMed abstracts. The output includes two parts: oligomeric state and symmetry. In the oligomeric state table, it states that 1Z77 (first column: PDB ID) has one oligomeric state prediction (second column: BA Number), and it is annotated as a monomeric structure in the PDB (third column: PDB). However, according to SC (fourth column: Sequence clustering), PISA (fifth column), EPPIC (sixth column), and TM (seventh column: Text Mining), 1Z77 is a dimeric protein. Therefore, our consensus result (eighth column) states that 1Z77 is a dimer. In the symmetry table one quaternary structure prediction is shown (second column: BA Number), having C1 symmetry (third column: PDB). However, according to SC (fourth columns: Sequence cluster), PISA (fifth column) and EPPIC (sixth column), 1Z77 has a C2 symmetry. Therefore our consensus result (seventh column) states that 1Z77 has a C2 symmetry. Red denotes divergence between current PDB annotation of oligomeric state and results provided by each of the four evaluation methods. https://doi.org/10.1371/journal.pone.0197176.g008 Investigation of protein quaternary structure via stoichiometry and symmetry information
Benchmark dataset results
We applied sequence clustering, text mining, PISA, and EPPIC to our 543 structure benchmark dataset to test the performance of each approach. For the sequence clustering approach we evaluated the trade-off between the error rate and the number of non-applicable cases as a function of sequence identity. While the cluster size at 40% sequence identity (S1 Fig) is larger than at higher levels of sequence identity, quaternary structure may not be conserved at this level of sequence identity. For 70%, 90%, and 95% sequence identity, the trade-off curves are very similar for small cluster sizes (Fig 5, S2 Fig, S3 Fig) with an optimum cluster size around 3. We chose the 70% sequence identity threshold for sequence clustering because quaternary structure is generally preserved at this level of sequence identity. In comparison, EPPIC uses a 60% sequence identity cutoff. For the text mining approach, we selected the SVM classifier over BLR since it performed overall better according to the metrics in Table 1.
Then, we aggregated all available results to arrive at a consensus result. S3 Table lists oligomeric state, stoichiometry, symmetry, and the individual predictions. Text mining predictions are based on stoichiometry data, whereas sequence clustering, PISA, and EPPIC use both stoichiometry and symmetry data. S4 Table summarizes the results for individual methods and the consensus predictions. Accuracy rates for individual methods ranged between 46% and 81%. We achieved 83% accuracy rate using the consensus approach (Fig 9). Investigation of protein quaternary structure via stoichiometry and symmetry information The number of correct predictions for method-agreements are summarized in Fig 10A. The 4-method consensus correctly predicted the oligomeric states of 111 structures. The combination of SC, PISA and EPPIC (i.e. 3-method consensus) correctly predicted 140 cases. In 2-method consensus, PISA and EPPIC correctly predicted 45 cases, while SC and EPPIC correctly predicted 32 cases, and SC and PISA predicted 19 cases correctly. When we checked the number of wrong predictions (Fig 10B), there is no single structure that is incorrectly predicted by all methods. The most incorrect results occurred between PISA and EPPIC methods with 30 cases. Three methods, TM, PISA, and EPPIC, and two methods, TM and EPPIC, incorrectly predicted 7 cases. Inconclusive results (Fig 10C) mostly occurred in single methods. PISA gave 41 inconclusive results, while 36 of the TM predictions were inconclusive. Finally, the results where predictions were not-applicable (Fig 10D) occured only in TM and SC.
TM gives high inconclusive results and it can not predict stoichiometry and symmetry. Therefore, we also provide a 3-method consensus based on SC, PISA and EPPIC results. After exclusion of TM, the accuracy rate decreased to 81%, while incorrect rate increased to 11%. Thus, even though its highly inconclusive nature, TM provides useful insights regarding correct quaternary structure information.
To apply our consensus approach to a larger data set and to provide a PDB-wide result, we applied the 3-method consensus to almost 77,000 protein structures that are solved by X-ray crystallography (S5 Table). We excluded TM since it is only applicable to the subset of stuctures where the full text of the primary citation is available. According to the PDB-wide results, we detected 16.1% of the structures as possible outliers.
Discussion
The oligomeric state (or states) of a protein represents the essential biological unit(s) that carries out a biological function in a living organism [4]. Therefore, it is crucial to determine with high reliability the biologically-relevant oligomerization state(s) of a macromolecule. This problem has been extensively studied and the scientific literature abounds with various methods and approaches. An extensive analysis of the problem can be found in Capitani et al. [21], wherein the authors reviewed main concepts of different approaches, including thermodynamic estimation of interface stability, evolutionary approaches and interface co-occurrence across different crystal forms.
The difficulty of the protein interface classification problem has led to incomplete or ambiguous experimental data, which in turn has resulted in incorrect annotations of the oligomerization states of macromolecules in the PDB. Another source of error are the simple mistakes during the deposition process. For example, sometimes authors simply identify the quaternary structure of the asymmetric unit. To the best of our knowledge, there are only two studies in the literature, which have investigated quaternary structure annotation errors in the PDB [22,23].
In this study, we focused on two main issues: (i) detection of the incorrectly annotated biological macromolecules in the PDB and (ii) assignment of the most probable stoichiometry and symmetry annotations for those incorrect structures. We used two of the most popular and widely used methods to evaluate the protein structures through their stoichiometry and symmetry, PISA and EPPIC. To investigate author annotations, we provided a text mining approach for searching and extracting correct quaternary structure information from primary papers of the macromolecules. In addition, we utilized homology, by looking at consensus within sequence clusters, for detection of incorrectly annotated protein structures in the PDB. Finally, we obtained the results from each method, including oligomeric state and symmetry information, and calculated a consensus result for each entry in the PDB. The consensus result approach, which uses the method agreements, has improved the prediction reliability. The number of correct predictions is increased, while both the number of incorrect predictions and the number of inconclusive predictions are decreased. The best combination of the methods is found as the three methods combination of SC, PISA and EPPIC, which correctly predicted 140 cases together. On the other hand, the 4-method consensus correctly predicted 111 cases. The reason for the lower number of correct predictions in the 4-method consensus is the high rate of inconclusive and not-applicable results in the text mining approach, which is almost 40%.
The most important issue in the text mining method is the difficulty in the full-text publication access. Even though, 42% of the primary articles in the PDB are open-access publications, it is hard to get these publications using the PMC database or publishers' websites for text mining, because these services have their own restrictions based on their respective policies. PMC does provide a service for bulk downloading of the publications for an open-access subset. However, this subset only covers 1/26 of all articles in PubMed, and that makes it impractical for text mining purposes. Some publishers, such as Elsevier and Wiley, provide their own text mining services through a specific API (application program interface). However, these services are not stable and not working all the time and more importantly they are not allowing bulk access to many papers. In addition, they do not provide all the publications through these services; instead they only provide a small subset of them. We found CrossRef TDM services as the most useful service for text mining purposes. They provide full-text links from various publishers in a user friendly and easy-to-access way. However, since some publishers only share a small part of their paper repositories, we are only able to reach 30% of the primary articles of the PDB entries.
In addition to the 4-method consensus, we also provided a 3-method consensus based on SC, PISA and EPPIC results by excluding TM. The 3-method consensus has the advantage that it can be applied to most structures in the PDB. On the other hand, TM results are useful to guide the user/biocurator to the specific papers/sentences that contain statements about the experimental evidence, as well as a resource to verify the author annotations. The performance of the methods reported here is based on a well-curated dataset. Predictions may be worse for cases outside the testing and training sets.
Conclusion
Determination of a 3D macromolecular structure is crucial to understand the fundamental mechanisms of biological processes, such as enzymatic reactions, ligand binding, or signalling. It is also important to reveal the underlying mechanisms of diseases, such as genetic variations. Furthermore, the 3D structures of macromolecules are vital for drug design and development studies, especially in structure-based drug design. Because of these reasons, the correct quaternary structure has a critical importance.
In this study, we developed a consensus approach by aggregating predictions from three and four different methods in order to detect incorrectly annotated protein structures in the PDB through their stoichiometry and symmetry, and assign the most likely quaternary structures annotations for the possibly incorrect annotations. For this task, we first benefited from homology to cluster similar PDB entries based on a certain sequence identity threshold and to predict a representative protein structure for the cluster through a consistency score calculation. Second, we searched through the primary articles associated with the PDB entries to validate the author deposition by extracting oligomeric state and experimental evidence information using a text mining approach. Moreover, we used the PISA and EPPIC to predict the stoichiometry and symmetry of structures. Then, we combined predictions from different approaches to achieve a consensus result and to predict the most probable annotations for the possibly incorrect structures. To test the performance of our consensus approach, we created a benchmark dataset using Ponstingl et al. [24], Bahadur et al. [25], Bahadur et al. [26] and Duarte et al. [5] datasets. Our consensus approach outperformed single methods and achieved 83% correct, 9% incorrect, and 8% inconclusive predictions, respectively. Therefore, our method provides more reliable evaluation than any single approach. Rare exceptions are cases where both stoichiometry and symmetry are consistent, but subunits form alternative interfaces. In those cases, the EPPIC results list predictions for alternative assemblies. Finally, we developed a web-based tool in order to make this approach usable for researchers in the field and PDB Biocuators.
Supporting information S1 Fig. Cluster size vs. correct, incorrect, and N/A (not available cases) Fig. Cluster size vs. correct, incorrect, and N/A (not available cases) | 2018-06-18T01:35:04.348Z | 2018-06-04T00:00:00.000 | {
"year": 2018,
"sha1": "0d65ed71307ec4922c5d7fc26e1e784484aa8acd",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1371/journal.pone.0197176",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "0d65ed71307ec4922c5d7fc26e1e784484aa8acd",
"s2fieldsofstudy": [
"Chemistry",
"Biology"
],
"extfieldsofstudy": [
"Medicine",
"Computer Science"
]
} |
93002878 | pes2o/s2orc | v3-fos-license | The toxicity of coated silver nanoparticles to Daphnia carinata and trophic transfer from alga Raphidocelis subcapitata
Nanoparticles (NPs) are causing threats to the environment. Silver NPs (AgNPs) are increasingly used in commercial products and may end up in freshwater ecosystems. The freshwater organisms are vulnerable due to water-borne and dietary exposure to AgNPs. Surface properties play an important role in the fate and behavior of AgNPs in the aquatic environment and their effects on organisms. However, effects of surface properties of AgNPs on organisms are poorly understood. In this study, we explored the effects of AgNPs coated with three different ligands; Tyrosine (T-AgNP), Epigallocatechin gallate (E-AgNP) and Curcumin (C-AgNP) in relation to the toxicity to a key aquatic organism; Daphnia carinata. The study focused on how coatings determine fate of NPs in the medium, mortality, feeding behaviour, bioaccumulation and trophic transfer from the freshwater alga, Raphidocelis subcapitata to daphnids. NP stability tests indicated that T-AgNPs were least stable in the ASTM daphnia medium while C-AgNPs were most stable. 48 h EC50 values of AgNPs to D. carinata were in the order of E-AgNP (19.37 μg L-1) > C-AgNP (21.37 μg L-1) > T-AgNP (49.74 μg L-1) while the 48 h EC50 value of Ag+ ions was 1.21 μg L-1. AgNP contaminated algae significantly decreased the feeding rates of daphnids. However, no significant differences were observed in feeding rates between algae contaminated with differently coated AgNPs. Trophic transfer studies showed that AgNPs were transferred from algae to daphnids. The bioacumulation of AgNPs in algae and the diet-borne bioaccumulation of AgNPs in daphnids varied for differently coated AgNPs. Bioaccumulation of C-AgNPs in algae was 1.5 time higher than T-AgNPs. However, the accumulation of T-AgNPs in daphnids via trophic transfer was 2.6 times higher than T-AgNPs. The knowledge generated from this study enhances the understanding of surface property dependent toxicity, bioaccumulation and trophic transfer of AgNPs in aquatic environments.
observed transfer of Carboxyl quantum dots (QD) from QD-exposed alga Raphidocelis subcapitata to Ceriodaphnia dubia. Elsewhere, trophic transfer of ENPs were demonstrated from algae to mussels [46], daphnids to zebrafish [47] and biofilms to snails [48]. Exposure to AgNPs and Ag + ions contaminated algae also cause changes in feeding behaviour in daphnids. McTeer, Dean [42] also reported a significant reduction in feeding when daphnids were fed with AgNP and Ag + ion contaminated algae compared to untreated algae. Zhu, Chang [49] observed dose dependent reductions in ingestion and filtration rate when D. magna was exposed to TiO 2 NPs.
In this study, we studied the toxicity of AgNPs coated with different organic ligands; Tyrosine (T-AgNP), Curcurmin (C-AgNP) and Epigallocatechin gallate (E-AgNP) to the freshwater filter-feeding cladoceran, Daphnia carinata. In addition, the effects of associated Ag with alga on daphnid feeding behaviour and trophic transfer from the alga diet to daphnids were investigated. Tyrosine, Curcurmin and Epigallocatechin gallate have different number of phenol structures and classified as mono-phenol, bi-phenol and poly-phenols respectively. Since they are organic compounds, their usage is considered as a green and ecofriendly approach to produce NPs while they are biocompatible which is a useful characteristic in therapeutic applications [50]. These coatings are used to produce NPs for different applications, mainly being found in medical applications [50][51][52][53][54]. Though there are studies on the effects of AgNPs with most commonly used coatings, studies on organic coatings used in this study are lacking. Daphnia sp. is one of the most sensitive species used in aquatic toxicological studies and is recommended as a model test organism by international agencies. The most common species used for studies is D. magna which is considered as an invasive species in some parts of the world. Therefore, it is required to improve test species and protocols to better reflect species sensitivity in different ecosystems [55], for environmental risk assessment and to protect native species. Also, daphnids represent the bottom level of the freshwater food chain and any qualitative or quantitative effect to the population will affect higher organisms [16]. To the best of our knowledge, this is the first study where the effects of differently coated NPs were studied against D. carinata while other studies so far have used D. magna. Also, the coating materials are quite novel and have not being used in similar kind of studies.
Preparation of NPs
L-tyrosine, epigallocatechin gallate (EGCG) and curcumin (Fig A1, A2 and A3 in S1 File) were obtained from Sigma-Aldrich. NP synthesis was performed in-house as previously described by Selvakannan, Swami [53]. Briefly, 10 mL of 1 mM Ag 2 SO 4 solution was mixed with 10 mL of 1 mM aqueous solution of Tyrosine, EGCG and Curcumin and diluted to 100 mL with MilliQ water. To each solution, 1 mL of 100 mM KOH solution was added, and the mixture was allowed to boil until the colour of the solution turned yellow which indicates NP formation. The AgNP solutions were allowed to age for 1 day and then concentrated by rotary evaporation. The solutions were then dialyzed for 48 h in a dialysis tube (MWCO: 3 kDa) which was submerged in copious amounts of MilliQ water with stirring to remove any uncoordinated silver ions, excess KOH and unbound coating materials. Water was replaced twice with fresh MilliQ water after 6 and 24 h. The dialyzed AgNP solutions were stored in the dark. Each type of NP (0.1 mL) was acid digested with ultra-pure grade 70% HNO 3 (Thermo Fisher Scientific, NSW, Australia) on a heating block at 105 o C for 12 h. The digested samples were then diluted with MilliQ water and the silver (Ag + ions) concentrations were measured by inductively coupled plasma mass spectrometry (ICP-MS) (7700x, Agilent Technologies).
Characterization of NPs
The zeta potential and hydrodynamic diameter (HDD) of the AgNPs were measured using a folded capillary cell and glass cuvette, respectively, on a Zetasizer (Dynamic light scattering; Malvern Zetasizer Nano series, NanoZS). The particles were drop cast on carbon copper grids and images were taken using a transmission electron microscope (TEM) operated at an accelerating voltage of 100 kV (TEM, JEOL 1010) equipped with a Gatan imaging system. The mean core size of NPs was determined from TEM images using ImageJ software. The surface plasmon resonance (SPR) was recorded as absorbance spectra using a UV-visible spectrophotometer (Varian Cary 50) operated at a resolution of 2 nm, from 200 to 800 nm, using a quartz cuvette with a path length of 1 cm. Concentrations of stock solutions were measured with ICP-MS (7700x, Agilent Technologies) after acid digestion with concentrated HNO 3 .
Culturing organisms
R. subcapitata pure cultures were maintained in the lab and the algal cells were inoculated in the MLA medium to prepare the algae stock cultures. MLA medium was prepared as described by Bolch and Blackburn [56] with slight modifications. Briefly, concentrate nutrient stock solutions were prepared by filter sterilization and the medium was prepared in MilliQ water by autoclaving for 15 min at 121˚C. The algae stock cultures were maintained axenically as per the OECD guidelines [57]. Cultures were aerated and incubated in a light-temperature controlled chamber at 23 ± 1˚C under continuous illumination (6000 lux). White fluorescent tubes were used as the light source while the light intensity in the test setup was measured using a LI-COR light meter (model LI-189). The pH was maintained at 7.5 ± 0.5. Algal cells were counted periodically with an automated cell counter (TC20, Bio-Rad Laboratories, Hercules, CA) to make sure the culture is at exponential growth stage when it is treated with NPs for the daphnid feeding and bioaccumulation test. A widespread Daphnia species in Australia, D. carinata was used to assess the acute toxicity, feeding behaviour and bioaccumulation experiments. ASTM standard medium [58] was used for stock culturing and in all experiments. Daphnid stock culture was maintained at 20 ± 1˚C with 16:8 h light and dark photoperiod to obtain neonatal cladocerans. They were fed with R. subcapitata (5 × 10 5 cells mL -1 ), a highly suitable nutrient source for daphnids. The medium was aerated to make it saturated with oxygen before addition of daphnids. The medium was renewed three times a week, and the pH was maintained at 7.5 ± 0.2 throughout and cultures were maintained in glass beakers which were loosely covered to prevent any contamination and evaporation.
AgNP temporal stability and dissolution in the test medium
The stability of AgNPs in the ASTM medium and MilliQ water was investigated as previously described by Tejamaya, Römer [59] with some modifications. Nanoparticle solutions of 5,000 μg L -1 were incubated in glass vials for 24 h at similar environmental conditions to the daphnid acute toxicity test. The SPR, HDD and zeta potential of the suspensions were investigated, and pH was monitored. The release of Ag + ions from NPs was investigated in the ASTM medium and MilliQ water over 24 h as previously described by Xia, Kovochich [60] with some modifications. Briefly, 1 mL from each NP suspension was extracted into Eppendorf tubes and centrifuged at 21,000 rpm for 15 min (Sigma 3-KL centrifuge). Supernatant (0.75 mL) from carefully removed tubes was transferred to 15 mL tubes, acidified with HNO 3 and diluted with MilliQ water. The Ag + ion concentrations were measured by ICP-MS (7700x, Agilent Technologies). The samples need to be in ionic form prior to entering the mass analyser in order to be detected. The coated NPs were not detectable as validated by independent experiments by using undigested NPs. Therefore, the influence of NPs which may be present in the supernatant to the readings was insignificant.
Acute toxicity bioassay with D. carinata
AgNP test solutions of relevant nominal concentrations for the acute test were prepared based on ICP-MS results just before the test began by sonicating and dispersing relevant volumes of stock solutions in daphnid culture medium. Ionic silver stock solution was prepared by dissolving Ag 2 SO 4 in culture medium followed by ICP-MS analysis and required concentrations for acute tests were prepared by dissolving relevant volumes from the stock solution in daphnid culture medium. Controls contained only the culture medium. Semi-static renewal acute toxicity tests were performed according to the OECD standard procedure [61]. Less than 24 h old third brood progeny was used for experiments. Daphnid neonates were exposed to seven different concentrations of each test substance for 48 h. Test concentrations were chosen based on results obtained from range finding tests. Concentrations employed for the tests were in the range of 10.0-40.0 μg L -1 for E-AgNP, 30.0-90.0 μg L -1 for T-AgNP, 10.0-35.0 μg L -1 for C-AgNP and 0.6-1.8 μg L -1 for Ag + ions. For each concentration, 5 neonates (age: < 24 h) were placed in a 35 mL glass vial containing 15 mL of test solution. The test solution was renewed after 24 h and all experiments were conducted in quadruplicate (n = 4). Daphnids were not fed during the 48 h time period. The tests were conducted at 20 ± 1˚C with 16:8 h light and dark photoperiod, pH was maintained at 7.5 ± 0.2 while the dissolved oxygen concentration in all the test solutions exceeded 3 mg L −1 . Immobilization was recorded after 24 and 48 h while they were considered immobile if they couldn't move within 15 s of gentle agitation of the test container. The toxicity tests were considered valid if the mortality was not >10% in the control.
Feeding analysis and AgNP trophic transfer
Algae feeding experiment was conducted as described by McTeer, Dean [42] and Grintzalis, Dai [62] with some modifications. Cell density of the algae culture which was in exponential growth stage was adjusted to 5 × 10 4 cells mL -1 . Algae cells were propagated in a 1 L Erlenmeyer flask containing 500 mL algae culture in the presence of particles (50 μg L -1 ) or absence (blank control) in triplicate. The flasks were incubated as per the OECD guideline [57] on orbital shakers (OM6, RATEK, Aus) at 100 rpm under the same environmental conditions used for algae culturing. Algal cell counts of each flask were taken with the automated cell counter and the specific algal growth rate of all treatments were calculated for each day. Algae cells were fed to daphnids after 6 days when the cells had reached stationary growth phase. Since the cells were no longer dividing, number of cells consumed could accurately be measured. A volume of 350 mL from each culture was centrifuged (3000 g, 5 min, 20˚C) and washed three times with MilliQ water to remove any loosely adsorbed contaminants from algae. The algae were then resuspended in ASTM medium in a 50 mL Falcon tube as it contained approximately 5 × 10 6 cells mL -1 and stored at 4˚C. Feeding experiments were carried out under similar environmental conditions used for the acute toxicity test in 100 mL glass beakers in triplicate with 50 mL ASTM medium in each beaker. Daphnids (15, age: < 24 h) were placed in each beaker and fed with AgNP treated and untreated algae for 5 days. Daphnids were transferred to fresh media after every 24 h and the starting and final algal cell numbers were determined with the automated cell counter. The solutions were stirred on a magnetic stirrer for 20 s before samples were taken for cell counting. The 5 day feeding phase was followed by a 3 day depuration phase in fresh medium where daphnids were fed with unexposed algae. Then, daphnids were oven dried at 60˚C for 48 h and acid digested on a block heater for 6 h at 100˚C with 70% HNO 3 and 30% H 2 O 2 . Digested samples were diluted with MilliQ water and analysed using ICP-MS. A volume of 50 mL from each remaining culture was harvested by centrifugation (3000 g, 5 min) and washed three times with MilliQ water. The resulting algal pellets were oven dried at 60˚C for 48 h and acid digested separately as described above. The digested samples (0.5 mL) were diluted with MilliQ water and the silver (Ag + ions) concentrations were measured using ICP-MS (7700X, Agilent Technologies).
Data analysis
The 24 and 48 h EC 50 values and their associated 95% confidence intervals (95% CI) were calculated using the TOXRat software (TOXRat solutions GmbH, version 3.0). The statistical method used was the probit analysis using linear maximum likelihood regression. The daily feeding rate was determined by dividing the number of algal cells consumed by the number of daphnids that were alive in each beaker. Trophic transfer was quantitatively determined by calculating the transfer of Ag to a daphnid from 10 5 algal cells consumed. For each dataset, mean and SD are presented and data were considered statistically significantly different at P < 0.05. All data were normally distributed according to Shapiro-Wilk test. Differences between treatments were analysed using one-way analysis of variance (ANOVA). When significant differences were detected at a 95% level of confidence, the Tukey's multiple comparisons were applied. ANOVA was performed using Sigmaplot statistics version 13. Differences of treatments over time for repeated measures were determined by 2-factor ANOVA followed by Holm-Sidak method.
Particle characterization
Strong, well defined absorptions peaks at 406 (T-AgNP), 400 (E-AgNP) and 414 nm (C-AgNP) in the SPR curves ( Fig 1A) indicate the presence of AgNPs [53, 63-66]; these peaks were absent in controls which included media, pure coating material and coating material plus KOH. The presence of NPs was further confirmed by TEM analysis (Fig 1B, 1C and 1D) which revealed spherical shaped NPs which were reasonably uniform in size. Intensity weighted mean HDD and zeta potential values in MilliQ water and core sizes of AgNPs are shown in Table 1. Zeta potential values below -30 mV suggest that organic molecules formed strong coatings on the AgNPs [50]. The secondary peaks of SPR spectra at 272 (T-AgNP), 270 (E-AgNP) and 256 nm (C-AgNP) are due to the Ag-bound tyrosine [53,67], EGCG [68] and curcumin molecules [69] respectively. The concentrations of T-AgNP, E-AgNP, C-AgNP and Ag 2 SO 4 stock solutions as measured by ICP-MS were 45.59, 48.84, 42.66 and 10.52 mgL -1 respectively.
Stability of coated AgNPs in ASTM medium
Suspensions of AgNPs in MilliQ water were initially yellowish in colour with absorption peaks at 406, 400 and 414 nm for T-AgNPs, E-AgNPs and C-AgNPs respectively. Once the AgNPs were dispersed in ASTM medium, visual observation revealed that the T-AgNP and E-AgNP suspensions turned to yellowish brown within minutes, but the yellowish colour of C-AgNP suspension was stable even after 24 h. SPR analysis revealed a larger decrease in the absorption peaks of T-AgNPs (34 & 70%) and E-AgNPs (33 & 57%) after 5 min and after 24 h respectively in comparison to the initial relevant absorption peaks of AgNPs in the MilliQ water measured after 5 min (Fig 2A and 2B). In contrast, the decrease of the absorption peak was 4 & 5% for C-AgNPs after 5 min and 24 h (Fig 2C). In addition, the SPR bands of T-AgNPs and E-AgNPs significantly broadened which was not observed for C-AgNPs. However, the SPR bands of AgNPs didn't change considerably in the MilliQ water ( Fig B1, B2 and B3 in S1 File). In the ASTM medium, the percentage change of dissolution after 24 h in comparison to the dissolution after first 5 min was 0.06, 0.04 and 1.51% for T-AgNPs, E-AgNPs and C-AgNPs respectively while it was 0.04, 0.36 and 4.39%, respectively in MilliQ water. The mean HDD of T-AgNPs and E-AgNPs in the ASTM medium increased approximately by 2.7 and 4.7 times respectively after 5 min and by 9 and 8 times after 24 h. However, the mean HDD of C-AgNPs did not change considerably ( Table 2). In comparison, none of the AgNPs showed considerable change in HDD in the MilliQ water (Table A in S1 File). The percentage intensity distribution of particle size (weighted according to the scattering intensity of each particle fraction) in media revealed that approximately 63, 79 and 87% of aggregates of T-AgNPs, E-AgNPs and C-AgNPs remained at the sub-micron level after 24 h. In comparison to the zeta potential of NPs in the MilliQ water (Table A in S1 File), it sharply increased by 13.4, 23.9 and 24.3 mV of T-AgNPs, E-AgNPs and C-AgNPs respectively after 5 min but did not change considerably thereafter ( Table 2). The polydispersity index (PdI) of T-AgNPs and E-AgNPs slightly increased in both ASTM medium ( Table 2) and MilliQ water (Table A in S1 File). A broader absorbance peak, a large background signal and increased HDD indicate aggregation of NPs [59,70]. As per the results, T-AgNPs and E-AgNPs aggregated rapidly in ASTM Table 2. HDD, zeta potential and PdI of AgNPs measured after 5 min and 24 h. AgNPs were dispersed in the ASTM media at Ag concentration of 5,000 μg L -1 . Standard deviations (± SD) are from triplicates.
Substance
Medium HDD (nm) Zeta Potential (mV) PdI medium in comparison to C-AgNPs. Aggregation of NPs depends on particle concentration, pH, ionic strength, ionic composition, concentration and composition of natural organic matter, and other characteristics of the aqueous media [71,72]. The observed differences in aggregation among coated AgNPs may be due to different coating materials since other characteristics of AgNPs such as initial HDD, shape and zeta potential are not significantly different (p > 0.05) [73,74]. Also, all NPs showed no signs of aggregation in MilliQ water within the same time duration. The observed higher aggregation of AgNPs in ASTM medium in comparison to the MilliQ water could be attributed to the higher ionic strength of the medium. Decreased zeta potential values in ASTM medium with high ionic strength compared to MilliQ water with relatively low ionic strength are in accordance with the classical colloid theory [75,76]. A reduction of the thickness of the diffuse double layer with increased ionic strength allows for the attractive van der Waals interactions to dominate and increases the particle-particle interaction resulting in increased aggregation [77,78]. In general, the dissolution of AgNPs was higher in MilliQ water than in the ASTM medium. Köser, Engelke [79] and Levard, Mitra [80] reported that the percentage dissolution of AgNPs correlates to the Clion content (Cl/Ag ratio) in the medium which was contradictory to our results. This may be due to the precipitation of Ag ions by halides (Cl -) in the medium [81] which is not expected in the MilliQ water. Halides may reduce the exposure of organisms to any free ions released from NPs. Also, the size of the AgNPs is one of deciding factors for the release of Ag + ions [82]. Small size particles have high surface to volume ratios and therefore, more atoms on the surface come into contact with oxidants in comparison to larger particles [83]. Aggregation of NPs results in larger particles with increased HDD reducing the surface area of particles available to release free Ag ions [34,84]. Higher dissolution of AgNPs in MilliQ water than in ASTM medium may also be due to the less aggregation of AgNPs. The observed higher dissolution of C-AgNPs in both ASTM medium and MilliQ water than T-AgNPs and E-AgNPs is attributed to the less aggregation of C-AgNPs, while dissolution depends on several other factors [85]. Aggregation, dissolution and change in NP characteristics such as HDD as observed in this study may influence the bioavailability of NPs and hence, play an important role in determining toxicity [59, [86][87][88]. Also, culture medium impacted the behaviour and properties of AgNPs which may ultimately lead to various toxicological responses [89]. This study shows that the type of NP coating and medium significantly influence the degree of aggregation and the behaviour of AgNPs, which are required to consider in environmental risk assessment [90].
Acute toxicity of AgNPs to D. carinata
Evaluation of acute toxicity is crucial in environmental risk assessment of NPs in protecting the organisms and setting up water quality guidelines. According to the results in this study, the toxicity of AgNPs and Ag + ions correlate with the concentration while the toxicity of Ag + ions is significantly higher than AgNPs as per the 48 h EC 50 values (Table 3). Among 84,91]. The shape of all NPs is spherical, and the initial size and charge do not significantly differ (p > 0.05) which may not explain the observed differences in acute toxicity. Therefore, the difference in toxicity is presumed to be the effects of different coating materials which reaffirm that different types of coating of NPs do have different effects on toxicity of NPs. T-AgNPs were least stable in the ASTM medium exhibiting the highest aggregation and settling of particles. Thus, the reduced toxicity of T-AgNPs compared to the other two types could be due to the lower bioavailability of NPs [71,92]. However, the toxicity of E-AgNPs was similar to that of C-AgNPs which showed less aggregation and high dissolution compared with E-AgNPs. Therefore, the observed difference in toxicity could be a result of several factors which can't be exclusively explained from the results of this study. Several previous studies have assessed the acute toxicity of AgNPs to Daphnia sp. and the reported EC 50 values fall in the range of 0.26 to 236.3 μg L -1 for AgNPs and 0.16 to 12.9 μg L -1 for Ag + ions [19, 29, [93][94][95][96]. The toxicity of coated AgNPs in this study is comparatively less than the values reported by majority of studies for other coated AgNPs. The broad range of toxicity values among published studies can be explained by different test scenarios and particle characteristics [84,97]. D. magna has been the preferred species in many previous studies while this study used D. carinata.
Many research findings support the idea that the toxicity comes exclusively from Ag + ions as the main source of toxicity [35, 98,99]. Ag + could prevent the absorption of Na across the membranes of gills by inhibiting the Na + , K + -ATPase activity and this could lead to ionoregulatory failure causing death [100]. The significantly lower EC 50 value for Ag + in comparison to NPs shows that the ionic silver is much more toxic than AgNPs. Coatings control the release of ions from AgNPs to the surrounding medium [34,98]. As per the percentage dissolution of AgNPs, the computed concentrations of dissolved fractions of Ag + ions from T-AgNPs, E-AgNPs and C-AgNPs at EC 50 concentrations in the ASTM medium were 0.21, 0.13 and 1.06 μg L -1 respectively which were below the 48 h EC 50 value of Ag + ions. The toxicity may therefore not exclusively come from released Ag + ions from AgNPs; NPs may have other toxic effects such as generation of reactive oxygen species (ROS) causing oxidative stress [26, 93,101,102], attachment to daphnid's body surface or appendages leading to physical impairment and behavioral changes [103].
Feeding behaviour of D. carinata
Daphnid feeding rate significantly increased from day 1 to 5 for all treated algae and in controls due to higher food consumption with ageing (Fig 3). The mean 5 day feeding rate was highest in the control followed by C-AgNP, E-AgNP and T-AgNP treated algae fed daphnids. Compared with the control, there was no significant deviation (p > 0.05) in feeding rate in the day 1 while only the feeding rate of E-AgNP treated algae was significantly different in the day 2. In contrast, the feeding rates of AgNP treated algae were significantly different to the relevant control from day 3 to 5 (p < 0.05). When compared with each treatment group, Daphnid feeding rates of algae treated with E-AgNPs in day 2 and T-AgNPs in day 5 were significantly different from other two types (p < 0.05). However, a consistent difference was not observed and therefore, the data obtained is not sufficient to prove any significant variations of feeding rates among daphnids based on algae treated with AgNPs with different coatings. Besides, less favoritism for untreated algae compared to treated algae shows that AgNPs have some effect on feeding behaviour. Previous studies also have shown that the feeding rate of daphnids became depressed when contaminants were associated with algae [104][105][106]. McTeer, Dean [42] observed a reduction in feeding rates when daphnids were fed with algae treated with polymer-coated AgNPs. Nutritional characteristics of all algae samples were similar and therefore, they emphasized that the feeding reduction was due to Ag toxicity but not due to the nutritional quality of the algae diet. Zhao and Wang [93] speculated that the reduced feeding rate was due to accumulation of NPs in the gut or due to higher sedimentation of contaminated algae to the bottom of the vessel causing less availability for filter feeding.
Inputs of energy from food is critically important for population growth and survival and therefore, the feeding rate in primary consumers like Daphnia sp. could have profound implications at the population level [104]. Feeding inhibition may cause reduction of growth, targeted inhibition of internal organs and reproduction of daphnids [93,[107][108][109]. It may also have effects on water clarity, altered nutrient regeneration rates, population size of predators [110] and elevated phytoplankton biomass due to reduced grazing [111]. Therefore, the observed feeding inhibition of algae associated with NPs by daphnids in this study is a cause of concern and should be considered in assessing aquatic NP pollution.
Trophic transfer of AgNPs from R. subcapitata to D. carinata
The mean AgNP accumulation in algae treated with 50 μg L -1 C-AgNP (0.68 ± 0.09 ng 10 5 cells -1 ), E-AgNP (0.61 ± 0.08 ng 10 5 cells -1 ) and T-AgNP (0.44 ± 0.06 ng 10 5 cells -1 ) Algal feeding rates of D. carinata over a 5 d feeding period exposed to R. subcapitata cells that were treated with 50 μg L -1 concentrations of AgNPs. Data are mean ± SD from three independent experiments, each with 15 daphnids. The error bars indicate the SD (p < 0.05, n = 3). The p-values for multiple pairwise comparisons were obtained from two-way ANOVA followed by Holm-Sidak method using Sigmaplot. Letter sign denotes comparison of p-values of feeding rates for each day separately while the � sign denotes comparison of p-values of feeding rates over 5 days. Treatments that do not share lowercase letters or number of � signs are significantly different.
https://doi.org/10.1371/journal.pone.0214398.g003 concentration was significantly different to the control (0.09 ± 0.04 ng 10 5 cells -1 ) (Fig 4A). C-AgNP accumulation in algae was significantly higher than T-AgNPs. However, the Ag accumulation profile in algae does not mirror the Ag accumulation profiles in daphnids. The mean Ag content in daphnids treated with T-AgNPs (0.016 ± 0.006 ng daph -1 10 5 cells -1 ) was significantly higher than the Ag content in daphnids treated with C-AgNPs (0.006 ± 0.0004 ng daph -1 10 5 cells -1 ) (Fig 4B). Also, Ag content in daphnids treated with T-AgNPs and E-AgNPs was significantly higher than the control. Though the T-AgNP treated algae accumulated lowest amount of NPs, it led to the highest mean percentage of Ag retention (3.6%) in daphnids followed by E-AgNP (2.1%) and C-AgNP treated algae (0.95%) (Fig 4C). In contrast, McTeer, Dean [42] found metal accumulation profiles in D. magna from trophic transfer correlated with the metal accumulation profiles in algae treated with ZnNPs. The Ag bioaccumulation in daphnids did not correlate with the daphnid survival percentage (Fig 4D) where no significant difference was observed between each treatment group.
Any remaining treated algae in the digestive tract may lead to overestimation of bioaccumulation. Gillis, Chow-Fraser [112] determined the length of time required to completely depurate metal contaminated sediments from the digestive track of D. magna in the presence of algae and recommend a minimum of 8 h. Petersen, Akkanen [113] observed a limited depuration of carbon nanotubes from D. magna after 48 h whereas Zhu, Chang [49] observed complete depuration of food associated TiO 2 NPs from the gut of D. magna after 26 h. Following exposure to the treated algae, daphnids were allowed to feed on fresh algae for 72 h in fresh medium and it was assumed all NPs in the digestive track were removed. The algae treatment concentration of AgNP (50 μg L -1 ) was chosen to ensure sufficient Ag concentrations were available for detection in daphnid tissues without causing mortality to algae, based on EC 50 values obtained through an algae acute test for all types of NPs in a different study (Table B in S1 File). However, since the toxicity of Ag + ions from silver salt (Ag 2 SO 4 ) to the alga was very high (72 h EC 50 : 51 μg L -1 ), the feeding and trophic transfer experiments were not conducted for algae contaminated with silver salt at this concentration.
Several other studies have also shown the potential transfer of different types of NPs along the food chain [41,47,114]. However, it is not possible to conclude whether Ag was transferred from algae to daphnids in the form of NPs or Ag + ions. McTeer, Dean [42] hypothesized that Ag + ions liberated from AgNPs by dissolution were accumulated by algae and then transferred in to the daphnids. Van Hoecke, De Schamphelaere [90] demonstrated that NPs could not cross the double cell layer of R. subcapitata when exposed to silica NPs as confirmed by TEM images. Piccapietra, Allué [115] found AgNP internalization was limited when C. reinhardtii was exposed to carbonate coated AgNPs. However, NPs less than 20 nm may pass through the algal cell walls since the cell walls are porous (5-20 nm in size) and their permeability changes during mitosis. Also, high concentrations of AgNPs may increase the permeability of algae cell wall resulting in more internalization [39]. Miao, Luo [116] confirmed internalization of NPs in the cell after exposing Ochromonas danica to AgNPs and Kalman, Paul [39] found AgNP localized in starch granules within the chloroplast of Chlorella vulgaris as determined by TEM images. Therefore, the trophic transfer of NPs from algae to daphnids may occur in the forms of NPs or Ag + ions which depend on several factors such as the type of algae, life stage of algae and size of NPs. Data generated from this study clearly show that the type of coating affects the NP accumulation in algae and trophic transfer from algae to daphnids. However, it is not possible to predict the toxicity to daphnids based on bioaccumulation through trophic transfer from this study, and hence, further studies are recommended. The transfer of AgNPs along the aquatic food chain could have adverse implications and therefore there is a need to take this into consideration in protecting aquatic organisms. In doing so, great caution must be taken when assessing the risk of differently coated NPs.
Conclusion
The type of AgNP coating and medium significantly influenced the degree of aggregation and the behaviour of AgNPs. Based on the 48 h EC 50 values of D. carinata, we found that the Ag + ions are significantly more toxic than AgNPs. The toxicity of E-AgNP and C-AgNP were not significantly different, but T-AgNPs were comparatively about 2.5 times less toxic. Since other characteristics such as shape, size and charge are quite similar, the difference in toxicity could be attributed to the effect of different coatings. Feeding experiments revealed that the ingestion rates of NP treated algae were significantly lower than untreated algae revealing associated AgNPs with algae change daphnid feeding behaviour which could have longer term negative effects on D. carinata population. However, findings from this study are not sufficient to conclude the cause of changed behaviour. Ingestion rates of algae treated with differently coated NPs were not markedly different showing that different types of coatings had little effect on D. carinata feeding. Our findings also demonstrated the diet-borne transfer of AgNPs from AgNP contaminated R. subcapitata to D. carinata. In the algae exposed to AgNPs, T-AgNP bioaccumulation was the highest while C-AgNPs were the lowest. However, bioaccumulation of Ag in daphnids through trophic transfer did not correlate with the accumulation profiles of Ag in algal cells. The percentage Ag retained in daphnids was highest for T-AgNP treated algae while it is lowest for C-AgNPs. These results demonstrate that type of coating may have effects on AgNP accumulation profiles at different trophic levels. The behaviour of differently coated NPs in medium, their toxicity profile and trophic transfer data generated in this study demonstrate the importance of considering type of coating in environmental risk assessment. | 2019-04-05T03:28:42.027Z | 2019-04-03T00:00:00.000 | {
"year": 2019,
"sha1": "c0ed27074f664f5077458c4c662535fc7a67beab",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0214398&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "c0ed27074f664f5077458c4c662535fc7a67beab",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Medicine",
"Chemistry"
]
} |
3047239 | pes2o/s2orc | v3-fos-license | Carbon dynamics and CO2 air-sea exchanges in the eutrophied coastal waters of the Southern Bight of the North Sea: a modelling study
. A description of the carbonate system has been incorporated in the MIRO biogeochemical model to investigate the contribution of diatom and Phaeocystis blooms to the seasonal dynamics of air-sea CO 2 exchanges in the Eastern Channel and Southern Bight of the North Sea, with focus on the eutrophied Belgian coastal waters. For this applica-tion, the model was implemented in a simplified three-box representation of the hydrodynamics with the open ocean boundary box ‘Western English Channel’ (WCH) and the ‘French Coastal Zone’ (FCZ) and ‘Belgian Coastal Zone’ (BCZ) boxes receiving carbon and nutrients from the rivers Seine and Scheldt, respectively. Results were obtained by running the model for the 1996–1999 period. The simulated partial pressures of CO 2 (pCO 2 ) were successfully compared with data recorded over the same period in the central BCZ at station 330 (51 ◦ 26 . 05 0 N; 002 ◦ 48 . 50 0 E). Budget calculations based on model simulations of carbon flow for BCZ a low annual sink of atmospheric CO − 0.17 mol C m − 2 − WCH of river in BCZ on annual scale, atmospheric CO
Introduction
Coastal waters are heavily impacted by nutrients and carbon inputs from rivers and estuaries that stimulate production and remineralization of organic matter. They also intensively exchange nutrients, and inorganic and organic carbon with the open ocean across marginal boundaries (e.g. Thomas et al., 2004a). Consequently, these areas play a key role in the global carbon cycle by linking the terrestrial, oceanic and atmospheric reservoirs (Gattuso et al., 1998;Mackenzie et al., 2004). There is nowadays a great uncertainty on the overall role of coastal ecosystems to act as a sink or a source for atmospheric CO 2 due to the large variability of recorded airsea CO 2 exchanges across latitudinal and ecosystem gradients. Existing data of CO 2 air-sea fluxes suggest that temperate marginal seas act as sinks for atmospheric CO 2 (Tsunogai et al., 1999;Frankignoulle and Borges, 2001a;Borges and Frankignoulle, 2002a;DeGranpre et al., 2002;Thomas et al., 2004a, b). On the contrary sub-tropical marginal seas (Cai et al., 2003) and near-shore aquatic ecosystems influenced by terrestrial inputs such as inner and outer estuaries Cai et al., 2000;Raymond et al., 2000;Borges and Frankignoulle 2002b;Bouillon et al., 2003), mangroves and non-estuarine salt marshes (Wang and Cai, 2004) act as sources of CO 2 to the atmosphere. However, the paucity of field data prevents the comprehensive biogeochemical description of the very diverse coastal ecosystems that is needed for the integration of CO 2 fluxes on a global scale. Furthermore, little is known on the seasonal and interannual variability of CO 2 dynamics and fluxes in coastal ecosystems, although it has been shown to be highly significant in open oceanic waters in relation to large scale climatic forcings acting at various time-scales (e.g. Takahashi et al., 2003). In the absence of long term time-series, these aspects of CO 2 cycling in coastal ecosystems can only be approached with modelling tools (e.g. Ianson and Allen, 2002). Up to now, little mechanistic modelling effort has been devoted to the description of the dynamics of dissolved inorganic carbon (DIC) in coastal environments (Walsh et al., 1994(Walsh et al., , 1996Mackenzie et al., 2004).
The North Sea is amongst the best-studied coastal areas in the World, with respect to its physical, chemical and biological characteristics. However, the spatio-temporal monitoring of DIC has been limited to near-shore waters such as the German Bight, the Wadden Sea or the Belgian coastal zone (Hoppema, 1991;Frankignoulle, 1999, 2002b;Brasse et al., 2002). Recently the seasonality of pCO 2 has been investigated in the whole North Sea (Thomas et al., 2004b). On this basis a detailed carbon budget that identifies the major players in the large scale air-sea CO 2 fluxes has been established (Thomas et al., 2004b). Results point the Southern Bight as distinct from the Northern North Sea in terms of organic and inorganic carbon cycling. This is related to both their different hydrographic features (permanently well-mixed shallow versus seasonally-stratified water column, in the South and North, respectively) and the influence of rivers inputs that are concentrated in the Southern Bight.
The Belgian coastal zone (BCZ) located in the Southern Bight of the North Sea is a highly dynamic system with water masses resulting from the variable mixing between the inflowing southwest Atlantic waters through the Strait of Dover and the Scheldt freshwater and nutrient inputs. The inflowing Atlantic waters are themselves enriched with nutrient inputs from the river Seine (Lancelot et al., 1987). The river inputs characterized by a large excess of nitrate over phosphate and silicate are shaping the structure and the functioning of the ecosystem characterized by the dominance of undesirable recurrent blooms of non-siliceous phytoplankton like Phaeocystis colonies (Lancelot, 1995).
In this paper, we test a complex biogeochemical model (MIRO-CO 2 ) to investigate the present-day impact of landbased nutrients and carbon inputs from the Seine and Scheldt rivers on the functioning of the Phaeocystis-dominated ecosystem of the Southern Bight of the North Sea and the related air-sea CO 2 exchanges. The model is run over the 1996-1999 period when existing field data of CO 2 partial pressure (pCO 2 ) allow model validation. We further conduct model scenarios to evaluate the respective contribution of biological, chemical and physical processes to the seasonal and annual variability of pCO 2 . Finally, we investigate the impact of the hydro-climatic constraints (river inputs, temperature, wind stress) on the pCO 2 seasonal dynamics by running scenarios with contrasting but realistic meteorological conditions.
Model description
The MIRO-CO 2 biogeochemical model results of the coupling of the detailed biogeochemical MIRO model (Lancelot et al., 2004) with the physico-chemical module of Hannon et al. (2001) detailing the seawater carbonate system and air-sea CO 2 exchange. MIRO is a mechanistic model that describes C, N, P and Si cycling through aggregated components of the Phaeocystis-dominated ecosystem of the North Sea. It includes thirty-two state variables assembled in four modules expressing the dynamics of phytoplankton, zooplankton, organic matter degradation and nutrients (NO 3 , NH 4 , PO 4 and SiO), regeneration by bacteria in the water column and the sediment. The phytoplankton module considers three phytoplankton groups (diatoms, free-living autotrophic nanoflagellates and Phaeocystis colonies) the growth physiology of which is described according to the AQUAPHY model of Lancelot et al. (1991). The latter model considers 3 intracellular constituents (small metabolites, reserve material, functional and structural metabolites) and distinguishes different processes: photosynthesis, reserve synthesis and catabolism, growth and associated nutrient uptake, respiration and lysis. The zooplankton module details the dynamics of two groups: the microzooplankton feeding on free-living autotrophic nanoflagellates and bacteria and the mesozooplankton grazing on diatoms and microzooplankton. Phaeocystis colonies escape grazing, but are submitted to colony disruption which releases in the ambient nanoflagellated cells and organic matter. The degradation of organic matter by planktonic bacteria is described according to the HSB model (Billen and Servais, 1989), considering two classes of biodegradability for both dissolved and particulate organic matter. Benthic organic matter degradation and nutrient recycling are calculated using the algorithms developed by Lancelot and Billen (1985) and Billen et al. (1989).
The physico-chemical module of Hannon et al. (2001) details the carbonate system in seawater and calculates CO 2 exchange between the surface water and the atmosphere. The speciation of the carbonate system (in particular pCO 2 ) is calculated based on the knowledge of DIC and total alkalinity (TA), using stoichiometric relationships and apparent equilibrium constants, which are function of temperature, pressure and salinity (Weiss, 1974;Millero et al., 1993). DIC and TA are computed taking into account, respectively, the biological uptake or release of carbon and the phytoplankton assimilation of nitrate, all provided by the MIRO model. Airsea CO 2 fluxes are calculated from the pCO 2 gradient across the air-sea interface and the gas transfer velocity estimated from wind speed and using the parameterisation of Nightingale et al. (2000). The latter was chosen among several existing empirical formulations since it was established from dual tracer experiments in the Southern Bight of the North Sea.
Model runs
The MIRO-CO 2 model was applied in the coastal domain of the eastern English Channel and Southern Bight of the North Sea, between the Baie de Seine and the northern limit of BCZ (Fig. 1). The model was implemented in a multibox frame delineated on the basis of the hydrological regime and the river inputs. In order to take into account the cumulated nutrient enrichment of Atlantic waters by the Seine and Scheldt rivers, two successive boxes, assumed to be homogeneous, were chosen from the Baie de Seine to the BCZ (Fig. 1). Each successive box was characterized by its own area, depth, water temperature and average salinity, light conditions and water residence time and was treated as an open system, receiving water from the southern adjacent box and exporting water to the northern box (Lancelot et al., 2004). The boundary conditions were provided by the results of the calculations performed for the conditions existing in the Western Channel area, considered as a quasioceanic closed system. Initial conditions of nutrients were extracted from the database of Radach et al. (1995). Initial conditions of DIC (2070 mmol C m −3 ) and total alkalinity (2290 mmol m −3 ) were estimated from existing data in the area for the 1996-1999 period . The seasonal variation of the state variables was calculated by solving the equations expressing mass conservation, according to the Euler procedure, with a time step of 15 min. Conservation equations and parameters are detailed in Lancelot et al. (2004). Model simulations were run with 1996-1999 climatological forcings for hydro-meteorological conditions and river inputs. These functions were computed from recorded daily global solar radiation (Oostende Station, Institut Royal de Météorologie, Belgium), seawater temperature and monthly nutrient loads for the rivers Seine (Cellule Antipollution de Rouen du Service de la Navigation de la Seine, France) and Scheldt [Institute for Inland Water Management and Waste Water Treatment, The Netherlands) and Department of Environment and Infrastructure (Ministry of Flemish Community, Belgium). Organic carbon loads by the Scheldt were retrieved from the Dutch water base (http://www.waterbase.nl). For the river Seine, we used data described in Servais et al. (2003). Land-based fluxes of DIC and TA were estimated based on a compilation of DIC and TA concentrations in the Seine and the Scheldt rivers (Frankignoulle et al., 1996Frankignoulle and Borges, 2001b;Abril, personal communication) and river discharges, making use of the "apparent zero end-member" method (Kaul and Froelich, 1984). Atmospheric pCO 2 was extracted from the Mace Head (53 • 33 N 9 • 00 W, Southern Ireland) and the National Oceanic and Atmospheric Administration/climate Monitoring and Diagnostics Laboratory/Carbon Cycle Greenhouse Gases Group (NOAA/CMDL/CCGG) air sampling network (available at http://www.cmdl.noaa.gov/). Wind speed at 50.0 • N 6.0 • W was provided by the Pacific Fisheries Environmental Laboratory (PFEL) and is based on Fleet Numerical Meteorology and Oceanography Center (FNMOC) synoptic pressure fields.
Validation data
Model results were daily-averaged and compared with available data sets in the central BCZ at the station 330 (51 • 26.05 N; 02 • 48.50 E) for the 1996-1999 period. Over this period, nutrients, phytoplankton and chlorophyll data www.biogeosciences.net/bg/1/147/ Biogeosciences, 1, 147-157, 2004 were collected at a weekly frequency except during winter when bimonthly (Rousseau, 2000). DIC was calculated from TA and pCO 2 measured between 1996 and 1999 in the BCZ and WCH (Borges and Frankignoulle, 1999, 2002b. Due to the inherent discrepancy between in situ temperature at the time of pCO 2 measurement and that imposed by the climatological forcing of the model (Fig. 2), field data of pCO 2 were corrected with respect to the modelled temperature. Similarly, DIC and TA data obtained at station 330 were normalized to the mean BCZ salinity of 33.5 imposed by the chosen box-model implementation in order to overcome the effect of salinity discrepancy (Fig. 2). The normalized DIC and TA are written DIC (33.5) and TA (33.5) . describes reasonably well the main trends (timing and magnitude) of the diatom-Phaeocystis-diatom succession: spring diatoms (Fig. 3a) initiate the phytoplankton bloom in early March and are followed by Phaeocystis colonies (Fig. 3b) which reach concentrations of 20 mg Chl a m −3 in April (Fig. 3c). Summer diatoms bloom after Phaeocystis decline and maintain up to fall. In agreement with observations, MIRO-CO 2 results show non-negligible concentration of diatoms all over the year (Fig. 3a) contrasting with the unique bloom of Phaeocystis in spring (Fig. 3b). Phytoplankton decline in spring stimulates the development of diverse heterotrophs such as bacteria (Fig. 3d), microzooplankton (Fig. 3e) and copepods (Fig. 3f). The seasonal succession and magnitude of spring bacteria and microzooplankton biomass are well captured by the model. However, biomass
Results and discussion
Biogeosciences, 1, 147-157, 2004 www.biogeosciences.net/bg/1/147/ reached by copepods and bacteria in spring and fall, respectively, are not properly simulated. These weaknesses of the model are discussed in Lancelot et al. (2004) and is mainly attributed to some mismatch between the simulated seasonal succession of copepods and their prey which might be related to some unreliable description of either the mortality process and fate of Phaeocystis colonies and derived matter and/or the copepod feeding function. Experimental work in this matter is in progress.
Carbonate chemistry
Figure 4 compares MIRO-CO 2 simulations of DIC (33.5) (Fig. 4a), TA (33.5) (Fig. 4b) and surface water pCO 2 ( Fig. 4c) with field observations. The seasonal evolution of DIC (33.5) and TA (33.5) is reasonably well captured by the model but concentrations obtained overestimate field data, especially in winter and late fall (Fig. 4a, b) when biological activity is low (Fig. 3). The simulated DIC (33.5) clearly reaches lowest values at the time of Phaeocystis bloom maximum ( Fig. 3b; Fig. 4a) and increases in mid-May when heterotrophs prevail over autotrophs ( Fig. 4a; Fig. 3d, e, f). In agreement the highest TA (33.5) values ( Fig. 4a) are modelled in end April-early May when phytoplankton biomass is elevated (Fig. 3a, b, c) and consumes NO 3 . Lowest TA (33.5) are simulated in late fall (Fig. 4b) when phytoplankton growth has ceased but heterotrophic process and nitrification are still operating. The observed seasonal signal of TA (33.5) is well represented (Fig. 4b) but the modelled amplitude (110 mmol m −3 ) is weak compared to that suggested by field data (∼200 mmol m −3 , Fig. 4b), mainly because of the failure of the model to capture the observed late fall-winter minima ( Fig. 4b). At the annual scale, Atlantic waters bring 96% of TA in BCZ and river loads and biological processes contribute 3.9% and 0.1%, respectively. Sensitivity tests with varying initial TA corresponding to the range of observed values in WCH indicate that a reduction of WCH initial conditions of TA by less than 2% improves significantly model simulations of DIC (Fig. 4a) and TA (Fig. 4b) in winter while fall observations remain overestimated by the model (Fig. 4a, b). Elevated modelled TA in fall could result from an overestimation of river inputs during this period. At that time of the year (October to January), river discharge is high and inputs of TA account for 45% of annual river loads. Due to the importance of biological processes on magnitude and seasonal variability of DIC, modification of DIC initial value in the Western Channel has no impact on DIC magnitude in the Belgian coastal zone.
The discrepancies in the BCZ between observed and modelled DIC and TA are due to the structure of the model since initial conditions are not forced in the BCZ but result from the transformations of chemical water properties from the WCH (model initial conditions) due to river inputs and biological activity during transport through the FCZ and BCZ boxes. Altogether, these discrepancies between observed and modelled DIC and TA have no influence on the magnitude and seasonal evolution of the modelled pCO 2 .
The comparison between modelled and measured pCO 2 (Fig. 4c) shows that our model is able to describe the observed seasonal variability, both in time and amplitude. The latter (defined as the difference between the maximum and minimum values) is evaluated at 300 ppm with pCO 2 values ranging between 445 ppm in summer and 145 ppm in spring (Fig. 4c). Model pCO 2 values generally agree fairly well with observations except in fall when field data are significantly higher than modelled values. Over the winter season, modelled surface pCO 2 are close to saturation (Fig. 4c). A marked decrease of pCO 2 is simulated in early March and coincides with the onset of the diatom bloom ( Fig. 4c; Fig. 3a) and pCO 2 reaches its minimum (145 ppm) during the Phaeocystis bloom ( Fig. 4c; Fig. 3b). Modelled pCO 2 increases as Phaeocystis declines and values above saturation with respect to atmospheric CO 2 are reached in July up to the end of October (Fig. 4c). The model early decrease of surface water pCO 2 in September compared to observations might be due to some underestimation of modelled bacterial biomass (and related activity) during fall (Fig. 3d).
Carbon budget
An annual budget of C flows was constructed for BCZ based on time integration of MIRO daily simulations of biological activities (photosynthesis and planktonic and benthic respiration), Atlantic inputs, Scheldt loads and airsea CO 2 fluxes. Results of these calculations (Fig. 5) www highlight the importance of Atlantic inputs of inorganic (226.26 mol C m −2 y −1 ) and organic (10.63 mol C m −2 y −1 ) carbon representing, respectively, 95 and 99% of total inputs to BCZ. River Scheldt inputs contributes to 4.5 and 1% of total DIC and TOC inputs, respectively. The uptake of atmospheric CO 2 is low (−0.17 mol C m −2 y −1 ) i.e. less than 0.1% of the total input of DIC. Autotrophic processes fix some 8% of the annual input of DIC (18.79 mol C m −2 y −1 ) from which 77 and 20% is remineralzed in the water column and sediment, respectively (Fig. 5). At the annual scale BCZ exports to the North some 0.38 mol C m −2 y −1 as TOC while little carbon is buried in the sediment (0.01 mol C m −2 y −1 ).
Physical and biological controls of surface water pCO 2 and air-sea CO 2 fluxes in central BCZ at the seasonal and annual scales
Physical and biological mechanisms regulating pCO 2 (Fig. 4c) and air-sea exchanges of CO 2 (Fig. 6a) in central BCZ were first investigated based on the comparative analysis of MIRO-CO 2 daily forcing of temperature (Fig. 6b) and predictions of biological features (gross primary production GPP, community respiration R (including autotrophic and heterotrophic planktonic and benthic respiration) and net ecosystem production NEP (the difference between GPP and R), Fig. 6c). In winter (November to late February), NEP and R are close to zero (Fig. 6c) and pCO 2 fluctuations ( Fig. 4c) are modulated by river inputs of carbon and seawater temperature (Fig. 6b). The system is near equilibrium and air-sea CO 2 fluxes are close to zero (Fig. 6a). From mid-March to end April, NEP is high enough to reverse the positive effect of temperature increase on pCO 2 that reaches extremely low values with respect to saturation ( Fig. 4c; Fig. 6b, c). In accordance, our model calculates for this period an important sink of atmospheric CO 2 in central BCZ (up to −20 mmol C m −2 d −1 ; Fig. 6a). In May, NEP is strongly negative (i.e. the system is heterotrophic) and surface water pCO 2 increases due to both respiration and temperature increase ( Fig. 4c; Fig. 6b, c). As a consequence the system evolves progressively from a sink of atmospheric CO 2 during spring to a source during the period end of June to November (Fig. 6a). After November, the system is again close equilibrium (Fig. 6a). Overall MIRO-CO 2 estimates a weak annual sink of atmospheric CO 2 of −0.17 mol C m −2 y −1 . As a first attempt to appraise the relative importance of physical and biological controls of surface water pCO 2 , we used the approach of Takahashi et al. (2002) that separates the seasonal effect of 'net biology' from that of temperature. As defined by these authors the 'net biology' effect includes the effect of biological processes, TA variation due to nitrate utilization, air-sea exchange of CO 2 and an addition of DIC and TA by external forcing (river inputs in the present study). Removal of the temperature effect on pCO 2 simulations is obtained after normalization of modelled daily values of pCO 2 to the annual mean temperature (11.9 • C) using the equation: where T mean and T mod are, respectively, the climatological annual mean and daily temperature (in • C) used in model runs.
Biogeosciences, 1, [147][148][149][150][151][152][153][154][155][156][157]2004 www.biogeosciences.net/bg/1/147/ The seasonal effect of temperature on pCO 2 is further estimated by attributing to the mean annual pCO 2 deduced from model runs (i.e. 360 ppm) a correction factor based on the difference between T mod and T mean as above. According to these calculations, the effect of biology on pCO 2 , represented by the seasonal amplitude of pCO 2 values corrected to the mean annual temperature, is estimated at 350 ppm (150 ppm in May to 500 ppm in February). The effect of temperature change is less, 230 ppm (260 ppm in January-February to 490 ppm in July-August). The relative importance of the temperature and 'net biology' effect (T/B) is given by the ratio between the above seasonal amplitudes and is evaluated at 0.66. Thus, based on MIRO-CO 2 results we conclude that temperature contributed less to the seasonal pCO 2 variability of central BCZ in 1996-1999 than biological processes. However in areas influenced by river inputs of carbon and nutrients such as BCZ, the approach of Takahashi et al. (2002) includes their impact in the so-called 'net biology' effect.
As an alternate approach to resolve these contributions, we compared MIRO-CO 2 reference 1996-1999 simulations of the seasonal evolution of surface water pCO 2 in BCZ and the related annual air-sea CO 2 fluxes (Fig. 7a) with those obtained by closing separately biological processes, Phaeocystis only (Fig. 7b) and the river inputs of organic and inorganic carbon (Fig. 7c). For the latter scenario the river nutrient inputs were maintained. A supplementary MIRO-CO 2 scenario with closing both biological processes and river inputs of carbon was conducted to extract the solely thermodynamic effect of temperature change on pCO 2 (Fig. 7d).
Suppressing biological processes predicts in BCZ a significant annual source of atmospheric CO 2 (+0.53 mol C m −2 y −1 , Fig. 7b). This suggests that biology is responsible for the annual sink predicted in BCZ (Fig. 7a) by taking up to 0.7 mol C m −2 y −1 of atmospheric CO 2 . In the absence of biology, seasonal evolution of water pCO 2 is mainly determined by temperature and water CO 2 super-saturation with respect to the atmosphere is simulated in summer (Fig. 7b) when water temperature is elevated (Fig. 6b). Additional comparison with scenarios closing separately Phaeocystis colonies estimates at more than 50% the contribution of their bloom to this uptake (−0.47 mol C m −2 y −1 , Fig. 7b). Comparison with the reference run ( Fig. 6a) indicates that the simulated water CO 2 under-saturation in spring is mainly associated with Phaeocystis colony growth. Spring diatoms also contribute to spring water CO 2 undersaturation ( Fig. 7b) but to a lesser extent. Suppression of river inputs of organic and inorganic carbon has no detectable influence on the seasonal evolution of pCO 2 but increases significantly the annual CO 2 sink obtained with the reference run (Fig. 7a, c). River inputs of carbon thus represent a significant source of atmospheric CO 2 (+0.87 mol C m −2 y −1 ) that is partly counteracted by biology on an annual scale (−0.7 mol C m −2 y −1 ). Finally our scenarios predict an annual sink of atmospheric CO 2 of −0.32 mol C m −2 y −1 associated to the seasonal signal of temperature only (Fig. 7d). All together these scenarios shows the predominance of different controls of pCO 2 and air-sea CO 2 exchanges in central BCZ along the season. Clearly river inputs of carbon are responsible for the simulated emission of CO 2 to the atmosphere in late winter, counteracting the effect of decreasing temperature (Fig. 7c d). NEP mostly that associated to Phaeocystis colonies explains the predicted CO 2 sink in spring which, under these conditions of temperature should have been a CO 2 source towards the atmosphere (Fig. 7a, b). During summer, temperature first and then C remineralization are the most important controls (Fig. 7b, d).
Role of land-based nutrients
In order to better appraise the role of land-based nutrients and carbon in modulating the seasonal cycle of surface water pCO 2 and air-sea CO 2 fluxes, we compared model results in central BCZ with those obtained for the WCH boundary conditions (Fig. 8). For this purpose simulations of pCO 2 (Fig. 8a) and air-sea CO 2 exchanges (Fig. 8b) in both areas were compared with the temperature forcing ( Fig. 8c) and the simulated NEP (Fig. 8d). In the WCH, a fairly good agreement is observed between daily simulations of pCO 2 and available field data (Fig. 8a). As a general trend surface water pCO 2 in WCH is driven by temperature variation, being under-saturated with respect to the atmosphere between October and mid-March and oversaturated during spring-summer (Fig. 8a). Over the latter period, two small decreases in pCO 2 are simulated, coinciding with events of elevated NEP (Fig. 8d) that are counteracting the effect of temperature increase (Fig. 8a, c, d). However and contrary to BCZ, the simulated pCO 2 never reaches under-saturated values in spring due to the low productivity of WCH compared to BCZ (Fig. 8a). At the annual scale indeed BCZ is net autotrophic (integrated NEP=+0.38 mol C m −2 y −1 ) while WCH is close to metabolic balance (integrated NEP ∼0 mol C m −2 y −1 ). Accordingly, WCH is almost in equilibrium with respect to atmospheric CO 2 (−0.01 mol C m −2 y −1 ; Fig. 8b) when BCZ acts as a sink for atmospheric CO 2 (−0.17 mol C m −2 y −1 ).
Applying the approach of Takahashi et al. (2002) to WCH simulations allows us to conclude that temperature variation over the year contributes more to pCO 2 seasonal variation than biology (T/B=1.21). Comparing this result with T/B obtained in BCZ (0.66) points to the major contribution of nutrient river inputs (i.e. coastal eutrophication) in counteracting the temperature control on surface water pCO 2 .
Sensitivity to external present-day forcings
In this section we explore to what extent changing meteorological conditions (rainfall (i.e. river discharge), temperature, wind stress) could alter the seasonal and annual dynamics of pCO 2 in coastal waters. For this application, we choose to compare MIRO-CO 2 reference simulations relative to the 1996-1999 climatological year with results obtained when running the model with changing one by one either the river inputs of carbon or the water temperature or the wind speed forcing corresponding to the most contrasted "meteorological" years for the period (the dry 1996 and wet 1999 year). For each simulation, the other forcing were maintained equal to the reference 1996-1999 mean value. Over 1996-1999, nutrient loads were very variable, modulated by freshwater discharge that was higher in 1999 compared to 1996 (Fig. 9a). Nutrient concentrations in the Seine and Scheldt rivers were indeed unchanged over this period. Similarly, concentrations of organic and inorganic carbon were assumed constant. Significantly higher temperatures were observed in 1999 compared to 1996, especially between January and August (Fig. 9b). Daily wind speeds were quite variable with however a similar annual average of 5.05 and 5.16 m s −1 in 1996 and 1999, respectively (Fig. 9c). The relative impact of each forcing was explored based on a comparison of modelled pCO 2 along the season (Fig. 10) and MIRO-CO 2 annual budget of biological activity (GPP, R and NEP) and air-sea CO 2 flux (Table 1). Generally, the seasonal trends and magnitude of pCO 2 variations were not significantly modified. However, significant changes in the intensity and direction of the annual air-sea CO 2 flux were simulated (Table 1).
Wind speed forcing
We found a negligible effect of wind speed on the seasonal cycle of pCO 2 (Fig. 10a), since this forcing does not control NEP and temperature in our model (Table 1). Still, these model scenarios reveal that the daily variations of wind speed impacted significantly on the annual budget of air-sea CO 2 (18%; Table 1). The largest sink is predicted in 1996 (−0.2 mol C m −2 y −1 , Table 1) and corresponds to lower gas transfer velocity, (8.38 and 8.58 cm h −1 (normalized to a Schmidt number of 600) in 1996 than in 1999, respectively).
River forcing
Comparison of seasonal simulations of water pCO 2 obtained with 1996 and 1999 river loads, shows for 1996 lower values between January and March, similar in May and higher from May to August (Fig. 10b). The lower late winter pCO 2 simulated in 1996 are mostly due to the lower DIC river inputs as biological activity in winter is low. The initial winter pCO 2 value in 1996 also probably explains why minimal pCO 2 values simulated in spring are very similar for both years (Fig. 10b), although the maximum NEP value obtained during the Phaeocystis bloom is higher in 1999 than 1996 (0.28 mol C m −2 d −1 in 1999 compared to 0.21 mol C m −2 d −1 in 1996, not shown).
On an annual base we calculated a lower sink for atmospheric CO 2 in 1999 compared to 1996 (Table 1) which contrasts with the elevated NEP simulated in 1999 (Table 1) in response to the higher riverine nutrient inputs. This has to be related to the higher DIC river inputs in 1999 that induce higher initial pCO 2 values in winter (Fig. 10b). Altogether this scenario points the dual role of river inputs on the seasonal dynamics of pCO 2 in eutrophied coastal waters by modulating their carbon and nutrient conditions.
Temperature forcing
Clearly, the difference in time and magnitude of surface water pCO 2 simulated in 1996 and 1999 (Fig. 10c) corresponds with temperature records deviation over the period Januaryto-July (Fig. 9b). Over this period temperature is significantly higher in 1999 than in 1996 (∼3 • C, Fig. 9b). In order to eliminate the temperature effect, modelled pCO 2 were normalized to the mean BCZ temperature (11.9 • C, Fig. 10d). From January to February, the lower pCO 2 simulated in 1996 are clearly due to lower temperature compared to 1999 (Fig. 10d). The 9 days delay obtained for the minimal simulated pCO 2 values in spring (Fig. 10d) is related to the late onset of Phaeocystis bloom in 1996 as a consequence of the temperature dependence of GPP. On an annual base, NEP is lower in 1999 than in 1996 (Table 1), in spite of a higher GPP in 1999 than in 1996 (Table 1), due to the stronger temperature dependence of heterotrophic processes.
The shift from an annual sink in 1996 to a source in 1999 of atmospheric CO 2 (Table 1) is related to a combination of temperature effect on the carbon chemistry and biology (increased respiration at higher temperature). Their relative importance can be assessed through the calculation of the T/B i.e. the ratio between pCO 2(temp) and pCO 2(bio) (Table 1). This suggests a higher temperature effect on biology than carbon chemistry. Indeed the T/B increase from 0.68 in 1996 to 0.69 in 1999 (Table 1) is due to a higher difference between pCO 2(bio) (362 ppm in 1996 compared to 353 ppm in 1999) than pCO 2(temp) (247 ppm in 1996 and 244 ppm in 1999).
Conclusions
The MIRO-CO2 simulations, by showing stronger seasonal signals of pCO 2 and related air-sea CO 2 fluxes in the carbon and nutrient-enriched BCZ compared to more oceanic waters (WCH), is giving additional support to observationbased conclusions on the key role of coastal areas in the global carbon cycle (e.g. Frankignoulle and Borges, 2001a). We also demonstrated the advantage of model simulations and sensitivity analysis for the identification of the major physical and biological mechanisms controlling the seasonal dynamics of surface water pCO 2 and their external forcing. In particular, we highlighted the significance of the Phaeocystis bloom in the spring-time drawn-down of atmospheric CO 2 . The model also allows us to explore how the intensity and direction of the annual air-sea CO 2 flux was changed in relation to realistic variations of environmental forcing such as water temperature, fresh water discharge or wind speed.
Model results revealed that the seasonal variation of pCO 2 in the BCZ is strongly driven by nutrient-enriched biological activity in spring that counteracts the effect of temperature increase. On an annual scale for the 1996-1999 period, the solely river inputs of organic and inorganic carbon are driving a source of CO 2 of +0.87 mol C m −2 y −1 that is balanced by an atmospheric CO 2 uptake of −0.7 and −0.32 mol C m −2 y −1 due to nutrient stimulated NEP and temperature change, respectively. Globally an annual weak sink is predicted for BCZ. However, based on a sensitivity analysis, we also showed that the present-day interannual variations of fresh water discharge (and related land-based nutrient and carbon inputs) and water temperature might have a significant impact on the air-sea CO 2 fluxes in the BCZ.
These simulations clearly illustrate the potentially significant but complex variations of CO 2 fluxes in nearshore coastal ecosystems in response to small environmental changes within the range of present-day variability. These variations in CO 2 fluxes are a small glimpse of those that can be expected from environmental changes predicted in the near future in context of global change and/or of management strategies. | 2014-10-01T00:00:00.000Z | 2004-12-23T00:00:00.000 | {
"year": 2004,
"sha1": "9ba9407384f20a3a4de2fc514181ea8965ed1564",
"oa_license": "CCBYNCSA",
"oa_url": "https://bg.copernicus.org/articles/1/147/2004/bg-1-147-2004.pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "22a13c5d7f8b9f8f392c4aa043d88c82339ae3ac",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Environmental Science"
]
} |
252763329 | pes2o/s2orc | v3-fos-license | Fundamental Analysis based Neural Network for Stock Movement Prediction
“Stock movements are influenced not only by historical prices, but also by information outside the market such as social media and news about the stock or related stock. In practice, news or prices of a stock in one day are normally impacted by different days with different weights, and they can influence each other. In terms of this issue, in this paper, we propose a fundamental analysis based neural network for stock movement prediction. First, we propose three new technical indicators based on raw prices according to the finance theory as the basic encode of the prices of each day. Then, we introduce a coattention mechanism to capture the sufficient context information between text and prices across every day within a time window. Based on the mutual promotion and influence of text and price at different times, we obtain more sufficient stock representation. We perform extensive experiments on the real-world StockNet dataset and the experimental results demonstrate the effectiveness of our method.”
Introduction
Stock Movement Prediction aims to predict the future price trend of a stock based on its historical price or related information. Stock price prediction can help investors, ordinary users and companies to predict the stock trend in the future, which has good application value.
The high randomness and volatility of the market make the task of Stock Movement Prediction a big challenge (Adam et al., 2016). However, with the development of neural network technology, stock movement prediction has achieved good results in recent years (Nelson et al., 2017;Hu et al., 2018;Xu and Cohen, 2018;Feng et al., 2019a;Sawhney et al., 2020;Tang et al., 2021;Zhao et al., 2022). Based on fundamental and technical analysis, existing methods can be roughly grouped into two categories, namely methods based on price factors only and methods based on price and other factors (e.g., news of the stock.). Nelson et al. (2017) used the LSTM (Hochreiter and Schmidhuber, 1997) network to predict future stock price trends based on historical price and technical analysis indicators. Feng et al. (2019a) used the adversarial training as perturbations to simulate the randomness of price variables, and trained the model to work well with small but intentional perturbations. They also extracted 11 related price features to effectively help the model to predict future changes.
According to the Efficient Market Hypothesis (EMH) (Fama, 1970), price signals themselves cannot capture the impact of market accidents and unexpected accidents, while social media texts such as tweets could have a huge impact on the stock market. Based on this idea, different models have been proposed to model relevant news texts to improve the overall performance of stock movement prediction. Hu et al. (2018) proposed to use the hierarchical attention mechanism to predict the trend of stocks based on the sequence of recent related news. Xu and Cohen (2018) integrated signals from social media which reflected the opinions of general users and used Variational Autoencoder(VAE) to capture the randomness of prices and the importance of different time steps by adding temporal attention. Sawhney et al. (2020) introduced a novel architecture for efficient mixing of chaotic temporal signals from financial data, social media, and inter stock relationships in a hierarchical temporal manner through Graph Attention Neural Network.
Although previous studies have achieved good results, whether it is a purely technical approach based on historical prices or a fundamental approach based on multiple factors such as prices and news, they can be improved in terms of the full integration of the two important factors of texts and prices. We found that previous works usually encode news and prices separately according to time series, and then fuse them through simple concatenation operation, similar to the work of Sawhney et al. (2020). In fact, in practice, prices on a given day can be influenced by different news at different times (e.g., previous day or after two days). Similarly, some news about a stock on a given day may be influenced by stock prices at different times. As is shown in Figure 1, if we can capture the context information of each price and text by different days, we can get more sufficient information for predicting the stock price accurately. Figure 1: Contexts of prices and texts across the history captured by co-attention. Left shows each price representation of one day captures context of all news about the stock from day 1 to day T by different attention weights. Right shows each texts representation of one day captures context of all prices about the stock from day 1 to day T by different attention weights.
To this end, in this paper, we propose a fundamental analysis based neural network for stock movement prediction. More specifically, we first use Bi-GRU to encode the original texts of each day. Then, we use text-level attention to get a text representation of each day. As for the prices of each day, we use the existing 11 indicator features and 3 indicators we proposed in this paper as price representation of each day. Then we use the coattention mechanism (Xiong et al., 2016) to capture more information between texts and prices across every day within a time window. Finally, we incorporate a Bi-GRU to encode the fully integrated texts and prices representation according to the time window, so that it can obtain various prices and text-related information of the stock, and obtain the final effective representation of the stock.
The contributions of this work are as follows: • We propose a fundamental analysis based neural network for stock movement prediction. The model introduces the coattention mechanism into text and price features of a stock to learn the effective context information of them. The method can obtain sufficient stock representation based on the mutual promotion and influence of texts and prices at different times.
• We also introduce three technical indicators based on raw prices in the financial field as their input features to better reflect the fluctuation information of the market. We perform multiple experiments on the StockNet dataset and the results demonstrate the effectiveness of our model.
Related Work
In this section, we will review the related work about stock movement prediction from technical analysis based approach and fundamental analysis based approach.
Technical Analysis based Approach
Technical analysis based approach is to predict the trend of a stock based on its historical price features such as close price and movement percent of price, which follows the assumption that future price changes are the result of historical behavior. Most recent stock movement prediction methods are based on deep learning. Among them, recurrent neural networks such as LSTM and GRU have become a key part for capturing the temporal patterns of stock prices. This is beacuse they can further capture long-term dependencies in time series. Nelson et al. (2017) used LSTM networks to study future trends, predicting stock prices based on historical stock prices and technical analysis indicators. These indicators are mathematical calculations designed to determine or predict the characteristics of a stock based on its historical data. A total of 175 technical indicators are generated each period, and they are designed to represent or predict a very different set of characteristics of a stock, like the future price, volume to be traded and the strength of current movement trends. Feng et al. (2019a) proposed to use adversarial training and add perturbations to simulate the randomness of price variables, and trains the model to work well with small but intentional perturbations. In addition, they extracted 11 related price features that effectively help the model predict future changes. Feng et al. (2019b) proposed the Temporal Graph Convolution (TGC) model combining historical prices for predicting movement of stock, which dynamically adjusts the predefined firm relations before feeding them into Graph Convolution Network (GCN) (Kipf and Welling, 2017). As LSTM struggles to capture extremely long-term dependencies, such as the dependencies across several months on financial time series. Transformer-based employs multi-head self-attention mechanism to globally learn the relationships between different locations, thus enhancing the ability to learn long-term dependencies. Ding et al. (2020) proposed various enhancements for Transformer-based models, such as enhancing locality of Transformer with Multi-Scale Gaussian Prior, avoiding learning redundant heads in the multihead self-attention mechanism with Orthogonal Regularization and enabling Transformer to learn intra-day features and intra-week features independently with Trading Gap Splitter. However, in reality, it is often difficult to find clear pattern of change from the market historical data. Furthermore, it fails to reveal the rules governing market volatility beyond stock price data.
Fundamental Analysis based Approach
Efficient Market Hypothesis tells that textual information can be used to extract implicit information for helping predict the future trend of stock prices, such as financial news and social media. Fundamental analysis based approach is able to capture information that is not available in traditional price-based stock prediction. A hybrid attention network (Hu et al., 2018) is proposed to predict stock trends by imitating the human learning process. In order to follow three basic principles: sequential content dependency, diverse influence, and effective and efficient learning, the model builds news-level attention and temporal attention mechanisms to focus on key information in news, and applies self-paced learning mechanisms to automatically select suitable training samples for different training stage improves the final performance of the framework. Different from the traditional text embedding methods, Ni et al. (2021) proposed Tweet Node algorithm for describing potential connection in Twitter data through constructing the tweet node network. They take into account the internal semantic features and external structural features of twitter data, so that the generated Tweet vectors can contain more effective information. Financial news that does not explicitly mention stocks may also be relevant, such as industry news, and is a key part of real-world decision-making. To extract implicit information from the chaotic daily news pool, Tang et al. (2021) proposed News Distilling Network (NDN) which takes advantage of neural representation learning and collaborative filtering to capture the relationship between stocks and news. Xie et al. (2022) conducted adversarial attacks on the original tweets to generate some new semantically similar texts, which are merged with the original texts to confuse the previously proposed models, proving that text-only stock prediction models are also vulnerable to adversarial attacks. This also reflects that the model obtained only by text training is less robust, so it is still necessary to incorporate knowledge such as relevant historical price features and the relationship between stocks to better improve the performance of the model. Therefore, some studies fuse price and text data to build models, and even add the relationship between stocks to improve the performance of the model. A novel deep generation model that combines tweets and price signals is proposed by (Xu and Cohen, 2018). They introduced temporal attention to model the importance of different time steps and used Variational Autoencoder(VAE) to capture randomness of price. Recent studies have attempted to simulate stock momentum spillover through Graph Neural Networks(GNN). Sawhney et al. (2020) introduced an architecture for efficient mixing of chaotic temporal signals from financial data, social media, and inter stock relationships in a hierarchical temporal manner. Cheng and Li (2021) proposed a momentum spillover effect model for stock prediction through attribute-driven Graph Attention Networks (GAT) (Veličković et al., 2017), and the implicit relations between stocks can be inferred to some extent. Zhao et al. (2022) constructed a market knowledge graph which contains dual-type entities and mixed relations. By introducing explicit and implicit relationships between executive entities and stocks, dual attention network is proposed to learn stock momentum overflow features.
Since stock prices have temporal characteristics, that is, the price of a day will be affected by the price and news text of previous days, in this paper, we propose to use coattention mechanism to obtain the context information of stock prices and news text under different timestamp, so as to improve the final representation of the stock and the prediction performance.
Task Definition
Similar to the previous work Xu and Cohen (2018), we define the stock movement prediction task as a binary classification problem. Given a stock s, we define the price movement of the stock from day T to T + 1 as: where p c T represents adjusted closing price on day T , −1 represents stock price goes down and 1 represents the stock price goes up. The goal of the task is to predict the price movement Y T +1 of a stock s according to its historical prices collections P and news text collections L in a time sliding window of T days, where P = {P 1 , P 2 , ..., P i , ..., P T }, L = {L 1 , L 2 , ..., L j , ..., L T }, where P i is the price features of the stock s on day i and L j is the news text collection of the stock s on day j .
Overall Architecture
The whole architecture of our method is shown in Figure 2. As is shown in Figure 2, we first encode raw text for each stock across every day over a fixed time window. As for the price, the existing price features and the three new proposed indicators are concatenated together as the price representation. Then richer information will be captured by our introduced coattention mechanism. In order to obtain the integrated information of various prices and texts within the time window, we adopt a Bi-GRU for final encoding.
In the following sections, we will describe text and price features encoding in Section 3.3 and 3.4. And we will introduce temporal fusion to handle prices and text in Section 3.5 and introduce global fusion by sequential modeling in Section 3.6. Finally, model training will be introduced in Section 3.7.
Text Encoding
As each text contains rich semantic information, we use a Bi-GRU to encode the text and get the representation of each text in one day. Besides, different texts within the same day about the same stock may also be different (e.g., one text contains important information about the stock while other texts don't have valuable information about the stock.). For addressing that, we use a soft-attention operation to get the weighted representation of the texts of one day.
Following the work of Xu and Cohen (2018), we incorporate the position information of stock symbols in texts to handle the circumstance that multiple stocks are discussed in one single text. Given stock s contains K number of related texts on day m, which is denoted as L m = {l m1 , l m2 , ..., l mi , ...l mK }, where l mi denotes the i-th text of stock s on day m. For each text l mi = {w 1 , w 2 , ..., w n }, suppose that the location where the stock symbol appears first is denoted as z, we use two GRUs to encode the words sequence from w 1 to w z to get the hidden representations − → h f and words sequence from w z to w n to get the hidden representations ← − h b , respectively. We use the average of the last hidden states of the two GURs − → h z and ← − h z as the hidden representation of the text h l mi : − Where e f ,e b is the word embedding using pre-trained Global Vectors for Word Representation(GloVe) (Pennington et al., 2014) for words of the text, f ∈ [1, . . . , z] , b ∈ [z, . . . , n]. After that, we can get all the text representations M i =[h l m1 ,h l m2 ,...,h l mK ]. Since the text quality is different, we use a text-level attention mechanism to identify texts that could have a more substantial impact on the market every day, and finally obtain a final representation of all texts. The calculation formula is as follows: where α K is the attention weight, W m and W u are the parameters to be learned, b m is the bias terms. h T exts dm is the representation of the news text of stock s on m-th day(day m ). According to the time sliding window defined previously, the text data in the window is finally recorded as H t = [h T exts d1 , h T exts d2 , ..., h T exts dT ].
Price Features
As mentioned in Section 2.2, the models that predict stock trends only based on text are often fragile, while price features have been shown to effectively reflect market volatility. In this paper, we introduce three new relevant price features to be used in our method. The three new technical indicators are from financial domain and are used to describe fluctuation of stock, namely Average True Range (ATR) (Bruni, 2017), Bias Ratio (BIAS) and Momentum (MTM) (Lin et al., 2017). The detailed calculation of the three indicators is shown in Table 1. We describe the tree indicators as follows: • ATR: ATR is a volatility indicator that was developed by Wilder (1978) and is used to measure the volatility or the degree of price movement of security. It was originally designed for commodity trading, which is frequently subject to gaps and limit moves. As a result, ATR takes into account gaps, limit moves, and small high-low ranges in determining the true range of a commodity, and it also applies to the stock market.
• BIAS: BIAS is the deviation between the closing price and moving average. When the stock price moves drastically to deviate from the trend, the possibilities for a pullback or rebound increase; When the stock price movement does not deviate from the trend, it is likely that the trend will continue.
• MTM: MTM is an indicator that shows the difference between today's closing price and the closing price n days ago. Momentum generally refers to the continued trend of prices. Momentum shows a trend, staying positive for a sustained uptrend or negative for a sustained downtrend. An upward crossing of zero can be used as a signal of buying, and a downward crossing of zero can be used as a signal of selling. How high the indicator is (or how low when negative) indicates how strong the trend is. Following previous work, We adopt 11 temporal price features based on the raw price (Feng et al., 2019a), denoted as F 1 = {p 1 , p 2 , ..., p 11 } and our proposed three new price features, denoted as F 2 = {p atr , p bias , p mtm }, as our final price features. The two are concatenated together to get the final price features of m-th day, recorded as h P rices dm = [F 1 , F 2 ]. According to the time sliding window defined above, the price features in the window are finally recorded as H p = [h P rices d1 , h P rices d2 , ..., h P rices dT ].
Temporal Fusion by Coattention Neural Network
After Section 3.3 and Section 3.4, the coding features of price and text were obtained as H p and H t respectively. To effectively blend text and price, we use the coattention mechanism (Xiong et al., 2016) to learn the fusion between text and price to obtain richer implicit information. First, we use a nonlinear projection layer to convert the dimension of the price feature into the same dimension as the text with the following formula: Applying the coattention mechanism to focus on both text and price, and learn about fusion. We first compute an affinity matrix that contains the corresponding affinity scores of all prices hidden states and texts hidden state pairs. Then the affinity matrix is normalized by Softmax, attention weights are generated for text features by row, and attention weights of price features are generated by columns. The calculation formula is as follows: Next, we calculate the attention context of price features based on the attention weight of text features. The calculation formula is as follows: Meanwhile, we compute the attention context of the text features as A p H t based on the attention weights of the price features. Following Xiong et al. (2016), we also calculate A p C t which maps text feature encoding into the space of price feature encoding. The calculation formula is as follows: Where h d is interdependent representation of the text and the price. The [ ] denotes for concatenation operation.
Global Fusion by Sequential Encoding
We input h d obtained from Section 3.5 into the bidirectional GRU to obtain the hidden states for each time t. To capture past and future information as its context, we connect the hidden states from the two directions to construct a two-way encoding vector h i with the following formulas: In addition to its own information, h i also contains information about its adjacent contexts. In this way, we encoded its time series. Since news releases on different dates contributed unequally to stock trends, we employed soft attention mechanism which is calculated as follows: where β i is the attention weight, W h and W o are the parameters to be learned, b h is the bias terms. Finally, h f inal is input into a classic three-layer preceptron (MLP) to predict the future trend of stocks.
Model Training
We use cross entropy for model training, which is calculated by equation (20), where N is the total number of stocks, y t i andŷ t i represent the ground truth and predict stock trend of stock i at t day, respectively.
Dataset
We use the SotckNet 1 dataset (Xu and Cohen, 2018) to train and validate the model. The dataset contains historical data on the high trading volumes of 88 stocks in the NASDAQ and NYSE stock markets. We annotate the samples based on the movement percent of the adjusted closing price of stock, and label the samples as up and down when movement percent ≥ 0.55% or ≤ −0.5%, respectively. We split the dataset temporarily with 70/20/10, leaving us with date ranges from 2014-1-1 to 2015-8-1 for training, 2015-8-1 to 2015-10-1 for validation and 2015-10-1 to 2016-1-1 for testing. Similarly, we adjusted trading days by removing samples with missing prices or texts and further aligned data for all trading day windows to ensure that data is available for all trading days in all windows.
Experiment Settings
We use a 5-day trading day sliding window to build the samples. Following the setting of Xu and Cohen (2018), we set the maximum number of texts in a day to 30, and each text has a maximum of 40 words. Glove word embedding was also used to initialize words into 50-dimensional vectors. We train the model using the Adam optimizer, with an initial learning rate set to 5e-5. The bidirectional GRU hidden dimensions for encoding tweets and sequential modeling were set to 100 and 64, respectively. Each model is trained for 40 epochs with a batch size of 32. We report the best average test performance of the model on the validation set at 5 different runs. Following previous studies (Xu and Cohen, 2018;Sawhney et al., 2020), we use Accuracy (Acc), F1 score, and Matthews Correlation Coefficient (MCC) as evaluation metrics for this classification task.
Compared Models
To demonstrate the effectiveness of our proposed model, we compare the results with the following comparative models.
• RAND. A simple predictor to make random guess about the rise and fall.
• ARIMA. Autoregressive Integrated Moving Average, an advanced technical analysis method using only price signals. (Brown, 2004).
• Adversarial LSTM. Feng et al. (2019a) proposed a deep model using an adversarial attention LSTM mechanism, which exploits adversarial training to simulate randomness during model training. They propose the use of adversarial training to improve the generalization of neural network prediction models, since the input feature for stock prediction is usually based on stock price, which is essentially a random variable that naturally changes over time. They added perturbations to their stock data and trained the model to work well with small but intentional perturbations.
• RandForest. Pagolu et al. (2016) implemented a sentiment analysis model based on Twitter data. The authors used Word2vec to analyse the polarity of sentiments behind the tweets and directly assessed tweets related to stock and tried to predict the price of the stock for the next day.
• TSLDA. A new topic model, Topic Sentiment Latent Dirichlet Allocation (TSLDA), which can obtain new feature that captures topics and sentiments on the documents simultaneously and use them for prediction of the stock movement (Nguyen and Shirai, 2015).
• HAN. A hybrid attention network that predicts stock trends by imitating the human learning process. Follows three basic principles: sequential content dependency, diverse influence, and effective and efficient learning. The model includes news-level attention and temporal attention mechanisms to focus on key information in news (Hu et al., 2018).
• StockNet. A Variational Autoencoder (VAE) to encode stock inputs to capture randomness and use temporal attention to model the importance of different time steps (Xu and Cohen, 2018). We compare with the best variants of StockNet.
• MAN-SF. Multipronged Attention Network (MAN-SF) jointly learns from historical prices, tweets and inter stock relations. MAN-SF through hierarchical attention captures relevant signals across diverse data to train a Graph Attention Network (GAT) for stock prediction. And the study considers one pre-built graph from Wikidata (Sawhney et al., 2020).
Experimental Results
We conduct several experiments to evaluate the performance of our method. In this section, we analyze the benchmark performance and the results of our model on the StockNet dataset. The experimental results of the different models are shown in Table 2. First, we compare the first three baseline models presented in this paper. All three baseline methods use only historical price information, although Adversarial LSTM with more representative features and
Finally, compared to StockNet, which also uses texts and prices, our model is 4.4, 3.6 and 0.147 higher on Acc, F1 values and MCC, respectively. Compared to another MAN-SF using the same data, our model contains no additional knowledge of stock relations. But the result still demonstrates that our model is 1.8, 0.6, and 0.033 higher than the MAN-SF on Acc, F1 values, and MCC, respectively. Overall experimental results demonstrate the effectiveness of the proposed model.
Ablation Study
In order to better demonstrate the different effects of components of our method, we conduct ablation studies to investigate the different contribution of coattention mechanism and the three proposed financial indicators. The results are shown in Table 3. We mainly design two variants: ours w/o coattention and ours w/o ATR-BIAS-MTM.
For w/o coattention, we change the method of learning effective implicit information between price and text from the coattention mechanism to the direct concatenation of the two. This model drops 1.7, 0.7 and 0.014 compared to the full model on Acc, F1 value and MCC, respectively, proving that the coattention mechanism can effectively improve the performance of the model and obtain richer information between price and text.
For w/o ATR-BIAS-MTM, We remove the three features proposed earlier in this paper and only use the 11 features proposed in previous studies (Feng et al., 2019a). The experimental results of the model decreased by 0.3, 0.3 and 0.007 on Acc, F1 values and MCC, respectively, which also prove that these three features help the performance of the model by reflecting the volatility of the market. Here we take ATR as an example to analyze, it can simply be understood as the expectations and enthusiasm of traders. Large or increasing volatility indicates that traders may be prepared to continue buying or selling stocks during the day. A reduction in volatility indicates that traders are not showing much interest in the stock market.
Model
Acc F1
Case Study
As mentioned before, we use the coattention mechanism in the model to capture richer information, which in turn help to learn more precise attention weights of intra-day tweets (Tweet-level attention) and inter-day of time slide window (Temporal attention). In order to investigate how the coattention mechanism guides the learning of attention weights, we conducted a case study on a sample of $FB(FaceBook) between Nov 5 th and Nov 9 th , 2015, which is finally used to predict the rise or fall of Nov 10 th , 2015. As shown in Figure 3, a row represents a day. For example, the first row represents texts of 5 th . And we use the trading day alignment, because the 7 th and 8 th are weekends, so the text data for the three days from the 6 th to the 8 th were merged together. Each rectangle inside each row represents the content of a text. All texts within a day are denoted as [T ext1, T ext2, . . . , T extK]. And We present the attention weights learned by our model (Ours) and without coattention mechanism (denoted as Ours(∆)).
First, we can see that the closer to the target day, the more weight Ours gives to that day. This is also in line with the laws of the real world, and the newer news can have a greater impact. Specifically, Ours pays more attention to the positive signals from the 6 th to the 9 th . On the 5 th , it pays too much attention to a neutral T ext3 whose impact is uncertain. However, because of giving it a lower weight on the day, it can help its correct prediction for the rise. On the contrary, Ours(∆) has a greater weight than Ours on the 5 th . At the same time, the tweet texts with negative signals in the 5 th and 9 th are more concerned by Ours(∆), and finally make a wrong prediction.
Next we analyze the texts for each day in more detail. For a more intuitive understanding, we artificially add different background colors to each rectangle to represent different tendencies of the text, such as green, red and grey backgrounds representing signals with positive, negative and neutral respectively. On the 5 th day, we can see that Ours(∆) has higher attention than Ours on the two negative texts T ext1 and T ext2. During the period from the 6 th to the 9 th , Ours gives a higher weight value to the texts with positive signals than Ours(∆), such as the T ext2 from the 6 th to the 8 th and the T extK of the 9 th , which all reflect the good development prospects of FaceBook. In particular, Ours has a smaller weight than Ours(∆) on the T ext1 with negative influence in 9 th . Although this negative news appears on the day closest to the target prediction, because the model combined with coattention can fuse the information of the entire window, and analyzes that Facebook stock is still showing an upward trend in general.
The observation shown in Figure 3 indicates that the coattention mechanism can guide the model to pay more attention to texts with tendencies and can effectively model the temporal. With more accurate attention weights, Ours can capture more effective representation, thus it can achieve better performance than Ours(∆).
Conclusion
To effectively fuse texts and prices to predict future stock movements, in this paper, we propose a fundamental analysis based neural network for stock movement prediction. Our model introduces the coattention mechanism to capture richer implicit information between text and price as a better representation of a stock. We also introduce three new technical indicators in the financial field as price features. We perform the extensive experiments on the StockNet dataset and the experimental results show the effectiveness of our proposed method. In the future, we plan to use more data other than stock prices, such as financial reports, relationships between stock, to better capture market dynamics. In addition, extracting features that can better reflect trend changes is still a direction worth exploring. | 2022-10-10T13:21:36.537Z | 2022-01-01T00:00:00.000 | {
"year": 2022,
"sha1": "865f5a669df80a0782be9e681c3b12b637a32d0e",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "ACL",
"pdf_hash": "865f5a669df80a0782be9e681c3b12b637a32d0e",
"s2fieldsofstudy": [
"Computer Science",
"Business"
],
"extfieldsofstudy": []
} |
231641802 | pes2o/s2orc | v3-fos-license | S100 Genes are Highly Expressed in Peripheral Leukocytes of Type 2 Diabetes Mellitus Patients Treated with Dietary Therapy
Background and Objectives We demonstrated that the mRNA induction of S100s in rat peripheral leukocytes by severe hyperglycemia was reduced by inhibiting postprandial hyperglycemia. Here, we compared inflammatory gene expression in peripheral leukocytes between type 2 diabetes mellitus (T2DM) patients undergoing dietary therapy alone and healthy volunteers, and between T2DM patients undergoing dietary therapy alone and those undergoing such therapy in combination with drug therapy using the α-glucosidase inhibitor miglitol. Methods T2DM patients who had undertaken dietary therapy alone or in combination with drug therapy using miglitol for ≥ 8 weeks and healthy volunteers were subjected to a meal tolerance test and glucose concentration, neutrophil elastase concentration, and mRNA expression analyses of peripheral leukocytes by microarray and quantitative reverse-transcription polymerase chain reaction (qRT-PCR) immediately before and 180 min after a meal. Results Blood glucose concentrations 60 min after a meal were lower in T2DM patients with dietary + miglitol therapy than in those with dietary therapy alone. Neutrophil elastase concentrations at 60 and 120 min after a meal were lower in T2DM patients with dietary + miglitol therapy than in those with dietary therapy alone. Expression levels of S100A8 in a fasting state and S100A6, S100A8, and S100A9 180 min after a meal were higher in T2DM patients with dietary therapy alone than in healthy volunteers. Expression levels of S100A12 in a fasting state and 180 min after a meal were higher in T2DM patients with dietary therapy alone than in T2DM patients with dietary + miglitol therapy. Conclusions S100 genes were more highly expressed in T2DM patients with dietary therapy than in healthy volunteers. Supplementary Information The online version contains supplementary material available at 10.1007/s40268-020-00334-z.
Introduction
A double-blind, randomized, placebo-controlled trial (RCT), called the Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM), including 1429 participants aged 40-70 years, showed that the administration of the α-glucosidase inhibitor (α-GI) acarbose, which suppresses blood glucose increases after meal intake by inhibiting α-glucosidases in the small intestine, tracked for an average of 3.3 years, significantly reduced the incidence of cardiovascular disease (CVD) and hypertension [1]. In the Acarbose Cardiovascular Evaluation (ACE) double-blind RCT, a phase 4 trial for 176 patients in China with coronary heart disease and impaired glucose tolerance, acarbose reduced the incidence of diabetes, but not major CVD events [2]. Regarding food intake, a cohort study of 75,521 middleaged American women also showed that higher intake of foods with a lower glycemic index, an indicator of an increase in blood glucose concentration after a meal compared with reference foods, such as white bread and white rice, reduced the incidence of CVD [3]. Therefore, the suppression of postprandial hyperglycemia after meal intake in people with impaired glucose tolerance or type 2 diabetes mellitus (T2DM) can reduce the development and progression of T2DM and associated complications such as CVD. However, the mechanism by which repeated blood glucose increases after meal intake induce the onset of arteriosclerosis-related diseases in T2DM patients remains unclear.
Our recent studies have demonstrated that inflammatory genes such as interleukin-1β (Il1b) and tumor necrosis factor-α (Tnfa) as well as S100 protein family members such as S100a4, S100a6, S100a8, S100a9, and S100a11 were highly expressed in peripheral leukocytes in streptozotocin (STZ)-induced severely hyperglycemic rats. It was also indicated that such expression was reduced by inhibiting the postprandial hyperglycemia associated with dietary intake via the administration of the α-GI miglitol for 20 days [4]. Additionally, repeated single-dose daily sucrose oral loading for 4 days led to increased expression of Il1b and Tnfa in peripheral leukocytes of moderately postprandially hyperglycemic rats [5]. The reduction of glucose fluctuations by switching the α-GI from acarbose or voglibose to miglitol, the latter of which strongly reduces postprandial hyperglycemia compared with the other two, in 43 T2DM patients for 3 months reduced the mRNA expression of IL1B, TNFA, and S100A4/6/9/10/11/12 in peripheral leukocytes [6]. These results indicate that the repeated induction of postprandial hyperglycemia by meal intake may induce the development of T2DM and subsequent arteriosclerosisrelated diseases including CVD by enhancing the expression of inflammatory genes in peripheral leukocytes. It has been reported that proinflammatory cytokines such as IL1B and TNFA, as well as chemokines, which are frequently induced by oxidative stress, promote the migration of leukocytes of the innate immune system, such as neutrophils, monocytes, and macrophages [7]. Leukocyte migration is the first step of atherosclerosis because the neutrophils secrete reactive oxygen species (ROS) and proteolytic enzymes, resulting in damage to blood vessels. Subsequently, macrophages, which are differentiated from monocytes, promote atheroma formation by incorporating cholesterol at the sites of the vessels damaged by neutrophils [8][9][10]. Regarding the S100 protein family, several studies have demonstrated that the heterodimer S100A8-S100A9 induces the expression of inflammatory chemokines, macrophage activation, and infiltration of macrophages into the vascular endothelium [11][12][13]. Additionally, S100A12 binds receptors for advanced glycation end products and enhances TNFA and vascular cell adhesion molecule 1 (VCAM1) expression in mice [14]. It is very likely that these S100 proteins are related to the progression of T2DM and arteriosclerosis-related diseases, including CVD, and can be more predictive of the development and progression of these diseases than proinflammatory cytokines because S100 proteins are located upstream of proinflammatory cytokine expression, as mentioned above. However, it remains unclear whether the expression levels of S100 family proteins in peripheral leukocytes are higher in T2DM patients than in healthy individuals and whether such expression is induced by meal intake in T2DM patients and healthy individuals. Additionally, it remains unclear whether the expression of S100 family proteins in peripheral leukocytes is lower in T2DM patients administered drugs for reducing postprandial hyperglycemia than in those without this.
Therefore, in the present study, a meal tolerance test (MTT) was conducted in healthy volunteers and in T2DM patients with dietary therapy alone for 8 weeks and in those with this therapy in combination with drug therapy using miglitol for 8 weeks. Using blood samples from these participants, we examined whether the mRNA expression of inflammatory genes including those encoding S100 proteins in the peripheral leukocytes differed among the groups.
Ethical Approval of the Study Protocol
This study was a prospective exploratory trial conducted in a hospital setting (Naka Kinen Clinic, Ibaraki) in Japan. The University Hospital Medical Information Network registration number for the trial is UMIN000018370 (http://www. umin.ac.jp/engli sh/). The study protocol was approved by the Ethics Committee of Naka Kinen Clinic. Study participants provided written informed consent for the use of their personal and health information in our analysis.
Inclusion and Exclusion Criteria
We reviewed the clinical records of potential participants and identified those who met the inclusion criteria but not the exclusion criteria for our study.
Type 2 Diabetes Mellitus Patients
Inclusion criteria were male and female T2DM patients, aged 20-74 years, with glycated hemoglobin (HbA1c) (National Glycohemoglobin Standardization Program) of 6.5-8.5%. Those treated with dietary therapy alone and those treated with drug therapy using miglitol (50 or 75 mg) at each meal in combination with dietary therapy for ≥ 8 weeks before study commencement in the clinic were included (those with dietary therapy alone: N = 125; those with dietary + miglitol therapy: N = 28). In accordance with the exclusion criteria, we excluded individuals who were or could have been pregnant or were trying to become pregnant; those with severe nephropathy; hepatic disorders; CVD (diagnosed within 24 weeks before study commencement); impaired pulmonary function, pancreatopathy, cancer, myocardial infarction, or a cardiac arrest or stroke within 24 weeks of study commencement; diabetic ketosis or were in a diabetic coma; those with inadequate management of blood pressure; a need for an immediate change of medications; those undergoing insulin therapy; and those who required modification of medications within 24 weeks of study commencement. We also excluded T2DM patients with severe diabetic complications (neural disorder, retinopathy, nephropathy); severe endocrine disorders, severe anemia, and diseases of the intestine with malfunction of digestion and absorption; a history of surgery within the digestive system (excluding appendectomy and cholecystectomy) > 1 year before study commencement; malfunction of the pituitary gland; dysfunction of the adrenal glands; a history of hypersensitivity to miglitol, or lactic acidosis or digestion problems; and those who were alcoholics; malnourished, undernourished, or did not eat meals regularly; those who had not altered their diet and/or exercise for ≥ 8 weeks before study commencement; had severe infectious diseases; had an external injury; those who would be undergoing surgery; and those who had gastrointestinal disorders (e.g., diarrhea, vomiting). We also excluded patients who were considered unsuitable for the trial by doctors. In total, 30 patients in the group with dietary therapy alone and six patients with miglitol therapy were removed. Using these inclusion and exclusion criteria, we recruited patients based on numerical targets (N = 15 for each group) (Fig. 1).
Healthy Participants
We recruited healthy volunteers in the hospital. The volunteers, who were relatives of patients receiving medical treatment in the hospital, were approached, and we enrolled the participants in order. The inclusion criteria for healthy volunteers were a body mass index (BMI) of 18.5-24.9 kg/m 2 and aged 20-59 years. These volunteers were among those who visited the hospital between July 2015 and November 2015. Regarding the healthy volunteers, we excluded those who could not provide their medical history; had a surface wound; had impaired mental capacity; were dependent upon alcohol or drugs; were addicted to drugs; had a food allergy; smoked > 10 cigarettes per day, and were positive for human immunodeficiency virus antigen and antibody, hepatitis B antigen, hepatitis C virus antibody, or syphilis.
Study Cohort
This study involved open-labeled exploratory research. Therefore, we did not calculate the sample size. However, our previous study demonstrated that a single dose of acarbose in a meal test reduced postprandial hyperglycemia and tended to reduce IL1B expression in peripheral leukocytes in 16 Japanese diabetic patients with similar HbA1c levels [15]. Therefore, in this study, we set a target sample size of 15 in each group (healthy volunteers, T2DM patients with dietary therapy alone, and T2DM patients treated with miglitol therapy). Finally, we successfully recruited 15 participants in each group. We recruited 30 T2DM patients and 15 healthy volunteers. Participants were grouped as healthy volunteers (C group), T2DM patients with dietary therapy alone (N group), and T2DM patients treated with miglitol therapy (M group). All patients and healthy volunteers were enrolled before starting the trial. Patients had been undergoing stable therapy (including miglitol treatment) for ≥ 8 weeks before entering the study. All 45 individuals completed the trial, in whom, we analyzed the relationship between blood glucose concentrations and mRNA levels in peripheral leukocytes. Among the participants, one individual in the M group was removed because of an HbA1c level of over 8.5% at the trial and enrollment periods, which failed to meet the inclusion criteria at enrollment. Additionally, one participant in each group was removed because of poor amplification of mRNA. Therefore, we finally analyzed the data of 14 individuals in the C group, 14 in the N group, and 13 in the M group (Table 1).
Meal Tolerance Test
The height (AD-6225A; A&D Medical, Tokyo, Japan) and body weight (AD-6207A; A&D Medical) of the participants were measured (X-SCAN PLUS; Owa Medical, Fukuoka, Japan), and HbA1c levels as well as serum concentrations of glycoalbumin (GA) were determined using MTTs were undertaken using a test meal (total energy of 505 kcal, comprising 64% carbohydrate, 12.5% protein, and 23.5% fat) with/without a single dose of miglitol. The miglitol was taken within 1 min before meal intake. The miglitol doses in the test meal were the same as the doses used before the trial. Anthropometric data were obtained just before the MTT in the morning, and the blood glucose concentration was measured at 60, 120, and 180 min after the meal using an automatic glucose analyzer (GA-1171; Arkray, Kyoto, Japan). Serum elastase concentrations were measured by a kit measuring granulocyte elastase (Fagnos Elastase Dip kit; Sanwa Kagaku Kenkyusho Co., Ltd.). --Blood samples were collected immediately before (0 h) and 180 min after consuming the meal. Blood samples were taken for measurement of the blood glucose concentration and mRNA expression in peripheral leukocytes.
Microarray Analysis and Quantitative Reverse-Transcription Polymerase Chain Reaction (qRT-PCR)
Microarray and quantitative reverse-transcription polymerase chain reaction (qRT-PCR) analyses including S100 genes were performed with approval from the Ethics Committee of Naka Kinen Clinic to determine gene expression (3 h before meal and after meal) of the peripheral blood leukocyte fraction. In this study, we assessed mRNA levels of inflammatory genes in peripheral leukocytes upon fasting and at 180 min after a meal because our previous animal studies demonstrated that induction of the expression of inflammatory genes in peripheral leukocytes in rats by single postprandial hyperglycemia can be assessed at 180 min after oral glucose loading [5,16,17]. Blood samples were obtained in a fasting state and at 180 min after a meal for total RNA extraction from peripheral leukocytes. Immediately after collection, these samples were fixed in a PAXgene RNA tube (Qiagen/ BD, Hombrechtikon, Switzerland), which included fixation buffer, for 4 h at room temperature as per the manufacturer's directions. Total RNA from the fixed blood samples was extracted using an RNeasy Kit purchased from Qiagen. In each microarray, eight samples in each group (C group at 0 min and N group at 0 min) were selected at random from a random number table. Aliquots containing 100 ng of total RNA were converted individually to cDNA, fractionated, and labeled with a Gene ChIP ® Poly-A RNA Control kit, WT Amplification kit, and Gene ChIP ® WT Terminal Labeling kit (Affymetrix, Santa Clara, CA, USA), in accordance with the manufacturer's instructions. Hybridization, washing, and staining were carried out using Affymetrix ® HuGene2.1 ST Array Strips and a GeneAtlas ® Hybridization Wash and Stain kit for WT Assay Strips (Affymetrix), in accordance with the manufacturer's instructions. After washing, the strips were analyzed using a GeneAtlas Imaging Station (Affymetrix). Data analyses were performed using Expression Console (Affymetrix), Transcriptome Analysis Console (Affymetrix), and Excel™ (Microsoft, Redmond, WA, USA) software. We considered significant differences between two groups to be those with P < 0.05 and a fold change greater than ± 1.3. The microarray data were confirmed by qRT-PCR, and the primer sequences used in this study are listed in S- Table 1 (see the electronic supplementary material).
Total RNA samples were converted to cDNA by reverse transcription using SuperScript™ III RT (Invitrogen, Carlsbad, CA, USA), in accordance with the manufacturer's instructions. Samples of total RNA were converted to cDNA by reverse transcription using SuperScript III RT (Invitrogen, Carlsbad, CA, USA), in accordance with the manufacturer's instructions. To estimate mRNA expression of the selected genes, PCR amplification using a universal probe was undertaken on a LightCycler™ instrument (Roche Diagnostics, Rotkreuz, Switzerland). The CT (Threshold Cycle) values of each gene determined by real-time RT-PCR were converted into signal intensities based on the ΔΔCT method, which recognizes a difference of one CT value as a twofold difference between samples. The RNA levels were normalized by the level of transcription factor II B (Tf2B). The primer sequences used in this study are listed in S- Table 1 (see the electronic supplementary material).
Statistical Analyses
Values are presented as the mean ± standard deviation (SD) or number (n). Statistical analyses were undertaken using JMP version 15 (SAS Institute Inc., Cary, NC, USA). The significance of differences among the three groups and between two groups was determined using unpaired Tukey-Kramer's multiple range test and unpaired Student's t test, respectively, as appropriate. Chi-squared tests were performed between N and M groups of each diabetic complication (retinopathy, nephropathy, or neuropathy) and between the N and M groups of each category of "drugs prescribed" (i.e., those for hypertension, hyperlipidemia, or hyperuricemia). P < 0.05 was considered significant.
Basic Clinical Parameters
The healthy volunteers (C group) comprised six men and eight women, the T2DM patients with dietary therapy alone (N group) comprised 12 men and two women, and the T2DM patients with dietary + miglitol therapy (M group) comprised eight men and five women. The mean age (years) in the C group was 33.1 ± 8.85, whereas it was 56.7 ± 10.3 in the N group and 60.5 ± 9.16 in the M group.
There was no significant difference in the duration of diabetes mellitus (weeks) between the N and M groups. The BMI, systolic blood pressure, diastolic blood pressure, uric acid, and blood glucose concentration at the fasting and postprandial periods (60-180 min) were significantly lower in the C group than those in the N group. The fasting high-density lipoprotein (HDL) cholesterol concentrations were significantly higher in the C group than in the N group. The BMI, HbA1c level, GA level, and fasting blood glucose concentration did not differ significantly between the N and M groups. No differences were observed between the N and M groups for each diabetic complication (retinopathy, nephropathy, or neuropathy) or between the N and M groups for each drugs prescribed category (i.e., those for hypertension, hyperlipidemia, or hyperuricemia). The serum concentration of neutrophil elastase was higher in the C group than in the N group at 60 and 120 min after a meal. The blood glucose concentration at 60 min after a meal in the MTT was significantly lower in the M group than that in the N group (Table 1).
Microarray and qRT-PCR Analyses in Peripheral Leukocytes
mRNA expression in peripheral leukocytes was compared between the C and N groups in a fasting state, using microarray analyses. Genes with upregulated expression in the N group compared with the levels in the C group upon fasting were S100A12, CLEC17, CD22, CLEC10, F5, PLCG2, MMP9, and CLCO4C1 (S- Table 2, see the electronic supplementary material). qRT-PCR was undertaken on the upregulated and downregulated genes detected in microarray analyses and genes related to them. Differences of S100A12 between C and N groups were not observed, but the S100A12 mRNA level was higher in the N group than in the M group. Other genes such as CLEC17, CD22, and F5 were not significantly differently expressed between the C and N groups or between the N and M groups.
We performed qRT-PCR for S100 genes and proinflammatory cytokines in peripheral leukocytes because our previous animal studies showed their higher expression in diabetic animals [5,16,17]. S100A6 in a fasting state, S100A8 in a fasting state and at 180 min after a meal, and S100A9 at 180 min after a meal showed higher mRNA expression in the N group than in the M group 180 min after a meal, but not between the N group and the M group. Regarding the proinflammatory cytokines, there were no significant differences in mRNA expression for interferon-γ (IFNG), IL1B, TNFA, or IL8 between the N and C groups, or between the N and M groups ( Table 2).
Discussion
The blood glucose concentration at 60 min after a meal in the MTT was significantly lower in T2DM patients with dietary + miglitol therapy than in those with dietary therapy alone. Researchers have reported that miglitol administration in T2DM patients reduces blood glucose concentrations markedly at 30 and 60 min after a meal [18]. In this study, we first demonstrated that the mRNA expression of S100A6, S100A8, S100A9, and S100A12 in peripheral leukocytes at fasting and/or 180 min after a meal was higher in T2DM patients undergoing dietary therapy alone than in healthy volunteers (S100A6, S100A8, and S100A9) and/or T2DM patients with dietary + miglitol therapy (S100A12). S100A8 and S100A9 have been reported to be elevated in the plasma of those with type 1 diabetes [11]. A population-based cohort study following up 106 patients who survived without amputation and 93 who survived without signs of peripheral arterial disease showed that higher plasma protein concentrations of S100A12 were associated with an increased risk of amputation or death in patients with type 2 diabetes [19]. Additionally, a cohort study following 264 patients with diabetes revealed that an increased plasma S100A12 concentration was associated with subsequent death from any cause and cerebrovascular events in diabetic retinopathy patients without a previous diagnosis of macrovascular disease [20]. Furthermore, reducing glucose fluctuations by switching the α-GI from acarbose or voglibose to miglitol in T2DM patients has been shown to reduce the mRNA expression of S100A6/9/11/12 in peripheral leukocytes in a fasting state [6]. Considering these results, the higher expression of S100A6, S100A8, S100A9, and S100A12 in T2DM patients with dietary therapy alone than in healthy volunteers and/ or T2DM patients with dietary + miglitol therapy may contribute to the higher risk of CVDs.
Notably, IFNG, IL1B, IL8, and TNFA, which are major proinflammatory cytokines, were not altered between healthy volunteers and T2DM patients with dietary therapy alone, or between T2DM patients with dietary therapy alone and those with dietary + miglitol therapy. Previous studies have demonstrated that IL1B mRNA expression in peripheral leukocytes was higher in T2DM patients than in healthy participants [21]. One reason why proinflammatory cytokines were not altered between T2DM participants with dietary therapy and healthy volunteers or between T2DM participants with dietary therapy alone and T2DM participants with dietary + miglitol therapy is that fasting blood glucose was higher in T2DM participants in the previous study (179.8 ± 71.46 mg/mL) than in this study (137.21 ± 49 mg/mL). Additionally, the diabetic background, such as duration of diabetes and prescribed drugs, was not described in that study. The T2DM patients with dietary therapy in the present study had undergone an MTT without taking antidiabetic drugs, but those with dietary + miglitol therapy had undergone an MTT upon taking a single dose of miglitol. The reason why the inhibitory action of miglitol was mild in the present study is that the duration of T2DM and fasting blood glucose concentration tended to be greater in patients with dietary + miglitol therapy than in those with dietary therapy, although these differences were not significant. In this study, we cannot consider the effects of a single dose of miglitol on gene expression in peripheral leukocytes because the miglitol group was treated with miglitol for over 8 weeks. Therefore, in further study, the expression of these inflammation-related genes should be examined in peripheral leukocytes and compared between healthy participants and T2DM patients without pharmacological therapy who have visited the hospital for diabetes therapy for the first time and been diagnosed at that time. It should also be examined whether such expression and the incremental data in peripheral leukocytes are reduced by treatment with a single dose of miglitol. In addition, it should be examined whether the protein expression of these inflammatory genes in peripheral leukocytes and the serum protein concentrations differ between healthy participants and T2DM patients without pharmacological therapy, and whether such expression is reduced by treatment with longer or a single dose of miglitol. Interestingly, we found that the expression of S100A6, S100A8, and S100A10 was higher in T2DM patients with dietary therapy alone than in healthy volunteers, despite the expression of proinflammatory cytokine genes in peripheral leukocytes not differing between the two groups. Additionally, the expression of S100A12 was lower in T2DM patients with dietary + miglitol therapy than in those with dietary therapy alone.
These results indicate that the mRNA levels of S100 proteins in peripheral leukocytes are responsive to repeated postprandial hyperglycemia rather than proinflammatory cytokine mRNA levels in peripheral leukocytes. Since S100 genes are related to inflammation and/or atherosclerosis, as mentioned above, and control blood glucose concentration, even if proinflammatory cytokine expression is not higher in T2DM patients than in healthy participants, reduction of S100 genes in peripheral leukocytes may lead to reducing complications related to inflammation and/or atherosclerosis such as macrovascular and microvascular diseases. Further studies should examine whether the mRNA levels of the S100 protein family in peripheral leukocytes predict the development and progression of T2DM and related complications including CVD. It should be noted that there were significant differences in age, BMI, systolic blood pressure, diastolic blood pressure, HDL cholesterol, and uric acid as well as blood glucose concentration between healthy participants and T2DM patients with dietary therapy alone. In addition, the recruited healthy participants were younger than the T2DM patients. This was because we recruited healthy volunteers who visited the hospital and enrolled the participants in the order in which they appeared. Regarding the other parameters besides age, in general, T2DM patients tended to have higher BMI, blood pressure, and uric acid, and lower HDL cholesterol. Therefore, in this study, it is impossible to assess confounding effects of blood glucose, blood pressure, and/or lipid profiles on the expression of inflammatory genes in peripheral leukocytes. Previous studies demonstrated that the expression of S100 genes in peripheral leukocytes was higher in diabetic rats than in healthy rats and that reducing postprandial hyperglycemia in diabetic rats by miglitol suppressed expression of the S100 genes [4]. In addition, reducing glucose fluctuation in T2DM patients by changing acarbose to miglitol reduced the expression of S100 genes in peripheral leukocytes [6]. In this study, we found that T2DM patients with dietary + miglitol therapy had lower expression of S100A12 in peripheral leukocytes at fasting and 180 min after a meal and tended to have lower expression of S100A6 at 180 min (P = 0.0987) and S100A8 in the fasting period (P = 0.0692) than T2DM patients with dietary therapy. Therefore, the higher expression of S100 genes in T2DM patients with dietary therapy than in T2DM patients with miglitol could in part have been caused by elevation of blood glucose concentrations. It should be examined whether double-blind RCTs for reducing postprandial hyperglycemia using miglitol can reveal a reduction in the expression of inflammatory genes including S100 genes. Further studies are also needed to examine whether the expression of genes related to inflammation in peripheral leukocytes is higher in T2DM patients than in agematched healthy participants. It should also be examined whether anti-hyperglycemic drugs besides miglitol, antihyperlipidemic drugs, and anti-hyperuricemic drugs affect the expression of genes related to inflammation in peripheral leukocytes in T2DM patients with hyperlipidemia and/ or hyperuricemia.
It should be mentioned that the serum neutrophil elastase concentration was lower in T2DM patients with dietary therapy alone at 60 and 120 min after a meal than in healthy participants at the same time points. A recent animal study demonstrated that knockout of neutrophil elastase reduced Western diet-induced nonalcoholic steatohepatitis in mice [22]. Our results appear to be the opposite of those observed in the animal study. However, in general, the functions of neutrophils, including chemotaxis, phagocytosis, and adhesion, were reported to be lower in diabetic patients than in healthy participants [23]. Neutrophil elastase has also been shown to contribute to the neutrophil extracellular traps [24]. Therefore, a decrease of neutrophil elastase concentration in T2DM patients with dietary therapy may contribute to an increased risk of infectious diseases. This should also be investigated in further studies.
Conclusion
S100 genes such as S100A12, S100A6, S100A8, and S100A9 were more highly expressed in T2DM patients with dietary therapy than in healthy volunteers and/or T2DM patients undergoing drug therapy using miglitol. | 2021-01-19T15:04:28.993Z | 2021-01-19T00:00:00.000 | {
"year": 2021,
"sha1": "4b6a3ba55a6ad7d696367885343ddca1fdc84ec2",
"oa_license": "CCBYNC",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s40268-020-00334-z.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "4b6a3ba55a6ad7d696367885343ddca1fdc84ec2",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
120124450 | pes2o/s2orc | v3-fos-license | Physics of Finance
We give a brief introduction to the Gauge Theory of Arbitrage. Treating a calculation of Net Present Values (NPV) and currencies exchanges as a parallel transport in some fibre bundle, we give geometrical interpretation of the interest rate, exchange rates and prices of securities as a proper connection components. This allows us to map the theory of capital market onto the theory of quantized gauge field interacted with a money flow field. The gauge transformations of the matter field correspond to a dilatation of security units which effect is eliminated by a gauge transformation of the connection. The curvature tensor for the connection consists of the excess returns to the risk-free interest rate for the local arbitrage operation. Free quantum gauge theory is equivalent to the assumption about the log-normal walks of assets prices. In general case the consideration maps the capital market onto lattice QED.
We give a brief introduction to the Gauge Theory of Arbitrage [1]. Treating a calculation of net present values (NPV) and currencies exchanges as a parallel transport in some fibre bundle, we give geometrical interpretation of the interest rate, exchange rates and prices of securities as a proper connection components. This allows us to map the theory of capital market onto the theory of quantized gauge field interacted with a money flow field. The gauge transformations of the matter field correspond to a dilatation (redefinition) of security units which effect is eliminated by a proper tune of the connection. The curvature tensor for the connection consists of the excess returns to the risk-free interest rate for the local arbitrage operation. Free quantum gauge theory is equivalent to the assumption about the lognormal walks of assets prices. In general case the consideration maps the capital market onto QED, i.e. quantum system of particles with positive (securities) and negative ("debts") charges which interact with each other through electromagnetic field (gauge field of the arbitrage). In the case of a local virtual arbitrage opportunity money flows in the region of configuration space (money poor in the profitable security) while "debts" try to escape from the region. Entering positive charges and leaving negative ones screen up the profitable fluctuation and restore the equilibrium in the region where there is no arbitrage opportunity any more, i.e. speculators washed out the arbitrage opportunity.
Informal sketch
There is a common belief that nature (and a human society, in particular) could be described in some unified terms and concepts of physics may find their use here. The present paper follows the last point and tries to draw parallels between the theory of financial markets and quantum gauge theory. Since it is difficult to be equally comprehensible for both physicists and economists we give the introduction which is basically intuitive and designed to explain main logic of the consideration. Further sections are more mathematical and use notions of quantum field theory.
Net present value as a parallel transport
First of all let us recall what the NPV method is. The NPV investment method works on the simple, but fundamental, principle that money have a time value. This time value has to be taken into account through so-called discounting process.
To elucidate the idea we use here an easy example from Ref. [2]: "Suppose you were made an offer: if you pay 500 pounds now, you will immediately receive 200 French Francs, 200 Japanese Yen and 200 German Marks. How would you go about deciding whether the offer was worthwhile? What you certainly would not do is to say: I have to give up 500 pieces of paper and, in return, I will get 600 pieces of paper. As I end up with 100 more pieces of paper, the deal is worthwhile! Instead, you would recognize that in order to evaluate the offer, you have to convert all the different currencies to a common currency, and then undertake the comparison" (please, do not consider this as a Euro propaganda!). In the same way that money in different currencies cannot be compared directly, but first has to be converted to a common currency, money in the same currency, but which occur at different points in time cannot be compared directly, but must first converted to a common point of time. This reflects the time value of money.
Intuitively it is clear that given the choice between L100 now or L100 in one year time, most people would take the L100 now since the money could be placed on deposit (risk free investment) at some interest r rate. Then, in one year time the L100 turn itself into L(1 + r)100 instead of keeping L100 only. Therefore r -the interest raterepresents the time value of money.
Skipping here some details (difference between simple and compound interest rate, continuous compounding, flat and effective interest rates [3]) we now are ready to formulate what the NPV is: if an amount of money F is to be received in T years' time, the Present Value of that amount (NP V (F )) is the sum of money P (principal) which, if invested today, would generate the compound amount F in T years' time: The interest rate involved in this calculation is known as the discount rate and the term (1 + r) −T is known as T-year discount factor D T : In a similar way, to calculate the present value of a stream of payments, the above formula is applied to each individual payment and the resulting individual present values are then summed. So, the NPV method states that if the NPV of an investment project has zero or positive value, the company should invest in the project; if it has negative value, it should not invest. There is also some use of the NPV for comparative analysis of several projects but we do not stop for the details (see [2]). What is really important here for our goals is the following geometrical interpretation: discounting procedure plays a role of a "parallel transport" of a money amount through a time (though in fixed currency). The discounting factor (1) is then an element of a structural group of a fibre bundle (which still has to be defined) and the discount rate coincides with the time component of the connection vector field. Let us remind that in the differential geometry the connection field is responsable for pulling a fibre element from on point of a base space to another. Moreover, it is obvious that the "space" components of the connection has something to do with the exchange rates (see the example above). Indeed, the exchange rates or prices are responsible for converting money in different currencies or different securities (read points of discrete "space") to the same currency (point of the space) at a fixed moment of time. So, they can be interpreted as elements of the structural group which "transport" the money in "space" directions and are space analogues of the discount factor. Summing up, we see that there is some analogue between elements of a fibre bundle picture with some connection field and a capital market. We make the analogue precise below.
Arbitrage as a curvature of the connection
Next keyword is an arbitrage. What then the arbitrage is? Basically it means "to get something from nothing" and a free lunch after all. More strict definition states the arbitrage as an operational opportunity to make a risk-free profit [4] with a rate of return higher than the risk-free interest rate accured on deposit (formalized version can be found in [5]).
The arbitrage appears in the theory when we consider a curvature of the connection. In more details, a rate of excess return for an elementary arbitrage operation (a difference between rate of return for the operation and the risk-free interest rate) is an element of curvature tenzor calculated from the connection. It can be understood keeping in mind that a curvature tensor elements are related to a difference between two results of infinitasimal parallel transports performed in different order. In financial terms it means that the curvature tensor elements measure a difference in gains accured from two financial operations with the same initial and final points or, in other words, a gain from an arbitrage operation.
In a certain sense, the rate of excess return for an elementary arbitrage operation is an analogue of the electromagnetic field. In an absense of any uncertanty (or, in other words, in an absense of walks of prices, exchange and interest rates) the only state is realised is the state of zero arbitrage.However, if we place the uncertenty in the game, prices and the rates move and some virtual arbitrage possibilities to get more than less appear. Therefore we can say that the uncertanty play the same role in the developing theory as the quantization did for the quantum gauge theory.
Money flows as matter fields
Last principle ingredient to enters the theory is "matter" fields which interact through the connection. It is clear now that the "matter" fields are money flows fields which have to be gauged by the connection. Indeed, we started the introduction of the concept with the example of NPV method which shows how an amount of money units changes while the payment is "pulled" in time or one currency is converted to another.
Dilatations of money units (which do not change a real wealth) play a role of gauge transformation which eliminates the effect of the dilatation by a proper tune of the connection (interest rate, exchange rates, prices and so on) exactly as the Fisher formula does for the real interest rate in the case of an inflation [2]. The symmetry of the real wealth to a local dilatation of money units (security splits and the like) is the gauge symmetry of the theory.
Following a formal analogy with U(1) gauge theory (electrodynamics) case, we can say that an amount of a certain currency units at a particular time moment is analogues to a value of a phase cross section of the U(1) × M fibre bundle at some space-time point. If we want to compare values of the U(1) × M cross section at different points of the space-time we have to perform a parallel transport of the phase from one point of the base to another exactly as we have to convert one currency to another or discount money in the financial setting.
A theory may contain several types of the "matter" fields which may differ, for example, by a sign of the connection term as it is for positive and negative charges in the electrodynamics. In the financial stage it means different preferances of investors. Thus, in the paper we deal with cash flows and debts flows which behave differently in the same gauge field: cash tries to maximize it-self while debts try to minimize them-selves. This is equivalent to a behaviour of positive and negative charges in the same electric field where they move in different directions.
Investor's strategy is not always optimal. It is due to partially incomplete information in hands, choice procedure, partially because of investor (or manager) internal objectives. It means that the money flows are not certain and fluctuates in the same manner as the prices and rates do. So this requires a statistical description of the money flows which, once again, returns us to an effective quantization of the theory.
All together is QED-like system
Collecting this all together, we are ready to map the capital market to a system of particles with positive (securities) and negative ("debts") charges which interact with each other through electromagnetic field (gauge field of the arbitrage). In the case of a local virtual arbitrage opportunity cash-debts flows in the region of configuration space (money poor in the profitable security) while "debts" try to escape from the region. Entering positive charges and leaving negative ones screen up the profitable fluctuation and restore the equilibrium so that there is no arbitrage opportunity any more. It means speculators wash up the arbitrage.
Taking into account the uncertainty mentioned above, we come to the quantum field theory with the gauge field and "matter" fields of opposite charges. At this point standard machinery of quantum field theory may be applied to obtain distribution functions of the interest/exchange rates and cash-debts flows correlators (which are essential for the response of the system). It may answer questions about dynamical response of a financial market, dynamical portfolio theory and to be an approach to option pricing and other problems.
Last point to note in the subsectioin is a principal lattice nature of the capital market theory, since there always exists a natural minimal time interval (transaction time) and the "space" is a graph. This remove usual problems of the quatum field theory which has to deal with divergences due to contiuous character of its space-time base.
Basic assumptions
In the previous subsections we discussed the main assumption (or postulate) which is a local gauge invariance with the dilatation gauge group. It can be shown that the postulate essentially dictates dynamical rules of the theory. However, there is a number of other assumptions which we want to list here. For the convinience we devide all assumptions into two parts. First of them deals with the prices and rates , i.e. contains assumptions concerning the connection field.
Exchange rates, prices of various securities and interest rate fluctuate and it bears
local arbitrage opportunities.
2. The most probable configuration of the random connection is the configuration with a minimal absolute value of the excess rate of return.
3. Arbitrage opportunities, in absense of money flows, are uncorrelated in different space-time points, i.e. the events are statistically independant.
4. We suppose exponential distribution of the rate of return on a local arbitrage operation. Its characteristics, in general, depend on both "space" and time coordinates. However, for sake of simplicity, below we do not consider the details. 5. When a continuous limit will be taken, we assume all required smothness properties of relevant objects.
Second part of main assumtions consists of ones about a behavior of cash-debts flows i.e. the "matter" fields.
1. We assume the perfect capital market enviroment, i.e. there are always a possibility to place money on a deposit and to borrow without any restrictions and at the same interest rate [6].
2. There are transaction costs. Their presense is not just unimportant complication. The transaction costs play a role of inertia for the cash-debts flows and stabilize the system.
3. Investors are (limited) rational and try to maximize their gain from the securities, minimizing debts at the same time.
4. Exactly as the rates and prices may fluctuate and are uncertain, flow trajectories fluctuate around the most profitable trajectory ("classical trajectory"). This reflects the fact that investors not always behave optimally.
Below we repeat the assumptions in more strict form which is useful for further formalisation of the theory.
Formal constructions
In this section we formalize the previous consideration. More precise, we give a description of the relevant fibre bundles, we construct the parallel transport rules using for this elements of the structural group and give an interpretation of the parallel transport operators. The corresponded curvature is also defined and it is shown that it is equal to the rate of excess return on the elementary plaquette arbitrage operation. This opens a way to a construction of the dynamics of the parallel transport factors what gives a lattice gauge theory formulation. The procedure of the dynamics construction is repeated then for the case of "matter" fields which represent cash-debts flows.
Fibre bundle construction
It is well-known that many important concepts in physics can be interpreted in terms of the geometry of fibre bundles [7]. Maxwell's theory of electromagnetism and Yang-Mills theories are essentially theories of the connections on principal bundles with a given gauge group G as the fibre. Einstein's theory of gravitation deals with the Levi-Civita connection on the frame bundle of the space-time manifold. In the section we show how the construction of fibre bundle can also be applied to describe a framework to develop the capital market theory.
Construction of the base Now we have to construct a base of the fibre bundle. Let us order the complete set of assets (which we want to analyze) and label them by numbers from 0 to N. This set can be represented by N (asset) points on a 2-dimensional plane (the dimension is a matter of convinience and can be choosen arbitrary). To add the time in the construction we attach a copy of Z -lattice (i.e. set of all integer number {..., −1, 0, 1, 2, ...}) to the each asset point. We use discretized time since anyway there is a natural time step and all real trades happen discretely. All together this gives the prebase set L 0 = {1, 2, ..., N} × Z.
Next step in the construction is to define a connectivity of the prebase. To do this we start with an introduction of a matrix of links Γ : L 0 × L 0 → {0, ±1} which is defined by the following rule for any x ≡ (i, n) ∈ L 0 and y ≡ (k, m) ∈ L 0 : Γ(x, y) = 0 except for 1. i = k and n = m − 1 accepting that the i-th security exists at the n-th moment and this moment is not an expiration date for the security; 2. n = m − 1 and at n-th moment of time the i-th asset can be exchanged on some quantity of k-th asset at some rate (we assume that the trasaction takes one unit of time).
Using the matrix Γ(., .) we define a curve γ(x, y) in L 0 which links two points x, y ∈ L 0 . We call the set γ(x, y) ≡ {x j } j=p j=1 a curve in L 0 with ends at points x, y ∈ L 0 and p − 1 segments if The whole L 0 can be devided into a set of connected components. A connected component is a maximal set of elements of L 0 which can be linked by some curve for any pair of elements. The base L is defined now as the connected component contained US dollars at, say, 15.30 of 17-th of June 1997. This complete the construction of the base of fibre bundle.
Structural group
The structural group G to be used below is a group of dilatations. The corresponding irreducable representation is the following: the group G is a group of maps g of R + ≡]0, +∞) to R + , which act as a muliplication of any x ∈ R + on some positive constant λ(g) ∈ R + : Transition functions of a fibre bundle with the structure group correspond below to various swap rates, exchange rates, discount factors for assets.
Fibres
In the paper we use fibre bundles with the following fibres F : 1. F = G, i.e. the fibre coincides with the structure group. The corresponding fibre bundle is called Principle fibre bundle E P . A gauge theory in the fibre bundle in the next section corresponds to random walks of prices and rates.
2. F = R + . This fibre bundle will be important to describe cash-debts flows. Indeed, a cross section (or simply a section) s (a rule which assigns a preferred point s(x) on each fibre to each point x ∈ L of the base) of the fibre bundle is a "matter" field. In the context s(x ≡ (i, m)) gives a number of units of i-th assert at the moment of time m.
Actions and the corresponding functional integrals will be written in terms of cross sections of the fibre bundles. The main property of the objects (actions and measures of the integrations) will be local gauge invariance, i.e. independance on a local action of the structural group.
The fibre bundle E we use below is trivial, i.e. E = L × F and we do not stop for the definition of projections. Construction of the fibre bundles for the simple stock exchange, FX-market and financial derivatives can be found in Ref [1].
Parallel transport, curvature and arbitrage
A connection is a rule of the parallel transport of an element of a fibre from one point (x) of a base to another point (y). It means that an operator of the parallel transport along the curve γ, U(γ) : F x → F y is an element of the structural group of the fibre bundle [7]. Since we do not deal with continuous case and restrict our-selves on a lattice formulation, we do not need to introduce a vector -field of the connection but rather have to use elements of the structural group G. By the definition, an operator of the parallel transport along a curve γ, U(γ), defined as a product of operators of the parallel transport along the links which constitute the curve γ: It means that we need to define only the parallel transport operators along elementary links. Since U(γ) = U −1 (γ −1 ), this restrict us on a definition of those along an elementary links with a positive connectivity. Summing up, the rules of the parallel transport in the fibre bundles are completely defined by a set of the parallel transport operators along elementary links with a positive connectivity. The definition of the set is equivalent to a definition of the parallel transport in the fibre bundle. Since in subsection 2.1 the connectivity was defined by a possibility of assert movements in "space" and time, it allows us to give an interpretation of the parallel transport. In the subsection two principle kinds of links with positive connectivity were defined. First one connects two points (i, n) and (i, n + 1) and represents a deposition the i-th assert for one unit of time. This deposition then results in a multiplication of the number of assert units by an interest factor (or internal rate of return factor) calculated as: where ∆ is a time unit and r i is an appropriate rate of return for the i − th assert. In the continuous limit r i becomes a time component of the corresponding connection vector-field at the point (i, ∆n). By the same way the parallel transport operator is defined for the second kind of the elementary links, i.e. links between (i, n) and (k, n + 1) if there is a possibility to change at n-th moment a unit of i-th assert on S i,k n units of k-th assert: U((i, n), (k, n + 1)) = S i,k n ∈ G . In general, an operator of the parallel transport along a curve is a multiplier by which a number of assert units is multiplied as a result of an operation represented by the curve.
Results of parallel transports along two different curves with the same boundary points are not equal for a generic set of the parallel transport operators. A measure of the difference is a curvature tensor F . Its elements are equal to resulting change in multiplier due to a parallel transport along a loop around an infinitasimal elementary plaquette with all nonzero links in the base L: The index m runs over all plaquette links, {U m } are corresponding parallel transport operators and an agreement about an orientation is implied. Now we show that the elements of the curvature tensor are, in fact, an excess returns on the operation corresponded to a plaquette. Since elements of the curvature tensor are local quantities, it is sufficient to consider an elementary plaquette on a "space"-time base graph. Let us, for example, consider two different assets (we will call them for the moment share and cash) which can be exchanged to each other with some exchange rate S i (one share is exchanged on S i units of cash) at some moment T i , and the reverse rate (cash to share) is S −1 i . We suppose that there exists a transaction time ∆ and this ∆ is taken as a time unit. So the exchange rates S i are quoted on a set of the equidistant times: Interest rate for cash is r 1 so that between two subsequent times T i and T i+1 the volume of cash is increased by factor e r 1 ∆ . The shares are characterized by a rate r 2 . As we will show latter, due to the gauge invariance we can fix r 1 to be risk free interest rate and r 2 related to the average rate of return of the share.
Let us consider an elementary (arbitrage) operation between two subsequent times T i and T i+2 . There are two possibilities for an investor, who posseses a cash unit at the moment T i , to get shares by the moment T i+2 . The first one is to put cash on a bank deposit with the interest rate r 1 at the moment T i , withdraw money back at the moment T i+2 and buy shares for price S i+1 each. In this way investor gets e r 1 ∆ S −1 i+1 shares at the moment T i+2 for each unit of cash he had at the moment T i . The second way is to buy the shares for price S i each at the moment T i . Then, at the moment T i+2 investor will have S −1 i e r 2 ∆ shares for each unit of cash at the moment T i . If these two numbers (e r 1 ∆ S −1 i+1 and S −1 i e r 2 ∆ ) are not equal then there is a possibility for an arbitrage. Indeed, suppose that e r 1 ∆ S −1 i+1 < S −1 i e r 2 ∆ , then at the moment T i an arbitrager can borrow one unit of cash, buy S −1 i shares and get S −1 i e r 2 ∆ S i+1 units of cash from selling shares at the moment T i+1 . The value of this cash discounted to the moment T i is S −1 i e r 2 ∆ S i+1 e −r 1 ∆ > 1. It means that S −1 i e r 2 ∆ S i+1 e −r 1 ∆ − 1 is an arbitrage excess return on the operation. In the other hand, as we have shown above, this represent lattice regularisation of an element of the curvature tensor along the plaquette. If e r 1 ∆ S −1 i+1 > S −1 i e r 2 ∆ then arbitrager can borrow one share at the moment T i , sell it for S i units of cash, put cash in the bank and buy S i e r∆ S −1 i+1 shares at the moment T i+2 . We have an arbitrage situation again.
We consider the following quantity It is a sum of excess returns on the plaquette arbitrage operations. In continuous limit this quantity converges as usual to a square of the curvature tensor element. The absense of the arbitrage is equivalent to the equality S −1 i e r 2 ∆ S i+1 e −r 1 ∆ = S i e r 1 ∆ S −1 i+1 e −r 2 ∆ = 1 , and we can use quantity (2) to measure the arbitrage (excess rate of return). In more formal way the expression (2) may be written as In this form it can be generalized for other plaquettes such as, for example, the "space"-"space" plaquettes. As we saw above, excess return is an element of the lattice curvature tensor calculated from the connection. In this sense, the rate of excess return for an elementary arbitrage operation is an analogue of the electromagnetic field. In the absence of uncertainty (or, in other words, in the absence of walks of prices, exchange and interest rates) the only state is realized -the state of zero arbitrage. However, if we introduce the uncertainty in the game, prices and the rates move and some virtual arbitrage possibilities appear. Therefore, we can say that the uncertainty plays the same role in the developing theory as the quantization does for the quantum gauge theory.
Last point to add in the section is a notion of the gauge transformation. The gauge transformation means a local change of a scale in fibres: together with the following transformation of the parallel transport operators: It is easy to see that the parallel transport operation commutes with a gauge transformation: and the curvature tensor is invariant under the transformation:
Gauge field dynamics
In the previous subsections we showed that the exchange rates and interest rate (or, more general, internal rate of return) discount factors are elements of the structural group of the fibre bundle. Moreover, they are responsable for the parallel transport in "space" and time directions correspondingly. In the present subsection we address a question about a dynamics for the exchange/discount factors. At a first sight the dynamics is difficult to specify since it is not restricted and any attempt to formulate the dynamics seems to be voluntary and not enough motivated. However, as we show below the dynamics can be derived from a few quite general and natural ssumptions. The main postulate of the present analysis is an assumption about local gauge invariance with the dilatation group as a gauge group.
Postulate 1: Gauge invariant dynamics
We assume that all observable properties of the financial enviroment (in particular, rules of dynamical processes) do not depend on a choice of units of the asserts. It means that all effects of, say, change of currency units or shares splittings may be eliminated by a corresponding change of interest rates, exchange rates and prices [8]. This is a very natural assumption which allow us, however, to make a step to a specification of the dynamics. Indeed, due to the gauge invariance an action which governs the dynamics has to be constructed from gauge invariant quantities.
Postulate 2: Locality
Furthermore, we assume local dynamics of the exchange/interest rate factors. This locality means that in the framework the dynamics of an assert is influenced by connected (in the sense of Γ connectivity on the base graph L) asserts only.
These two postulates allow us to make the following conclusion: the action s gauge has to be a sum over plaquettes in the base graph of some function of the (gauge invariant) curvarture of the connection. It means that the action is a sum over plaquettes of a function of the excees return on the plaquette arbitrage operation as it was shown in the previously.
Postulate 3: Free field theory -correspondance principle We postulate that the action is linear on the plaquette curvatures on the base graph since this is a simplest choice. It will be shown later that the postulate is equivalent to the free field theory description in an absense of matter fields and produce quasi-Brownian walks for the exchange/interest rates in the continuous time limit. This means that the approach generalizes standard constructutions of the mathematical finance. This fact serves as a correspondance principle.
Postulate 4: Extremal action principle In fully rational and certain economic enviroment it should be no possibility to have "something from nothing", i.e. to have higher return than the riskless rate of return. In more general form, the excess rate of return (rate of return above the riskless rate) on any kind of operations takes the smallest possible value which is allowed by external economic enviroment. Together with the locality of the action this give the extremality principle for the action.
Postulate 5: Limited rationality and uncertainty The real enviroment is not certain and fully rational and there exist nonzero probabilities to get other excess rates of return (exchange rates, prices and interest rates fluctuate and its bears local arbitrage opportunities). We assume that the possibilities to have the excess return R(x, T ) at point of "space" x and time moment T = 0 are statistically independant for different x, T and are distributed with an exponential probability weight e −βR(x,T ) with some effective measure of rationality β. If β → ∞, we return to a fully rational and certain economic enviroment.
Formally, we state that the probability P ({U i,k }) to find a set of the exchange rates/ interest rates {U i,k } is given by the expression: Now we are ready to writing down the general action for the exchange/interest rate factors. However, before writing we would like to consider in more details a very simple example which gives some insight of the general framework. Let us, once again, consider two-assert system (cash-shares).
The first moment to mention is a gauge fixing. Since the action is gauge invariant it is possible to perform a gauge transformation which will not change the dynamics but will simplify further calculation. In lattice gauge theory [9] there are several standard choices of the gauge fixing and the axial gauge fixing is one of them. In the axial gauge an element of the structural group are taken constant on links in time direction (we keep them e r 1,2 ∆ ) and one of exchange rates (an element along "space" direction at some particular choosen time) is also fixed. Below we fix the price of the shares at moment T = 0 taking S 0 = S(0). It means that in the situation of the ladder base the only dynamical variable is the exchange rate (price) as a function of time and the corresponding measure of integration is the invariant measure dS i S i . From above derivation, the definition of the distribution function for the exchange rate (price) S = S(T ) at the moment T = N∆ under condition that at the moment T = 0 the exchange rate was S 0 = S(0) is given by: It is not difficult to see that in the limit ∆ → 0 the expression in brackets converges to the integral −β Here we introduced so-called volatility σ as β = σ −2 and the average rate of share return µ as µ = r 1 − r 2 . It is easy to give an interpretation of the last relation. The system at whole is not concervative and both the interest rate r 1 and the rate r 2 come outside of the system (from banks and the corresponding company production). Let us imagine that the world is certain and that because of the company production the capitalization of the firm increased. For this amount new shares with the same price S 1 have been ussued (no dividends have been payed). The number of new shares for each old share is equal e r 2 ∆ . It means that the cumulative (old) share will have a price S 1 e r 2 ∆ while the original price (at zero moment) was S 0 . Taking into account discounting and certainty, we end with the following expression: which tells that the rate of return on the share is equal r 1 − r 2 . After introducing an uncertainty last expression turns into an average rate of return on the share. Eqn (6) returns us to a justification of Postulate 3 which, together with other Postulates, are equivalent to a log-normal model for a price walks in an absence of matter fields, which we consider in details in the next section. Now we give the general expression for probability distribution of a given exchange rates and internal rates of return profile: where the sum is calculated over all plaquettes in the base graph with all links with nonzero connectivity Γ, the index m runs over all links of a plaquette, {U m } are corresponding elements of the structural group, which perform the parallel transport along the links.
To complete the gauge field consideration we want to return once again to the gauge invariance principle. Next section is devoted to the consideration of the "matter" field which interact through the connection. It is clear now that the field is a field which represents a number of assets (as a phase in the electrodynamics) and has to be gauged by the connection. Dilatations of money units (which do not change rules of investors behaviour) play a role of gauge transformation which eliminates the effect of the dilatation by a proper tune of the connection (interest rate, exchange rates, prices and so on) exactly as it is in the Fisher formula for the real interest rate in the case of inflation [2]. The symmetry to a local dilatation of money units (security splits and the like) is the gauge symmetry of the theory.
Effective theory of cash-debts flows: matter fields
Now let us turn our attention to "matter" fields. These fields represent cash-debts flows on the market. Importance of the cash-debts flows for our consideration is caused by their role in a stabilization of market prices. Indeed, if, say, some bond prices eventually go down and create a possibility to get bigger return than from other assets, then an effective cash flows appear, directed to these more valuable bonds. This causes an upward shift of the prices because of the demand-supply mechanism. All together these effects smoothen the price movements. The same picture is valid for debts flows if there is a possibility for debts restructurisation. As we will see all this features finds their place in the constructed framework.
To formulate an effective theory for the flows we will assume that: 1. Any particular investor tries to maximize his return on cash and minimize his debts.
2. Investor's behaviour is limited rational i.e. there are deviations from pure rational strategy because of, for example, lack of complete information, specific financial manager's objectives [2] and so on.
We start with a construction of an effective theory for the cash flows and the generalize it for the case of debts presence. The first assumption tells us that an investor tries to maximize the following expression for the multiplier of a value of his investment (in the case of cash, shares and securities): by a certain choice of the strategy which is resulted in the corresponding trajectory in "space"-time for assets. Here {U i } N i=i are exchange (price) or interest factors which came from a choice of the investor behaviour at i-th step on the trajectory C and boundary points (at times T = 0 and T = N) are fixed. We assume that there is a transaction time which is smallest time in the systems equal to ∆. In other words, a rational investor will choose the trajectory C 0 such that Choosing the best strategy, fully rational investor maximizes his return s. However, as we assumed limited rationality, in analogy with the corresponding consideration of the connection field probability weights, we define the following probability weight for a certain trajectory C with N steps: with a some "effective temperature" 1/β which represents a measure of average unrationality of investors in a unit time.
It is possible to generalize the approach to a case of many investors operating with cash and debts. The corresponding functional integral representation for the transition probability (up to a normalization constant) has the form [1]: with the actions for cash and debts flows: We do not stop here to describe the gauge invariant boundary conditions which are discussed in details in Refs [1,10]. It is interesting to note that the Eqn(10) may be transformed to Eqn(9) inverting signs of r i and inverting exchange rates S in the absense of the transaction costs. This corresponds to the fransformation from negative to a positive charge. The transaction costs make this symmetry only approximate.
In the absence of restructuring of debts (i.e. one kinds of debts cannot be transfromed to another kind of debts) last terms containg S have to be dropped out. Then the action takes esspecially simple form: All said above was related to the particular case of two assets problem. Now we give a form of the action for the most general situation. As it was shown before, a general system of assets is described by a form of the "space"-time base graph L of the fibre bundle. Elements of the graph are labeled by (i, k) where the index i, i = 0, .., N, represents a particular asset (a point in "space") and the index k labels time intervals. By the definition of the base L any two points of the base can be linked by a curve γ, each formed by elementary segments with a nonzero connectivity Γ. In its own, each of the segments is provided with an element of the structural group U which performs a parallel transport along this (directed) link. These allow us to give a most general form of the action s1 ((i 12 , k 12 ) ≡ (i 1 , k 1 ), (i 2 , k 2 ) ∈ L : Γ((i 1 , k 1 ), (i 2 , k 2 )) = 1): s1 = 1 β (i 12 ,k 12 ) (ψ + (i 1 ,k 1 ) U β (i 1 2,k 1 2) (1 − tδ i 1 −1,i 2 ) β ψ (i 2 ,k 2 ) − δ (i 1 ,k 1 ),(i 2 ,k 2 ) ψ + (i 1 ,k 1 ) ψ (i 1 ,k 1 ) ) . (12) In the same way the action for the debts flows can be written: s1 ′ = 1 β (i 12 ,k 12 ) (χ + (i 1 ,k 1 ) U −β (i 1 2,k 1 2) (1 + tδ i 1 −1,i 2 ) −β χ (i 2 ,k 2 ) − δ (i 1 ,k 1 ),(i 2 ,k 2 ) χ + (i 1 ,k 1 ) χ (i 1 ,k 1 ) ) . (13)
Conclusion
In conclusion, we proposed a mapping of the capital market theory in the lattice quantum gauge theory where the gauge field represents the interest rate and prices and "matter" fields are cash-debts flows. Basing on the mapping we derived action functionals for both the gauge field and "matter" fields assuming several postulates. The main assumption is a gauge invariance of the dynamics which means that the dynamics does not depend on particular values of units of assets and a change of the values of the units may be compensated by a proper change of the gauge field. The developed formalism has been applied to some issues of the capital market theory. Thus, it was shown in Ref [10] that a deviation of the distribution function from the log-normal distribution may be explained by an active trading behavior of arbitragers. In this framework such effects as changes in shape of the distribution function, "screening" of its wings for large values of the price and non-Markovian character (memory) of price random walks turned out to be consequences of damping of the arbitrage and directed price movements caused by speculators. In Ref [11] consequences of the bid-ask spread and the corresponding gauge invariance breaking were examined. In particular, it turned out that the distribution function is also influenced by the bid-ask spread and the change of its form may be explained by this factor as well. So, the complete analysis of the statistical characteristics of prices have to account both these factors. Besides, it is possible to show [12] that the Black-Scholes equation for the financial derivatives can be obtained in the present formalism in an absence of speculators (i.e. absence of the arbitrage game).
Let us now make two final remarks.
1. In all cited above references a very simple model of a stock exchange was considered, where the only one kind of security is traded. The consideration can be generalized on more realistic situation with a set of traded securities. Following this line, dynamical portfolio theory can be constructed and, in static (equilibrium) limit, will coincide with standard portfolio theory [13]. In the dynamical theory time-dependent correlation functions will play a role of response functions of the market to an external perturbation such as a new information or a change in macroeconomic environment. Account of virtual arbitrage fluctuations will lead to a time-dependent modification of CAPM [3].
2. Since the influence of the speculators leads to a non-Markoffian character of a price walks there is no possibility to eliminate a risk using arbitrage arguments to derive an equation for a price of a derivative. Then the virtual arbitrage and corresponding asset flows have to be considered. It will lead to a correction of Black-Scholes equation.
We will return to these points in our forthcoming papers. | 2019-03-29T11:50:04.356Z | 1997-10-18T00:00:00.000 | {
"year": 1997,
"sha1": "2fdfae7a9f94c148ce05baf202a3be00d88abb64",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "2fdfae7a9f94c148ce05baf202a3be00d88abb64",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Physics",
"Economics"
]
} |
2462847 | pes2o/s2orc | v3-fos-license | Neurological Manifestations of Brucellosis in an Indian Population
Brucellosis is a zoonotic disease causing serious public health problems in countries of the Middle-East and developing countries like India. Neurobrucellosis is one of the devastating complications of this re-emerging zoonosis. The objective of this review was to identify the neurological manifestations of Brucellosis in an Indian population and bring into light the effective modalities used for treating neurobrucellosis. A systematic review of the scientific literature reported in accordance with the preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines was conducted. Three databases (PubMed, IndMed, and ScienceDirect) were used to analyze retrospectively case reports of sufficient quality for data extraction (from the last 15 years, 2002-2016), and relevant literature was reviewed. Most of the cases had a definite history of exposure to Brucella through occupational contact with cattle, drinking raw milk, or living near unhygienic abattoir or even trips to epidemic areas outside India. The common presentations include fever, meningitis, brisk deep-tendon reflexes, extensor plantars, sensory deficit usually below the twelfth thoracic vertebral level, weakness of lower limbs, ocular signs of papilledema, and retrobulbar neuritis. The usual systemic findings associated were hepatosplenomegaly and weight-loss. Neurobrucellosis needs to be kept in mind in the differential diagnosis of fever of unknown origin involving neurological symptoms and systemic involvement. Prognosis is good if there is a combination of antibiotics, each with different mechanisms of action given in full dose. Suitable measures for its prevention are also suggested.
Introduction And Background
Brucellosis is an infectious, febrile disease affecting animals as well as humans. This globally common zoonosis is caused by the bacteria of several Brucella species [1]. Brucellosis-free countries have seen re-introductions of the disease associated with the movement of Brucella infected livestock [2]. According to Chromel BB et al. (1994), human brucellosis is not usually acquired through animal contact, but is transmitted more often by consumption of infected animal products [3]. However, human brucellosis has been associated with sporting activities, such as hunting, as well as with tours to endemic countries, where sometimes even the local delicacies (like aborted animal fetuses in Equador) are a source of brucellosis [2,4]. Brucellosis in humans presents as a diverse clinical picture, with neurobrucellosis being an uncommon but serious complication [5]. The treatment option for neurological manifestations of brucellosis has always been a combination of antibacterials of different mechanisms of action [5]. The duration of treatment depends on the severity of neurological involvement. In this review, we specifically discuss the neurological involvement of human brucellosis as per case reports from the developing country of India, where neurobrucellosis is rarely reported [6].
A systematic review of the scientific literature, reported according to the preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines, was conducted. Three databases were searched, which consisted of two international databases (PubMed & ScienceDirect) and one Indian database (IndMed). A total of 53 articles were found using the search words 'Neurobrucellosis' and 'India' with a date limit of the years 2002 to 2016 (the last 15 years). Two articles were found to be duplicate copies of one another in different databases, and from the remaining 51, the relevance criterion was fixed as papers with only case reports specifically with neurobrucellosis as the main manifestation. Thus, we classified them as: relevant articles=10; irrelevant articles=41.
Then, we did a retrospective analysis of the clinical manifestations, laboratory, and radiological findings and treatment modalities along with follow-up results of those ten case reports in great detail [5][6][7][8][9][10][11][12][13][14]. To our knowledge, this is the first review of neurobrucellosis from India so its clinical importance should be realized owing to the fact that a large number of such cases go undiagnosed/misdiagnosed as tuberculosis, especially in the villages of India [15].
Review
The clinical findings of ten patients from ten case reports related to neurological manifestations of brucellosis published between January 2002 and May 2016 were considered in this paper. The average age of the cases was 34 years (range: 6-59 years). Gender was reported for all of the patients; there were seven males and three females. There was a history of exposure to animals in 60% of the cases, and 40% of the patients had returned from recent trips outside of India as shown in Figure 1.
FIGURE 1: History of Exposure to Brucella Sp.
High-grade fever was reported in 80% of the cases. Headaches, impaired consciousness, and seizures were also reported in some cases. Sensory deficit was seen in 60% of the cases. It included impaired sensations of lower limbs, as well as the defects in the sensation of vision and hearing. Motor deficit was reported in 70% of cases which usually included loss of motor functions of lower limbs. Increased deep tendon reflexes, extensor plantars, anal sphincter abnormalities, and hemiparesis were also noted. Meningoencephalitis was found in 20% of the cases. The features studied are summarized in Figure 2 and
FIGURE 3: Clinical Presentations of Neurobrucellosis (Part 2)
The commonly accompanying systemic findings (as shown in Figure 4) with neurobrucellosis were hepatosplenomegaly and weight loss. In 70% of the patients, radiological findings of the brain/spine were positive. The raised blood titers against Brucella were found in 100% of patients, and in cases where a lumbar puncture was done, it showed increased proteins and lymphocytes with cerebrospinal fluid (CSF) culture positive for Brucella.
FIGURE 4: Associated Systemic Findings in Our Cases of Neurobrucellosis
The drugs used were always a triple combination of the following drugs: doxycycline, rifampicin, streptomycin, gentamicin, co-trimoxazole, and ceftriaxone. Diazepam and valproate were used for managing associated seizures (if any). In 20% of the cases, neurobrucellosis was confused with and misdiagnosed as disseminated tuberculosis. In 90% of the cases, the patients became asymptomatic after the respective course of treatment; though in one case, a patient died due to misdiagnosis and late starting of antimicrobials against Brucella. The follow-up details are summarized in Figure 5.
FIGURE 5: Follow-Up Results of Cases of Neurobrucellosis
Human brucellosis is caused by strains of Brucella which are small, gram-negative, unencapsulated, nonsporulating, non-motile rods, or coccobacilli. In vivo, Brucella bacteria behave as facultative intracellular parasites. Their resistance to drying renders Brucella stable in aerosol form facilitating airborne transmission. The majority of human disease has been due to Brucella melitensis [15]. Nervous system involvement in brucellosis is rare, but is a serious manifestation [5]. In an experimental animal model, the ganglioside-like molecules expressed on the surface of Brucella melitensis were found to induce anti-GM1 ganglioside antibodies resulting in flaccid limb weakness and ataxia-like symptoms [16]. In 2007, the incidence of neurobrucellosis in brucellosis patients was reported to be 18.86% from northwest India, i.e., Bikaner, from a hospital-based case series of 175 patients with serologically confirmed cases of brucellosis [17].
There is a wide variation in the clinical presentations of neurobrucellosis with more common ones being meningeal irritation, polyneuropathy/radiculopathy, and diffuse neural involvement. Also, fever, headache, sweating, sensory loss, motor deficit, weakness in extremities, and hearing and vision defects, which were reported from Turkey in a pooled analysis of 187 cases, were consistent with our review of ten cases from India. This study also revealed that their cases had hepatosplenomegaly and weight loss as the commonly associated systemic findings, which were found in 50% and 20% of our cases, respectively [18].
For the diagnosis of neurobrucellosis, the neurological symptoms (usually not accompanied by systemic manifestations), positive CSF culture/blood serology, and CT/MRI abnormalities are required. The latter were found by Al Sous MW et al., to be of three types: inflammation, white matter changes, and vascular insult [7,19]. In our review, 70% of the cases had positive radiological findings.
Treatment of neurobrucellosis includes more than one antibiotic but it should be able to cross the blood-brain barrier (BBB) and achieve good CSF concentration, e.g., doxycycline and rifampicin. The concentration of streptomycin/gentamycin in CSF is therapeutic only when meninges are inflamed. Ceftriaxone and co-trimoxazole are also being used. Diazepam and valproate have been used for seizure management. Short-term use of steroids has been implicated in associated hydrocephalus and conditions like papilledema, cranial neuropathies, and raised intracranial pressure [7,[19][20].
The duration of treatment has to be individualized [20]. The prognosis is usually good with an improvement of symptoms; however, if the treatment is started late, it may prove fatal [8].
The case reports we studied provided data from patients with neurobrucellosis having presented to health centers. It may be possible that cases who do not present to health centers are less severe. The result of this review may, therefore, be biased toward more severe cases.
Conclusions
Neurobrucellosis may mimic a large number of central and peripheral nervous system pathologies. Thus, we need to keep neurobrucellosis in mind in cases of fever of unknown origin, meningitis/meningoencephalitis with lymphocytic predominance in CSF, and other neurological manifestations especially when these occur with hepatosplenomegaly or a recent history of weight loss. Also, in neurotuberculosis or any similar condition seen in a patient having a history of travel to an endemic country, contact with cattle or consumption of infected dairy products, a differential diagnosis of neurobrucellosis should be considered. We strongly suggest that emphasis if given on the veterinary prevention and testing of infection in herds/flocks including their active immunization, control of animal movement, pasteurization of dairy products and most importantly, the reporting of cases to appropriate public health authorities, the incidence of a Brucella infection will decrease in both animals and humans.
Conflicts of interest:
The authors have declared that no conflicts of interest exist. | 2018-04-03T05:36:12.489Z | 2016-07-01T00:00:00.000 | {
"year": 2016,
"sha1": "d61f57176b8a5b7d304e997d764c8093651fa1e9",
"oa_license": "CCBY",
"oa_url": "https://assets.cureus.com/uploads/review_article/pdf/4802/1511801230-20171127-1339-n27izr.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "9fed4bb06296dd1ad86cc3543a98f9c19b1dcec3",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
212971152 | pes2o/s2orc | v3-fos-license | Recovering from A Disaster: A Study of The Relief and Response System
Mass natural disasters cause severe damage. Aiming at the multi-supply and multi-demand point problem of routes scheduling and packing configurations, we are ready to provide adequate and timely response to the emergencies.
Summary
Mass natural disasters cause severe damage. Aiming at the multi-supply and multi-demand point problem of routes scheduling and packing configurations, we are ready to provide adequate and timely response to the emergencies.
First, we identify the best locations of cargo containers to conduct the two required missions. We use ArcGIS software and create TCL model (Target-oriented comprehensive location model) to assist. TCL model aims to find the best location under restrictions with comprehensive methods. TCL model includes buffer analysis, Kernel Density Analysis, the Slope Algorithm and IDW Interpolation.
Then we design the flight routes with transferring it into a generalized TSP (GTSP) and solving it via the minimum-weight Hamilton cycle. We discuss two situations: the video reconnaissance is of the whole island or only the main roads. Besides, to maximize task completion with the limited flight distance, we consider two kinds of flight trips: the round trip and the one-way trip. After comparing the characteristics of different drones, we are supposed to choose 2 or 3 kinds of drones to implement the transport tasks.
Since the flight routes raise requirements for the amounts of drones and the medical packages, we design packing configurations for each cargo container. We employ the genetic algorithm to solve this 3D container-loading problem. Then we can find the best plan with the relatively largest space utilization.
Once we have depicted the flight routes and packing configurations meticulously, our IWE (Independence Weighted Evaluation) Model calculates how effective the candidate solutions are and provides the best drone fleet. Thinking about the real situation of a disaster area, the road condition may change even in a few days. So, we creatively design a kind of plan called loop, which means we send drones to finish a video reconnaissance once l days. After making a drone schedule with linear regression, we get all the details of the drone fleet and its schedule.
Next, we will take the worst hurricane which hit the United States territory of Puerto Rico in 2017 as an example. We are going to recommend a drone fleet and set of medical of the response system, identify the best location on Puerto Rico to position cargo containers to transport the system then provide the drone payload packing configurations and a drone flight plan eventually.
1)
Assume that a drone's flight speed remains unchanged during flying.
2)
If the drone needs to obtain videos, it will only fly along the road. (The reason is shown in to Appendix.)
3)
If the drone needs to deliver medicine, it can fly directly between two places.
Select a Drone
The relationship between drone endurance and payload is shown in the following figure.
According to the data, we figure out that type B, C and F have the highest ranks among the seven candidates with flight capability. B is especially suitable for long distance flight. C and F have high payload capability. Thus, to simplifying the model and maximize the utility of the drones, we will choose type B, C and F for further design.
Identify the Best Locations for the Cargo Containers with TCL Model
We introduce TCL model (Target-oriented comprehensive location model) to assist. TCL model aims to find the best location under restrictions with comprehensive methods. The model includes buffer analysis, Kernel Density Analysis, the Slope Algorithm and IDW Interpolation.
Figure. 3 Locations of the hospitals
For this problem, we use the kernel density analysis to get the probability density of the island to the highway and present in a thermodynamic chart. Then, we use the DEM data to calculate the slope. The flatter the area, the more suitable it is to locate the cargo container. Rarely will only one factor or one layer be sufficient in ordinary spatial analysis involving typical GIS analysis. So we choose the weighted overlay tool and raster data for a solution. The overlay result is as following: Figure. 4 Spatial Overlay of Raster Data Finally, we identify the best locations of the cargo containers as below (the brightest nodes). The geographic coordinates are as follows:
Design the Flight Routes
We regard the endpoints of each road as the vertices set. We regard the roads as the arcs between two vertices. We also regard the length or the flight time of each road as the weight of the arc, respectively. Then the problem is transferred into a Generalized Traveling Salesman Problem (GTSP). That is, given a set of target vertices and three set of drones located at distinct depots, find a Hamiltonian path for each drone such that each vertice is visited at least by one drone, the drone-target constraints are satisfied and the sum of the costs of the paths traveled by all the drones is minimized [3].
We employ graph theory to solve this problem. In graphical terms, the aim is to find a minimumweight Hamilton cycle in a weighted complete graph.
Design Packing Configurations for Each Cargo Container
The basic calculating process of GA (genetic algorithm) is as following: Figure. 6 Calculating Process of GA
Evaluate Solutions
To build the evaluation model, the first step is to select evaluation indicators and then do the dimensionless method. Moreover, it is supposed to determine the weighted vector of evaluated index with multi-correlation coefficient. Eventually, IWE model is | 2019-12-05T09:30:16.382Z | 2019-11-29T00:00:00.000 | {
"year": 2019,
"sha1": "5a6e40e7112d028580a58311473b60af881336c4",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1755-1315/384/1/012206",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "c2c3003480caa8b392401dd5ed5f1d05858202cb",
"s2fieldsofstudy": [
"Environmental Science",
"Engineering"
],
"extfieldsofstudy": [
"Physics",
"Computer Science"
]
} |
15413029 | pes2o/s2orc | v3-fos-license | A pharmacodynamic study of sorafenib in patients with relapsed and refractory acute leukemias
We report the results of a Phase I dose escalation trial of the multikinase inhibitor sorafenib in relapsed/refractory acute leukemias using an intermittent dosing regimen. Fifteen patients with advanced leukemia (Acute myeloid leukemia(AML), 2=Acute lymphoblastic leukemia(ALL), 1 Biphenotypic) and a median age of 63 (range 37–85) years were enrolled and treated on a dose escalation trial. Toxicities ≥grade 3 were present in 55% of cycles and the maximum tolerated dose (MTD) was determined to be 400mg BID × 21days in a 28 day cycle. Plasma inhibitory assays of kinase targets ERK and FLT3-ITD demonstrated excellent target inhibition, with FLT3-ITD silencing occurring below the MTD. The N-oxide metabolite of sorafenib appeared to be a more potent inhibitor of FLT3-ITD than the parent compound. Despite marked ex vivo FLT-3 ITD inhibition, no patients met criteria for complete or partial response in this monotherapy study. Eleven of fifteen patients experienced stable disease as best response. Although sorafenib demonstrated only modest clinical activity as a single agent in this heavily treated population, robust inhibition of FLT3 and ERK suggest there may be a potential important role in combination therapies.
INTRODUCTION
Acute leukemia in adults remains difficult to cure with conventional cytotoxic chemotherapy with approximately 70% of adult patients diagnosed with AML and ALL dying of the disease. (1)(2) One recent strategy is based on targeting cellular signal transduction pathways found to be mutated or overactive in the malignant clone. Acute leukemia has several potential targetable pathways and one pathway of great interest is the Ras/Raf/MEK/Erk pathway due to its role in cellular division, differentiation, and apoptosis. This pathway is constitutively activated in more than 50% of primary acute myeloid leukemia samples; (3) and constitutive activation is associated with an inferior clinical outcome. (4) This pathway can be activated by mutations in receptor tyrosine kinases such as FLT3, c-KIT and VEGF. (5) Sorafenib is a multi-targeted tyrosine kinase inhibitor, with activity against RAF kinase, VEGF receptors, both wild type and ITD(Internal Tandem Duplication)-mutated FLT3, PDGF receptors, c-KIT, and RET kinase. (6) Sorafenib was recently approved by the U.S. Food and Drug Administration for the treatment of renal cell cancer (7) and hepatocellular carcinoma. (8) Preclinical studies of sorafenib in acute leukemia have demonstrated downregulation of the MAPK pathway, sensitization to receptor-mediated apoptosis by downregulation of Mcl-1(Myeloid cell leukemia-1), (9)(10) and growth inhibition of AML cells with FLT3-ITD mutations. (11) Early published clinical studies of sorafenib in AML suggest continuous dosing at a dose approved for solid tumors (400mg twice daily) is not tolerated well in patients with AML/ MDS. (12) Sorafenib has been found to occasionally induce hematologic responses and complete remissions in select patients with FLT3-ITD AML. (13)(14) Based on these data and earlier work that suggested improved tolerance in intermittently dosed schedules, (15) we undertook a phase I dose escalation trial to determine the dose limiting toxicities (DLT) and maximum tolerated dose (MTD) of sorafenib given orally, twice-daily (BID), for either 14 days or 21 days of a 28-day treatment cycle in patients with refractory acute leukemia. We also examined pharmacokinetics, pharmacodynamics, and tumor response. Correlative studies included the assessment of target modulation via plasma inhibitory assay (PIA) of phosphorylated-ERK (p-ERK) and phosphorylated-FLT3-ITD (p-FLT3) using methods previously developed for FLT3 targeted therapies. (16)(17) Additionally, we investigated the metabolism of sorafenib during intermittent dosing to assess residual active compounds found after discontinuation of sorafenib. This examination focused on one metabolite, sorafenib N-oxide, which was thought to have biologic properties similar to the parent compound (18) and has been reported to represent the most common metabolite representing ~17% of circulating total drug. (19) Prior investigations of this metabolite have not studied its activity against mutated FLT3.
Patient selection
Patients over the age of 18 years with pathologic confirmation of relapsed or refractory AML, or ALL, were considered eligible. Standard end organ function and performance status criteria for Phase I investigations were used, including bilirubin <2.0 mg/dL, AST/ALT <5× upper limit of normal, and creatinine clearance >60mL/min/1.73m 2 . Patients were required to have a peripheral blast count <30,000/µL and no greater than a 50% increase in absolute blast count within the preceding week. Hydroxyurea was allowed up to 48 hours after starting sorafenib. Protocol and consent form were approved by the Johns Hopkins School of Medicine Institutional Review Board. All patients gave informed consent according to the Declaration of Helsinki.
Treatment plan
Patients were initiated on twice daily oral tablet dosing of sorafenib for either 14 or 21 consecutive days of treatment in each 28-day cycle and were managed in the outpatient clinic. Table 1 lists the planned dosing levels. All patients were evaluated for DLT for the purpose of determining the MTD.
Evaluation of response
Baseline evaluations, including an on-study bone marrow aspirate and biopsy, were conducted within 1 week prior to entry into the study. Bone marrow assessments were performed on or about cycle 1 day 8 (early treatment assessment), prior to initiation of cycle #2 and every two cycles thereafter. Clinical responses for AML and ALL were measured according to International Working Group definitions.(20)
Determination of DLT, MTD and Stopping Rules
All patients filled out medication and side effect/toxicity diaries that were reviewed weekly. Toxicities were graded according to the National Cancer Institute Common Toxicity Criteria, Version 3.0. Dose escalation continued until a DLT occurred in 2 patients out of the 3 patients in a cohort. When 1 DLT was observed in the first 3 patients during the first treatment cycle at a given dose level, 3 additional patients (up to 6 patients in total) were treated at that level. When DLTs occurred in the first 2 or 3 patients treated at a given dose level, no further dose escalation occurred. The dose immediately below the dose level that produced 2 DLTs was considered the maximum tolerated dose (MTD). Patients without evidence of disease progression or DLT secondary to therapy were allowed to continue on that dose for a total of 6 cycles. No intrapatient dose escalation was permitted.
Pharmacokinetic studies
Serial blood samples were collected in lithium heparin-containing tubes prior to and at 0.25, 0.5, 1, 2, 4, 6, and 8 hours after the administration of the first dose of sorafenib. Additional blood samples were collected prior to administration (C min ) on day 2, 3, 8, 15 of continuous dosing and prior to the start of cycle 2. Samples were processed within 30 minutes of collection by centrifugation for 10 minutes at 1,500× g under refrigeration (~4°C). The resultant plasma was stored at −70°C until subsequent analysis for sorafenib and sorafenib N-oxide concentrations using a validated liquid chromatography/tandem mass spectrometry (LC/MS/MS) method. (21) Briefly, sorafenib and sorafenib N-oxide were extracted from plasma using acetonitrile precipitation. Separation of sorafenib, sorafenib N-oxide, and the internal standard, [ 2 H 3 15 N] sorafenib, was achieved on a Waters X-Terra™ C 18 (150 mm × 2.1 mm i.d., 3.5 µm) analytical column using a mobile phase consisting of acetonitrile/10 mM ammonium acetate (65:35, v/v) containing 0.1% formic acid and isocratic flow at 0.2 mL/min for 6 minutes. The analytes were monitored by tandem mass spectrometry with electrospray positive ionization. Linear calibration curves were generated over the range of 0.007-7.26 µg/mL(0.02-15.6µM) for sorafenib and 0.01-2.5 µg/mL(0.2-5.2µM) sorafenib N-oxide. Plasma samples that were diluted 1:10 (v/v) with pooled plasma were accurately quantitated. The accuracy and within-and between-day precisions were within the acceptance criteria for bioanalytical assays. (22) Pharmacokinetic variables were calculated by standard noncompartmental methods using WinNonlin professional (version 5.2).(23) C max was the observed maximum concentration. T max was the time point at which the C max was observed. C min was the observed minimum concentration or pre-dose sampling. C min was determined to be valid if the interval between evening dose and the sample was within an allowed range of 12 ± 2 hours and obtained predose. AUC (0-8h) was the observed AUC, calculated using a combination of linear and log trapezoidal rules. Pharmacokinetic parameters were summarized using descriptive statistics.
Correlative studies
Inhibitors-Sorafenib was provided by Bayer Pharmaceuticals through the Investigational Drug Branch, Cancer Treatment Evaluation Program, National Cancer Institute. Sorafenib N-oxide was obtained from Toronto Research Chemicals. Sorafenib and sorafenib N-oxide were dissolved in DMSO and stored at −80°C as 10 mM stock solutions. All samples in any given experiment contained identical concentrations of DMSO.
Cell Culture-All cell lines were cultured in RPMI/10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA) at 37°C in 5% CO 2 . The TF/ITD cell line was derived by transfecting TF-1 cells (growth factor dependent) with an expression vector containing the FLT3 coding sequence containing an ITD mutation from an AML patient, as described. (16) The resultant TF/ITD line is growth factor independent and expresses constitutively phosphorylated FLT3 and ERK 1/2.
Plasma Inhibitory Activity (PIA) Assay-The PIA assay was performed as described previously using the TF/ITD cell line as target cell line.(17) Phosphorylated ERK1/2(Cell Signaling, Danvers, MA) was assessed in whole cell lysates.
Sorafenib phase I patient characteristics
The patient characteristics are listed in Table 2 and were typical of patients in early phase leukemia trials with a median of 3 (range 1-6) prior therapies.
Sorafenib associated toxicities and Maximum Tolerated Dose
Toxicities (any grade) that were potentially related to therapy were seen in 23 out of 31 cycles (74%) ( Table 3). Grade ≥3 toxicities were experienced in 17 (55%) cycles. Dose limiting toxicities at 600mg BID × 14 days were demonstrated in 2/3 patients and included elevated transaminases (n=1), and musculoskeletal back pain unimproved with standard measures (1). Based on the dose limiting toxicities described at 600mg BID ×14d, the dose of 400mg BID × 21d was determined to be the MTD and further dose escalation did not occur.
Sorafenib and sorafenib N-oxide pharmacokinetics
All patients were evaluable for pharmacokinetic analysis (Table 4). Sorafenib exhibited a variable plasma concentration-time profile with a slow absorption phase followed by a long terminal elimination phase thus resulting in a relatively flat concentration-time profile as previously described. (24)(25)(26)(27) Sorafenib N-oxide exhibited a similar profile with the maximum concentration occurring at the same time of after the T max for sorafenib. (19) Moderate inter-individual variability in pharmacokinetic parameters was noted, with a coefficient of variation for the sorafenib AUC (0-8h) and C max of up to 95% and 116 %, respectively. The variability was higher for the sorafenib N-oxide metabolite with the coefficient of variation for AUC (0-8h) and C max of up to 129% and 124 %, respectively. Sorafenib concentrations were detectable in 33% (1/3) of patients after 14 day break in treatment and in 80% (4/5) of patients after a 7 day break. Sorafenib N-oxide was only detectable in 40% (2/5) of patients after a 7 day break.
FLT3 and ERK inhibition
We prepared dose response curves assessing inhibitory potency of sorafenib on FLT-3 ITD autophosphorylation and ERK phosphorylation using TF-ITD cells in RPMI with 10% FBS. Immunoblot analysis revealed an IC 50 of 1.2 nM for inhibition of phosphorylated FLT3 in media (data not shown). The IC 50 for inhibition of phosphorylated ERK was similar (0.91nM, data not shown). Previous work with tyrosine kinase inhibitors has shown most inhibitors in development are highly protein bound. (16)(17)28) We therefore determined the IC 50 values of sorafenib for inhibition of phosphorylation of FLT3 and ERK using TF-ITD cells, substituting normal human plasma for culture medium. In plasma, the IC 50 of sorafenib for P-FLT3 inhibition shifts to approximately 308nM ( Figure 1A). ERK signaling in plasma was more resilient with an IC 50 rising to 841nM ( Figure 1A).
FLT3 inhibition at trough time points
Pharmacodynamic analysis of FLT3 and ERK targeting was performed on patients completing one cycle of therapy (n=9). Direct analysis of leukemic phosphoprotein status was difficult as sufficient circulating leukemia cells were not available from most patients at each correlative time point due to low white blood count. We examined serial plasma specimens collected pretreatment on day 1, day 8, day 15, and day 29 in an inhibitory assay (PIA) using a FLT-3 mutant cell line(TF-ITD) to assess target inhibition potential in an ex vivo setting. PIA data takes into account protein binding, active metabolite levels, and cytokine levels which may influence target sensitivity to inhibition. Each of the patients studied (n=9) achieved complete inhibition of FLT3-ITD phosphorylation in the PIA on trough samples drawn while on therapy ( Figure 2) suggesting inhibition of FLT3-ITD occurs below 400mg BID.
Sorafenib inhibits FLT3 partially through a metabolite
To better understand the activity of sorafenib against FLT3 and ERK, we plotted the results of the PIA assays with the pharmacokinetic values obtained at those time points ( Figure 3A & 3B). We then overlaid the standard curves for sorafenib suppression of FLT3-ITD in normal human plasma. Interestingly, we found that nearly all PIA assessments appeared to have greater inhibitory activity than predicted by the standard curves with sorafenib alone.
With the possibility of a drug metabolite contributing to the biologic activity of sorafenib on FLT3, we assayed serial plasma samples for sorafenib's major metabolite, sorafenib Noxide. Sorafenib N-oxide on average was found at levels approximately 12% of sorafenib levels, (median=8%, range 2%-44%, standard deviation 12%) with marked inter-patient variability. We found the N-oxide metabolite to be more potent than sorafenib(~14.59 fold) at inhibiting autophosphorylation of FLT3-ITD and ERK in human plasma when assessed by western blotting in cell based assay using TF/ITD cells ( Figure 1B). To examine the additive effect of the presence of the N-oxide metabolite we then replotted the PIA/PK data to account for the activity of the N-oxide metabolite of sorafenib by including the sum of the parent and metabolite PK values corrected for the increased potency (14.59 fold) of sorafenib N-oxide. This created an "adjusted sorafenib concentration" value which more closely approximated the sorafenib standard curve for P-FLT3 inhibition and P-ERK inhibition ( Figure 3C & 3D) for samples in the inhibitory range.
FLT3 inhibitory activity significant and extends up to 7 days post dosing
Plasma samples at trough obtained during the trial resulted in complete FLT3 silencing ( Figure 2) and the average drug concentrations (Table 4) were well above the FLT3 predicted inhibitory range ( Figure 1). Pharmacodynamic examination of FLT3 inhibitory activity in the PIA also revealed 4/5 samples with inhibitory activity at day 29, seven days after the last dose ( Figure 2) and all four patients received more than one cycle of therapy. In two samples the primary FLT3 inhibitory compound at Day 29 when adjusted for potency was sorafenib N-oxide (pt016 and pt 17) with levels of sorafenib N-oxide of 0.20 and 0.09 µg/mL (0.42 and 0.19µM) respectively.
Clinical activity of sorafenib
The best response demonstrated in 11/15 of patients on this trial was stable disease. The longest duration of SD was 3 months, experienced by 2 of the patients (13%). While no patients met the criteria for complete or partial response, bone marrow blast counts decreased in 6/15 (40%) patients after one cycle by an average of 18%. Table 5 represents pre and post treatment for all patients on trial. Five of the 9 patients treated on a three week schedule of drug showed a decrease in their marrow blasts while only 1 of five evaluable patients treated on either of the 2 week schedules showed a decrease. Interestingly, this patient was the only one on the 2 week dosing schedules with a FLT3-ITD mutation.
One of the two ALL patients cleared their peripheral blood blasts while on sorafenib and demonstrated progressive improvement in bone marrow blasts from 89% pretreatement to 47% after cycle 1 and 45% after cycle 2. Interestingly, the patient had a MLL rearranged ALL with translocation (4;11). This translocation has been associated with over expression of WT FLT3 and in vitro sensitivity to FLT3 inhibitors. (29)(30)
DISCUSSION
Sorafenib can inhibit numerous potential pathways in acute leukemia, but despite such broad biologic activity, we found relatively limited single-agent clinical activity. The clinical activity in our study population was limited to the observation of reduced bone marrow blasts in 56% (5/9) of patients treated at 400mg BID × 21 days and 1 patient with FLT3-ITD AML treated on the 14 day schedule. Dose escalation beyond 400mg BID was limited due to grade 3 and 4 toxicities. Similar clinical activity and toxicity has been reported with single agent sorafenib given continuously where 11 out of 20 patients were found to have a >50% reduction in peripheral blood blasts and 9 of 11 patients with FLT3-ITD AML had measured responses, including one morphologic marrow CR.(31) One potential explanation for the lack of improved tolerance to our intermittent dosing was the demonstration of prolonged biologic activity after therapy completion.
Interestingly in AML, one target of great importance is FLT3-ITD, and in our study, sorafenib demonstrated suppression of FLT3-ITD at dosing below the MTD. Prior studies targeting FLT3-ITD with lestaurtinib documented the association of sustained complete or near complete inhibition of FLT3-ITD with clinical response;(32) but unlike the use of lestaurtinib, sorafenib uniformly suppressed FLT3-ITD in all samples assessed by PIA. This finding was somewhat surprising based on sorafenib pharmacokinetics and preclinical studies of sorafenib alone; however our correlative studies suggest the sorafenib N-oxide metabolite contributes significantly to in vivo FLT3-ITD inhibition. This inhibitory activity persisted up to seven days after the completion of drug dosing in several patients. This observation is clinically important with preclinical modeling of FLT3 inhibitors in combination with cytarabine and daunorubicin demonstrating antagonism when the FLT3 inhibitor was used prior to the conventional therapy.(33) There may be a need for a wash out period prior to the use of cell cycle dependent salvage or even consolidative treatments with the concomitant use of sorafenib.
The targeting of signal transduction pathways therapeutically has yet to be broadly successful. Even attempts to target a pathway thought to be as tissue specific as mutated FLT3 in AML, has proven to be more complicated than many first appreciated. For example, the individual type of mutation is certainly critical as preclinical studies suggest that patients with a D835Y mutation in FLT3 are unlikely to be sensitive to some FLT3 inhibitors such as sorafenib.(11) Also, there is evidence that allelic burden of FLT3-ITD is important for ex vivo sensitivity of primary leukemia blasts to FLT3 inhibition, and perhaps those with high allelic ratio may be a subset that benefits the most from FLT3 targeted therapy. (34)(35) Additionally, the clinical activity of targeted agents can be influenced by protein binding and drug-drug interactions. (32)(33) Our study, like others has demonstrated the activity of metabolites of the primary agent, may in fact, play a major role in an agent's biologic activity. (16) Finally, the disease state must also be factored into the equation as targeting mutated pathways at the time of minimal residual disease such as post induction, or following a stem cell transplant might have the best opportunity to suppress the leukemic clone long term.(13) Taken together, future clinical studies of targeted agents must include biologic correlatives if we hope to fulfill the hope that the new agents can impact clinical outcomes in a more discriminate way.
Figure 1. Sorafenib N-oxide is a potent FLT3 inhibitor
A. Standard curve generated as described previously (17), from western blot of TF-ITD cells in plasma exposed for one hour to increasing concentrations of sorafenib. B. Standard curve generated with sorafenib N-oxide in plasma.
Figure 2. PIA results for patients receiving sorafenib
Plasma was isolated from whole blood samples obtained from patients receiving increasing doses of sorafenib on the clinical trial. Samples were obtained immediately prior to dosing on Days 1, 8, and 15 and 29 of each cycle. Dose levels 1, 2, and 3 correspond to total daily doses of 800, 800, and 1200 mg, respectively (see Table 1). Shown are the results from representative patients on successively higher dose levels using the PIA assay on TF-ITD cells for phosphorylated FLT3 (left) and ERK (right). For dose level 2 and 3, extra time points on Day 1 show complete silencing of FLT3 activity within 2 hours of the first dose with maintenance of this inhibition throughout the treatment cycle. Vertical lines have been inserted to indicate a repositioned gel lane. In parallel, plasma from the same time points were examined in PIA assays for assessment of FLT3 and ERK inhibition potential. On the × axis the results of the pharmacokinetics are plotted for sorafenib. On the y axis, the degree of FLT3 inhibition, as assessed by PIA, is plotted as a percent of control. This data is overlaid by the standard curve for sorafenib in plasma as generated in TF-ITD cells(solid line, see Figure 1A). B. PIA results, as described in panel A, of P-ERK compared to standard curve for sorafenib inhibition of P-ERK(solid line). C. The PIA assay values for FLT3 inhibition were replotted after adjusting the "effective" sorafenib concentrations by adding the amount sorafenib N-oxide multiplied by its potency factor using the equation: Adjusted sorafenib concentration=Sorafenib + (Sorafenib N-Oxide*14.59). D. The same experimental data generated from analysis of P-Erk and corrected for sorafenib N-oxide as described in panel C. | 2016-05-12T22:15:10.714Z | 2010-06-10T00:00:00.000 | {
"year": 2010,
"sha1": "227109abeed602ee6555b25601c5b28466fcfcdc",
"oa_license": null,
"oa_url": "https://www.nature.com/articles/leu2010132.pdf",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "227109abeed602ee6555b25601c5b28466fcfcdc",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
269523283 | pes2o/s2orc | v3-fos-license | Comparison of clinical efficacy between robotic-laparoscopic excision and traditional laparoscopy for rectal cancer
Abstract Backgrounds: Laparoscopic surgery, robot-assisted surgery and open surgery are the most commonly consumed surgical techniques in daily living. Considering that in recent years, the situation of choosing laparoscopic surgery and robot-assisted surgery to treat rectal cancer in China is prosperous. Meanwhile, researches lacked in the comparison part between the 2, so we will systematically compare the clinical efficacy of robot-assisted resection and traditional laparoscopic resection for rectal cancer. Methods and analysis: We will search Clinical research literature published before January 2020 in PubMed, Embase, the Cochrane library, Science Network, Wan Fang database, Chinese national knowledge infrastructure, and Chinese biomedicine that evaluate the correlation of rectal cancer with Leonardo's robot and traditional laparoscopy, from inception to July 2019. Weighted mean difference and odds ratio were used to compare the efficacy of robot-assisted resection versus conventional laparoscopic resection for rectal cancer, and the main indicators are operation time, complication rate, conversion rate, blood loss, and length of stay. Results and conclusion: This study will systematically evaluate the clinical efficacy of robot-assisted resection and traditional laparoscopic resection for rectal cancer, thus providing evidence to the clinical application. The results will be published in a peer-reviewed journal. Ethics and dissemination: No ethical approval and participant consent are required, since this study data is based on published literature. The results of the study will be submitted to a peer-reviewed journal. PROSPERO registration number: CRD42020172161
Introduction
Rectal cancer, the morbidity and mortality in our country are increasing year by year. Affected by diet and environment, the number of participants with rectal cancer has been high in China. Every year, a large number of participants with rectal cancer receive surgical treatment in the hospital. Moreover, the disease is also a major cause of death as well as a major public health problem in China, second only to lung cancer and gastric cancer participants, with an estimated 191,000 deaths in 2015. [1] With the development and progress of the society, the demand for accurate treatment and minimally invasive surgery for tumors is on the increase.
Since the 1980s, minimally invasive techniques, represented by laparoscopy, have been developed and widely used due to their good clinical performance. Many data and records have proved that laparoscopic surgery is obviously superior to traditional open surgery in the effect, showing the advantages of less trauma and faster postoperative recovery. Many participants with rectal cancer also use laparoscopic surgery as their first choice for surgery. However, there are also some limitations, such as 2-dimensional plane imaging of the surgical field, and the increasing demands of counterintuitive reverse instrumentation. Last but not the least, the movement degree of freedom of the instrument is less, so it is difficult to complete fine separation, suture, anastomosis, and other operations. [2] The Leonardo's robot surgical system is an advanced robotic platform that uses minimally invasive methods to perform This work was supported by the clinical study of early ERCP ± EST therapy for predicting moderate and severe acute biliary pancreatitis (18PJ506). complex surgery, radical prostatectomy, and a variety of gynecological procedures that currently make up the vast majority of robotic surgery. Since the robot system was first used for the treatment of rectal cancer in 2001, the robot system has been described as an effective tool for precise tissue separation and easier internal suture, and has been highly praised by surgeons. [3] The robotic system allows the surgeon to operate without direct contact with the participant through 3D vision system (high definition, 10 times the magnification of visual stable camera) and operation control of action calibration system, Its flexible telescopic motion reduces physiological tremors, provides superior dexterity, and increases ergonomic comfort, perfectly addressing some of the drawbacks of traditional laparoscopic surgery. [4] At present, more and more hospitals in China have begun to use the da Vinci surgical system for the treatment of rectal cancer. However, there is a major disadvantage of robot system in China, which is the high cost. Therefore, there are still a large number of hospitals in China that have not introduced robot system, and adopt traditional laparoscopic surgery instead. Consequently, there are a limited number of reports documenting the clinical outcomes of robotic versus laparoscopic surgery.
The purpose of our study was to conduct a systematic review and meta-analysis of robot-assisted and laparoscopic hepatectomy.
Since the rise and popularity of robotic systems from 2001 to 2010 was relatively short, the sample size of many previous metaanalyses was relatively small, which limited its statistical ability. So we took some studies published after 2014 as a reference, [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] including more records and samples, which can significantly improve the sample size and statistical power of meta-analysis. Finally, it will be clear which approach is more beneficial to participants in terms of treatment effect and treatment cost.
Data sources and search strategy
This protocol is conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocol (PRISMA-P) statement guidelines. The search terms will include "Leonardo's robot", "rectal cancer", "laparoscopy", "excision", "minimally invasive surgery". No language exclusions will be applied. The references of the identified studies will be manually searched
Study selection and inclusion criteria
We will include participants with any of the following conditions: rectal cancer (carcinoma of rectum or colorectal cancer) participants, we will include systematic reviews and metaanalyses that investigate the clinical efficacy between roboticlaparoscopic excision and traditional laparoscopy for rectal cancer. We will include any type of rectal cancer as experimental intervention (robotic-laparoscopic excision is used as adjunctive therapy while traditional laparoscopy as controls.).
Study design and outcomes
Clinical RCTs will be considered eligible for our study. Articles will be excluded from the current meta-analysis if they are duplicate articles, cohort studies, retrospective studies, case reports, letters, editorials, conference abstracts, or animal experimental studies. The primary outcome will be operation time, complication rate, conversion rate, blood loss, and length of stay.
Data extraction
Data extraction will include characteristics of systematic reviews and meta analyses (first author, publication year, the number of trials included, the number of participants in each meta-analysis, and methods used for pooled analysis), the interventions they received (name, dose, frequency, and the total duration of treatment), the monitoring for efficacy or adherence, and the measure of outcome (specifically defined as event or measure and time frame for the ascertainment of this outcome). For studies with more than 1 follow-up period, we will select the longest. In the event of missing data, we will attempt to contact the corresponding authors for details.
Assessment of methodologic quality
Two reviewers will independently appraise the methodological quality of the included studies using the Cochrane Collaboration tool for assessing the risk of bias. Each study will be reviewed and scored as having a high, low, or unclear risk of bias according to the following domains: selection bias (random sequence generation and allocation concealment), performance bias (blinding of participants and personnel), detection bias (blinding of outcome assessments), attrition bias (incomplete outcome data), reporting bias (selective reporting), and other bias (other sources of bias). The corresponding author would arbitrat any discrepancies between the findings of the reviewers.
Discussion
In this study, we evaluated the clinical efficacy of robot-assisted and conventional laparoscopic resection of rectal cancer. In general, we found that there was no significant difference in the safety of the 2 surgical methods. Except for the large number of low rectal tumors in the robot group, there was no statistically significant difference in surgical indicators among participants. [5] However, the time of robot-assisted surgery is significantly higher than that of traditional laparoscopic surgery, the use of anesthetic painkillers is higher, and intestinal obstruction is significantly higher, which also increases the risk of complications of robot-assisted surgery. However, since the emergence of laparoscopic surgery is earlier than the manual assisted system, and surgeons have more time and experience to use it, it is not clear whether this finding is directly related to the early experience of surgeons and the fixed docking, start-up and disassembly time of robots. We plan to reassess the results in the future as we gain more experience and record more data on whether the time of robot-assisted surgery will be significantly reduced or even lower than that of traditional laparoscopic surgery. [20,21] Conversion rate is one of the important parameters for the feasibility of this minimally invasive technique. Our metaanalysis showed the same results as the previous meta-analysis, that the conversion rate of robot-assisted surgery in rectal cancer resection was lower than that of traditional laparoscopic surgery. Our study shows that participants undergoing robot-assisted surgery generally spend more time in hospital than participants undergoing laparoscopic surgery. There was no significant difference between the long-term assessment indicators, including the overall survival rate and the 2year local recurrence rate. In none of our studies did we report details of participants who lost follow-up. Future studies should focus on long-term follow-up and evaluation of long-term outcomes of the da Vinci surgical system in participants with rectal cancer. Robot assisted surgery of rectal cancer participants need higher required upfront costs than traditional laparoscopic surgery participants. But the study of the evaluation results should be based on long-term treatment of participants with tumor and function. As a result, the overall cost efficiency has not yet been determined, and longer follow-up data is needed for further study. As a result, the overall cost efficiency has not yet been determined, and longer follow-up data is needed for further study.
Author contributions
Zhen Chen and ZhuoLi Zhu conceived the idea for this study; Zhen Chen, ZhuoLi Zhu and Pingxi Wang designed the metaanalysis; Zhen Chen and Fanwei Zeng provided statistical advice and input; Zhen Chen drafted the protocol. Pingxi Wang and ZhuoLi Zhu reviewed the protocol and provided critical feedback. All authors approved the article in its final form. | 2020-07-15T06:00:57.092Z | 2020-01-01T00:00:00.000 | {
"year": 2020,
"sha1": "7094ac9e49cea5d0360e74e6a96b47951c9fccac",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "7094ac9e49cea5d0360e74e6a96b47951c9fccac",
"s2fieldsofstudy": [
"Medicine",
"Engineering"
],
"extfieldsofstudy": [
"Medicine"
]
} |
269956319 | pes2o/s2orc | v3-fos-license | Nonlinear geometric decomposition of airfoils into the thickness and camber contributions
In the thin airfoil theory, the camber line and the thickness distribution of general airfoils are mainly extracted by a linear combination of the upper and lower surfaces, giving rise to geometric distortions at the leading edge. Furthermore, despite the recent effort to obtain analytic expressions for the zero-lift angle of attack and quarter-chord moment coefficient, more analytic generalization is needed for the camber line component in the trigonometric series coefficients. This paper presents a straightforward algorithm to extract the camber line and thickness distribution of general airfoil shapes based on a finite differences method and the Bézier curve fitting. Integrals in the thin airfoil theory involving a Bernstein basis are performed, leading to series coefficients involving Gegenbauer polynomials. The algorithm is validated against analytical expressions of the NACA airfoils without introducing or adapting geometric parameters, and the results demonstrate good accuracy. In addition, the present work indicated a significantly different geometric behavior for the SD7003 airfoil camber slope at the leading edge obtained by the proposed algorithm and the classical linear approximation. Moreover, the method can be coupled conveniently in recent reduced-order models established on the thin airfoil theory to obtain expressions in closed forms for general airfoils.
I. Introduction
A lthough the advancement of computational fluid dynamics (CFD) in the last decades has permitted numerical modeling of complex aerodynamic designs, there is still room for analytical theories and simplified models to shed light on aerodynamic physics.Applying the classical first-order small-disturbance approximation for a thin wing leads to a linear equation for the boundary condition, which can be solved separately as mean-camber line and thickness problems [1].As revealed by Prandtl, the main aerodynamic features of wings with infinite span -the airfoilare significant to aerodynamic investigations of finite-span wings, establishing the relevance of two-dimensional flow analyses around airfoils [2].In this sense, Laplace's equation solutions for inviscid, incompressible, and irrotational flow problems are extensively employed, especially in the study of lifting surfaces [3].
The results obtained in CFD calculations are valid as long as the governing equations and boundary conditions are correctly formulated and implemented [4].In potential flow problems, the linear distribution of singular solutions as sources and vortices on the problem boundaries is an elementary approach.Earlier steady low-speed aerodynamic models have one of its classical thin airfoil theories developed by Glauert [5].By assuming a small-disturbance problem, the model distributes a continuous vortex sheet along the chord and then calculates its vorticity using the no-penetration boundary condition at the camber line.As the solution is not unique, an additional physical condition -known as the Kutta condition -is required [6].Expanding the airfoil vorticity in a trigonometric series gives the relation between the camber line and the main aerodynamic characteristics in Glauert's model.
Despite the proper establishment of steady aerodynamics, the unsteady phenomena represent a branch that still requires substantial investigation compared to the steady one.Examples are the flapping wings of insects and birds and the revolving blades of helicopters and wind turbines [7].Since then, numerous authors in the past century have developed analytical and computational approaches to model and predict unsteady effects, especially for two-dimensional potential flows.Wagner [8] obtained lift and circulatory responses due to a step change in the angle of attack of a flat plate by using the complex analysis of conformal mappings, while Theodorsen [9] solved for small-amplitude harmonic oscillations.Rosenhead [10] introduced the method of approximating the continuous wake's vortex sheet by an array of discrete vortices.Afterward, Katz and Weihs [11] improved Glauert's model by implementing the time-stepping method and discrete vortex shedding for unsteady thin airfoils, and Ramesh et al. [12] have enhanced this model by assuming an intermittent leading-edge vortex (LEV) shedding governed by the maximum instantaneous suction peak.
Recently, an approach approximates the leading edge as a parabola and uses pressure distribution from potential-flow solutions to identify inflow quantities such as airspeed and angle of attack [13,14].
Mostly, these reduced-order methods are established on the thin airfoil theory and obtain the camber line through the airfoil's surface set of points.In this regard, two perspectives exist for the airfoil section construction through the thickness distribution and camber line.While the British convention plots the thickness distribution over the camber line perpendicular to the chord, the American convention plots the distribution perpendicular to the camber slope [15].
The difference between the methods is assumed to be small as traditional airfoils have a maximum camber, usually in the range of 0% (symmetrical sections) to 5%.Nevertheless, much higher camber slopes are used in turbine blades.
Using cubic B-spline curves, Li et al. [16] applied mathematical morphology to reconstruct the blade section from a point-sampled turbine surface.Thus, introduced the method to extract the mean-camber curve, in which the process becomes a nonlinear least-squares problem.
Several applications of curve parametrization, such as the cubic B-spline, were developed for shape design and aerodynamics optimization.Olha et al. [17] constructed a naturally parameterized curve with cubic curvature using an optimization algorithm to solve nonlinear integral equations in aerodynamic problems.Boehm [18] discussed the errors in Bézier and cubic spline presentations of 4-digit NACA airfoils, and Jaiswal [19] studied Bezier curve approximation of airfoils using fourth-order, sixth-order, and eighth-order Bezier curves, where it was solved the linear least squares parametric functional problem for the control points and the nonlinear least squares problem for the nearest points.Ribeiro, Awruch, and Gomes [20] employed genetic algorithms and artificial neural networks coupled with Bézier curves in airfoil optimization for wind turbines.In contrast, Lepine et al. [21] investigate the performance of a nonuniform rational B-spline (NURBS) geometrical representation for the aerodynamic design of wings.
Regarding direct methods of resolving an airfoil as lifting and thickness problems, Barger [22] presented an adaptation of the Theodorsen theory of wing sections of arbitrary shape [23].While Theodorsen uses thin airfoil approximations, Barger exploits the full potential theory to represent the airfoil as a combination of lifting line and thickness distribution.Their model applies a complex analysis of conformal mappings to transform the airfoil potential field into the potential field around a circle.Recently, Cibin and Ranjith [24] have presented analytical expressions of aerodynamic characteristics for arbitrary wing shapes.The airfoil's camber line is obtained as a linear combination of the upper and lower surfaces and is represented in Bernstein polynomials to generate closed-form expressions for the zero-lift angle and the moment coefficient in terms of the beta function.
Despite the precision of the thin airfoil theory in simulating conventional shapes, the thickness effect plays a role in the aerodynamic features.Motta, Guardone, and Quaranta proposed modifying the Theodorsen model to accurately predict the effects of the airfoil thickness on unsteady loads [25].Typically, the classical zero-thickness consideration gives good accuracy for airfoil thickness ratios up to 12% [26].The recent research effort to comprehend the aerodynamics of flapping wings is related to the development of micro-aerial vehicles and power extraction devices, in which thicker shapes generate higher thrust and propulsive efficiency for plunging airfoils [27].Nevertheless, experimental data for plunging maneuvers are historically scarce, which leads to adapting well-known pitching results and methods for predicting flapping airfoil responses [28], where a fundamental concept is the airfoil's effective angle of attack [29].
The proper analysis of the nonlinear geometric decomposition of airfoils into thickness and camber contributions for the thin airfoil theory is deficient in the literature.Numerical computations are focused mainly on obtaining the zero-lift angle of attack and the quarter-chord moment coefficient by a linear approximation and do not present any analytical expressions for the generalized trigonometric series coefficients.The present work aims to develop a straightforward method to extract the camber line and the thickness distribution of general airfoils without resorting to small-angle assumptions.Integrals in the classical thin airfoil theory involving the Bernstein basis are performed, giving rise to series coefficients of Gegenbauer polynomials.Although static aerodynamic validations are conducted, the proposed method is easily coupled to the Fourier series coefficients of many recent unsteady aerodynamic models [30][31][32][33][34][35][36][37][38][39][40].
II. Methodology
This Section is structured as follows.In Section II.A, the first-order small-disturbance approximation is applied to derive the lifting and thickness problems.Section II.B is dedicated to defining the Bernstein polynomials and Bézier curve fitting.In Section II.C, the algorithm based on the finite differences method and Bézier curves is introduced, and its implementation into the thin airfoil theory is explored.
A. Thin airfoil theory
The airfoil's upper and lower surfaces complex functions, and respectively, are prescribed for a given thickness distribution () applied perpendicular to the camber line () = + (), ∈ [0, ], as illustrated in Fig. 1: where is the airfoil chord and () = tan −1 ′ ().Furthermore, Eq. ( 1) can be written as Applying the classical first-order small-disturbance approximation leads to a linear equation for the boundary condition.Considering the two-dimensional airfoil problem, the boundary condition can be transferred to the -axis: where (, ) is the velocity potential, is the freestream velocity, and is the geometric angle of attack.One can solve both lifting and thickness problems through the linear boundary condition (4) by distributing singularity elements in the -axis as detailed next.
Lifting problem
Assuming a vorticity distribution () over the -axis (see Fig. 2), the resulting velocity potential is given by and the boundary condition (4) considering the lifting problem becomes The Glauert variable transformation [5] is appropriate, where and are related as and the vorticity distribution over the airfoil is taken as a trigonometric series: where the coefficients 0 , 1 , . . ., are given by By making use of the Bernoulli equation [3], the pressure coefficient is To obtain the airfoil's lift coefficient, one must integrate the pressure difference over the chord: From Eqs. ( 9), ( 10) and ( 12), the angle of zero-lift 0 is given by The moment coefficient is where From Eqs. ( 10) and ( 14), the quarter-chord moment coefficient is
Thickness problem
The thickness problem is modeled by a source distribution () as shown in Fig. 3.The velocity potential is given by Hence, the induced velocity at = 0 is where is the -component of the induced velocity .The boundary condition (4) leads to the source strength formula From Bernoulli's equation and Eq. ( 17) the pressure coefficient is
B. Bézier curve fitting
Exploring the Bernstein polynomials in describing complex geometries was first made by Casteljau and Bézier in the 1960s [41].Although the slow convergence in approximating continuous functions, Bernstein polynomials found an application in automotive designing.Currently, the so-called Bézier curves often arise in wing design and aerodynamic optimization, and their main properties are detailed next.
Bernstein polynomials
The Bézier curves are described by a sum of control points weighted by a Bernstein basis of degree and parameter ∈ [0, 1], defined by Applying the binomial theorem on (1 − ) − gives Therefore, the Bézier curve z() is obtained by the Bernstein basis as: where ∈ C, = 0, 1, . . ., , are the control points, and the coordinates x and ỹ are the real and imaginary parts of z(), respectively.
Curve fitting
To find the curve that best fits a given set of points = [ 1 , 2 , . . ., ] , the following optimization problem should be solved: To solve the linear problem for , an initial set of nodes is required.In this way, the initial guess (0) = (−1)/(−1), = 1, 2, . . ., , can be taken.Hence, the vector is calculated as where ★ = −1 is the Moore-Penrose pseudo-inverse of .Therefore, one can apply Newton's method to solve (26).As the Hessian () of the function is a diagonal matrix, the method proceeds through the iterations where "⊘" is the Hadamard division (element-wise division).The process restarts with the new control points ( +1) obtained from Eq. ( 27) until ()/ ≤ , where is a suitable error criterion.
C. Finite difference method based on Bézier curves (FDM-B)
Bézier curve shapes the airfoil surface from the given discrete points to proceed with the finite difference method, which starts with an initial guess = + , = 1, 2, . . ., , for the camber line as depicted in Fig.
The FDM-B algorithm
Bézier curve fits the surface through the approach described in Section II.B.2 and the set of parameters = [ 1 , 2 , . . ., ] is obtained.The parameter , 0 ≤ 1 < < ≤ 1, related to the leading edge is obtained by solving the root-find problem using a Newton-Raphson iteration: To constrain the problem for the upper and lower surfaces, the change of variable is taken: where = ( + )/( − ).In this way, we have ( , ) = ( 1 , ) for the upper surface and ( , ) = ( , ) for the lower surface.Considering the camber line with inner points, we have Δ = 1/( + 1) and = Δ, = 1, 2, . . ., .Hence, an initial guess (0) for the set of points is taken.The normal line at is given by where a second-order central finite difference is used to approximate the derivative, and its intersections z, with the airfoil surfaces are obtained solving the root-finding problem: and the parameters and are obtained from the transform (30).
The proposed method assumes (0) = () = 0 and it is taken into account in Eq. (32).Moreover, the camber slopes at the airfoil edges are calculated from the FDM-B algorithm as From the geometry of the problem shown in Fig. 4 the following root-find problem can be solved: which enforces the camber line to lay at the midpoint of each vector z − z .In this way, the Jacobian ( ) of ( 38) is the diagonal matrix given by where x′ , = ℜ( z′ ( , )) and ỹ′ , = ℑ( z′ ( , )), and the camber-line inner points are adjusted at each iteration: until the stopping criteria || ( )||/ ≤ ε is achieved.Thus, we have the discrete points ( ) ≈ = ( ), and the thickness distribution as One can take the generalized thickness distribution function as where , = 0, 1, . . ., , are shape coefficients -which can be calculated by the Moore-Penrose pseudo-inverse method.The radius of curvature at the leading edge is given by
Implementation in thin airfoil theory
The FDM-B algorithm allows obtaining the discrete points of the camber line and the camber slope (the latter from Eq. ( 32)) of general airfoil sections.In this way, the Bézier curve fitting provides an analytic expression for the camber-line slope.Thus, from Eq. ( 22) we have: where the parameter is defined from the leading edge to the trailing edge i.e., to .From Eqs. ( 21), ( 30) and (44) and setting ( , ) = ( , ), we have: From Eqs. ( 10) and ( 45), the following general integral must be solved (see Appendix B) in order to couple the FDM-B method into the thin airfoil theory: where () is given by Eq. (C27).Therefore, the Fourier coefficients (Eqs.( 9) and ( 10)) become: where are the camber slope coefficients.Thus, different from an infinite Fourier series -which needs truncation of termsthe proposed method uses the exact terms for the aerodynamic contribution of the camber line.From Eq. ( 13) the angle of zero-lift is and from Eq. ( 14) the quarter-chord moment coefficient is Furthermore, the pressure coefficient (11) becomes Regarding the aerodynamic analysis of the thickness problem, the pressure coefficient ( 19) is obtained from Eqs. ( 17), ( 18) and ( 42) as (see Appendix A): where and are related through Eq. ( 7).
Thus, the upper and lower surface pressure coefficients in the first-order approximation problem are the linear superimposition of the lifting and thickness contributions ( 52) and ( 53).The complete methodology is summarized in the flowchart presented in Fig. 5, and its subprocesses are illustrated in Fig. 6.
Yes
No
III. Results & Discussion
During the 1930s, the National Advisory Committee for Aeronautics (NACA) investigated a group of airfoils in the NACA variable-density wind tunnel [42].The related airfoils were developed by combining a thickness distribution and the camber line where the camber-line slope is given by for NACA 4-digit airfoils with the nomenclature "NACA ", where = /100 is the maximum camber, = /10 is the position of the maximum camber, and = /100 is the maximum thickness.In this way, the referred airfoils become practical in the present analysis as they have closed-form expressions.Thereafter, numerical computations are performed to validate the FDM-B algorithm against analytical results for NACA 4-digit airfoils.
Theoretically, the first-order linear approximation gives good accuracy for airfoil thickness ratios lesser than 12% [26].As the approximation holds as thinner the airfoil is, the present analysis is focused on airfoil thicknesses close to 12%.Then, Bézier curves of degree = 16 fitted the surface of the NACA 4409 and 2412 airfoils, containing 60 points, with a stopping criterion = 2 × 10 −9 .The FDM-B algorithm is employed with = 30 and ε = 1 × 10 −10 for obtaining the main geometric parameters.As shown in Fig. 7, the proposed method demonstrated good accuracy in predicting the camber line, whereas the linear approximation deviates considerably from the exact equation (55), especially in the first quarter of the chord.A reasonable error is found even for small maximum camber airfoils, as in the NACA 2412 airfoil (Fig. 7b), where (0)/ = 1.52 × 10 −3 in the linear approximation, leading to distortions of the effective angle of attack [22].For the NACA 4409 airfoil, the linear approximation gives (0)/ = 1.66 × 10 −3 .
One should notice that the proposed algorithm imposes the camber line at the leading and trailing edges to be zero i.e., (0) = () = 0. Overall, the FDM-B algorithm kept the relative errors under 0.5% for the camber line.
After obtaining the camber-line slopes at the inner points, second-order accuracy forward and backward difference approximations give the first derivatives ′ (0) and ′ (), respectively, as detailed in Eqs.(36) and (37).Fig. 8 compares the camber slope obtained by the FDM-B algorithm and the linear approximation.As in the camber line analysis, the linear approximation contrasts with the exact formula (56) in the quarter-chord region.The comparison of the camber slope at the leading edge obtained from both methods is presented in Table 1, where the linear approximation gives Although satisfactory accuracy when predicting the thickness distribution along the chord (see Fig. 9), the linear approximation fails at the leading edge, giving (0)/ = 1.66 × 10 3 and (0)/ = 1.52 × 10 3 for NACA 4409 and 2412 airfoils, respectively.As the FDM-B algorithm fits the inner points, special attention must be paid to the airfoil edges.In the present analysis, Eq. ( 42) imposes the leading edge to be zero, whereas no assumption is made to the trailing edge.Assuming the generalized thickness distribution function (42) with = 4 (as in the NACA thickness formula (54)), one can get the shape coefficients , = 0, 1, . . ., 4, by the Moore-Penrose pseudo-inverse method.In this way, the FDM-B algorithm predicts the trailing-edge thickness with 2.9% and 1.5% errors for NACA 4409 and 2412 airfoils, respectively.One should add the equation () = to the process of obtaining the shape coefficients to impose a specific thickness at the trailing edge.Furthermore, the radius of curvature at the leading edge (from Eq. ( 43)) is presented in Table 1.Finally, Fig. 10 shows the Bézier fitting and the resulting composition of the geometric outcomes from the FDM-B algorithm by making use of Eq. ( 1), demonstrating good accuracy when compared against the airfoil data.Even though many more geometric parameter validations would be carried out, the radius of curvature and camber-line slope at the leading edge find applications in recent aerodynamic analysis, as the parabola-airfoil matching [13,14,39].In addition to the geometric verification against analytical results, aerodynamic validations were performed considering the FDM-B algorithm implemented into the thin airfoil theory, as described in Section II.C.2.While Joseph and Mohan [24] proposed analytical expressions for two main aerodynamic characteristics in terms of the beta function, the present work expresses the Fourier coefficients ( 9) and (10) as Gegenbauer polynomials [43].In this way, not only closed-form expressions for the zero-lift angle of attack and the quarter-chord moment coefficient are obtained as for any aerodynamic quantities based on the thin airfoil theory.The analysis method compares the FDM-B algorithm and the linear approximation results against analytic functions obtained from the integration of Eqs. ( 13) and ( 15) by directly using the NACA camber-line slope (56) and the variable transformation (7).Thus, the zero-lift angle of attack and the approximation predicts the camber-line slope at the leading edge 68% greater than the FDM-B algorithm.Furthermore, the linear approximation generates a monotonically non-increasing camber slope -similar to the NACA airfoils.In contrast, the FDM-B algorithm predicts a non-monotonic function, with a leading edge slope ′ (0)/ = 0.092 and a peak of 0.115 at / = 0.027.This behavior in the camber slope leads to a concave shape in the camber line leading edge, whereas usual airfoils have a convex contour.Bézier curves of degree = 16 fitted the camber line and camber-line slope points given by the FDM-B algorithm and the linear approximation, with stopping criterion = 2 × 10 −9 .For the thickness fitting, the Moore-Penrose pseudo-inverse method is applied to obtain the shape coefficients, considering = 10 and () = 0. Thus, a composition of these geometric fitted functions is applied in Eq. ( 1) to reconstruct the SD7003 airfoil surface, as depicted in Fig. 13.Despite the good accuracy of the proposed method, one can notice a slight discrepancy in the leading edge composition.Indeed, the FDM-B algorithm is based on a uniform grid, and the edges may need particular chord discretization as in the SD7003 airfoil case, once the camber-line slope presents high gradients at the leading edge.
As illustrated, the linear approximation underestimates the camber slope coefficients when compared to the FDM-B algorithm.The difference between the coefficients obtained from both methods increases from = 0 to = 5, decreasing to nearly zero for = 16, affecting especially the zero-lift angle of attack and the quarter-chord moment coefficient (Eqs.( 50) and ( 51)), as presented in Table 3.The geometric characteristics such as maximum camber and thickness obtained by both methods are also in Table 3 and show excellent agreement with Selig [44].
Airfoil data Bézier fitting FDM-B and Eq. ( 1) Fig. 13 Fitted surfaces of the SD7003 airfoil.Concerning the zero-lift angle of attack and quarter-chord moment coefficient, no detailed experimental data is found in [44] for the SD7003 airfoil.In this way, the XFOIL code [45] is employed for the inviscid aerodynamic analysis.Considering the pressure distribution coefficient as the composition of Eqs. ( 11) and ( 19), Fig. 14 compares the results from both methods against the obtained by the XFOIL.Both methods demonstrate a similar performance in evaluating the pressure distribution along the chord, where the leading-edge pressure tends to be infinity -as expected in the thin airfoil theory.Yet, the results at the lower and the aft part of the upper surfaces are in excellent agreement with the XFOIL outcomes.As no analytic expressions are available for the SD7003 geometric composition, the FDM-B algorithm is a more suitable alternative than the linear approximation to obtain the main aerodynamic features based on the thin airfoil theory.On this wise, a second-order subsonic airfoil theory [46] may be applied to investigate the accuracy of the proposed method at the leading edge.Furthermore, the zero-lift angle and moment coefficient obtained from the XFOIL code are −1.737• and −0.0413, respectively.Thus, the relative errors of the FDM-B algorithm and linear approximation are 1.2% and 1.6% for the zero-lift angle and 3.6% and 4.6% for the moment coefficient, respectively.However, one should point out that the XFOIL code is founded on the panel method, whereas the proposed method uses the thin airfoil theory approximations.
IV. Conclusion
A straightforward method was implemented to obtain the camber line and the thickness distribution of general airfoils.The model is based on a finite differences method coupled with Bézier curve fitting without resorting to small-angle assumptions or additional geometric parameters.Thus, integrals in the aerodynamic thin airfoil theory involving the Bernstein basis are performed, giving rise to series coefficients of Gegenbauer polynomials.The FDM-B
Fig. 2
Fig. 2 Induced velocity at a point (, ) by the vortex sheet.
Fig. 3
Fig. 3 Induced velocity -and its Cartesian components -at the point (, ) by the source sheet.
Fig. 5
Fig. 5 Flowchart of the nonlinear geometric decomposition method.
Fig. 7 Fig. 8
Fig. 7 Comparison of the camber line obtained by the linear approximation and the FDM-B algorithm.
Fig. 9
Fig. 9 Comparison of the thickness distribution obtained by the linear approximation and the FDM-B algorithm.
Fig. 11 Fig. 12
Fig. 11 Comparison of the camber line and the thickness distribution obtained by the linear approximation and the FDM-B algorithm for the SD7003 airfoil.
Fig. 14
Fig. 14 Comparison of the pressure distribution obtained by the linear approximation, the FDM-B algorithm, and the XFOIL code. | 2024-05-23T15:10:13.559Z | 2024-05-21T00:00:00.000 | {
"year": 2024,
"sha1": "2042fd6b29ea4de4d8e758280c5656af911cb6c4",
"oa_license": "CCBY",
"oa_url": "https://www.researchsquare.com/article/rs-3473763/latest.pdf",
"oa_status": "GREEN",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "15a0f57bd48e5f05ccdbb06b2681ee9543995838",
"s2fieldsofstudy": [
"Engineering",
"Physics",
"Mathematics"
],
"extfieldsofstudy": []
} |
114497666 | pes2o/s2orc | v3-fos-license | Hydraulic Phenomena Frequency Signature of Francis Turbines Operating in Part Load Conditions
With the integration of renewable energies into the electricity grid, new requirements have been defined by power station operators. These changes bring new challenges for hydraulic turbine manufacturers, such as more flexibility during machines operation. In recent years, investigations have been focused on off-design conditions, since unsteady phenomena occur far from the classical Francis turbine operating range. This is especially the case at partial load, where dynamic stresses on the runner could impact the machine lifetime. The main objective of the study presented in this paper is to gain a better understanding of Francis turbine partial load flows. Thus, different runners are compared, based on test rig measurement campaigns already realized on model scale turbines. Pressure sensors and strain gauges signals are compared, using Fast Fourier Transform (FFT) analysis and Spatial Harmonic Decomposition (SHD). Finally, an attempt to classify these sets of frequencies is realized, according to their dynamical loading on the structure operating in off- design conditions and their impact on the turbine lifetime.
Introduction
As modern electricity grids include new renewable energy sources, hydroelectric power station operators need to meet new requirements in terms of flexibility and adaptability. The impact on hydraulic turbines manufacturers is to consider extended operating ranges for their machines. To address this request, off-design hydraulic and mechanical conditions are now better investigated. Francis turbines are known for their high efficiency at their Best Efficiency Point (BEP), and the wide range of heads and flow rates they cover worldwide. For more than a decade, the extension of their operating zone has led to the study of the partial load vortex rope [1]. This hydraulic phenomenon causes pressure fluctuations in the machine, mainly in the cone of the draft tube. Most of the technical specifications for Francis turbines still require low level of pressure fluctuations.
Going even deeper at partial load is a new challenge. In such regimes, Francis turbines encounter even more misadapted flows they were not designed for in the past. By increasing the time of operation at partial load, Francis runners are more often confronted to dynamic stresses on the blades, which may lead to lifetime reduction issues [2]. To develop a better understanding of such unsteady loadings and partial load flows, experimental data from reduced scale Francis turbines (called "model" turbine) have been compared. The objective of this study is to describe characteristics of the phenomena occurring at deep part load and their structural impact on the turbine runner.
Experimental Setup
The tests were performed in the GE Hydro Laboratory. Two kinds of instrumentation are considered in the present paper. Unless otherwise indicated, all the models presented thereafter are provided with similar configuration of sensors ( Figure 1). The acquisition frequency is 2.4 kHz.
Firstly the study makes use of classical flush-mounted pressure fluctuation sensors on non-rotating parts of the scale model. Sensors in the vaneless space and cross sections of the draft tube cone will be used for comparisons. Secondly, strain gauges are positioned directly on runner blades. The gauges are positioned at the connections of blades to the crown and the band of the runner. These zones are critical due to the stress concentration occurring in the fillet radii of the junctions when the blades are loaded. These sensors give information of the static and dynamic loading on the blades.
In addition, flush mounted pressure fluctuations sensors on the runner blades are available for one of the designs. Such embedded instrumentation gives access to information on the pressure field inside one rotating interblade channel.
Comparison of several runner designs at the rated head
In this first part, the purpose is to find if there are similarities between fluctuations signals recorded while different turbine models are operating at part load. The frequency content is analyzed by using Fast Fourier Transform (FFT).
Turbine configurations
Several designs are compared. Selected ones are noted here A, B, C, D and E, from low to high-head turbines. Two types of manufacturing process of the model runner are commonly used for such investigation, achieving either only hydraulic similitude with the industrial turbine, or both hydraulic and mechanical similitude.
Operating points
Dimensionless parameters are used to describe hydraulic turbines hill chart [3]: the net head (m), the discharge (l/s), the gravitational acceleration, and the peripheral velocity (m/s) of the runner outlet, which section is (m 2 ).
For each design, the load variation at the rated head and at sigma plant value is performed ( Table 1). At each steady operating point, a FFT analysis of the dynamic signals is performed. For the sets of measures, waterfall diagrams [3] are used, showing the evolution of one sensor spectrum amplitude and frequency along with the discharge coefficient. For each design, the frequency, the discharge coefficient and the amplitude are made dimensionless by dividing respectively by the rotation frequency of the runner , the discharge coefficient at the best efficiency point !" and the maximum spectrum amplitude obtained for the whole load variation. In the next section, waterfall diagrams from different sensors signals are compared, to highlight several frequency signatures of phenomena occurring at deep part load. Figure 2 shows the waterfall diagram of the gauge with the highest levels of dynamic strain for each runner. Since the blades are thinner near the trailing edge junctions, all these strain gauges are positioned between 6% and 12% from the trailing edge in the chordwise direction. When reducing the discharge, large-band frequency signatures can be seen on designs A, B and C strain signals. The shape of this signature was generally correlated with the presence of interblade vortices (IBV). It consists of an increase of the spectrum amplitude on a whole range of frequencies. For design A, the signature is around 0.4 !" , while for the designs B and C, it is around 0.3 !" . For the runners D and E, the region of higher amplitude extends over 30 , it is therefore not clear if the same phenomenon is occurring for these designs.
"
Claw-like" symmetric frequency signature. On pressure sensors located in the vaneless gap between the guide vanes and the runner, pressure fluctuations show a "claw-like" frequency signature on design B and C ( Figure 3). Below 0.35 !" , two sets of several amplitude peaks are developing in opposite frequencies when the flow rate is increasing. Frequency differences between the peaks of each group seem to remain constant, and peaks of opposite frequency trend with the discharge can be identified. Both sets seem symmetrical compared to a vertical line corresponding to half of the blade passing frequency. All these elements suggest that a coherent phenomenon is causing this claw-like signature.
The design A shows some high amplitudes too, however a lack of discretization in the load variation records prevents from any conclusion on the frequency signature. For the higher head designs D and E, no signature of this type was found. Firstly, on the strains gauges ( Figure 4 on the bottom) two different trends are observed for the higher peak appearing at deep part load. The first one illustrated with the design B consists of a decreasing frequency peak around from 0.15 !" to 0.3 !" . The second trend concerns the design E for which the peak frequency stays rather constant and around 0.2 on the same discharge range than mentioned before. Another interesting peak at 0.8 will be worth mentioning hereinafter.
Concerning draft tube cone pressure fluctuations ( Figure 4 on the top), both designs are different. For design B, deep part load spectrums contain only small amplitudes peaks at lower frequencies, unlike what was on strain gauges signals. An upper part-load resonance can be found around 0.8 !" , but has only a small impact on the runner (see the bottom-left diagram of Figure 4). For design E, a constant frequency peak at about 0.2 is observed from roughly 0.45 !" to 0.15 !" . For design E, it was possible to use the Spatial Harmonic Decomposition [4] on four sensors located in the same horizontal cross section of the draft tube cone. A frequency analysis of the 3 and 3 56 terms, representing respectively in phase pressure pulsation and a pattern with one-nodal diameter rotating in the runner direction, showed that the fluctuations peak at 0.2 in the draft tube was decomposed by the runner strain gauge in two parts, just like it was done for the vortex rope [4]. The 0.8 peak on the embedded strain gauge is linked to the rotating part of this phenomenon, while the 0.2 peak is present on strain signals, without changes in frequency from the stationary to the rotating frame instrumentation. Contrary to the vortex rope for which the rotating component is usually of major amplitude, the in-phase amplitude of the fluctuations ( 3 ) is higher than the rotating one ( 3 56 ) for this 0.2 pulsation.
Frequency insight on design B test rig results
When trying to extend Francis turbines operating range at deep partial load below the part load helical vortex rope, the first phenomenon that the runner will experience, is the one associated with the large band frequency observed previously. This signature has been studied on the design B, since this reduced scale runner is equipped with on-board pressure sensors and strain gauges.
Signal frequency content analysis methodology
Classical representations used to describe fluctuations at partial load are the peak-to-peak values or the root mean square of the signals, as a function of the discharge. For finite energy signals, the signal power can be derived from the root mean square value (assuming an ergodic process):
Frequency content of the dynamic signals for the design B
To see the evolution of the signal frequency content during load variation, the signal spectral power (SSP) is plotted as a function of the relative discharge coefficient (thick black line) and placed next to the corresponding waterfall diagram ( Figure 6). In addition, five grey lines show the spectral power evolution of each frequency range during the load variation. This allows to see at the same time the evolution of the total power and the frequency content of the signal.
This analysis is derived for the gauge with the highest levels of dynamic strain of the design B (Figure 6), which is located at the junction of the trailing edge with the band. When decreasing the discharge, a first local maximum of dynamic loadings is reach at 0.65 !" due to the presence of the part load vortex rope. It is worth mentioning that the majority of the spectral power is contained in the lower range of frequencies R0 ' 1S , which includes the rope frequency. Then the loadings reach their absolute maximum around 0.31 !" , which corresponds to higher level of the large-band frequency signature identified earlier (Figure 2). For this operating point, the major part (roughly 40%) of the signal spectral power is coming from the range of frequencies R7 ' 20S which encloses the largeband frequency signature. On the right of the Figure 6, the runner outlet flow can be visualized, and developed cavitating interblade vortices can be seen near the band for this point of maximum strain of the blades.
The same analysis is applied on the embedded pressure sensors of the runner B (Figure 7). For the inlet sensor, a large-band signature can also be seen at 0.31 !" . It corresponds to a local maximum of pressure fluctuations but, unlike the dynamic strain, the spectral power is lower than the one of the vortex rope regime. Even if there is a small increase of power caused by the range R7 ' 20S , the lower range R0 ' 1S is dominant in this case. For the outlet sensor, no large-band signature was found; although interblade vortices are clearly visible in the runner channels near the sensor location (see Figure 6). Both pressure signals show clearly an upper part load resonance around 0.8 !" but once again no impact is seen on the strain gauge. The great majority of the spectral power comes from the range R1 ' 4S , which encloses this phenomenon frequency.
View of the runner outlet at maximum dynamic loadings &0.31 !" (
Impact on turbine lifetime
The aim of this last section is to estimate the structural impact of dynamic loadings on the runner in deep partial load operating conditions. The dynamic stresses due to the vortex rope will be taken as a reference for the damage at partial load.
Load controlled fatigue testing
Fatigue analysis is based on material characterization. Load controlled fatigue testing are performed on different materials to derive their Wöhler curve (or S-N curve), representing the stress amplitude as a function of the number of cycles to failure. As the stress increases, the allowed number of cycles decreases: with C F the number of cycles for a given probability of survival (usually 95 or 99.9 %), Δ) the dynamic stress amplitude, and Y and Z are the tested material constants. To describe the prototype material behavior, the Wöhler curve that is used is appropriate for welded structure in cast steel. Load controlled fatigue tests were conducted in water, under in-plane bending conditions with non-zero mean stress value.
The Rainflow counting method
To describe the blade loading temporal evolution, the commonly used Rainflow method is applied [5]. With this method, the signal is decomposed into several amplitude ranges and a number of cycles are determined for each range of stress amplitude. Then, the Miner's assumption of linear cumulative damage is used: for each previously mentioned amplitude value, an elementary damage is defined as the ratio of the number of cycle encountered during the signal record and the number of cycles at failure. The damage is finally calculated by adding the elementary ones. This linear rule of cumulative damage has some limitations to reproduce such a complex physics, however it is a reasonable engineering tool used in industrial dimensioning processes.
Generally, the value of 1 is chosen as a limit for the damage during dimensioning steps, which corresponds to a chosen probability of failure. For the full scale turbine, the rate of damage per unit of time can be derived by taking into account the ratio of runner rotation frequencies between the model and the prototype. With this methodology, the damage is calculated at the strain gauge location. The endurance limit was neglected to preserve information on the smaller amplitudes.
Comparing the vortex rope and deep part load phenomena.
For the rest of the analysis, the most sensitive strain gauge of the design B will be considered. Four operating points are analyzed ( Table 2): the best efficiency point (BEP), the part load maximum dynamic loadings in the vortex rope regime (PL), which is taken as a reference for the damage value; the deep part load occurrence of the large-band frequency signature (DPL), and the low variable frequency at quasi-no load (QNL). For these four operating points, the previously exposed fatigue damage method is applied. The results are listed in Table 2.
According to these results, the rate of damage increases during partial load operation. The vortex rope phenomenon has a real impact on the runner lifetime, since the rate of damage is more than 100 times higher than at the best efficiency point where the damage is practically negligible. At deep part load, when the dynamic loadings are at their maximum, the rate of fatigue damage gains roughly two orders of magnitude, while the stress peak-to-peak value is doubled. At lowest flow rate, the rate of damage is still high compared to the vortex rope, but still lower than for the large-band signature 8 which occurs at higher frequency. As an order of magnitude, one hour of operation at 0.31 !" is equivalent to about ninety hours at 0.65 !" . The reason for higher rates of damage at deep part load can be explained by examining the Figure 8. When deriving the Rainflow counting method, the recomposed prototype stress signal is divided in several ranges of peak-to-peak amplitudes (on the horizontal axis). For each range, the number of counted cycles can be read on the vertical axis. The four colors are corresponding to the four operating points previously introduced (Table 2). For smaller stress amplitudes on the left of the green dashed line, the difference in damage between the vortex rope and the deeper partial load regimes is mostly linked to the number of cycles, which is multiplied by roughly 100, due to higher frequencies in the signal. However, the main changes are coming from the higher amplitudes that are not occurring when operating in the vortex rope regime (on the right of the dashed green line). The severe damage at deep partial load is mostly due to these increased amplitudes of the dynamic loadings on the runner blades.
Conclusion
To address Francis turbine partial load issues, a better understanding of the flow inside the machine for this type of operating points is required. Several turbine designs were investigated by comparing signals acquired during test rig measurements campaigns. At partial load, it was possible to find similarities between pressure fluctuations in the flow and dynamic loadings on the runner blades of the different designs. The most complete instrumented runner was studied in detail by building specific tools to characterize the frequency content of partial load signals recorded in both stationary and rotating frames. For this design, the dynamic loadings on the blades reach a maximum when operating at deep partial load. At this operating point, the loadings are correlated with the occurrence of a largeband frequency signature in strain gauges signals, which contains the major part of the signal spectral power. A similar signature was found in signals of a pressure sensor located at the runner inlet. Investigations on the phenomenon causing such signature confirmed the presence of developed interblade vortices at the point of maximum dynamic stresses. Finally, a fatigue analysis based on the Rainflow counting method and Miner's cumulative damage theory was realized. The rate of damage for the point of maximum large-band loadings is roughly two orders of magnitude higher than for the point of maximum vortex rope stresses. The same result was found for deeper part load lower frequency phenomena, with a rate of damage slightly reduced compare to the large-band frequency one. The phenomenon involving this large-band frequency signature is a major blocking factor that must be overcome in order to develop deep partial load operation and further investigations are recommended. | 2019-04-15T13:08:34.281Z | 2016-07-04T00:00:00.000 | {
"year": 2016,
"sha1": "80ee64e3f0a2546fc094de8c0aeb076e1930b3b2",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1755-1315/49/8/082001",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "a9e17a6667a6c475187ffafff229b8f17d78a5af",
"s2fieldsofstudy": [
"Engineering"
],
"extfieldsofstudy": [
"Engineering"
]
} |
232159166 | pes2o/s2orc | v3-fos-license | Remote Measurement in Rheumatoid Arthritis: Qualitative Analysis of Patient Perspectives
Background Rheumatoid arthritis (RA) is characterized by recurrent fluctuations in symptoms such as joint pain, swelling, and stiffness. Remote measurement technologies (RMTs) offer the opportunity to track symptoms continuously and in real time; therefore, they may provide a more accurate picture of RA disease activity as a complement to prescheduled general practitioner appointments. Previous research has shown patient interest in remote symptom tracking in RA and has provided evidence for its clinical validity. However, there is a lack of co-design in the current development of systems, and the features of RMTs that best promote optimal engagement remain unclear. Objective This study represents the first in a series of work that aims to develop a multiparametric RMT system for symptom tracking in RA. The objective of this study is to determine the important outcomes for disease management in patients with RA and how these can be best captured via remote measurement. Methods A total of 9 patients (aged 23-77 years; mean 55.78, SD 17.54) with RA were recruited from King’s College Hospital to participate in two semistructured focus groups. Both focus group discussions were conducted by a facilitator and a lived-experience researcher. The sessions were recorded, transcribed, independently coded, and analyzed for themes. Results Thematic analysis identified a total of four overarching themes: important symptoms and outcomes in RA, management of RA symptoms, views on the current health care system, and views on the use of RMTs in RA. Mobility and pain were key symptoms to consider for symptom tracking as well as symptom triggers. There is a general consensus that the ability to track fluctuations and transmit such data to clinicians would aid in individual symptom management and the effectiveness of clinical care. Suggestions for visually capturing symptom fluctuations in an app were proposed. Conclusions The findings support previous work on the acceptability of RMT with RA disease management and address key outcomes for integration into a remote monitoring system for RA self-management and clinical care. Clear recommendations for RMT design are proposed. Future work will aim to take these recommendations into a user testing phase.
Introduction
Background Rheumatoid arthritis (RA) is an autoimmune disease that affects approximately 1% of the adult population in the United Kingdom [1]. Its primary symptoms are recurrent joint pain, swelling, stiffness, and deformities, contributing to fatigue, reduced ability to function [2], increased prevalence of depression [3], reduced quality of life [4], and premature mortality [5]. Symptoms vary from day to day, and disease progression is unpredictable [6]. Currently, clinical status is monitored by regular clinical appointments at fixed intervals, typically every 6 months. As symptom severity fluctuates between visits, appointments may not capture the critical time points of symptom exacerbation.
Remote measurement technologies (RMTs), including smartphone apps and wearable devices, have recently emerged as useful tools for supporting health management [7][8][9]. Health tracking apps enable patients to actively log changes in symptoms as well as to collect passive data from built-in smartphone sensors or wearable devices. RMT offers opportunities to track symptom severity continuously and in real time, allowing the collection of rich amounts of data in naturalistic settings and overcoming difficulties in exploring symptoms during time-limited appointments [10]. There is growing evidence to support the cost-effectiveness of mobile health interventions. A number of studies have shown positive outcomes, including improved attendance rates at health promotion centers and medication adherence, as well as positive costing outcomes on economic evaluators (eg, a score of 79.6% on the Consolidated Health Economic Evaluation Reporting Standards checklist) [11].
There is a huge appetite for the integration of RMTs into clinical care for RA. A recent systematic review reported the availability of 19 Android or iOS apps for symptom measurement in RA, representing a range of self-reported and passively collected features [12]. Along with this ambition is a growing body of evidence examining engagement with RMT for symptom tracking; the concept of RMT has a good level of face validity, with an estimated 86% of patients with RA reporting a willingness and interest in using apps for symptom monitoring [13].
Despite this ambition, evidence suggests huge variability in engagement with RMT, with adherence levels ranging between 11% and 65%, depending on the requirements for the patients, burden of questionnaire completion, and length of follow-up [14,15]. Barriers for engagement are extensive and specific to the individual; systematic review evidence suggests that perceived clinical value, symptom severity, and convenience are key drivers of uptake [16].
When developed in close collaboration with patients and clinicians, the use of apps in clinical care may positively affect the health outcomes of patients with RA. Co-design involves target end users working with researchers through development, pilot testing, and dissemination [17]. The REMORA (Remote Monitoring of Rheumatoid Arthritis) project, which carefully co-designed a symptom measurement app with patients, has demonstrated excellent levels of patient engagement and has identified temporal changes that might have been previously missed by consultants [18]. In another study, patients with RA who tracked their disease through validated questionnaires and digitally recorded joint counts better adhered to medication, better managed activities of daily living, and reported less worry about the future [19]. Incorporating accelerometer and objective gait balance, alongside symptom reporting, can also accurately predict in-clinic RA activity [20,21].
However, this level of patient input for RMT development and testing is uncommon. Multiple systematic reviews have highlighted the lack of patient involvement in the majority of the apps available, calling for more user experience research to feed into the design and development of app-based symptom measurement systems [12,22]. There is an even greater dearth of patient involvement in the development of platforms to merge subjective symptom reporting with passive data collection [23].
Objectives
Accordingly, the aim of this body of research is to develop an app-based symptom measurement system that fully meets the needs of the patient and clinician users. This study represents the first step in this program, which is to elicit the views of patients to start developing the requirements for a system that captures both subjective and objective symptoms through patient-reported assessments and data collected via passive sensors. The objectives of this project are (1) to identify the symptoms prioritized by patients with RA for inclusion in a multiparametric RMT data collection platform and (2) to identify the key requirements that the platform would need to have for maximized utility, uptake, and long-term engagement with symptom management.
Design and Ethics
This study used qualitative methodology to understand patient views on the aspects of RA health and clinical care which may be most amenable to measurement via digital technologies. Semistructured focus groups explored the key symptoms and key requirements of an RMT platform. A topic guide (Multimedia Appendix 1) was developed based on recent systematic review evidence [16], which provided a loose framework for discussion. The focus groups were comoderated by a service-user researcher, experienced in qualitative research methods (RW), and the lead investigator (FM). Neither of the moderators were involved in patient care.
The study protocol and topic guide were approved by the Office for Research Ethics Committees Northern Ireland (ORECNI, REC number: 17/NI/0179).
Study Participants and Recruitment
The eligibility criteria for inclusion in this study were as follows: (1) clinically verified RA, (2) aged 18 years or older, (3) able to speak English fluently, and (4) able to give informed consent. Patients were considered ineligible if they were unable to physically attend a focus group discussion at King's College Hospital or had major cognitive impairment or dementia.
Eligible patients attending outpatient appointments at the King's College Hospital National Health Service (NHS) Foundation Trust rheumatology service were consecutively approached and invited to participate. Patients were initially approached by a clinical trials practitioner, who was not directly involved in the patients' clinical care. Patients were provided with an information sheet and provided verbal information about the nature of the project. With the patient's consent, their contact details were provided to the lead researcher (FM), who then approached participants separately to discuss the study in detail.
All participants provided written informed consent to participate and were informed that their data would be anonymized and that they were free to withdraw at any time with no consequences to their clinical care.
Data Collection
Two semistructured focus groups were conducted in November 2017 and January 2018. Both were conducted by two facilitators: facilitator 1 (RW), a lived-experience service user, and facilitator 2 (FM), a postdoctoral researcher and health psychologist at King's College London. Each focus group lasted approximately 60 min, and participants also completed brief assessments to establish key demographics such as age, gender, comorbidities, and RA disease duration.
Data Analysis
The focus groups were recorded and transcribed verbatim. Each participant was anonymized by assigning a unique number throughout the transcripts. Data-driven thematic analysis was used by two researchers (KW and AI), who independently analyzed both transcripts using NVivo 12 software (QSR International). Codes were discussed between the two researchers and grouped into overall themes. Owing to the data-driven approach used, it was anticipated that topics would emerge that deviated from the original interview guide. These were also included in the thematic analysis to highlight the concepts that the research team may not have considered in advance. Figure 1 shows the flow of participants from identifying eligible participants to the final sample group of 9.
Themes
A total of four overarching themes emerged from these data. These include the following: (1) important symptoms and outcomes in RA, (2) ways in which patients manage RA symptoms, (3) views on the current RA health care system, and (4) views on remote measurement in RA.
Symptoms Experienced and Important Outcomes
Patients with RA experience a host of symptoms related to both their physical and mental health. Both groups revealed pain and mobility to be the most important outcomes to assess and target during treatment: Further discussions showed that the onset of RA had an impact not only on people's health but also on their lifestyle. Important outcomes here related to maintaining social life and the ability to work and/or be a parent: The inability to maintain a professional work life was also linked with the outcome of mood, with the same patient stating:
Yeah, even when I was diagnosed, first diagnosed, I was off work for four months. Initially, it's like, yeah, I can relax, but then it's, I don't know what to do with myself. Then I know, personally, I'll go down. I'm under the psychologist here as well. [P6]
Although there was a general acknowledgment of the importance of these health and lifestyle outcomes in RA management, there were some individual differences between patients. They explained that these personalized impacts of RA are often the outcomes that are overlooked in the medical profession:
Symptom Management
Symptom management emerged as the second key point of discussion. Ways of dealing with symptoms were first discussed in terms of medication and then in relation to the use of alternative symptom management strategies.
There was a general sense of uncertainty surrounding the effectiveness of medication in both focus groups. Some patients expressed positive experiences of using RA medication to improve outcomes, whereas others described medication as having had no effect on their symptoms. A couple of patients expressed reluctance to use traditional disease-modifying antirheumatic drugs (DMARDs) and steroid medications because of their effectiveness and side effects, respectively. They also felt that there was a lack of support in dealing with the confusion or fear surrounding taking different medications: Complementary treatments emerged as more acceptable than just taking medication alone. Patients reported improved outcomes from dietary changes, meditation, physiotherapy, and exercise. There was agreement that integration of medication and complementary treatments was important, with suggestions for a more holistic approach to RA management. Several barriers to this approach were highlighted, one being the perceived reluctance of health professionals to prescribe alternatives: Other barriers to incorporating complementary treatments into RA management include the allocation of services based on location and restricted access to information. One patient suggested the creation of a website that personalizes exercise class availability based on location. That said, patients also tended to agree that pain or feelings of self-consciousness presented the biggest barrier to using exercise:
Current Views on the System and Ideas for Improvement
Key aspects in effective disease management include the importance of continuity of care, patient-centered care, and the impact of symptom fluctuations on appointment effectiveness.
Continuity of care was highlighted several times as a major area of importance. Perceived lack of coordination, both between departments and external health care providers, was cited as a barrier to continuity of care. Patients discussed their differing experiences in facilitating the link between their consultant, pharmacist, and prescription delivery service: Following this, patient-centered care was also discussed as being of equal importance. Patients reported not seeing the same consultants at each stage of their treatment, meaning that information is often unnecessarily repeated. One individual explained how their experience with an unfamiliar registrar made them feel as though their longstanding story and experience of RA was devalued: Patients went on to suggest how patient-centered care could be improved. For some, this involved building trust with medical professionals; for others, being able to provide some background information before appointments would suffice: I know that sometimes you feel if only they had a card to give to all the persons before, if the person could report and that they could even tick, maybe, things so that, yes, that person has that or that or that-that would make it much easier for them. [P1] The third key issue questioned the effectiveness of current appointments. Both focus groups consistently noted that symptom fluctuations had a major impact on whether their appointments were effective. In general, it was agreed that symptom fluctuations, especially in relation to pain, are often not captured in appointments: . One patient gave an example of the time that their appointment coincided with a period of heightened pain, which resulted in hospitalization. However, this was described as an exception, rather than the rule: Discussions surrounding how this could be improved were two-pronged. On the one hand, patients thought it a good idea to be able to conduct appointments or blood tests on days immediately after their pain symptoms. However, there was consensus that being able to log symptom flares for the purpose of reporting at the time of an appointment would be beneficial as well. It was even suggested that this could be in the form of a digitalized message, for example, a text or an email:
Remote Measurement in RA
Patients with RA might benefit both from providing personal information to new clinicians before appointments and keeping a log of symptom fluctuations to review during appointments. As such, each of the focus groups was asked about the potential of an RMT platform, such as a wearable and smartphone app, to track changes in their health.
Patients had varied views on which specific aspects of RA management should be tracked remotely. These generally culminate in two main symptoms: pain and mobility. With pain, patients were keen to highlight that pain can fluctuate by body part, as well as by time of day. With mobility, it was noted that this can vary both within and between days. These were all perceived as important aspects to capture in remote measurements:
I suppose I would imagine being able to sort of put in particular parts of the body which tend to have pain: joints, or in my case tendons as well. And then monitor maybe daily how they were doing. Some things that come and go, would be quite interesting. [P8]
Maybe just the, maybe the periods of stiffness, when they are, is it different in the morning. [P9] One patient went as far as to describe a way in which their pain fluctuations could be visualized:
But if we had a -and we just ticked, for instance, with a timetable and ticked such a time that was crucial. You could even do it in color, I suppose. If it's very bad, you could have it in red or black or green or whatever. [P1]
It became apparent that it might be of interest to track not only the main symptoms but also the triggers of these symptoms. It was thought that this measurement might act towards preventing symptom flares. There was much deliberation over the symptom triggers that patients were already aware of, and these tended to differ by individual. In general, these included sleep, tiredness, diet, exercise, stress, mood, and other psychological factors. Often, these factors are present comorbidly, and it is difficult to determine the direction of causality: Although there seems to be a general sense that remote measurement in RA might be beneficial in some formats, there were a few differing views on whether encompassing this into an app would be helpful. This seemed to be dependent on the individual's established coping mechanisms:
I'm an app freak, I've literally got an app for everything, anything around technology. So I thought, if there's a way I can manage my arthritis through using an app then that would just be brilliant! [P7]
The problem with the app, as we're talking about it I think, is that the more you put in to sort of cover everything the more daunting it's going to look like [
Principal Findings
This study aimed to provide the first step in developing a multiparametric RMT system for RA symptom tracking by eliciting views on (1) key symptoms prioritized by patients for inclusion in the system and (2) key requirements for maximized utility, uptake, and long-term engagement with RMT symptom management. In addition to these two original objectives, useful information about how RMT systems may be incorporated into existing health care services was elicited. A total of four key themes emerged: (1) key symptoms experienced and important outcomes to consider; (2) clinical and self-management strategies for RA symptoms; (3) ways in which the current health care system can be improved; and (4) facilitating RMT use in patients with RA. Figure 2 breaks down each of these themes into recommendations for consideration when designing a multiparametric RMT system for RA symptom tracking.
Comparison With Previous Work
The results of this study align with previous findings supporting the acceptability of RMT apps for use in RA disease management [13][14][15]24]. Patients approved the development of an app for the purpose of individual symptom tracking. In terms of specific symptoms to include, results reinforce work led by Crouthamel et al [14] and El Miedany et al [19] in the admission that pain and mobility fluctuations are key symptoms to track. In particular, in the study by Crouthamel et al [14], use of a novel joint pain map for the interactive recording of the number and severity of painful joints seems particularly relevant. Our results suggest that patients might find it useful to be able to record these data at multiple periods throughout the day, alongside at weekly intervals. Interestingly, our results show that patients are also interested in tracking symptom triggers, such as sleep, diet, exercise, and mood. Reade et al [15] go some way toward capturing this in their study app, which collects self-reported measures across 10 variables, but crucially this work suggests that patients might prefer to personalize the symptom triggers they are reporting on, as well as to adapt these over time. Much of this work has focused on symptom tracking to improve health outcomes in patients with RA. Although our results suggest these to be of utmost importance, patients also placed value on lifestyle-related outcomes, such as work, parenthood, and socialization. It is less clear at this stage how these aspects of RA management would be incorporated into a symptom-tracking app.
In addition to tracking RA symptoms for self-management, these results suggest that an RMT system should also include a means of transferring data to a health care provider. Logistical barriers exist with respect to sharing data between third-party apps and secure NHS servers, so it is important to consider the value of linking patient symptom trackers with electronic health records (EHRs). Patients proposed the benefits of providing information before appointments, both for the purpose of aiding understanding of their condition and to accurately recall symptom fluctuations. This sentiment resonates well with qualitative work in this area [24]; participants were keen to use RMT tools to communicate with their physician, provided they felt that this would be incorporated into their care. In a similar sense, our findings suggest that provision of information to consultants seems dependent on the trust held for that professional. Previous work demonstrated that integrating remotely captured symptom fluctuations into EHRs presents an effective way for consultants to identify temporal changes and provide tailored disease management [18].
Our findings suggest that an app for patients with RA should include both symptom tracking and data transmission components. At present, the most popular apps available to download for RA management are those that combine symptom tracking and educational content [22]. Educational content, defined as information on disease pathology, diagnosis, or explanation of inputted symptom data, was not explicitly mentioned in our focus groups as an important feature of an app. However, discussions regarding disease management revealed uncertainty surrounding the use of RA medication and complementary treatments. An app that could provide dedicated information and recommendations for a range of treatment options might be the best fit. Interestingly, our results show that perceived reluctance to consider medication alternatives by health care providers is a barrier to treatment adherence. It might also be useful to consider combining educational content with data transmission features, such that patients could research their personal preferences and report back to consultants for review in upcoming appointments. In addition, patients reported motivation and access as key barriers to accessing complementary treatments. Inclusion of a social support feature in an app would go some way toward combating them. These results are in line with recommendations of Luo et al [22] to implement data transmission and access to social communities, alongside symptom tracking and education, into an app for RA management.
Limitations
Several evaluation points must be considered. First, the sample size was small; therefore, additional themes related to this topic might have been missed, and there is limited generalizability to the wider RA population. Second, all patients across the two focus groups, although presenting a variation in age and disease experience, were female. Both facilitators were also women. Male participants were originally recruited but did not attend on the day. Including a female service user as a facilitator might have been more relatable to patients, yet results may still be biased toward a female experience of RA. It is not yet clear in the literature whether the perception of remote symptom measurement varies by gender. This is a common limitation, and the gender ratio is similar to other studies in the field [24]. Third, sampling for qualitative work inevitably results in a selection effect, whereby certain individuals might be more motivated to participate. Similarly, consecutive sampling through a single center excluded patients who were not currently attending outpatient appointments or could not physically attend a focus group. Given the fluctuations in pain and mobility that are apparent in RA, this might have skewed attendance toward those who were feeling well on the day. Had the groups included those presenting with symptom flares, certain themes may have been exacerbated. Fourth, the topic guide (Multimedia Appendix 1) was intentionally vague. This has elicited some important areas for discussion that were not previously considered by researchers, but it has also resulted in the underdevelopment of intended topics. For example, the concept of passive data monitoring has not been widely discussed in the groups, perhaps owing to the relative novelty of the technologies.
Applications for Future Research
This work provides the foundation for developing a multiparametric RMT system for symptom management in RA. Having explored patient views on key symptoms and concepts for consideration, there are some clear applications for app design (Figure 2). An RMT system should include, at the very least, options to track changes in pain and mobility. Symptom severity may be best tracked visually, via a color chart, with the option of tracking changes by body part and over time. Ideally, users should also be able to add additional personalized symptoms and symptom triggers that they feel are pertinent to their experience of RA. There should also be some informational content available through the app regarding the use of medication or the local availability of complementary therapies. Developers should also consider the concept of data transmission of such information to relevant health care providers.
The next steps for future work in this body of research are to develop an RMT system that can undergo subsequent user testing with a similar group of patients with RA. Service-user workshops offer opportunities to facilitate the co-design of aspects such as user interface and usability [18]. Alongside app development, this should also include the provision of a passive, wearable device that can complement active symptom tracking. Given that our discussions found mobility to be a key outcome of importance, passive monitoring is likely to offer additional, unobtrusive insight into symptoms. These sessions could run over several days and include the option of attendance via video call to allow the inclusion of participants experiencing symptom flares or not attending clinic on the day. The purpose of such user testing should be to assess the usability and feasibility of the system and to understand how to maximize the utility of the data collected while minimizing the burden on patients. In turn, this would provide further insight into the barriers of uptake and long-term engagement with using RMT for symptom tracking.
A key requirement for implementing such technologies in clinical care is to assess the viewpoints of all stakeholders [10]. Parallel work should, of course, look to incorporate the views of rheumatology professionals into these discussions. This is especially relevant given that our findings highlight a clear desire to use an RMT system to send information to clinicians ahead of appointments, in the form of personalized details and symptom fluctuations. It is of upmost importance to assess the feasibility of such data transmission. This work could encompass a combination of both patient and clinician stakeholder views in single focus group sessions, discussing how the data would be incorporated into appointments.
Conclusions
This paper has provided an in-depth exploration of the clinical outcomes valued by a group of patients with RA and, as a result, the key areas of consideration for inclusion in a disease management system. Static time point assessments miss important information for patients with fluctuating disease symptoms. Patients are interested in symptom tracking, and there is a clear and consistent message from patients that remote monitoring via an RMT system has a place in RA self-management and clinical care. This work has helped pave the way for the initial design of such an app, the success of which will be contingent on further co-design with patients. It should capture relevant and personalized outcomes with the possibility of integration with EHRs. Future work in this program aims to combine this app with passive symptom monitoring to create an optimal RMT system for RA symptom tracking. In the current climate of the COVID-19 pandemic, health services are witnessing a rapid shift toward remote management of disorders through telemedicine [25]. This work represents a step toward creating an acceptable and engaging remote system for use as an interface between self-management and clinical care during unprecedented times and beyond. | 2020-12-24T09:07:46.535Z | 2021-03-01T00:00:00.000 | {
"year": 2021,
"sha1": "413620a42534aaf08c1a77bc59b6580aabf04035",
"oa_license": "CCBY",
"oa_url": "https://formative.jmir.org/2021/3/e22473/PDF",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "09506f8e0e6ae94fa9523ae258afc98fce05ebce",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
16469205 | pes2o/s2orc | v3-fos-license | Human Papillomavirus Detection in Head and Neck Squamous Cell Carcinomas and Its Clinical Implications
antibody directed against epidermal growth factor receptor, and radiotherapy in head and neck squamous cell carcinoma. They demonstrated improved patient survival compared with radiation alone. The use of this combination increased skin irritation, but otherwise it had the same side effects as radiotherapy alone. An analysis of patients in this trial revealed that those with oropharyngeal squamous cell carcinoma who were male and younger, a group that mirrors the human papillomavirus-positive population, benefited most from the combination therapy. These results suggested that radiation plus cetuximab, instead of cisplatin-based chemotherapy, may reduce treatment toxicity without compromising cancer control for patients with human papillomavirus-positive oral squamous cell carcinoma. For this reason, Radiation Therapy Oncology Group (RTOG) has initiated a phase III randomized study of radiotherapy with cisplatin or cetuximab in patients with human papilloma virus-associated oropharyngeal cancer (RTOG-1016). So, factor head neck
Introduction
Squamous cell carcinomas of the head and neck are a biologically heterogeneous group of cancers with a variable clinical course (Tran et al., 2007).
Human tumor viruses account for approximately one-fifth of all cancers worldwide (Psyrri & Tsiodoras, 2008). The first association between human papillomavirus and head and neck cancer was observed during the 1960s (Rabbett, 1965). A possible role for human papillomavirus in the etiology of cancers at other sites within the head and neck was first suggested by Löning et al., in 1985. Since then, mounting epidemiological, molecular, and clinical evidence indicates that high-risk human papillomavirus (especially human papillomavirus-16) account for the development of head and neck carcinoma in some individuals who do not have the classical risk factors for this disease (Psyrri & Dimaio, 2008).
Distinguishing human papillomavirus positive from human papillomavirus negative head and neck squamous cell carcinoma can provide prognostic information, because different studies have shown better clinical outcome among patients with human papillomavirus positive head and neck squamous cell carcinoma. Although there are innumerable options for human papillomavirus detection in head and neck squamous cell carcinoma, there isn´t any standardization of procedures to use in clinical practice. Several authors propose a testing algorithm of first screening for human papillomavirus using p16 immunohistochemistry, after positive p16 results confirmatory testing with polymerase chain reaction or similar technique is carried out (Pannone et al., 2012;Smeets et al., 2007).
The demonstration that human papillomavirus have a role in human carcinogenesis has allowed the development of preventive and therapeutic strategies aimed at reducing the incidence and mortality of human papillomavirus-associated cancers (Psyrri & Dimaio, 2008).
This chapter reviews the human papillomavirus detection in head and neck squamous cell carcinoma and its clinical implications. Our search strategy included an electronic search of MEDLINE (pubmed), to identify all published articles about this issue. We use the key words "Human papillomavirus", "head and neck neoplasm". We checked the titles and abstracs retrieved. Each author independently assessed the full text of studies relevant to this review.
Risk factors in head and neck cancer
The main risk factors for head and neck cancer globally are tobacco and alcohol (Dobrossy, 2005). These agents act by inducing mutations in key genetic pathways that govern normal cell turnover such as p53 and the product of the retinoblastoma gene (pRb) (Pfeifer et al., 2002) Approximately 20% of head and neck cancers occur in people lacking these established risk factors (Wiseman et al., 2003). There is strong epidemiologic and experimental evidence indicating that human papillomavirus accounts at least partly for this subset of cancers (Shanta et al., 2000), and it has suggested that human papillomavirus may be an independent risk factor for oropharyngeal carcinoma, as well as a modulator the malignancy process in some tobacco-and alcohol-induced oropharynx tumors (Turner et al., 2011).
Human papillomavirus: Concept
Human papillomavirus is a member of the papillomaviridae family. They are small, nonenveloped, DNA viruses. They may be found integrated into the host genome, nonintegrated or episomal, or as a combination or mixture of these types in infected tissue (Turner et al., 2011) [ figure 2].
Human papillomavirus life cycle and its role in the pathogenesis of head and neck squamous cell carcinoma (head and neck squamous cell carcinoma)
Through wounds or abrasions, the papillomaviruses infect basal epithelial cells. The viral DNA is maintained in the nuclei of infected epithelial cells (Stubenrauch & Laimins, 1999). human papillomavirus-DNA replicates to a high copy number only in terminally differentiated cells near the epithelial surface (Stubenrauch & Laimins, 1999). The late viral genes, which encode the L1 and L2 proteins that constitute the virus particle, are expressed only in the highly differentiated cells (Bedell et al., 1991).
Replication of the human papillomavirus genome is critically dependent on the host-cell DNA replication machinery (Cheng et al., 1995). The papillomavirus E1 and E2 proteins are required for viral DNA replication and papilloma formation (Wu et al., 1994). E1 is an ATPdependent helicase that initiates viral replication in cooperation with the E2 protein. In addition, the E2 protein can function as a transcriptional repressor of E6 and E7 oncogene expression among other functions (Psyrri & Dimaio, 2008). E2 loss of function allows upregulation of E6 and E7 oncoproteins (Pannone et al., 2012).
Transcription of human papillomavirus-16 E6/E7 mRNA in tonsillar carcinomas is not necessarily dependent on viral DNA integration, and the viral DNA is predominately in episomal form (Mellin et al., 2002). It has been also demonstrated that high risk human Episomal DNA Mixed DNA Integrated DNA papillomavirus episomal DNAs up-regulate the activity of E6/E7 promoter, which in turn gives rise to elevated E6 and E7 protein expression in cancer cell (Pannone et al., 2012). Mellin et al (Mellin et al., 2002) concluding that in oropharyngeal carcinomas human papillomavirus is almost exclusively not integrates and its carcinogenic activity is due to E6/E7 oncoproteins expressed from episomal viral sequences. It is unknown whether the physical state of the virus influences tumor biology (Tran et al., 2007;Koskinen et al, 2003). However, the data suggested that a higher viral load cloud be a favourable prognostic indicator and that tumours with episomal DNA had larger tumours than patients with mixed or integrates forms of viral DNA. Higher copy number of episomal viral DNA was able to induce more rapid growth, perhaps by higher expression of the viral oncogenes (Pannone et al., 2011).
Human papillomavirus encode E6 and E7 proteins that create a state competent for DNA replication. The E6 protein of the high-risk human papillomavirus binds and induces the degradation of the p53 tumor suppressor protein via an ubiquitin-mediated process. E6 also activates telomerase allowing the regenesis of the ends of chromosomes after cell division. While, the human papillomavirus-E7 protein binds and destabilizes the retinoblastoma (Rb) tumor suppressor protein and related proteins. The molecular consequence of the expression of these viral oncoproteins is cell cycle entry and inhibition of p53-mediated apoptosis (figure 3). The E6 and E7 proteins also interact with other cellular targets. Together, these effects promote cell-cycle progression and viral DNA replication in differentiated keratinocytes (Tran et al., 2007;Leemans et al., 2011;Hobbs et al., 2006). As a result, somatic mutation in TP53 (encoding p53), cyclin D1, and deletion or silencing CDKN2A (encoding p16) are established cancer genes in human papillomavirus-negative head and neck squamous cell carcinoma. In contrast, human papillomavirus-associated tumors are less likely to harbor TP53 mutation and the genes encoding the Rb family are established cancer genes in human papillomavirus-positive head and neck squamous cell carcinoma. In addition, human papillomavirus-positive head and neck squamous cell carcinoma has strong expression of p16 (as a component of the retinoblastoma tumor suppressor pathway) (Snow & Laudadio, 2010;Leemans et al., 2011). In the other hand, p16 expression loss defines a subgroup of head and neck squamous cell carcinoma patients with human papillomavirus-negative tumors.
So, the etiology of head and neck cancer is complex. Human papillomavirus, tobacco and alcohol represent three independent risk factors for head and neck carcinoma in the oral cavity and oropharynx.
The different risk factors can be combined. Smith et al (Smith et al., 2012) found that cancer in oral cavity or oropharyngeal risk was different among patients with several risk factors (Table 1). This investigation suggests that while risk of head and neck squamous cell carcinoma by tumor site is both different between oral cavity and oropharynx, both sites are nonetheless associated with independent effects for each of the three major head and neck squamous cell carcinoma risk factors.
The association between tobacco/alcohol, human papillomavirus, and tumor site is complex.
Oral Cavity/Oropharynx
Human
Epidemiologic and experimental evidence of an etiologic role for human papillomavirus in head and neck squamous cell carcinoma
Certain subsets of head and neck squamous cell carcinoma have fallen in parallel with the reduction in smoking, rates of oropharyngeal squamous cell carcinomas have risen by 2.1% and 3.9% among men and women respectively, from 1973 to 2001, particularly tongue and tonsillar cancers (Shiboski et al., 2005). Similarly, the incidence of tonsillar cancer increased by approximately 2-3% per year among men younger than 60 years from 1975 through 1998 (Canto & Devesa, 2002). In addition, the incidence of human papillomavirus-associated oropharyngeal cancer has increased between 1973 and 2004 (Chaturvedi et al., 2008).
These data suggest that human papillomavirus has emerged as an increasingly important cause of oropharyngeal cancer not only because tobacco-associated head and neck squamous cell carcinoma have decreased, but also because the incidence of human papillomavirus-associated oropharyngeal cancer is increasing (D´Souza & Dempsey, 2011).
This increase in the incidence of oropharyngeal cancer was paralleled by an increase in certain sexual behaviors. This change in the demographics of patients with head and neck squamous cell carcinoma is consistent with a role for genital human papillomavirus in the pathogenesis of oropharyngeal squamous cell carcinoma in individuals whose sexual practices are typically associated with sexual transmission of the virus (Psyrri & Dimaio, 2008). An elevated risk of oropharyngeal cancer has been associated with increasing number of sexual partners, younger age of first sexual intercourse, the practice of oral sex, and a history of genital warts (Trans et al., 2007).
One of the most important studies establishing the causal relationship between human papillomavirus and head and neck cancer was a multi-center case control study conduced by the International Agency for Research into Cancer (IARC) (Herrero et al., 2003). Findings confirmed that human papillomavirus-positive tumors cluster among non-smokers and nondrinkers.
There has been wide variation in human papillomavirus positivity rates in cancers at different sites within the head and neck. Approximately 25% of oropharyngeal cancers have tested human papillomavirus-positive, with rates in tonsillar cancer considerably higher (Trans et al., 2007). In fact, tonsillar crypts seem particularly susceptible to transformation by human papillomavirus, which is similar to the transformation zone of the uterine cervix, the location in which most cervical cancers originate (Psyrri & Dimaio, 2008).
Human papillomavirus detection
Since Syrjänen´s initial observations in 1983 (Syrjänen et al., 1983), there have been numerous reports on human papillomavirus-DNA detection in head and neck squamous cell carcinoma with rates varying from 0% to 100% of tumors studied (Clifford et al., 2003;Campisi et al., 2007). These differences in detection rate are due to at least two principal factors (Pannone et al., 2012): 1. Differences in the epidemiological distribution of oncogenic high risk human papillomavirus in the world 2. Different analytical methods utilized So, there are nearly innumerable options for human papillomavirus detection in head and neck squamous cell carcinoma and no standardization of procedures to be used in clinical practice. The method choice depends greatly upon the desire information (test directed at identifying a broad group of high risk human papillomavirus or targeted at specific human papillomavirus genotypes), available tissue type (fresh tissue, fixed tissue, incision biopsy, brush cytology, saliva, serum, fine needle aspiration biopsy), the ubiquity and preservation of the candidate target molecule (DNA, RNA, and protein), and resources in (Snow & Laudadio, 2010;Robinson et al., 2010) The Southern blot has long been considered the gold standard for detection of specific DNA sequence, however, with its technical demand, necessity for large quantities of DNA… Its use in clinical applications for human papillomavirus detection is rare (Snow & Laudadio, 2010).
Several amplification techniques (polymerase chain reaction [PCR]) have been developed for human papillomavirus type-specific using a specific primer set or for wide-spectrum human papillomavirus detection. Some of them adequately and equivalently amplify the target of interest, as L1 (late gene that encodes the viral capsid). However, multiple portions of the human papillomavirus genome, including L1, may be deleted in the process of integration to false negative results. For this reason, assays have been developed, which amplify portions of E6 and E7 (Snow & Laudadio, 2010). Many studies have shown reproductible results and high sensitivity with RNA-based assays (reverse transcriptase polymerase chain reaction) when using frozen tissue, but this material is not always available for testing. Multiple studies have compared RNA extraction from fresh or frozen tissue with that from formalin-fixed-paraffin-embedded tissue. The greatest decrease in RNA quality occurs immediately after fixation and processing (Snow & Laudadio, 2010).
Consensus polymerase chain reaction and genotyping is applicable to formalin-fixed-paraffinembedded material and it has high sensitivity, however, it can detect of biologically irrelevant human papillomavirus, and the sample can be contaminated during biopsy acquisition. Type specific polymerase chain reaction has similar characteristics to consensus polymerase chain reaction. Real time polymerase chain reaction is applicable to formalin-fixed-paraffin-embedded material, it has high sensitivity and specificity, and it gives an estimate of the viral load, however, it requires tissue microdisecction and DNA extraction (Robinson et al., 2010).
The human papillomavirus-DNA test may be used in head and neck pathology departments with the following diagnostic and prognostic purposes (Reimers et al., 2007): 1. Distinguish human papillomavirus positive from human papillomavirus negative head and neck squamous cell carcinoma and providing prognostic information 2. Distinguish human papillomavirus positive metastases to the loco-regional lymph nodes derived from oropharyngeal cancers versus metastases of other origins 3. Furnish potentially useful indications for cancer treatment options 4. Contribute to the differential diagnosis of rhino-pharynx undifferentiated carcinoma (Worl Health Organization type I potentially related to human papillomavirus infection whereas Type II and III potentially related to Epstein Barr Virus) 5. Provide valuable information for head and neck cancer research Table 2 shows the different types of primers (Hunsjak et al., 2000;Gravitt et al., 2003;Snow & Laudadio, 2010;Micalessi et al., 2012). Figure 4 shows the genomic structure of human papillomavirus.
Immunohistochemistry for the expression of viral human papillomavirus proteins as p16, E5, E6, E7 as surrogate markers of human papillomavirus infection. In the case of p16, human papillomavirus independent pathways of oncogenesis can lead to increased expression of p16 and the specificity is only 79% (Snow & Laudadio, 2010;Pannone et al., 2011). In fact, the immunohistochemistry detection of p16 protein has been proposed as surrogate marker of human papillomavirus infection in head and neck squamous cell carcinoma (Reimers et al., 2007). In situ hybridization is a reproducible technique applicable to detection of a wide of human papillomavirus types particularly from formalin-fixed-paraffin-embedded tissues. However, in situ hybridization is considered method with a low sensitivity, because the low applicability in clinical routine for the long and hard technical word required in (Snow & Laudadio, 2010;Pannone et al., 2011).
A recent study (Pannone et al., 2012) has tested the reliability of a triple method which combines evaluation of p16 expression of viral human papillomavirus proteins by immunohistochemistry, human papillomavirus-DNA genotyping by polymerase chain reaction, and viral integration into the host by in situ hybridization. All the head and neck HPV squamous cell carcinoma confirmed human papillomavirus positive by polymerase chain reaction and/or in situ hybridization were also p16 positive by immunohistochemistry. So immunohistochemistry showed a very high level of sensitivity as single test but lower specificity level. The double method, in situ hybridization and polymerase chain reaction increased significantly the specificity, but reduced the sensitivity. They observed different levels of p16-immunohistochemistry accuracy in the different cancer subpopulation studied. So, in a cohort of prevalently alcohol/tobacco associated cancers, p16-immunohistochemistry test showed a lower level of specificity in detecting human papillomavirus positive cases. In addition, a recent literature report demonstrates different p16 accuracy according to different anatomical sub-sites of head and neck region (Doxtader & Katzenstein, 2011). In this context, the p16-immunohistochemistry test alone could be used only as a screening method and need to be associated with molecular tests in order to detect human papillomavirus-DNA and to assess its integration status.
The hybrid capture technique is used extensively among pathology labs to detect 13 highrisk human papillomavirus genotypes in cervical cytology specimens. The use of this method is limited for head and neck squamous cell carcinoma human papillomavirus testing, but the technique has potential for screening oral brushings. However, at this time, oral brush cytology has not achieved a sensitivity or specificity sufficiently competitive with surgical biopsy for diagnosis and prospective studies are necessary to determine the clinical use of screening in (Snow & Laudadio, 2010;Pannone et al., 2011).
Luminex system combines PCR with hybridization to fluorescence-labeled polystyrene bead microarrays. This technology provides a new platform for high-throughput nucleic acid detection and is being utilized with increasing frequency. It is a sensitive, reproducible technique for the simultaneous genotyping of all clinically relevant genital HPV types. However, these Luminex assays have shown low ability for type-specific genotyping and have missed variants with the type-specific probes. Multiple infections may occur in 20-40% of specimens. Luminex-based HPV genotyping can be used to differentiate between newly acquired HPV types and pre-existing infections when applied over time. Nevertheless, a limitation of the assay is the reduction of signal that occurs for a plasmid target in low abundance when it is amplified with another target that is 2 or 3 logs higher in abundance. This technology has been tested in cervical samples (Oh et al., 2007;Lowe B et al., 2010).
Human papillomavirus serology. The immune to human papillomavirus infection involves both the cell-mediated and humoral responses. Human papillomavirus seropositivity is
potentially indicative not only of current oral infection but also of any past infection not limited to the oral cavity or oropharynx (Pannone et al., 2011). Antibodies to human papillomavirus E6 and E7 proteins are markers for an invasive human papillomavirusassociated cancer. The use of human papillomavirus viral load in conjunction with serological markers may serve to identify a subset of human papillomavirus-associated head and neck tumors in which human papillomavirus is biologically active (Ragin et al., 2007). Table 3 resumes the characteristics of different methods for human papillomavirus detection (Dobrossy, 2005). Establishing a diagnosis of head and neck cancer requires the acquisition of adequate biopsy material. Typically, tissue samples are fixed in formalin, processed in the laboratory and formalin-fixed-paraffin-embedded, whereas fine needle aspiration biopsy samples are usually treated with an alcohol-based fixative. So, for an human papillomavirus test to be useful it should be capable of reliably classifying 'human papillomavirus related' cancers in fixed cell and tissue samples. The techniques used should be reproducible, subject to standardization and quality assurance and be economically viable (Robinson et al., 2010).
Method
The presence of the viral DNA does not establish causality, since the majority of human papillomavirus infections may be transient rather than persistent (Ragin et al., 2007). The important issue is that human papillomavirus is transcriptionally active (Trans et al., 2007;Hobbs et al., 2006). In fact, cases that are human papillomavirus positive but negative for p16 expression (or negative for E6/E7 mRNA) are molecularly more similar to human papillomavirus negative cases suggesting that in these instances human papillomavirus is not directly involved in carcinogenesis (Snow & Laudadio, 2010;Weinberger et al., 2006).
Detection of high-risk E6/E7 mRNA or protein would be the ideal test for classifying a tumor as truly human papillomavirus-associated, while it´s possible to perform quantitative polymerase chain reaction on formalin-fixed-paraffin-embedded samples the maximum accuracy is found using fresh frozen tissue (Pannone et al, 2012). Determination of p16 expression status by immunohistochemistry could serve as a reasonable surrogate marker for biologically relevant high-risk human papillomavirus infection (Psyrri & Dimaio, 2008).
Smeets et al (Smeets et al., 2007) propose a testing algorithm of first screening for human papillomavirus using p16 immunohistochemistry, after positive p16 results confirmatory testing with polymerase chain reaction is carried out. This approach had almost 100% sensitivity and specificity, with 2% risk of false positive. Others authors as Westra (Westra, 2009) propose confirmatory testing by in situ hybridization (Robinson et al., 2010). The majority of pathology laboratories have the capability of delivering the first algorithm.
Some authors are searching for other defining molecular characteristics. There is evidence that human papillomavirus positive head and neck squamous cell carcinoma tends to contain normal copies of the p53 gene (Braakhuis et al., 2004). However, p53 mutations have been described in these tumours (Westra et al., 2008). This presence of mutant p53 along with human papillomavirus infection in the same tumour raises the possibility that human papillomavirus infection is simultaneous and has no influence on pathogenesis (Robinson et al., 2010). Human papillomavirus viral oncoproteins are known to have epigenetic effects. They can silence the expression of key tumour suppressor genes by promoter methylation (Henken et al., 2007). The emergence of global genome methylation assays represents novel ways of refining the molecular classification of head and neck cancers in the future (Robinson et al., 2010).
Clinical implications
Several lines of clinical evidence also suggest that human papillomavirus-associated head and neck squamous cell carcinoma could be biologically distinct from classical head and neck squamous cell carcinoma (Table 4). Class IV, human papillomavirus-negative/p16 expressing. Small number of apparently human papillomavirus negative carcinomas that over express p16. There are two major reasons for this entity: a] misclassification as human papillomavirus negative, because the human papillomavirus test chosen lacks sensitivity, b] tumours where accumulation of p16 has been caused by perturbation of other cellular signaling pathways, or due to possible an as yet unidentified infectious agent.
CHARACTERISTICS
Class III had the highest viral loads. The 5-year survival in class III was 79%, significantly higher than in the other two classes (20% and 18%, P = 0.0095). Disease free survival for class III was 75% compared with 15% and 13% for classes I and II, respectively (P = 0.0025). The 5year local recurrence was 14% in class III compared with 45% and 74% (P = 0.03). Multivariate survival analysis confirmed the prognostic value of the three class model. It is clear that head and neck squamous cell carcinoma human papillomavirus-positive and p16 expressing is different from classic head and neck squamous cell carcinoma, but it is not clear whether head and neck squamous cell carcinoma human papillomavirus-positive and p16 non expressing (probably, tobacco/alcohol-related tumors that are infected by high-risk human papillomavirus) represents a group biologically distinct from human papillomavirus-negative tumors (Cheng et al., 1995). Other studies in (Smith et al., 2012;Harris et al., 2011) have confirmed better disease-specific and recurrence-free survival in human papillomavirus and/or p16 positive tumors.
About class IV, several studies (Reimers et al., 2007;Harris et al., 2011;Shah et al., 2009;Weinberger et al., 2004) have shown that patients with human papillomavirus negative and p16 positive tumors had better outcomes than patients with p16 negative tumors. So, p16 status could be the truly important prognostic marker in head and neck squamous cell carcinoma, independent of human papillomavirus infection.
For all this, p16 positivity has been proposed to be a more reliable and reproducible prognostic marker in head and neck squamous cell carcinoma (Harris et al., 2011).
On the basis of these results, we can refine a model for human papillomavirus-associated oropharyngeal cancer. The favorable outcome of human papillomavirus-induced oropharyngeal cancers might be attributable to the absence of field cancerization or enhanced radiation sensitivity (Lindel et al., 2001). In 1953, the term "field cancerization" was proposed to explain the high propensity to develop local recurrence after treatment of head and neck squamous cell carcinoma and the high likehood that multiple independent tumours will develop in the head and neck mucosa. This phenomenon is due to the presence of carcinogen induced early genetic changes in the epithelium from which multiple independent lesions arise (Slaughter et al., 1953).
Disrupting E6 and E7 in oropharyngeal cell lines results in increased levels of p53 and pRB and increased levels of p53-activated genes (Rampias et al., 2009). These findings indicate that in human papillomavirus-induced carcinoma the p53 and pRB pathways remain intact.
So, unlike tobacco associated oropharyngeal cancers that harbor mutant TP53, the apoptotic response of human papillomavirus-associated tumors to radiation and chemotherapy might be intact. Some authors have proposed that p16-expressing cells are less hypoxic and respond with less accelerated repopulation when irradiated (Lassen et al., 2009).
Given that the rate of p53 mutation is quite low in human papillomavirus-associated tumors, the addition of p53 mutation sequencing could have added valuable information had sufficient tissue been available (Harris et al., 2011). There have been conflicting data on p53 expression in human papillomavirus positive head and neck squamous cell carcinoma tumor cells. Some studies have observed high expression of nuclear p53 in some human papillomavirus-containing tumors with wild-type p53 (Hafkamp et al., 2003), and other studies demonstrating low p53 expression (Wilczynski et al., 1998). Mechanism of overexpression of wild-type p53 in the presence of the virus is known (Tang et al., 2011).
It has been suggested that less intensive treatment modalities should be examined in order to decrease treatment-toxicities. For that, the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (USA) (Pfister et al., 2011) recommend that oropharyngeal squamous cell carcinoma is tested for high risk oncogenic human papillomavirus.
Furthermore, epidermal growth factor receptor expression has been suggested to be correlated with human papillomavirus status (Almadori et al., 2001). There are data suggesting a direct link between human papillomavirus-encoded proteins and epidermal growth factor receptor expression (Kim et al., 2006). Kumar et al (Kumar et al., 2008) reported the phenotype human papillomavirus positive and epidermal growth factor receptor high to be associated with poorer survival after chemotherapy and radiation than human papillomavirus positive and epidermal growth factor receptor low tumors.
Bonner et al (Bonner et al., 2010) tested the combination of cetuximab, a monoclonal antibody directed against epidermal growth factor receptor, and radiotherapy in head and neck squamous cell carcinoma. They demonstrated improved patient survival compared with radiation alone. The use of this combination increased skin irritation, but otherwise it had the same side effects as radiotherapy alone. An analysis of patients in this trial revealed that those with oropharyngeal squamous cell carcinoma who were male and younger, a group that mirrors the human papillomavirus-positive population, benefited most from the combination therapy. These results suggested that radiation plus cetuximab, instead of cisplatin-based chemotherapy, may reduce treatment toxicity without compromising cancer control for patients with human papillomavirus-positive oral squamous cell carcinoma. For this reason, Radiation Therapy Oncology Group (RTOG) has initiated a phase III randomized study of radiotherapy with cisplatin or cetuximab in patients with human papilloma virus-associated oropharyngeal cancer (RTOG-1016).
So, epidermal growth factor receptor, and p53 are also relevant markers that modify the prognostic effect of human papillomavirus and may help guide the development of targeted therapy in head and neck squamous cell carcinoma.
However, not all patients with human papillomavirus positive tumors respond well to therapy and the reasons for failure in some cases are not known (Maxwell et al., 2010). The variability of high risk human papillomavirus containing cell lines enhances our ability to study the role that human papillomavirus plays in head and neck squamous cell carcinoma development and response or resistance to therapy (Tang et al., 2011). Combination of several risk factors could explain this. A positive tobacco history in patients with human papillomavirus positive tumors may represent a distinct group of head and neck squamous cell carcinoma when all head and neck squamous cell carcinoma are divided by etiologic factors: human papillomavirus negative smokers, human papillomavirus positive never smokers, human papillomavirus positive ever smokers, etc… TPV status likely provided an additive and possibly synergistic effect with others risk factors (Tang et al., 2011).
In addition, in the patient with metastatic head and neck carcinoma of unknown origin, the presence of human papillomavirus in a fine needle aspiration biopsy sample can be used to direct the search to the oropharynx (Zhang et al., 2008).
All of this could lead to a new diagnosis and treatment algorithm ( Figure 5).
Others implications are the preventions actions: 1. Screening studies have been performed in healthy adults using biopsy samples or less invasive saliva and oral lavage-based testing methods to identify human papillomavirus. These techniques revealed prevalence rate between 0-25% (Turner et al., 2011).
Detection of high risk human papillomavirus DNA may help identify individuals, including those with: a) any genetic predisposition to acquire high risk human papillomavirus infection and/or b) a limited immunologic ability to eliminate the virus. Whether oral exfoliated high risk human papillomavirus status is predictive of cancer before invasion or progression in patients with head and neck squamous cell carcinoma is unknown.
Quantitative measurement of salivary human papillomavirus16 DNA can be promise for surveillance and early detection of recurrence. Detection of high risk human papillomavirus in oral exfoliated cells may serve as clonal markers to monitor the presence of residual tumor after surgery or radiation, cancer recurrence, and progression (Pannone et al., 2011).
A recent study (Turner et al., 2011) recruited patients and screened saliva samples for high risk human papillomavirus using quantitative polymerase chain reaction. They confirmed human papillomavirus16, but not human papillomavirus18 in a small subset of the healthy adult patients. These patients were female and minority (2.6%).
2. Prophylactic vaccines that prevent persistent cervical human papillomavirus-16 infections might be effective in preventing these cases of head-and-neck cancer as well, either indirectly by eliminating an anogenital source of virus or directly by protecting the oropharyngeal epithelium itself from infection (Psyrri & dimaio, 2008).
However, these vaccines do not alter the prognosis of established human papillomavirus infection. Therapeutic vaccines based on the viral oncogenes are still in the developmental stage, but they may eventually prove beneficial if used in association with conventional approaches for the management of advanced disease (Tran et al., 2007).
Conclusion
Emerging evidences suggest that human papillomavirus-associated head and neck squamous cell carcinoma is a separate subgroup and biologically distinct from classical head and neck squamous cell carcinoma.
Human papillomavirus positive head and neck squamous cell carcinoma is typically found in the oropharynx and have been associated with younger patients who are less likely to be smokers or drinkers and with improved response to therapy and overall survival (Harris et al., 2011).
So, recognition that human papillomavirus has an etiologic role in head and neck squamous cell carcinoma has important implications for prognosis, treatment, disease prevention, and screening tests which are still being developed. Several authors have suggested that patients with human papillomavirus positive head and neck squamous cell carcinoma can be treated with chemo-radiotherapy or cetuximab-radiotherapy instead surgery.
Although high risk human papillomavirus detection is of utmost importance in clinical setting of head and neck squamous cell carcinoma, there is no agreement about the "golden standard" considering the number of molecular methods or combinations available.
There is evidence that detection of high risk human papillomavirus by consensus polymerase chain reaction alone is insufficient to accurately classify tumours, however, there is convincing evidence that the detection of p16 protein by immunohistochemistry can be used as a surrogate marker for the elaboration of oncogenic human papillomavirus proteins (Robinson et al., 2010). So, this is feasible as part of a routine diagnostic process using either a combination of p16 immunohistochemistry and in situ hybridization (Westra, 2009) or p16 detection and consensus polymerase chain reaction (Smeets et al., 2007).
Several studies have shown p16 expression status as a predictor of prognostic marker in head and neck squamous cell carcinoma, independent of human papillomavirus infection. In addition, epidermal growth factor receptor, and p53 are also relevant markers that modify the prognostic effect of human papillomavirus and may help guide the development of targeted therapy in head and neck squamous cell carcinoma. | 2017-09-17T19:13:55.590Z | 2012-12-05T00:00:00.000 | {
"year": 2012,
"sha1": "e121c5ca13e21f4b11d348a8c720378b0a4bb57b",
"oa_license": "CCBY",
"oa_url": "https://www.intechopen.com/citation-pdf-url/41352",
"oa_status": "HYBRID",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "1fcac8d4752f171ae5f36ffd3bae8f518a0e097e",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
268957262 | pes2o/s2orc | v3-fos-license | The functional and psychological impact of delayed hip and knee arthroplasty: a systematic review and meta-analysis of 89,996 patients
This systematic review and meta-analysis aimed to determine the impact of presurgical waiting times on pre-/post-operative joint specific pain and function, health-related quality of life (HRQOL) and perspectives of patients awaiting primary elective total hip (THR) and knee (TKR) replacements. MEDLINE, EMBASE, PUBMED, and CENTRAL databases were searched from inception until 30th January 2023 (CRD42022288128). Secondary literature and unpublished datasets containing paediatric, non-elective, partial, or revision replacement populations were excluded. PRISMA 2020 reporting and GRADE certainty of evidence guidelines were followed. Residual maximum likelihood meta-analysis and linear meta-regression was performed to elucidate the influence of presurgical waiting time. Twenty-six studies were eligible for systematic review and sixteen for meta-analysis, capturing 89,996 patients (60.6% female, mean age 67.4 years) between 2001 and 2022. A significant deterioration in joint function (mean difference (MD):0.0575%; 95% CI 0.0064, 0.1086; p = 0.028(4d.p.); I2 = 73.1%) and HRQOL (MD: 0.05%; 95% CI − 0.0001.0009; p = 0.011(4 d.p.); I2 = 80.6%) was identified per additional day of waiting. Despite qualitative evidence, meta-analysis could not observe a relationship with postoperative outcome data. Patient responses to delayed THR and TKR surgery were unanimously negative. Immediate action should seek to reduce the increased patient anxiety and significant reductions in pre-operative joint functionality and HRQOL associated with prolonged pre-surgical waiting time, whilst mitigating any potential deleterious post-operative effects.
Primary total hip (THR) and knee replacements (TKR) are amongst the most common elective orthopaedic procedures 1 .In the UK, the mortality-adjusted lifetime risk for a THR and TKR at fifty years-old is 11.6% and 10.8% and 7.1% and 8.1% for females and males, respectively, whilst, in the USA, over 50% of patients diagnosed with symptomatic knee osteoarthritis will undergo TKR 2,3 .Most frequently, these surgeries are performed in the management of end-stage osteoarthritis-a degenerative process which accounts for 2% of global disability years, for example impacting 10% of UK adults 4,5 .
Following the COVID-19 pandemic, the total number of THRs and TKRs performed in the UK has halved and not yet recovered 6,7 .Concurrently, demand has continued to rise, resulting in extended waiting lists 6,7 .Similar trends can be seen in comparable health systems including, Canada, the Netherlands, and Denmark 8,9 .Consequently, understanding how increased waiting-times for elective primary THR/TKR influences patientcentred outcomes will be essential for future healthcare planning.
Previously a systematic review of 1,646 patients with osteoarthritis awaiting THR and TKR by Hoogeboom et al. identified no evidence of deterioration in self-reported pain status 10 .However, this analysis was limited to the first 180 days of waiting for THR or TKR 10 .More recently, Patten et al. concluded that the pain levels of patients with osteoarthritis remained stable for the first year after addition to a surgical waiting list, although only reported a median follow-up of 13.6 weeks 11 .Furthermore, both reviews did not measure changes in joint functionality, quality of life, wider patient perceptions, nor explore how prolonged pre-surgical waiting time might impact postoperative outcomes.
Sex (% female) [WT groups]
Kelly et al. 13 Prospective cohort Patients waiting for > 1 month for www.nature.com/scientificreports/ between the rate of decreasing postoperative functional gain 53 .The symmetrical funnel plot (Fig. 2D) alleviated concerns discounts publication bias when considering outcome certainty.Five studies were unsuitable for quantitative synthesis due to outcome reporting 18,20,25,37,38 .Of these three associated decreasing postoperative functional gains with increasing time awaiting surgery 18,20,25 .Apart from Holzapfel et al., all study outcomes were captured 6 to12 months post-operatively.Holzapfel explored the impact of presurgical postponement, e.g., following patient admission for primary elective THR and TKR, over 10 years at a single German centre (N = 10,140) and identified an increased risk of complications and revisions within the postponed surgery group 37 .
Within multivariate regressions of their reported THR (n = 29,303) and TKR (n = 32,602) populations, Nikolova and colleagues identified statistically significant deterioration in 6-month post-operative OHS(0.0951,p < 0.01) and OKS(0.0385,p < 0.01) per additional day of waiting time, exceeding the estimated effect sizes proposed in this review (Fig. 2C) 30 .This apparent inconsistency between univariate effect and a sufficiently powered multivariate analysis raised concerns around imprecision and unmeasured confounding, and limited certainty in the proposed effect size to "moderate" (Supplementary Table 6).Consequently, it is likely that a negative correlation between preoperative waiting time and postoperative joint function exists.
Changes in health-related quality of life whilst awaiting total hip or knee replacement surgery
Eleven studies, representing 7,831 patients commented on changes in HRQOL in the preoperative period 13,16,17,19,22,24,26,27,32,33,35 .Seven of these studies (containing ten cohorts and capturing 2,153 patients) were suitable for meta-analysis 13,16,17,22,27,32,35 .Figure 3A presents a forest plot of the change in single preference-based health related quality of life index associated with waiting for total hip or knee replacement.Whilst awaiting www.nature.com/scientificreports/surgery, there was a significant deterioration in patients' HRQOL (MD 0.04; 95% CI 0.00,0.09;p = 0.04; I 2 = 88.8%) on a scale of 0-1 (worst-to-best).Linear meta-regression (Fig. 3B) showed a statistically significant daily deterioration coefficient of 0.0005 (95% CI − 0.0001.0009;p = 0.011 (to 4 d.p.);I 2 = 80.6%).Notably, this analysis excluded the non-fixed waiting time cohort of Hivronen et al.Despite 143/183 patients breaking protocol, the reported per-protocol and intention-to-treat admission 15D score and standard deviations were identical to 3 decimal points.In the absence of author clarification, this raised significant concerns around both reporting errors and a negative compliance bias 22 .
Although publication bias was not of concern, (Fig. 3C), risk of bias, primarily due to limited control of confounding, enabled only moderate certainty in this conclusion (Supplementary Table 6). 16,24,35.Notably, McHugh et al., the study which contributed to this risk of bias, was not included within the quantitative synthesis, and therefore did not influence the slope of the meta-regression (0.0005; 95% CI − 0.0001,0009; p = 0.011 (to 4 d.p.);I 2 = 79.8%).Thus, this risk of bias is immaterial when considering the certainty of the proposed effect size.In summary, HRQOL deteriorates every day whilst awaiting primary elective THR or TKR surgery.
To explore heterogeneity with regards to presurgical waiting time, Fig. 4B presents a linear meta-regression, identifying a coefficient for daily reduction in post-operative HRQOL of 0.0003 (95% CI − 0.0012,0.0017;p = 0.715 CI − 0.0018,0.0009;p = 0.5554 (to 4 d.p.);I 2 = 61.4%)had no significant deterioration with waiting time.Clogg's method did not identify significant differences between these slopes 53 .extraction 14 .Desmeules et al. reported a significant reduction in the 6-month postoperative SF-36 role physical domain within a subgroup of patients waiting longer than 9 months for TKR, relative to the 3-6 and 6-9 month subgroups, with no statistically significant changes in the additional reported SF-36 subdomains (physical functioning and bodily pain only) 28 .Six-months postoperatively, Nikolova et al. identified significant EQ-5D reductions of 0.0620 (n = 29,303) and 0.0587 (n = 32,602) per additional day spent waiting for THR and TKR, respectively (rescaled 0-100, worstto-best) 30 .Subsequently, a post-hoc sensitivity analysis of Nikolova et al. demonstrated negligible influence on the results of the postoperative HRQOL synthesis (N = 596; MD 0.0003; 95% CI − 0.0015,0.0022;p = 0.7437 (to 4d.p.);I 2 = 96.0%).Although the magnitude of Nikolova's proposed multivariate effect size was consistent with the estimated effect (Fig. 4C) for TKR, the THR outcome showed marked discrepancy with this synthesis.This presented a challenge to certainty when considering this analysis's univariate meta-regression, particularly around unmeasured confounding within long-term, post-operative outcomes (Supplementary Table 6).As such, whilst THR patients reported greater improvements in HRQOL at 6-12 months post-operatively relative to TKR, there is likely an additional relationship with presurgical waiting time that could not be observed in this analysis.The influence of publication bias, as assessed by funnel plot, (Fig. 4D) was neglibile.
Patient perspectives on delayed elective total hip or knee replacement surgery
Only five studies that reported patients' qualitative responses to and perceptions of delayed THR or TKR were identified, summarised in Table 4 (below).All studies (3/3) commenting on psychological responses reported increased patient anxiety, whilst two studies (2/2) reported worsening patient perspectives of both THR and TKR services.
However, it was noted that there were several limitations to this outcome.Firstly, the qualitative nature of these outcomes introduced diverse outcome measures, limited both the directness and comparability of the outcome synthesis.Consequently, this precluded quantitative synthesis and assessment of publication bias, reducing certainty in this outcome to "moderate".Despite these, all reported outcomes captured negative patient perspectives on delaying THR and TKR for patients, speaking to an underlying deleterious patient experience associated with delayed surgical care.
Summary of findings
This review presents a contemporary and high-quality (Oxford Centre for Evidence-Based Medicine, OCEBM, Level 1) analysis exploring the impact of presurgical waiting time on primary, elective THR and TKR outcomes 55 .In this systematic review and meta-analysis a significant association between increased presurgical waiting time and deteriorations in patient joint-related function and HRQOL for primary elective THR and TKR were identified.Given reported thresholds, this synthesis estimates that clinically meaningful deterioration (i.e., more than the minimum clinically important difference) in joint specific outcomes and HRQOL likely occur within 6 months of waiting for surgery 32,56,57 .Furthermore, qualitative synthesis indicated that pre-surgical waiting time may deleteriously influence clinical outcomes up to 12-months post-operatively.In alignment with current evidence, post-operative meta-analysis identified a decreased gain in joint specific outcomes and HRQOL at 6 to 12 months post-operatively for primary elective TKR patients, relative to THR 58 .Patients also indicated strongly deleterious perspectives on prolonged presurgical wait for these procedures.
Limitations
There are several considerations when interpreting this review.Firstly, limitations within the reported literature prevented the use of gold-standard individual patient meta-analysis and/or multivariate meta-regression approaches.
Secondly, the outcome measures synthesised for the primary outcomes of this meta-analysis were diverse.However, only validated outcome measures were utilised in quantitative analysis, and recommended approaches to comparing these were utilised to preserve the validity of this review.Interestingly, there was no trend in preference for particularly outcome measures over time, indicating broad consensus on the utility and validity of each system.Furthermore, 61,905 of the 89,996 patients reported were drawn from one study 30 .Although this study had a low risk of bias and, through its own multivariate regression, estimated a greater effect size in post-operative joint-function and HRQOL outcomes than was observed in the current meta-analysis, it is possible that this review was underpowered to directly observe any relationship between pre-surgical waiting time and postoperative joint function and HRQOL 30 .Consequently, a deleterious relationship remains plausible.Reassuringly, this meta-analysis independently replicated previous findings in literature, showing a greater improvement in joint functionality and HRQOL following THR relative to TKR in patients 6-12 months post-operatively 58 .Despite this, these postoperative outcomes also likely carried greater exposure to confounding due to the multifactorial nature of postoperative recovery (e.g., when considering the impact of independent predictors of postoperative outcomes, such as capacity for self-care, comorbid diseases (e.g.diabetes), and exercise) 59,60 .
Finally, all reported surgeries were performed in Europe, North America, or New Zealand.This limits the applicability and generalisability of these findings to African, Asian, and South American patient populations [61][62][63] .Indeed, future research should seek to expand this evidence-base for patients in underserviced health systems.However, when considering the influence of patient heterogeneity and confounding on the reported analyses, both these geographical restrictions and the broader homogeneity of included patient populations serve to mitigate against unmeasurable confounders, (e.g., socioeconomic circumstance), and preserve the internal validity of this work.In alignment with this, whilst included studies capture a 21-year study period with consequent practice variability, reported effect sizes show broad temporal agreement.Interestingly, the significant reduction in preoperative HRQOL associated with prolonged pre-surgical wait time was driven by more recent studies, perhaps indicating the influence of additional risk factors and multimorbidity within modern patient populations.
Conclusions
This is the first meta-analysis and systematic review to explore both post-operative outcomes and patients' perspectives to delayed primary elective THR and TKR surgeries.Despite variable outcome assessment, patient voice was unanimous that delays to surgery carried deleterious impacts on both their psychosocial wellbeing and perceptions of care.Indeed, out with this context this finding is reflected within broader elective orthopaedic literature 39,42,44 .Importantly given its influence on HRQOL, future work should seek to further explore patient perspectives on delayed surgical care provision.
The findings of this review differ from previous systematic reviews and meta-analyses on the effect of waiting for THR or TKR.When considering previous reviews, this reported population (preoperative outcomes N = 9,020) exceeds both previous sample sizes (N = 1,646 and N = 2,490 patients for Hoogeboom and Patten, respectively), enhancing the relative power of this meta-analysis (10,11].Given this larger patient population, and coupled with this review's implementation of meta-regression to prevent false negative errors from arising around binary discrimination, the pre-operative arm of this analysis is likely powered. Whilst this meta-analysis clearly shows the deleterious effect of prolonging pre-surgical waiting time in primary elective THR and TKR, no explicit relationship between waiting time and postoperative HRQOL and joint functionality was identified.However, as pre-surgical wait times continue to increase, and patient pre-operative joint function (itself an independent predicator for postoperative joint function) continues to deteriorate, this effect may become more apparent in future 25,45 .
In conclusion, this systematic review and meta-analysis of 89,996 patients undergoing primary elective THR and TKR demonstrates a significant association between prolonged pre-surgical waiting and deleterious preoperative joint-specific outcomes and HRQOL.Furthermore, patient voice unanimously condemned delayed care.Postoperatively, there was a plausible relationship between waiting time and 12-month joint function and HRQOL, whilst patients undergoing THR experience greater joint functionality and HRQOL gains from surgery, relative to TKR.Future work should elucidate the determinants of post-operative joint function and HRQOL following primary THR and TKR, across the global economic spectrum.However, in lieu of this, urgent action should be taken to minimise the ongoing deterioration of patients' joint functionality, HRQOL, and psychological status whilst awaiting elective primary THR/TKR.
Protocol registration and ethics
The study protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO) database on 01/03/2022 (CRD42022288128) and was undertaken in accordance with the Preferred Reporting Items for Systematic Review and Meta-analyses (PRISMA, 2020) and Cochrane Handbook for Systematic Reviews of Interventions (2021) 64,65 .Ethical approval was not required.
Information sources and search strategy
Reported populations of patients undergoing elective primary THRs or TKRs were interrogated, to explore how the duration of the pre-surgical waiting period, defined as the period from placement on waiting list until surgery, influenced joint-specific outcome scores, indices of health-related quality of life, and psychological responses.On 30/1/2023, this search strategy (Supplementary Table 1) was executed on the MEDLINE(R) and In-Process, In-Data-Review & Other Non-Indexed Citations 1946 to January 30th, 2023", "EMBASE 1980 to 2023 Week 5″, PubMed, and Cochrane CENTRAL databases.No limitations were placed these searches.
Selection and eligibility criteria
Primary published observational and randomised-controlled literature reporting elective primary TKR and THR populations, their time-to-treatment on presurgical waiting lists, and inclusion of relevant outcome data were eligible.The primary outcomes of this systematic review were validated measures of joint specific outcomes, HRQOL, or psychosocial perspective, with qualitative data being sought secondarily.Individual case-studies and unpublished data were ineligible.To ensure internal validity and preserve external validity, reported populations containing patients aged 16 or under and those undergoing non-elective (e.g., unplanned trauma surgery), partial, or secondary revision of THR/ TKR were also excluded.
Following execution of the search strategy, identified records were de-duplicated and independently screened by two authors to assess relevance.Inter-rater reliability was thus assessed using Cohen's kappa and disagreements resolved by the senior author 12,65 .Subsequently, relevant abstracts were reviewed to establish final eligibility.All relevant records were reviewed to identify further references of interest.Non-English language manuscripts were only to be excluded after attempting to contact the corresponding authors.
Data collection and analysis
Following identification, primary outcome data was extracted by the primary author under supervision from the final author.This data was then analysed in prespecified pre-and post-operative subgroups to prevent confounding.Where duplicate data was reported in literature, the revised analysis was retained.Preferred primary outcome measures included clinically validated joint-specific outcome scores (Oxford Hip and Knee scores (OHS, OKS), health-related quality of life scores (HRQOL, e.g., Short Form 36 (SF-36) and EuroQol-5D (EQ5D)), for meta-analysis.However, qualitative, subjective, non-validated, or otherwise incompatible outcome measures were also recorded and manually tabulated for systematic review.
Potential sources of ascertainment bias and heterogeneity were identified through further collection of descriptive datasets (e.g., patient eligibility criteria, duration of presurgical waiting time, study period, study location(s), sample sizes, proportion female sex, and mean patient age was sought for reported populations].Where missing data was identified or further information required for analysis, clarification was sought from the relevant corresponding author.
Given the diverse outcome measures within scoping literature, comparable data was transformed and represented as mean and standard deviation to facilitate quantitative synthesis.Joint-specific pain and function scores were converted into a common weighted score (0-100, worst-to-best of pain and function subdomains) 65 .Similarly, SF-36 HRQOL (0-100 scale) was converted into a single preference-based EQ-5D index (0-1, worstto-best) following Ara and Brazier's method 54 .Where necessary, the standard deviation was imputed using Wan et al. 's adaptive method 66 .Effect sizes were summarised as either mean difference (MD, percentage change) or odds ratios (OR), where possible.
In the absence of individual patient data, summary meta-analysis was undertaken using a random effect residual maximum likelihood model of mean difference.95% confidence interval (95% CI), and heterogeneity (I 2 ) were also calculated.Linear meta-regressions were performed to explore heterogeneity due to variable reported mean pre-surgical waiting times, whilst sensitivity analyses of regression curves were performed using Clogg's method to enable comparison of postoperative hip and knee arthroplasty data 53,65 .Mixed model metaanalyses and meta-regressions were performed using the "Metafor" R software package 67,68 .Publication bias of quantitative syntheses was assessed using Egger's funnel plot tests 65,68,69 .
Risk of bias and certainty assessment
Risk of bias from randomised, non-randomised, and cross-sectional studies following gold-standard Risk of Bias 2 (RoB2), Risk of Bias in Non-Randomised Studies (ROBINS-I), and Joanna Briggs Institute (JBI) frameworks, respectively 47,48,70 .In each instance, the overall risk of bias for sought outcomes was ascribed following the highest constituent domain.Sensitivity analysis was indicated following identification of maximal overall risk of bias.The Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework was also used to assess the certainty of evidence for each outcome synthesised as either "high", "moderate", "low" or "very low" 50 .
Figure 2 .
Figure 2. Postoperative Joint-Specific Outcome Data (N = 66,836].(A) Forest plot of mean difference showing Total Hip; (B) Meta-regression by mean waiting time; (C) Meta-regression sensitivity analysis of Nikolova et al. 2016 by mean waiting time.(N = 4,873); (D) Funnel plot.Outcome measures reported included: Western Ontario and McMaster Universities Osteoarthritis Index score (Yellow); Oxford Hip Score (Orange); Oxford Knee Score (Blue), Knee Society Clinical Rating System (Purple), and Knee Injury and Osteoarthritis Outcome Score (Magenta].
Table 1 .
Summary characteristics of included studies.
Table 2 .
Summary of findings from studies reporting changes in clinical functionality and health-related quality of life between baseline and preoperative endpoint.Summary of preoperative reported findings.Bold, italics, bolditalics indicates significant findings pertaining to the impact of waiting time on reported outcomes.
Table 3 .
Summary of findings from studies reporting clinical functionality and health-related quality of life with a postoperative endpoint.Summary of postoperative reported findings.Bold, italics indicates significant findings pertaining to the impact of waiting time on reported outcomes.
Table 4 .
Summary of findings from studies exploring patient perspectives on prolonged waiting times.Summary of reported patient perspectives.NHS: National Health Service; TKR: Total Knee Replacement; THR: Total Hip Replacement; PHQ-9: 9-Item Patient Health Questionnaire; OHS: Oxford Hip Score; OKS: Oxford Knee Score; PT: Physiotherapy. | 2024-04-07T06:16:56.157Z | 2024-04-05T00:00:00.000 | {
"year": 2024,
"sha1": "8ca1df4fb47e9486e0a4e3bf3180b4df3a153c48",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "b2a42627aed49417ff642bad5022d9114b51c945",
"s2fieldsofstudy": [
"Medicine",
"Psychology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
235666753 | pes2o/s2orc | v3-fos-license | Research on Usability of the Toolbar of Teaching Interactive Smart Tablet
The work aims to find out the usability problems in the design of the toolbar of teaching interactive smart tablet, and provide an improvement basis and direction for improving the user experience of the teaching interactive smart tablet. Usability testing and user interview were used to research the toolbar of Seewo teaching interactive smart tablet. Quantitative analysis of task completion time, task success level and task completion efficiency were measured in usability testing. The thesis combines the results of user interviews with the results of usability testing to find out the specific factors that affect the usability of the toolbar of teaching interaction smart tablet. The icon design of the toolbar of teaching interactive smart tablet has a great impact on usability. The hiding of frequently used functions may make it difficult for users to find them. The larger screen makes it difficult to find icons outside the field of view. The usability of interactive smart tablet can be improved by improving icon recognition, optimizing common interactive processes, and setting important functions in the best visual field.
Introduction
More and more multimedia equipment has been served in the teaching process. Traditional multimedia classrooms are gradually transforming into intelligent classrooms, and the demand for intelligent teaching equipment is gradually increasing. Many companies have launched specialized intelligent interactive devices for teaching, such as Seewo (about 40% market share), HiteVision (about 20% market share), ChangHong (about 10% market share). However, it is found that there are still some usability problems in the use of these devices. In order to improve the operation convenience and use efficiency of teaching interactive smart tablet, this paper chooses Seewo teaching interactive smart tablet which occupies a high market share to study the usability of its toolbar.
For the concept of usability, different scholars have given different definitions. The International Organization for Standardization (ISO 924-11) have defined usability as "the effectiveness, efficiency and satisfaction of a product when a specific user accomplishes the target to be determined under a specific use situation" [1] . Usability testing, on the one hand, is very helpful in improving the system interface design, discovering the main problems in the interactive framework and refining button labels, operation sequence and priority, etc. [2] On the other hand, it is also an important reflection of user-centred design idea [3] . Nielsen J believes that usability evaluation can be regarded as an irreplaceable practical study, which directly reveals how users use the interactive design system through testing. In contrast to questionnaire, user interview and user diary, usability testing can avoid the subjective influence of users as much as possible [4] . For the evaluation of usability, several scholars proposed different evaluation systems [5][6][7][8] . The existing research mostly focused on the 2 teaching application and system design of teaching interaction smart tablet and similar products [9][10][11] , barely on their usability. This paper mainly used the task completion time, task success level, task completion efficiency and other relevant indicators which is proposed by Tom Tullis, Bill Albert [8] , and combined with the user interview to measure and make a comprehensive evaluation the usability of the toolbar of teaching interactive smart tablet.
Experimental environment and materials
Seewo teaching interactive smart tablet is used as the experimental object in the experiment. The screen size is 86 inches, and the length, width and thickness are 1958 mm, 1160 mm and 89 mm, respectively. This experiment was carried out in a smart classroom, and the teaching interactive smart tablet was installed in the form of a bracket, without any other interference in front of it. Seewo digital tablet is the main interface. When not used, the toolbar is hidden on both sides of the teaching interactive smart tablet, and the left and right hidden ICONS are symmetrical to each other. The hidden ICON on the right toolbar was shown in Figure 1 (a). The toolbar expansion diagram is shown in Figure 1 (b). The circular icons from the top to the bottom of Figure 1 (b) means "Drawing Board", "Paint Brush", "Screenshot", "Zoom In", "Other Tools" and "Half Screen", respectively. The experimental process was recorded by video recording equipment, so that the experimental results could be viewed and supplemented when analysed.
(a) The hidden icon of the toolbar (b) The expanded view of the toolbar Figure 1. the toolbar of Seewo teaching interactive smart tablet
Selection of subjects
In order to ensure the accuracy of the usability test of the Seewo teaching interactive smart tablet, people who have never used the Seewo teaching interactive smart tablet were selected as the subjects in this experiment. Seventeen subjects, who have normal physiological functions, and aged between 22 and 30, were selected for the experiment. All of them were undergraduates or graduate students. The subjects were familiar with electronic teaching equipment, but they had never used the Seewo teaching interactive smart tablet. So they could meet the experimental needs. All the subjects signed the informed consent, and the final valid data of each task in the experiment was more than 15.
Preparation of experimental tasks
Before the experiment, we summarized the functions of Seewo teaching interactive smart tablet toolbar and communicated with 5 teachers who often used the teaching interactive smart tablet (among whom, one is a primary school teacher, two are junior high school teachers and two are university teachers). According to the results of communication, 9 frequently used functions were selected as experimental test tasks, as shown in Figure 2.
2.2.Experimental process 2.2.1. Equipment adjustment and task explanation
Relevant parameters of the video recording equipment and the teaching interactive smart tablet were adjustment before the experiment, so that the initial parameters of the teaching interactive smart tablet (such as the shape and colour of the paint brush) at the beginning of the experiment were the same for each subject. The subjects were numbered 001 to 017, and the experiment content was explained to them. While one of the subjects was doing the experiment, other subjects took a rest in another classroom to avoid the interference of bystanders on the subjects in the experiment and the influence on the following subject.
Task releasing and execution
After the experimenter confirmed, the experiment would start, and the subjects were informed of the experiment task by voice. During the testing, the subject can try the uncertain operation. When the subject completed a task by himself, completed the task under a prompt, or indicated to give up the task, this task testing ended, and the experimenter informed the next experimental task. After all the 9 tasks were completed, the subjects could explore the tasks which they found confusing in the Figure 2. A total of 17 experimental records were obtained.
User interview
The subjects were interviewed after experiment. The contents of the interview mainly included the overall evaluation of the usability of the toolbar of teaching interactive smart tablet, the confusion in the process of using the toolbar and the suggestions for improvement of the toolbar. After the interview, the usability test ended.
Task completion time
According to the observation in the experiment and the video review, the completion of each task was sorted out, and the completion time and standard deviation of each task were sorted out in Table 1. The subjects who did not complete the task or completed the task after asking for help were defined as task failure. The time of task failure was not included in the statistics since the completion time of task failure did not need to be considered in the analysis of the success level of the follow-up task and the completion efficiency of the task, meanwhile, the completion time of task failure was of little reference significance. It can be seen from the average completion time of Task 1 (find the toolbar), Task 4 (find other tools) and Task 6 (open the drawing board), the average completion time is longer and the standard deviation is larger, indicating the completion of these tasks took a long time in the task process and the difference in time was large. The average completion time of Task 2 (find the blue brush), Task 3 (find the page zoom) and Task 8 (move the whole page) is relatively short. Since when the task fails the time data is not included, the average task completion time and standard deviation cannot be used to directly judge the usability of each task-related function of the toolbar. In the analysis of task success level and task completion efficiency, the usability of task-related functions will be evaluated by combining the average completion time of each task with the completion situation or completion rate.
3.2.
Task success level When dividing the task success level, the 4-point scoring method proposed by Tom Tullis and Bill Albert will be referred to [8] . On the four-point success scale, the number 1 means no problems, 2 means minor problems (completing a task after taking a short detour), 3 means major problems (completing a task after taking a long detour), and 4 means failure or giving up. Based on the actual requirements of usability testing of teaching interactive smart tablet toolbar, this paper formulated the task success grading standards as Table 2. According to the classification criteria, the task success levels of the nine tasks were divided, and the proportion of success levels of each task was shown in Figure 3. The average success level of each subject in the same task was calculated according to Formula 1, and the average success level obtained was shown in Table 3. The task success level is sequential data, the levels are not strictly equidistant. Therefore, when the average task success level of the two tasks is not quite different, it cannot be judged which task is better completed. Only when the average task success level of two tasks differs greatly, the usability of the corresponding functions of the two tasks can be compared through the average success level. The task is completed successfully and took less than 10s 2 The task is completed by self after one error, or it takes between 10s-20s 3 The task is completed by self after multiple mistakes, or take more than 20s 4 The task is incomplete or completed after seeking tips from the experimenter Average success rating ∑ Task success rating of subjects Number of valid data 1 Figure 3. Percentage of success level for each task As can be seen from Figure 3 and Table 3, task 1 (find the toolbar), task 2 (find the blue brush), task 4 (find other tools) and task 6 (open the drawing board) have the small percentage of success level 1 (that is, successfully completed), and have a high value of success level and poor usability. Task 3 (find page zoom), Task 5 (run the countdown), Task 8 (move the whole page.), and Task 9 (add a page) have a lower success rate and better usability.
Task completion efficiency
This study adopted the efficiency measurement combining task completion rate and task completion time to measure task completion efficiency [8] of the toolbar of Seewo teaching interactive smart tablet. This calculation method is the task completion rate/the average task completion time (for those who successfully complete the task). The higher the efficiency value is, the better the usability is. The time unit was selected according to the actual situation. After selection, the efficiency percentage must be between 0-100. This paper chose 10s as a time unit, and the efficiency measurement results of each task were shown in Table 4. It can be seen from the efficiency measurement results that the efficiency of task 1 (find the toolbar), task 2 (find the blue brush), task 4 (find other tools), and task 6 (open the drawing board) is low, that is, the usability is poor. Task 3 (find page zoom), task 5 (run the countdown), and task 8 (move the whole page) have higher task efficiency, that is, better usability. The task success level and task completion efficiency of each task were combined for comparison as Table 5. It can be seen that there is a reliable correlation between the ranking of task success level and task completion efficiency, but they are not completely consistent. The Pearson correlation coefficient of task success ranking and task completion efficiency ranking is represented by r. The calculation method of r is as in formula 2, where is the success rank ranking of task i, is the average number of success ranks of each task, represents the ranking of the completion efficiency of the task i, is the average of the completion efficiency of each task, and the value of Pearson's correlation coefficient r and the definition of correlation were shown in Table 6 [12] . Pearson correlation analysis was performed with SPSS software, and the value of r was 0.8. It can be considered that there is a strong correlation between the two, and the results of task success level and task completion efficiency can be mutually verified. The task success level and the completion efficiency ranking are not completely consistent. This is because the task success level considers more of errors and the completion status of the task completion process. When more attention is paid to errors during task completion, usability can be measured in terms of task success rating, and when more attention is paid to task completion time, usability can be measured in terms of efficiency. ∑ ∑ ∑
User interview results
In order to further understanding the reasons why the subjects encountered problems when using the Seewo teaching interactive smart tablet toolbar, user interviews were conducted with each subject after the experiment. The opinions and suggestions put forward by the subjects were shown in Table 7. It can be seen that tasks with poor performance in task success level and completion efficiency were mentioned more frequently in user interviews, and the interview results are consistent with the usability test results. According to the analysis of usability experiments and user interview records, the main reason for the poor usability of task 1 (find the toolbar) is that the tool bar prompt icon is small when hidden, so it is difficult to find. The reason for the poor usability of task 4 (find other tools) and task 6 (open the artboard) is that the meaning of the slogan is not clear, and it is difficult for the subjects to intuit the actual function corresponding to the icon. The main reason for the poor usability of Task 2 (find the blue mark) is that the colour switching function of the brush is hidden, and it is difficult for participants to find the colour switching. At the same time, during the interview process, problems that were difficult to find in the usability testing will be found, such as the difficulty of seeing the content of both parties due to the large screen of the teaching interactive smart tablet, and the need to walk back and forth in the process of completing the tasks.
Suggestions for usability improvement
Based on the results of usability tests and user interviews, the following suggestions has been proposed to improve the Seewo teaching interactive smart tablet toolbar : (1) Enlarge or change the hidden icons in the toolbar to make it more eye-catching; (2) Redesign the icon styles of "other tools", "drawing board", and "zoom" to make their semantics more clear and more compatible with actual functions; (3) Cancel the hidden form of the brush colour change, and set the colour change next to the brush; (4) Change the name of "Roaming" for moving the entire screen to "Mobile Screen"; (5) Place important functions in a prominent position.
Conclusion
Through the study on the usability of Seewo teaching interactive smart tablet toolbar, it can be found that the icon design of teaching interactive smart tablet toolbar has a significant impact on the usability. The unreasonable interactive process will reduce the usability. Ergonomic problems caused by large screen and page layout problems will also affect the usability. In the future, when improving the usability of the toolbar of the teaching interactive smart tablet, we can start from three aspects: improving the icon recognition, optimizing the interaction structure, and setting important functions in the best field of view. There are more and more applications of teaching interactive smart tablets. Research on usability of interactive smart tablets for teaching will provide guidance for the design of teaching interactive smart tablets, so that users can get a better teaching experience. | 2021-06-29T20:03:34.704Z | 2021-06-01T00:00:00.000 | {
"year": 2021,
"sha1": "1b92dc0d9242bf88677501da116857261a5f603d",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1742-6596/1948/1/012150",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "1b92dc0d9242bf88677501da116857261a5f603d",
"s2fieldsofstudy": [
"Computer Science",
"Education"
],
"extfieldsofstudy": [
"Physics"
]
} |
9362007 | pes2o/s2orc | v3-fos-license | Quantitative assessment of cardiac load-responsiveness during extracorporeal life support: case and rationale
We describe a case of a patient assisted by extracorporeal life support, in which we obtained the dynamic filling index, a measure for venous volume during extracorporeal life support, and used this index to assess cardiac load-responsiveness during acute reloading. While reloading, the obtained findings on cardiac pump function by the dynamic filling index were supported by trans-esophageal echocardiography and standard pressure measurement. This suggests that the dynamic filling index can be used to assess cardiac load-responsiveness during extracorporeal life support.
Background
The successful use of extracorporeal life support (ELS) for cardiopulmonary assist and as bridge to myocardial recovery has been shown in many cases [1][2][3]. Unfortunately, weaning from extracorporeal life support often relies on preconceived protocols, if any, and is largely based on trial-and-error and limited data on actual cardiac function.
The common procedure to assess cardiac recovery during ELS, is the use of stress tests. These tests induce an increased volume load on the myocardium and address the Frank-Starling response. This response can be visualized using trans-esophageal echocardiography (TEE), and further quantified by measurement of blood pressure and cardiac output [4,5]. Volume loading is achieved by reducing ELS flow, upon which, depending on the degree of recovery, the heart will increase its output and maintain arterial and venous blood pressure ( Figure 1). In patients whose heart has not (yet) recovered the response is blunted, which causes low output, decreased arterial blood pressure, and increased venous volume.
Recently, we developed a technique to quantitatively assess venous volume that can be (potentially) drained by the centrifugal pump-based ELS circuit, and introduced the dynamic filling index (DFI, in ml/rotation) for the purpose of optimizing ELS flow [6][7][8]. DFI measurement uses periodical, transient reductions (-100 rpm) of pump speed, each lasting approximately 10 seconds, superimposed on the steady state pump speed. The resultant changes in bypass flow are used to calculate the DFI as Δflow/Δspeed (Figure 3 in [7]). When venous volume increases the DFI increases, and vice versa. We found that the DFI is more sensitive to detect changes in venous volume than other routinely recorded hemodynamic and pump related parameters, like central venous pressure. This case shows a patient supported by ELS in which several events of cardiac reloading were performed. In two events, we added DFI measurement to the monitoring of the patient. Before, during, and after reloading, changes in venous volume were quantified using DFI measurement. Under these circumstances, the DFI adequately represented cardiac load-responsiveness as confirmed by TEE.
Case Presentation
We present the case of a patient with a severe myocardial infarction of the anterior wall. Despite urgent percutaneous coronary intervention and the institution of intraaortic balloon support, the patient developed cardiogenic shock refractory to medical therapy. To stabilize the
Observations during cardiac reloading
We show a less successful and a more successful outcome of on-pump cardiac load-responsiveness. Cardiac reloading was performed by reducing bypass flow by approximately 40-45% with concomitant inotropic support. DFI measurements were performed during full unloading (bypass flow >4 l/min), during reloading (bypass flow ≈2 l/min), and again during subsequent full unloading (bypass flow >4 l/min). The number of DFI measurements for each level of support ranged from 5 to 15, and was 12 on average. The hemodynamic condition and corresponding mean DFI in the two events of reloading are presented in Figure 2.
As support flow was reduced during the first event on day 3, from 4.2 to 2.2 l/min (Figure 2, left), the patient became hemodynamically unstable despite intravenous dobutamine administration. TEE showed poor ventricular function. Radial artery pressure dropped (13%) and central venous pressure rose (44%). The DFI increased by 30%, from 1.8 ± 0.2 to 2.3 ± 0.2 ml/rotation (p<0.001). To stabilize the patient again, flow was increased back to 4.1 l/min. The DFI subsequently decreased by 35%, returning to 1.7 ± 0.2 ml/rotation (p<0.001). The second reloading event described here was started at day four (Figure 2, right). ELS flow was reduced from 4.0 l/min to 2.4 l/min. Parallel TEE indicated improved ventricular wall contrac- Figure 1 Rationale behind the use of venous volume measurement (dynamic filling index, DFI) in testing cardiac load-responsiveness. With acute reduction of pump flow (e.g. from 4 to 2 l/min), the patient remains hemodynamically stable if cardiac load-responsiveness is sufficient (upper left). If cardiac load-responsiveness is blunted, the heart is not able to take over output, and venous volume will increase (upper right). Venous volume modulates the relation between pump speed (s) and flow (f) in the assisted circulation, which can be quantified by the DFI, i.e. the slope of the curve Δf/Δs. The DFI will not increase with diminished ELS flow when cardiac load-responsiveness is good (lower left), but will increase if cardiac load-responsiveness is decreased (lower right). CVP, central venous pressure; ABP, arterial blood pressure; RV, right ventricle; LV, left ventricle; SVR, systemic vascular resistance. tility, and the patient remained hemodynamically stable with a dobutamine infusion set to 5 μg/kg/min. This time, the DFI did not change significantly (p>0.05). Although the DFI readings were not directly included in the evaluation of load-responsiveness, they did independently and quantitatively confirm the increased cardiac load-responsiveness as judged from the TEE and pressure data. After the positive result of this stress test, pump flow was set back to 4.0 l/min to give the heart additional time to recover.
During the next four days, central venous and radial artery pressure indicated improved hemodynamic stability. An improvement of left ventricular function was confirmed by TEE. Based on the observation that cardiac load-responsiveness improved, further DFI measurements were stopped. After three more days of ELS, the patient was weaned from the extracorporeal system with the balloon pump (1:1) in place for additional afterload reduction, and on inotropic support (milrinone 25 μg/kg/ h).
Discussion
This report presents a patient assisted by extracorporeal life support (ELS) in which we demonstrated the use of the dynamic filling index for the quantitative assessment of cardiac load-responsiveness.
The common procedure to assess weanability from cardiac support consists of reduction of device flow, while intensivists, cardiac surgeons, cardiologists and perfusionists are consulted to confirm the rather subjective findings on cardiac recovery. This clinical routine is addressed by Schmid et al. [9], who state that weaning protocols are more exemption than routine, and that patient selection, diagnosis of adequate myocardial recovery, and timing of explant surgery are still unresolved. Pitsis et al. described intermittent unloading and loading of the heart in LVAD-supported patients [10], which can be regarded as cardiac muscle training prior to weaning. In such settings, TEE can be used to visualize cardiac function [4]. TEE, however, provides only a snap shot of the current situation, whereas pressure and cardiac output are measured continuously. In that context, Hoshi et al. described quantitative assessment of cardiac function by continuous measurement of the eccentric displacement of the centrifugal pump impeller [11]. Their measuring principle, however, can only be applied to fully magnetically levitated assist pumps. In all these cases, DFI measurement may support the TEE and hemodynamic observations by providing instant and quantitative information on cardiac pump function. Moreover, DFI measurement is by design applicable to all types of centrifugal pumps. When performed in a regular fashion, the DFI might even enable optimizing the actual weaning process.
In the case presented here, central venous pressure and DFI showed comparable changes in response to ELS reduction and indicated similar changes in cardiac loadresponsiveness. Central venous pressure measurement, however, depends on catheter placement, i.e. the location of pressure measurement (jugular vein, caval veins, atrium). When pressure is measured near the inlet of the venous ELS cannula, a change in central venous pressure read-out will, along with hydrostatic pressure, also reflect the rate of suction by the ELS circuit, rather than cardiac preload. Therefore, changes in central venous pressure may in some cases not reflect true changes in cardiac preload. The DFI is a measure of on-pump venous volume [7], and changes in DFI during reduced unloading (i.e. reduced ELS flow) could be a more reliable measure than central venous pressure. We previously found that DFI is more sensitive to small changes in venous volume than pump inlet pressure/central venous pressure [8]. Therefore, if a reliable central venous pressure is unavailable, DFI can be an alternative to assess cardiac load-responsiveness during ELS. We are aware that we show only a single case in which the DFI was used to support the regular hemodynamic monitoring. More clinical investigations are necessary to establish the additional value of the DFI in the presence of regular ICU parameters. Considering the ease-of-use of DFI measurement and the promising findings in this patient, we intend to study regular DFI measurement in patients assisted by ELS, which, considering the incidence of ELS, will likely require a multicenter study.
Consent
Informed consent was obtained, and the use of patient data for this case report was approved by the medical ethical committee of the Maastricht University Medical Center. | 2014-10-01T00:00:00.000Z | 2010-04-27T00:00:00.000 | {
"year": 2010,
"sha1": "13e0ceae03feb735b995ddb71fe0c4f3dbfdef7e",
"oa_license": "CCBY",
"oa_url": "https://cardiothoracicsurgery.biomedcentral.com/track/pdf/10.1186/1749-8090-5-30",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "c2795c7a8535f2f76dea61067a3d8c687cfb8046",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
13192994 | pes2o/s2orc | v3-fos-license | To what extent are surgery and invasive procedures effective beyond a placebo response? A systematic review with meta-analysis of randomised, sham controlled trials
Objectives To assess the quantity and quality of randomised, sham-controlled studies of surgery and invasive procedures and estimate the treatment-specific and non-specific effects of those procedures. Design Systematic review and meta-analysis. Data sources We searched PubMed, EMBASE, CINAHL, CENTRAL (Cochrane Library), PILOTS, PsycInfo, DoD Biomedical Research, clinicaltrials.gov, NLM catalog and NIH Grantee Publications Database from their inception through January 2015. Study selection We included randomised controlled trials of surgery and invasive procedures that penetrated the skin or an orifice and had a parallel sham procedure for comparison. Data extraction and analysis Three authors independently extracted data and assessed risk of bias. Studies reporting continuous outcomes were pooled and the standardised mean difference (SMD) with 95% CIs was calculated using a random effects model for difference between true and sham groups. Results 55 studies (3574 patients) were identified meeting inclusion criteria; 39 provided sufficient data for inclusion in the main analysis (2902 patients). The overall SMD of the continuous primary outcome between treatment/sham-control groups was 0.34 (95% CI 0.20 to 0.49; p<0.00001; I2=67%). The SMD for surgery versus sham surgery was non-significant for pain-related conditions (n=15, SMD=0.13, p=0.08), marginally significant for studies on weight loss (n=10, SMD=0.52, p=0.05) and significant for gastroesophageal reflux disorder (GERD) studies (n=5, SMD=0.65, p<0.001) and for other conditions (n=8, SMD=0.44, p=0.004). Mean improvement in sham groups relative to active treatment was larger in pain-related conditions (78%) and obesity (71%) than in GERD (57%) and other conditions (57%), and was smaller in classical-surgery trials (21%) than in endoscopic trials (73%) and those using percutaneous procedures (64%). Conclusions The non-specific effects of surgery and other invasive procedures are generally large. Particularly in the field of pain-related conditions, more evidence from randomised placebo-controlled trials is needed to avoid continuation of ineffective treatments.
INTRODUCTION
Surgery and other invasive procedures such as endoscopy and percutaneous procedures are widely used in medicine but their specific efficacy and risk-benefit profile are rarely assessed in rigorous and systematic ways. The development of minimally invasive procedures has expanded the use of such interventions for treating a variety of conditions such as low-back pain, 1 arthritis, 2 endometriosis, 3 Parkinson's disease, 4 gastro-oesophageal reflux 5 and obesity. 6 Strengths and limitations of this study ▪ This is the first systematic review using a meta-analysis approach to estimate both specific and non-specific components in sham-controlled surgical trials, and to what extent those effects differ among conditions and procedures. ▪ All sensitivity analyses showed similar results as the main analysis, except one, namely the sensitivity analysis for large studies (≥100 patients), which showed a smaller non-significant effect size. ▪ Our results have implications for clinical research and practice by arguing against the continued use of ineffective invasive treatments, especially in the field of chronic pain. ▪ One limitation might be that the conclusions from our meta-analysis are restricted to available published data on surgical interventions that have been tested in sham-controlled clinical trials.
Rarely are these procedures evaluated using rigorous research designs involving randomisation, allocation concealment and blinding or placebo controls, which are considered gold standards for medical interventions. In the absence of controls for common sources of bias, studies on these procedures may give a false impression of their true efficacy. Is it possible to test invasive procedures using rigorous methods? Blinding of outcome assessment is challenging since mimicking a complex, invasive procedure such as surgery, or insertion of a scope or a needle, requires an elaborate sham procedure. Moreover, there is significant controversy over the ethics of using sham procedures, even with carefully informed patients, further restricting the number of such studies being carried out. 7 8 However, can we justify widespread use of these procedures without rigorous testing?
The use of blinded, sham procedures permits rigorous assessment of treatment efficacy by comparing the outcome in the treatment and sham groups. Specifically, sham procedures control for a variety of observed outcomes in the sham group that are distinct from the specific efficacy of the surgery or invasive procedure under investigation. These 'non-specific' outcomes include placebo responses (also sometimes called placebo effects), which we define here as the observed outcome changes in the sham groups. These changes are due to the natural history of the patient's condition or regression to the mean and a response to the ritual of medical treatments. Such rituals include the type of procedure ( pill, needle, knife or touch), the status, authority and communication style of the provider, the setting and context of the treatment and the patient's and practitioners expectation about the outcome. 9 Yet, invasive procedures are thought to incorporate many factors that may contribute to the placebo responses including use of a hospital-like setting; multiple, authoritative providers; frequent and repeated suggestions about expected outcomes; a physical invasion of the body; and an elaborate ritual of treatment delivery and recovery. 10 Thus, one would expect a significant contribution from surgical ritual and other non-specific factors to the observed outcomes during invasive procedures in clinical practice and in randomised trials without sham control groups. Several high profile studies support this hypothesis in which sham procedures involving only superficial anaesthesia were compared to the more invasive true procedure. [11][12][13] For example, Moseley et al 11 reported no greater pain improvement in patients with osteoarthritis of the knee that underwent arthroscopic knee surgery compared to a sham procedure in which a cut was made over the knee without introducing the arthroscope. Two more recent controlled studies of vertebroplasty for painful osteoporotic vertebral fractures reported similar degrees of pain relief from sham procedures involving only superficial anaesthesia compared to the more invasive active procedures. 12 13 In contrast, a systematic review comparing surgical with non-surgical treatments for painful osteoporotic vertebral fractures came to the conclusion that vertebroplasty and kyphoplasty are superior to non-surgical treatments. 14 Since invasive interventions frequently go along with larger non-specific effects than non-invasive treatments 15 16 surgical trials that do not include a sham surgery arm may give biased results. Thus, the efficacy of invasive procedures, for example, for chronic pain conditions, remains controversial. 17 In addition, many invasive procedures involve the risk of anaesthesia and high cost. 17 Therefore, it is important to estimate to what degree the observed outcomes from invasive procedures are due to specific efficacy of the treatments or to other factors.
To better understand these issues we conducted a systematic review and meta-analysis of studies on surgery and invasive procedures in which a parallel sham procedure was included for comparison. Our study aims were to: (1) assess the quantity and quality of such studies; (2) estimate the magnitude of specific effects over sham procedures; and, (3) estimate the contribution of the surgical ritual and other non-specific factors to outcomes from these procedures.
Identification of studies
The following online databases were searched from their inception through January 2015: PubMed, EMBASE, CINAHL, CENTRAL (Cochrane Library), PILOTS, PsycInfo, DoD Biomedical Research, clinicaltrials.gov, NLM catalog, as well as NIH Grantee Publications Database. We used as our initial search terms: 'Diagnostic Techniques, Surgical' OR 'Orthopedic Procedures' OR 'Specialties, Surgical' OR 'Surgical Procedures, Operative' OR 'surgery' (Subheading) or surgery) AND ('Placebos' OR 'Placebo Effect' or sham surg* or placebo surg* or mock surg* or simulated surg* or placebo proc* or sham proc* or mock proc* or simulated proc*). We restricted our search to humans and randomised controlled trials. Variations of these search terms were made for MESH terms, where necessary, and are available on request from the first author.
The 'Grey literature' was searched by looking for relevant dissertations, conference proceedings, Google Scholar and searching the internet using the keyword scheme as well as searching all relevant reference lists of identified articles and related reviews. We also contacted and consulted with leading experts in the fields of surgery and placebo, and shared databases that these experts have collected over the years relating to placebo to make sure we captured all the relevant literature.
Eligibility criteria
Studies were included in the systematic review if they: (1) were randomised controlled trials; (2) involved a population for which there was a symptom-driven medical condition for which an invasive procedure or classical surgery as defined below was being performed; and (3) had a comparison group that used a sham procedure to mimic the real procedure.
Classical surgery was defined as a procedure that followed the typical surgical experience that uses preoperative preparation, anaesthesia, an incisional trauma (usually through muscle and fascia and into the peritoneum) and a postoperative recovery process. Invasive procedures were defined as when an instrument was inserted into the body (either endoscopically or percutaneously) for the purpose of manipulating tissue or changing anatomy. In all cases we selected studies where when these procedures were compared to a sham procedure that used the same surgical or invasive procedure, instrument and ritual, but eliminated the hypothesised active component of tissue manipulation. We excluded studies in which the procedure was used simply as a delivery mechanism for another ongoing active treatment such as a pacemaker, brain or cardiac stimulation, or delivery of a drug or biological product. Studies where an invasive procedure was implemented for prevention of a medical condition or there was no symptom-driven condition were also excluded.
Four investigators (CC, LC, KL and KM) screened titles and abstracts for relevance in two phases based on the inclusion criteria: phase one eliminated all clearly irrelevant studies, phase two applied all inclusion/exclusion criteria listed above for the remaining studies. Any disagreements about including a study were resolved through discussion and consensus, and approved by the first author (WJ). All reviewers were fully trained in systematic review methodology. At least two reviewers had to review each citation in order for it to progress to the next phase of the review. A Cohen's κ on agreement was attained for both phases above 88%.
Quality assessment and data extraction
The methodological quality of the individual studies (sequence generation, allocation concealment, was assessed independently by three reviewers using the Cochrane Risk of Bias (ROB) tool. 18 Descriptive data was independently extracted on the following items: population; condition for which surgery was performed; sample ( population) entered; dropout rate; informed consent details; whether a power calculation was performed and achieved; intervention and sham procedure used; primary and secondary outcomes and the statistical data associated with these; whether expectation was reported; author conclusions; adverse events reported; funding source, and reviewer comments. We also extracted from each study, if available, a continuous and a dichotomous main outcome at two time points (intermediate and late), and a continuous and a dichotomous pain outcome (when applicable). The most important outcome measure (miOM) was defined as either: (1) the primary main outcome measure ( pMOM) at a time point as predefined in the trial; or (if not 1), (2) the only major outcome of a trial at the latest available time point; or (if neither 1 nor 2), (3) the clearly most relevant outcome determined by two independent reviewers at the latest available time point. Secondary outcomes were intermediate time points of the most important outcome measure; pain outcomes at the latest available time point; or, pain outcomes at the intermediate time point. All discrepancies were tracked by the review manager and were resolved by consensus and discussions during team meetings. Data were entered into a webbased, secure, systematic review management programme called Mobius Analytics SRS (Mobius Analytics Inc, Ottawa, Ontario, Canada).
Data synthesis and analysis
According to our analysis plan, the meta-analyses focused on continuous outcomes. The primary analysis was based on trials reporting a most important continuous outcome measure in sufficient detail to be included in the meta-analysis. Secondary analyses were based on trials reporting (1) mentary table 1, and a sensitivity analysis was computed for these outcomes (see below).
Within-group and between-group effect sizes were based on Cohen's 19 d for change within one group, and Cohen's d for between-group effect measures, respectively, correcting for small-sample bias. 20 In order to keep the effect size framework coherent for within-group and between-group designs, change from baseline was used throughout. When SD was not reported, it was calculated from pre-SD and post-SD, 21 using r=0.5 for the product-moment correlation between pre and post measures.
Analyses of continuous data were performed with the generic inverse variance module of the Cochrane Collaboration's Review Manager software (V.5.1), using standardised mean difference (SMD) as the effect size measure. As we expected heterogeneity, a random effects model was used. Within-group effect sizes were pooled in such a way that positive values indicate improvement, while positive values of between-group effect sizes indicate superiority (more pronounced improvement) of the intervention group over the control (sham) group. To estimate the relative contribution of non-specific outcomes to treatment effects, the per cent ratio of the pooled within-group treatment effects in the sham and the treatment groups was calculated. We used Cochrane's Q test and calculated I 2 to examine statistical heterogeneity, with low, moderate and high I 2 values of 25%, 50%, and 75%. 22 Egger's test was used to assess funnel plot asymmetry. 23 A p value of less than 0.05 was set as the level of significance. 24 25 Subgroup analyses were performed according to predefined categories of target diseases and types of surgery. To check the robustness of results, we performed sensitivity analyses with four criteria: (1) studies specifying a primary main outcome measure ( pMOM); (2) imputing 0.3 and 0.7 for pre-post correlation coefficient r, when missing; (3) studies with total sample sizes ≥100; and, (4) studies with low risk of allocation concealment. An additional sensitivity analysis was performed for dichotomous outcomes of 12 studies that provided no continuous outcome (see online supplementary figure 1).
Eligible studies
Our search identified a total of 7360 citations. After excluding clearly irrelevant references the full text of 113 publications were obtained. Of these, 46 were excluded, mainly for not including an instrumental or surgical intervention or a sham procedure as defined above. A total of 55 studies (in 67 publications) involving a total of 3574 enrolled patients met our inclusion criteria for systematic review (figure 1). 26 Characteristics and quality of included studies Characteristics of the included studies are summarised in online supplementary table 1. About half (25) of the studies were carried out on pain-related conditions with back pain (7) being the most frequent 11 12 27-31 followed by arthritis (4), 13 32-34 angina from coronary artery disease (4), [35][36][37][38][39] (2) 48 49 and migraine (2). 50 51 ) The most frequently studied non-pain condition was obesity, especially when using balloon insertion (11). [52][53][54][55][56][57][58][59][60][61][62] Other conditions that had more than one study included gastroesophageal reflux disease (GERD) (5), [63][64][65][66][67] Parkinson's Disease (2), 68-74 sleep apnoea (2), 75 76 dry eye (2) 77 78 and asthma (2). [79][80][81] Some other conditions were also studied (see online supplementary table 1). [80][81][82][83][84][85][86][87][88][89][90] of the studies involved endoscopic or percutaneous procedures in which tissue was removed or altered or a material (eg, dye, cement, balloon) was inserted. 11-13 28 31 34 38 40-43 52 54 56 61 63 65 67 77-79 90 Some of these procedures used a catheter to reach an internal organ (such as the heart or gall bladder) or a needle to inject a material or cell (often into the lumbar spine or Figure 1 Flow chart of included studies. RCT, randomised controlled trial. brain). 27 29 30 32 53 55 57 59 60 62 64 66 85 89 Five studies evaluated more classical surgical procedures in which the body was opened with a scalpel or drill. 50 51 74-76 In most studies, blinding was achieved using elaborate sham procedures. Those mimicking classical surgical procedures usually cut the body, leaving a scar but causing less damage than the real surgery. Sham percutaneous and endoscopic procedures often involved superficial insertion of a needle or a scope. For example, in the Parkinson's studies on surgical interventions on the brain, sham procedures involved placing burr holes without penetration of the skull. 68-74 Sham surgery for endometriosis would often involve 'diagnostic laparoscopy' with no internal tissue destruction. Sham balloon insertion for obesity treatment usually involved inserting the balloon but not inflating it. [52][53][54][55][56][57][58][59][60][61][62] Overall, the risk of bias was low in these studies, with some exceptions. Of the 55 studies (67 publications) included in the systematic review, 34 studies (62%) reported an adequate method for generating the allocation sequence, however only 23 (42%) had adequate concealment of allocation. Blinding of the patients and outcome assessors was adequate in 48 (87%) studies and incomplete data was adequately addressed in 52 (95%). Fifty-two (95%) of the studies were free from suggestion of selective outcome reporting and 53 studies were judged to be free of other sources of bias.
Overall analyses
Thirty-nine studies (2902 patients) with continuous data were included in the main analysis. The overall effect of surgery compared to sham surgery was highly significant (SMD 0.34, 95% CI 0.20 to 0.49; p<0.00001), while heterogeneity was large (I 2 =67%, p<0.00001). Excluding one outlier 52 reduced I 2 to 57% (SMD, 0.30, 95% CI 0.17 to 0.43; p<0.00001), indicating moderate heterogeneity. Sensitivity analyses provided comparable effect sizes (figure 2), except for studies with overall sample sizes of 100 participants or more, for which the SMD was nonsignificant at 0.15 (n=10; 95% CI −0.02 to 0.32; p=0.09; I 2 =66%). Inspection of the funnel plot suggests the presence of biases in the meta-analysis, such as small study bias or publication bias (figure 3). Asymmetry in the funnel plot was confirmed by the Egger's test (asymmetry coefficient 1.7, p=0.017).
Subgroups by type of procedure
Between-group SMD did not differ significantly between classical surgery, endoscopic surgery and percutaneous procedures (χ 2 =1.10, p=0.58; results not shown).
Changes from baseline within sham and active groups
The pooled SMD for changes from baseline was 0.61 in the sham groups (95% CI 0.47 to 0.75, p<0.00001, n=39, I 2 =76%) and 0.92 (95% CI 0.74 to 1.09, p<0.00001, n=39, I 2 =86%) in the treatment groups. Thus, on average, the changes in the sham groups accounted for 65% of the overall improvement from the treatments. This proportion of specific to non-specific treatment effects was larger in pain-related conditions (78%) and obesity (71%) than in GERD (57%) and other conditions (57%), and was considerably smaller in classical surgery trials (21%) than in endoscopic trials (73%) and those using percutaneous procedures (64%; figure 4). Changes in the sham groups accounted for 89% and 82% of overall improvement in intermediate and late pain outcomes. not strong and could be changed with more and better research. In addition, the contribution of non-specific effects is even more substantial for certain conditions and procedures. While non-specific effects accounted for approximately 65% of the effects from all invasive procedures, they made up to 78% of the active treatment effects in chronic pain conditions and 71% of the active treatment effects in obesity. These percentages are substantially higher than those observed in non-surgical trials, namely 40% for chronic pain conditions and 33% for obesity. 91 The higher contribution of non-specific effects in surgical trials could well be the result of higher placebo effects. However, the lack of no-treatment groups in our data set (and other data set) 92 allows no firm conclusion. 91 Our subgroup analyses indicate that the current evidence does not support the specific efficacy of invasive procedures for chronic pain conditions ( p=0.08) and was borderline for obesity ( p=0.05), but does support these procedures for GERD ( p=0.0002). However, please note that the analysis of dichotomous outcomes showed a somewhat larger specific effect for pain studies (see online supplementary figure 1). There is insufficient data to make recommendations about the other conditions examined.
DISCUSSION
Strengths and weakness of this study This study has several limitations. First, both the central strength and limitation of our study is that we pooled effect estimates of the included studies. We consider this a strength at is allows us to: (1) make an estimate of the overall effects of invasive procedures in sham-controlled surgical studies, (2) estimate the strength of confidence in the currently available data as to the specific efficacy of those procedures; and, (3) empirically investigate to what extent results differ between conditions and procedures. Obviously, it is not reasonable to expect that surgery has similar specific effects across conditions and outcomes so our subgroup estimates should not be interpreted clinically without considering how the interventions and outcomes varied. This is also indicated by the moderate-to-large heterogeneity in our meta-analyses, indicating more variation of effect sizes than would be expected by chance. Second, it is difficult to fully double-blind invasive procedures. While most studies successfully blinded patients and outcome assessors, physicians doing these procedures could not be blinded. Thus, it is possible that they communicated information to patients that biased the studies. Price and others have shown that physician expectations can influence pain outcomes even when restrictions are placed on verbal communication. 93 94 Third, publication bias may play a role in the accuracy of our estimates. It is known that negative studies (in this case, studies showing no difference between real and sham procedures) are not published as frequently as positive studies. However, our search strategy was comprehensive and the study selection process was reliable. We also conducted a thorough search of the grey literature, as described above, and had input by experts in placebo research, increasing the likelihood of capturing all studies in this area. This activity allowed for a cross-check in the end to ensure we captured most of the relevant published randomised controlled trials for this review. We did not find any unpublished reports that met our inclusion criteria appropriate for this review, however there were some publications that were not readily accessible through the search engines commonly accessed that we were able to capture through these methods. Our sensitivity analyses on study quality factors did not change our primary findings, except restricting the analyses to large studies with 100 participants and above, revealed a considerably smaller, non-significant SMD at 0.15 (95% CI −0.02 to 0.32; p=0.09). Egger's test for funnel plot asymmetry, however, suggested a small study bias in our data set. While our combined estimates of effect size must be considered crude for the overall meta-analysis, they are reasonable estimates for the pain, GERD and obesity subgroups. Meta-analyses of placebo-controlled drug studies in pain, depression, hypertension, ulcer treatment and other areas often report a similar magnitude of specific treatment effects compared to non-specific effects. [95][96][97][98] Those studies, however, usually have much larger sample sizes, increasing confidence in their estimates. Finally, we found only one three-armed study that included no treatment, active and sham groups. 67 Therefore, it is not possible to estimate the contribution that the ritual and context make to outcomes in invasive procedures compared to no treatment. Especially in the field of pain and obesity such three-armed studies would seem to be essential for making good evidence-based decisions.
Our findings are consistent with a systematic review published in the BMJ in 2014. 92 That study, however, used vote count and reported that 74% of 55 trials showed improvement in the placebo arm with 51% reporting no difference between surgery and placebo and 49% reporting surgery was superior to placebo. We have built on that study by doing a more comprehensive literature search and meta-analysis which allowed us to estimate the magnitude of surgical effects, the confidence in the current findings and to examine that magnitude across various quality parameters, conditions, procedures and outcomes. We can now conclude that at least chronic pain conditions lack clear evidence for the efficacy of the explored surgical interventions (eg, classic surgery and endoscopic procedures. Since these conditions represent a high public health burden worldwide we need to obtain better evidence for the use of these procedures. In addition, it is clear that the evidence from placebo controlled trials in the field as a whole is poor.
Implications for practice, research and policy
These results have a number of implications for practice, research and policy. The evidence from available shamcontrolled trials indicates that invasive procedures are not clearly more effective than sham procedures for various chronic pain conditions including endometriosis, back pain, arthritis, angina and migraine. There is evidence to support surgical interventions for GERD and limited evidence to support the use of balloon insertion for obesity.
Given the large number of invasive and surgical procedures being performed, it is noteworthy that we could identify only 55 sham-controlled studies in the literature.
Certainly, not all invasive procedures warrant shamcontrolled comparisons; for example, when results demonstrate indisputable changes in objective parameters the risks of sham procedures would be excessive. However, given that non-specific factors make a large contribution to the effects from invasive procedures for conditions like pain, more rigorous evaluation is needed before their widespread use is recommended for these conditions. A recent survey of surgeon's attitudes about sham surgery may provide an opportunity to conduct more such research. Surgeons generally agreed that a placebo component to surgical intervention might exist. 99 Furthermore, results of a recent systematic review indicate that the risks of adverse effects associated with sham surgical procedures are small. 92 Thus, more welldesigned sham-controlled surgical trials are warranted to avoid the continued use of ineffective invasive treatments.
Contributors WBJ served as the PI on this project and was responsible for the conception and design of the project, obtaining funding, acquisition of data and interpretation of the data, drafting and final revision of the article, and final approval of the version submitted. CC served as the project manager and reviewer, and contributed to the conception and design of the systematic review, acquisition of data and analysis and interpretation of data, drafting the article, and approval of the version to be submitted. LC and KM served as study quality reviewers and contributed to the conception and design, acquisition of data, analysis and interpretation of data, drafting the manuscript and approval of the version to be submitted. In addition, KM led the meta-analysis section of the project. TJK and FGM served as subject matter experts, and were involved in the conception and design and interpretation of the data, revising the manuscript critically for important intellectual content and approval of the version to be submitted. LK served as the statistical expert on the project and was involved in the design and conduct of the meta-analysis, acquisition of data, analysis and interpretation of the data, contributing to the manuscript in statistics, meta-analysis techniques, the results section of the manuscript and approval of the version submitted. KL served in the conception and design, acquisition of the data, analysis and interpretation of the data, design of the meta-analysis technique for extracting data, assisted in drafting and revising the article for important intellectual content, and approval of the version to be submitted.
Funding This work is supported by the US Army Medical Research and Materiel Command under Award number W81XWH-08-1-0615. The views, opinions and/or findings contained in this report are those of the author(s) and should not be construed as an official Department of the Army position, policy or decision unless so designated by other documentation. The funding source had no role in the design and conduct of the study, in the collection, analysis and interpretation of the data, or in the preparation, review, or approval of the manuscript.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement No additional data are available.
Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work noncommercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http:// creativecommons.org/licenses/by-nc/4.0/ | 2016-05-12T22:15:10.714Z | 2015-12-01T00:00:00.000 | {
"year": 2015,
"sha1": "50083f39c24072de71dcb380ea7f1d6f2624f209",
"oa_license": "CCBYNC",
"oa_url": "https://bmjopen.bmj.com/content/5/12/e009655.full.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "b10f933de1a7a5217b2533b88666ea0f99618b5b",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
233704501 | pes2o/s2orc | v3-fos-license | Measurement Engineering to Design a Truncated Ground Plane Compact Circular Ring Monopole Patch Antenna for Ultra Wideband Applications
In this paper, using in-depth simulations and measurements, a simple and compact design is engineered for making a circular ring microstrip patch antenna radiating element which is suitable for different ultra wide band (UWB) applications. This design approach is different because it has not utilized the usual method of using a set of electromagnetic equations and calculations to make the radiating antenna. Measurements and simulations were performed on Microwave CST. Using this measurement engineering approach, novelty of proposed antenna structure is obtained by making the required changes in the ground plane. The measurements showed that truncating the ground plane by a square shape structure of 2.5 mm by 2.5 mm size at the feed point was practically significant to provide an impedance bandwidth S11<-10dB\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left( {S_{11} < - 10\;{\text{dB}}} \right)$$\end{document} ranging from 2.75 to 32.035 GHz with a VSWR which is less than 2. For this entire bandwidth the directivity has shown a variation from 0.8 to 7.9 dBi. The compact size (33 mm × 28 mm × 1.57 mm), low design complexity, very high bandwidth, good directivity and satisfying VSWR has made this antenna unique among all previously presented UWB antennas.
Introduction
Ultra-wideband (UWB) is a radio technology that has the advantages of using very low power spectral density (PSD) with fractional bandwidth ≥ 0.2-0.25 or a bandwidth of at least 500 MHz. This range is usable for high bandwidth applications in engineering, medical sciences and other areas of science [1][2][3][4][5]. Designing a UWB communication system, entails state-of-the-art design of transmitters, receivers and radiating elements, i.e. antennas. In this paper the focus is the design of a contemporary UWB antenna with desired characteristics low volume and size.
3
Antennas are the radiating elements with the purpose of sending and receiving signals to and from wireless channel. In addition to many other features, speed of transmission and reception process carries a challenging research interest for a high speed communication system. Of the many factors, the high speed communication needs a high bandwidth to cover mobile and other wireless facilities. With this need of broad bandwidth, UWB antennas with low profile are the best resource for high speed communications.
In designing a UWB antenna it is important to consider the design principles, procedures and final antenna characteristics. Hence there should be no ambiguity in them. Because of their inherent advantages, the most popular of UWB antennas are microstrip patch antennas [6][7][8]. Microstrip Antenna (MSA) has a three layer structure namely radiating patch, ground plane and substrate (dielectric). The substrate lies between patch and ground plane. Literature study showed that microstrip antenna's type, shape, design analytics, and feeding procedures are the commonly accepted methods to improve the antenna's characteristics and performances [9][10][11][12][13][14][15]. These attributes help to improve the antenna parameters including return loss, gain, directivity, and bandwidth. The design process involves that the antenna designer first determines any of these parameter(s) as the required output then tries to optimize their antenna design accordingly.
Most importantly in the design of the UWB antenna, the working bandwidth should be considered according to the Federal Communications Commission (FCC).The UWB standard used in USA and Canada is unlicensed band of approximately 7.5 GHz ranging from3.1 to 10.6 GHz. In addition, FCC has also limited the emissions for the maximum power-efficiency which is − 41.3dBm/MHz [16,17].
A conventional microstrip patch antenna has the intrinsic properties of narrow impedance bandwidth which means that the antenna will show resonance on a single (ideally) or to a narrow band of frequencies (in practical). Several techniques have been used to enhance the impedance bandwidth of the microstrip patch antenna. Some of the commonly stated techniques in the literature study included an insertion of a slot in a patch in the form of square, ring or U-shaped slot [18][19][20][21][22], truncating the ground plane and designing aperture coupled patch antennas [23,24], using integrated band pass filter [25], making use of optimally designed impedance matching network. In all these configurations, parts of the system that a designer has to model in finding antenna radiation are the surfaces of the metals, having surface charge density, which radiate electromagnetic fields. A low Q factor or alternatively high bandwidth antenna has also been achieved by increasing the thickness of the dielectric substrate [26], thus making more room for radiated energy than stored energy. However, high bandwidth antenna has a lower gain value. But by using an array configuration for the elements, the problem of lower gain and lower power handling capacity can be overcome.
In this research, we have proposed the antenna design and its measurement engineering of a high band width UWB antenna along with a moderate structure size and good directivity bandwidth.
Background/Motivation from Previous Work
In the beginning the parametric values were achieved after the theoretical and numerical analysis of various types of Microstrip Antennas (MSA's). In rectangular patch we have two parameters (Length and Width) for changing their different modes. Similarly in our design, the patch is circular ring type also have two parameters (outer radius R and inner radius r) by which we can change their modes. So we put all the values of our proposed antenna in well known equations of rectangular patch antenna to find the length and width. Here we supposed that the value of width and length is equal to the outer radius (R) and inner radius (r) respectively of the proposed antenna which is circular ring type. The width of rectangular patch can be found as: Here c is the velocity of light and is equal to 3 × 10 8 m∕s. f 0 is the resonance frequency and is equal to 10 GHz. And ε r is the relative permittivity of FR-4 which is used for substrate material, its value is 4.3.
Substituting the values in Eq. (1) give us, Here we have supposed that the value of width in Eq.
(2) will be equal to outer radius ( R ) of our proposed antenna Similarly to find actual length ( L ) the following equations were used, Here h is the height of substrate and is equal to 1.57 mm and ε eff is the effective permittivity.
Here L eff is the effective length of rectangular patch Here ΔL is the change in length of rectangular patch Finally L is the actual length of rectangular patch. After solving the above Eqs. (4), (5), (6) and (7), L is found as, (9) we also calculated the length and width of ground plane and truncation was performed just by rectangle shape in a simple way that one side was made twice to the other side. After getting all initial parametric values, we simulated the antenna and found the bandwidth 0.8512 GHz with reflection coefficient S 11 < −10 dB . The simulation is shown in Fig. 1.
These initial parametric values made it possible to get a bandwidth by simulating our antenna. The obtained bandwidth was quite narrow but provided us with the values to initiate the design process.
Problem Definition
The bandwidth of Ultra-wide band antenna depends on different design techniques e.g. patch shapes and feeding techniques. In this context, lots of configurations have been studied but the large size was the main problem to achieve the appropriate bandwidth. The antenna size and the bandwidth are dependent on each other. If bandwidth needs to be increased, then this requires the increase in the antenna size gradually. However we want antenna size to be compact and minimal.
Several antennas for suitable Ultra wideband applications have been tested such as stacked patch antenna wide slot antenna [27], tapered slot [28] and also other more antennas are designed and investigated in order to get the best fit antenna satisfying UWB requirements.
Some of the major issues of concern in the present available UWB antennas include the following: (9) r = L = 6.26 mm Fig. 1 Frequency versus reflection coefficient ( S 11 ) graph, when outer radius R = 8.89 mm and Inner radius r = 6.26 mm 1. Most of the designs are multifaceted which lead to complexity in the fabrication process. 2. Many tested designs have low band width and for achieving high band width, antenna size has to be increased implicitly [29]. 3. Some designs have good band width but unsatisfactory antenna characteristics and also have greater VSWR values [30].
Considering the above mentioned problems, the aim of this work was research to obtain an antenna which is simple in design, have high bandwidth, satisfactory antenna characteristics, low VSWR and have compact antenna size too; all suitable for ultra wideband applications.
Research Methodology
As discussed above, in the design of an antenna there are many parameters. For a patch antenna, these include the parameters of substrate, patch and ground plane. A complex combination of them can give varying characteristics of beam width, bandwidth, radiation patterns, directivity and gain to an antenna. In this research we developed the research methodology for measuring the effect of the antenna parameters to identify most suitable values of them for meeting the requirements. In this approach of measurement engineering we explored the effect of individual parameters on antenna characteristics. At a single instant only one parameter is made to vary while others remained constant. Following section shows this detailed analysis.
Analysis on the Simulations of Antenna's Parameters
The geometry of the proposed optimized antenna is shown in Fig. 2 and Table 1. Antenna has been oriented in xy-plane whereas its normal lies in the direction of z-axis. The overall size of the antenna is 33 mm × 28 mm × 1.57 mm, with FR4 substrate (lossy) having permittivity, r = 4.9.Various feed types can be used to feed the antenna The designed values shown in Table 1 are final optimized values. However they were obtained after quite lengthy simulations.
Simulations and measurements performed to reach these values are presented in this section. Simulations were performed on the following parameters.
Effect of the Patch Outer Radius (R)
For the optimization of the patch, we have first optimized the outer radius of the circular patch, Simulations performed and as shown in Fig. 2 showed that by increasing the value of " R " we get a lower value of the lower frequency band of UWB, i.e.f L frequency. It was also noted in the simulations that after a very specific value (here R = 8.3mm ) if outer radius is increased, then there is no change in f L frequency. Simulations carried out are shown in Fig. 3. They show changes in reflection coefficient ( S 11 ) versus entire frequency graph for different outer radius of the circular patch (R).
It is clearly observable that in this design both the achieved bandwidth and lowest resonance frequency are dependent on the radius of the circular patch ( R ). From the Fig. 3 it can be seen that when R = 8.4 mm then the reflection coefficient is below then − 10 dB over the entire frequency band of interest. For values of R ≥ 8.4 mm , S 11 < −10 dB for the whole bandwidth of UWB.
Hence, considering the requirement of compact size, the optimized value of outer radius was taken as R = 8.4mm.
Effect of the Patch Inner Radius ( )
For different values of the inner radius r = 5 mm, 4.75 mm, 4 mm, 3 mm and 2 mm, simulations were carried out to explore significant change in the reflection coefficient of antenna. This is as shown in Fig. 4. From Fig. 4, it is seen that changing effective inner radius of circular patch has significant effect on resonance frequency and return loss. Increasing of the effective inner patch radius doesn't change notably the other characteristics except reflection behavior of planar antenna. It is also seen in Fig. 3 that at marker 3 ( r = 4.75 mm ) the reflection coefficient is exactly equal to − 10 dB and at marker 4 ( r = 5 mm ) reflection coefficient is − 9.5 dB. Hence below r = 4.75 mm reflection coefficient would be less than − 10 dB in magnitude and above r = 4.75 mm reflection coefficient would be greater than − 10 dB in magnitude. It is seen from the Fig. 4 that when r = 4 mm then the reflection coefficient is less than −10 dB over the entire frequency band of interest and the resonating frequency is 3.5 GHz.
Hence, considering the requirement of compact size, the optimized value of patch inner radius was taken as r = 4 m.
Effect of the Ground Plane Length ( )
For different values of the ground plain length, L g = 12.5 mm, 12 mm or 11.5 mm and 11 mm , respective variations in the reflection coefficient were obtained. From the simulations it is concluded that as the value of L g is decreased, the antenna started behaving more effectively with respect to reflection coefficient ( S 11 ). The plot of Reflection coefficient ( S 11 ) vs. entire frequency graph for different ground plane length L g is provided in Fig. 5. From the graph it is seen that the changing ground plane length has also significant effect on f L and f R . It is observable from Fig. 4 that when L g =11.5 mm then the reflection coefficient S 11 < −10 dB over the entire frequency band of interest and the resonating frequency is 3.5 GHz. At lower value of L g , S 11 is not consistent to be less than − 10 dB. Hence, considering the requirement of compact size, the optimized value of ground plane length was taken as L g = 11.5mm.
Effect of the Ground Plan Width ( )
In this simulation, effect of the ground plane width on the reflection behavior of the proposed antenna was explored for different values of the ground plane width W g = 28 mm, 24.5 mm and 22 mm . It was investigated that as W g was reduced, a corresponding change appeared in the reflection coefficient of antenna. For W g = 28 mm, provided the best behavior of proposed antenna. At this specific value we obtained S 11 = − 11.06 dB. This showed that it is a good selection for the performance of antenna. The simulations are shown in Fig. 6. From the graph it is very clear that there is no change in f L as the value of ground plan width in increased but only there was improvement in reflection coefficient of antenna as shown by marker 3, 4 and 5 in Fig. 5. It was measured that when W g = 28 mm then the reflection coefficient was less than − 10 dB over the entire frequency band of interest and at this value maximum band width was obtained. At W g = 24mm reflection coefficient was exactly equal to − 10 dB; so this value was measured to be a boundary value for the improvement of reflection coefficient. If W g is increased above 24 mm , good reflection coefficient was obtained and below this value there was poor reflection coefficient.
Hence, considering the requirement of compact size, the optimized value of ground plane width was taken as W g = 28 mm.
Effect of Feed-Height ( )
For different values of the feed width F h = 12.7 mm, 12.5 mm, 12.3 mm, 12 mm, 11.7 mm . Simulations were carried out to measure that increase in the feed-line height, good performance of antenna with respect to its reflective behavior was obtained. This behavior increased as feed height was increased from its appropriate value. Simulations are shown in Fig. 7.
Effect of Feed-Line Width ( )
For different values of the feed width, F w = 2.55 mm, 3 mm and 2 mm it wasmeasured that as the feed-line width was increased from 2 to 2.5 mm good performance of antenna was obtained with respect to its reflective behavior. Simulations of these measurements analysis are shown in Fig. 8. Figure 8, also investigated that the changing feed-line width has effect on resonance frequency ( f r ) and f L . Hence, considering the requirement of compact size, the optimized value of ground plane width was taken as F w = 2.55mm.
Effect of Truncation of the Ground Plane
There is a very useful and effective technique for reducing the reflection coefficient. This is done by truncating the ground plane into rectangular or square shape geometry. This behavior of antenna is simulated and measured in this section. Fig. 9. Simulations in Fig. 9 showed that f L remained constant for all the different dimensions of square shaped ground plane while reflection coefficient was less than − 10 dB and f H on each curve varied. Bandwidth was increased as we dimensions of square were decreased. When the dimensions of square shaped ground plane were made L × W = (2.5 × 2.5) mm 2 , the bandwidth was achieved up 28.7 GHz. The value of bandwidth continuously increased as dimension or areas of square were decreased but when the dimension was made L × W = (1.0 × 1.0) mm 2 then again bandwidth started decreasing. This behaviour and relationship of bandwidth against the dimensions of ground plane are summarized in Table 2.
Simulations in Fig. 10 showed that f L remained constant for all the different dimensions of rectangular shaped ground plane while reflection coefficient was less than − 10 dB and f H on each curve varies. Bandwidth was decreased as the dimensions or area of rectangle was decreased. But when the dimensions of recangle shaped ground plane were made L × W = (2.5 × 3.5) mm 2 , L × W = (1.5 × 3.0) mm 2 the increment in bandwidth were achieved beyond the simulation frequncy (30 GHz). When we continuously decreases the dimension or area of rectangle then at L × W = (075.5 × 1.5) mm 2 the value of bandwidth again decreased. This behaviour and relationship of bandwidth against the dimensions of ground plane are summarized in Table 3.
Simulating for both rectangular and square ground plane shows that high bandwidth can be achieved by either of them. But comparatively square shaped ground plane provided higher bandwidth and practically they are easier to be fabricated. Hence therefore square shaped ground plane structure was considered for the final design.
Results
Simulations were carried out for bandwidth measurements using both non-optimized and optimized truncated ground plane antenna structures. The simulation results for S 11 (dB) are shown in Fig. 11. In Fig. 12 are shown the simulation results when the truncated ground plane optimized to square shape was used. This clearly shows the increase in bandwidth as compared to non-optimized truncated plane.
The vector network analyzer (VNA) of maximum range of 24 GHz was used to verify the simulated result. The verified or measured result are shown in Fig. 13.
The directivity of the optimized antenna and integrated plot of E and H plane are shown in Figs. 14 and 15 respectively.
The voltage standing wave ratios are shown in Fig. 16.
Since in the proposed antenna the primary concern was increase in bandwidth along with moderate structure size of antenna, below in Table 4 some previous reported antennas are listed with the comparison in term of size and bandwidth with this design (Fig. 17). In this paper, we have successfully designed and developed a compact in size and high bandwidth UWB antenna using simulations and measurement engineering. In this study, in-depth simulations were carried out to show the design process and setting of antenna parameters. The final antenna structure is engineered to be compatible in low volume requirements such as in body area networks and others. He is the author of over 30 peerreviewed international journal and conference publications. He is also external supervisor in various international universities. He is the reviewer of various international indexed journals. He has also conducted various workshops and tutorials. | 2021-05-05T00:08:25.591Z | 2021-03-23T00:00:00.000 | {
"year": 2021,
"sha1": "df147838d17c2535824c436a5098aca332f5be88",
"oa_license": "CCBY",
"oa_url": "https://www.researchsquare.com/article/rs-207953/latest.pdf",
"oa_status": "GREEN",
"pdf_src": "SpringerNature",
"pdf_hash": "bd6b1fc8ded7aaef79d21cab5b0d277005df5174",
"s2fieldsofstudy": [
"Business"
],
"extfieldsofstudy": [
"Computer Science",
"Physics"
]
} |
260436342 | pes2o/s2orc | v3-fos-license | A major asymmetric ice trap in a planet-forming disk: III. First detection of dimethyl ether
The complex organic molecules (COMs) detected in star-forming regions are the precursors of the prebiotic molecules that can lead to the emergence of life. By studying COMs in more evolved protoplanetary disks we can gain a better understanding of how they are incorporated into planets. This paper presents ALMA band 7 observations of the dust and ice trap in the protoplanetary disk around Oph IRS 48. We report the first detection of dimethyl ether (CH3OCH3) in a planet-forming disk and a tentative detection of methyl formate (CH3OCHO). We determined column densities for the detected molecules and upper limits on non-detected species using the CASSIS spectral analysis tool. The inferred column densities of CH3OCH3 and CH3OCHO with respect to methanol (CH3OH) are of order unity, indicating unusually high abundances of these species compared to other environments. Alternatively, the 12CH3OH emission is optically thick and beam diluted, implying a higher CH3OH column density and a smaller emitting area than originally thought. The presence of these complex molecules can be explained by thermal ice sublimation, where the dust cavity edge is heated by irradiation and the full volatile ice content is observable in the gas phase. This work confirms the presence of oxygen-bearing molecules more complex than CH3OH in protoplanetary disks for the first time. It also shows that it is indeed possible to trace the full interstellar journey of COMs across the different evolutionary stages of star, disk, and planet formation.
Introduction
Complex organic molecules (COMs) are the precursors of prebiotic molecules, and thus understanding their formation and evolution will help us gain more insight into how life originated in our own Solar System (Caselli & Ceccarelli 2012). With new facilities such as the Atacama Large Millimeter/submillimeter Array (ALMA) and the Rosetta Orbiter Spectrometer for Ion and Neutral Analysis (ROSINA) on the Rosetta mission, we are now able to compare chemistry across a range of astronomical environments and get a better understanding of the chemical history of COMs throughout the entire star and planet-formation process, including comets (e.g. . It is crucial to study COMs in planet-forming disks in order to understand how the material in the disk is incorporated into planets and what degree of complexity is present at the epoch of planet formation (Öberg & Bergin 2021).
The formation of most of these COMs is thought to occur in cold molecular clouds (Boogert et al. 2015). During this time, atoms and simple molecules such as CO will stick to the dust grains forming an ice layer and undergo chemical reactions (e.g. Herbst & van Dishoeck 2009;Chuang et al. 2018;Ioppolo et al. 2021). The products are subsequently released back into the gas phase if there is an increase in temperature resulting in thermal desorption. Additionally, molecules will also return to the gas phase via other processes such as UV photodesorption, but this can lead to the fragmentation of the molecule when it enters the gas phase Cruz-Diaz et al. 2016). COMs are therefore especially abundant in the gas phase in young warm systems where they are easily detected because of thermal sublimation (T dust > 100 K) (e.g., Bergner et al. 2017;Jørgensen et al. 2018;van Gelder et al. 2020;Mercimek et al. 2021;Belloche et al. 2020). This is in contrast to older protoplanetary disks which are colder and thus the COMs remain frozen on dust grains in the bulk of the disk and are therefore more often undetectable in the gas phase with ALMA (van 't Hoff et al. 2020). However, COMs are expected to be abundant in protoplanetary disk ices and there is some evidence for this in the outbursting protostellar source V883 Ori which is rich in COMs (van 't Hoff et al. 2018;Lee et al. 2019).
The current situation is that, in protoplanetary disks of more than 1 Myr old, even the most abundant COM, methanol (CH 3 OH), is difficult to detect. CH 3 OH is a cornerstone in the chemistry leading to many larger complex organic molecules (Öberg et al. 2009(Öberg et al. ). Walsh et al. (2016 presented the first detection of CH 3 OH in the TW Hya protoplanetary disk. However, the fractional abundance relative to H 2 is very low (3 × 10 -12 -4 × 10 -11 ), indicating a chemical origin in the gas phase via inefficient and fragmenting non-thermal desorption of the ices rather than thermal sublimation (Walsh et al. 2017). Carney et al. (2019) also provided an upper limit on the abundance of CH 3 OH in the Herbig Ae disk HD 163296 of < 1.6 × 10 -12 relative to H 2 . For comparison, CH 3 OH abundances in hot protostellar cores are typically of order 10 −6 , comparable to those in ices (Boogert et al. 2015 More recently, detected CH 3 OH for the first time in a warm Herbig transition disk. In comparison to the ringed CH 3 OH emission in TW Hya the CH 3 OH in the HD 100546 disk originates from the inner 50 au of the disk and its likely origin is thermal desorption. Because Herbig Ae/Be sources like this one are inherently warm, which prevents freezeout of the precursor CO, in situ formation of the CH 3 OH is unlikely. Instead the presence of CH 3 OH in the disk can be explained via the inheritance of COM-rich ices from colder parent molecular clouds. Also, another Herbig source was revealed to have a rich observable chemistry: the IRS 48 transition disk (van der Marel et al. 2014;van der Marel et al. 2021b;. What makes this disk particularly interesting is the fact that it contains a highly asymmetric dust trap of large grains ( 0.1 mm) on the southern side of the star, making it the most asymmetric disk detected to date (van der Marel et al. 2013;van der Marel et al. 2021a). van der Marel et al. (2021b) report the detection of CH 3 OH and formaldehyde (H 2 CO) in this disk. The emissions have the same crescent shape as the dust continuum, showing for the first time the direct link between a dust trap and COMs. This coincidence was hinted at with low signal-to-noise H 2 CO observations (van der Marel et al. 2014) but is now confirmed. The bulk of the ice reservoir of the IRS48 disk is constrained to the larger dust grains, and because of UV irradiation from the central star, the dust temperature increases enough to liberate the CH 3 OH from the ices. additionally report the detection of SO 2 in the IRS 48 dust trap, the first detection of this molecule in a protoplanetary disk, along with detection of SO. The detection of these molecules supports the presence of oxygen-rich gas where the C/O < 1 because of sublimated ices.
In this paper, we report the analysis of ALMA data of IRS 48 including the first detection of dimethyl ether (CH 3 OCH 3 ) in a protoplanetary disk and a tentative detection of methyl formate (CH 3 OCHO). CH 3 OCH 3 is the largest complex organic molecule that has been detected in a protoplanetary disk to date. We also report the first detection of nitric oxide (NO) in a protoplanetary disk, which will be analysed in a future paper. Our paper is structured as follows: In Section 2 we describe our observational methods and in Section 3 we show our data analysis and provide the values for the derived column densities. In Section 4 we discuss the chemistry of the detected species, compare abundances to other astronomical environments, and determine upper limits for other molecules covered in the data. Finally, in Section 5 we give a short summary and provide conclusions. In papers I and II, we cover the detections of CH 3 OH, H 2 CO, SO, and SO 2 (van der Marel et al. 2021b; and in this paper we present the detection of CH 3 OCH 3 and NO, and investigate other tentative detections and upper limits.
Observations
Data reduction was done using the Common Astronomy Software Applications (CASA) 1 version 5.7.0. The spectral windows have channel widths of ∼ 1.7 km s −1 and a beam size 1 https://casa.nrao.edu/index.shtml of 0 . 55×0 . 42 (PA = 80 • ). The spectral windows have central frequencies of 349.7, 351.5, 361.6, and 363.5 GHz, respectively, with SPW1 from 349.79 to 350.66 GHz, SPW2 from 350.60 to 352.47 GHz, SPW3 from 360.68 to 362.55 GHz, and SPW4 from 362.61 to 364.47 GHz. We imaged the data with the tclean function in CASA using a Briggs weighting with a robust value of 0.5. The image was recentred to the star position using the phase centre parameter in CASA and was set to ICRS 16:27:37.17 -24:30:35.55. We used a Keplerian mask over the region of emission at a distance of 136 pc (Gaia Collaboration et al. 2021), an inclination angle of 50 • , and a position angle of 100 • (van der Marel et al. 2021a).
The cleaned images were subsequently stacked using GoFish version 1.3.6 (Teague 2019) in order to increase the signal-tonoise ratio. This method makes it possible to identify potential weak lines and also distinguish between lines that are blended (very close in frequency). We extract spectra over the whole azimuth of the disk although the lines are co-spatial with the dust trap. This is done because the observations are not well spatially resolved. The spectra for the four spectral windows are shown in the Appendix Figures B1-B4.
Spectral analysis
The stacked, continuum-subtracted spectra were analysed using the CASSIS 2 spectral analysis tool version 5.1.1 (Vastel et al. 2015) in a similar way to that used by Nazari et al. (e.g., 2021). The flux densities were first converted to brightness temperature units and local thermodynamic equilibrium (LTE) conditions were assumed in order to derive the column densities and excitation temperatures. We made use of the Cologne Database for Molecular Spectroscopy (CDMS) (Müller et al. 2001(Müller et al. , 2005 and the Jet Propulsion Laboratory (JPL) database (Pickett et al. 1998) for molecular information. In Table A.1 we list the transitions of the detected species. The integrated intensity maps of selected lines are presented in Figure 1. The search for other potential features in the spectra was carried out by making a selection of commonly detected COMs in other environments and only taking into account lines with E up ≤ 400 K and A ul ≥ 1× -6 10 s -1 . For this selection of molecules, we modelled the spectra in CASSIS by assuming an excitation temperature of 100 K motivated by the rotational temperature derived by van der Marel et al. (2021b) for the CH 3 OH. We also calculated the best-fit column density at both 70 K and 250 K to have an estimate of the column density error for the detected species, which is typically a factor of two. The absolute calibration error is much smaller, of order 10%, and this will cancel out in abundance ratios. We used a FWHM of ∼ 7 km s −1 based on the line width of a strong CH 3 OH line and a source size of 1.4 × 10 -11 sr based on the 5 σ emission continuum of the disk (Figure 1), the same as in van der Marel et al. (2021b); . Using these variables we derived column densities and upper limits. We note that the inferred column density, N, scales inversely with the assumed emitting area, Ω source (Goldsmith & Langer 1999), that is, In the case where the source does not fill the beam, the column density will be underestimated by a dilution factor: where Ω beam is the beam size (e.g. van Gelder et al. 2020).
Revising the CH 3 OH column density
We first modelled the CH 3 OH lines in our spectra based on the parameters derived in van der Marel et al. (2021b). Using CAS-SIS we find models consistent with the data using a column density for CH 3 OH of 5 × 14 10 cm -2 and an excitation temperature of 100 K which is in agreement with the rotational diagram analysis of van der Marel et al. (2021b). These results are shown in Figure 2. However, in this work, additional CH 3 OH transitions are detected and these are highlighted in Table A1. Figure 1 shows the integrated intensity maps of a CH 3 OH line with an upper energy level of 333 K. We also detect two weaker lines in the stacked spectra that are better fit at a higher column density of 2 × 10 15 cm −2 and still at a temperature of 100 K (see Figure 2 for a comparison of the two models). These two lines are the 9 -5,4 -9 -4,6 and 3 1,2 -4 2,2 at 351.236 GHz (E up = 241 K, log 10 (E A ) = −4.44 s −1 ) and 361.236 GHz (E up = 339 K, log 10 (E A ) = −3.58 s −1 ) respectively. Both lines were visible in our stacked spectra at the 2.5-3σ level but neither one was detected in the channel maps above the 3 σ level. These weak lines are reproduced at a different column density, which likely indicates that the emission traced by the stronger lines is optically thick. van der Marel et al. (2021b) calculate the optical depth of these lines and show they are optically thin for the assumed emitting area. The difference between this and our result can be resolved if the lines are optically thick and beam diluted, because the column density is inversely proportional to the assumed emitting area. We also derived a 3σ upper limit for the column density of 13 CH 3 OH of < 5.5 ×10 14 cm −2 . This gives a strict upper limit on the CH 3 OH column density of ≈ 3.3 ×10 16 cm −2 assuming a 12 C/ 13 C ratio of 60. This upper-limit is consistent with the column density found via the weakest CH 3 OH lines. We use N(CH 3 OH) = 2 × 10 15 cm −2 as a reference for comparisons.
Detection of dimethyl ether and methyl formate
We detect two sets of blended lines for CH 3 OCH 3 . See Table A.1 for the transition information, Figure 1 for an intensityintegrated map of one of the sets of blended lines, and Figure C.1 for the channel maps of both. We derive a column density of 1.5 × 15 10 cm -2 at an assumed excitation temperature of 100 K; see Figure 2 for the CASSIS model fits. The excitation temperature of CH 3 OCH 3 may be lower than that of CH 3 OH (e.g. Jørgensen et al. 2018) but given that only a few transitions are detected, we calculate the column densities over a range of excitation temperatures from 80 to 250 K (listed in Table 1). We also found a tentative detection of CH 3 OCHO after modelling several features in our spectra. One such emission feature with a S/N above 3σ noise level can be seen at 363.48 and 363.49 GHz (Figure 2). We show a clear > 5 σ detection of a line in the channel maps at this frequency range ( Figure C.2) and the integrated intensity map is shown in Figure 1. While the modelled spectrum of CH 3 OCHO does not provide an exact fit for the emission feature, it is the closest fit found after considering other possible candidates. Furthermore, the model spectrum for CH 3 OCHO fits several other smaller features in the spectra (see Figures B.1,B.2,B.3,B.4). From these models we derive a best-fit column density of 1.3 × 10 15 cm -2 at an excitation temperature of 100 K.
Other line detections and upper limits
We detect a total of five transition lines for nitric oxide (Table A.1). This is the first detection of NO in a protoplanetary disk. We were alerted to the possible presence of NO in our disk when we encountered a difficulty in fitting the bright CH 3 OH line at 350.68 GHz (Figure 2). NO has two transitions at this frequency, but a single line was not enough to confirm a definitive detection of the molecule because this line is also blended with CH 3 OH. We are able to prove the presence of NO in the disk after successfully fitting an additional double feature at 351.04 and 351.05 GHz (Figure 2). From the CASSIS spectral analysis models, we derive a best-fit total column density for NO of 3 × 15 10 cm -2 at an excitation temperature of 40 K. The NO lines have a low excitation temperature (36 K) compared to many of the COM lines detected and the lower temperature best fits the multiple lines. The NO lines will be more quantitatively analysed in a future paper.
We also detect an additional SO 2 line at 363.16 GHz ( Figure B.1) that was not reported in . The column densities from our spectral analysis are in agreement with their value.
Finally, we also derive upper limits for species that remain undetected in the IRS 48 disk but that have been observed in younger sources and other older disks such as formic acid (t-HCOOH), acetaldehyde (CH 3 CHO), and methyl cyanide (CH 3 CN) (e.g., Bergner et al. 2017;Favre et al. 2018;van Gelder et al. 2020;Ilee et al. 2021). These upper limits are listed in Table 1.
Discussion
We find a wealth of molecular complexity in the IRS 48 disk, including the first detections of multiple molecules in disks. In this section, we discuss the chemical origin of the COMs, compare relative abundances to other environments, and consider the prospects for further complexity in the disk.
Chemical origin of the COMs
The observed CH 3 OH emission in the IRS 48 disk first presented by van der Marel et al. (2021b) is azimuthally co-spatial with the dust trap and peaking at slightly smaller radius. van der Marel et al. (2021b) proposed that the presence of CH 3 OH in the disk is due to thermal ice sublimation and that the ice reservoir is constrained to the larger millimetre-sized grains. Vertical mixing in the vortex may also help in lifting icy dust grains to the warm surface. CH 3 OH forms on CO ice via a sequence of H-addition reactions with key intermediates HCO and H 2 CO (Fuchs et al. 2009;Chuang et al. 2017). Because the grain surface chemistry of CH 3 OH is related to both CH 3 OCH 3 and CH 3 OCHO, particularly in the presence of UV radiation, we expect that both of these COMs also originate from the sublimating ices (Öberg et al. 2009), and Garrod et al. (2008) provide a theoretical model in which complex organic molecules, including CH 3 OCH 3 and CH 3 OCHO, can form via cold grain-surface reactions (≤ 50 K) involving radicals: A&A proofs: manuscript no. irs48_coms_corrected This model shows a common formation route from the methoxy precursor CH 3 O. These pathways have also been shown to be present in laboratory experiments (Chuang et al. 2016). These COMs could nevertheless be further enhanced due to UV irradiation of the ices from the central star resulting in photodissociation of CH 3 OH (Öberg et al. 2009;Walsh et al. 2014), where the dissociation products can then recombine via reactions (1) and (2).
Comparisons to other environments
CH 3 OCH 3 is the largest complex organic molecule detected in a protoplanetary disk. It has been detected in several other, younger sources (e.g. Taquet et al. 2015;Soma et al. 2018;Bergner et al. 2018). We compare our results as summarised in Table 1 (100 K column) and Figure 3 with the observed abundances in other astronomical environments including the class 0 protostellar binary IRAS 16293 A and B (Jørgensen et al. 2018;Manigand et al. 2020), the outbursting source V883 Ori (Lee et al. 2019), and the comet 67P . The high observed CH 3 OCH 3 /CH 3 OH ratio in the IRS 48 disk, of order unity, is different from that in the other sources by a factor of 5-10 ( Figure 3). The abundance derived for CH 3 OCHO also shows a similar trend to CH 3 OCH 3 in that CH 3 OCHO seems to be more abundant compared to the other sources. The high derived column-density ratios of CH 3 OCH 3 /CH 3 OH and CH 3 OCHO/CH 3 OH in IRS 48 (Figure 3) compared to other environments may be due to optically thick CH 3 OH emission that is beam diluted resulting in over-estimated abundance ratios. To increase the optical depth in CH 3 OH to the amount that would make the ratio consistent with other sources, an area of order 10 −12 sr would be needed. If the emission is constrained to the inner edge of the dust cavity, this would require a crescent shape for the emitting area with a length of ≈1 . 0, corresponding to a width of ≈0 . 1 in COM emission, which could be resolved in future higher resolution data. Detections of CH 3 OH isotopologues are needed to determine whether optical depth is indeed the cause of the difference or chemical processing in the disk relative to ices in dark clouds and young stars could be responsible for the enhanced chemical complexity in the UV-irradiated ice trap. Table 1 and the grey bar marks the ±1σ error calculated from the line-free channels in each spectral window. Panels A and B show the two weak CH 3 OH lines 9 -5,4 -9 -4,6 and 3 1,2 -4 2,2 with the 100 K CASSIS models at 5 × 10 14 cm −1 (blue) and 2 × 10 15 cm −1 (green). Panels C and D show the best-fit models for the CH 3 OCH 3 20 1,20 −19 0,19 and 11 3,8 -10 2,9 transitions and the CH 3 OCHO 32 3,30 -31 3,29 and 32 3,30 -31 2,29 transitions. In panel D the negative dip in the spectrum at ≈363.5 GHz may be an atmospheric absorption feature (https://almascience.eso.org/about-alma/atmosphere-model). Panels E and F show the best-fit models for NO covering the 4 1 -4 3 and 4 1 -4 4 transitions. Panels C, D, and E also show the CH 3 OH model for the strong lines.
The CH 3 OCH 3 /CH 3 OCHO ratio in our disk is approximately 0.9 and this is consistent with what is observed in other sources across a full range of environments from star and disk formation to comets (Coletta et al. 2020). This adds further evidence that these two species are likely chemically related to one another and points towards ice formation of both molecules and therefore the inheritance of ices in the IRS 48 disk.
Upper limits
We also derived upper limits for the column densities of COMs that were previously detected in other sources. These molecules are listed in Table 1; see also Figure 3. Although CH 3 CHO, t-HCOOH, and CH 3 CN have been detected in several sources, we note that they remain absent in our disk despite having formation routes via grain-surface chemistry (Walsh et al. 2014). In particular, HCOOH can form via the HCO or HOCO radicals and CH 3 CHO via the CH 3 and HCO radicals. The nondetection of formic acid is potentially interesting, as in disk chemical models it is predicted to have a similar fractional abundance to CH 3 OH in the gas phase (Walsh et al. 2014). From the upper limit, we constrain this ratio to <10 %. However, Walsh et al. (2014) predict that the CH 3 OH ice column density is approximately ten times higher than HCOOH. This could explain the non-detection of HCOOH in our data if the sublimating ice reservoir is the primary origin of both molecules. TW Hya also has a detection of formic acid (Favre et al. 2018) and in this disk the t−HCOOH/CH 3 OH abundance ratio is approximately unity. This is at least an order of magnitude higher than the 1%-10% seen in young stars and comets (e.g. . In comparison to IRS 48, where the observable chemistry appears to be dominated by ice sublimation, TW Hya is a cold disk where small amounts of COMs in the gas phase are due to nonthermal desorption and/or gas-phase chemistry. Similarly, with the detection of CH 3 OCH 3 we might also expect to have detected CH 3 CHO. The abundance of CH 3 CHO with respect to CH 3 OH is found to be about ten times lower thanthat of CH 3 OCH 3 . This difference is consistent with the results of van Gelder et al. (2020) who find lower abundances of CH 3 CHO compared to CH 3 OCH 3 in young protostellar envelopes.
Another molecule that is particularly interesting to look at is CH 3 CN as it has been detected in multiple protoplanetary disks Bergner et al. 2018;Ilee et al. 2021). The formation of CH 3 CN seems to be dominated by gas-phase chemistry but grain surface processes cannot be neglected . In particular, gas-phase CH 3 CN is enhanced in environments with high C/O ratio. In TW Hya, the only disk with detections of both CH 3 OH and CH 3 CN the CH 3 CN/CH 3 OH column density ratio is approximately unity. Unlike what we expect for IRS 48, in TW Hya the observable CH 3 CN and CH 3 OH likely do not have the same chemical origin. The CH 3 CN is primarily formed via gas-phase routes whereas the CH 3 OH most likely originates from the ices Walsh et al. 2016Walsh et al. , 2017. In IRS 48 we have an upper limit of ≈10%. This is in better agreement with the 1%-10% seen in comets and young stars (see Figure 3 and e.g. Bergner et al. 2017).
We also obtained upper limits on the deuterated form of methanol CH 2 DOH. The upper limit on this column density at an excitation temperature of 100 K is 6.0 × 14 10 cm −2 . This then gives an upper limit on the D/H of 10% and this is consistent with the ratios seen in protostellar cores, young low-mass stars, and comet 67P (≈1%-10%, e.g. van Gelder et al. 2020;Drozdovskaya et al. 2021).
Prospects for further complexity in the IRS 48 ice trap
Confirmation of the CH 3 OCHO detection is needed because our models do not fit the emission feature very well and we only have one significant feature to fit given the frequency coverage of the observations. There are several other COMs that remain undetected in our disk and which should be the focus of future work. The molecules listed in Table 1 are examples of species that should be searched for in future studies due to the fact that most have been detected in multiple protostellar sources. The detection of CH 3 OCH 3 and CH 3 OCHO alongside CH 3 OH implies a rich ice chemistry in the IRS 48 dust trap. Other COMs which have related formation routes via the radials HCO, CH 3 O and CH 2 OH including ethylene-glycol, acetaldehyde, ethanol, and glycolaldehyde, should be subject of follow-up observations.
Conclusions
We analyzed ALMA data of the IRS 48 transition disk, revealing a wealth of molecular complexity.
-We report the first detections of dimethy ether (CH 3 OCH 3 ), nitric oxide (NO), and a tentative detection of methyl formate (CH 3 OCHO) in a protoplanetary disk. -We report an additional detection of a SO 2 transition in the disk with an upper energy level of ≈250 K. -The emissions of the detected species show a direct link with the asymmetric dust trap in the southern region of the disk, further suggesting that molecular complexity in this disk is due to ice sublimation. -The abundance ratios of CH 3 OCH 3 and CH 3 OCHO compared with CH 3 OH are high relative to other environments. This either means that these molecules are enhanced relative to CH 3 OH in this disk or that the CH 3 OH column density we derive is underestimated. The latter situation could be due to the lines being optically thick but beam diluted. A higher CH 3 OH column density would mean a COM-emitting area smaller than the assumed area, the 5σ extent of the millimetre dust trap. With further high-angular-resolution observations, we will be able to determine whether or not the emitting area is truly just the thin inner edge of the dust trap. -The detection of CH 3 OCH 3 and CH 3 OCHO in such a warm disk and the agreement in the CH 3 OCH 3 /CH 3 OCHO column density ratio with other environments strengthens the case for an origin inherited from the cold cloud phase, but the abundances with respect to CH 3 OH may be enhanced because of UV irradiation.
Hopefully future observations of the IRS 48 icy dust trap will allow for the detection of other COMs and more robust constraints on the column density and excitation conditions. This work is an important puzzle piece in tracing the full interstellar -Molecular information was obtained using the JPL and CDMS database (Müller et al. 2001(Müller et al. , 2005Pickett et al. 1998). The spectra were modelled using the CASSIS spectral analysis tool. For the detected lines, a range of T ex from 70 K to 250 K was modelled. The 3 σ upper limits were derived assuming a T ex of 100 K. journey of COMs across the different evolutionary stages of star, disk, and planet formation. | 2022-01-30T16:06:09.466Z | 2022-01-31T00:00:00.000 | {
"year": 2022,
"sha1": "cbf385429c435c03bef1fc562c54b6130c9a46d5",
"oa_license": null,
"oa_url": "https://www.aanda.org/articles/aa/pdf/2022/03/aa42981-21.pdf",
"oa_status": "BRONZE",
"pdf_src": "Arxiv",
"pdf_hash": "4716feb47a820e8f8075dadd91962ff7e3bcbb76",
"s2fieldsofstudy": [
"Environmental Science",
"Physics",
"Chemistry"
],
"extfieldsofstudy": [
"Physics"
]
} |
253284993 | pes2o/s2orc | v3-fos-license | Role of accuracy of data in the functionality of primary health care system of Pakistan: A comparative study.
… Objective: To evaluate the Quality and accuracy of existing Data, to know about the current situation of Infrastructure and Functionality of Primary Health Care (PHC), and to assess the District Health Information System (DHIS) Operational Needs. Study Design: Mixed Method Retrospective Descriptive study. Setting: Seven Tehsils of Rawalpindi District. Period: January to February 2020. Material & Methods: One BHU was randomly selected from each Tehsil to cover all geographical areas. Data was collected using questions derived from the Procedures Manual of DHIS and comparing with the monthly DHIS reports of PHC facilities. Data was analysed after entering into SPSS software (V-22). Results: The 57 % of the facilities’ in-charges (4 out of 7) had DHIS training, while the remaining 43% were observed as either not trained or self-trained. The overall accuracy rate in DHIS was 71.43% but variance was observed in different variables. Conclusion: The current situation of DHIS demands initiative for additional collaborations among multiple vertical health programs. Some promising interventions in government policies, including skilled support programs and training of DHIS teams through regular sessions and workshops are highly recommended.
INTRODUCTION
Health Information System is a system that collects data, analyses and use the results for improving quality and efficiency of health services, along with better management at all levels specifically PHC. 1 A proper DHIS is vital for improving the health indicators of a developing country like Pakistan with limited physical, human and financial resources. Timely and reliable data is crucial for appropriate planning and organization of health services. DHIS was introduced to replace Health Management Information System (HMIS). 2 Currently out of 140 districts 126 districts are reporting regularly in Pakistan. 3 DHIS is nationally standardized system of data collection, analysis and feedback. Currently this is operational in public sector health facilities only.
The data provided by DHIS is used for planning, application and monitoring of diseases along with implementation of preventive services by utilizing the available resources. 4 In our setting, PHC facilities comprises of two main setups, Basic Health Units (BHU) and Rural Health Centers (RHC). Here, Medical Officer is responsible for supervision of the overall work of the facility, its employees and health personnel involved in outreach/ community-based services. A standard checklist is provided at all PHC facilities. 5 For administrative purposes, district Rawalpindi is divided into seven tehsils and has ninety-eight BHU's. 6 Report compliance of health facilities in Punjab is ninety nine percent (99%). 7 Lot quality assurance sampling (LQAS) has been anticipated as an effective tool for monitoring health programs https://doi.org/10.29309/TPMJ/2022.29.11.7091 2 and assessing the quality of data at health center. 8 However, there are very limited studies, who have analyzed the accuracy of data reporting from the BHUs in Rawalpindi. The deficiency in basic management skills of health teams is a major constraint in functionality of PHC in Pakistan Therefore, we aimed to assess the quality of available data and evaluate the accuracy of reports (LQAS) of seven Basic Health Units of Rawalpindi district to know the current situation of infrastructure & functionality in these areas, highlighting the DHIS operational needs.
MATERIAL & METHODS
This study was conducted in seven tehsils of Rawalpindi district. One BHU was randomly selected from every tehsil to cover all geographical areas of the district. The list of names of all BHUs as shown in Table- This was a mixed method retrospective descriptive study on the data recorded in the months of October, November and December 2019, while the data was gathered in January-February 2020.
Monthly reports and data recorded in DHIS was used to compare and check for accuracy and application of reports.
A self-constructed DHIS questionnaire, adapted from different tools of DHIS was used. Qualitative data collection was done through semistructured interviews, observation of participants and analysis of documents, checklists and reports. Data was analyzed by content analysis methods and reorganized according to seven themes developed during the study. These themes are data generation; data management; data analysis, transmission, and reporting; data interpretation; feedback and; training on DHIS method. Quantitative data was collected from DHIS registers and its reporting on DHIS forms. Results were analyzed using SPSS 22, using descriptive statistics and frequency distributions.
RESULTS
The participants of the study consisted of professionals and qualified staff, which compliments the credibility of data. This is shown in Figure-1. During the survey, it was found that only one facility was in charge of filling the reporting form of DHIS. Among them 57.1% were trained in DHIS instruments; they had acquired training from a local office. About 85% (6 out of 7) of the health facilities recommended having refresher training in DHIS. The data analysis pointed out that 14% BHUs faced the shortage of tools and medical equipment.
LQAS test of Hayal Shaarif was in lowest accuracy (56.42%) with a missing data ratio of 12.82%. BHU Dakhali's data was the most accurate among all the BHUs (95%) and missing data percentage was 25.64%. BHU Tret had an accuracy ratio of 82% but had the highest missing data ratio (35.89%), BHU Karor had an accuracy of 70% and Sahang had an accuracy of 74% but missing data ratio was higher in Karor (35.89%) than Sahang (28.2%). The overall accuracy in this study was 75% and missing data ratio was 25%. The LQAS results are shown in Table-III. DISCUSSION DHIS provides the basic data for planning, execution and monitoring on main indicators of disease pattern, preventive services and available resources. 11 At PHC level, Medical Officer in-charge is responsible for the overall supervision although supervisory role has not been clearly mentioned in their job descriptions. 14 This was depicted in our study that the personnel responsible for filling out the registers were not all medical officers; hence they were not properly trained for doing this task.
One of the significant indicators to assess functioning of health facility is the outpatient attendance per capita. 15 If Out Patient Department (OPD) attendance is found to be high in the public health facilities, it suggests that the population is extremely satisfied with the services in these facilities. 16 The accuracy ratio in OPD data was 80.95%, matching the results of other developing countries. As in Uganda, 85.3% of health facilities used OPD registers while only 61.8% were correctly filled. Reporting trend in our study was same as in Uganda but we had better socioeconomic indicators. 17 Pakistan ranks among the ten countries having highest under-5 mortality burden, both for diarrhea and lower respiratory infection. 18,19 A survey conducted in a rural Pakistani community showed that for childhood diarrhea and pneumonia, caregivers prefer to seek consultation from doctors (97%), of which 75% caregivers pursued care from private and 45% from public sector. 20 However, our study showed no patients of Pneumonia in 71.42% BHUs, possibly because they refer Pneumonia cases to other hospitals. Among those BHUs, where Pneumonia cases were documented, the accuracy ratio was 90.47% and during peak months no patient was recorded in these seven BHUs. In addition, there was highest number of diarrheal patients in BHU Tret, with accuracy ratio of 66%.
Our study showed a 95% accuracy in reporting of malaria cases. This is because of a high number of malaria cases in our community with estimated 1.5 million annually. 17 Tuberculosis (TB) is one of the major public health issues in Pakistan accounting for 61% of the TB burden in the WHO Eastern Mediterranean Region. 21 Our study showed careless behavior in DHIS data, as two BHUs were not reporting the data properly and a huge variation was observed in remaining BHUs, where the registered cases of TB were higher than reported data.
For immunization coverage, one such study reported that the coverage of individual vaccines was 76% for BCG, 61% for DPT 1, 49% for DPT 2, 45% for DPT 3 and about 27% for measles. 22 Comparatively, UNICEF reports, BCG coverage is less at 67%, DPT 3, is much higher at 63% and measles also higher at 57%. 23 From these discrepancies, it is obvious there is problem with the reporting of vaccination status.
No data is available on the accuracy of family planning and ANC visits. A study conducted in Punjab, Pakistan showed that overall, 28% of BHUs and RHUs had poor infrastructure. 16% of the health facilities were deficient in equipment. Essential Laboratory items, such as urine strips for albumin, blood sugar testing strips, and hemoglobin reagents, were particularly shortly stocked. 24 A study conducted in Lahore, Pakistan showed regarding data recording and reporting tools, that out of 40 lady health workers (LHWs) interviewed, 32 (80%) had good knowledge, 6 (15%) had satisfactory knowledge, while 2 (5%) had an unsatisfactory knowledge. Only 47.5% of reports were found accurate, according to our scoring system. Additionally, 35% of reports had missed data entries, misrepresented data and simulated material. 25 Moreover, a study conducted in African countries, showed that data gathering has low priority and is unreliable. 26 This is reflected in our study as well, where there was low accuracy ratio of data in most BHUs and missing data frequency was high.
CONCLUSION
The current situation of DHIS demands that the policy makers have to take initiative for collaborations among various health programs. Some promising interventions for appointing and retaining staff along with improving government regulations are highly needed. We highly recommend training of health teams through workshops and sessions, along with on-site support on regular basis. Copyright© 14 July, 2022. | 2022-11-04T18:24:34.735Z | 2022-10-31T00:00:00.000 | {
"year": 2022,
"sha1": "2a94bbd30598564f16ee388a675d8c27b6a0bc30",
"oa_license": "CCBYNC",
"oa_url": "http://theprofesional.com/index.php/tpmj/article/download/7091/5021",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "0a5902ca18d47d2891f729402f010b860328d90b",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
247025094 | pes2o/s2orc | v3-fos-license | PNP Hydrogel Prevents Formation of Symblephara in Mice After Ocular Alkali Injury
Purpose To create an alkali injury symblephara mouse model to study conjunctival fibrosis pathophysiology and test polymer nanoparticle (PNP) hydrogel as a preventative therapeutic. Methods Mice were injured using NaOH-soaked filter paper to determine the optimal NaOH concentration to induce the formation of symblephara. Injured mice were observed for 7 days to detect the formation of symblephara. Forniceal shortening observed on hematoxylin and eosin (H&E)-stained tissue sections was used as a symblephara marker. Alpha-smooth muscle actin (α-SMA) expression, Masson's trichrome assay, and periodic acid–Schiff (PAS) staining were used to determine myofibroblast expression, collagen deposition, and goblet cell integrity. PNP hydrogel, with multivalent, noncovalent interactions between modified biopolymers and nanoparticles, was applied immediately after alkali injury to determine its ability to prevent the formation of symblephara. Results Forniceal shortening was observed in H&E images with 1N NaOH for 2 minutes after 7 days without globe destruction. PNP hydrogel prevented forniceal shortening after alkali injury as observed by H&E histology. α-SMA expression and collagen deposition in eye tissue sections were increased in the fornix after injury with 1N NaOH compared with uninjured controls. PNP hydrogel treatment immediately after injury reduced α-SMA expression and collagen deposition in the forniceal region. Mucin-secreting goblet cells stained with PAS were significantly lower in alkali-injured and PNP hydrogel–treated conjunctivas than in uninjured control conjunctivas. Conclusions We observed that 1N NaOH for 2 minutes induced maximal forniceal shortening and symblephara in mice. PNP hydrogel prevented forniceal shortening and conjunctival fibrosis after injury. This first murine model for symblephara will be useful to study fibrosis pathophysiology after conjunctival injury and to determine therapeutic targets for cicatrizing diseases. Translational Relevance This mouse model of symblephara can be useful for studying conjunctival scarring disease pathophysiology and preventative therapeutics. We tested PNP hydrogel, which prevented the formation of symblephara after injury.
Introduction
Each year in the United States, 2.4 million ocular injuries occur, of which 11.5% to 22.1% are due to alkali burns. 1,2 These injuries often lead to ocular scarring and vision loss, with lifelong decreased quality of life, lower economic productivity, and chronic neuropathic pain. 3 Alkali injuries cause conjunctival inflammation, and the resultant fibrosis may cause adhesions between the eyelid and globe, known as symblephara. 4 Symblephara are one of the most challenging eye problems faced by physicians and are debilitating for patients. Alkali burns are more damaging than acid burns, as alkali causes fatty acid saponification, which allows deeper penetration through the different eye layers. 5 The severity of symblephara ranges from subconjunctival inflammation and fibrosis to conjunctival forniceal obliteration, where the eyelid is completely adhered to the globe. 6 The fornix is required for ocular surface health by maintaining the tear reservoir, thus preventing dry eye disease, and it is the home for the accessory lacrimal glands. Symblephara may also cause ocular motility restriction, which may lead to double vision and headaches, as well as direct neuropathic pain. If left unchecked, symblephara may cause blindness due to dryness and tear insufficiency. When symblephara have formed, corneal decompensation and corneal transplantation often follow, the latter of which frequently fails due to an inadequate ocular surface environment to maintain a graft. 7 Current clinical treatments are ineffective to prevent or durably treat symblephara, and little is known about the molecular mechanisms behind their formation. An anti-adhesion therapy that inhibits the formation of symblephara would revolutionize how patients with ocular injuries are treated. 8 Ocular surface alkali injury has been used to study corneal injury in mice [9][10][11][12][13][14] ; however, the formation of symblephara in these mice has not been explored thus far. An alkali injury model described the formation of ocular symblephara in rabbits. 15 The rabbit ocular surface was injured with 2N NaOH for 90 seconds, and forniceal shortening and the formation of symblephara followed for 4 weeks. Although this model is useful for developing a basic understanding of the formation of symblephara, it may be difficult and expensive to use for high-throughput studies for testing therapeutics, with a long timeline until the formation of symblephara.
Years ago, the formation of symblephara after alkali burns in patients was treated with a glass rod to manually break the adhesion strands 16 ; however, this approach was ultimately found to be ineffective at keeping the conjunctival surfaces sufficiently apart to prevent the reoccurrence of symblephara. Current treatment methods for alkali ocular burns include lamellar keratoplasty along with amniotic membrane transplantation (AMT) and autologous conjunctival or oral mucosal transplantation, 17,18 AMT with symblephara ring, 19 and soft contact lenses to keep the palpebral and bulbar surface apart. 16 However these methods are not completely effective because they may not reach the deep fornix region, where the conjunctiva can still adhere and cause forniceal shortening. Therefore, there is a need for alternative, more effective methods to prevent ocular adhesions. We describe the first, to our knowledge, murine alkali-injury model that develops symblephara within 7 days. We also test a novel supramolecular polymer nanoparticle (PNP) hydrogel 20,21 that prevents conjunctival adhesions and forniceal shortening after alkali-burn injury in mice.
Animals
All experimental protocols were approved by the Stanford University Institutional Animal Care & Use Committee (APLAC #33501), and all procedures were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and in compliance with the Animal Research: Reporting of In Vivo Experiments guidelines. C57BL6 male mice between the ages of 3 and 6 months were used for this study. Age-matched littermates were used for comparison. Experiments were repeated at different ages with similar results.
NaOH Dose Response
The mice were divided into five groups, each treated with increasing NaOH concentrations. Round filter papers disks (3.5-mm diameter) were soaked in 0, 0.25N, 0.5N, 1N, and 2N NaOH. They were subsequently placed under the superior and inferior mouse eyelids. The eyes were closed shut for 2 minutes, after which the filter paper was removed, the eyes were irrigated with phosphate-buffered saline (PBS), and pH was measured using pH test strips (M1095350001; MilliporeSigma, Burlington, MA) after absorbing excess PBS with a cotton sponge. Irrigation with PBS was continued until the pH measured 7.0.
Formation of Symblephara Over Time
Mice were divided into four groups, each treated with 1N NaOH as described above. The mice were allowed to recover and observed for 7 days. The globes and eyelids were together enucleated from the orbit by cutting around eyelid and under the eyeball such that the fornix structure was preserved. Tissue resection was done at days 1, 5, 7, and 14 after injury; fixed in 10% formalin; and sectioned into 5-μm-thick sections for hematoxylin and eosin (H&E) staining analysis.
PEG-PLA Synthesis
A procedure was followed and analyzed as described previously. 22 PEG (0.25 g, 4.1 mmol) and N,Ndibutylurea (DBU; 10.6 mg, 10 mL, 1.0 mol% relative to lactic acid) were dissolved in dichloromethane (DCM; 1.0 mL). Lactic acid (1.0 g, 6.9 mmol) was dissolved in DCM (3.5 mL) with mild heating. The lactic acid solution was then added rapidly to the PEG/DBU solution and was stirred rapidly for 10 minutes. The PEG-PLA copolymer was then recovered from the reaction medium by precipitation from excess 50:50 mixture of cold diethyl ether and hexanes, collected by filtration, and dried under vacuum to yield a white amorphous polymer.
HPMC-C 12 Synthesis and Characterization
HPMC (1.5 g) was dissolved in Nmethylpyrrolidone (NMP; 60 mL) by stirring at 80°C for 1 hour. When the polymer had completely dissolved, the solution was heated to 50°C. A solution of 1-dodecylisocyanate (0.5 mmol, 10% dodecyl modification by weight) was dissolved in NMP (5 mL) and added to the reaction mixture followed by 150 μL of N,N-diisopropylethylamine as a catalyst. The solution was then stirred at room temperature for 16 hours. This solution was then precipitated from acetone, and the HPMC-C 12 polymer was recovered by filtration, dialyzed within a 3.5-kDa cutoff dialysis bag for 3 days in water, and then lyophilized, yielding a white amorphous material.
PEG-PLA Nanoprecipitation
A procedure was followed and analyzed as described previously. 22 A solution (1 mL) of PEG-PLA in acetonitrile (50 mg/mL) was added dropwise to water (10 mL) at a stir rate of 600 rpm. Nanoparticles were purified by ultracentrifugation over a filter with a molecular weight cutoff of 10 kDa (MilliporeAmicon Ultra-15 Centrifugal Filter Unit; MilliporeSigma) followed by resuspension in water to a final concentration of 250 mg/mL. Nanoparticle size and dispersity were characterized by dynamic light scattering (D H = 35 nm, polydispersity index (PDI) = 0.02).
PNP Hydrogel Formulation
HPMC-C 12 was dissolved in PBS at 6 wt% and loaded into a 1-mL Eppendorf tube. A 20 wt% PEG-PLA nanoparticle solution in PBS was then added to PBS and loaded into the tube. The HPMC polymer solution (150 mL) and nanoparticle solution (300 mL) were added together and mixed well by vortexing to create the PNP 1:10 hydrogel formulation.
PNP Hydrogel Treatment Regimen
The right and left eyes of each mouse were injured with 1N NaOH as described above. After irrigation and pH 7 was reached, 50 μL PNP hydrogel was applied on the ocular surface of the left eye and under the eyelid using a spatula. The right eye was left untreated. Eyelids were blinked so that the PNP hydrogel got evenly spread and reached the deep fornix of the eyes.
Histological Examination
For H&E-stained sections, 10% formalin fixed eyes were embedded in paraffin and 5-μm sections. Paraffin sections were deparaffinized using xylene and rehydrated in a graded series of ethanol and then water, and they were stained with H&E according to standard protocol. Images were taken on an EVOS XL Core Imaging System (Thermo Fisher Scientific, Waltham, MA) at 10× and 20× magnifications. For immunostaining, sections were blocked with 5% bovine serum albumin (BSA) in PBS for 1 hour, then incubated at 4°C overnight in primary antibody, alpha-smooth muscle actin (α-SMA) (C6198; Sigma-Aldrich, St. Louis, MO) or MSLN (250519; Abbiotec, Escondido, CA) diluted at 1:100 and 1:200, respectively, in 1% BSA in PBS. Sections were incubated in secondary antibody diluted at 1:250 for 1 hour, and nuclei were stained with 4 ,6-diamidino-2-phenylindole (DAPI) for 5 minutes. Fluorescent images were taken using the THUNDER Imager EM Cryo CLEM microscope (Leica Microsystems, Wetzlar, Germany) at 10× and 20× magnifications. For Masson's trichrome and periodic acid-Schiff (PAS) stain, 5-μm sections were cut and stained according to the manufacturer's protocol. Quantitative analysis of collagen deposition and goblet cell numbers was done in Photoshop (Adobe, San Jose, CA).
Rheology Measurement and Fluorescent PNP Hydrogel Imaging
A 1:10 PNP hydrogel was used according to published methods. 20 Rheology measurements were performed on a DHR-2 rheometer (TA Instruments, New Castle, DE) using a 20-mm serrated plate geometry with a gap of 600 μm. Flow sweeps were performed at shear rates from 0.01 to 100 s −1 . Frequency sweeps were performed at a constant strain of 1% from 0.1 to 100 rad/s. dibenzocyclooctyne group (DBCO)functionalized Alexa Fluor 647 dye (DBCO-AF647; Jena Bioscience, Jena, Germany) was conjugated to azide-functionalized PEG-PLA nanoparticles through a copper-free click reaction. A 1.5× molar excess of dye to nanoparticle was added to the nanoparticle solution overnight at room temperature. The nanoparticles were then concentrated through centrifugation to remove any excess dye. To prepare the fluorescent PNP hydrogel, 50% dye-functionalized PEG-PLA nanoparticles were used in combination with 50% non-functionalized PEG-PLA nanoparticles. An in vitro imaging system, the Lumina Imager (PerkinElmer, Waltham, MA), was used to image fluorescent PNP hydrogel retention in the mouse eye. An exposure time of 0.1 second was used to quantify PNP hydrogel retention as the total flux of photons in the region of interest over time.
Statistical Analysis
Results in the figures represent mean ± SEM. Statistical analysis was done in Excel (Microsoft, Redmond, WA) and R (R Project for Statistical Computing, Vienna, Austria). Unpaired two-tailed Student's t-tests and one-way analyses of variance were performed to determine P values. Additionally, Kruskal-Wallis nonparametric tests were performed wherever indicated. P ≤ 0.05 was considered significant.
Conjunctival Injury With 1N NaOH Induced Forniceal Shortening and Formation of Symblephara in Mice After 7 Days
In control mouse eyes, the palpebral and bulbar conjunctiva line the eyelid and globe, respectively, forming the fornix at the junction (Fig. 1). The conjunctival fornix is required for ocular surface health. It provides a deep tear reservoir, allows smooth contact between the eyelid and globe, and permits full ocular motility. Alkali injury to the ocular surface causes forniceal shortening, leading to dry eye disease, formation of symblephara, and possible ocular motility restriction. To evaluate and test anti-scarring therapies for dry eye disease and symblephara, we first created a mouse conjunctival injury model using NaOH at increasing concentrations. The resulting conjunctival injury was evaluated based on the extent of forniceal shortening and conjunctival inflammation without globe destruction. Filter paper discs (3.5mm) soaked in increasing NaOH concentrations of 0.25N, 0.5N, 1N, and 2N were placed under the mouse eyelids such that the bulbar and palpebral conjunctiva were in full contact with the soaked filter paper for 2 minutes. H&E-stained paraffin sections 7 days after injury showed no significant injury up to 0.5N NaOH (Figs. 2a-2c). At 1N NaOH, forniceal shortening was observed accompanied by conjunctival inflammation and increased immune cells at the injury site (Fig. 2d). At 2N NaOH, there was complete globe destruction after injury (data not shown). To determine the injury extent over time, we evaluated the mice at 1, 5, and 7 days after injury with 1N NaOH for 2 minutes (Fig. 3). There was minimal conjunctival injury in the H&Estained paraffin sections after 1 day (Fig. 3a), such that the fornix and conjunctiva appeared normal. Five days after injury, the conjunctiva appeared damaged with adhesions between the palpebral and bulbar conjunctiva (Fig. 3b). Seven days after injury, the conjunctival adhesions were clearly evident, resulting in forniceal shortening, fibroblast infiltration, conjunctival inflammation, and fibrosis (Fig. 3c). Mice observed 14 days after alkali injury showed globe destruction (data not shown); thus, 7 days was deemed the optimal formation time for symblephara.
PNP Hydrogel Prevented Forniceal Shortening After Alkali Injury
PNP hydrogels have proven to be an effective therapy for pericardial and abdominal fibrotic adhesions in rats and sheep following surgery. 20,21 The primary formulation evaluated in this study was comprised of 1 wt% of dodecyl-modified HPMC-C 12 polymers and 5 wt% PEG-PLA nanoparticles (D H ∼ 35 nm), which is denoted as PNP-1-5. PNP hydrogels of different formulations exhibit tunable shear-thinning, self-healing, yield-stress, and viscoelastic properties that enable them to create a lubricious barrier between organs and tissues that reduces the incidence and severity of adhesions. [22][23][24] We have previously demonstrated that hydrogels that are too soft and weak flow too easily when perturbed in the body, leaving target tissues too quickly while adhesions are still developing. Hydrogels that are too stiff are easily dislodged from target tissues as the yield stress of such materials exceeds their adhesion strength to the tissue. For these reasons, we sought to make use of modified HPMC as a functional polymer component, as this cellulosic derivative is known to be strongly muco-and tissue-adhesive. 20 Rheological characterization of PNP-1-10 hydrogels has demonstrated that these materials exhibit solid-like characteristics with G > G across a broad range of frequencies, suggesting that they can form a robust barrier over the eye for significant timescales. At a frequency of 10 rad/s and 1% strain, the stiffness (G ) of these materials was determined to be 13 Pa (Fig. 4a). These PNP-1-10 materials exhibited extreme shear thinning (3 orders of magnitude reduction in viscosity with increased shear rates), suggesting that these hydrogels can be easily spread over the eye by manual application (Fig. 4b).
Additionally, the PNP hydrogel exhibits a functional yield stress, which is required for robust local persistence upon administration, as demonstrated by a flow sweep measurement where the stress was observed to asymptote to approximately 1 Pa at low shear rates (Fig. 4c).
To evaluate the anti-scarring and anti-fibrotic properties of the PNP hydrogel in the conjunctival adhesion mouse models as a preventative therapy for symblephara, we tested fluorescently labeled PNP hydrogel at a 1:10 concentration in control mice. Fluorescently labeled PNP hydrogel was applied on the ocular surface and under the eyelids to form a barrier between the palpebral and bulbar conjunctiva at time 0 (Fig. 4d). In vivo imaging under the PerkinElmer IVIS fluorescent imaging system showed that fluorescent PNP hydrogel was present on the ocular surface at 3 hours (Fig. 4e) and 6 hours (Fig. 4f); however, by 12 hours (Fig. 4g), it was not detectable. Quantification of the region of interest at 6 hours and 12 hours showed a significant decrease in total flux (Fig. 4h).
To determine the efficacy of the PNP hydrogel in preventing conjunctival adhesions after alkali injuries, we applied the PNP hydrogel to the ocular surface and under the eyelids directly after 2-minute ocular surface injury with 1N NaOH. Seven days after injury, mice with the PNP hydrogel were compared to mice without PNP hydrogel 7 days after their injury (Figs. 5a-5e). As observed in H&E-stained sections, the injured mice that received PNP hydrogel had less severe adhesions and did not show forniceal shortening compared with injured mice without PNP hydrogel (Figs. 5f-5i).
Corneal injury after 1N NaOH injury, as observed by H&E-stained sections, was found to be more severe than after PNP hydrogel application (Figs. 5j-5o). Injured mice showed acute and chronic inflammatory infiltration, corneal edema, and loss of corneal clefts (Figs. 5k, 5n). Corneas with PNP hydrogel application showed milder corneal edema, as observed by corneal clefts and lesser inflammation in the anterior chamber.
PNP Hydrogel Did Not Prevent Conjunctival Goblet Cell Damage After Alkali Injury
Dry eye disease can be caused by alkali injury to the conjunctiva and from forniceal shortening. Conjunctival goblet cells produce mucins that keep the ocular surface hydrated and lubricated. To measure goblet cell damage after injury and to assess the protective potential of the PNP hydrogel, we stained paraffinembedded sections with PAS, which stains mucinproducing conjunctival goblet cells (Fig. 6). Control mice conjunctiva showed healthy goblet cells stained in purple at the conjunctival epithelial surface (Fig. 6a). These goblet cells were absent from the badly damaged mouse conjunctivas from 2-minute injury with 1N NaOH for 7 days (Fig. 6b). PNP hydrogel application neither restored nor prevented goblet cell damage (Fig. 6c). PAS-stained goblet cell quantification showed no significant difference between goblet cell numbers in the injured compared with the PNP hydrogel-treated conjunctivas (Fig. 6d).
PNP Hydrogel Reduced Collagen Deposition, Myofibroblast Infiltration, and MSLN Expression After Alkali Injury
Collagen deposition is a hallmark of fibrotic disease. To determine the collagen deposition extent after injury, we performed Masson's trichrome staining, which turns collagen blue, on paraffin-embedded sections. Although collagen is intrinsically expressed in mouse cornea, there is minimal collagen surrounding the forniceal region in control eye sections (Fig. 7a). Seven days after injury with 1N NaOH, collagen deposition was observed in the forniceal region along with fibroblasts (Fig. 7b). In paraffin sections from mice that were given topical PNP hydrogel after injury, there was less collagen deposition (Fig. 7c). Quantification of the distance from fornix to the levator superioris muscle in the Masson's trichrome-stained sections showed a significant increase in the collagen deposition in injured mouse tissue compared with control and PNP hydrogel-treated tissues. Additionally, we stained for the expression of α-SMA, which is a myofibroblast marker at the injury site (Fig. 7d). We observed an increase in α-SMA-positive cell expression in the injured tissue that was lower in PNP hydrogel-treated tissues, indicating that fibrosis may be prevented by PNP hydrogel application (Figs. 7e-7g). To evaluate the potential mechanism of adhesion formation in the mouse alkali injury model, we stained the injury site with mesothelin (MSLN), a marker of mesothelial cells. We observed an increase in MSLN expression in the forniceal region with alkali injury compared with control (Figs. 7h, 7i), which was reduced in PNP hydrogel-treated eyes (Figs. 7j, 7k).
Discussion
Ocular surface alkali injury can lead to palpebral and bulbar conjunctival adhesions that form symblephara. To better understand how symblephara form and to evaluate preventive therapies, we have created an ocular conjunctival injury mouse model using NaOH. Alkali has been used to injure the cornea in mice 9,11,13 ; however, these studies did not focus on the conjunctival effects as well as subsequent formation of symblephara. Prior studies have explored alkali injury on canine 25 and rabbit 26 conjunctiva using NaOH; however, genetic and therapeutic studies can be challenging in large animal models, in addition to the longer 4 to 5 weeks required for symblephara to develop. In our model, alkali injury caused the formation of symblephara with forniceal shortening without globe destruction within 7 days after injury. This finding provided a unique model for studying ocular alkali burns and cicatrizing conjunctival diseases and for evaluating preventative therapies. We also found that a novel topical PNP hydrogel can prevent forniceal shortening and the formation of symblephara up to 7 days after alkali injury. An ideal animal model for symblephara should have tunable ocular surface injury without globe disruption and forniceal shortening due to adhesions. 26 One rabbit model of symblephara used a 10-mm filter paper crescent soaked in 2N NaOH for 90 seconds to induce symblephara. 26 In this study, symblephara formed in the late stages of alkali burns, around 4 weeks after NaOH injury, with fibroblast recruitment in the subconjunctival collagen fibers. Similarly, in a canine model of symblephara, symblephara formed 5 weeks after injury with 1N NaOH for 90 seconds. 25 In our mouse model, we observed forniceal shortening within 7 days of injury with 1N NaOH for 2 minutes. H&E and α-SMA stained tissue sections show fibroblast and myofibroblast infiltration at the injury site with eyelid margin deformation and crusting. The histological and molecular findings for mouse symblephara are similar to those for the rabbit and canine models, with the advantages that the symblephara form more quickly in mice and the ease of animal handling.
New symblephara preventive therapies would greatly benefit patients with alkali burns or cicatrizing disease. Previous studies have evaluated antiadhesion materials made of cross-linked gelatin films, 27 sodium hyaluronate, and sesame oil 28 for their anti-inflammatory and anti-oxidative properties. Other studies have tested amniotic membrane and seprafilm as a barrier to prevent adhesions and symblephara. 19,29,30 Anti-adhesive materials have been mostly studied in the context of postoperative adhesions that form due to normal wound healing after surgery.
The molecular mechanisms involved in ocular adhesions are currently not well understood. JUN expression has been implicated in abdominal adhesions 31 where JUN was suppressed to prevent abdominal adhesions. Adhesions may form after mesothelial cell loss or damage at the injury site marked by MSLN and anti-MSLN antibodies. 32 MSLN has been explored as a therapeutic target for adhesion prevention. Although PNP hydrogel prevents the formation of symblephara after alkali injury, it may also prevent other cicatrizing diseases that involve ocular adhesions. We found an upregulation of MSLN after alkali injury, indicating that adhesion might originate due to loss of mesothelial cells at the site of injury. Our mouse model for symblephara may also be used to test molecular targets, such as JUN and MSLN, as the ocular surface contains mesothelial cells 33 that are involved in adhesion formation.
In the mouse model for symblephara, we found that PNP hydrogel prevented forniceal shortening. Although both of the products used clinically to prevent peritoneal adhesions, INTERCEED (Johnson & Johnson, New Brunswick, NJ) and Seprafilm (Baxter, Deerfield, IL), contain hyaluronic acid and carboxymethylcellulose, 34 these materials exhibit very poor performance, as they are easily dislodged from the site of administration. 20,34 Yet, Seprafilm was nevertheless found to be superior to polylactic acid, hyaluronic acid, chitosan, and PEG materials when evaluated in a rat cecum abdominal adhesion model. 35 In contrast, PNP hydrogels were found to be far superior to both INTERCEED and Seprafilm in rat and sheep models of abdominal and pericardial adhesions. 20,21 PNP hydrogels are composed of hydrophobically modified HPMC and PEG-PLA nanoparticles that imbue them with the unique ability to adhere to tissue surfaces, as demonstrated in rat and sheep models of postoperative adhesions. 20,21 In the present study, PNP hydrogels more easily reached the deep forniceal because of their superior tissue adherence and unique yield-stress and viscoelastic properties. 20,21 Furthermore, the PNP hydrogel shearthinning and self-healing properties make them easy to handle and apply to the eye.
The presence of myofibroblasts (activated fibroblasts) is a key feature of pathological tissue repair and are upregulated in the conjunctival epithelial cells during fibrosis. 36 α-SMA is a myofibroblast marker expressed at the injury site. 37 α-SMA was significantly expressed in the forniceal region of our alkali-injury mouse model. This fits with other conjunctival injury models where α-SMA was upregulated in the conjunctiva after injury. 38 Collagen deposition is known to occur in fibrotic disease. 38,39 In the eye, collagen deposition after conjunctival injury has been observed in mice using allergic eye disease models 40,41 and a physical injury model 42 where collagen fibers were detected in the subconjunctival wounded area 7 days after injury. We observed collagen deposition in the alkali-injury mice with symblephara 7 days after injury which was reduced by treatment with PNP hydrogel. Goblet cell density was evaluated in a keratoconjunctivitis sicca mouse model 43 where goblet cells were reduced in the injured model, causing decreased mucoproteins. Goblet cells were almost completed ablated after alkali injury and were not protected by the PNP hydrogel, perhaps because they were immediately destroyed after NaOH application. Testing different PNP hydrogel concentrations may improve this effect in the future.
In conclusion, we describe, to the best of our knowledge, the first model for murine alkali-injury symblephara that can be useful for studying the pathophysiology of conjunctival scarring diseases. We tested a novel supramolecular PNP hydrogel that successfully prevents the formation of symblephara when applied after injury. | 2022-02-23T06:23:24.192Z | 2022-02-01T00:00:00.000 | {
"year": 2022,
"sha1": "5d13021fe988efb19660bef9116ec03487370c5e",
"oa_license": "CCBYNCND",
"oa_url": "https://doi.org/10.1167/tvst.11.2.31",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "7b95f26b981a76c9dff9b5f4dcc9221aa578b5c6",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
247061371 | pes2o/s2orc | v3-fos-license | Real-Time Visualization of the Infection and Replication of a Mouse-Lethal Recombinant H9N2 Avian Influenza Virus
H9N2 avian influenza viruses (AIVs) continuously cross the species barrier to infect mammalians and are repeatedly transmitted to humans, posing a significant threat to public health. Importantly, some H9N2 AIVs were found to cause lethal infection in mice, but little is known about the viral infection dynamics in vivo. To analyze the real-time infection dynamics, we described the generation of a mouse-lethal recombinant H9N2 AIV, an influenza reporter virus (VK627-NanoLuc virus) carrying a NanoLuc gene in the non-structural (NS) segment, which was available for in vivo imaging. Although attenuated for replication in MDCK cells, VK627-NanoLuc virus showed similar pathogenicity and replicative capacity in mice to its parental virus. Bioluminescent imaging of the VK627-NanoLuc virus permitted successive observations of viral infection and replication in infected mice, even following the viral clearance of a sublethal infection. Moreover, VK627-NanoLuc virus was severely restricted by the K627E mutation in PB2, as infected mice showed little weight loss and a low level of bioluminescence. In summary, we have preliminarily established a visualized tool that enables real-time observation of the infection and replication dynamics of H9N2 AIV in mice, which contributes to further understanding the mechanisms underlying the pathogenic enhancement of H9N2 AIV to mice.
INTRODUCTION
H9N2 avian influenza virus (AIV), one of the primary subtypes of influenza virus circulating in poultry, has been widely distributed around the world (1-4). Since the H9N2 AIV was first isolated from turkeys in the US in 1966, it has been rapidly transmitted in all continents (5)(6)(7). In Asia, the H9N2 AIVs have become endemic in poultry in different countries, posing a huge economic loss to the poultry farming industry (8)(9)(10)(11)(12). Furthermore, H9N2 AIVs have conduced to some zoonotic events by providing the internal segments to reassortment viruses such as H7N9, H5N6, and H10N8 (13,14). Strikingly, it has been reported that H9N2 AIVs continuously cross the species barriers to infect mammalians, including dogs, pigs, and humans (15)(16)(17). As of December 17, 2021, 93 cases of human infection with H9N2 influenza viruses have been reported worldwide, since the first human infection case was confirmed in 1998 (18,19). Previous studies have shown that the H9N2 AIVs generally caused mild infections in poultry, even in humans. Notably, some H9N2 AIVs were found to replicate efficiently and possess lethality in mice (20)(21)(22), but the viral infection dynamics in vivo and the detailed mechanisms of increased pathogenicity of H9N2 AIVs to mice remain unclear, posing a potential threat to public health.
Currently, multiple influenza reporter viruses with different characteristics have been successfully used for in vitro or in vivo studies, such as the screening of antiviral compounds or neutralizing antibodies, the development of vaccines, or the identification of host factors (23)(24)(25)(26). One of the common methods for generating a valid influenza reporter virus is to insert a reporter gene into the middle of the NS segment because it is the minimal genome of the influenza virus and this strategy can prevent any adverse effect on viral packaging signals (24,25,27,28). Furthermore, different reporter genes serve different purposes, and the choice of the best reporter gene depends on the type of study. For instance, a fluorescent gene like GFP is most useful to observe localization in cells (24,26), but luciferase is more valuable for quantitative purposes (29,30). For in vivo imaging, luciferase reporters are preferred over fluorescent proteins because the sensitivity and specificity of fluorescent imaging are regularly interfered with by tissue auto fluorescence, resulting in a large amount of background (31). To analyze the viral infection dynamics in mice, the NanoLuc luciferase that possesses ∼150-fold greater activity than either firefly (Photinus pyralis) or Renilla luciferases (32) was selected to generate our reporter virus.
Based on the strategy of the NS segment generating a reporter virus, we produced a recombinant H9N2 AIV expressing the NanoLuc luciferase (V K627 -NanoLuc virus) by reverse genetics. V K627 -NanoLuc virus possessed similar virulence and replicative capacity in mice to its parental virus. Bioluminescent imaging of the V K627 -NanoLuc virus allowed continuous observations of viral dynamics in infected mice. Overall, these results provide an effective tool to study the real-time infection dynamics of H9N2 AIV in mice, facilitating the exploration of the mechanisms of enhanced virulence of H9N2 AIV to mice.
Plasmids and Cells
The eight-plasmid reverse genetics system for H9N2 avian influenza virus A/Chicken/Guangdong/V/2008 (V K627 ) was previously described (21). The non-structural (NS) gene of the influenza A virus encodes an mRNA transcript that is alternatively spliced to express two viral proteins, the nonstructural protein 1 (NS1) and the nuclear export protein (NEP). The method to generate the NS1-NanoLuc-NEP segment referred to the previously reported study (27). Briefly, the NanoLuc coding sequence was placed downstream of the NS1 coding sequence via a GSGG linker, followed by a sequence of 2A peptide from porcine teschovirus (PTV-1 2A) and the NEP coding sequence. Furthermore, silent mutations were introduced into the endogenous splice acceptor site of the NS1 ORF to prevent splicing (33). The NS-NanoLuc-NEP segment maintained the non-coding sequence of the NS segment at both ends. Finally, the complete NS1-NanoLuc-NEP segment was cloned into plasmids pHW2000 with BsmBI restriction sites. Madin-Darby canine kidney (MDCK) and Human embryonic kidney cells (HEK293T) were maintained in Dulbecco's modified essential medium (DMEM) with 10% fetal bovine serum (FBS) at 37 • C in 5% CO 2 .
Generation of Reassortant Viruses
V K627 -NanoLuc virus was rescued by reverse genetics techniques (34). Briefly, 500 ng of each plasmid encoding the seven gene segments of V K627 and the NS-NanoLuc-NEP segment were transfected into HEK293T cells in six-well plates by using Lipofectamine 2000 (Invitrogen). After 6 hours, the medium was replaced with Opti-MEM (Gibco) including 1 µg/ml L-1-Tosylamide-2-phenylethyl chloromethyl ketone (TPCK) treated trypsin (Sigma). 48 h later, the HEK293T cells were resuspended and harvested in the medium, and the mix was injected into specific pathogen-free (SPF) chicken embryonated eggs. Then V K627 -NanoLuc virus was confirmed by performing a hemagglutination assay. The V K627E -NanoLuc virus was rescued by a single K627E mutation in the PB2 protein of V K627 -NanoLuc virus with the same protocol. V K627 parental virus was as a control, which contained PB2 residue K627 and was lethal to mice in our previous study (21).
RNA Extraction, RT-PCR, and DNA Sequencing
The rescued recombinant virus (V K627 -NanoLuc virus) was labeled as P0 and then passaged in SPF embryonated eggs for three generations (passage 1-3, P1-P3). RNA obtained from viral stocks (P0-P3) was extracted with the RNeasy minikit (Qiagen) as directed by the manufacturer. The sequences of the NanoLuc gene in the viral stocks were confirmed by reverse transcription-PCR (RT-PCR) reported by Hoffmann et al. (35) and sequencing. The specific primers used in this study were as follows (5 ′ -3 ′ ): NanoLuc gene (Forward: ATGGTCTTCACACTCGAA; Reverse: CGCCAGAATGCGTTCGC).
Western Blot Analysis
MDCK cells grown in six-well plates were infected with V K627 -NanoLuc virus at a MOI of 1, and cells were lysed using RIPA Lysis Buffer (Beyotime) at 24 hours post-infection (hpi). The protein samples were followed to SDS/PAGE and transferred to a nitrocellulose membrane. NS1 protein and NS1-NanoLuc fusion protein were detected by immunoblotting with a mouse polyclonal anti-NS1 antibody (GeneTex, dilution 1:1,000) and the viral nucleoprotein (NP) was detected by immunoblotting with anti-NP antibody (SinoBiological, dilution 1:1,000), and followed by IRDye 800CW, goat anti-mouse IgG (LC-COR, dilution 1:10,000). The membrane was imaged using an Odyssey infrared imaging system (Li-CoR, United States).
Virus Growth Kinetics
MDCK cells grown in six-well plates were infected at a multiplicity of infection (MOI) of 0.001 at 37 • C, as described previously (36). One hour later, cells were washed with PBS, and then incubated with DMEM containing 1 µg/ml TPCK trypsin at 37 • C with 5% CO 2 . Culture supernatants were collected at 12, 24, 36, and 48 hpi. Titers were determined by performing 50% tissue culture infective dose (TCID 50 ) assays on MDCK cells.
Mice Experiments
Pathogenicity of Viruses: 5-week-old female BALB/c mice (n = 5/group) (Guangdong Medical Lab Animal Center) were anesthetized with isoflurane and inoculated intranasally with V K627E -NanoLuc and V K627 -NanoLuc virus at a dose of 10 6 EID 50 /50 µl. The body weight and survival of mice were monitored daily for 14 days. Mice were humanely euthanized when they lost more than 25% of their body weight. Determination of LD 50 : 5-week-old female BALB/c mice (n = 5/group) were anesthetized with isoflurane and inoculated intranasally with V K627 virus and V K627 -NanoLuc virus at doses of 10 3 , 10 4 , 10 5 , or 10 6 EID 50 /50 µl. The body weight and survival of mice were monitored daily over a period of 14 days. Mice were humanely euthanized when they lost more than 25% of their body weight. LD 50 values were calculated by the method of Reed and Muench. To detect the viral replication in the lungs of mice, another three mice from the group of 10 6 EID 50 /50 µl were euthanized at 5 days post-infection (dpi) and viral titers in the lungs were determined by an EID 50 assay.
In vivo Imaging
The same mouse from each group was imagined at the appointed time by using an IVIS imaging system. Briefly, infected mice were anesthetized with isoflurane and 100 µl Nano-Glo reagent (Promega, dilution 1:25) was injected into retro-orbitally of mice. The Living Image software was used for image acquisition and analysis. Flux measurements were automatically calculated from the signal of mice. All data of composite images used the same scale.
Statistics
Data represent means ± standard deviations (SD) (n ≥ 3) unless otherwise noted. Multiple comparisons were performed by using an unpaired t-test in the GraphPad Prism software (GraphPad Software Inc.). Significance is defined as P < 0.05 and is indicated with an asterisk ( * ).
Generation of a Mouse-Lethal Influenza Reporter Virus
The NS segment of the influenza virus encodes NS1 protein and NEP protein produced from unspliced mRNA and spliced mRNA, respectively ( Figure 1A). It has been previously reported that NS1 and NEP can be expressed and separated effectively by the cleavage site of 2A from porcine teschovirus (PTV-1 2A) during translation (27). In this study, we altered the NS segment of the V K627 virus, which led to NS1-NanoLuc as a fusion protein with PTV-2A cleavage site, allowing NEP protein to be separated from the NS1-NanoLuc fusion protein during translation ( Figure 1A). By using reverse genetics techniques, a recombinant H9N2 AIV encoding the NanoLuc luciferase (V K627 -NanoLuc virus) was rescued (marked as P0). The amino FIGURE 2 | The pathogenicity of V K627 -NanoLuc virus in vivo. BALB/c mice (n = 5) were infected intranasally with 10 3 , 10 4 , 10 5 , or 10 6 EID 50 /50 µl of V K627 virus. Survival (A) and body weight (B) were monitored for 14 days. BALB/c mice (n = 5) were infected intranasally with 10 3 , 10 4 , 10 5 , or 10 6 EID 50 /50 µl of V K627 -NanoLuc virus. Survival (C) and body weight (D) were monitored for 14 days. (E) Virus titers in the lungs of infected mice with 10 6 EID 50 /50 µl at 5 days post-infection (dpi). The data are shown as means ± standard deviations (SD) (n = 3). Statistical significance was determined by an unpaired t-test (*P < 0.05, ns: no significant difference).
Frontiers in Veterinary Science | www.frontiersin.org acid at position 627 of the PB2 protein on both V K627 -NanoLuc virus and parental virus V K627 is lysine (K).
The stability of the NanoLuc reporter gene was tested following amplification in SPF embryonated eggs. The rescued V K627 -NanoLuc virus (P0) was propagated in SPF embryonated eggs for three generations (passages 1 to 3, P1-P3). RT-PCR and Sanger sequencing were performed on RNA obtained from viral stocks. As shown in Figure 1B, the NanoLuc gene was detectable using specific primers in all the V K627 -NanoLuc viruses. To detect the expression of NS1-NanoLuc fusion protein, Western blot analysis was performed in MDCK cells (Figure 1C). In the V K627 virus-infected cells, NS1 protein (∼ 25 kDa) was detected. In V K627 -NanoLuc virus-infected cells, the band corresponding to NS1-NanoLuc fusion protein (∼ 43 kDa) was observed. These results showed that the V K627 -NanoLuc virus expressed NS1 protein and NanoLuc luciferase as a fusion protein.
To test whether the presence of a longer NS1-NanoLuc-NEP segment in V K627 -NanoLuc virus affected the viral replication in culture, we compared the growth kinetics of V K627 -NanoLuc virus and parental virus V K627 in MDCK cells ( Figure 1D). MDCK cells were infected with V K627 -NanoLuc and V K627 virus at a MOI of 0.001, and the viral titers in the supernatant at various hpi were determined by TCID 50 assay. Compared with the V K627 virus, V K627 -NanoLuc virus showed attenuated replication in MDCK cells, with maximum titers reaching up to 5.60 lgTCID 50 /Ml. This indicated that V K627 -NanoLuc virus replicated efficiently in MDCK cells, though reaching about 10fold lower titers than the parental virus.
V K627 -NanoLuc Virus Causes Severe Pathogenicity in Mice V K627 virus caused significant pathogenicity in mice as previously described in our study (21,37). To test whether the V K627 -NanoLuc virus was comparable pathogenicity to its parental virus, five mice of groups were infected with 10 3 , 10 4 , 10 5 , or 10 6 EID 50 /50 µl of V K627 virus and V K627 -NanoLuc virus, monitored daily for survival and weight loss. We found that survival and weight loss showed an intense dose-dependent effect (Figures 2A-D). In the parental virus-infected mice, all mice inoculated with 10 5 EID 50 /50 µl or higher and a minority of those infected with 10 4 EID 50 /50 µl succumbed to infection (Figures 2A,B). In the V K627 -NanoLuc virus-infected mice, all mice inoculated with 10 5 EID 50 /50 µl or higher succumbed to infection, whereas all mice infected with 10 4 EID 50 /50 µl survived (Figures 2C,D). According to the survival data, the LD 50 value of V K627 -NanoLuc virus was determined as 10 4.5 EID 50 /50 µl, only 2.3-fold higher than those of the parental virus (10 4.13 EID 50 /50 µl).
We determined the viral titers in the lungs of three mice infected with 10 6 EID 50 /50 µl at 5 dpi. The result showed that the V K627 -NanoLuc virus replicated efficiently in the lungs of mice, with mean titers reaching 4.33 lgEID 50 /200 µl, slightly lower than V K627 virus (4.83 lgEID 50 /200 µl) (Figure 2E). Combined, these data indicated that V K627 -NanoLuc virus replicated to high levels and caused significant pathogenicity in mice, which was close to those of the parental virus.
Real-Time in vivo Imaging of V K627 -NanoLuc Virus Infection
It has been proven that in vivo imaging of reporter viruses is useful to study the viral infection dynamics for different viruses, including influenza A virus (38,39), dengue virus (40,41), and vaccinia virus (42). To analyze the real-time infection dynamics of a mouse-lethal H9N2 AIV, V K627 -NanoLuc virus was used to visualize the replication. We performed infections with mice at indicated doses of V K627 -NanoLuc virus and the same mouse from each group was imagined at 3 and 5 dpi by using an IVIS imaging system. Longitudinal imaging of the same infected mouse from each group showed that the bioluminescent signal from V K627 -NanoLuc virus-infected lungs increased as the dose of the infection increased ( Figure 3A). To better explore the viral infection dynamics, a single mouse was infected with the V K627 -NanoLuc virus at a sublethal dose of 10 4 EID 50 /50 µl. Then we imagined the infected mouse and monitored its body weight at 3, 5, 7, 9, and 11 dpi, respectively. The correlation could be observed between the body weight and the bioluminescent signal from the same mouse infected with the V K627 -NanoLuc virus ( Figure 3B). The bioluminescent signal was mainly detected in the left lung accompanied by weight loss in the mouse at 3 dpi. The bioluminescent signal peaked and displayed on both sides at 5 dpi, suggesting the spread of the influenza virus to the right lung. Bioluminescent intensity decreased at 7 dpi, suggesting the viral infection began to decline. The bioluminescent signal was undetectable at 9 dpi, accompanied by a rise in body weight, indicating the clearance of infection was continued. In brief, the V K627 -NanoLuc virus was capable of permitting serial observation of viral replication and clearance in real time.
More evidence has demonstrated that PB2 residue K627 is a critical factor of virulence and host range for influenza A viruses, which can significantly facilitate the adaptability of viruses and contribute to enhanced virulence in mice (43,44). In this study, our reporter virus harboring PB2 residue K627 causes high pathogenicity in mice, showing similar results to those of the previous study. To test whether the K627E mutation in PB2 protein had an impact on the virulence of our reporter virus, the V K627E -NanoLuc virus was rescued by a single K627E mutation in the PB2 protein of the V K627 -NanoLuc virus. We performed infections in mice with 10 6 EID 50 /50 µl of V K627E -NanoLuc virus and V K627 -NanoLuc virus, monitoring daily for survival and weight loss, and imagined the infected mice at 3 and 5 dpi by an IVIS imaging system. As expected, mice infected with V K627 -NanoLuc virus lost weight and succumbed to infection, displaying a strong bioluminescence signal. In contrast, mice infected with V K627E -NanoLuc lost little weight and survived, showing a low level of bioluminescence signal (Figure 4).
DISCUSSION
H9N2 AIVs have been distributed in various avian species worldwide and are repeatedly spread to mammals. In general, H9N2 infections in poultry or human are mild. However, some H9N2 AIVs were reported to cause lethal infection in mice (20)(21)(22). To better understand the mechanisms, we generated a mouse-lethal recombinant H9N2 AIV (V K627 -NanoLuc virus) to analyze the real-time infection dynamic of H9N2 AIV in mice. Our work has demonstrated that the virulence and replicative capacity of the V K627 -NanoLuc virus in mice were very close to its parental virus. Furthermore, in vivo imaging of our reporter virus was able to visualize viral replication in mice models.
The creation of replication-competent influenza reporter viruses is interfered with by the complex structure of segmented genomes (26). All viral genes are essential in vivo, so large insertions or a simple replacement of reporter genes may severely attenuate viral replication. Most importantly, the inappropriate location of the reporter gene insertion can interfere with viral packaging signals and subsequently affect virus assembly. In our reporter virus, a small NanoLuc reporter gene was inserted into the middle of the NS segment to circumvents the repetition of packing signals. In the same strategy, various recombinant influenza viruses carrying foreign genes in their NS segments have been successfully used in vitro and in vivo (24,27,28). Although V K627 -NanoLuc virus was attenuated for replication in MDCK cells, it replicated effectively in the lungs and showed near-native pathogenicity properties in mice. These improvements allowed us to monitor the infected mice in vivo during infection with our reporter virus.
More evidence has demonstrated that PB2 protein 627 plays a key role in virulence and host range for influenza A viruses, facilitating the adaptability of viruses in mammals (8,43,44). It's reported that the higher pathogenicity of H9N2 viruses in mice is associated with the PB2 E627K mutation (21,45). Adaptation of H9N2 AIV to mice results in multiple amino acid substitutions, including E627K in PB2 protein (43). Our previous study indicated that the PB2 E627K mutation of H9N2 AIV contributed to enhanced virulence in mice by inducing a higher level of glucocorticoids (GCs) (46). However, the detailed mechanisms that increase the pathogenicity of H9N2 AIV in mice are still poorly understood. In this study, the characteristics of our reporter virus harboring PB2 residue K627 were consistent with those of previously reported viruses, and our reporter virus could provide a practical tool for further studies on understanding the mechanisms of increased virulence of the H9N2 AIV to mice.
In summary, we have generated an influenza reporter virus encoding the NanoLuc luciferase, which was lethal to mice and was usable for in vivo imaging. Our report virus was able to track the real-time dynamics of infection and replication of H9N2 AIV in mice. Besides, our reporter virus could be applied to rapidly screen and assess the efficacy of antiviral therapies and vaccines.
DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.
ETHICS STATEMENT
The animal study was reviewed and approved by South China Agriculture University Institutional Animal Care and Use Committee. | 2022-02-24T15:11:01.938Z | 2022-02-24T00:00:00.000 | {
"year": 2022,
"sha1": "0f68068a1905f8297c88af888718753cad360675",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Frontier",
"pdf_hash": "0f68068a1905f8297c88af888718753cad360675",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
236320907 | pes2o/s2orc | v3-fos-license | The Influence of the Exclusion of Central Necrosis on [18F]FDG PET Radiomic Analysis
Background: Central necrosis can be detected on [18F]FDG PET/CT as a region with little to no tracer uptake. Currently, there is no consensus regarding the inclusion of regions of central necrosis during volume of interest (VOI) delineation for radiomic analysis. The aim of this study was to assess how central necrosis affects radiomic analysis in PET. Methods: Forty-three patients, either with non-small cell lung carcinomas (NSCLC, n = 12) or with pheochromocytomas or paragangliomas (PPGL, n = 31), were included retrospectively. VOIs were delineated with and without central necrosis. From all VOIs, 105 radiomic features were extracted. Differences in radiomic features between delineation methods were assessed using a paired t-test with Benjamini–Hochberg multiple testing correction. In the PPGL cohort, performances of the radiomic models to predict the noradrenergic biochemical profile were assessed by comparing the areas under the receiver operating characteristic curve (AUC) for both delineation methods. Results: At least 65% of the features showed significant differences between VOIvital-tumour and VOIgross-tumour (65%, 79% and 82% for the NSCLC, PPGL and combined cohort, respectively). The AUCs of the radiomic models were not significantly different between delineation methods. Conclusion: In both tumour types, almost two-third of the features were affected, demonstrating that the impact of whether or not to include central necrosis in the VOI on the radiomic feature values is significant. Nevertheless, predictive performances of both delineation methods were comparable. We recommend that radiomic studies should report whether or not central necrosis was included during delineation.
Introduction
Tumour morphology might be heterogeneous with alternating regions of relatively vital tumour tissue, mild hypoxia, severe hypoxia and necrosis [1]. Tumour hypoxia manifests itself predominantly in solid tumours [2]. Central necrosis of tumours occurs as a result of hypoxia and is caused by uncontrolled oncogene-driven proliferation without efficient vasculature, inducing a nutrient and oxygen shortage [1]. As a morphological marker, central necrosis is associated with poor prognosis in a variety of cancers [3][4][5], including non-small cell lung carcinomas (NSCLC) [6,7]. Larger regions of necrosis can be detected on 2-[ 18 F]fluoro-2-deoxy-D-glucose positron emission tomography ([ 18 F]FDG PET/CT) as an often centrally located region with little to no tracer uptake.
Radiomics aims to quantify the geometry and tracer uptake, including uptake heterogeneity, of tumours by using first order, shape and texture features and hypothesizing that these features can be used for tumour characterisation, prognostic stratification and response prediction in precision medicine [8]. It is uncertain whether regions of central necrosis should be added to the delineation of the tumour, since the effect of delineation methods on the predictive value of the radiomic signature remains unknown [9]. Semiautomatic tumour delineation methods used in radiomic analysis apply isocontours by using fixed or adaptive thresholds [10,11] or more advanced algorithms, such as the fuzzy locally adaptive Bayesian (FLAB) algorithm [12]. Although these methods have shown to be highly reproducible [10,13], they often underestimate the true (anatomical) tumour volume by excluding (up to a certain degree) regions of low tracer uptake. Some studies manually add the excluded regions of low tracer uptake to the volume of interest (VOI), but this is not always clearly reported. It is hypothesised that the addition of a region of central necrosis to the VOI may influence all three radiomic feature classes numerically. First order features might be affected by the addition of voxels with low grey levels and this skews the intensity histogram. Shape features might be influenced by the different 3D morphology of the VOI when central necrosis is included. Texture features, representing spatial relationships between voxels in terms of run lengths or size zones of the same voxel values or combinations of neighbouring voxel values, might change as well. The introduction of an area of central necrosis might, for instance, result in long runs with low values that will change the run length matrix and, as a result, the feature values. The Image Biomarker Standardisation Initiative (IBSI), which is an independent international collaboration working towards standardising the extraction of image biomarkers, provides reporting guidelines for radiomic studies but, up to this point, does not specify the need to report on the inclusion/exclusion of necrosis while describing the used segmentation method [14]. Moreover, the effect of the delineation method, including whether or not to include central necrosis, on the performance of the radiomic signature for predicting underlying tumour biology remains unknown [9].
This study explores how central necrosis influences PET radiomic analysis by assessing the differences in radiomic features and the predictive performance of features extracted from VOIs delineated using an isocontour method with and without the manual addition of the region of central necrosis for two datasets of NSCLC and pheochromocytomas or paragangliomas (PPGL), catecholamine-producing neuroendocrine tumours that arise from the chromaffin cells of the adrenal medulla and extra-adrenal sympathetic paraganglia [15].
Patient Population, Data Acquisition and Image Reconstruction
Subjects from two cohorts of patients who underwent an [ 18 F]FDG PET/CT in a single academic centre were retrospectively included to study the effect of different aspects of central necrosis on the radiomic analysis. A cohort of patients with non-small cell lungcarcinomas (NSCLC, n = 35), generally presenting a high tumour-to-background ratio, and a cohort of patients with pheochromocytomas or paragangliomas (PPGL, n = 77), generally presenting a low tumour-to-background ratio, were included.
The NSCLC cohort is a previously published prospective cohort [16]. Patients underwent a dynamic [ 18 F]FDG PET/CT scan with the primary tumour located centrally in the field of view using the Biograph Duo or Biograph 40 mCT (Siemens Healthineers, Erlangen, Germany) at the Radboud University Medical Center between 2009 and 2014. Only tumours with a diameter larger than 30 mm were included to minimise the influence of partial volume effects and to be able to reliably quantify uptake heterogeneity [17]. Imaging was in accordance with European Association of Nuclear Medicine (EANM) guidelines for tumour PET imaging [18]. Patients fasted for at least 6 h before imaging and serum glucose levels were below 8 mmol/L. Directly after the start of the acquisition, a standardised infusion of 3.45 MBq of [ 18 F]FDG per kilogram of body weight started. The final time frame (50-60 min p.i.) of the dynamic series was used in the current study. Voxel sizes were 2.56 × 2.56 × 3.38 and 1.59 × 1.59 × 2.03 mm 3 for the Biograph Duo PET/CT and Biograph 40 mCT PET/CT, respectively. This study has been reviewed and approved by the Commission on Medical Research Involving Human Subjects Region Arnhem-Nijmegen, the Netherlands. All patients signed an informed consent form.
The PPGL patients who underwent a [ 18 F]FDG PET/CT scan in the Radboud University Medical Center between 2011 and 2018 were retrospectively included. A selection of these patients has previously been described [15,19]. Static PET/CT images were acquired using the Biograph 40 mCT (Siemens Healthineers, Erlangen, Germany), in accordance with aforementioned EANM guidelines [18]. Patients fasted for at least 6 h and serum glucose levels were below 8 mmol/L. Image acquisition (3 or 4 min per bed position) started 60 (55-75) minutes after intravenous administration of [ 18 F]FDG (dosage according to a non-linear dosage regimen based on body weight; details can be found in Supplementary File S1). The reconstructed voxel size was 3.18 × 3.18 × 3.00 mm 3 . This retrospective database study has been reviewed and approved by the Commission on Medical Research Involving Human Subjects Region Arnhem-Nijmegen, the Netherlands. Informed consent was waived due to the retrospective nature of the study. Patients that objected to the use of their anonymised data were excluded.
Additional details on patient preparation, data acquisition, image reconstruction, image processing and radiomic analysis can be found in Supplementary File S1: the IBSI reporting guidelines [14].
Volumes of Interest Delineation
VOIs were delineated semi-automatically using 3DSlicer version 4.11 (www.slicer.org, accessed on 1 March 2021) [20] and in-house built software implemented in Python version 3.7 (Python Software Foundation, Wilmington, Delaware). The often peripheral region of the tumour showing increased [ 18 F]FDG uptake (VOI vital-tumour ) was delineated using a semi-automatic threshold-based method and corrected for local background [10]. A threshold of 41% of the peak standardised uptake value (SUV peak ) that was obtained using a sphere of 12 mm diameter [21] was selected, since the delineated tumour sizes of this method agreed best with pathological tumour sizes [22]. As tumours in the PPGL cohort showed low contrast between the tumour and surrounding tissue, boxing was applied to exclude the surrounding [ 18 F]FDG-avid tissues. VOI gross-tumour was generated by manually adding the volumes of central necrosis to VOI vital-tumour , using the low-dose CT as a visual reference. The necrotic tumour fraction (NTF) was determined using Equation (1): NTF = 1 − VOI vital-tumour /VOI gross-tumour (1) where a NTF of 0 indicates no central necrosis and a higher NTF indicates larger volumes of necrosis. Patients were selected when NTF > 0.
Radiomic Feature Extraction
For each selected patient, 105 radiomic features were extracted from VOI vital-tumour and VOI gross-tumour using PyRadiomics version 3.0 in Python version 3.7 (Python Software Foundation, Wilmington, DE, USA) [23]: 18 first order features, 14 shape features, 22 grey level cooccurrence matrix (GLCM) features, 16 grey level run length matrix (GLRLM) features, 16 grey level size zone matrix (GLSZM) features, 14 grey level dependence matrix (GLDM) features and 5 neighbouring grey tone difference matrix (NGTDM) features. A fixed bin size of 0.5 g/mL was applied.
Statistical Analysis
Statistical analyses were performed in SPSS version 25 (IBM Statistics, Chicago, IL, USA). Per cohort and for both cohorts together, differences in radiomic features extracted from VOI vital-tumour and VOI gross-tumour were assessed using a Wilcoxon signed-rank test or a paired t-test after testing for (log-)normality. Since over one hundred features are tested simultaneously, some features may show a significant difference between both delineation methods by chance, which increases the false discovery rate [24]. Therefore, the Benjamini-Hochberg multiple testing correction was performed [25]. The Benjamini-Hochberg correction determines the significance level for specific feature (p i ) using Equation (2): where i is the ranking of a feature when ranking all features based on the significance level of the paired t-test from smallest to largest, n is the total number of features and a is the original significance level (a = 0.05). Additional subset analyses of all patients based on the NTF and SUV max were performed by creating three equally-sized groups for low, medium and high values: NTF: ≤0.12, 0.12 < NTF ≤ 0.36, >0.36; SUV max : ≤4.61, 4.61 < SUV max ≤ 12.09, >12.09 g/mL. Differences in numbers of affected features per cohort and subgroup were assessed using the Fisher's exact test. Overlaps in the affected features per cohort and subgroup were visualised using Venn diagrams. For the PPGL cohort, the predictive performance for the underlying tumour biology of the radiomic models based on features derived from the different delineation methods was assessed by binary logistic regression in R version 3.6.0 (R Foundation for Statistical Computing, Vienna, Austria). Moreover, a radiomic model was created out of features from both delineation methods, assuming that both features contain different information. The response variable in regression was the noradrenergic biochemical profile of the PPGLs. Unsupervised feature selection or dimension reduction was performed to deal with multicollinearity and high dimensionality, which occurs when the number of features largely exceeds the number of patients. As a rule of thumb, 1 feature was selected for every 10 subjects [26] and 3 features were selected to be tested (PPGL dataset: n = 31 patients). The predictive performance of the radiomic models was not assessed for the NSCLC cohort since this cohort consisted of only 12 patients, which corresponds to only 1 feature to be tested and is inadequate to explain sufficient variance of the dataset. Dimension reduction in the PPGL dataset using redundancy filtering and factor analysis was performed using the FMradio (Factor Modeling for Radiomics Data) R-package version 1.1.1 [27]. Features were scaled (centred around 0, variance of 1), avoiding that features with the largest scale dominated the analysis. Redundancy filtering of the Pearson correlation matrix of features is performed with a threshold of τ = 0.95 and, from each group, one feature is retained. Factor analysis of the redundancy filtered correlation matrix with an orthogonal rotation was executed so that the first factor explained the largest possible variance in the dataset; the succeeding factors explained the largest variance in orthogonal directions. The sampling adequacy of the model, which is quantified by the Kaiser-Meier-Olkin (KMO) statistic, was predefined to be ≥0.9. The feature with the highest loading on a single factor was selected for regression analysis. The three selected features are associated with the noradrenergic biochemical profile using multiple binary logistic regression. Areas under the curve (AUC) of the receiver operating characteristic (ROC) of the radiomic models based on the three selected features for VOI vital-tumour , VOI gross-tumour and combined were computed and compared using DeLong's test for paired ROC curves. A sham experiment was conducted to validate the findings by randomisation of the outcome labels (noradrenergic biochemical profile) [28]. This takes into account the prevalence of the outcome and the distributions and multicollinearity of the radiomic features but uncouples their hypothesised relation. Binary logistic regression was performed and the sham experiment was repeated 100 times to calculate the mean AUCs.
Results
Patient characteristics of included patients with central necrosis (n = 43) are presented in Table 1. In the NSCLC cohort, central necrosis was observed in 12 out of 35 patients (34%). In the PPGL cohort, central necrosis was observed in 31 of 77 patients (40%; Figure 1). The three selected features are associated with the noradrenergic biochemical profile using multiple binary logistic regression. Areas under the curve (AUC) of the receiver operating characteristic (ROC) of the radiomic models based on the three selected features for VOIvital-tumour, VOIgross-tumour and combined were computed and compared using DeLong's test for paired ROC curves. A sham experiment was conducted to validate the findings by randomisation of the outcome labels (noradrenergic biochemical profile) [28]. This takes into account the prevalence of the outcome and the distributions and multicollinearity of the radiomic features but uncouples their hypothesised relation. Binary logistic regression was performed and the sham experiment was repeated 100 times to calculate the mean AUCs.
Results
Patient characteristics of included patients with central necrosis (n = 43) are presented in Table 1. In the NSCLC cohort, central necrosis was observed in 12 out of 35 patients (34%). In the PPGL cohort, central necrosis was observed in 31 of 77 patients (40%; Figure 1). At least 65% of the features were affected by the choice of delineation ( Table 2). The At least 65% of the features were affected by the choice of delineation ( Table 2). The PPGL population was influenced the most with 79% of affected features, compared to 65% in the NSCLC population, which is a significant difference between the populations (p = 0.031). Out of the 105 features, 61% were affected in the NSCLC cohort as well as the PPGL cohort. For all patients taken together, even 82% of the features were affected.
First order features were affected substantially, with at least 72% of features, followed by texture features with at least 66% of features. Shape features were affected the least, with 50% of affected features in both datasets. Of all texture feature classes, GLCM features were affected the least, with a maximum of 68% of features. For all other classes, the maximum number of affected features was at least 80%. The size of the NTF appeared to influence the number of affected shape features (nonsignificant; Table 3). For a small NTF, 36% of the features were affected, increasing to 57% and 71% for medium and large NTFs, respectively. Moreover, for small and medium NTFs 100% of the NGTDM features were affected compared to only 40% for a large NTF. Although nonsignificant, the value of the SUV max appeared to influence the number of affected features (Table 4) and a higher value resulted in more affected features (56%, 66% and 70% for low, medium and high values, respectively). This increasing trend could also be observed for all texture feature classes except for GLCM features, where the number of affected features decreased with an increasing SUV max .
Overlap in affected features between cohorts and subgroups is generally high (Figure 2), yet the affected shape features varied largely between cohorts and for different-sized NTFs. It can be observed that many features that were affected in one subset or cohort, were also affected in most of the other subsets or cohorts (Supplementary Table S1). Table 4. Numbers and percentages of features affected by the delineation method (VOI vital-tumour vs. VOI gross-tumour ) per feature class for the different subgroups based on SUV max . Differences in the number of affected features between subgroups were assessed using the Fisher's exact test. SUV max : maximum standardised uptake value, GLCM: grey level co-occurrence matrix, GLRLM: grey level run length matrix, GLSZM: grey level size zone matrix, GLDM: grey level dependence matrix, NGTDM: neighbouring grey tone difference matrix. For each of the three radiomic models evaluating predictive performance (VOI vital-tumour , VOI gross-tumour and combined) for the PPGL dataset, three factors were retained and the three best corresponding features were selected ( Table 5). The KMOs of the models were excellent (>0.96). AUCs varied 0.791-0.829, but were not significantly different between the radiomic models (VOI vital-tumour vs. VOI gross-tumour : p = 0.775; VOI vital-tumour vs. combined: p = 0.625; VOI gross-tumour vs. combined: p = 0.874; Figure 3). The mean AUCs of the sham experiments were lower (0.645-0.655) than the AUCs of the radiomic models, indicating the validity of the findings. Figure 3. ROC curves and AUCs for the different radiomic models: blue: VOIvital-tumour (features: first order minimum, shape surface area and GLCM informational measure of correlation 2); green: VOIgross-tumour (features: shape surface area, NGTDM complexity and GLDM dependence entropy); black: combined (features: VOIgross-tumour GLCM sum entropy, VOIvital-tumour shape maximum 3D diameter and VOIvital-tumour shape surface volume ratio). ROC: receiver operating characteristic, AUC: area under the ROC curve. first order minimum, shape surface area and GLCM informational measure of correlation 2); green: VOI gross-tumour (features: shape surface area, NGTDM complexity and GLDM dependence entropy); black: combined (features: VOI gross-tumour GLCM sum entropy, VOI vital-tumour shape maximum 3D diameter and VOI vital-tumour shape surface volume ratio). ROC: receiver operating characteristic, AUC: area under the ROC curve.
Discussion
In this study, we assessed the effect of the inclusion of central necrosis during tumour delineation on radiomic analysis in two cohorts of patients with NSCLC and PPGL. Around two-third of radiomic features showed significant differences between adaptive threshold delineation with and without manual addition of the region of central necrosis. Nevertheless, the predictive performance of radiomic models with and without central necrosis for the noradrenergic biochemical profile of PPGLs was not significantly different. Due to the low number of subjects, the predictive performance was not assessed for the NSCLC cohort.
At least 65% of all features were significantly affected after adjustment for multipletesting by the difference in delineation method. Less features were affected in the NSCLC cohort compared to the PPGL cohort (65% versus 82%, respectively), which is likely a result of lower power of the test due to a smaller cohort (12 versus 31, respectively).
More than 72% of the first order features, describing the distribution of voxel intensities in a histogram, significantly changed when central necrosis, i.e., lower intensity values, was added to the VOI. The number of affected first order features increases with a higher SUV max , which can be explained by the larger range of voxel values in the intensity histogram in the case of a high SUV max after the addition of a region with central necrosis.
At least 66% of the texture features, describing the spatial relationships between individual voxels in terms of run lengths, size zones of the same voxel values or combinations of neighbouring voxel values, were affected by central necrosis, which resulted in a change of the spatial relationships between the voxels. It is beyond the scope of this study to dive into the mathematical definition of all texture features, but we will highlight some of our findings and possible explanations.
Similarly to first order features, the number of affected texture features also increased with increasing SUV max . The introduction of a region with low grey levels might result in longer run lengths and size zones with low values. In a tumour with relatively high grey levels, this might result in a larger run length or size zone matrix, with high incidences for the low grey values (central necrosis) and for the high grey levels (edge of the tumour) and low incidences in the middle range. For a tumour with a lower SUV max , the matrices remain smaller, with incidences in the low and middle ranges and this results in different feature values.
Furthermore, it is remarkable that almost all normalised texture features were significantly different for both delineation methods for the NSCLC cohort, the PPGL cohort and all patients combined: The normalised GLCM inverse difference and inverse difference moment, the GLRLM grey level non-uniformity and run length non-uniformity and the GLSZM grey level non-uniformity and size zone non-uniformity were significantly different between both delineation methods. The normalised GLDM dependence non-uniformity was only different in the PPGL cohort and the combined cohort. Normalisation of GLCM features is performed to improve classification accuracy [29]. Normalised features are standardised for the number of elements in their respective matrix, i.e., the GLCM consists of the square of the number of discretised grey levels and the GLRLM consist of the product of the number of discretised grey levels and the maximal run length [14]. Since the number of discretised grey levels and the maximal run length increase by the addition of the region of central necrosis, it could result in a decrease in feature values, resulting in differences in feature values between delineation methods.
Compared to other features classes, shape features were affected least frequently by the choice of the delineation method, but 50% of shape features were still affected. Some shape features consider the outer diameter or morphology of the VOI, which, in most cases, did not expressively change when adding the region of central necrosis. Nevertheless, in some cases the region of central necrosis touched the outer surface of the volume of interest (3D U-shape) and caused some features to change. The number of affected shape features increased with the NTF as a result of a larger additional region of necrosis.
Several studies on repeatability and reproducibility showed that radiomic feature values are affected by delineation methods [30]. Unfortunately, overviews of repeatability and reproducibility on a feature level are scarce and are often limited to feature classes. Traverso et al. wrote a systematic review on repeatability and reproducibility of radiomic features and assessed to what extent (highly likely/probable/less likely) the different feature classes were affected by different processing steps of the radiomic pipeline [31]. They found that it is probable that semi-automatic VOI delineation exerts an adverse effect on repeatability and reproducibility of texture features. Moreover, in the case of shape features, this adverse effect is probable but when compared to shape features derived from CT, PET shape features are more reproducible. According to Traverso et al., first order features are less likely to be affected by the delineation method, which is in sharp contrast with our study that shows that first order features are affected in particular by the delineation method. They also present that entropy was consistent among the most repeatable and reproducible first order features [31]. However, in our study it can be observed that entropy is one of the features that is significantly affected by the delineation method in all cohorts and subgroups. Coarseness and contrast (GLCM as well as NGTDM), on the other hand, are considered among the least reproducible features [31,32], whereas our results show that coarseness and GLCM contrast are affected in only two and zero out of nine subgroups (Supplementary Table S1), respectively. This shows that non-repeatable or non-reproducible features are not the only features affected by central necrosis and therefore the choice of the delineation method concerning central necrosis should be considered in the design of radiomic studies.
Our study analysed the differences in radiomic features in two cohorts with different tumour types, showing a different tumour-to-background ratio. While the same image analysis (VOI delineation and radiomic feature extraction) was performed, acquisition and reconstruction settings were different between cohorts even though both protocols were in accordance with the EANM guidelines. The features affected by the delineation method, however, were highly similar between both cohorts, indicating that the effect of the delineation method concerning central necrosis is independent of the tumour type. Therefore, whether or not to include central necrosis in the tumour delineation is an important factor to consider when performing clinical radiomic studies. We hypothesise that this might also apply to other tumour types, but the number of affected features might vary as a result of tumour characteristics such as the tracer uptake and distribution, tumour geometry and NTF.
While almost two-third of radiomic feature values were significantly affected by the choice of delineation method, the predictive performances of the radiomic models, as assessed in the PPGL cohort, were not affected accordingly. The predictive performances, as assessed by the AUCs for the noradrenergic biochemical profile of PPGLs and found valid in a sham experiment, were not significantly different between radiomic features derived from VOIs with and without central necrosis. An explanation for this could be that the radiomic feature set describes many different types of heterogeneity and, as a result, feature sets from both delineation methods contain useful features in terms of predictive performance. Multicollinearity within one feature set is high, but might also be high between feature sets of different delineation methods. Additionally, the combination of radiomic features from both delineation methods resulted in an AUC similar to the ones of the different delineation methods seperately. This indicates that, in this small dataset, the combination of features from the two delineation methods does not result in additional information that is favorable for the predictive performance. It should be taken into account that, in order to prevent overfitting of the model, only three features could be retained in factor analysis. In other tumour types, multicollinearity in and between feature sets of different delineation methods is expected to be high as well, but further research is needed to confirm this effect in a larger population as well as for different tumour types.
This study showed that the effect of inclusion or exclusion of the region of central necrosis in the delineation significantly impacts radiomic feature values in PPGL and NSCLC, but does not impact the predictive performance of the PPGL radiomic model. A guideline on the choice to add or leave out central necrosis in delineation could not be provided. From a biological perspective, regions of central necrosis are part of the tumour and should therefore be included in the VOI, especially considering that central necrosis is associated with poor prognosis [6]. On the other hand, from a data-driven perspective, some features might already capture the presence of central necrosis without our awareness, since the features are investigated exploratively and without biological rationale. Both delineation methods can be used in radiomic studies, but feature values vary largely between both methods. For reproducibility purposes, especially in the setting of external validation [33], future studies should report whether regions of central necrosis were included in the delineation.
Conclusions
Central necrosis of tumours on [ 18 F]FDG PET significantly impacts radiomic feature values. Almost two-third of the features were affected, demonstrating that the influence of whether or not to include regions of central necrosis in the delineation of the tumour on the radiomic feature values is significant. However, no significant difference in the predictive performance of both delineation methods was observed. In order to advance reproducibility of radiomic research, radiomic studies should report on whether or not central necrosis was (manually) included during delineation. Informed Consent Statement: All NSCLC patients signed an informed consent form. For the PPGL study, informed consent was waived due to the retrospective nature of the study. Patients that objected to the use of their anonymised data were excluded.
Data Availability Statement:
The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request. | 2021-07-26T05:27:26.720Z | 2021-07-01T00:00:00.000 | {
"year": 2021,
"sha1": "1fde4e8677d1fb58c430098bd2133386022b7225",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2075-4418/11/7/1296/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "1fde4e8677d1fb58c430098bd2133386022b7225",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
56412315 | pes2o/s2orc | v3-fos-license | Grape and Apple Pomaces from Southern Brazil : Valorization of By-Products through Investigation of Their Antioxidant Potential
The extraction process of bioactive compounds from grape and apple pomaces was optimized using a response surface methodology to evaluate the effect of ethanol concentration, time and temperature on the extraction of total phenolic content, antioxidant activity by radical scavenging 2,2-diphenyl-1-picrylhydrazyl hydrate (DPPH) method and individual phenolics by high performance liquid chromatography (HPLC). Optimizing suggests that the best condition for total phenolic and gallic acid extraction from grape pomace was obtained in intermediate level (60% ethanol, 60 °C, 30 min) while 80% ethanol, 75 °C and 39 min was more efficient for p-coumaric acid extraction. For apple pomace the extract obtained with 50% ethanol at 75 °C and 39 minutes was more active by DPPH method and content of caffeic acid. Therefore, these abundant and inexpensive agro-industrial residues produced in Southern Brazil, are an important source of bioactive compounds that may be used as antioxidant substances in food and pharmaceutical industries.
Introduction
Phenolic compounds in food have received especial attention due to their beneficial effects to human health. 1,2ost of the biological properties of phenolics are related to their antioxidant activity.They are able to scavenge free radicals as reactive species of nitrogen and oxygen, naturally produced during aerobic metabolism. 3,4Studies have shown that daily consumption of antioxidants can protect the body against many diseases as cancers and cardiovascular diseases. 2][7][8] When processing grapes and apples to produce wine, juice and jams, significant amount of waste is generated (peel, seeds, pulp and stems).It is estimated that only 20% of fruit weight is used and the remaining is discarded 9,10 as animal feed and to make compost, destinations that do not generate economic gains for agribusiness and represent problems for disposal. 11Nevertheless, large amounts of the residual quantities of bioactive compounds are maintained into the vegetal tissues.Previous studies 12 have shown that approximately 58% of phytochemical remain in the pomace during industrial processing.
Therefore, pomaces of grapes and apples become valuable sources by representing significant economic gains for producers, becoming raw materials for other activities and offer a new possibility of use by preventing or reducing the problems and environmental impacts related to the accumulation and disposal of these products. 13,14here is a possibility of using these residues as natural antioxidant in the food and pharmaceutical industries 15 and as an antimicrobial agent, prolonging the shelf life of products by inhibiting the lipid oxidation, [16][17][18] enhancing color, flavor and aroma of foods. 19Shirahigue et al. 20 showed a delayed lipid oxidation in processed chicken meat stored under refrigeration added with grape residue extracts (mixture of seed and peel) from Isabel and Niagara varieties.
][23][24] The chemical composition of grape and apple pomaces is varied and complex and it is intimately related to the geographical region and climate where it is produced. 25owever, the available information on the polyphenolic and antioxidants from pomaces of Southern of Brazil, that has a wet temperature climate, with well-defined seasons is still limited and scarce, hindering the valorization of these by-products.
Extraction of bioactive compounds from plant materials is an important process and various techniques have been studied in this regard. 26The design of experiments (DOE) has been used in several studies of extractions for being considered a useful analytical tool, with application in the screening of the most relevant variables of a given analytical system. 27Therefore, it is important to apply DOE in studies covering many variables, as it makes possible the planning and carrying out in an organized way from a minimum number of experiments, saving time and financial resources and allows scientists to evaluate the influence of input and output variables aiming at increasing the process, improving the efficiency and quality of results. 28,29herefore, the aim of the present study was to optimize the extraction of phenolic compounds with antioxidant activity from grape and apple pomaces using central composite design (CCD).
Preparation of materials
The grape pomace samples (Vitis labrusca, Bordeaux variety) from the red wine processing were collected at a winery in Mariopolis city, Paraná state, Brazil and apple pomace samples (Malus domestica, Fuji variety) from juice processing were obtained in Treze Tílias city, Santa Catarina state, Brazil.Both collected in the year of 2014.The samples were lyophilized, grounded, weighed, packed and stored at −6 o C until analysis.
Statistica software was used for designing experiments and statistical data analysis (ANOVA).Experimental data were analyzed to fit a second order polynomial regression containing linear and quadratic coefficient and two-factor interaction effects.Models and regression coefficients were considered significant when p-values were lower than 0.05.The model equation of response (Y) of three independent variables (X1, X2 and X3) is: (1) where Y is the dependent variable (total phenolic compounds-TPC, DPPH, caffeic acid, gallic acid, vanillic acid, p-coumaric acid, ferulic acid and trans-resveratrol stilbene), β 0 is the constant coefficient, β i is the linear coefficient (main effect), β ii is the quadratic coefficient and β ij is the two-factor interaction coefficient.
A representative sample lyophilized of grape and apple pomaces (2 g) was extracted with a mixture of 25 mL of X3 ethanol:water mixture.The samples were homogenized using a vortex (Phoenix Luferco, model Ap56).After, the samples were heated at X1 ºC for X2 minutes and then cooled to room temperature and solvent was added if necessary to bring the final volume back to 25 mL.Ethanolic extract of grape pomace (EEGP) and ethanolic extract of apple pomace (EEAP) were obtained.All extracts were filtered through a qualitative paper for total phenolic content (TPC) and DPPH analysis and then filtered using 0.45 µm polytetrafluoroethylene (PTFE) syringe filter (Chromafil ® ) for chromatographic analysis and stored at −4 ºC until analysis.
Total phenolic content (TPC)
The total phenolic content was determined by the colorimetric analysis using Folin-Ciocalteau reagent as described by Singleton et al. 30 The reaction medium consisted of 0.5 mL of EEGP (1.6 g L -1 ) and EEAP 3.2 g L -1 , 2.5 mL of Folin-Ciocalteau reagent diluted 10 g L -1 and, after five minutes, it was added 2.0 mL of sodium carbonate (Na 2 CO 3 ) 40 g L -1 .The mixture was incubated at room temperature in the dark for 2 hours against ethanol as blank.Subsequently, the absorbance was measured using a spectrophotometer (model UV-Vis lambda 25, PerkinElmer) at 740 nm.Gallic acid was used as reference at concentration ranging from 10 to 80 mg L -1 and the results of total phenolic content were expressed as mg gallic acid equivalents (GAE) g -1 of dry weight (DW).
Antioxidant activity using 2,2-diphenyl-1-picrylhydrazyl hydrate (DPPH) free radical scavenging method DPPH free radical scavenging activity was measured according to the method described by Brand-Williams et al. 31 The reaction medium consisted of 0.5 mL of EEGP (0.5 g L -1 ) and EEAP (1.6 g L -1 ), 3 mL of ethanol and 0.3 mL of DPPH radical solution 0.5 mmol L -1 in ethanol.The mixture was incubated at room temperature in the dark for 45 min and the absorbance was measured using a spectrophotometer (model UV-Vis lambda 25, PerkinElmer) at 517 nm.Ethanolics solutions of Trolox (15 to 100 µmol L -1 ) were used for calibration, and the results were expressed as µmol of Trolox g -1 of dry weight (µmol Trolox g -1 ).The measurements were conducted in triplicate.
High-performance liquid chromatography with photodiode array detection (HPLC-PDA)
The high-performance liquid chromatography with photodiode array detection (HPLC-PDA) method was performed according to Oldoni et al. 28 using a liquid chromatography equipment (Varian 920-LC) coupled to an analytical column Varian C-18 RP (250 × 4.6 mm, 5 µm) and a detector of photodiode array (PDA).It was injected 10 µL of EEGP and EEAP at a concentration of 5 and 6 g L -1 , respectively.The mobile phase was composed of water:acetic acid (99:1, v/v) (solvent A) and methanol:acetic acid (99:1, v/v) (solvent B) in gradient mode at a constant flow of 1 mL min -1 .The gradient started with 5% solvent B to 7% of B in 7 minutes, 20% B in 15 minutes, 50% B in 30 minutes, 90% B in 50 minutes, and 95% B in 55 minutes.The column was maintained at a constant temperature of 30 o C.This work used authentic standards of vanillic acid, gallic acid, caffeic acid, p-coumaric acid, ferulic acid and trans-resveratrol stilbene.The identification was performed by comparison of absorption in ultraviolet at 280 and 320 nm wavelengths and retention times (Table 1).
Results and Discussion
Optimization by RSM was carried out to verify that the independent variables-temperature (X1), time (X2) and ethanol concentration (X3)-significantly influenced the process of extraction of phenolic compounds and antioxidant activity present in EEGP and EEAP.Influence was verified through the responses obtained in the analysis of total phenolic content (TPC), antioxidant activity by DPPH radical scavenger method and determination of phenolic compounds by HPLC for both extracts.
Based on the chromatographic analysis by HPLC, in EEGP it was identified and quantified gallic acid, caffeic acid, p-coumaric acid and trans-resveratrol and in EEAP caffeic acid (Figure 1).In both chromatograms some compounds were observed but not identified.Their ultraviolet (UV) spectra correspond to phenolic families analyzed as flavan-3-ol and dihydrochalcone (280 nm), hydrocinnamic acid (320 nm) and flavonol (350 nm). 32arhadi et al. 22 studied different cultivars, species and parts of grapes (seed, skin, peel and pulp) and identified other compounds with antioxidant capacities beyond those found in this work, particularly catechin, epicatechin, quercetin and rutin flavonoids.Tournour et al. 8 conducted comparative studies of different varieties of Portuguese grapes using different methods of extraction, and their studies have revealed the presence of gallic acid, caffeic acid, syringic acid, catechin and epicatechin, and the major compounds were syringic acid and catechin.
Leyva-Corral et al. 5 evaluated apple pomace and a product formulated from that and identified phenolic acids and flavonoids in both, especially caffeic acid.Among the major phenols identified in apple pomace, García et al. 33 described cinnamic acids.After that, RSM was used to evaluate the influence of independent variables on dependent variables and regression coefficients.It was possible to demonstrate the terms of the model, quadratic and linear adjustments, as well as models of significant interactions between the terms.RSM analysis is performed by combining two independent factors related to dependent variable.
Optimum extraction conditions for grape pomace
For caffeic acid, trans-resveratrol and antioxidant activity by DPPH method no variable or interaction was statistically significant.The values of antioxidant activity and phenolic compounds for EEGP varied from 102.16 to 307.2 µmol of Trolox g -1 and from 17.98 to 33.52 mg GAE g -1 of dry grape pomace, respectively.The runs 10 and 11, corresponding to the combination of intermediate levels of three input variables showed the highest responses for DPPH and TPC, respectively (Table 2).
The main effect, time of extraction had significant effect (p < 0.05) on the TPC (Table 3).The content decreased by 49.7% when going from 30 min, run R11 (33.52 mg GAE g -1 ) to 39 min of extraction, run R16 (17.98 mg EAG g -1 ).It was observed that the TPC of the extracts initially increased with increasing time until reaching a maximum, after which the TPC decreased with increasing time.It suggests that after the optimum extraction time occurred degradation of some thermolabile phenolic compounds, thus leading to a lower concentration of phenolic compounds.
The developed model can be considered statistically significant and predictive.F-value calculated was 7.05 times higher than F-tabulated with 95% of confidence (Table 3) for TPC.The models had satisfactory levels of adequacy (R 2 ).
The extraction of phenolic p-coumaric acid (CUA) was influenced by interactions between time versus temperature and ethanol concentration versus temperature (Figure 2 and Table 3).At EtOH concentration of 50%, the content of p-coumaric acid at 45 and 75 °C was 21.2 and 19.4 µg g -1 , runs R2 and R15, respectively (Table 2).In comparison, the p-coumaric acid of the extract set 45 and 75 °C using 80% EtOH was approximately 18.1 and 34.2 µg g -1 , runs R3 and R13, respectively.The samples extracted at 45 and 75 °C using 80% EtOH were significantly different.An increase in temperature from 45 to 75 °C caused a 47% increase in content of p-coumaric acid when extracting with 80% EtOH (Table 2 and Figure 2).Ethanol is less polar (1.69 dipole moment) than water (1.85 dipole moment) so by increasing the concentration of ethanol the polarity of the solvent extractor is decreased.The extraction of gallic acid has a maximum at around 65% of ethanol and p-coumaric acid at 80%.The difference between extractions of these compounds can be explained by the fact that gallic acid is a more polar compound than p-coumaric acid.
The interactions between the extraction time and temperature also influenced the extraction of phenolic p-coumaric acid.At an ethanol concentration of 80% and 75 °C, the content at 21 and 39 min was 34.2 and 17.7 µg g -1 , runs R13 and R16, respectively.In comparison, the p-coumaric acid of the extracts at 21 and 39 min using 45 °C was approximately 18.1 and 10.2 µg g -1 , runs R3 and R4, respectively.The samples extracted at 21 and 39 min using 45 and 75 °C were significantly different.An increase in time from 21 to 39 min caused a 48.4% decrease in content of p-coumaric acid when extracting with 75 °C and EtOH 80%.The maximum predicted CUA (47.13 µg g -1 ) was obtained under the extraction conditions of 50% of ethanol, 45 °C and 39 min for EtOH concentration, temperature and time, respectively (Table 2).
Two-way joining analysis (Figure 3) confirms the maximum conditions for extraction of phenolic compounds and antioxidant activity from EEGP.This analysis evaluates the clustering of samples related to dependent variables.The intensity of color shows which dependent variable is more related to extracts produced in different conditions.For EEGP, runs R2, R3, R4 and R16 showed lowest results for all variables analyzed.Runs R9, R13, R11 and R10 showed the highest content of phenolic compounds and gallic acid, indicating a strong correlation between these results.The run R5 showed the highest content of p-coumaric acid and trans-resveratrol.
Because three variables (p-coumaric acid, phenolic compounds and galic acid) were significant, presented distinct values and there is interest in obtaining the optimal operating condition of the independent variables, it was calculated the Desirability function.The calculated global value was 0.7784 for the conditions of 60 °C, 30 min and ethanol concentration of 65%, which corresponds to the central point, simultaneously satisfying all the requirements necessary for dependent variables.Dependent variables that were influenced by the evaluated parameters (X1, X2 Karacabey and Mazza 3 using RSM in grape pomace also observed the effect of ethanol concentration and temperature on the extraction of compounds with antioxidant activity and the ethanol concentration was the most significant variable for the extraction of bioactive compounds.Other authors have also reported that the polarity of the solvent used in the extraction directly affects not only the amount of phenolic compounds but also the composition of phenolic compounds. 34
Optimum extraction conditions for apple pomace
The variations observed for antioxidant activity and TPC obtained for the EEAP were 2.34 to 4.06 µmol Trolox g -1 and 5.04 to 8.01 mg GAE g -1 of dry apple pomace, respectively.
In extraction of phenolic compounds no independent variables were significant; however, the antioxidant activity by DPPH method was influenced by all tested variables (Table 3) with quadratic and linear effects.It was observed that when time and extraction temperature variables varied from lower level to higher level occurred an increase of 42.4% in antioxidant activity (Table 4).
The main effect, ethanol concentration was significant (p < 0.05) to extract compounds with antioxidant activity by DPPH method.The activity increased more than 40.0%when going from 80 (2.34 µmol g -1 ) to 50% (4.06 µmol g -1 ) EtOH concentration (Table 4).The effect of ethanol concentration can be related to decrease in polarity of solvent extraction due to the increase of ethanol concentration.Therefore a change in polarity of solvent coupled with temperature and time increase provided the maximum extraction of compounds with antioxidant activity from apple pomace which indicates the heat stability of these compounds.
In the extraction of caffeic acid only the time factor was not significant, while the independent variable temperature and ethanol concentration, as well as the interaction between ethanol concentration and temperature were significant (Table 3).It is possible to suggest that caffeic acid shows heat stability and a better affinity with medium less polar.The maximum predicted caffeic acid (61.83 µg g -1 ) evaluated by HPLC was obtained under extraction conditions of 50% of ethanol concentration and 75 °C (Table 4 and Figure 4).Two-way joining analysis for EEAP (Figure 5) showed that runs R3, R4 and R12 presented the lowest content of phenolic compounds and antioxidant activity while R9 extracted lower content of p-coumaric acid.
Wijngaard and Brunton 35 reported in apple pomace that only the temperature had a significant effect on the model response for antioxidant activity using DPPH method while concentration of ethanol plays the most important role to extract individual polyphenol groups as phenolic acids and flavonoids.
Using ANOVA, it was observed that models were significant and predictive.F calculated was 14.27 and 153.80 times higher than F tabulated for DPPH method and caffeic acid, respectively, with 95% confidence (Table 3).The models had satisfactory levels of adequacy (R 2 ).
For apple pomace, DPPH and caffeic acid variables were significant and from the Desirability function (0.7025), optimal extraction conditions were 60 °C, 30 min and ethanol concentration of 65%, corresponding to the central point, same condition obtained for grape pomace.The lack of fit calculated was not significant for DPPH model (p > 0.05) while for the caffeic acid model the value was less than 0.05, indicating that the model is inadequate.
Conclusions
Effects of time, temperature and ethanol concentration on antioxidant activity and phenolic compounds of grape and apple pomaces were determined by response surface methodology (RSM).The optimization procedure was conducted in order to simultaneously maximize the antioxidant activity by DPPH method, total phenolic From desirability function it was possible to establish that optimal conditions of extractions for both pomaces were the independent variables in central point.On the whole, it is interesting to note that grape and apple pomaces from Southern of Brazil are a valuable source of natural antioxidants and have properties that suggest applications in the food and pharmaceutical industries.
Figure 2 .
Figure 2. Responses surfaces obtained from EEGP: (a) p-coumaric acid as a function of ethanol concentration and time; (b) gallic acid as a function of time and ethanol concentration; and (c) TPC as a function of temperature and time.
Figure 4 .
Figure 4. Responses surfaces obtained from EEAP (a) caffeic acid as a function of temperature and ethanol concentration and (b) DPPH as a function of ethanol concentration and time.
Table 1 .
Chromatographic parameters of phenolic compounds analyzed by HPLC
Table 3 .
Analysis of variance (ANOVA), regression coefficients of the models, R 2 values and F-value of the optimized second order polynomial models of antioxidant activity as measured by DPPH, phenol content (TPC), level of gallic acid, level of p-coumaric acid and caffeic acid level for ethanolic extracts of grape (EEGP) and apple pomace (EEAP) a not significant p > 0.05; b significant at p < 0.01; c significant at < 0.001; d significant at p < 0.05; e F ratio (F-value/tabulated F-value); GA: gallic acid; CUA: p-coumaric acid; DPPH: antioxidant activity; CA: caffeic acid; DW: dry weight.
Table 4 .
Central composite rotatory design and corresponding response values for ethanolic extract from apple pomace (EEAP) | 2018-12-17T20:04:26.015Z | 2017-10-01T00:00:00.000 | {
"year": 2017,
"sha1": "4f5af7cd50d2c298ab361b764d34bf2c8a3aab4d",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.21577/0103-5053.20170014",
"oa_status": "GOLD",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "4f5af7cd50d2c298ab361b764d34bf2c8a3aab4d",
"s2fieldsofstudy": [
"Agricultural And Food Sciences"
],
"extfieldsofstudy": [
"Chemistry"
]
} |
21725652 | pes2o/s2orc | v3-fos-license | Nonsurgical treatment and stability of an adult with a severe anterior open-bite malocclusion
OBJECTIVES: A skeletal anterior open-bite is a challenging malocclusion for the orthodontist due to the difficulty and instability of correction. Treatment options for the adult patient include extractions, anterior extrusion with intermaxillary elastics, posterior intrusion using skeletal anchorage, occlusal adjustment, and orthognathic surgery. Patient compliance plays a key role in posttreatment stability. The present case report demonstrates the orthodontic treatment of an adult patient who presented with a complex open-bite malocclusion. MATERIALS AND METHODS: Treatment involved the placement of four miniscrews to assist intrusion of maxillary molars by applying posterior vertical maxillary elastics and extrusion of the anterior segments using anterior vertical interarch elastics. RESULTS: Ideal intercuspation was successfully achieved and good stability was maintained during 3 years following treatment. CONCLUSION: The intrusion of the maxillary molars with miniscrews is an interesting option in selected cases of skeletal anterior open bite. The retention protocol should be specific in these cases.
Introduction
A complex anterior open-bite malocclusion is usually caused by a combination of habits and skeletal, dental, or functional influences. Besides treatment difficulty, the long-term stability of the open-bite correction is considered challenging because of the great potential for vertical relapse regardless of the treatment modality or retention protocol. [1][2][3][4][5][6][7][8] In adult patients, the most common nonsurgical treatment option is the orthodontic extrusion of anterior teeth by the use of interarch elastics. In cases treated with tooth extractions the increase in the interincisal angle during space closure also aids in the correction of the open-bite due to the "drawbridge principle." [1] Orthognathic surgery is the likely treatment option when an open-bite is associated with severe skeletal disharmony. In selected cases; however, in which facial aesthetics is not compromised, the intrusion of the maxillary posterior teeth using skeletal anchorage can provide satisfactory occlusal results, and comparable stability to other open-bite treatment modalities which may involve tooth extractions, interarch elastics, and orthognathic surgery. [1,3,4] The aim of the present case report is to describe the nonsurgical orthodontic treatment of an adult patient who presented with a complex anterior open-bite malocclusion which was treated using a combination of skeletal anchorage, fixed appliances, and vertical intermaxillary elastics. The correction of the open-bite, achieved mostly by the intrusion of posterior teeth, was satisfactory and stable 3 years after the end of active treatment.
Etiology and diagnosis
A healthy 40-year-old Caucasian woman presented concerned about her compromised speech and masticatory efficiency as a result of a long-standing open-bite. The patient reported that she could not incise food with her anterior teeth and had difficulty in chewing. A facial analysis showed a symmetrical face, a straight profile, and lip competence. A smile analysis showed a decreased maxillary incisor display. The dental midlines were coincident with the midsagittal plane. Intraorally, she had a half-unit Class II molar relationship on the right side and a Class I relationship on the left side, with generalized maxillary and mandibular spacing. An anterior open-bite extended from first molar to first molar [ Figures 1 and 2]. Only the second and third molars were in occlusal contact [ Figures 2 and 3]. The overjet was 2.1 mm and the overbite was -4.1 mm. A functional evaluation revealed an adaptive static and dynamic forward tongue posture, which was evident during speech and swallowing. The patient's speech was distorted during the pronunciation of the sibilant sounds [s] and [z]. In addition, the tongue presented an increased size which approached macroglossia.
A panoramic radiograph showed no maxillomandibular abnormalities. Cephalometrically, there were good basal sagittal and vertical relationships and a balanced soft-tissue profile. The palatal plane was rotated counterclockwise [SN.PP = 0.5°]. The maxillary and mandibular incisors were proclined and protruded [ Table 1]. General dental treatment was required and completed prior to the commencement of orthodontic treatment. The volunteer signed a release form authorizing the use of his image in scientific research.
Treatment options
The main objective of orthodontic treatment was to eliminate the open-bite while achieving satisfactory smile aesthetics and masticatory function. To provide stability of the treatment results, the elimination of the abnormal tongue function was proposed during the active phase of treatment. Maintenance of the original facial aesthetics was planned because it was considered satisfactory and The first proposed treatment alternative was a combined orthodontic-surgical plan. The open-bite would be corrected by posterior maxillary impaction and the consequent mandibular auto-rotation would close the open bite. [3] A second treatment option was the extraction of the first molars or premolars followed by anterior retraction and mesialization of the posterior teeth during space closure. This would assist open-bite correction by to the uprighting and extrusion of the incisors and the counterclockwise rotation of the mandible. [9] Extrusion of the anterior and posterior teeth with intermaxillary elastics was considered as a third option to manage the open bite. This alternative was likely to have limited or no skeletal effect and would require a large amount of dental extrusion and patient compliance. [10] A fourth alternative involved the extraction of the third molars, the intrusion of the maxillary molars using skeletal anchorage, and the simultaneous extrusion of the anterior teeth with vertical intermaxillary elastics. [4,10,11,12] This option had the potential to provide some minimal positive facial changes at the same time that the open-bite was being corrected.
The patient did not accept surgery because of the impact it would have on her facial appearance. The second option was discarded because of the risk of exacerbating the patient's lip retrusion during the retraction of the anterior teeth. The third option was also eliminated because extensive extrusion of the anterior teeth would have a high susceptibility to relapse and could also produce excessive gingival display. Although, the fourth option presented the most challenging biomechanics, it was considered that it would provide the best skeletal and dental results.
Treatment progress
The crown fracture of the maxillary right central incisor was restored and all third molars were extracted before starting the orthodontic treatment. Figure 4] during the middle of which, the left buccal miniscrew needed to be repositioned more apically, because it approached the molars. The total intrusion force used was 250 g per side (right or left).
After 12 months of treatment and 8 months of molar intrusion, the maxillary and mandibular interdental spaces were closed by elastic chains acting from second molar to second molar. Elastic chains were initially used with rectangular and finally with 0.020-inch stainless-steel arch wire. Figure 5a shows the elastic chain in the mandibular arch, and the anterior teeth tied-together in the maxillary arch to avoid reopening of the spaces. The elastic chains were maintained throughout treatment, particularly in the mandibular arch to avoid reopening of the spaces due to the anterior tongue posture and function.
After 20 months of treatment, considerable reduction of the open-bite had occurred [ Figure 5a]. At this time, 0.018-inch stainless steel arch wires were inserted in the maxillary and mandibular arches, and bilateral vertical interarch elastics were added to aid in the closure of the remaining open-bite [ Figure 5b]. The vertical elastics were used 20 h/day for 12 months and the results can be observed in Figure 5b (open-bite closure). After this period, the intermaxillary elastics were maintained only overnight for 10 months, as an active retention period.
The fixed appliances were removed after 4.1 years of active treatment. Wrap-around retainers were provided for the maxillary arch and a 0.028-inch stainless-steel arch wire segment was bonded to the mandibular anterior teeth. Two maxillary appliances with posterior bite blocks to retain the posterior intrusion were constructed for day and night-time use [ Figure 6a]. The daytime retainer incorporated an orifice close to the incisive papillae which guided correct tongue position [ Figure 6b].The night-time retainer contained a palatal crib to prevent lingual pressure on the anterior teeth [ Figure 6c]. A loop was made in the mandibular Figure 7]. The maxillary anterior teeth had aesthetic restorations placed after orthodontic treatment was completed. The patient was also referred for a maxillary labial frenectomy to prevent reopening of the diastema between the maxillary central incisors. [13] A myofunctional evaluation was also undertaken to investigate and correct, if necessary, the resting tongue posture and function, to minimize the risks of open bite and anterior diastema relapse.
Treatment results
A harmonious facial balance, improved smile, and good interdigitation were achieved at the end of treatment. The intraoral photographs and cephalometric analysis showed that there was a 5.8 mm improvement in the overbite, from -4.1 mm to 1.7 mm [Figures 7-9 and Table 1]. There was improvement in the apical base relationship [ANB angle] from 2.7 o to 0.7 o , and an increase in mandibular protrusion [SNB angle] from 81.7 o to 83.4 o , as a result of the counterclockwise rotation. The maxillary molars were intruded by 1.9 mm and the maxillary incisors were extruded by 1.6 mm. The lower incisors remained stable. There was no obvious evidence of clinically-significant root resorption.
The patient was satisfied with the treatment results, and the outcome was stable 3 years after the end of active treatment [ Figures 10 and 11]. No return of the anterior open-bite relapse was observed. The dentoskeletal changes that occurred during treatment and after 3 years posttreatment are presented in the cephalometric superimposition [ Figure 12].
The maxillary removable retention was discontinued after a year but the mandibular retention was recommended for a lifetime. During the retention period, there was a tendency for the interproximal space of the maxillary central incisors to reopen. Therefore, a 0.020-inch stainless steel wire was bonded to the palatal surfaces of these teeth [ Figure 10]. There was space opening distal to the left canine in the mandibular arch. These changes may be explained by the patient's refusal of the upper labial frenectomy and the recommended myofunctional therapy after treatment. Some anterior tongue pressure likely remained which may have tipped the incisors forward and as evident by the increase in incisor mandibular plane angle (IMPA).
After treatment, the patient reported an improvement in function, especially in incising food. Clear improvement during speech pronunciation was also apparent. However, this was not confirmed by a speech therapist because the patient refused.
Discussion
In the present case, the patient's chief complaint was primarily functional, related to speech and mastication. Aesthetic-related complaints were minimal, but concerned the insufficient display of the maxillary anterior teeth upon smiling. The patient's balanced facial appearance and profile played an important role in the decision for a non-surgical treatment option consisting of pre-adjusted Roth prescription fixed appliances aided by skeletal anchorage. The correction of the anterior open-bite resulted mostly from a counterclockwise rotation of the mandible (caused by the extraction of the third molars and the intrusion of the maxillary posterior teeth) plus the extrusion and uprighting of the maxillary anterior teeth (due to the use of vertical intermaxillary elastics).
The amount of miniscrew molar intrusion was 1.9 mm and is similar to the average intrusion previously reported. [12,14] It is expected that the anterior open-bite may close at a rate of 3-4 mm for each millimeter of molar intrusion obtained. The 5.8 mm of overbite correction obtained was within this ratio (5.8/1.9 = 3.05) and confirms previous reports. [15] The correction of the dental Class II relationship on the right side was a result of the mandibular counterclockwise rotation, which caused a reduction of 2° in the maxillomandibular relationship (ANB angle). [4,12] Despite the dental Class II relationship correction, there was an increase in the initial overjet. This may be explained by the presence of larger diastemas in the mandibular arch at the beginning of treatment. Consequently, the Table 1]. This result is consistent with a recent study demonstrating that overjet decreases in only 65% of the cases when the mandible undergoes a counterclockwise rotation. [12] The stability of the open-bite correction was excellent at the 3-year posttreatment review, considering that 80% of relapse occurs in the first year after treatment. [4] Previous reports have recommended that careful attention is required during the retention phase because of the significant amount of tooth movement that can occur when miniscrews are used as part of the biomechanics. [11] For this reason, posterior occlusal coverage was included in the wrap-around maxillary retainer to minimize extrusion of the posterior teeth. [4,12] However, other appliances might also be necessary to modify tongue posture and function and to avoid an anterior resting tongue position at nighttime. [16] Myofunctional therapy has also been advocated as an adjunct in the reduction of open-bite relapse. [17][18][19] Several authors have reported that the risk of incisor protrusion is greater in cases in which there is altered tongue posture at rest, and it is smaller when there is only altered function (speech and swallowing). Therefore, speech therapy is indicated especially in cases in which altered tonguerest posture is present. [17] The presented patient refused myofunctional and speech therapy as recommended. Fortunately, there was little impact on overbite stability as only a small diastema opened between the left mandibular canine and first premolar and between the maxillary central incisors. The upper diastema was corrected and retained with a bonded segment of coaxial wire. Another possible factor that may have caused the opening of the midline diastema was that the patient did not accept the frenectomy of the upper lip as requested at the end of orthodontic treatment. [13] In addition, perhaps the maxillary retainer should have been used for a longer period of time. The clinical stability at 3 years posttreatment can be confirmed by the cephalometric superimposition, which shows a reduction of the skeletal vertical component due to the mandibular counterclockwise rotation, which increased the SNB angle and reduced the ANB beyond that which was achieved during orthodontic treatment [ Figure 12]. It may be speculated that the vertical control promoted by the occlusal coverage of the maxillary retainers contributed to these posttreatment changes. The reduction of the maxillary and mandibular incisor proclination achieved during treatment was partially lost at 3 years posttreatment, possibly due to tongue function and that full upper and lower arch control was not established on a long-term basis. These changes in incisor inclination during the treatment and retention periods also explain the reduction and increase of lip projection during these periods.
Conclusions
The present case report details the non-surgical orthodontic treatment of an adult patient with a complex anterior open-bite treated using a combination of skeletal anchorage for upper molar intrusion, fixed appliances, and vertical intermaxillary elastics. Molar intrusion might be effectively achieved by using miniscrews as anchorage in patients with an anterior open-bite. Therefore, this treatment modality can be considered an acceptable alternative to orthognathic surgery in selected cases in which the dental malocclusion-rather than facial aesthetics-is the primary patient concern. However, retainers to minimize extrusion of the posterior teeth should be used to improve stability. The outcome of this case was stable after 3 years, although a longer follow-up is necessary to assure long-term stability.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest. | 2018-05-21T22:38:44.947Z | 2018-01-01T00:00:00.000 | {
"year": 2018,
"sha1": "fdce323fd49e724d0cc088843505f4c241b16302",
"oa_license": "CCBYNCSA",
"oa_url": "https://doi.org/10.4103/jos.jos_69_17",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "96917cd8943ca73c975b64bb22b4d18227121e95",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
237940074 | pes2o/s2orc | v3-fos-license | Overcoming Conflicting Data when Updating a Neural Semantic Parser
In this paper, we explore how to use a small amount of new data to update a task-oriented semantic parsing model when the desired output for some examples has changed. When making updates in this way, one potential problem that arises is the presence of conflicting data, or out-of-date labels in the original training set. To evaluate the impact of this understudied problem, we propose an experimental setup for simulating changes to a neural semantic parser. We show that the presence of conflicting data greatly hinders learning of an update, then explore several methods to mitigate its effect. Our multi-task and data selection methods lead to large improvements in model accuracy compared to a naive data-mixing strategy, and our best method closes 86% of the accuracy gap between this baseline and an oracle upper bound.
Introduction
Most work in semantic parsing (and NLP in general) considers a scenario where the desired outputs of a model are static and can be specified with a large, fixed dataset. However, when deploying a semantic parsing model in a real world virtual assistant, it is often necessary to update a model to support new features or to enable improvements in downstream processing. For example, creators of a media assistant may want to add new commands specialized towards new media types like podcasts, or those of a navigation assistant might like to add a new feature to allow users to specify roads to avoid. Such changes require that the structures output by the model are updated, either by introducing a new intent (a representation of a desired action) or re-configuring arguments of existing intents. In this work, we investigate the best way to make updates to the intents and arguments of a task-oriented neural semantic parsing model. * Work performed during an internship at Google.
To make an update, new data annotations must be collected to specify the new form that is desired for model outputs. Because changes can be quite frequent, we would like to be able to collect a small amount of data for each update (on the order of 50 examples) and merge the new information with a much larger existing dataset. Naively, we might hope to simply combine new data in with the old and train on the combination. However, this approach has the problem that some of the older data with out-of-date labels may conflict with the new labeling. These conflicts occur whenever inputs that would be affected by a change appear in the original dataset. For example, when introducing a new intent for podcasts, the original dataset may have included podcast examples labeled with a more generic media intent or a label indicating that the feature is unsupported. When introducing a new argument, say 'roads to avoid', there may be instances in the original dataset that should have this argument labeled but do not because they were annotated before the argument was introduced. This conflicting data can confuse the model and cause it to predict labels as they were before an update rather than after. Unfortunately, this problem of conflicting data during model updates is understudied in the academic literature.
To enable exploration of this problem on publicly available datasets, we propose a method to easily create simulated updates with conflicts (Section 3.3) and release our update data. The idea behind our method is to form updates in the reverse direction, relabeling instances of a particular intent or argument to simulate out-of-date labels. Using our proposed setup, we demonstrate how conflicting data greatly hinders learning of updates (Section 5).
In addition, we explore several methods for mitigating the negative effects of conflicting data by modifying how we combine the old and the new data (Section 6). One approach is to keep the old and new datasets separate, but to share information indirectly with either fine-tuning or multi-task learning. Another approach is to explicitly filter data that is likely to conflict using a learned classifier. Each of these methods substantially improves performance compared to naively mixing in new data, establishing strong baselines for future work on this task.
In summary, the contributions of this work are 1) establishing an experimental setup to test updates with conflicting data, 2) demonstrating that conflicting data leads to large losses in model performance if left unmitigated, and 3) exploring several possible approaches to mitigate this problem.
Related Work
There has been a substantial amount of prior work on making updates to neural models (Xiao et al., 2014;Rusu et al., 2016;Li and Hoiem, 2016;Kirkpatrick et al., 2017;Castro et al., 2018), demonstrating a recognition in the community that the ability to update models is important. However, most of these works consider a setting where none of the original data conflicts with the new labels. Thus these works, and the general continual learning and class-incremental learning literature, assume that the space of inputs affected by a change does not appear at all in the original data. In many scenarios, this assumption does not hold because the original dataset will aim to cover the full distribution of inputs a model might encounter. Because of the non-conflicting assumption, this body of prior work focuses on other questions such as what can be done when the original data is no longer available.
One paper that does consider updates with label conflicts is Chen and Moschitti (2019). Although they do not intentionally set out to study conflictingdata updates, in their NER task locations where new labels apply are tagged with special "outside" labels prior to the update, which cause conflicts with the new labels. While their work avoids the conflicting data problem by considering a setting where the original data is no longer available, our experiments show that it can be advantageous to instead keep the original data around and use more direct methods to avoid problems of conflicting data.
Our work also has parallels to the concept drift literature (Tsymbal, 2004;Lu et al., 2018). However, concept drift work focuses on unintentional changes to a natural data distribution over time, while our work is concerned with intentional restructuring of the output representation space, leading to very a different setup and choice of methods. In particular, that work operates over a stream of examples where changes often occur gradually, does not generally include structured prediction tasks, and does not allow the practitioner to introduce additional structure on the types of annotation given (as we do in Section 3.2).
Finally, our work relates to work on training with noisy labels (Sukhbaatar et al., 2014;Veit et al., 2017;Jiang et al., 2018), since the incorrect labels from conflicting data could be viewed as a type of noise. However, it is important to evaluate the problem of conflict-causing updates separately from other noisy label problems because the distribution of incorrect labels due to an update will be very different from most other sources of label noise. While not all noisy-label methods can directly apply to our task (as many are designed for classification as opposed to structured prediction) and may not take full advantage of the additional structure of our problem, we believe this line of work can still serve as a source of inspiration for future exploration on our task.
Preliminaries -Task-oriented Semantic Parsing
The experiments in this paper focus on a semantic parsing task where the goal is to generate a tree structure conditioned on an input sentence. We use the task formulation and data from Gupta et al. (2018) as the basis of our setup. Output trees are made up of intents and arguments (aka slots), where intents come from a fixed inventory of labels, and arguments consist of an argument-type label along with a value. Argument values may either be freeform text selected from the input sentence, or a nested intent to form a hierarchical structure. In this work, we represent these trees with a linearized form using nested brackets, which allows for the use of standard sequence-to-sequence models (see Section 4 for details of our base model and Figure 1 below for some example inputs and outputs).
Data Updates
In this paper, we focus on the task of making a single update to the intent and argument structure output by a model, where the update is specified by collecting a small amount of additional data. 1 Accordingly, our task setup expects two sets of data: a large amount of data from before a change, which we will call the V1 (version one) set, and a small amount of data from after a change, which we call the V2 set. The V1 set represents the current state of the system with any data collected in the past, while the new V2 set is collected specifically for the purpose of introducing a particular update. For the purposes of testing methods in this paper, we will form these two sets synthetically, as described in Section 3.3 below.
Because the V2 set is gathered explicitly for the purpose of introducing a particular update, we would like the input distribution of this data to be targeted rather than uniformly covering the full input space. Ideally, a substantial portion of this data should be examples whose labels will actually change after the update. We will call this portion of data the changed set. However, it is also useful to specify some examples that are not affected by a change, so that we can accurately determine the scope of a change. We will call inputs whose label would be the same under the new and old label scheme unchanged examples. We will include a set of unchanged examples in the V2 data to show that these labels have been confirmed under the V2 labeling scheme.
Many types of changes that we care about only apply to examples labeled in a particular way in the original V1 data. For example, a new argument can often only be used for specific intents, or a new intent may only apply to examples previously labeled as unsupported. By taking advantage of this information, we may be able to avoid some unwanted side effects of our model updates. To help us handle this information, we define a third category of examples: trivially-unchanged. The trivially-unchanged partition contains all examples which we can determine to be unaffected based solely on the original V1 labels and some simple hand-defined rule like a list of affected intents. By identifying these examples, we can directly include them in the updated training set without causing label conflicts. In the remainder of this paper, we reserve the term unchanged to refer specifically to unchanged examples that do not fall into the trivially-unchanged category. Thus, the unchanged partition represents the remaining hard examples that are difficult to distinguish from the changed set in the V1 data. See Figure 1 for examples of how our three data partitions apply to particular updates.
In some instances, it is also useful to talk about the data in the original V1 set in terms of the three partitions (changed, unchanged, and triviallyunchanged). In this case, these labels refer to whether an example would change if we had gathered new labels for them. In actuality, the changed subset will have out-of-date labels in V1, and we call these examples conflicting data. Our experiments show that conflicting data causes substantial problems when learning an update (see Section 5 and 7), but unfortunately, the examples that make up the conflicting set cannot be easily identified in the V1 data.
Creating Synthetic Changes
To enable exploration of the conflicting data problem, we demonstrate a method for easily simulating updates with conflicts on an existing publicly available dataset. We form our synthetic updates in the reverse direction -the data from the original dataset represents the final V2 form, and we modify some examples to represent a V1 (pre-update) form. We aim to make changes that can be done automatically in the reverse direction, while still being interesting to learn in the forward direction.
To form our updates, we select a particular intent or argument from the dataset to simulate the introduction of. We sample a set of examples labeled with the selected intent or argument to form the primary part of the V2 training set. For other examples with this intent or argument, we will keep them as part of the V1 training set, but re-label them to some form they may have taken before the new intent or argument was introduced. These re-labeled examples act as conflicting data.
For example, suppose we would like to simulate the creation of the intent GET_INFO_TRAFFIC, as shown at the top of Figure 1. After selecting a subset of GET_INFO_TRAFFIC examples for the V2 training set, we form a conflicting set in the V1 data by re-labeling GET_INFO_TRAFFIC examples to use the related intent GET_INFO_ROAD_CONDITION. For this particular change, we kept the arguments the same when updating, but for others we remove Examples not labeled with one of the seven intents that allow the SL:OBSTRUCTION slot are trivially-unchanged. Figure 1: Examples from different types of updates that we simulate. To simulate updates, we use the original dataset as the V2 form, and modify some examples to a V1 form in a way that is not easily reversed. We describe our data in terms of three partitions: changed, unchanged, and trivially-unchanged -as described in Section 3. In total, we form five different synthetic updates to use in our experiments. 2 Figure 1 also shows examples for two other types of updates: introducing a new intent for previously unsupported inputs and introducing a new argument. The updates are formed from the TOP dataset, which contains over 40,000 English queries about navigation or events labeled with tree-based semantic parses (Gupta et al., 2018).
Base Model
The semantic parsing model we use for our experiments is a sequence-to-sequence model based on the transformer architecture (Vaswani et al., 2017). We use a sequence-to-sequence model because of their flexibility and widespread use in semantic parsing (Jia and Liang, 2016;Dong and Lapata, 2016;Rongali et al., 2020) and NLP in general. Our model encodes a language input using a pretrained 12-layer BERT model (Devlin et al., 2019), then decodes a parse tree flattened by depth-first traversal. At each step, the decoder can generate either 1) a labeled bracket representing an intent or argument label 2) a closing bracket or 3) an index of an input token to be copied. The hyperparameters of our model architecture and training can be found in Appendix A.
2 Data for our synthesized updates can be downloaded from https://github.com/google/ overcoming-conflicting-data/, and Appendix B summarizes the partition sizes for each of these changes. Figure 2: Accuracy as a function of data size with conflicting data compared to accuracy when an oracle removes the conflicts, averaged across five different updates.
Effect of Conflicting Data
Before we describe and test our methods for mitigating conflicts, this section will briefly explore how conflicting data affects learning. We evaluate model updates both with and without conflicting data, and compare accuracies as we vary the amount of new V2 data being introduced. The non-conflicting setting represents an oracle where all conflicting data is removed, which we can easily simulate in our synthetic data-creation process but is generally not achievable on real-world data without additional manual annotation.
More precisely, when evaluating updates with conflicting data, we include 50 examples with outof-date labels in the changed category. We mix these examples with a full set of unchanged and trivially-unchanged examples to represent a V1 training set from before an update. We then introduce different amounts of changed examples with updated labels to act as the V2 training set, with sizes ranging from from half the conflicting set size (25) to four times as much (200). For each data size, we measure accuracy using an exact match metric, meaning the entire tree output by the model must match the reference to be considered correct. For the non-conflicting setting, we do not include the 50 examples with out-of-date labels, but otherwise use the same setup.
The results for this experiment are shown in Figure 2, after averaging over five different changes (a detailed breakdown of results can be found in Appendix C). To the left of the graph, we see that when the amount of conflicting data is greater than the size of the new data being added, we get less than half of the accuracy we would get without conflicting data. While the gap narrows somewhat with more data, even when we introduce four times as much new data as there are conflicting examples, the presence of conflicting data still leads to a loss of over 10% accuracy. These results show just how detrimental conflicting data can be to the learning of a model update.
Mitigation Methods
In this work, we consider three methods to alleviate the problems caused by conflicting data: finetuning, multi-task learning with separate decoder heads, and data filtering with a learned classifier.
Fine-tuning
Our first and simplest method for handling conflicting data is fine-tuning. For this approach, we first train a model on only the V1 training data, then after training completes, we take the final parameters and use those as initialization for training with the V2 data. This approach can alleviate the conflicting data problem because the second stage of training does not include any conflicting data and the model will have less confusion about how to label changed examples. By first training on the V1 data, we are also able to benefit indirectly from the larger amount of data contained in it.
During the second stage of training, we train on the changed and unchanged data in the V2 training set, as well as the trivially-unchanged examples from the V1 training set. Trivially-unchanged data from the V1 data can be included in V2 training because it can be easily identified and is known to not conflict.
Multi-task
For our next method, we use an approach from multi-task learning where multiple decoder heads are used for different sets of data (Caruana, 1997;Fan et al., 2017). Our two "tasks" correspond to the V1 data and the V2 data, and we use a separate set of parameters for the final pre-softmax layer of the decoder for each of the versions (as illustrated in Figure 3). The V1 head is only trained with V1 data and the V2 head is only trained with V2 data, but the encoder layers and decoder transformer are shared between both. This way, the V2 head is never trained on conflicting examples from V1, but the overall V2 model can still benefit from Figure 3: In our multi-task method, we feed the new V2 data to a separate decoder head to separate it from the possibly-conflicting data in the original V1 training set. The encoder layers and decoder transformer layers are shared between V1 and V2. some information in V1 data indirectly through the shared encoder and decoder transformer layers. After training, the V1 head can be discarded, and the V2 head is used to make decisions. We train on both versions simultaneously, with each batch containing some of both types of data. As with our other methods, the V2 head is also trained on trivially-unchanged data from the V1 data.
The goal behind this approach is similar to that of fine-tuning: to avoid training the V2 model directly on the possibly-conflicting V1 data while still sharing some amount of information through model parameters. However, unlike fine-tuning, which is liable to forget information from V1 as training progresses, the simultaneous training for the multi-task method keeps the V1 information active for better sharing.
Classifier-based Data Selection
The final method we explore in this work is classifier-based data selection. The idea behind the data selection strategy is to explicitly select examples from the original V1 data that we don't think will conflict. Using the small amount of V2 data, we train a classifier to predict whether an example will be changed or unchanged, and then apply this classifier to the V1 data, as illustrated in Figure 4. We can then include the selected examples in our updated training set, allowing us to take advantage of more information from the original training set while filtering out many of the problematic conflicting examples.
We first train a classifier on the V2 data to learn a binary decision between changed and unchanged examples. This training requires that we can distinguish between which examples are changed and unchanged, which can either be specified as part of the annotation process, or can be estimated by running an existing V1 model to predict old labels for the provided V2 examples (in this work we use the annotation method, as part of our synthetic data-creation process). Our classifier uses the same BERT encoder as our sequence-to-sequence parsing model and is initialized with parameters from a parsing model for the V1 data. Representations of this encoder are averaged across time before feeding into a small feedforward network with a hidden dimension of 512. After training on the V2 set, the binary classifier is run on the V1 training set (excluding the trivially-unchanged examples, which can be automatically included as-is for V2 training). This creates a categorization of predictedchanged and predicted-unchanged, which we hope will closely approximate the true changed and unchanged sets. For many of the changes we tested, the classifiers performed quite well, with accuracies above 90% on held-out data. For examples in predicted-unchanged, we will include them in the set of training examples used to train our updated model. For examples in predicted-changed, we do not want to directly include them, and consider two possible solutions: 1) remove them completely or 2) include the examples with an intent-only loss, as described below.
Intent-only Loss
The idea of the intent-only loss variant of the data selection method is to try to take advantage of more information about the predicted-changed examples without requiring us to know the full form of the tree after a change. While we know that these examples are likely to have changed, in general we do not know if or how the full argument structure will change for all examples. However, it is usually possible to know what the intent should be for the
Evaluation
In this section, we describe the evaluation of our methods for mitigating the conflicting data problem. We report results in terms of exact match accuracy between the predicted and target parse structure. Results are aggregated across 5 different updates, giving us a total of 1500 test examples for each method (5 updates × 3 partitions × 100 examples). We also average across 5 different runs for each method to reduce variance.
We compare against three baselines: training only on the original V1 data, training only on the new V2 data, and directly mixing the two data versions together into a single dataset. 3 We also compare to an upper bound where the entire V1 training set is re-annotated with updated labels. For many updates, this upper bound requires thousands of new annotations, as compared to the one hundred labels used by our methods.
Results
The results of our evaluation are shown in Table 1. All of our methods substantially outperform the baselines. Our best method, the selection classifier with an intent-only loss on changed examples ( § 6.3), obtains an accuracy of 71.5%, covering 86% of the gap between the best baseline and the oracle upper bound.
To see a more clear picture of what is happening, we also break down results by data partitions, as shown in Figure 5 (an even more detailed breakdown across different updates is provided in Appendix D). In this chart, we can see that the baseline that mixes the V1 and V2 data without accounting for conflicts performs extremely poorly on the changed examples, echoing our results in Section 5. On the other hand, training only on the small set of V2 data throws away all the information from the original V1 training set, limiting its performance (particularly on unchanged examples). Our methods provide an effective way to combine information in both datasets without overwhelming changed data with out-of-date labels.
Conclusion
This work has shown that in order to make effective updates to the outputs of a neural semantic parsing model by adding new data, it is important to consider the effect of conflicting examples in the original data. Conflicting data is likely a problem in many different scenarios where outputs to a model must be updated, and we believe that further study into methods for mitigating the effects of conflicting data is an important direction to allow practitioners to handle the constantly-changing needs of real-world machine learning applications.
D Main Results Detail
The following table breaks down our main results across the different updates tested. These results are described in Section 7 and summarized in Figure 5. | 2020-10-27T01:01:16.955Z | 2020-10-23T00:00:00.000 | {
"year": 2021,
"sha1": "599c80dac054f835ff08db894c77392a2f93da2f",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "ACL",
"pdf_hash": "482f6d59c37b4116b173635a696e6be8d6e943a0",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
226846016 | pes2o/s2orc | v3-fos-license | Clinical and Molecular-Genetic Insights into the Role of Oxidative Stress in Diabetic Retinopathy: Antioxidant Strategies and Future Avenues
Reactive oxygen species (ROS) overproduction and ROS-signaling pathways activation attack the eyes. We evaluated the oxidative stress (OS) and the effects of a daily, core nutritional supplement regimen containing antioxidants and omega 3 fatty acids (A/ω3) in type 2 diabetics (T2DM). A case-control study was carried out in 480 participants [287 T2DM patients with (+)/without (−) diabetic retinopathy (DR) and 193 healthy controls (CG)], randomly assigned to a daily pill of A/ω3. Periodic evaluation through 38 months allowed to outline patient characteristics, DR features, and classic/OS blood parameters. Statistics were performed by the SPSS 24.0 program. Diabetics displayed significantly higher circulating pro-oxidants (p = 0.001) and lower antioxidants (p = 0.0001) than the controls. Significantly higher plasma malondialdehyde/thiobarbituric acid reactive substances (MDA/TBARS; p = 0.006) and lower plasma total antioxidant capacity (TAC; p = 0.042) and vitamin C (0.020) was found in T2DM + DR versus T2DM-DR. The differential expression profile of solute carrier family 23 member 2 (SLC23A2) gene was seen in diabetics versus the CG (p = 0.001), and in T2DM + DR versus T2DM − DR (p < 0.05). The A/ω3 regime significantly reduced the pro-oxidants (p < 0.05) and augmented the antioxidants (p < 0.05). This follow-up study supports that a regular A/ω3 supplementation reduces the oxidative load and may serve as a dietary prophylaxis/adjunctive intervention for patients at risk of diabetic blindness.
Introduction
Diabetes mellitus (DM) is a multifactorial-polygenic disease characterized by chronic hyperglycemia and altered metabolism of carbohydrates, lipids, and proteins. There are two major clinical forms: Insulin-dependent DM (type 1: T1DM) and non-insulin-dependent DM (type 2: T2DM), being the latter the most common presentation, accounting for 90% of all diabetics [1]. DM displays alarming growth rates worldwide, being a serious health concern with an estimated number of affected people of 600 million by the year 2030 [2,3].
Microvascular changes are clinically manifested during the natural course of DM. Population studies reported that 20% of diabetics had retinopathy (DR) at the time of diagnosis [4]. The main pathogenic mechanisms in the diabetic retina include the activation of oxidative stress (OS), advanced glycation end-product generation, inflammation, activation of protein kinase C, polyol, and hexosamine pathways, etc. [5]. In addition, a series of endogenous/exogenous risk factors have been pointed as contributing highly to the DR initiation and progression [6].
The incomplete reduction of molecular oxygen (O 2 ) leads to the formation of reactive oxygen species (ROS) with unpaired electrons in its outermost orbital, making these molecules extremely unstable and reactive. The generation of superoxide anion (O 2 − ) and hydrogen peroxide (H 2 O 2 ) is prone to start the free radical chain reaction due to hydroxyl radical formation (OH − ), by the Fenton and Haber Weiss chemical reactions [7,8]. Furthermore, nitric oxide (NO) is a non-reactive free radical implicated in a variety of biological functions, such as vasodilation, immune response, neurotransmission, and apoptosis [9,10]. It is well known that redox status influences the regulation of intracellular signaling with enormous repercussions in health and disease [11]. In this context, epidemiologic and experimental studies demonstrated that DR is noticeably related to OS [5,[12][13][14][15].
The concern of nutritional supplements in ocular health and disease has been largely addressed [16,17]. Nutritional control has been proposed as a potential intervention for slowing DR initiation/progression, but studies on dietary assessments in diabetics showed contradictory results [18][19][20][21][22][23]. Moreover, nutraceuticals cannot replace eating adequate amounts of a high variety of foods, but more than half of Americans and many Europeans take at least one dietary supplement on a regular basis [16,17]. These may be especially beneficial to individuals with poor nutrition caused by a variety of processes, including DM. International science-based regulatory considerations in nutrition and nutraceutical supplements are needed to update its recommendations for the ophthalmological population [18][19][20][21][22][23]. Antioxidant (A) vitamins (vit) are essential for a wide variety of biological functions. Among them, the reduced form of ascorbic acid (AscH2; L-ascorbic acid; vit C), a hydrophilic vitamin, enzymatic cofactor, and ROS scavenger, is involved in the protection against a variety of disorders.
Abnormal dietary intake and/or plasma levels of vit C (deficiency defined as a concentration equal of less than 35 µmol/L) have been associated to a higher risk of getting sick, including the development of xerosis and other ophthalmic pathologies [24]. In this concern, vit C absorption and distribution to tissues and organs is achieved by active transport across the cell membranes. Two specific protein transporters encoded by the sodium-dependent vit C transporter genes, SLC23A1 and SLC23A2 [25], are involved in this fact. The latter is responsible for the regulation of the intracellular levels of vit C to protect cells from OS, as well as to promote the type I collagen maturation. In fact, genetic variations in the SLC23A2 gene expression modify the vit C availability [24,25].
Given that early diagnosis can delay or prevent DR-related blindness, there is considerable interest in exploring the role of its molecular diagnostics [17,18]. Up to today, studies on DR biomarkers have dealt with pathogenic processes and risk factors involved in DR development/progression [2,3,5,26,27]. The need for more accurate diagnostics that allow advance ophthalmologic and vision care is currently providing a critical window of opportunity for diabetics. The present study deals with improving knowledge about the classical and new potential diagnostic/prognostic candidate biomarkers to verify/validate them by using various endpoints like ophthalmic indicators, classical blood biochemistry, oxidative and antioxidant parameters, as well as the SLC23A2 gene. The secondary endpoint was to analyze the effect of a course of A and omega 3 fatty acids (A/ω3) in the form of mono-combination on the OS presumptive biomarkers, in order to better managing the diabetic eyes.
Community-Based Study Design
The study adheres to the Declaration of Helsinki (Edinburgh, 2000), the Ethics Committee standards of the study centers (no. 12/33), and the Spanish Agency of Medicines and Medical Devices (AEMPS). All requirements for clinical research and to maintain the privacy of the data obtained were met. A prospective interventional open-label multicenter case-control study was conducted between 2014-2015/2018-2019. Part of the study design and sampling methodology has been described in detail [19]. The main purpose was outlining T2DM patient characteristics, risk factors, precise blood-based biochemical parameters, and typical DR features to identify novel clinical and molecular-genetic biomarkers for DR. The second objective was to look at the effects of a course of (A/ω3) in the health and vision care among T2DM patients with (+) and without (−) DR.
Participants
575 potential participants of both sexes aged 25-80 years [300 T2DM patients and 275 healthy controls (CG)] were carefully recruited by general practitioners and ophthalmologists (listed at the end of the manuscript) to ensure consistency and efficacy of the study sample, interviewed according to the study inclusion/exclusion criteria (Table 1) and appointed to the first selection step.
Selection and Appointment Schedules
Baseline data were recorded as: Socio-demographics, personal/familial background (DM duration and treatment) and personal characteristics [nutrition-mediterranean diet (MedDiet) adherence by the validated 14-items questionnaire; Prevención con Dieta Mediterránea (PREDIMED), as previously reported [19], body mass index (BMI) and regular physical activity classified as to practice (1) or no (2) exercise. At the first visit, patients underwent an ophthalmological examination in both eyes: best-corrected visual acuity (BCVA), intraocular pressure (IOP), ocular fundus examination/retinography, and spectral domain optical coherence tomography (SD-OCT). Provided written informed consent before starting any related activities.
Participants able to attend the visits and to follow the study guidelines during the study period.
EXCLUSION
Males and females, aged < 25 years and > 80 years.
Patients with proliferative diabetic retinopathy, diabetic macular edema, or ocular or systemic diseases or aggressive treatments. Previous ocular surgery or laser for 12 months (at least). Other oral supplements with antioxidants and/or omega 3 fatty acids, including vitamins in eyedrops.
No acceptance for the study participation and/or not signing the informed consent. Unable to attend the visits or to follow the study guidelines during the study period. Of the initial recruitment, 480 suitable participants were verbally informed and signed on the consent form. The participants were then classified into: T2DM patients (n = 287) and CG (n = 193) ( Figure 1). Diabetics were also sub-classified according to having retinopathy "+DR" (n = 129) or not "−DR" (n = 158). (Figure 1). Antioxidants 2020, 9, x FOR PEER REVIEW 4 of 20 Of the initial recruitment, 480 suitable participants were verbally informed and signed on the consent form. The participants were then classified into: T2DM patients (n = 287) and CG (n = 193) ( Figure 1). Diabetics were also sub-classified according to having retinopathy "+DR" (n = 129) or not "−DR" (n = 158). (Figure 1). Diabetics and controls were homogeneously divided into those randomly assigned to "a pill of A/ω3 per day: +A/ω3" or "no-pill condition: −A/ω3", as reflected in the Figure 1 Diabetics and controls were homogeneously divided into those randomly assigned to "a pill of A/ω3 per day: +A/ω3" or "no-pill condition: −A/ω3", as reflected in the Figure 1. Pills containing vit (E, C, B1, B2, B3, B6, B9, B12), lutein, zeaxanthin, glutathione (GSH), hydroxytyrosol, trace elements (Se, Mn, Zn, Cu) and docosahexaenoic acid (DHA), commercialized as Nutrof Omega ® in packs of 35 capsules of 33.49 g each (Laboratorios Thea SA, Barcelona, Spain) were gently provided by the manufacturer and given to the participants without any charge ( Figure 2). Ophthalmic visits were scheduled every 6 months throughout the 38-months follow-up. Patients with proliferative DR (PDR) or diabetic macular edema (DME) were excluded by its special vulnerability to severe complications and the need to receive specific treatments immediately.
Antioxidants 2020, 9, x FOR PEER REVIEW 5 of 20 capsules of 33.49 g each (Laboratorios Thea SA, Barcelona, Spain) were gently provided by the manufacturer and given to the participants without any charge ( Figure 2). Ophthalmic visits were scheduled every 6 months throughout the 38-months follow-up. Patients with proliferative DR (PDR) or diabetic macular edema (DME) were excluded by its special vulnerability to severe complications and the need to receive specific treatments immediately.
Ophthalmologic Procedures
Ocular examination in each eye included: BCVA, ocular fundus inspection/photographs (ImageNet; Topcon, Barcelona, Spain), and SD-OCT (Carl Zeiss Meditec, Madrid, Spain). OCT parameters were measured by the direct cross-sectional imaging device. Retinographies and SD-OCT data were evaluated by retina specialists to obtain optimal nonproliferative DR (NPDR) diagnosisgrading in each eye, according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) [28] severity scale. The presence and number of microaneurysms/hemorrhages, venous beading, and intraretinal microvascular abnormalities were considered for DR diagnosis/progression. Worsening was considered an augmentation of the above signs. As reflected in Table 1, those patients with new vessel growth on the retinal surface (PDR) and the presence of focal-diffuse DME were also excluded from final data processing.
Ophthalmologic Procedures
Ocular examination in each eye included: BCVA, ocular fundus inspection/photographs (ImageNet; Topcon, Barcelona, Spain), and SD-OCT (Carl Zeiss Meditec, Madrid, Spain). OCT parameters were measured by the direct cross-sectional imaging device. Retinographies and SD-OCT data were evaluated by retina specialists to obtain optimal nonproliferative DR (NPDR) diagnosis-grading in each eye, according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) [28] severity scale. The presence and number of microaneurysms/hemorrhages, venous beading, and intraretinal microvascular abnormalities were considered for DR diagnosis/progression. Worsening was considered an augmentation of the above signs. As reflected in Table 1, those patients with new vessel growth on the retinal surface (PDR) and the presence of focal-diffuse DME were also excluded from final data processing.
Biosample Processing
Blood sampling was scheduled from baseline and every 6 months (during the 38-month follow-up) to biochemical and molecular-genetic probes. Antecubital vein blood was collected into 4.5 mL ethylene-diamine-tetra-acetic acid (EDTA) or sodium citrate vacutainer tubes (Becton Dickinson, Auckland, New Zealand), as an anticoagulant, under fasting conditions at 8:00 a.m. One EDTA tube (purple cap) was used to determine glucose, glycosylated hemoglobin (HbA1c), and plasma vit C concentrations. The other EDTA tube was used to analyze gene expression. Finally, the sodium citrate tube (purple cap) was centrifuged at 3000 rpm/10 min to obtain the plasma fraction, which was aliquoted and stored at −80 • C until processing. This plasma was used to analyze malondialdehyde (MDA), total antioxidant capacity (TAC), and total glutathione (GSH). All experiments were performed in duplicate. The determinations of glucose and HbA1c were carried out in the Department of Clinical Analysis of the University Hospital Dr. Peset (Valencia, Spain). To complete the analytical proceedings, the remaining experiments were performed in the laboratories of the Ophthalmic Research Unit "Santiago Grisolía" (Valencia, Spain).
-Determination of lipid peroxidation by-products. MDA/thiobarbituric acid reactive substances (TBARS): MDA/TBARS. It was assayed at high temperature (90-100 • C) under acidic conditions and extracted with butanol. Fluorescence was measured in duplicate at 544 nm excitation, 590 nm emission in relation to standard samples fluorescence. The concentration was calculated by extrapolating all data in the standard curve, as reported [19,29].
-Determination of total GSH. A modification was done [19] of the method, firstly reported by Tietze [31]. The OxiSelectTM Total GSH (GSSG/GSH) kit was utilized (Cell Biolabs, INC, Ref: STA-312. Madrid, Spain). Global thiol reagent, 5-5 -dithiobis [2-nitrobenzoic acid] (DTNB) reacts with GSH to form both the 412 nm chromophore, 5-thionitrobenzoic acid (TNB), and the disulfide product (GS-TNB). The GS-TNB was reduced through an enzymatic reaction catalyzed by the GSH reductase and β-nicotinamide adenine dinucleotide phosphate (NADPH). Thus, a second TNB molecule was released by recycling the GSH. Any oxidized GSH (GSSG) initially presented in the reaction mixture or formed from the mixed disulfide reaction of GSH with GS-TNB is reduced to GSH, and measured, as described [19,31].
-Determination of the glycemic profile [fasting glucose and HbA1c were performed by 2 different automated chemistry analyzers in the Department of Clinical Analysis of the main study center, as follows: (1) Abbott kits manufactured for use with the Architect c8000 (Abbott Laboratories; Abbott Park, IL, USA) and (2) Arkray AU 4050 (Arkray Global Bunisess Inc., Kyoto, Japan), respectively.
-Determination of plasma vitamin C (vit C). Fresh frozen plasma aliquots were thawed and acidified by adding perchloric acid and diethylenetriaminepentaacetic acid (DTPA; a strong trace metal chelator) for avoiding the ascorbate to destabilize. After centrifugation, supernatants were treated with Tris (2-carboxyethyl) phosphine (TCEP; a potent thiol-free reducing agent) to gather any rest of the ascorbate that became oxidized during previous proceedings. The concentration of vit C was determined by high-performance liquid chromatography (HPLC) with electrochemical detection, as previously reported [32,33]. Briefly, a Shimadzu HPLC System (Shimadzu Scientific Instruments, Columbia, MD, USA) that was equipped with a 5 µM YMCPack ODS-AQ column (Waters Corp., Milford, MA, USA) and a Coulochem III electrochemical detector (ESA, Chelmsford, MA, USA), under reversed-phase conditions was used. Sampling injection volume was 5 µL, and compounds were eluted over an 18 min runtime at a flow rate of 0.6 mL/min, following the method described by Li [32] with minor modifications [33]. Plasma vit C concentrations were expressed as mean (SD) in µmol/mL, taking into consideration the Linus Pauling Institute recommendations (https://lpi.oregonstate.edu/ mic/vitamins/vitamin-C) of consuming sufficient vit C to obtain at least a circulating concentration of 60 µmol/L, as well as international guidelines as follows: <11 µmol/L indicate severe deficiency; 23-50 µmol/L inadequate; 50-70 µmol/L adequate; and >70 µmol/L is deemed saturating.
-Gene expression assays. Whole blood samples were obtained from each participant and collected into EDTA tubes. Total RNA was isolated from blood samples by the Trizol method. Then, 300 ng of total RNA (integrity number-RIN > 7) were converted into cDNA by reverse transcription using the High-Capacity RNA-to-cDNA™ Kit (Applied Biosystems, Foster City, CA, USA). The relative SLC23A2 gene expression was analyzed by real-time PCR, using a 7900HT Sequence Detection System (SDS; Applied Biosystems ® , Madrid, Spain). TaqMan gene expression assays were used for both target (SLC23A2) and internal control (18S rRNA) genes (Applied Biosystems ® , Spain). Samples were assayed in duplicate. The expression values were calculated by the double delta Ct formula, as previously reported [34,35], and the results were expressed as fold changes in gene expression for each group and subgroup, at baseline.
Statistical Analysis
The "Microsoft Excel" program and "IBM Statistical Package for the Social Sciences" SPSS V.24.0 program (IBM SPSS Statistics for Windows, Version 24.0. Armonk, NY, USA) were used. For continuous variables, non-parametric (Mann-Whitney U) and parametric (t-test) statistics were utilized, and the results were expressed as mean ± standard deviation (SD). Categorical variables were expressed as percentages. A p-value of less than 0.05 was considered statistically significant, which was adjusted by a Bonferroni correction when pairwise comparison in multiple groups was conducted. The non-parametric Spearman's Rho correlation test (or correlation coefficient) was used to evaluate the power of association between 2 specific variables (the value r = 1 means a positive correlation, and the value r = −1 means a negative correlation).
Results
Out of the suitable initial 480 participants, 365 (rate of response from baseline: 76%) completed their 38-month study course ( Figure 1). The overall dropout rate was 24%. Deterioration of health/lack of motivation (7%), confusing ocular fundus or retinographies, and/or progression to PDR or DME (10%), and blood samples failing to show any data (7%) constituted common causes of withdrawn. Table 2 shows the main socio-demographic data, medical history, and laboratory parameters of the study participants. 26% of the baseline population had DR. Lead time between the DM diagnosis and the last study visit was 22 (5) Eating habits were addressed during the study course to check whether the participants had nutrient adequacy and adherence to the MedDiet. Higher nutritional control was observed in the T2DM − DR (41%) than in the T2DM + DR (30%) (p = 0.001). Moreover, a higher percentage of diabetics +DR (32%) displayed a lower MedDiet questionnaire score as compared to the T2DM − DR patients (19%) (p = 0.0001). Significantly higher BMI was seen in diabetics vs. The CG (p = 0.001).
No significant differences were observed between the T2DM patients and the CG, regarding age, gender, and physical exercise routine, but statistically significant differences (p = 0.000001) in glucose and HbA1c determinations were detected between the main study groups (Table 2).
Ocular fundus was explored and photographed in the CG participants, showing a normal appearance without any DR signs ( Figure 3A), in contrast to the diabetics with mild-to-moderate retinal complications, as reflected in the ocular fundus of Figure 3B.
Furthermore, the series of ocular fundus examinations of the T2DM group revealed a mild increment in NPDR development and progression with respect to baseline. Among the participants who completed the study altogether, those with retinopathy showed impairment of retinal affectation grading (54%: Mild DR; 35%: Moderate DR; 11%: Severe DR) from baseline (76% mild; 21% moderate; 3% severe). Table 2. Demographics, lifestyle, and classical biochemical blood traits in the study participants. Similar age and distribution by gender, but significant differences regarding the glycemic profile were observed between the T2DM patient vs. The CG throughout the follow-up.
T2DM
CG Morphologic and morphometric evaluation of the OCT images was performed on the T2DM patients and the CG individuals ( Figure 3C,D). Representative OCT images, as shown in Figure 3C,D, did not show noticeable differences in central retinal thickness from each eye when the diabetic patients and the control individuals were compared ( Figure 3E). We have to emphasize that this tool improved our capacity to document diabetic retinal lesions in the posterior eye pole. Table 3 shows the main parameters of the ophthalmologic examination in the study participants. These data revealed noticeable differences in structural/functional values between groups, distributed into participants assigned to the A/ω3 regimen versus the no supplemented ones.
A relevant point of this follow-up study was the evaluation of DR progression. In fact, the ETDRS grading scale [28] (that is more popular among the researchers that other similar DR classification systems) was used in the present work to define DR progression. The presence of hemorrhages, hard exudates, and cotton-wool spots was equivalent to the ETDRS level 35 of a mild NPDR [28]. Moderate and severe NPDR forms have also been conveniently staged, as reflected in Figure 3.
From the 129 T2DM patients with DR (NPDR forms) at baseline, 22% showed a cumulative incidence of DR impairment at the end of follow-up, the majority of them (13%) corresponding to mild NPDR stages, 5% to moderate stage, and 4% corresponding to the severe NPDR form. Among the diabetics that showed clinical signs and symptoms of retinopathy progression, 65% pertained to the −A/ω3 subgroup. From the 158 initial T2DM patients without NPDR, (27%) developed DR at the end of the study, 12% mild; 9% moderate; 5% severe. Among the diabetics that initially did not suffer DR, 58% pertained to the −A/ω3 subgroup as shown in the Table 4. Table 3. Ophthalmologic data in the supplemented vs. The non-supplemented participants. The OS markers were assayed in plasma samples from the groups and subgroups of participants. . Therefore, diabetics displayed significantly higher circulating pro-oxidant markers and lower antioxidant levels than the non-diabetic study population. Furthermore, significantly higher plasma levels of MDA/TBARS, lower TAC, as well as decreased vit C were observed in the T2DM + DR with respect to the T2DM − DR patients, both at baseline and at the end of follow-up. Table 5 illustrates the effects of the A/ω3 administration in mono-combination on biochemical blood traits, as the patients were categorized in relation to their response to the nutraceutical regimen. Therefore, participants randomly assigned to the oral +A/ω3 improved their OS background with statistical differences between the T2DM + DR vs. T2DM − DR patients throughout the 38-month follow-up (Table 5). A regular A/ω3 supplementation course reduced the OS by conferring a potential neuroprotective effect to the diabetic retina, which is a priority and challenging topic for our current research and the foreseeable future.
Baseline 38-Months Baseline 38-Months End of Study Variables
In our study population, we observed that aging, higher DM length, bad metabolic control (as reflected by the elevated levels of glycemia and HbA1c) in the presence of significant oxidative load were frequent findings that helped to identify diabetics at risk of DR and DR progression. We also found that overweight patients had poorer overall glycemic control also suffered a more advanced stage of NPDR than the normal-weight diabetics throughout the 38-month follow-up. Regarding the gene expression assays, we detected a differential expression profile of the SLC23A2 gene (responsible for the regulation of the intracellular levels of vit C) between the main groups and subgroups of the study participants at baseline (Figure 4).
Discussion
The main results from this population study from baseline to the 38-month follow-up were: (1) DR progression was more likely with higher DM duration, BMI, and Hb1Ac in the context of outstanding OS, and (2) Noticeable amelioration over six consecutive screening intervals, was detected in major hematological parameters and OS markers in the T2DM + DR patients randomly assigned to the oral A/ω3 regimen, as compared to the non-supplemented participants.
Several authors reported that metabolic and risk factors control ameliorated DR prognosis [1][2][3][4][5][6]19,28]. In the current study, we planned to analyze classic risk factors (T2DM duration, sedentary lifestyle, eating habits, and BMI), as well as blood parameters (glucose/HbA1c) and findings from the circulating candidate DR predictors (MDA/TBARS, TAC, GSH, vit C, and the SLC23A2 gene). Recent reports established that DR prevalence in Spain in the T2DM patients depends on the study population, ranging from 15% [27] to 26% [28,36]. From our initial 480 participants, 26% had DR. In addition, the mean T2DM lastingness at 38-months follow-up was 15 (3) years in −DR patients vs. 22 (5) years for the +DR patients. Nevertheless, additional factors also play critical roles in managing DR at any stage [1][2][3][4][5][6]19,[26][27][28], with a special interest in diabetics without clinically detectable retinopathy and those suffering from the early DR stage [37,38]. When corrected, these data for multiple testing, significantly lowered SLC23A2 gene expression (fold change) in the T2DM vs. The CG (p < 0.05) and in the T2DM +DR vs. −DR (p < 0.05) was detected. Furthermore, we found that the SLC23A2 gene was down-regulated in T2DM patients, and it was associated with significantly lower vit C plasma levels (p = 0.001). The above association remained statistically significant (p < 0.05) even after a multivariate adjustment for major potential confounders (sex, age, eating habits, and BMI) was carried out in the data processing.
In addition, we analyzed the correlation between the main OS markers and the SLC23A2 gene expression in blood samples of the study participants at baseline. Spearman Rho correlation between the plasma MDA/TBARS values and SLC23A2 blood gene expression was statistically significant The final rate of compliance with the A/ω3 regimen was 89% for the T2DM and 80% for the CG. The highest response was for the T2DM + DR group than those from the T2DM − DR patients. There were no observed/reported side effects of the oral supplementation, except a few cases of mild gastrointestinal complaints.
Discussion
The main results from this population study from baseline to the 38-month follow-up were: (1) DR progression was more likely with higher DM duration, BMI, and Hb1Ac in the context of outstanding OS, and (2) Noticeable amelioration over six consecutive screening intervals, was detected in major hematological parameters and OS markers in the T2DM + DR patients randomly assigned to the oral A/ω3 regimen, as compared to the non-supplemented participants.
Nutrition is a pivotal component for diabetic patients care. Regarding DR, researchers have described the role of micronutrients (including A/ω3), macronutrients, several food groups, and dietary aspects and lifestyle, such as the MedDiet, in the initiation and progression of DR [4][5][6][18][19][20][21][22][23][24]. Despite this, findings remain inconclusive, and current evidence does not inform about the specific dietary components that are likely to reduce (or increase) the risk of DR. Moreover, new pharmacologic therapies for DR are urgently needed [4,19,[26][27][28].
A growing paradigm in managing DR is the relevance of endogenous pathogenic mechanisms, such as OS and its downstream effectors [4,[12][13][14][15]19]. A variety of plasmatic OS biomarkers (MDA/TBARS, TAC, vit C, vit E, and ω3) have been reported in diabetics [7,[12][13][14]19,26,27,39,40]. Furthermore, Pan et al. interestingly reported a combinatorial molecular signature that may potentially be used as a marker of PDR [41]. Regarding the latter report, we similarly have described a differential profile of circulating OS markers in the T2DM patient's with respect to the non-diabetic individuals, but our diabetic population was only constituted by NPDR patients.
There is growing interest in identifying genes that could trigger OS in diabetics. Our experiments allowed us to demonstrate a statistically significant differential expression profile of the SLC23A2 gene between the T2DM patients and the non-diabetic participants, as well as between the T2DM + DR vs. −DR patients. Being this gene the regulator of vit C intracellular levels (for protecting the cells from OS) and the promoter of type I collagen maturation [24,25], we also found that the SLC23A2 gene was down-regulated in T2DM patients and highly associated with significantly lower vit C plasma levels. Multivariate adjustment for major potential confounders (sex, age, eating habits, and BMI) was carried out in the data processing, and the association of lower SLC23A2 expression levels with lower vit C plasma concentration remained statistically significant in our study population (p < 0.05). Shaghaghi et al. [42] emphasized the necessity to consider the genetic variation of vitamin C transporters in clinical and epidemiologic studies regarding complex diseases. Encouraged by our results, we strongly suggest the connecting role of the SLC23A2 gene down-regulation and the inappropriate vit C bioavailability, with redox imbalance in the context of chronic hyperglycemia and DR. In this concern, Murgia and Adanski [43] also reported that genetic variation alters nutritional requirements, and that nutritional genomics provides an outstanding challenge for biomedical research, as in DR.
In addition, a significant plasmatic decrease in MDA/TBARS, as well as an increase in TAC, was found in the T2DM +A/ω3 and the CG +A/ω3 during the study period. Inclusion of multiple components in the tested formula [vit (E, C, B1, B2, B3, B6, B9, and B12), lutein, zeaxanthin, GSH, hydroxytyrosol, trace elements (Se, Mn, Zn, and Cu) and DHA] theoretically targeting multiple and overlapping pathogenic mechanisms implicated in DR, strongly support a wider counter-response [7,[12][13][14]19,[44][45][46]. In fact, the European Union ROS consortium (EU-ROS) stated in 2017 that many diseases had been found to be associated with OS, and subsequently, the theory that OS-related pathologies can be corrected by antioxidant therapy was launched [47]. However, while experimental studies support this theory, clinical studies still generate controversial results [48,49]. In our opinion, the possibility remains that biotherapies based on the redox changes occurring in diabetics may help to protect the retina. In the latter concern, an interesting spectrum of presumptive factors for DR protection has been ultimately proposed, including outstanding antioxidants [50][51][52][53].
From a clinical viewpoint, retinal microvascular changes found in the NPDR eyes are related to basement membranes thickening, pericyte loss, capillary occlusion, and microaneurysms, that can progress towards ischemia, macular edema, and neovascularization [4,5,19,[36][37][38]. In this context, image analysis of retinal lesions can provide qualitative and quantitative data to accurately diagnosing and grading DR. Our data suggest relatively lower rates of progression of DR in our cohort during the 38-month follow-up. Regarding the participants that completed the study, only 11% of the T2DM patients +DR progressed to severe NPDR forms. Among the participants assigned to the A/ω3 supplementation, the impairment was noticeably less than in the non-supplemented participants. From the 129 T2DM patients with DR (NPDR forms) at baseline, 22% showed a cumulative incidence of DR impairment at the end of follow-up, the majority (18%) corresponded to the mild-to-moderate NPDR stages. Among the diabetics that showed clinical signs and symptoms of retinopathy progression, 65% pertained to the −A/ω3 subgroup. Moreover, from the 158 initial T2DM patients without NPDR, 27% developed DR at the end of the study. In addition, the majority of these pertain to mild-to-moderate stages of the disease (21%), but 5% progressed to the severe NDPR. Among the T2DM that initially did not suffer DR, 58% were concerned with the −A/ω3 subgroup.
Data also showed that a reduction of hyperglycemia, HbA1c, and BMI may have a beneficial effect on the progression to more severe DR forms. We also confirmed that a regular A/ω3 supplementation course reduces the oxidative load, confers beneficial effects as opposed to detrimental ones, and may potentially act as a dietary prophylaxis/adjunctive intervention for high-risk diabetics. The OCT imaging technique provides retinal cross-section images within the micrometer resolution, allowing the evaluation of structural characteristics of the healthy and pathologic retina [54]. It is widely known that the OCT examination qualitatively and quantitatively improves the preclinical diagnosis of retinal changes occurring in diabetics [13,19,[55][56][57][58][59][60]. It was reported in T2DM -DR patients that the foveal and temporal retinal thickness appeared significantly reduced and that a higher HbA1c level was associated with a thinner retina in the temporal perifoveal area in these patients [58]. Data comparison at the end of follow-up in our study cohort showed a significant trend to stabilization of the ophthalmologic parameters, mainly regarding the BCVA and OCT-derived measurements in the T2DM + DR + A/ω3 vs. The T2DM + DR − A/ω3 groups. This is an interesting point to consider because interventions directed to counteract several of the pathogenic mechanisms involved in the deleterious effects of chronic hyperglycemia on the retinal microvasculature are needed.
Moscos et al. [61] reported an increase in the foveal thickness of T2DM patients at two years of oral supplementation with carotenoids, suggesting that the above type of intervention may benefit the visual function in diabetics. It has also been said that early oral supplementation with nutraceuticals provides a significant health benefit to prevent DR progression because of its antioxidant and anti-inflammatory properties, reducing both the neural and vascular damage caused in the retina by the hyperglycemia [7,12,19,62,63]. The combined A/ω3 formula used herein strongly supported the descriptions of the above authors and our previous works [19,35,39,40,49] on the processes of retinal neuroprotection and vasculoprotection. Moreover, Amato et al. [64] emphasized the possible side effects of systemic neuroprotectants as well as doubts about the appropriate retinal bioavailability of these substances in contrast with the intravitreal injections. We hereby confirm that with the formulation used herein, no adverse effects were reported but scarcely mild gastrointestinal discomfort, while the intravitreal injections may potentially produce endophthalmitis, hemorrhages, rhegmatogenous retinal detachment, lens damage, and/or other complications [65]. These authors interestingly described the feasibility of delivering magnetic nanoparticles functionalized with a neuroprotectant substance by means of intravitreal injections in mice, suggesting an interesting role of these new approaches to protect the retina in diabetic patients. In this concern, Dolz-Marco et al. [66] also reported that intravitreal DHA administration (at the specific assayed doses) in rabbits is safe and potentially useful for treating retinal diseases.
No data from the interim time points were presented in this work because the main results from the 18-months follow-up were previously published [19]. Regarding the above data, the BCVA (each eye separately) was significantly higher in the CG as compared to the T2DM group [(CG at baseline: RE: 0.95 ± 0.10 and LE: 0.96 ± 0.08; p < 0.001/CG; at 18 months: RE: 0.95 ± 0.11 and LE: 0.95 ± 0.10; p < 0.001) (T2DMG at baseline: RE: 0.81 ± 0.30 and LE: 0.83 ± 0.22; p < 0.001/T2DMG; at 18 months: RE: 0.78 ± 0.22 and LE: 0.76 ± 0.22; p < 0.001)]. Regarding the redox status of the study population at 18 months, on the one hand, plasmatic pro-oxidants (MDA/TBARS levels) significantly decreased in the T2DM + A/ω3 subgroup with respect to the T2DM − A/ω3 patients, and on the other hand, the TAC significantly increased in the T2DM + A/ω3 subgroup, while the non-supplemented participants did not show any change. These data are similar to those obtained from the 38-month follow-up during the present work.
The authors are confident that the results obtained during this follow-up study can be useful in designing an integrative document with the clinical (ocular fundus multimodal imaging) and molecular-genetic biomarkers (MDA/TBARS, TAC, vit C, and SLC23A2 gene) in the form of practical guidelines that can be helpful for primary care physicians as well as to ophthalmologists and ophthalmic researchers. In this concern, classical unmodifiable risk factors, such as the older age and/or longer DM duration have been widely involved in the development and progression of DR, altogether with the modifiable biochemical parameters of higher glycemia and HbA1c as well as the increased BMI and the lack of physical exercise [1][2][3][4][5][6]19]. All these, and the oxidative stress parameters reported herein (MDA/TBARS, TAC, vit C), can help outline the importance of integrating patient characteristics, as well as clinical and biochemical biomarkers better to identify diabetics at risk of retinopathy and visual disability. This follow-up study strongly emphasizes the benefits of the effects of the A/ω3 administration in mono-combination on both the DR and biochemical blood traits in T2DM patients. Up to today, our group continues to pursue research, consistently looking for new solutions to avoid diabetic blindness.
A study limitation is the relatively small final sample of participants distributed in each group and subgroup reaching the 38-month follow-up. Patients who developed PDR or DME were excluded from data processing to avoid the bias of statistical procedures, mainly because these patients needed to be immediately treated with specific medical, laser, or surgical interventions. Statistical processing of the GSH data showed surprising results. No significant differences were found at the end of the study regarding the A/ω3 supplementation (containing GSH). We may hypothesize that sample handling may induce artificial oxidation of GSH, producing GSSG that interfered with our results. It is needed to proceed with a rapid -SH alkylation with N-ethylmaleimide to improve these results for further research. The study has provided a large amount of data for easy analyses. Because of this, we have focused on the main objectives of our work, and some of the data have been omitted for the present study. In fact, the expression of the SLC23A2 gene was only assayed at baseline in the main groups and subgroups of participants. No comparison was made of the SLC23A2 gene expression between the baseline and the end of follow-up and nor in the participants assigned or not to the oral supplement regimen due to unsolvable problems at the end of follow-up. Nevertheless, the differential expression profile of the SLC23A2 gene in diabetics, as compared to the controls, as well as in diabetics with and without NPDR, provided interesting data at the beginning of the study (as reflected in Figure 4), showing outstanding correlation with the plasma vit C levels. In fact, Hierro et al. [67] demonstrated a SLC23A2 down-regulation in the kidney and brain from a model of streptozotocin Zucker rats.
Conclusions
In summary, the results of the present study reveal the importance of multicenter collaborative works in DR research to better managing the diabetic retina. As retinal microangiopathy is a sight-threatening complication of DM, early diagnosis, proper supervision, and appropriate treatment of DR are pivotal for optimizing eye and vision care in the affected patients. We confirm that OS plays relevant roles in DR initiation and progression, also suggesting that OS can be detected with easily accessible and inexpensive blood analyses of T2DM patients. We have identified the MDA/TBARS, TAC, vit C, and SLC23A2 gene as presumptive prognostic biomarkers to improve knowledge on DR and ophthalmological outcomes. Large scale studies are needed to better estimate its efficiency and further usability in clinical practice. Finally, this 38-month follow-up study supports that a regular A/ω3 supplementation reduces the oxidative load and may serve as a dietary prophylaxis/adjunctive intervention for patients at risk of diabetic blindness
Acknowledgments:
The authors thank all physicians, ophthalmologists and researchers who have helped us to complete the present study for their work and outstanding interest in diabetic retinopathy. We especially wish to thank D. Corella (Valencia, Spain), G. Virgili (Florence, Italy), and J. F. Arévalo (Baltimore, MD, USA) for peer interactive communication, and for the thoughtful comments and critical reading of the manuscript. The authors and the involved professionals in the present study want to thank Laboratorios Thea, S.A., (Barcelona, Spain) for gently providing the Nutrof Omega ® formulation that was given to the participants during the study course.
Conflicts of Interest:
All authors of this work have disclosed that they have no significant financial relationships or financial interests in the commercial companies that are related to this study or paper. | 2020-11-12T09:05:01.699Z | 2020-11-01T00:00:00.000 | {
"year": 2020,
"sha1": "cb7b240d7e9d7cfc1e0238bac3c93ce2b1369a27",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2076-3921/9/11/1101/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "55c1ceed6eff9928465a8d4e02e2e57702edc3be",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
1894104 | pes2o/s2orc | v3-fos-license | Pathogen-Specific T Cell Polyfunctionality Is a Correlate of T Cell Efficacy and Immune Protection
Introduction Understanding the factors that delineate the efficacy of T cell responses towards pathogens is crucial for our ability to develop potent therapies against infectious diseases. Multidimensional evaluation of T cell functionality at the single-cell level enables exhaustive analysis of combinatorial functional properties, hence polyfunctionality. We have recently invented an algorithm that quantifies polyfunctionality, the Polyfunctionality Index (Larsen et al. PLoS One 2012). Here we demonstrate that quantitative assessment of T cell polyfunctionality correlates with T cell efficacy measured as the capacity to kill target cells in vitro and control infection in vivo. Methods We employed the polyfunctionality index on two datasets selected for their unique ability to evaluate the polyfunctional imprint on T cell efficacy. 1) HIV-specific CD8+ T cells and 2) Leishmania major-specific CD4+ T cells were analysed for their capacity to secrete multiple effector molecules, kill target cells and control infection. Briefly, employing the Polyfunctionality Index algorithm we determined the parameter estimates resulting in optimal correlation between T cell polyfunctionality and T cell efficacy. Results T cell polyfunctionality is correlated with T cell efficacy measured as 1) target killing (r=0.807, P<0.0001) and 2) lesion size upon challenge with Leishmania major (r=-0.50, P=0.004). Contrary to an approach relying on the Polyfunctionality Index algorithm, quantitative evaluation of T cell polyfunctionality traditionally ignores the gradual contribution of more or less polyfunctional T cells. Indeed, comparing both approaches we show that optimal description of T cell efficacy is obtained when gradually integrating all levels of polyfunctionality in accordance with the Polyfunctionality Index. Conclusions Our study presents a generalizable methodology to objectively evaluate the impact of polyfunctionality on T cell efficacy. We show that T cell polyfunctionality is a superior correlate of T cell efficacy both in vitro and in vivo as compared with response size. Therefore, future immunotherapies should aim to increase T cell polyfunctionality.
Introduction
Understanding the factors that delineate the efficacy of T cell responses towards pathogens is crucial for our ability to develop potent therapies against infectious diseases. Multidimensional evaluation of T cell functionality at the single-cell level enables exhaustive analysis of combinatorial functional properties, hence polyfunctionality. We have recently invented an algorithm that quantifies polyfunctionality, the Polyfunctionality Index (Larsen et al. PLoS One 2012). Here we demonstrate that quantitative assessment of T cell polyfunctionality correlates with T cell efficacy measured as the capacity to kill target cells in vitro and control infection in vivo.
Methods
We employed the polyfunctionality index on two datasets selected for their unique ability to evaluate the polyfunctional imprint on T cell efficacy. 1) HIV-specific CD8 + T cells and 2) Leishmania major-specific CD4 + T cells were analysed for their capacity to secrete multiple effector molecules, kill target cells and control infection. Briefly, employing the Polyfunctionality Index algorithm we determined the parameter estimates resulting in optimal correlation between T cell polyfunctionality and T cell efficacy.
Results
T cell polyfunctionality is correlated with T cell efficacy measured as 1) target killing (r=0.807, P<0.0001) and 2) lesion size upon challenge with Leishmania major (r=-0.50, P=0.004). Contrary to an approach relying on the Polyfunctionality Index algorithm, quantitative evaluation of T cell polyfunctionality traditionally ignores the gradual contribution of more or less polyfunctional T cells. Indeed, comparing both approaches we show that Introduction Pathogens compose a major socio-economic challenge to modern society. Humans are able to develop pathogen-specific immunity, which is induced either naturally (pathogen infection) or artificially (vaccination). Such immunity is supposed to confer protection by 1) antibody mediated neutralisation and elimination of pathogens, or to control infection through 2) T cell mediated elimination of infected host cells. Understanding the factors that delineate the efficacy of antibody and T cell responses towards pathogens is crucial for our ability to develop potent therapies.
T cells play important roles in the series of highly coordinated immune events that lead to pathogen clearance. Indeed, they are directly involved in the eradication of infected host cells, but they are also inherently communicating with innate immunity and pathogen-specific antibody development, which are crucial for pathogen clearance. It is custom to analyse the effect of T cells at different levels, 1) pathogen clearance and clinical recovery, 2) target killing, cellular help and recruitment of innate immune cells and 3) effector molecules expressed by T cells. Whereas T cell efficacy is typically evaluated extrinsically (level 1 and 2), their functionality is more often analysed intrinsically (level 3). Indeed, T cell functionality assays have the advantage of being applicable to large cohorts as well as many cell types and subsets in a standardized manner, with readouts that can be highly multiparametric. Here, we focus on how to associate or even predict extrinsic T cell efficacy from intrinsic T cell functionality. Using highly multiparametric datasets of T cell polyfunctionality we also propose a widely applicable analytical strategy, which objectively identifies the importance of individual and combinatorial effector functions.
Functional evaluation of T cell responses has in recent years advanced from single-parameter (e.g. IFN-γ-secretion) to more complex multidimensional measurements. Numerous studies have successfully associated single-parameter functional assays of T cells with their efficacy. [1] Furthermore, it is becoming increasingly clear that ex vivo functional polyvalency of T cells is an important correlate of T cell efficacy. [2,3,4,5,6] Of note, it is still debated if ex vivo T cell polyvalency is directly [7] or indirecty [8] associated with T cell efficacy in vivo. Presently, no standardized or objective analytical technique exists to properly define the relationship between functional polyvalency and biological efficacy.
T cell polyfunctionality, defined as the co-expression of multiple functional molecules (such as cytokines and chemokines) at the single-cell level, is measurable with high-throughput technologies, such as multiparametric flow cytometry. Multiparametric cell analysis generates very large combinatorial datasets. For instance, 3 bimodial parameters divide cells into 2 3 = 8 socalled boolean combination subsets. In the case of T cell polyfunctionality, the complexity of such boolean datasets has been reduced to enable associations with clinical and biological parameters. Two main approaches have been applied in attempt to reduce complexity of T cell polyfunctionality, 1) polyfunctionality profiles characterized by the frequency of cells positive for 0, 1, 2, until n simultaneous parameters presented as pie diagrams (n+1 dimensions) and 2) quantitative polyfunctionality characterized by the frequency of cells that produce a defined number of functions ( Fig 1A).
Polyfunctionality profiles have been instrumental for our understanding of cellular functionality, but remain difficult to interpret with appropriate statistical inference. Quantification of cellular polyfunctionality would largely facilitate statistical analysis, but so-far there is no widely accepted quantification methodology that integrates the contribution of all functional layers of T cells. Indeed, the quantitative approaches most frequently used [9,10] are restricted to the enumeration of highly polyfunctional cells and do not consider pauci-functional cells. More specifically, they do not take into account functional cells positive for a number of functions inferior to a defined lower limit and do not differentiate cells according to the number of functions for which they are positive (Fig 1A).
We have recently proposed a novel algorithm, which quantifies polyfunctionality while taking into account all functional T cells (polyfunctionality index) (Fig 1B). [11] The polyfunctionality index is easy to implement with a variety of statistical techniques, such as comparative and correlative statistics (Fig 1C). [12,13] Here, we use the algorithm to determine the importance of T cell polyfunctionality with regards to T cells capacity to kill target cells in vitro and control infection in vivo, hence T cell efficacy. We finally develop an approach with which we The principles of single-cell polyfunctionality analysis for modelling associated variables, such as T cell efficacy. 3 (n) bimodial effector molecules (A, B and C) were measured at the single-cell level identifying 2 3 = 8 distinct combinatorial cell subsets, which can be stratified according to the number of simultaneous functions at the single-cell level (i). A) "Classical" polyfunctionality analysis generally assess only cells positive for a defined minimal number of simultaneous effector molecules (i.e. 3 simultaneous functions). B) Contrarily, the polyfunctionality index considers all 8 functional subsets and ingeniously parameterizes the influence of individual (φ A >φ B >φ C ) as well as combined (q) functionalities. C) Polyfunctionality assessed with the polyfunctionality index algorithm is therefore modulable, contrary to "classic" polyfunctional analysis, enabling a more optimal fit (green line) of model variables. D) Using regression analysis it is feasible to obtain proper parameter estimates (φ A , φ B , φ C and q), which have biological significance. Indeed, interpretation of such parameters enables an objective evaluation of the influence of individual as well as combinatorial functions on the predictive capacity of polyfunctionality with regards to a desired model variable, such as T cell efficacy.
were able to accurately discern and grade individual functional parameters, such as effector molecules, which are important for describing the association between T cell polyfunctionality and T cell efficacy (Fig 1D). Our approach enables an objectively based exclusion of redundant functional parameters and thus reduce both cost and complexity of experimental procedures. The parameter estimates identified by our approach may serve as predictive biomarkers and identify key targets to be modulated by therapeutic interventions.
Selection of T cell polyfunctionality data
Contrary to the analytical approaches so-far utilized for T cell polyfunctionality analysis, the polyfunctionality index provides several components enabling an objective and quantitative assessment of the degree to which T cell efficacy is associated with T cell polyfunctionality. Moreover, we can robustly grade the impact of individual functional parameters. In two of our previous studies we showed association between T cell polyfunctionality and T cell mediated target killing [9] as well as vaccine induced control of Leishmania major infection [10]. We therefore selected these two datasets to more thoroughly understand the contribution of individual functional parameters with respect to T cell efficacy.
Polyfunctional CD4 + T cell responses against Leishmania major A second experimental dataset was composed of polyfunctionality profiles of Leishmania major-specific CD4 + T cells from the spleen of mice vaccinated against Leishmania major. These data were associated with vaccine induced protective capacity measured as the lesion size post infectious challenge as previously described. [10] Briefly, mice were vaccinated with a recombinant leishmanial polyprotein (MML) either expressed by a replication-deficient adenovirus or as protein with CpG as adjuvant. Vaccination with adenovirus was delivered once either intramuscular (leg) or subcutaneous (foot). Protein vaccination was administrated three times in intervals of two weeks. A group of mice were vaccinated with sub-lethal doses of live Leishmania major. T cell polyfunctionality defined by co-expression of IFN-γ, TNF-α and IL-2 was analysed by intra-cellular cytokine staining at day 28 post vaccination by in vitro stimulation with MML for 6 hours. Stained cells were acquired on a BD LSRII flow cytometer. Maximum lesion size was measured post intradermal challenge in both ears with metacyclic Leishmania major promastigotes.
Flow cytometry data analysis
Data analysis was accomplished with FlowJo (TreeStar Inc) software. Polyfunctionality analysis was performed using Pestle and Spice softwares (Mario Roederer, ImmunoTechnology Section, VRC/NIAID/NIH) [14] as well as FunkyCells Boolean Dataminer software (www.FunkyCells. com). Primary polyfunctionality data for the two datasets analysed in the present manuscript can be found in S1 and S2 Tables.
Ethics statement
Human blood samples were obtained following acquisition of the study participants' written informed consent and the study protocol was reviewed and approved by local ethics committee of Pitié-Salpêtrière Hospital, Paris. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of The National Institutes of Health. The protocol was approved by the Vaccine Research Center Animal Care and Use Committee (protocol number VRC-04-072).
Polyfunctionality index algorithms
In the analysis herein, we define three different, but related algorithms to represent polyfunctionality. The formulae and their application in validating their relationship with T cell efficacy are summarized in Table 1. 1. General polyfunctionality index. We previously developed the general polyfunctionality index algorithm. [11] Briefly, ::::::
General PI
X 1 :::::: Effector molecules positively associate with T cell efficacy.2) Importance of effector molecules are graded according to their relevance with respect to the predicted parameter (i.e. T cell efficacy).
3) Accounts for increasing levels of polyfunctionality
Importance of effector molecules are NOT graded according to their relevance with respect to the predicted parameter (i.e. T cell efficacy). 2) Accounts for increasing levels of polyfunctionality Modified general PI X 1 :::::: None PI/PI max e PI / (e PI +1) Apply general PI X 1 :::::: n > 0 is the number of functions studied. x j indicates in a binary fashion (2) if the combinatorial T cell subset is the frequency of cells performing the particular combination of functions Þ is a factor assigned to a T cell subset performing the particular combination of functions . q!0 is the parameter that modulates the effect of increasing levels of polyfunctionality, be it in a monotonic (q = 1) or exponential (q6 ¼1) manner, according to the number of simultaneous functions expressed by the cell subset. The algorithm requires that the sum of all F x 1
2. Special polyfunctionality index. The general polyfunctionality index can be simplified to the special polyfunctionality index (7), when all effector molecules are considered to contribute equally (φ j = 0 for all j).
where F i is the frequency of cells performing i simultaneous functions.
3. Modified general polyfunctionality index. Some parameterizations of the general polyfunctionality index may not be fully adapted to certain experimental conditions. The general polyfunctionality index assumes that all effector molecules contribute positively to T cell efficacy (6). Indeed, this may not be the case for anti-inflammatory and regulatory cytokines or in the case where particular cytokines define detrimental differentiation pathways. In order to account for effector molecules that have no effect or even a negative effect on T cell efficacy, we ignore assumption (6), thus letting all φ j take on any value from negative to positive infinite. As a result, the denominator ð1 þ φ 1 Þ no longer ensures that the polyfunctionality index ranges from 0-1 and was hence removed from Eq (1). Instead, the distribution of the polyfunctionality equation must be evaluated a posteriori based on the given inputs, in which other operations may be warranted (i.e. adjustment via dividing by the maximum PI value or inverse logit transformation-cf. Table 1). The following Eq (8) represents the modified general polyfunctionality index: :::::: Interpretation of polyfunctionality index parameters q and φ From these equations, the parameters of most interest are q and φ i and can be estimated using appropriate regression techniques. In the instance of q, the values it takes are interpreted as follows: q ! 0: polyfunctionality is not an immune correlate. 0 < q < 1: polyfunctionality is a moderate immune correlate. q > 1: polyfunctionality is a strong immune correlate.
In the instance of φ i , the values it takes on are interpreted as follows: φ i < 0: the function i contributes negatively to the polyfunctionality index. φ i = 0: the function i contributes neutrally to the polyfunctionality index. φ i > 0: the function i contributes positively to the polyfunctionality index. Of note, the φ i -values are relative measures. Indeed, comparing φ i with all other φ j permit interpreting which functions contribute more or less to the polyfunctionality index, and therefore which functions are more relevant to measure to model or predict a given target variable.
Modelling T cell efficacy using the special polyfunctionality index
For the experiments assessing the polyfunctionality of HIV-specific CD8 + T cells, [9] we employed non-linear least-squares regression with the dependent variable being percent killing and the independent variable being the resulting series expansion of the special polyfunctionality index (7), where F i was determined using FunkyCells software (www.FunkyCells.com) and i and n were constants. Both the slope (β 1 ) in relation to the polyfunctionality index and constant (β 0 ) parameter estimates were obtained for this model. The parameter q was also directly calculated from this model.
A total of four separate experiments were conducted to assess the polyfunctionality of vaccine induced CD4 + T cell responses towards Leishmania major. [10] For each experiment, separate non-linear least-squares regression models were constructed using mean lesion size as the dependent variable and polyfunctionality index values from the series expansion as the independent variable (described above). In order to determine q, a range of values were inputted into each of the series expansions of the polyfunctionality index, while the median squared correlation coefficient (r 2 ) across the four regression functions was calculated for the same q. This resulted in a description of median r 2 as a function of q, from which numeric first and second derivatives were calculated using a previously defined algorithm. [15] From these, the optimal q was then determined by the local maximum of the concave downward function. Conveniently, this is analogous to using the referenced output option of FunkyCells Boolean Dataminer (www.FunkyCells.com) in conjunction with the Generalized Reduced Gradient non-linear iterative computation included in the Frontline Systems Solver add-in for Microsoft Excel. Graphical representation of r 2 as a function of q can be accomplished using FunkyCells Track Function software (www.FunkyCells.com).
For both these methods, robustness of q was evaluated via bootstrapping by randomly selecting with replacement 90% of the original data (overall for experiments involving HIV-specific CD8 + T cell polyfunctionality and within experiments involving Leishmania major-specific CD4 + T cell polyfunctionality) and re-running the q computation above. 95% two-sided confidence intervals (CIs) were then constructed using the non-studentized pivotal method.
Modelling T cell efficacy using the modified general polyfunctionality index
To assess the differential importance of individual effector molecules we employed the modified general polyfunctionality index (8). Non-linear regression models were used in a similar manner as described above, only the independent variable was the resulting series expansion of the modified general polyfunctionality index. The effector molecule-specific factor, φ j , is then estimated from the terms φ 1 . This can be viewed as a parameter estimate of the individual slopes describing the relationship between a specific effector molecule and T cell efficacy while accounting for polyfunctionality.
Using co-linearity to assess the redundancy of HIV-specific CD8 + T cell effector molecules We first wanted to examine the probability distribution of polyfunctionality for T cells secreting a specific effector molecule. Using FunkyCells software (www.FunkyCells.com), mean frequencies of each boolean combination were calculated. These frequencies can be viewed as the conditional probability of cells secreting i number of effector molecules given that they secrete effector molecule j.
In the regression analysis, we viewed the redundancy of certain effector molecules as a colinearity problem. In order to assess this, we used linear regression to model the frequency of cells expressing one specific effector molecule against the frequency of all other effector molecules with intercept. This resulted in the correlation coefficient, b R 2 j , which was then used to calculate the centered variance inflation factor (VIF) for chemokine j: High VIF (>5) represents strong co-linearity with at least one other effector molecule and suggests that the specific cytokine can be removed from the model.
Statistical analysis
Statistical analysis was performed using STATA (v12.1, College Station, TX, USA) and a pvalue <0.05 was considered significant.
T cell killing is associated with T cell polyfunctionality
We previously showed that functional avidity of 10 CD8 + T cell clones specific for an HLA-B27 Ã 05 restricted p24 Gag HIV epitope (KK10) was associated with polyfunctionality and HIV-suppressive activity. Indeed, CD8 + T cell clones of high functional avidity reacted to stimulation by antigen presenting cells loaded with KK10 at lower concentrations than CD8 + T cell clones of low functional avidity, both with regards to target killing and capacity to produce effector molecules separately and simultaneously. [9] To further understand the relationship between polyfunctionality and T cell efficacy, we employed the polyfunctionality index. In the first approach, we used the special polyfunctionality index, which enables an estimation of the impact of polyfunctionality, in the special condition where all effector molecules are considered equally important (cf. materials and methods section). Briefly, target killing was associated with polyfunctionality index values of individual T cell clones at different target concentrations. In the initial model, q was equal to 1, thereby reflecting a situation where the number of simultaneous effector molecules produced by a T cell adds to the polyfunctionality index in a monotonic manner. We observed a strong correlation between polyfunctionality and target killing (Fig 2A, r (Fig 2B). Using least-squares regression, we then estimated which q resulted in the optimal fit, obtaining a value of q = 1.24 with a correlation coefficient of 0.86 ( Table 2). The capacity of T cell polyfunctionality to predict T cell efficacy as measured by target killing is quantified by the parameter q. The fact that q>0 shows that polyfunctionality is a correlate of T cell efficacy, which contributes monotonically (q = 1) or slightly exponentially (q>1) to the predictive capacity of T cell polyfunctionality. Of note, for q!0 the polyfunctionality index defines the frequency of cells producing at least one effector molecule. Our model therefore allows a direct statistical comparison between the predictive capacity of functional T cell quality (q>0) and quantity (q!0).
Control of Leishmania major infection is associated with T cell polyfunctionality
Whereas the above dataset is based on robust and elaborate in vitro analysis of human T cell functionality, we wanted to investigate whether quantitative assessment of T cell polyfunctionality represents a marker of in vivo T cell efficacy. We previously reported that polyfunctionality of vaccine induced Leishmania major-specific CD4 + T cells were strongly correlated with the protective capacity against an infectious challenge with live Leishmania major. [10] Indeed, the size of the skin lesion induced by the Leishmania major infection was inversely correlated with the frequency of antigen-specific CD4 + T cells producing IFN-γ, TNF-α and IL-2, simultaneously.
Employing the same methodology as outlined for the in vitro generated dataset on T cell killing above, we used the polyfunctionality index to further understand the relationship between T cell polyfunctionality and protection against Leishmania major infection. As above, we used the special polyfunctionality index considering all effector molecules equally important (cf. materials and methods section). We first regressed polyfunctionality index values for Leishmania major-specific CD4 + T cells up-against resistance to Leishmania major infection (peak skin lesion size post challenge) for q = 1. We observed a strong negative slope of -5.29 (95%CI: [-6.26; -4.27], p<0.001) producing a correlation coefficient of 0.58 (P = 0.025) between polyfunctionality and Leishmania major induced skin lesion size (average statistics for 4 independent experiments- Fig 3A). The correlation coefficient changes with varying levels of q ( Fig 3B). Using least-squares regression analysis as outlined above, the optimal q-value with the highest median correlation coefficient was q = 1.33. The robustness of the optimal q-value Predicting T Cell Efficacy with the Polyfunctionality Index was analysed by bootstrapping (cf. materials and methods), giving an optimal q = 1.9 and a 95% confidence interval of [-2.04:5.84]. Interestingly, the Leishmania major-specific response size defined as the frequency of CD4+ T cells producing cytokines upon stimulation with Leishmania major, is less well correlated with protection as measured by lesion size (S1A Fig, r 2 = 0.36, P = 0.12). Moreover, response size and polyfunctionality of the Leishmania major-specific T cell response do not correlate (S1B Fig, r 2 = 0.15, P = 0.48). The capacity of CD4 + T cell polyfunctionality to predict vaccine efficacy, as measured by the resistance evoked towards Leishmania major infection, was quantified by the parameter q, and demonstrates that antigen-specific CD4 + T cell polyfunctionality is a superior correlate of vaccine efficacy compared with response size.
Differential evaluation of effector molecules does not increase the predictive power of T cell efficacy
Our analysis demonstrates that polyfunctionality is an important positive immune correlate for 1) T cell mediated killing of HIV infected target cells and 2) in vivo control of Leishmania major infection. We subsequently wanted to elucidate if certain effector functions would be more important for the correlates than others. We therefore performed a least-squares regression analysis of T cell efficacy as a function of polyfunctionality, defined by the polyfunctionality index. For T cell efficacy measured as target killing, the analysis was performed considering all effector molecules equally important (model 1-special polyfunctionality index as above) and by including effector molecules as parameter estimates (model 2-modified general polyfunctionality index). The two analysis were able to model 85.63% (model 1) and 91.44% (model 2) of the model variability ( Table 2). The parameter estimates showed that q was significantly larger than 0 for model 1 (q = 1.24; 95% CI: 0.85-1.63, p<0.0001). In concordance with model 1 we performed a second analysis taking into account the importance of individual effector molecules (model 2), which found q to be significantly larger than 1 (q = 3.05; 95% CI: 1.30-4.81, p = 0.02) (Fig 2B and Table 2). None of the φ-values associated with the 5 effector molecules, CD107a, IFN-γ, IL-2, MIP-1β and TNF-α, respectively, were significantly different from 0 (Model 2, Table 2). In other words, polyfunctionality is the primary predictor of T cell efficacy and all measured effector molecules contribute equally to the predictive capacity of the polyfunctionality index with regards to T cell efficacy.
We attempted a similar analysis for the Leishmania major vaccine study. However, the variability of this dataset, did not allow the polyfunctionality index parameters to converge. We performed a power analysis using fixed parameters from previous models (S2 Fig), demonstrating that 75 groups of mice would be needed to achieve 80% power in order to determine a significant difference for a true φ i = 1. The inherent data variability is most likely due to the in vivo nature of the mouse experiments. A larger study is required to reduce some of this variability, thus allowing full evaluation on whether certain effector molecules contribute more to the predictive capacity of polyfunctionality with regards to vaccine protection.
Information redundancy between effector molecules
From our previous studies it is clear that effector molecules expressed as a result of cellular stimulation, such as antigen-specific stimulation, appear in quantities ranked according to effector molecules but increasing with the intensity of stimulation in a synchronous manner. [9] Cellular pathways activated by antigen-specific stimulation and leading to expression of effector molecules have been shown to include a significant amount of redundancy. [16] We therefore wanted to evaluate if the effector molecules measured in our experimental dataset would be redundant in a polyfunctional setting.
We first analysed the probability of expressing 1, 2, until n simultaneous effector molecules for cells expressing a given effector molecule. This analysis was performed for all effector molecules within our datasets. For the dataset charting polyfunctionality of HIV-specific CD8 + T cell clones; CD107a, IFN-γ and TNF-α have very similar probability distributions, secreting equally zero, one or two other effector molecules. On the contrary, MIP-1β was mostly secreted alone or with another effector function, while IL-2 was almost always secreted by itself (Fig 4A). This analysis suggests that there might be redundancy between CD107a, IFN-γ and TNF-α. Consequently, we analysed the variance inflation factors (VIF) for each effector Predicting T Cell Efficacy with the Polyfunctionality Index molecule based on the modelling of target killing (Table 3 and Fig 4B). This analysis (model 1) showed that CD107a presents substantial co-linearity with at least one other effector molecule (VIF CD107a = 24.62). We consequently removed CD107a from the model and reanalysed the VIF associated with the remaining effector molecules (model 2). In model 2, TNF-α continued to be highly correlated with at least one other effector molecule (VIF TNF-α = 12.63). We therefore also removed TNF-α from the model (model 3). In this final model, the VIFs of IFN-γ, IL-2 and MIP-1β were all low (VIF<5), indicating that the remaining three cytokines are non-redundant. (Table 3 and Fig 4B) Using this information, we finally performed a non-linear regression model of target killing as a function of T cell polyfunctionality including the three non-redundant effector molecules (IFN-γ, IL-2 and MIP-1β) using the modified general polyfunctionality equation. The methodology was identical to the one applied to all 5 effector molecules above. The result confirmed the previous analysis, as none of the 3 effector molecules had φ-values significantly different from 0 (Table 4). Of note, in concordance with the VIF analysis, this model of target killing explained essentially the same amount of variability as the model including 5 effector molecules (88.99% versus 91.44%). Altogether, we demonstrate a robust methodology to identify redundant functional markers, which can consequently be excluded in future studies based on similar biological and clinical settings.
Discussion
Our study demonstrates the capacity of the polyfunctionality index to objectively quantify to which degree T cell polyfunctionality should be considered an immune correlate of T cell efficacy. We show that T cell polyfunctionality is indeed a superior correlate of T cell efficacy both in vitro and in vivo as compared with T cell response size. In one biological setting we analysed the Predicting T Cell Efficacy with the Polyfunctionality Index efficacy of HIV-specific CD8 + T cell clones to kill target-cells in vitro. In a second biological setting we analysed the efficacy of vaccine induced Leishmania major-specific CD4 + T cells to control Leishmania major infection in vivo. For both studies we demonstrate that our prediction model of T cell efficacy considers the number of simultaneously expressed effector molecules increasingly more informative. Of note, the importance of polyfunctionality with regards to both HIV-specific CD8 + and Leishmania major-specific CD4 + T cell efficacy is very similar, despite the two otherwise unrelated biological settings. This proposes that general trends across clinical and biological settings may be found, although this hypothesis will take further studies to verify.
Oftentimes, experimental assessment of T cell efficacy requires procedures that are technically challenging and/or non-ethical. E.g. human vaccination trials most often can only await clinical infectious events as it would be unethical to perform infectious challenge experiments on vaccinated subjects. In the same line, pre-clinical vaccination protocols in animals may not have a clinically relevant infectious model (e.g. SIV versus HIV infection). In the latter cases it may be necessary to create in vitro experimental procedures to assess T cell efficacy. It would therefore be of outmost importance to define robust immune correlates of T cell efficacy based on assays that are feasible to perform in terms of ethics, cost and standardization. Detection of a battery of effector molecules secreted by antigen-specific T cells fulfils these criteria, as it only requires a blood sample and a fairly labour non-intensive experimental procedure. From the inputs obtained in the study herein, the polyfunctionality index can be used to establish a comprehensive score for T cell activity, which, from our data, is a robust predictor of T cell efficacy. Of note, the polyfunctionality index is calculated from the frequency of all combinations of effector molecules, but does not presently integrate the quantity of effector molecules (accessible as signal intensity, e.g. fluorescent intensity). Our previous works demonstrate that the quantity of effector molecules defines a characteristic hall mark of CD4 Th-type subsets [17] and is associated with vaccine efficacy [10]. It might therefore be possible to further improve the polyfunctionality index by integrating the quantity of effector molecules. Future studies aiming to examine T cell efficacy among subpopulations of HIV-1 or Leishmania major infected individuals are facilitated by polyfunctionality as a proxy.
However, there is one major caveat for using polyfunctionality in that each infection and set of effector molecules must be associated with a clearly definable marker of T cell efficacy. This is all the more important as the impact of T cell polyfunctionality in various biological and clinical settings could be heterogeneous. Indeed, for HIV-specific CD8 + T cells we associated two very closely related measures, T cell killing and polyfunctionality of T cells presented to the same target in the same type of in vitro assay. Conversely, measuring the efficacy of Leishmania major-specific CD4 + T cells as the capacity of a mouse to control Leishmania major infection is more biologically distant to the T cell polyfunctionality assay. Noteworthy, our results observed for datasets derived from these two biologically and clinically distinct phenomena are very homogeneous. To know if our observations are widely applicable, the outlined approach should be applied in more studies associating T cell polyfunctionality with T cell efficacy. Single-cell polyfunctionality is measurable by multiple techniques for many cell types and many different panels of effector molecules. [9,13,18] The approach outlined in the present work is applicable to all types of single-cell polyfunctionality measures. Our study defines a robust methodology to identify the most important parameters (polyfunctionality and/or effector molecules) contributing to a given polyfunctionality analysis. We therefore believe that past and future studies including single-cell polyfunctionality measures could readily benefit from our analytical approach to increase biological insight.
Two important aspects of our work are to 1) identify potential redundancy between effector molecules and 2) define if non-redundant effector molecules contribute differentially or equally to the immune correlate defined by the polyfunctionality index.
To answer the first question we performed a variance inflation factor analysis of our datasets. For the five effector molecules analysed in our in vitro CD8 + T cell assay, we successfully identified three non-redundant effector molecules (IFN-γ, MIP-1β and IL-2). To define their contribution to the polyfunctionality immune correlate, we employed the modified general polyfunctionality index. In concordance with the variance inflation factor analysis this model demonstrated similar model fit compared to the model with five effector functions. None of the three effector molecules measured in the Leishmania major vaccination dataset were redundant. Of note, redundancy is here uniquely linked to our predictive model and by no means linked with the biological roles of the monitored effector molecules.
The second question was answered using regression model analysis. The regression models established for the two experimental settings presented here demonstrate that polyfunctionality (q) is an important immune correlate of T cell efficacy. Of note, our mathematical model permits the identification of the better immune correlate of T cell efficacy between polyfunctionality and response size. Response size is here defined as functional T cells irrespective of the number of simultaneously produced effector molecules. Our in vitro and in vivo data unanimously demonstrate that polyfunctionality is a superior immune correlate as compared with response size. From the HIV-specific target killing dataset we fully delineated all parameter estimates including estimates of the impact of individual effector molecules (φ-values). Of note, no effector molecule impacted the polyfunctionality index immune correlate significantly different compared to other effector molecules. We recognize that the in vitro nature of our dataset could hampen the effect of certain effector molecules, such as chemo-attractants, whose effect would be better represented in in vivo studies. A recent theoretic study performed kinetic mathematical modelling of HIV pathogenicity, studying among several other parameters HIVspecific CD8 + T cell polyfunctionality. Of note, they demonstrate that T cell efficacy measured as the control of viral burden is correlated with T cell polyfunctionality. [19] Moreover, in concordance with our findings they demonstrate that T cell efficacy is mediated by the simultaneous production of effector molecules with no single effector molecule being more responsible than others.
Another recent study dissected the efficacy of CD8 + T cells in protecting against pulmonary Yersinia pestis infection, showing that TNF-α and IFN-γ effector molecules were crucial for in vivo protection, whereas perforin-driven cytotoxicity was not. [7] The previous study elegantly demonstrated the feasibility and importance of identifying the effector molecules contributing to T cell efficacy, however their use of mouse knock-out models to discern important versus non-important effector molecules is laborious and non-applicable to human studies. Our study provides an alternative, generalizable, non-laborious and low-cost methodology to discern the ranking of effector molecules according to their contribution to T cell efficacy both for animal and more importantly human studies.
In conclusion, we propose a methodology to identify robust parameter estimates rendering the polyfunctionality index an immune correlate of T cell efficacy. Of note, the parameter estimates has an associated biological interpretation, which estimates the importance of individual as well as combined effector molecule expression at the single-cell level. Our data demonstrate that polyfunctionality represents a significantly more effective immune correlate of T cell efficacy compared with individual effector functions. Furthermore, the approach outlined here also allows an identification of non-redundant effector molecules, which will allow a more cost-effective T cell analysis and thus improve research and future health care budgets.
Supporting Information S1 Fig. Percentage of Leishmania major-reactive CD4 + T cells (response size) is only weakly associated with the degree of protection induced by vaccination and non-associated with their polyfunctionality. Mice were vaccinated with Leishmania major antigens employing different antigen preparations (recombinant leishmanial polyprotein (MML) with adjuvant (CpG) and replication-defective adenovirus expressing MML), different doses and different routes of injection. Mice were challenged 28 days post vaccination by intradermal ear injections of Leishmania major carrying parasites. Leishmania major protection levels were measured as peak lesion size. Spleen derived lymphocytes harvested 28 days post vaccination were stimulated with anti-CD28 and MML for 6 hours and analysed for the production of IFN-γ, TNF-α and IL-2 by multiparametric flow cytometry. Experiments were repeated in four independent experiments A) Scatter plots show the association between normalized lesion size and Leishmania major-specific response size quantified as the frequency of CD4 + T cells producing cytokines upon stimulation with Leishmania major antigens. B) Scatter plots show the association between Leishmania major-specific CD4 + T cell polyfunctionality quantified as the polyfunctionality index (q = 1.33) and Leishmania major-specific response size. Statistical analysis was conducted with parametric aggregate correlation statistics. The first set of simulations tested whether φ 1 was significantly different from 0, given that the true value of φ 1 is 1. Using the same simulated conditions, but only modifying the true values of the other φ's, there appears to be no influence on power (all solid markers). When φ 1 is increased two-fold, we see that power increases substantially at all levels of N (all hollow markers). Experimental units (N) in this case are groups of mice (each group conposed of 3-4 mice) under the same experimental conditions, since these groups were within-averaged during power analysis. (TIF) S1 Table. Dynamic table with raw polyfunctionality data for HIV-specific CD8+ T cell clones and corresponding target killing capacity. The dynamic table automatically calculate polyfunctionality index values for each sample upon modification of q-and φ-values. Further explanation can be found in the abbreviation datasheet included. (XLSX) S2 Table. Dynamic table with raw polyfunctionality data for Leishmania major-specific CD4+ T cells and corresponding protection levels measured as peak lesion size post challenge. The dynamic table automatically calculate polyfunctionality index values for each sample upon modification of q-and φ-values. Further explanation can be found in the abbreviation datasheet included. (XLSX) | 2016-05-12T22:15:10.714Z | 2015-06-05T00:00:00.000 | {
"year": 2015,
"sha1": "94d534e1edb6402ebf590d7148c47f01bd0da7a5",
"oa_license": "CC0",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0128714&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "5214a0f47555839ee55a9536da43374d66901027",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
253949446 | pes2o/s2orc | v3-fos-license | Pathogenicity of Salmonella enterica subspecies enterica serovar Enteritidis phage type 1 in one-day-old specific pathogen-free chicks
Background: The studies about Salmonella infection in newly hatched chicks were not extensive. Aim: The objective of this study was to determine the pathogenicity of Salmonella enterica subspecies enterica serovar Enteritidis (SE) phage type (PT) 1 in one-day-old specific pathogen-free (SPF) chicks. Methods: Seventy, one-day-old SPF chicks, were divided into SE group (30 chicks), mortality group (10 chicks), both orally inoculated (1.0 ml) with SE PT1 (1 × 108 colony-forming unit per 1.0 ml), and one control group (30 chicks). The chicks were sacrificed at 6 and 12 hours, and days 1, 2, 3, 5, 7, 10, 14, and 21 post-inoculation (pi). Samples were collected for bacterial isolation, histological examination, and ultrastructural examination. Results: Starting from day 2 pi, the body weight in the SE group significantly (p < 0.05) decreased. The SE isolation percentages from the liver, spleen, mid-intestinal content, cecal content, cecal tonsil, blood, and cloacal swab were 0.73, 0.77, 0.33, 0.33, 0.36, 0.40, and 0.30, respectively. The isolation percentage in the liver was significantly (p < 0.05) higher than the blood and cloacal swab. The villi heights and crypt depths in the SE group were significantly (p < 0.05) greater and smaller, respectively. Ultrastructurally, erosion and necrosis were observed in the microvilli of the cecal tonsil. The bacteria were engulfed by macrophages at the interepithelial clefts of the M-like M cells. Conclusion: It was concluded that the inoculation of SE PT 1 in one-day-old chicks caused a systemic infection with diarrhea, a decrease in the body weight and villi height in the duodenum, jejunum, and ileum, and high bacterial loading in the liver with mild gross and histological lesions of organs, erosion, and necrosis of microvilli and low mortality. The bacteria entered the body system from the intestinal tract through the interepithelial clefts of the M-like M cells of the cecal tonsil.
Introduction
Salmonella is a causative agent of cross-infection between humans and animals (Ajmera and Shabbir, 2020). It is characterized as a Gram-negative, motile, hydrogen sulfide-producing, an acid-labile facultative intracellular bacterium (Ajmera and Shabbir, 2020). Globally, Salmonella enterica and Salmonella typhimurium represent the major causative agents of human Salmonellosis (Havelaar et al., 2015).
The typing of Salmonella is important for the identification of the bacteria, tracing, epidemiological investigation, monitoring antibody susceptibility, food safety, and public health measures. Despite the low resolution, serotyping and multilocus sequence typing (MLST) analyses are considered the standard methods for typing Salmonella (Yan et al., 2021). Based on the phage typing method, Salmonella enterica subspecies enterica serovar Enteritidis (SE) phage type (PT) 1 is a causative agent of human Salmonella infection in Sweden (Nygård et al., 2004), Canada (Khakhria et al., 1991), Italy (Vencia et al., 2015), and Finland (Lukinmaa et al., 1999). However, recently, a highresolution molecular typing method was established. It was core genome MLST. Such a method underwent whole genomic sequencing for accurate bacterial tracing (Yan et al., 2021). Enteric pathogens, in general, and Salmonella in particular, attach and invade the intestinal epithelium. The infection brings on some adverse consequences Submitted: 26/07/2022Accepted: 20/10/2022Published: 14/11/2022 to the intestinal function and structure comprising a reduction in the length of the epithelial villi and damage in the intestinal crypts (Bellido-Carreras et al., 2019). This infection initiates activation of the natural killer cells accompanied by intraepithelial T cells. Then, it leads to macrophage infiltration and activation of dendritic cells. Also, there is a proliferation of T cells in the spleen and antibody release in the serum (Meijerink et al., 2021).
The studies, about this infection, in newly hatched chicks, were not extensive (McWhorter and Chousalkar, 2019). Thus, the investigation of the infection, with light microscopy and scanning electron microscopy (SEM), would be expected to reveal some consequent influences on the function and architectural physique of the intestine. Moreover, it might uncover some details about the entry mechanism of the microbe and the defensive reaction of the host. The objective of this study was to determine the pathogenicity of SE PT 1 in one-day-old specific pathogen-free (SPF) chicks.
Salmonella enterica isolate
Salmonella enterica was isolated from the liver samples of commercial broiler chickens in Melaka, Malaysia. The method of isolation and identification was described by Kyung-Min et al. (2015). In brief, the samples were collected in Rappaport-Vassiliadis (RV) broth incubated at 37°C for 24 hours and then cultured on both xylose lysin deoxycholate (XLD) and brilliant green (BG) agar at 37°C for 24 hours. The suspected colonies of Salmonella were cultured on urease and triple sugar iron (TSI) biochemical test. Positive cultures from the biochemical test were then identified serologically by Salmonella poly-O antisera using slide agglutination test (SAT).
Phage typing
Positive SE colonies were characterized by phage typing at the Laboratory of Enteric Pathogens Center for Infections Institute, 61 Colindale Avenue, London, UK by the method described by Demczuk et al. (2003). A loopful of frozen nutrient broth was cultured on nutrient agar at 37°C for 24 hours. Positive colonies were incubated in blood agar (BA) culture at 37°C for 24 hours. The colonies from BA were inoculated into 3 ml of phage broth for 2 hours with vigorous shaking. The phage broth was then poured on a phage agar plate after the removal of excess broth. Next, bacteriophages were left overnight to dry at room temperature. The phage lysis pattern of cultures was evaluated according to Demczuk et al. (2003) scheme. It was identified that the three samples were SE PT 1. Serotyping Salmonella polyvalent O (A-I, Vi) antiserum was used at room temperature. Some bacterial growth was emulsified and examined under light microscopy. Clumping, of Salmonella polyvalent O antiserum represented a positive result. Positive colonies were tested with monovalent group D antiserum. Positive growth of group D antisera was tested with H antisera by adding 1 ml of solid nutrient agar. The agar was melted by boiling water, and incubated in a water bath for 30 minutes at 45°C. 10 μl of heterologous H antiserum was added, shaken, and left to solidify. Heavy suspensions of the bacterial growth and normal saline were streaked and incubated overnight at 37°C. Using a loop, a speck of the bacterial growth was emulsified with normal saline. Then, 10 μl of H antiserum was mixed. Clumping was considered a positive result for SE (Vandepitte et al., 2003). Inoculum preparation SE was cultured with nutrient broth on nutrient agar and incubated for 24 hours at 37°C. The cultures were transferred to BA, incubated for 24 hours at 37°C, harvested, and suspended in normal saline, and eventually an inoculum of 1 × 10 8 colony-forming unit (cfu) per 1.0 ml was prepared. SPF chicks SPF embryonated chick eggs were incubated in Siam Incubator System, Thailand, and allowed to hatch. One-day-old chicks were provided with water and an antibiotic-free ration at all times. Experimental design Seventy, one-day-old SPF chicks were divided into three groups, namely SE group (30 chicks), control group (30 chicks), and mortality group (10 chicks). Chicks in SE and mortality groups were infected orally with 1.0 ml of 1 × 10 8 cfu per 1.0 ml of SE while the chicks of the control group were not infected. Sampling was carried out on the SE group at hours 6 and 12, and on days 1, 2, 3, 5, 7, 10, 14, and 21 post-inoculation (pi). Samples of liver, spleen, mid-intestinal contents, cecum, cecal tonsils, and cloacal swabs were collected and incubated in RV broth for 24 hours at 37°C for SE isolation. Mortality group chicks were not sacrificed but left either to record the pathogenicity of the microbe or to cause death throughout the trial when the chicks did not undergo any disturbance. Clinical signs, body weights, and gross lesions were recorded. Samples of liver, spleen, duodenum, jejunum, ileum, and cecum were fixed in 10% buffered formalin, embedded in paraffin, and processed for histopathological examination. Cecal tonsils were fixed in 2.5% glutaraldehyde cacodylate buffer, pH 7.2 at 4°C for ultrastructural examination by SEM.
Bacterial isolation
A loopful of RV broth was cultured onto XLD and BG agar and incubated for 24 hours at 37°C. The positive black colonies (due to the production of hydrogen sulfide) and rounded pink colonies from BG agar were then cultured onto TSI and urease for 24 hours at 37°C. The positive results from TSI were then transferred to BA and incubated for 24 hours at 37°C. The positive http: //www.openveterinaryjournal.com S. Ahmad et al. Open Veterinary Journal, (2022), Vol. 12(6): 839-850 colonies were then tested by SAT using poly-O antisera (Kyung-Min et al., 2015).
Morphometric analysis by measurement of villi height and crypts depth
The intestinal segments from the duodenum, jejunum, and ileum samples were processed and stained with hematoxylin and eosin (HE). The tissues were examined under light microscopy (Motic BA410) connected with a camera (Motic Moticam pro 285A) for determination of the villi height and crypt depth using the software of motic images plus 2.0 ML. The villi heights and the crypt depths were measured from the tip to the crypt junction and from the base to the villi junction, respectively (Gava et al., 2015) (Fig. 1). Three readings were recorded for every collected sample.
Scanning electron microscopy
The cecal tonsil tissues were fixed in 2.5% glutaraldehyde cacodylate buffer for 12 hours. Then, the tissues were washed with 0.1 M sodium cacodylate buffer, fixed with osmium tetroxide for 2 hours at 4°C, and washed with 0.1 M sodium cacodylate buffer. The next step was dehydration with acetone and run for critical point drying. After that, the cecal tonsil tissues were mounted with double-sided tape or colloidal silver. Eventually, they were coated with gold sputter or coater samples and viewed under SEM (Mogana and Patchamuthu, 2016).
Statistical analysis
Using IBM Statistical Package for the Social Sciences Statistics 25 software, the technique of T-test (independent sample) was applied for the result of body weight. While for the result of SE isolation, when SE was isolated 1(3) (one positive sample out of three examined samples), 2(3), or 3(3), the percentages were considered 33%, 66%, or 100%, respectively. Hence, for the calculation, the values, of SE isolation, were considered 0.33, 0.66, or 1.00, respectively. Then, one-way analysis of variance (ANOVA) with repeated measures was applied. The level of significance was considered when p < 0.05.
Ethical approval
All experimental procedures were undertaken in accordance with the Research Policy on Animal Ethics of Universiti Putra Malaysia.
Body weight
The body weight of the chicks in the SE group was fluctuating at the beginning of the experiment. Starting from day 2 pi till the end, every time there was an increase. The body weight in the SE group was lower (p > 0.05) in comparison with the control group before day 2 pi. However, it was also significantly lower (p < 0.05) in comparison with the control group starting from day 2 pi till the end of the experiment ( Table 1). The body weight of the chicks in the control group increased continuously starting from hour 12 pi till the end of the experiment. Starting from day 5 pi till the end of the experiment, it was significantly The value of body weight was expressed as mean ± standard error mean (SEM). The letters in the superscript before comma (a and b) indicates that the value was significantly different from other value in the same row. The letters in the superscript after comma (c, d, e…h) indicate that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05.
Clinical signs
In the SE group, at hour 6 pi, one chick was affected with diarrhea, then diarrhea was observed in all chicks from hour 12 pi till the end of the experiment. From day 2 pi till the end of the experiment, the chicks were affected by inappetence. From day 3 pi till the end of the experiment, there was a sign of ruffled feathers. From day 5 pi till the end of the experiment, there was stunting. In the mortality group, all the chicks were affected with diarrhea from hour12 till the end of the experiment. From day 3 pi till the end of the experiment, there were inappetence, ruffled feathers, and stunting.
No abnormal clinical signs were observed in the control group.
Gross lesions
In the SE group, at hours 6, 12 pi, and day 1 pi, gross lesions were not observed. From day 2 pi till the end of the experiment, mild airsacculitis and mild peritonitis were recorded in all chicks. Also, from day 2 pi until day 21 pi, congestion, in the liver and kidney, was observed in some chicks. One chick died on day 3 pi and another chick died on day 5 pi. Dead chicks were affected by dehydration.
In the control group, no gross lesions were recorded.
Bacterial isolation Liver
In the SE group, SE was isolated at hours 6, 12 pi from the liver tissues (33%). On days 1, 2, and 3 pi, it was isolated 100%. On days 5 and 7 pi, it was isolated at 33%. On day 10 pi, it was isolated 100%. On days 14 and 21 pi, it was isolated at 33% (Table 2).
Blood
In the SE group, SE was isolated at hour 6 pi (33%). At hour 12 pi, it was isolated 66%. On days 1, 2, and 7 pi, it was isolated at 66%. On days 3, 5, and 10 pi, it was isolated at 33%. On days 14 and 21 pi, it was not isolated (Table 2).
Mid-intestinal content
In the SE group, SE was isolated at hour 6 pi (33%). At hour 12 pi, it was isolated 100%. On day 1 pi, it was isolated at 33%. On day 2 pi, it was isolated at 66%. On days 5, 7, and 10 pi, it was isolated at 33%. On day 3 pi, it was not isolated (Table 2).
Cecal content
In the SE group, SE was isolated at hour 6 pi (100%). At hour 12 pi, days 1 and 2 pi, it was isolated at 66%. On days 7 and 21 pi, it was isolated at 33%. It was not isolated on days 3, 5, 10, and 14 pi (Table 2).
Cloacal swab
In the SE group, SE was isolated at hours 6 and 12 pi (66%). On days 1 and 2 pi, it was isolated at 33%. On day 3 pi, it was isolated at 66%. On day 10 pi, it was (Table 2).
Statistical analysis for the result of isolation of SE from different tissues of chicks from hour 6 till day 21 pi
One-way ANOVA with repeated measures was applied on the result of the isolation of SE from different tissues of the chicks from hour 6 till day 21 pi to determine the overall mean and standard error of the mean for the isolation of SE 10 times in the liver, the overall mean for the isolation of SE 10 times in spleen, the overall mean for the isolation of SE 10 times in mid-intestinal content, and so on. There were significances (p < 0.05) for the differences obtained when the mean of the liver was compared with the mean of blood and also when the mean of the liver was compared with the mean of the cloacal swab (Tables 3 and 4).
Villi height Duodenum
The villi height in the duodenum of the SE group was rising throughout the experiment. But there was a drop on day 2 pi. The height was significantly (p < 0.05) greater than the height in the previous time of tissue collection of the same group at hour 6, days 1 and 2 pi. The villi height in the SE group was significantly (p < 0.05) smaller than the height in the control group from day 2 pi till the end with the exception of day 7 pi ( Table 5). The villi height in the control group was significantly (p < 0.05) greater than the height in the previous time at hour 6, days 1,3, 5, 7, 10, and 21 pi (Table 5).
Jejunum
The villi height in the jejunum of the SE group was rising every time apart from days 1 and 3 pi. The height was significantly (p < 0.05) greater than the height in the same group (SE) at hour 6 days 1 and 2 pi. The villi height in the control group increased every time apart from day 3 pi. The height was significantly (p < 0.05) greater than the height in the same group (control) at hour 6, days 1, 2, 3, 5, 10, and 21 pi. When the height value in the SE group was compared with the height in the control group, it was significantly (p < 0.05) smaller than the height in the control group on days 1, 2, 3, 5, and 10 pi (Table 6).
Ileum
The villi height in the ileum of the SE group was fluctuating, in other words, sometimes, it increased and sometimes it decreased in comparison with the measurement of the time before. However, only on day 2 pi, there was a significant (p < 0.05) increase. When the height value in the SE group was compared with the height in the control group, it was significantly (p < 0.05) greater in the control group only on days 1 and 10 pi. The villi height in the control group increased significantly (p < 0.05) more than the height in the previous time in the control group only at hour 6, days 2, 10, and 21 pi (Table 7).
Crypt depth Duodenum
The crypt depth in the duodenum of the SE group was significantly (p < 0.05) greater than the depth in the duodenum of the control group at hour 6 and day 7 pi. The crypt depth in the SE group was significantly greater than the previous value in the same group (SE) at hours 6, 12, days 2, 3, 10, 14, and 21 pi. The crypt The value of villi height was expressed as mean ± standard error mean (SEM). The letters in the superscript before comma were different (either a or b), which indicated that the value was significantly different from other values in the same row. When the letters in the superscript after the comma were different (c, d, e…j), it indicated that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05. The value of villi height was expressed as mean ± standard error mean (SEM). When the letters in the superscript before comma were different (either a or b), it indicated that the value was significantly different from other values in the same row. When the letters in the superscript after comma were different (c, d, e…j), it indicated that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05. The value of villi height was expressed as mean ± standard error mean (SEM). When the letters in the superscript before comma were different (either a or b), it indicated that the value was significantly different from other values in the same row. When the letters in the superscript after comma were different (c, d, e…g), it indicated that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05. The value of crypt depth was expressed as mean ± standard error mean (SEM). When the letters in the superscript before comma were different (either a or b), it indicated that the value was significantly different from other values in the same row. When the letters in the superscript after comma were different (c, d, e…g), it indicated that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05.
Jejunum
The value of the crypt depths in the jejunum of the SE group was significantly (p < 0.05) greater than the value in the control group on day 7 pi. The crypt depth value in the SE group was significantly (p < 0.05) greater than the previous value in the SE group at hour 6 and day 21 pi. The crypt depth value in the control group was significantly (p < 0.05) greater than the previous value in the control group at hour 6, days 1, 10, 14, and 21 pi (Table 9).
Ileum
The value of the crypt depths in the ileum of the SE group was significantly (p < 0.05) smaller than the value in the control group at hour 12 pi. The crypt depth value in the SE group was significantly (p < 0.05) greater than the previous value in the SE group at hours 6, 12, and day 3 pi. The crypt depth in the control group was significantly (p < 0.05) greater than the previous value in the control group at hour 6, and day 21 pi (Table 10).
Microscopic lesions
In the liver, there was mild heterophilic infiltration and fatty degeneration at hour 6 pi. Then at hour 12 pi, there was congestion. On days 1, 2, and 3 pi, there was mild to moderate congestion, fatty degeneration, and heterophilic infiltration. On days 5, 7, 10, 14, and 21 pi, there were necrosis, inflammatory cell infiltration, and congestion. In the spleen, there was congestion at hour 6 and 12 pi. Then on days 1 and 2 pi, there were heterophilic infiltration and congestion. On days 3, 5, and 7 pi, there were heterophilic infiltration, edema, and congestion. On days 10, 14, and 21 pi, there were some hyperplasia and edema.
In the bursa of Fabricius, at hours 6, 12, days 1, 2, and 3 pi, there were degeneration and congestion. From day 5 pi until the end of the experiment, there were congestion, necrosis, and thickening in the interfollicular area. In the duodenum, at hours 6 pi, there were mild heterophilic infiltration and congestion. At hour 12 pi, there were congestion and necrosis. On days 1, 2, 3, and 5 pi, there were more heterophilic infiltration, congestion, and necrosis. On days 7, 10, 14, and 21 pi, there were inflammatory cell infiltration and necrosis. In the jejunum, at hours 6 and 12 pi, there were heterophilic infiltration and congestion. On days 1, 2, and 3 pi, necrosis was observed. On days 5, 7, 10, and 21 pi, the same previous lesions persisted. In the ileum, at hours 6 and 12 pi, there was mild heterophilic infiltration. From day 1 pi till the end of the experiment, there were heterophilic infiltration, congestion, and necrosis. In the cecum, at hours 6 and 12 pi, there was mild heterophilic infiltration in the lamina propria. On days 1, 2, and 3 pi, some mild congestion and moderate heterophilic infiltration in the lamina propria were recorded. On days 5, 7, 10, 14, and 21 pi, necrosis was detected. On day 5 pi, moderated heterophilic infiltration, congestion, and necrosis in the lamina propria were detected. On days 7, 10, 14, and 21 pi, moderated heterophilic infiltration in the lamina propria and muscular layer beside necrosis of the epithelial cells. The value of crypt depth was expressed as mean ± standard error mean (SEM). When the letters in the superscript before comma were different (either a or b), it indicated that the value was significantly different from other values in the same row. When the letters in the superscript after comma were different (c, d, e…g), it indicated that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05. The value of crypt depth was expressed as mean ± standard error mean (SEM). When the letters in the superscript before comma were different (either a or b), it indicated that the value was significantly different from other values in the same row. When the letters in the superscript after comma were different (c, d, e…g), it indicated that the value was significantly different from the previous value in the same column. The level of significance was accepted when p < 0.05.
Dead chicks
On days 3 and 5 post-challenged, severe infiltration of inflammatory cells, congestion, and multiple areas of focal coagulative necrosis were observed in the hepatic tissue of the dead chicks. Also, increased lymphocytic activity was recorded in the spleen tissue. The intestine was severely autolyzed. The bursa of Fabricius tissue was congested and infiltrated by heterophils. The interfollicular tissue was thickened, whereas, the lymphoid follicles were invaded by focal areas of necrosis.
Scanning electron microscopy Hours 6 and 12 pi
There was some roughening in the microvilli of the cecal tonsils and macrophage engulfment of some individual or bacterial clusters near the interepithelial clefts.
Day 2 pi
The microvilli surface was affected by erosion and necrosis. Bacterial infiltration through the erosions and bacterial engulfment by the microphage were observed (Figs. 2-4).
Days 5, 7, and 21 pi
There were necrosis and sloughing of the microvilli of the cecal tonsil's surface. Some individual bacteria were detected approaching the M-like M cells of the microvilli of the cecal tonsils (Fig. 5).
Discussion
The study demonstrated that the oral inoculation of 1.0 ml of SE PT 1 with a dose of 1 × 10 8 cfu/1.0 ml in one-day-old SPF chicks resulted in a moderate infection. In addition, some morphometric alterations were detected in the villi height and crypt depth of the intestine. The technique for typing Salmonella in this study, phage typing, is a phenotypical technique that has been applied for a long time. It is a rapid and low-cost technique (Mohammed, 2017). Whereas the technique itself necessitates specialized phage collections and bacterial strains for propagation. Thus, it is solely carried out in a limited number of laboratories (Bertrand et al., 2012). It was reported that the use of temperate phage 1, from the phage typing conducted by Ward et al. (1987), resulted in the conversion of PTs 1-20 (Rankin and Platt, 1995). In general, such a technique is unstable due to its limited reproducibility (Bertrand et al., 2012). There were 2 (6%) dead chicks. However, in a study conducted by Lima et al. (2016), it was found that the mortality rate, in one-day-old chicks experimentally infected with different strains of SE, was 69.5%. Also, in an outbreak of Salmonella in many poultry farms in four states in Nigeria, the mortality rate was ranging from 40% to 80% (Mshelbwala et al., 2017). While, in a study, conducted by Fernández et al. (2001), oneday-old chicks were inoculated with SE PT 4, and the mortality rate was 23%. The huge difference in the mortality rate between the chicks in this study and the chicks in other cited studies could be attributed to the differences in the virulence of different SE PTs (Dhillon et al., 2001), and different serovars (Cheng et al., 2019). The body weight value, of the chicks in this study, was significantly (p < 0.05) smaller in the SE group than the body weight value in the control group from day 2 pi till the end of the experiment. On day 21 pi, the body weight in the control group reached a peak, which was greater by 39% than the body weight in the SE group. Such reduction in the body weight of the infected chicks might be due to profuse diarrhea. The significantly lower values of body weight in the SE group in comparison with the control group reflected the economic importance of SE infection with respect to commercial loss which could affect the farms' broilers. In a research conducted by Alisantosa et al. (2000), the body weights in 15 groups of one-day-old SPF chicks were recorded on 14, 21, and 28 days pi of SE PTs 4, 8, and 23. It was found that in the majority of the groups, the average body weights were significantly lower than the uninoculated controls. Also, Dhillon et al. (1999) and Fernández et al. (2001) found a decrease in the body weights with percentages ranging from 10% to 24% in Salmonella inoculated newly hatched chicks compared with uninoculated chicks. Since the surface area of the small intestine is significantly enhanced by the presence of villi and microvilli (Kiela and Ghishan, 2016). So, the intestinal function can be evaluated with the villi height, cell area, and cell mitosis (Yamauchi et al., 1996). However, the villi are covered with columnar epithelial cells or enterocytes at the tip. These are mostly absorptive cells, whereas crypt cells are generally regarded as secretory. Furthermore, most of the nutrient transport occurs in the small intestine (Kiela and Ghishan, 2016). In this study, the villi height in the duodenum of the control group started to increase and became greater significantly (p < 0.05) than the height of the villi in the SE group from day 2 pi till the end of the experiment apart from day 7 pi. The villi height in the jejunum of the control group was greater significantly (p < 0.05) than the height of the villi in the SE group on days 1, 2, 3, 5, and 10 pi. In the ileum, the villi heights were greater significantly (p < 0.05) on days 1 and 10 pi than the height of the villi in the SE group. The depth, of the crypt in the duodenum of the control group, was smaller significantly (p < 0.05) than the depth of the crypts in the duodenum of the SE group at hour 6 and day 7 pi. In the jejunum, the crypt depth, in the control group, was smaller significantly (p < 0.05) than the crypt depth of the SE group at hour 12 and day 7 pi. However, in the ileum, the crypt depth in the control group was also smaller significantly (p < 0.05) than the crypt depth in the ileum of the SE group at hours 6 and 12 pi. The results of villi height and crypt depth in this study indicated that the intestinal absorption in the SE group decreased. In contrast, the secretion of crypts in the SE group increased. However, in one previous study conducted by Rebollada-Merin et al. (2020), in which 7-day-old chicks were challenged with 3.3 × 10 5 cfu/ ml of the monophasic variant of colistin-resistant S. typhimurium, the duodenum villi heights were found slightly lower (539.81 μm) on day 7 of life, higher (896.95 and 1,208.45 μm) on days 14 and 21 of life, respectively. The values obtained in the study of Rebollada-Merin et al. (2020), were also smaller than the values in the control of the present study. In addition, in the study of Rebollada-Merin et al. (2020), the duodenum crypt depths were found greater (78.14, 93.91, and 128.09 μm) on days 7, 14, and 21 of life, respectively. However, greater values in the study of Rebollada-Merin et al. (2020), in comparison with this study, were expected since the chicks in their study were challenged too late in comparison with the chicks in this study. In this study, SE was not isolated from the blood on days 14 and 21 pi. It might be an indication of some clearance of the microbe from the blood. However, Buck et al. (2004) found that fimD mutant of SE PT 4 was present in the blood for 3 weeks after infection of laying hens (ISA Warren Brown) at the age of 19 weeks. Whereas, the wild-type parent strain of the microbe was cleared from the blood within the first 3 days. Using SEM, SE PT 1 was detected in the tissue of the cecal tonsil at hour 6 pi. The bacteria were engulfed by the macrophages at the interepithelial clefts of the M-like M cells of the cecal tonsil which might provide a suggestion about the route of entry of the bacteria from the intestine to the body system. The bacteria also caused some erosion and necrosis of the microvilli which definitely lead to the disturbance and decreased absorption of the feed materials and some reduction in the body weight of the chicks. The significant reduction in the body weight of the SE group in comparison with the control group might also be due to diarrhea and a decrease in villi heights of the duodenum, jejunum, and ileum. The result of SEM supported some previous studies that M cells were the target cells invaded by Salmonella typhi and S. typhimurium (Jepson and Clark, 2001). Another action, induced by Salmonella, was programmed cell death in the macrophage (Bruno et al., 2009). Furthermore, Huang et al. (2020) reported that Salmonella crossed the intestinal epithelium. Then it was engulfed by some phagocytic cells such as dendritic cells and macrophages. It was recommended in future studies to subtype the causative agent, SE PT 1, with a more sophisticated, stable, and accurate technique.
Conclusion
It was concluded that the inoculation of SE PT 1 in 1-dayold chicks caused a systemic infection with diarrhea and a decrease in body weight and villi height in the duodenum, jejunum, and ileum, and high bacterial loading in the liver with mild gross and histological lesions of organs, erosion, and necrosis of microvilli and low mortality. The bacteria were engulfed by macrophages at the interepithelial clefts of the M-like M cells of the cecal tonsil and entered the body system from the intestinal tract.
Data availability
The data used to support the findings of this study are included in the article.
Conflict of interest
No potential conflict of interest was reported by the authors. | 2022-11-26T16:07:35.023Z | 2022-11-14T00:00:00.000 | {
"year": 2022,
"sha1": "3d0f2b7cccb4d758e78ef0ece943ccc80bd7544c",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.5455/ovj.2022.v12.i6.8",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "55bd09623e9a5be578ad2cdbe1f1e991c5bbce60",
"s2fieldsofstudy": [
"Medicine",
"Agricultural And Food Sciences"
],
"extfieldsofstudy": [
"Medicine"
]
} |
119525001 | pes2o/s2orc | v3-fos-license | Hubble Space Telescope analysis of stellar populations within the globular cluster G1 (Mayall II) in M31
In this paper we present a multi-wavelength analysis of the complex stellar populations within the massive globular cluster Mayall II (G1), a satellite of the nearby Andromeda galaxy projected at a distance of 40 kpc. We used images collected with the Hubble Space Telescope in UV, blue and optical filters to explore the multiple stellar populations hosted by G1. The $m_{\rm F438W}$ versus $m_{\rm F438W}- m_{\rm F606W}$ colour-magnitude diagram shows a significant spread of the red giant branch, that divides $\sim 1$ mag brighter than the red clump. A possible explanation is the presence of two populations with different iron abundance or different C+N+O content, or different helium content, or a combination of the three causes. A similar red giant branch split is observed also for the Galactic globular cluster NGC6388. Our multi-wavelength analysis gives also the definitive proof that G1 hosts stars located on an extended blue horizontal branch. The horizontal branch of G1 exhibits similar morphology as those of NGC6388 and NGC6441, which host stellar populations with extreme helium abundance (Y>0.33). As a consequence, we suggest that G1 may also exhibit large star-to-star helium variations.
INTRODUCTION
The Hubble Space Telescope (HST) Ultraviolet Legacy Survey of Galactic Globular Clusters (GCs, GO-13297, PI: Piotto) has identified and characterised the multiple stellar populations (MPs) from homogeneous multi-band photometry of a large sample of 58 GCs. All GCs host two or more distinct populations and the complexity of the multiplepopulation phenomenon increases with the cluster mass. Specifically, the MPs of the most massive Galactic GCs exhibit extreme chemical compositions -such as large helium and light-(and in some cases also heavy-)elements variations among the cluster members -and their present population is dominated by second-generation stars born from matter polluted by stars belonging to the first generation; these properties appear more extreme than those commonly observed in less massive GCs of the Milky Way Milone et al. 2017a).
In the majority of the analysed GCs, MPs have uniform metallicity and are characterised only by different helium content and light-element abundance. However, some massive clusters also show internal variations in heavy elements, including iron and s-process elements (e.g. Marino et al. 2015;Johnson et al. 2015). In particular, ω Cen, which is the most-massive Milky-Way GC (M = 10 6.4 M ⊙ , McLaughlin & van der Marel 2005), exhibits the most-extreme variation of iron, with [Fe/H] Norris et al. 1996;Johnson & Pilachowski 2010;Marino et al. 2011b).
Understanding the properties of MPs in GCs may constrain the models of formation of the GCs in the early Universe, the mechanisms responsible for the assembly of the halo, and the role of GCs in the re-ionization of the Uni-verse (e.g., Renzini et al. 2015;Renzini 2017 and references therein).
Evidence of MPs in GCs belonging to external galaxies are widely reported in literature. Photometric detection of MPs in four GCs of the Fornax dSph have been reported by Larsen et al. (2014). Multiple sequences have also been observed in the CMDs of young and intermediate age clusters of the Small and Large Magellanic Clouds (see, e.g., Niederhofer et al. 2017a,b;Correnti et al. 2017;Milone et al. 2018a, 2017b andreferences therein), but the phenomenon that generates these multiple sequences may be different from that observed for old stellar systems. Spectroscopic detection of MPs in GCs of external galaxies have been reported, e.g., by Larsen et al. (2012, in the Fornax dSph ), Schiavon et al. (2013, in M 31), Mayya et al. (2013, in M81), Mucciarelli et al. (2009), Dalessandro et al. (2016, and Martocchia et al. (2018, in the Large and Small Magellanic Clouds).
In this context, the investigation of massive clusters in nearby galaxies is of great importance to understand whether the occurrence of MPs and the dependence on the mass of the host cluster are a peculiarity of Milky-Way GCs or are an universal phenomenon.
To address this issue, we investigate the presence of MPs in the GC Mayall II (G1, Sargent et al. 1977), in the nearby galaxy M 31, a cluster three times more massive than ω Cen (M = 10 7.2 M ⊙ , Meylan et al. 2001). G1 has been previously investigated by using Wide Field and Planetary Camera 2 photometry in the F555W and F814W bands (Rich et al. 1996;Meylan et al. 2001;Ma et al. 2007;Federici et al. 2012). Early evidence of metallicity variations is provided by Meylan et al. (2001) on the basis of the broadened RGB. It is noteworthy that G1 exhibits kinematic hints for a central intermediate mass black hole (Gebhardt et al. 2005).
In this work we present exquisite multi-band photometry of the stars hosted by G1, extracting their luminosities in six HST filters, from UV to optical bands (Section 2). We identify major features of the CMD of G1 never shown before and we analyse them to extract the main properties of the cluster stellar populations (Section 3). Finally, we compare the results obtained from G1 with those for massive Galactic GCs that present similar features in the CMD (Section 4), and we discuss them (Section 5) in the context of multiple stellar populations.
OBSERVATIONS AND DATA REDUCTION
For this work we reduced the archival HST data of G1 collected with the Wide Field and Planetary Camera 2 (WFPC2), with the High-Resolution Channnel (HRC) of the Advanced Camera for Surveys (ACS), and with the UVIS imager of the Wide Field Camera 3 (WFC3). WFPC2 data in F555W and F814W bands were collected within programs GO-5464 (PI: Rich) and GO-5907 (PI: Jablonka); observations in F555W with ACS/HRC were obtained within GO-9767 (PI: Gebhardt); WFC3/UVIS images in F275W, F336W, F438W, and F606W bands were taken for GO-12226 (PI: Rich). Table 1 gives the journal of observations used in this work. Figure 1 shows a stacked three-colour image of the G1 field.
To obtain high precision photometry, we perturbed library point spread functions 1 (PSFs) to extract, for each image, a spatial-and time-varying array of PSFs, as already done in other works by our group (see, e.g., Bellini et al. 2017;Milone et al. 2018a;Nardiello et al. 2018a,b). These PSF arrays have been used to extract positions and fluxes of the stars on the images. We corrected the positions using the geometric distortion solutions given by Anderson & King (2003, WFPC2), Anderson & King (2004, ACS/HRC), Bellini & Bedin (2009) and Bellini et al. (2011, WFC3/UVIS). We used the Gaia DR2 catalogue (Gaia Collaboration et al. 2016 to transform the positions of the single catalogues in a common reference frame. We used the FORTRAN routine kitchen_sync2 (KS2, for details see Anderson et al. 2008;Bellini et al. 2017;Nardiello et al. 2018b) that takes as input the PSF arrays, the transformations and the images, and analyses all the images simultaneously to find the position and measure the flux of each source, after subtracting its neighbours. This routine 1 http://www.stsci.edu/∼jayander/STDPSFs allows us to obtain high precision photometry in crowded environments and of stars in the faint magnitude regime. We calibrated the output photometry into the Vega-mag system using, for WFPC2, the zero-points and the aperture corrections tabulated by Holtzman et al. (1995), for ACS/HRC the zero-points given by the "ACS Zeropoints Calculator" 2 and the aperture corrections tabulated by Bohlin (2016), and for WFC3/UVIS the zero-points and the aperture corrections listed by Deustua et al. (2017).
To perform our analysis, we selected only well measured stars as in Nardiello et al. (2018a) on the basis of several diagnostic parameters output of the routine KS2 (see Nardiello et al. 2018b for details). The procedure is illustrated in Fig. 2 for the filter F438W, but it is the same for all the other filters. Briefly, we divided the distributions of the parameters RMS (the standard deviation of the mean magnitude divided √ N − 1 with N the number of measurements) and QFIT (the quality-of-fit) in intervals of 0.5 magnitude and, in each magnitude bin, we calculated the 3.5σ-clipped average of the parameter and then the point 3.5σ above the mean value of the parameter. We linearly interpolated these points (red line) and we excluded all the points above (in the case of the RMS) or below (for the QFIT) this line (azure points). Moreover, we excluded all the sources that have RADXS> 0.5 (where RADXS is the shape parameter, defined in Bedin et al. 2008) and RADXS< −0.5 (bottom left-hand panel of Fig. 2), and for which the number of measurements is N < 2. Middle-panel of Fig. 2 shows the m F606W versus m F438W − m F606W CMD for the stars that pass the selections in F438W and F606W filters, while the right-hand panel shows the rejected stars.
Artificial stars and completeness
We estimated the completeness of our G1 catalogue using artificial stars (AS). We produced 100 000 ASs having magnitude 21.75 ≤ m F606W ≤ 28.25, flat luminosity function in F606W, colours that lie along the red giant branch (RGB) and the horizontal branch (HB) fiducial lines, and flat spatial distribution. For each AS in the input list, the routine KS2 added the star to each image, searched and measured it using the same approach used for real stars, and then removed the source and passed to the next AS (see Anderson et al. 2008 for details). Using this approach, if an artificial star is added on a crowded zone of the image (like the cluster center), comparing the input and output photometry of the artificial star we can estimate how much the crowding is affecting the measurement of a star located in that position.
To estimate the level of completeness, we considered an AS as recovered if the difference between the input and output positions is < 0.5 pixel and the difference between the input and output m F606W and m F438W magnitudes is < 0.75 mag. Figure 3 illustrates the completeness of our catalogue: right panel shows the completeness contours in the m F606W versus radial distance plane. Our catalogue has completeness > 50% for stars with m F606W 26.5 at all radial distances from the cluster centre.
Field decontamination
Due to the large distance of G1 and the short baselines of the observations, using proper motions to separate G1 members from field stars is not possible. We decontaminated the cluster CMD from background and foreground stars using the statistical approach (see, for example, Milone et al. 2018a). The procedure is illustrated in Fig. 4. We defined two circular regions: a cluster field containing G1 and a reference field nearby on the same CCD frame, covering the same area of the cluster field. The radius of each circular region is equal to 33.5 arcsec (∼ 1.5 × r t , Ma et al. 2007). The stars of cluster and reference fields are showed in the m F606W versus m F438W − m F606W CMDs of top-left and top-right panels of Fig. 4, respectively. For each star i in the reference field, we calculated the distance where C F438W,F606W,cf and C i F438W,F606W,rf are the m F438W − m F606W colours of the stars in the cluster and reference fields, respectively. We flagged the closest star in the G1 field as a candidate to be subtracted. We associated to these flagged stars a random number 0 ≤ r i ≤ 1 and we subtracted from the cluster field CMD the candidates with r i < c i r f /c i c f , where c i r f and c i c f are the completeness of the star i in the reference field and of the closest star in the cluster field. Bottom panels of Fig. 4 show the result of the decontamination: left panel show the m F606W versus m F438W − m F606W CMD of cluster field stars after the decontamination, while right panel shows the stars subtracted from the cluster field CMD. Note the similarity of G1 and field. Field is dominated by M31 stars whose CMD is similar to G1.
THE CMD OF G1
Left panel of Fig. 3 shows the m F606W versus m F438W −m F606W CMD of all the well measured stars in the field containing G1.
In panel (a) of Fig. 5 we plot the decontaminated m F438W versus m F438W − m F606W CMD for all the well measured sources of G1. We highlighted possible asymptotic giant branch (AGB), RGB, and HB stars in orange, red, and green, respectively. In the following we analyse the peculiarities of these sequences and we compare the main properties of the CMD of G1 with what observed in other massive Galactic GCs.
The upper part of the RGB
A visual inspection of the bright part of the m F438W versus m F438W − m F606W CMD (panel (a) of Fig. 5) reveals that the upper part of the RGB sequence is broadened in colour. The broadening becomes evident if we compare this sequence with the average observational errors in colour (magenta crosses 3 ) in the same range of magnitudes. Also note that the fainter (and therefore affected by larger errors) blue HB is narrower than the brighter RGB, strengthening that the RGB broadening is intrinsic. This broadening was already observed by Meylan et al. (2001), and was associated to a spread in metallicity among RGB stars.
Panel (b 1 ) of Fig. 5 shows that the upper part of the RGB of G1 splits in two sequences. We verticalized the RGB using the same procedure adopted by Bellini et al. (2013, see also Milone et al. 2009 for details). As a result of the verticalization process, for each star we have a transformed colour and magnitude in a new plane 'ordinate' versus 'abscissa' (panel (b 2 )). The distribution of the transformed colours of the RGB stars is shown in panel (b 3 ): the distribution is bimodal, with two clear peaks. We fitted the histogram with a bi-Gaussian function to derive the fraction of stars belonging to the two sequences. We found that the redder RGB (named RGB a) contains 32±2 % of the RGB stars, while the RGB b, that is on the blue side of the RGB, contains 68±2 % of the RGB stars. In Fig. 6 we demonstrate that the bimodality of the upper part of the RGB is a real feature: panel (a) shows the verticalized RGB sequence as obtained in panels (b) of Fig. 5. We divided the verticalized sequence in two groups, on the basis of the gap in the RGB: the stars having 'abscissa'< 0.5 (in blue, RGB b) and the stars with 'abscissa'> 0.5 (in red, RGB a). To investigate if the separation of the two sequences is real, we used the approach adopted in Nardiello et al. (2015Nardiello et al. ( , 2018a: if the colour spread is due to photometric errors, a star of the RGB a sequence that is redder than the RGB b sequence in the m F438W versus m F438W − m F606W CMD (panel (b) of Fig. 6), should have the same probability of being either red or blue in CMDs obtained with other combinations of filters. The fact that in the totally independent m F814W versus m F555W − m F814W CMD (panel (c) of Fig. 6) RGB a and RGB b form two well-defined sequences demonstrates that the stars belonging to these two groups have indeed different properties.
Metallicity of G1 and spread of the RGB
In literature there are many estimates of the metallicity for G1, obtained using photometry (see, e.g., Stephens et al. 2001;Meylan et al. 2001;Bellazzini et al. 2003;Rich et al. 2005), integrated spectra (e.g., Huchra et al. 1991), and spectro-photometric Lick indices (Galleti et al. 2009). Huchra et al. (1991) As already discussed in the previous section, the RGB of G1 is strongly widened. Meylan et al. (2001) attributed this spread to variations of [Fe/H] among RGB stars of G1. In this section we investigate if this hypothesis is valid, comparing the observed split of the upper part of the RGB with theoretical models.
The procedure adopted is illustrated in Fig. 7. First, we derived for the observed RGB a and RGB b the fiducial lines. We derived the fiducial lines in the m F814W versus m X − m F814W CMDs, where X=F438W,F555W,F606W. The procedure to obtain a fiducial line is based on the naive estimator: we divided the RGB a (RGB b) sequence in intervals of δ = 1 F814W magnitudes. On these intervals we defined a grid of N points separated by steps of width δ/10 and we calculated the median colour and magnitude within the interval m i F814W < m F814W < m i F814W + δ, with i = 1, ..., N; we interpolated these median points with a spline. An example of fiducial lines is shown in panel (a) of Fig. 7, where the red line is for RGB a and the blue line is for RGB b.
We obtained synthetic fiducial lines as follows: we considered a 12 Gyr α-enhanced BaSTI isochrone (Pietrinferni et al. 2004(Pietrinferni et al. , 2006(Pietrinferni et al. , 2009) with a given [Fe/H], and we obtained a first guess synthetic CMD interpolating the isochrone on a vector of random m F814W magnitudes, assuming a flat luminosity distribution. We broadened this synthetic CMD by adding to the colour of each synthetic star a random Gaussian noise, with a dispersion equal to the average colour error measured for upper RGB stars. Each syn- thetic CMD is composed by 10 000 synthetic stars. For each synthetic CMD we extracted synthetic fiducial lines using the same approach adopted for the observed CMD. Panel (b) of Fig. 7 For each m cut F814W , we compare δ(X − F814W) obs and δ(X − F814W) synt to search for the best ∆[Fe/H] that reproduce the observed colour difference in all the bands. Panel (c) of Fig. 7 shows, coloured from blue to red, the colour differences obtained from the models for m cut F814W = 22.0; black stars are the observed points at the same magnitude level. In Table 2 we listed the ∆[Fe/H] obtained for each m cut F814W . We computed the weighted average of all the ∆[Fe/H], using as weight 1/σ ∆[Fe/H] , and we did so for three different cases: (i) the two populations have the same helium and C+N+O abundance (Scenario 1); (ii) the two populations share the same C+N+O content, but RGB a is populated by stars having primordial helium, while RGB b stars have Y = 0.30 (Scenario 2); (iii) the two populations have the same helium content, but RGB b is populated by stars C+N+O-enhanced (Scenario 3). In the latter case, the C+N+O content is about a factor of two larger than the CNO sum of the mixture used for the standard α-enhanced case (Pietrinferni et al. 2009). Because of the uncertainties on the bolometric corrections in blue and UV bands for the isochrones C+N+Oenhanced, for Scenario 3 we excluded the δ(F438W − F814W) point in the computation of ∆[Fe/H]. For the Scenario 1 we found that the split of the RGB is well reproduced with a mean metallicity difference of ∆[Fe/H]= 0.15 ± 0.03, while in case of Scenario 2 and 3 we found ∆[Fe/H]= 0.12 ± 0.03 and ∆[Fe/H]= 0.14 ± 0.04, respectively.
We want to highlight that a similar behaviour of the RGB could also be reproduced by a combination of variations of metallicity, helium, and C+N+O content (see discussion in Sect. 4.1).
The HB of G1
We know that the metallicity is the parameter that mainly determine the morphology of the HB of a GC. However, some Galactic GCs, despite they have the same metallicity, present different HB morphologies. This is the so-called "2nd-parameter" problem. Many parameters have been suggested as candidates to explain the peculiarities observed for some Galactic GCs. Recently, Milone et al. (2014) have demonstrated that the extension of the HB correlates with the mass of the hosting GC. Milone et al. (2018b) show that the morphology of the HB is also correlated with the internal helium variations among the different populations hosted by GCs, demonstrating that HB morphology is strictly related to the presence of multiple populations (see also, e.g., D'Antona et al. 2002;Dalessandro et al. 2011Dalessandro et al. , 2013. The HB morphology of M31 GCs and the effects of the second parameter have been analysed in many works in literature (see, e.g., Rich et al. 2005;Perina et al. 2012).
Already Rich et al. (1996) discovered that the HB of G1 is very populated on the red side, and that a few stars popu-late the blue side (see also Rich et al. 2013). In this paper we confirm that the HB of G1 is composed by a very populated red clump and we show for the first time the evidence of a very extended blue HB (green points in panel (a) of Fig. 5), more than expected and observed until now, even though it is an intermediate metallicity GC ([Fe/H]∼ −0.95).
Taking into account of the completeness of our sample, we found that the red clump and the blue side of HB of G1 contain ∼ 85% and ∼ 15% of HB stars, respectively. It has been demonstrated that Galactic GCs that show HB morphologies similar to that of G1 also host a small fraction of very high helium-enhanced stars (see, e.g., Caloi & D'Antona 2007;Busso et al. 2007;Bellini et al. 2013;Milone et al. 2014;Tailo et al. 2017) that populate the blue HB. We suggest that also G1 might host a small fraction (∼ 15 %) of highly helium-enhanced stars, that evolve into the blue side of the HB.
COMPARISON BETWEEN G1 AND GALACTIC GCS
In this section we compare the massive GC G1 in M31 with Galactic GCs of similar mass. We compared G1 with two Galactic GCs, NGC 6388, and NGC 6441, using the HUGS 6 catalogues published by Nardiello et al. (2018b).
G1 versus NGC 6388
The GC NGC 6388 is a massive, metal-rich ( As already demonstrated by Bellini et al. (2013), the upper part of the RGB of NGC 6388 splits in two sequences in the m F438W versus m F438W − m F606W CMD. Panels (b) of Fig. 8 shows the verticalized RGB sequence (bottom) and the histogram of the "abscissa" distribution (top) for NGC 6388, obtained as in Sect. 3. By least-squares fitting a bi-Gaussian function, we found that the RGB on the red side contains 42 ± 2 % of RGB stars, while the remaining 58 ± 3 % stars are on the blue side. This result is very similar to what we found for G1.
In Sect. 3.2 we have tested the hypothesis that the RGB split observed for G1 is due to a metallicity variation of ∼ 0.15 dex between the two populations. For NGC 6388, Carretta et al. (2007) found no intrinsic metallicity spread from spectroscopic measurements of RGB stars. A possible explanation of the RGB split could be a variation of the C+N+O content between the two populations, that implies also an SGB split in the optical CMDs. The SGB split for NGC 6388 was noticed by Bellini et al. (2013). Unfortunately, on the dataset used in this work, we are not able to distinguish between metallicity or C+N+O variations as cause of the RGB split observed for G1, because the SGB stars of G1 are too faint to be measured.
The HB morphologies of G1 and NGC 6388 are also very similar (in green in Figs. 5 and 8). Both HBs present a populated red clump, containing ∼ 85-90 % of HB stars and well spread in magnitude. Tailo et al. (2017) showed that the red side of the HB of NGC 6388 is compatible with the presence of He-enriched stars (∆Y ∼ 0.08).
The extended blue HB of G1 is very similar to that of NGC 6388. In the case of NGC 6388, Busso et al. (2007) and Tailo et al. (2017) showed that stars on the blue side of the HB of NGC 6388 are very helium-enriched, up to ∆Y ∼ 0.13 for the bluest ones.
There are few differences between the two clusters that must be taken into account in this analysis. First of all, NGC 6388 is more metal-rich ([Fe/H]=−0.45 ± 0.04, Carretta et al. 2009) than G1 ([Fe/H]∼ −0.95). Moreover, the present-day mass of NGC 6388 (∼ 2 × 10 6 M ⊙ , Boyles et al. 2011), is almost one order of magnitude smaller than G1 mass (∼ 1.5 × 10 7 M ⊙ Meylan et al. 2001). Finally, the position of the two clusters in the host galaxy is totally different: NGC 6388 is a bulge Galactic GC, while G1 is projected at 40 kpc in the halo of M31.
Multi-band comparison of the HB morphologies
The dataset used in this work allows us to analyse the extended HB of G1 from the near UV to the visible.
To compare the HB morphology of G1 with that of other Galactic GCs, for each m F606W versus c X = m X − m F606W CMD, where X=F275W, F336W, and F438W, we performed a translation of the entire CMD as follows: we calculated the 96th percentile (on the red side) of the HB colour distribution, c 96th,X , and the associated magnitude, m 96th,F606W Then, we defined a new colour c * X = c X − c 96th,X and magnitude m * F606W = m F606W − m 96th,F606W . To compare the HBs of G1 and of the Galactic GCs NGC 6388 and NGC 6441, avoiding the very low completeness of the G1 catalogue in the center region, we selected only the HB stars between 2.5 × r h and 4.5 × r h from the cluster centers, where r h is the half-light radius (r h =1.7, 31.2, and 34.2 arcsec for G1, NGC 6388, and NGC 6441, respectively, Harris 1996;Ma et al. 2007).
The first row of Fig. 9 shows the multiband photometry of the HB of G1 (green points). In the left panel we also report the completeness level. We compared the HB morphology of G1 with that of two Galactic GCs that are massive ( 10 6 M ⊙ ), that host stars highly enriched in helium (Y ∼ 0.30 − 0.35) and that have a populated red clump: NGC 6388 (2nd row of Fig. 9) and NGC 6441 (3rd row of Fig. 9). We highlighted in green their HB stars.
The GCs NGC 6388 and NGC 6441 are [Fe/H]-rich clusters (∼ −0.45; Carretta et al. 2009) located in the bulge of the Milky Way. Both clusters show a peculiar HB morphology: a very populated red clump (85-90% of HB stars) and an extended blue HB (Rich et al. 1997). Even if these two clusters seem to be twins, the chromosome maps of RGB stars (Milone et al. 2017a) are very different.
A comparison between the first, the second and the third row of Fig. 9 illustrates how the HB morphology of G1 is similar to that of the two selected Galactic GCs in all the analysed filters: the populated and extended red clumps, and the blue HB that covers about the same colour ranges are common features to these three GCs.
We know that both the red clumps of NGC 6388 and of NGC 6441 host stars with primordial helium content and helium enhancement by ∆Y ∼ 0.08 − 0.1, while the blue HB is populated by stars that have helium content Y > 0.35 (Tailo et al. 2017). It is possible that a similar helium enhancement might plausibly explain the extended HB of G1.
The GCs NGC 6388 and NGC 6441 host stars that form an extreme HB (eHB, Bellini et al. 2013;Brown et al. 2016, see also Fig. 8). These stars are located mainly in the center of the clusters. The cluster G1 does not show eHB stars: we attribute the lack of such stars to the extreme crowding and consequent low completeness at these magnitudes in the central regions of the cluster.
SUMMARY
In this work we presented a multi-band analysis of the CMDs of the GC Mayall II (G1) located in the halo of M31. We used HST data collected with filters that cover wavelengths from ∼ 250 nm to ∼ 950 nm. This is the first time that G1 is observed in UV and blue filters.
The high accuracy photometry extracted from HST data with new, advanced tools, has allowed us to extract for the first time a deep CMD (magnitude limits V ∼ 28.5, ∼ 3 magnitudes below the HB level) for this cluster. The m F606W versus m F438W − m F606W CMD shows new features never observed until now: a wide, likely bifurcated RGB, an extended blue HB, and populated RGB bump, AGB and red clump.
We show that the upper part of the RGB of G1 splits in two real sequences in the m F438W versus m F438W − m F606W CMD, as for the Galactic GC NGC 6388. We derive the fraction of stars within each sequence and we found the the redder sequence (RGB a) includes ∼ 32 % of RGB stars, while the bluer RGB b sequence contains the remaining ∼ 68 %.
We tested the hypothesis of [Fe/H]-variations to explain the RGB split, comparing the observed fiducial lines of the two populations with synthetic fiducial lines obtained using models. We found that the RGB split could be repro-duced assuming that the two populations have a difference in [Fe/H] of ∆[Fe/H]= 0.15± 0.03 if the populations share the same helium and C+N+O content, ∆[Fe/H]= 0.12 ± 0.03 if the two populations have different helium content and the same C+N+O content, and ∆[Fe/H]= 0.14 ± 0.04 if the two populations share the same helium abundance but have different C+N+O content.
We compared the HB morphology of G1 with that of massive Galactic GCs that present a populated red clump. We found many similarities with the HBs of NGC 6388 and NGC 6441: a populated red clump (> 85 % of HB stars) and an extended blue HB, despite these GCs are [Fe/H]rich. We know that the red clump and the blue HB of NGC 6388 and NGC 6441 are populated by stars helium enriched by ∆Y ∼ 0.08 and ∆Y ∼ 0.13, respectively. Given the similarity between the HB morphologies of G1 and NGC 6388/NGC 6441, we can conclude that also HB stars of G1 may be helium enriched. An additional proof is given by the correlation found by Milone et al. (2018b) for Galactic GCs between the maximum internal helium variation and the mass of the cluster: more massive clusters have larger internal maximum helium variation. Because G1 is ∼ 3 times more massive than ω Cen, the most massive GC in the Milky Way (∼ 4 × 10 6 M ⊙ , D'Souza & Rix 2013), we expected that the maximum internal helium variation among G1 stars is high, with ∆Y 0.1.
All the results obtained in this work represent a proof that G1 hosts multiple stellar populations, characterised by different chemical properties. Moreover, some evidence tell us that G1 could belong to the group of "anomalous" GCs, that host populations with different C+N+O contents and/or [Fe/H]-content (as already suggested by Meylan et al. 2001). In the Milky Way a significant fraction of globular clusters belongs to the group of "anomalous" GCs (ω Cen, M22, NGC 5286, M54, M19, M2, NGC 6934, etc., see, e.g., Norris et al. 1996;Carretta et al. 2010;Marino et al. 2011a;Yong et al. 2014;Marino et al. 2015;Johnson et al. 2015Johnson et al. , 2017Marino et al. 2018), and the discovery of a GC in M31 with similar characteristics means that these kind of clusters are not a prerogative of our Galaxy. Unfortunately, the dataset used in this work is not sufficient to constraint the chemical properties of this GC and of its populations, and more multi-band photometric observations are required to trace the formation and evolution of G1. | 2019-03-01T19:00:02.000Z | 2019-03-01T00:00:00.000 | {
"year": 2019,
"sha1": "44590495e928fd2f972b88fce0969a27c7135d8f",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1903.00488",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "44590495e928fd2f972b88fce0969a27c7135d8f",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
235695259 | pes2o/s2orc | v3-fos-license | Protective Effect of Quercetin Nanoemulsion on 5-Fluorouracil-Induced Oral Mucositis in Mice
The target of this study was to evaluate the efficacy, histopathological, oxidative stress, and molecular effects of quercetin (QRC) in mice with oral mucositis induced by 5-fluorouracil (5-FU). Thirty-six albino male mice with oral mucositis induced by 5-FU as a chemotherapeutic agent were used in this study. The animals were randomly divided into 6 groups: control group, mucositis (MUC) group, pretreatment group, posttreatment group, and two last groups including nanoemulsion form of QRC with a dose of 5 mg/kg in both pretreatment and posttreatment. In the present evaluation, fewer oral lesions were observed in the QRC posttreatment groups compared to the pretreatment and nanoemulsion receiving groups. In the SOD assay, the most significant difference was observed in the posttreatment nanogroup (41.073 ± 1.24) and pretreatment nanogroup (43.453 ± 2.60) in comparison to the 5-FU group (30.897 ± 1.93). The results of CAT assay also showed a significant difference in nano-posttreatment (124.60 ± 10.85), posttreatment (135.4 ± 9.82), and nano-pretreatment groups (128.80 ± 7.20) compared to the 5-FU group (55.07 ± 8.91). The expression of inflammatory genes such as Hif-1α and NfκB in this group was lower than in the other groups, although this difference was not significant. It seems that the use of QRC can improve the treatment process of oral mucositis induced by 5-FU.
Introduction
ere are several drugs available for cancer chemotherapy that can be used alone or in combination with other agents to treat a wide range of malignancies. Although chemotherapy is an efficient process to treat many types of cancer, it usually has toxic side effects depending on the type and dose of the drug. Some side effects of chemotherapy are mild and treatable such as nausea and hair loss, while others can cause serious side effects like infection [1]. e basic mechanism of action of anticancer agents is working on cells with a high rate of division. Based on this mechanism of action, other normal cells with a high rate of division, such as the gastrointestinal epithelium, are affected too. 5-Fluorouracil (5-FU) belongs to the category of anticancer and antimetabolite drugs. 5-FU inhibits DNA synthesis during the S phase of the cell cycle by limiting the availability of thymidylate [2]. It is used alone or in combination with other common drugs in the treatment of various cancers, including breast, head and neck, anal, stomach, colorectal, and some skin cancers [3][4][5]. e mechanism of action of 5-FU is associated with interference with DNA synthase and inhibition of thymidylate synthase. 5-FU, like other anticancer drugs, has many side effects that reduce the patient's quality of life by causing various destructive effects and sometimes interrupt treatment [6][7][8].
e most common side effects are nausea, diarrhea, vomiting, oral and intestinal mucositis, mouth ulcers, loss of appetite, light sensitivity, metallic taste, neutropenia, and thrombocytopenia [7]. Oral mucositis (OM) is an important and common side effect of chemotherapy with 5-FU treatment. Its incidence is 40% in chemotherapy and almost 100% in combination with chemotherapy and radiotherapy [9][10][11]. Chemotherapyinduced OM with anticancer drugs may initiate swelling, erythema, or ulcers. It can also involve a wide range of changes, from mild burning sensitization to painful wounds. OM symptoms consist of sleep and eating disorders, communication barriers, and severe pain, which can affect patients' quality of life [12]. Also, these symptoms can even interrupt the course of treatment [13,14]. ese lesions start when the oral mucosa is exposed to chemical agents, causing DNA destruction and cell death. ey are commonly caused by the production of reactive oxygen species (ROS) and oxidative stress [15]. Various studies have shown that the origin and extent of these lesions can be controlled by using antioxidants. Antioxidants protect cells from oxidative stress damage by neutralizing the ROS and preventing their formation [16][17][18].
Preventing or improving OM can increase the patient's quality of life and uninterrupted therapeutic regimen. Currently, the most important methods used to improve OM include cryotherapy, prescribing nonsteroidal antiinflammatory drugs (NSAIDs), using mouthwash to disinfect the mouth brushing [12,19] local anesthesia, such as diphenhydramine and using sodium carbonate mixture of promethazine and milk of manganese [20,21]. ese methods do not seem to be effective enough [22]. Anticancer drugs cause OM by two main mechanisms. In addition to directly damaging the mucosa, these drugs also suppress the immune system and predispose to bacterial infections [23]. Anticancer drugs can reduce the secretion and function of mucosa and mucosal cells [24]. On the other hand, using NSAIDs to alleviate adverse reactions of chemotherapeutic agents is problematic and cannot be a good solution to treat and improve mucositis caused by cancer chemotherapy [25].
Quercetin (3,3′,4′,5,7-pentahydroxyflavone) (QRC) is a naturally occurring flavonoid found in large quantities in the diet and its main sources are tea, apples, red wine, onions, broccoli, kale, oranges, and blueberries ( Figure 1) [26]. e most important property of this flavonoid is its antioxidant effect, but its oral bioavailability is very low. is flavonoid can prevent cell death and oxidative damage by several mechanisms, such as inhibiting the activity of oxygen free radicals, lipid peroxidation, inhibiting xanthene oxidase, and chelating metal ions [27]. It is also known as an antiallergic, anti-inflammatory, and antiviral compound [28]. QRC can prevent the invasion of malignant tumors of the prostate, liver, lung, breast, colon, and cervix. QRC is easily metabolized by the tyrosinase enzyme to different metabolites which can enhance anticancer activity [29]. Recently, it has been reported that QRC in combination with cisplatin, has synergistic effects in cancer treatment [30]. Also, QRC was used in a double-blind, randomized, placebo-controlled clinical trial to prevent and treat OM due to chemotherapy [31]. QRC nanoparticles have been used in many studies to increase the effectiveness of chemotherapy drugs and improve cancer treatment. It has been reported that QRC nanoparticles can reduce injuries caused by intestinal mucositis caused by methotrexate [32]. In another study, QRC nanoparticles were used to improve abnormal cell growth in breast cancer. e results showed that treating with QRC nanoparticles can result in cell death or preventing from cell proliferation in breast cancer [33]. In another study, silica nanoparticles were used to load QRC and doxorubicin nanoparticles, which increased the quality of chemotherapy in gastric cancer [34]. Based on studies and according to the fact that nanoparticles have a large surface area despite their low weight, they can be more active than other molecules; therefore, we used QRC nanoemulsion to improve OM caused by chemotherapy with 5-FU.
Nanoemulsion Preparation.
A nanoemulsion of QRC formed spontaneously in an oil phase of GMO, PEG-40 hydrogenated castor oil, and PEG 400 (1 : 8 : 1) [35]. 50 mg QRC was added to 10 grams of oil phase. QRC, oil, surfactant, and cosurfactant were stirred at 100 rpm for 2 h. Further sonication for 1 h using a bath sonicator (Elmasonic Med 60) was applied to complete the mixing process. To obtain nanoemulsion, deionized water was added to the oil phase at a ratio of 5 : 1 and stirred gently.
Animal Model and Treatment.
irty-six adult male albino mice weighing around 25-30 g were accommodated in polyvinyl cages (six mice per cage) at controlled and standard temperature of 22°C ± 3°C with 12-hour light/dark cycle. A standard diet and water were also provided. All experiments and procedures of the study were approved and authorized by the Animal Ethics Committee of the National Institute of Medical Research Development, Tehran, Iran Oral mucositis Prevention and management (i) Figure 1: Quercetin is a flavonoid that is abundant in some of our food products that, in addition to treating and preventing cancer, can also be useful in reducing the side effects of chemotherapy such as oral mucositis. 2 Journal of Oncology (IR.NIMAD.REC.1398.120).
e animals were randomly allocated into six groups: (i) Control group (saline + without mucositis) (ii) Mucositis group (IP injection of a single dose of 300 mg/kg 5-FU in the 6 th day) [36] (iii) Pretreatment group (supplementation with nano-QRC with a dose of 5 mg/kg before the induction of the disease-day 2 to day 6) (iv) Pretreatment group (supplementation with QRC with a dose of 5 mg/kg before the induction of the disease-day 2 to day 6) (v) Posttreatment group (mucositis + supplementation with nano-QRC with a dose of 5 mg/kg after the induction of the mucositis-day 7 to day 13) (vi) Posttreatment group (mucositis + supplementation with QRC with a dose of 5 mg/kg after the induction of the mucositis-day 7 to day 13)
Drug-Induced Mucositis Model.
All the experimental mice were kept in the animal house for one day without any treatment to adapt to the environmental conditions. From the second to the sixth day, QRC 5 mg/kg was given intraperitoneally (IP) to pretreatment groups until the 6 th day, and on the 6 th day, all groups except for the control received a single intraperitoneal (IP) dose 300 mg/kg of 5-FU. After the sixth day of treatment, the pretreatment and posttreatment animals received 5 mg/kg of QRC until the 13 th day [37]. During this period, the rate of improvement of the OM of the animals with the scoring method was noted. On the 13 th day, the animals were anesthetized with chloroform and their tongues were removed for histological and molecular studies. e severity of inflammation of the oral mucosa and the rate of defecation from the fourth day after induction were assessed by 4 different stages based on previous studies ( Figure 2).
Oxidative Stress Measurement
2.5.1. Malondialdehyde (MDA). Six ml of whole blood from each mice was collected. After clot formation, it was centrifuged at 3000 × g for 10 minutes at 4°C and the serum was separated based on the protocols in the kit. e absorptions of MDA in the plate were read at 530-540 nm by ELISA-reader (TebPazhohanRazi, Tehran, Iran).
Catalase (CAT).
e blood sample was allowed to clot for 10 minutes at 25°C. In order to separate the serum, it was centrifuged at 3000 × g for 10 minutes at 4°C. According to the protocols, the CATactivity was read at 540 nm by ELISAreader (TebPazhohanRazi, Tehran, Iran).
Superoxide Dismutase (SOD).
e blood sample was allowed to clot for 10 minutes at 25°C. In order to separate the serum, it was centrifuged at 3000 × g for 10 minutes at 4°C. According to the protocols, the SOD activity was read at 440-460 nm wavelength by ELISA-reader (TebPazho-hanRazi, Tehran, Iran).
Macroscopic and Microscopic Histopathological Analysis.
For macroscopic analysis, the changes on the tongue during 4 days were examined. e severity of the injury in a completely blind manner for macroscopic analysis, erythema, vasodilatation, erosion, epithelial ulcerations was evaluated based on a scoring way as follows ( Figure 3) [38]: (i) Score 0: totally healthy without damage with no erosion or vasodilatation in the surface area. (ii) Score 1: erythema is present; however, there is no sign of surface erosion (iii) Score 1.5: existing severe erythema, surface erosion, and presence of vasodilation (iv) Score 2: focal ulcers in one or more faces of the mucosa are observed, but not exceeding 25% of the surface area, severe erythema and vasodilatation (v) Score 2.5: accumulative ulcers can be noticed of about 50% of the surface area (vi) Score 3: accumulative ulcers can be noticed of about 75% of the surface area Tissue staining was resolved to consume the hematoxylin and eosin (H&E) experiment. is was accounted for the percentage of H&E-positive cells divided by the total number of cells. Samples were fixed in 10% buffered formalin, fixed in paraffin, and sectioned (in the size of 5 μm). Hematoxylin and eosin were applied for staining. USA). All the RNA samples were transferred to −80°C until further experimentation.
cDNA Synthesis and Real-Time PCR.
Pars Tous cDNA synthesis kit was applied for cDNA synthesis; the mixture was including 250 ng RNA, 5 µl enzyme buffer 2x, 1 µl of reverse transcriptase enzyme, and the mixture reached the 10 µl volume with DEPC treated water. e thermal program was performed using FleXCycler2 by incubating the reaction mixture 10 minutes at 25°C for random hexamer primer annealing, 60 minutes at 47°C for reverse transcriptase reaction, and 5 minutes at 85°C for ending the reaction. In this study, the SYBR Green method was used for real-time PCR assay. e mRNA amplification was performed by ABI stepone plus PCR system (applied biosystem step-one plus PCR, USA) and using AmpliqonRealQ Plus Master Mix Greenhigh Rox (Ampliqon, Denmark). e specific primers were designed utilizing Oligo7 v 7.60 software and OligoAnalyzer online tool (http://www.idtdna.com); and the sequence of the primers is represented in Table 1. Every 10 µl PCR reaction mixture consisted of 6.25 µl master mix, 0.25 µl of each primer, 2.25 µl RNase free dH2O, and 1 µl cDNA templates. e PCR temperature protocol was started with 95°C for 15 minutes as the first activation temperature then 40 cycles of temperature were performed at 95°C for 15 s, 61, 62, and 63°C for 30 s (according to the appropriate temperatures for each primer), and 72°C for 30 s and then ramped from 60 to 95°C to achieve a melting curve. In this study, GPDH was used as the reference internal control gene.
Statistical
Analysis. All data were analyzed via Graph-Pad Prism v 8.0.2). Results were expressed as the mean-± SEM.
e Kruskal-Wallis test was applied to specify whether there were any statistically significant differences between the means of expression of Hif-1α and NfκB among the groups. Investigation of the mean difference among the groups was carried out by Dunn's multiple comparison test.
Weight Variation.
During the 13 days of the experiment, analysis of changes in weight over three days (days 6, 10, and 13) showed that the 5-FU group had less weight gain than the control groups (p < 0.001, Figure 4). ere was also a significant difference between group 5-FU and QRC nano 5 and QRC 5 (p < 0.05).
Biochemical Analysis.
e results of the MDA test analysis showed a significant decrease in posttreatment and pretreatment of QRC and nano-pretreatment of QRC in comparison to the 5-FU group (p < 0.05, 0.0001, and 0.01, resp.) ( Figure 5). In the biochemical analysis of the SOD enzyme, a significant increase was observed between nanopretreatment and posttreatment of QRC in comparison to the 5-FU group (p < 0.01). e results of serum catalase level showed that there was a significant increase between the treatment groups of nano-pretreatment and posttreatment of QRC compared to the 5-FU group (p < 0.0001).
Effects of QRC Pretreatment and Posttreatment on Histopathological Aspects of 5-FU-Induced OM.
e histology of the tongue was normal in the control group, whereas tongue tissue of the 5-FU treatment group had significantly degraded (Table 2) [36]. In the 5-FU group, in addition to hemorrhage, inflammatory cell infiltration and hyalinization occurred in the tongue tissue, although no signs of hemorrhage and infiltration of inflammatory cells were observed in the pretreatment and posttreatment groups ( Figure 6).
Evaluation of Hif-1α and NFκB Expression in Oral Mucosa
Using Real-Time PCR. As the results showed, the expression of Hif-1α and NFκB increased in the 5-FU injected group compared to the normal group. e expression of Hif-1α and NFκB downregulated in the posttreatment of the QRC group compared to the 5-FU group but not significantly that assign with p > 0.9 and p > 0.3, respectively.
Discussion
Chemotherapy, which is one of the main therapies of cancer treatment, works by interfering with the synthesis of protein, DNA, and RNA in cells that have a high rate of division [39]. erefore, normal tissue cells are damaged along with cancer cells. OM is one of the prevalent adverse effects of cancer chemotherapeutic drugs [40]. As mentioned earlier, one of the mechanisms that lead to the complication of OM is the formation of oxidative stress and free radicals. QRC is used in the present study as a powerful antioxidant and antiinflammatory that has been used against the toxicity of many neoplastic agents [32,41]. Induction of mucositis is difficult in the animal model, and the extent of the lesion depends entirely on the induction protocol. In this study, we induced OM according to the previous protocol [37]. In mice, the peak severity of oral mucosal lesions usually occurs 4 days after induction and the onset of recovery within 7 days [42]. In the present study, after the administration of 5-FU on day 6, the signs of oral mucositis appeared at around day 10.
Weight loss during the experiment was mostly observed on the fourth day after mucositis, in which QRC pretreatment and QRC pretreatment in nanoform groups showed less weight loss than the 5-FU group, and approximately seven days after mucosal induction, weight gain increased. Weight loss is a common result of treatment with antineoplastic drugs [43].
At the macroscopic level, the results on the 4 th day after mucosal induction in all treatment groups showed a significant difference compared to group 5-FU. However, regarding the macroscopic level on the 4 th and 6 th days after mucosal induction, only the posttreatment of the QRC group showed a significant difference with group 5-FU. However, in histopathology, the posttreatment group with QRC did not show a significant difference compared to the other groups.
CAT and SOD are endogenous antioxidants that are able to clear free radicals. ese enzymes' activity and their high levels indicate that the body is under oxidative stress [44]. e performance of the pretreatment groups was more significant in comparison to the posttreatment groups. In both pretreatment groups, a significant decrease in serum MDA level was observed compared to the 5-FU group. Since QRC is widely mentioned in sources as a powerful antioxidant, increasing the strength of the antioxidant system has not been unexpected. Following the results of the present study, some researchers have noted an increase in the number of various antioxidants in other mucosal models [32,41,42].
Recent studies have shown that QRC has anti-inflammatory properties and can downregulate the production of some inflammatory factors such as NFκB, COX-2, and NO [44]. In the molecular part of our study, an interesting event occurred. In the QRC posttreatment group, the NFκB level decreased. But in both the nano-pre-and posttreatment of QRC groups, the NF-Κb and Hif-1α level increased. To justify this event, we can refer to the synergistic effects of QRC and 5-FU, which we found in a recent study [45]. It means that although the role of certain doses of QRC as a flavonoid with therapeutic properties is undeniable, we can find out that pretreatment with QRC not only cannot prevent its effectiveness against OM damage but also can reduce the levels of antioxidant enzymes and increase the expression of inflammatory factors that lead to cell death and apoptosis.
On the other hand, by comparing the nano-and nonnanoforms of QRC in this study, it can be found that since the nanoform has more permeability to tissues and cells, it was observed that the group receiving 5-FU and nanoform of QRC as the same dose, had poor performance compared to the other groups receiving pre-and posttreatment of QRC. In this case, we found that, to use the nano-QRC form, we had to use a lower dose than the non-nanoform.
Our study focused on comparing the pre-and posttreatment protective effects of QRC against OM caused by 5-FU. QRC, as an effective flavonoid in the treatment of many disorders, if used in inappropriate doses can make disruption on the living body.
Conclusion
e results of the present study showed that QRC could be a useful compound to prevent the effects of chemotherapyinduced OM with 5-FU, since QRC has been shown to lessen the severity of lesions and inflammation. According to our results from the real-time PCR assay, histopathology, and oxidative stress measurement, using QRC in appropriate doses can be a suitable compound in combating OM.
Data Availability
Data used to support the study are available from the corresponding author upon request.
Conflicts of Interest
e authors declare no conflicts of interest. | 2021-07-02T05:20:03.399Z | 2021-06-09T00:00:00.000 | {
"year": 2021,
"sha1": "e62c62f3d65593af9e1e3a39960273e7cb4c7873",
"oa_license": "CCBY",
"oa_url": "https://downloads.hindawi.com/journals/jo/2021/5598230.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "e62c62f3d65593af9e1e3a39960273e7cb4c7873",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
10192122 | pes2o/s2orc | v3-fos-license | Smchd1 regulates a subset of autosomal genes subject to monoallelic expression in addition to being critical for X inactivation
Background Smchd1 is an epigenetic modifier essential for X chromosome inactivation: female embryos lacking Smchd1 fail during midgestational development. Male mice are less affected by Smchd1-loss, with some (but not all) surviving to become fertile adults on the FVB/n genetic background. On other genetic backgrounds, all males lacking Smchd1 die perinatally. This suggests that, in addition to being critical for X inactivation, Smchd1 functions to control the expression of essential autosomal genes. Results Using genome-wide microarray expression profiling and RNA-seq, we have identified additional genes that fail X inactivation in female Smchd1 mutants and have identified autosomal genes in male mice where the normal expression pattern depends upon Smchd1. A subset of genes in the Snrpn imprinted gene cluster show an epigenetic signature and biallelic expression consistent with loss of imprinting in the absence of Smchd1. In addition, single nucleotide polymorphism analysis of expressed genes in the placenta shows that the Igf2r imprinted gene cluster is also disrupted, with Slc22a3 showing biallelic expression in the absence of Smchd1. In both cases, the disruption was not due to loss of the differential methylation that marks the imprint control region, but affected genes remote from this primary imprint controlling element. The clustered protocadherins (Pcdhα, Pcdhβ, and Pcdhγ) also show altered expression levels, suggesting that their unique pattern of random combinatorial monoallelic expression might also be disrupted. Conclusions Smchd1 has a role in the expression of several autosomal gene clusters that are subject to monoallelic expression, rather than being restricted to functioning uniquely in X inactivation. Our findings, combined with the recent report implicating heterozygous mutations of SMCHD1 as a causal factor in the digenically inherited muscular weakness syndrome facioscapulohumeral muscular dystrophy-2, highlight the potential importance of Smchd1 in the etiology of diverse human diseases.
Background X inactivation is a developmentally regulated, epigenetic mechanism that results in monoallelic expression of X-linked genes in female mammals to achieve dosage equivalence between XX females and XY males [1]. Other examples of monoallelic expression include: (a) genomic imprinting where the parental origin of the alleles of specific genes or gene clusters are epigenetically marked during gametogenesis to be expressed exclusively from either the maternally or paternally inherited allele [2], (b) stochastic monoallelic expression, which includes allelic exclusion and is characteristic of multigene families (for example, genes encoding the olfactory [3] and pheromone receptors [4], interleukins [5], B and T cell receptors [6,7], and natural killer cell receptors [8]), (c) the random combinatorial and differential monoallelic expression of the clustered protocadherins [9][10][11], and (d) the apparently random widespread monoallelic expression of possibly hundreds of individual genes spread throughout the genome [12,13].
Smchd1 (structural maintenance of chromosomes hinge domain containing 1) was identified as a semidominant suppressor of variegation in an N-ethyl-Nnitrosourea (ENU) mutagenesis screen for epigenetic modifiers [30]. The ENU derived mutant allele, named MommeD1, resulted from a nonsense mutation in the Smchd1 gene that lead to dramatically reduced Smchd1 transcript levels, probably the result of nonsense mediated mRNA decay [31]. Our breeding studies showed midgestation lethality of female but not male Smchd1 MommeD1/MommeD1 embryos, suggestive of a failure of X inactivation in the female embryos. While the inactive X elect in female mutant embryos was decorated with Xist transcript and histone H3 trimethylated at lysine 27 (H3K27me3), many X-linked genes either failed to become inactivated or escaped X inactivation shortly after the initiation of silencing [31]. None of the genes tested showed the CpG island hypermethylation of the allele on the inactive X that is a normal characteristic of X inactivation [31]. Subsequently, it was shown that Smchd1 acts late in the epigenetic cascade driving X inactivation and that some X-linked genes undergo Smchd1-independent hypermethylation of the allele on the inactive X [32]. These findings demonstrated that Smchd1 function was critical for either the completion or maintenance, but not for the initiation, of X inactivation.
Several pieces of evidence suggest that Smchd1 has functions other than during X inactivation. The ENU mutagenesis screen in which Smchd1 was identified relied upon the detection of altered variegated expression from a green fluorescent protein (GFP) transgene array integrated into an autosomal location [30,33]. Haploinsufficiency for Smchd1 also modified expression of the Agouti viable yellow (A vy ) allele of the Agouti gene (A), which shows variegated expression due to varying CpG methylation of an intracisternal A particle inserted upstream of the gene's promoter [30]. In addition, although some Smchd1 MommeD1/MommeD1 males survived as fertile adults in the FVB/n background, when backcrossed onto the C57Bl6/J background all homozygous males displayed perinatal lethality [34]. Further, an independently generated Smchd1-null allele (Smchd1 Gt(AD0165)Wtsi ), which phenocopies the female Smchd1 MommeD1/MommeD1 phenotype, also displays perinatal lethality of males on a mixed 129/C57 genetic background [31]. We hypothesized that Smchd1 does not function solely in X inactivation but probably also epigenetically modulates expression of autosomal genes in both males and females. In this study, we have used genome-wide approaches to identify autosomal genes that have deregulated expression in the absence of Smchd1 function.
Genome-wide expression profiling of Smchd1 mutant embryos
We undertook genome-wide microarray expression profiling to screen for differences between Smchd1 MommeD1/MommeD1 and Smchd1 +/+ , male and female embryos (n = 4 for each genotype and sex combination) at E9.5 ( Figure 1, Additional files 1 and 2). Males and females were analyzed separately because the females are likely to have deregulated expression of autosomal genes as a secondary consequence of X inactivation failure in addition to autosomal genes whose expression is directly regulated by Smchd1. As expected, because the analysis involved comparison of Smchd1 MommeD1/MommeD1 versus Smchd1 +/+ , the most significantly decreased transcript levels in all comparisons was Smchd1.
After filtering out the genes represented by poor quality probes on the microarray, as determined by the Illumina BeadChip Probe Reannotation datasets [35], and counting genes represented by multiple probes only once, the list of unique genes showing significantly altered transcript levels in female Smchd1 MommeD1/MommeD1 embryos (356 genes, adjusted P < 0.05, Figure 1A, Additional file 1) was much more extensive than that for males (9 genes, adjusted P < 0.05, Figure 1B, Additional file 2). The differentially expressed genes in females included 66 X-linked genes with increased transcript levels and no X-linked genes with decreased levels. Autosomal differentially expressed genes in females (290 genes in total) were distributed between those with increased transcript levels (157 genes) and those with decreased levels (133 genes) in Smchd1 MommeD1/MommeD1 compared with Smchd1 +/+ embryos.
The cell cycle related genes Ccng1 and Cdkn1a were high in the ranked list of significantly upregulated transcripts in female mutant embryos but not in male mutants. A likely explanation for the altered expression levels of these genes is that the female Smchd1 MommeD1/MommeD1 embryos were beginning to fail at this stage of development despite looking normal at the macroscopic level. This was supported by gene set enrichment analysis of the autosomal gene classes altered in female mutants (Additional file 3). Gene categories associated with the apoptosis pathway were significantly altered, indicating that failure of X inactivation initiates cell death at this stage in at least a subset of embryo tissues. Gene categories involved in developmental processes normally active in this stage of development and biosynthetic categories were also significantly altered, consistent with failure of embryo growth and development. These results confirmed our previous study showing midgestation lethality due to failure of X inactivation in female Smchd1 mutants [31] and considerably extended the list of genes failing X inactivation (from 7 of 16 genes previously tested [31] to 66 genes in the current study, for a nonredundant list of 70 genes failing X inactivation). Most of the genes identified in this study were previously classified as dependent on Smchd1 function for methylation of their CpG island (48 of 66) [32], while three genes (Ebp, Magee1, and Ndufa1) were previously shown to have an intermediate level of methylation in Smchd1 mutants; the remaining 15 genes were not previously classified. Importantly, no X-linked genes showed significantly altered transcript levels in the male embryo comparison. The ability to detect statistically significant differential expression depends on several factors, including the signal intensity (A value) and intersample variability. Genes with tissue-specific expression, low absolute expression, or lower quality probes on the array are likely to produce false negatives, hence some X-linked differentially expressed genes could have been missed by this analysis. We therefore examined the observed fold change of all expressed X-linked genes (336 genes) compared with all expressed autosomal genes (9065 genes), where expressed genes were defined as having an A value > 7.0. The mean log2 fold change of expressed X-linked genes was significantly higher than zero (mean = 0.1730, n = 336, Wilcoxon signed rank test P value < 2.2x10 −16 , Additional file 4).
The nine genes showing significantly altered transcript levels in males were also altered to a similar degree (fold change and level of statistical significance) in females. Prominent within both male and female gene lists with increased transcript levels were Ndn and Mkrn3, two genes known to be subject to genomic imprinting, which are located in the Snrpn imprinted gene cluster on mouse chromosome 7 [15q11-13 in human beings] [36,37]. The Snrpn imprinted gene cluster contains both paternally and maternally expressed imprinted genes (Figure 2A), with Several of the clustered protocadherin genes had an altered expression. Pcdhβ22 was significantly differentially expressed in both males and females, with increased transcript levels in Smchd1 mutants. Pcdhβ3 and Pcdhβ16 also showed significantly altered expression, but only in the female list. Pcdhα, Pcdhβ, and Pcdhγ are located in a cluster on mouse chromosome 18 [38] and display an unusual form of monoallelic expression involving random and combinatorial expression with each individual neuron expressing a unique combination of the α, β and γ isoforms [9,10,39]. It has been proposed that this results in enormous combinatorial diversity of the protocadherins displayed on the surface of neurons and may confer a unique identity on each neuron [39]. To confirm the results obtained from the microarray analysis we performed qRT-PCR analysis for the top four deregulated genes in males using RNA derived from E9.5 embryos (Additional file 5). All genes tested displayed significantly increased transcript levels (P < 0.01 for Mkrn3, Pisd-ps1, and Pcdhβ22; P < 0.05 for Ndn) in both male and female Smchd1 MommeD1/MommeD1 mutant embryos compared with Smchd1 +/+ embryos.
Given the phenotypic variability displayed by male Smchd1 mutants on different genetic backgrounds, we subjected an independent set of male E9.5 embryos of a different genetic background to RNA-seq analysis (n = 3 for each genotype, Smchd1 +/+ and Smchd1 MommeD1/MommeD1 ). These embryos were generated from mice where the Smchd1 MommeD1 mutation The Snrpn imprinted gene cluster, where those genes expressed from the paternal allele are shown in blue and those expressed from the maternal allele are in red. CpG islands are represented by circles on the line and where these represent an sDMR (small circles) or imprint control region (ICR) (large circle) the methylation status is indicated (M inside the filled circle for the methylated allele or unfilled circle for the unmethylated allele). Quantification of the expression levels of imprinted genes or transcripts within the Snrpn cluster by qRT-PCR measured using RNA derived from: (B) male and female E9.5 embryos, (C) mouse embryonic fibroblasts (MEF) derived from male E14.5 embryos, and (D) whole brain from adult male mice. The synthesis of first-strand cDNA from E9.5 embryo RNA was primed with oligo dT, but for the analysis of transcript levels in MEFs and brain the cDNA synthesis was primed with a cocktail of the reverse primers used for qRT-PCR. In each case the qRT-PCR signal was normalized relative to that of Rala and plotted relative to the corresponding Smchd1 +/+ sample. The genotype, sex, and number of replicates are indicated in each case. Statistical analysis was performed using the t test. ** P < 0.01 and * P < 0.05 compared with wildtype. Error bars indicate standard error.
had been backcrossed to 129 T2/SvEms background for four generations. A partially overlapping, but still relatively short list of 24 genes showed significant differences in gene expression between the genotypes (Additional file 6). It is likely that the variation in gene expression differences detected using the microarray and RNA-seq analysis results from a combination of the different genetic background used in each study and the inherent sensitivity and selectivity differences of the two methods used to assay gene expression. The differentially expressed genes detected by both analyses again pointed to disruption of the Snrpn imprinted gene cluster, with Ndn (adjusted P = 0.0059) and Peg12/Frat3 (adjusted P = 0.0018) being significantly differently expressed approximately 2-fold in the RNA-seq analysis. The other gene detected by both analyses was Pisd-ps1, about which little is known.
What was evident in the RNA-seq analysis, and could not have been detected by the microarray, was a novel gene that flanked and spliced across Ndn, Magel2, and Mkrn3 (Additional file 7). This novel gene corresponded to [AK142799 Genbank], a 3478 bp mRNA identified in a RIKEN Mus musculus 15-day embryo head cDNA library. The mRNA encodes a hypothetical tyrosine-rich region profile/EGF-like domain-containing protein. Expression levels of this gene appeared to be increased in Smchd1 MommeD1/MommeD1 compared with Smchd1 +/+ embryos (although not to statistically significant levels, owing to relatively low sequence read count), similar to the imprinted genes Ndn, Magel2 and Mkrn3, which lie within its intron.
Specific analysis of the Snrpn cluster of imprinted genes
Since some genes within the Snrpn imprinted cluster showed statistically significant increased transcript levels in the absence of Smchd1 in both the microarray and the RNA-seq analyses, we used qRT-PCR to analyze further genes specifically from the Snrpn imprinted gene cluster in E9.5 embryos ( Figure 2B). We found that Ndn, Mkrn3 and Peg12/Frat3 were significantly deregulated but Snrpn, Ube3a and Atp10a were not. Snrpn normally displays imprinted expression in all tissues [40], while Ube3a has imprinted expression only in the brain [41][42][43]. There are conflicting reports regarding the imprinted status of Atp10a in both human beings [44][45][46] and mice [47][48][49]. The inclusion of Smchd1 MommeD1/+ female embryos in this analysis demonstrated that haploinsufficiency for Smchd1 was not sufficient to cause detectable disruption of expression for the affected genes. In addition, Magel2 transcript levels were not sufficient to allow reliable detection.
Similar results were obtained for Ndn, Mkrn3, Peg12/ Frat3, Snrpn, Ube3a, and Atp10a in mouse embryonic fibroblast (MEF) cultures ( Figure 2C) as for E9.5 embryo samples. Magel2 was more reliably detectable in MEFs but, while its transcript levels appeared to be increased in Smchd1 MommeD1/MommeD1 MEFs, the difference was not statistically significant (P = 0.081012) for the four independently derived MEF cultures used.
While most genes within the Snrpn cluster have their highest expression levels in the brain (for example, Peg12/ Frat3 [50,51]), some are only expressed in the brain (for example, snoRNAs [52,53]) or are only expressed in an imprinted manner in the brain (for example, Ube3a [43]). In addition, a recent study [54] has identified transcripts uniquely expressed in the brain that originate from imprinted loci located between Snrpn and Ndn. These transcripts encode either ncRNAs or predicted miRNAs, and are normally expressed from only the paternal allele. Analysis using RNA isolated from the brains of adult Smchd1 +/+ and Smchd1 MommeD1/MommeD1 males ( Figure 2D) also revealed similar results for Ndn, Mkrn3, Magel2, and Peg12/Frat3, with significantly increased transcript levels in the absence of Smchd1. A subset of the recently identified imprinted genes spread across the region between Ndn and Snrpn (that is, DOKist4, AK045535, and AK086712) were also tested and found to be significantly increased. The imprinted snoRNAs (SnoRD115 and SnoRD116) that lie between Snrpn and Ube3a were expressed equally in Smchd1 +/+ and Smchd1 MommeD1/MommeD1 brain.
Epigenetic signature and allelic expression of disrupted imprinted genes
We considered that the increased transcript levels seen in Smchd1 mutants (for example, Ndn, Magel2, Mkrn3, Peg12/Frat3, and the ncRNA transcripts originating from loci between Snrpn and Ndn) could result from loss of imprinting and the biallelic expression for these genes rather than an increased level of monoallelic expression. Imprinted genes can be identified by their overlapping permissive and nonpermissive epigenetic markings, where the expressed allele is marked by H3 dimethylated at lysine 4 (H3K4me2) and the nonexpressed allele by CpG methylation [55], or alternatively H3 trimethylated at lysine 4 (H3K4me3) and H3 trimethylated at lysine 9 (H3K9me3) [56], respectively. Thus, loss of imprinting and biallelic expression would result in easily detectable changes of these epigenetic signatures with increased levels of the permissive marks (that is, H3K4me2 or H3K4me3) and reduced levels of the nonpermissive marks (that is, CpG methylation or H3K9me3).
To analyze DNA methylation, we used methylated DNA immunoprecipitation (MeDIP) followed by qPCR to quantify the degree of enrichment. Differential methylation of the parental alleles of the imprint control region (ICR), associated with Snrpn [57], is the primary mechanism that directs imprinted expression within the cluster [58,59]. It is the maternally inherited allele of the Snrpn ICR that is methylated, while the paternally inherited allele is not methylated (Figure 2A). A secondary (or somatic) differentially methylated region (sDMR) is associated with the CpG island of Ndn, Mkrn3, Magel2, and Peg12/Frat3 [51,[60][61][62]. The sDMR becomes methylated on the maternally inherited allele at postzygotic stages, usually postimplantation, rather than being established in the gametes.
MeDIP analysis of DNA from E9.5 Smchd1 +/+ , Smchd1 MommeD1/+ , and Smchd1 MommeD1/MommeD1 male and female embryos ( Figure 3A) showed that the Snrpn ICR and the CpG island of Atp10a had no difference in DNA methylation levels between the genotypes. Recovery of MeDIP material indicated that in Smchd1 MommeD1/MommeD1 embryos the sDMRs associated with Ndn, Magel2, Mkrn3, and Peg12/Frat3 were almost completely unmethylated compared with Smchd1 +/+ embryos. The Smchd1 MommeD1/+ embryos had a level of methylation intermediate between Smchd1 +/+ and Smchd1 MommeD1/MommeD1 embryos for Ndn, Magel2, Mkrn3, and Peg12/Frat3. This intermediate methylation level ( Figure 3A) did not translate into a detectable effect on the expression of these genes ( Figure 2A). This result was similar to that previously shown in Smchd1 MommeD1/+ female embryos, where expression of many X-linked genes and X inactivation of an X-linked GFP transgene was similar to Smchd1 +/+ embryos despite X-linked genes showing a level of CpG island methylation intermediate between Smchd1 +/+ and Smchd1 MommeD1/MommeD1 embryos [31]. While Ube3a displays imprinted expression in the brain, its CpG island is not strongly differentially methylated and remains unmethylated on both alleles [63]. The MeDIP result for Ube2a is consistent with this.
These results indicated that in the absence of Smchd1 the levels of methylation of the Snrpn ICR were retained, but the sDMRs associated with Ndn, Magel2, Mkrn3, and Peg12/Frat3 were unmethylated. Similar results were obtained from MeDIP of MEF DNA ( Figure 3B). The result was confirmed by bisulfite sequencing of MEF DNA (Additional file 8), which also confirmed that the differential methylation of the Snrpn ICR was retained in the absence of Smchd1. The imprinted loci lying between Snrpn and Ndn do not contain any annotated CpG islands in MM10 and were not analyzed.
In parallel, we used chromatin immunoprecipitation (ChIP) to analyze H3K4me2, marking the same region in MEFs. Similar ChIP in human cells shows that the CpG island of NDN and the Prader-Willi syndrome imprinting center (located in exon 1 of the SNRPN gene) carry H3K4me2 and H3K4me3 marks on the expressed paternal allele but not on the nonexpressed maternal allele [64,65]. The results ( Figure 3C) showed that levels of H3K4me2 were approximately doubled for each gene with deregulated expression and not different for genes whose expression level was indistinguishable in Smchd1 MommeD1/MommeD1 compared with Smchd1 +/+ controls. This is consistent with loss of imprinting for Ndn, Mkrn3, and Magel2, and both alleles being marked with this permissive histone modification in the absence Figure 3 The epigenetic signature of deregulated genes in the Snrpn cluster is consistent with biallelic expression. The relative level of DNA methylation was quantified by qPCR of MeDIP recovered DNA in samples from (A) male and female E9.5 embryos, and (B) MEFs isolated from male E14.5 embryos. (C) The relative level of H3K4me2 for each gene was quantified by qPCR of ChIP recovered material in chromatin derived from MEFs isolated from male E14.5 embryos. The qPCR signal was normalized relative to that of Rhox6/9 for MeDIP or relative to input for the H3K4me2 ChIP. Results are plotted relative to the corresponding Smchd1 +/+ sample. The genotype, sex, and number of replicates are indicated in each case. Statistical analysis was performed using the t test. In (A) **P < 0.01 or *P < 0.05 for Smchd1 MommeD1/MommeD1 compared with Smchd1 +/+ . Statistical significance not shown for the comparison of Smchd1 +/+ compared with Smchd1 MommeD1/+ . In (B) and (C) ** P < 0.01 compared with wildtype. Error bars indicate standard error. ChIP, chromatin immunoprecipitation.
of Smchd1. H3K4me2 levels on Peg12/Frat3 were not tested. No difference in H3K4me2 levels was seen at the Snrpn ICR, consistent with our finding that loss of Smchd1 did not affect the Snrpn expression.
In combination, these changes in epigenetic signature and gene expression in Smchd1 mutants are completely consistent with loss of imprinting resulting in biallelic expression of Ndn, Magel2, Mkrn3, and Peg12/Frat3 but with imprinting being maintained at the other imprinted genes within the cluster (that is, Snrpn and Ube3a). From our gene expression analysis DOKist4, AK045535, and AK086712 are also likely to be subject to loss of imprinting but not the imprinted snoRNAs (that is, SnoRD115 and SnoRD116). To definitively demonstrate the allelic expression of genes within the Snrpn imprinted gene cluster we used RNA-FISH on MEFs derived from E14.5 Smchd1 +/+ compared with Smchd1 MommeD1/MommeD1 embryos. We compared the number of signals in the nucleus resulting from FISH probes that detect Snrpn and either Ndn or Magel2 transcript ( Figure 4A-D). In Smchd1 +/+ MEFs Snrpn, Ndn, and Magel2 were expressed monoallelically in approximately 80% of nuclei ( Figure 4E and F). In Smchd1 MommeD1/MommeD1 MEFs, monoallelic expression of these genes was only seen in 12 to 23% of nuclei, with between 46% and 63% of nuclei showing biallelic expression of Ndn or Magel2 but monoallelic expression of Snrpn.
Identification of other imprinted genes showing loss of imprinting in Smchd1 mutants
Lack of DNA methylation at the sDMR associated with Ndn, Magel2, Mkrn3, and Peg12 was the most obvious phenotype associated with homozygous mutation of Smchd1. The level of CpG methylation of the ICR associated with the Snrpn cluster showed no difference between Smchd1 MommeD1/MommeD1 and Smchd1 +/+ samples. We undertook qPCR MeDIP analysis of the ICR or sDMR associated with other imprinted genes using embryo-and male-MEF-derived DNA. In these analyses, only the sDMR associated with the Cdkn1c gene showed a significant decrease in CpG methylation levels in Smchd1 Mommed1/MommeD1 samples; however, this was only observed in embryo-derived DNA (Additional file 9A), not in the DNA derived from MEFs (Additional file 9B). The degree of DNA methylation loss at Cdkn1c in the Smchd1 Mommed1/MommeD1 samples was not as marked as for the sDMR associated with Ndn, Magel2, Mkrn3, and Frat3/Peg12. We were not able to demonstrate any significant loss of methylation at the ICR of any imprinted gene cluster that we tested.
The use of single nucleotide polymorphisms (SNPs) to determine the parental allele of origin of expressed alleles is the most widely accepted method for determining whether genes are subject to (or failing) genomic imprinting. We had previously backcrossed the Smchd1 MommeD1 mutation from the FVB/n background (in which the ENU mutagenesis screen that identified Smchd1 was undertaken) onto the C57Bl6/J background for more than ten generations. While none of the genes in the Snrpn imprinted gene cluster carries SNPs between these genetic backgrounds, several other imprinted genes do. This is particularly the case for imprinted genes that are expressed in an imprinted manner uniquely in the placenta. To identify more imprinted genes with loss of imprinting in Smchd1 mutants, we analyzed the placentas of F1 (C57Bl6/J × FVB/n) E14.5 male embryos (n = 5 for each genotype, Smchd1 +/+ and Smchd1 MommeD1/MommeD1 ). To ensure that we were able to dissect out the embryonic portion of the placenta free from contamination by maternal tissue, we also included a ubiquitously expressed GFP transgene (UBI-GFP) [66] in the sires. The genotypes of the mice set up in the cross were as follows: (C57Bl6/J; Smchd1 MommeD1/+ , UBI-GFP T/+ ) sire crossed to (FVB/n; Smchd1 MommeD1/+ ) dam. Only the brightly GFP fluorescent embryonic portion of the placenta was used for RNA-seq analysis. Very few genes showed differential levels of expression between the genotypes (Additional file 10), with only four RefSeq genes reaching statistical significance Smchd1, Mmp12, Cyp1a1, and Apoa1 (adjusted P < 0.05).
Using SNP analysis, we identified the Igf2r imprinted cluster as being disrupted in Smchd1 mutants. The cluster includes the imprinted genes Igf2r, Slc22a2, and Slc22a3, which are preferentially expressed from the maternally inherited allele, and the Airn ncRNA, which is preferentially expressed from the paternally inherited allele ( Figure 5A). Igf2r and Airn are imprinted in all tissues, but Slc22a2 and Slc22a3 are imprinted only in the placenta, with Slc22a3 becoming biallelically expressed from E15.5 onwards [67]. In Smchd1 mutant placenta, the allelic pattern of Igf2r expression is identical to that in wildtype placenta with expression predominantly from the maternal allele, but the Slc22a3 gene showed biallelic expression rather than the predominant maternal allele shown by wildtype placenta of the same age ( Figure 5B, Additional file 11). This finding was confirmed by analysis of placentas derived from a reciprocal cross: a (FVB/n; Smchd1 MommeD1/+ ) sire crossed to a (C57Bl6/J; Smchd1 MommeD1/+ ) dam ( Figure 5C). This formally demonstrates loss of imprinting in the absence of Smchd1. No other imprinted genes that were amenable to analysis in this experiment (that is, showing imprinted expression in the placenta and carrying suitable SNPs) showed similar statistically significant loss of imprinting (Additional file 11). Bisulfite analysis confirmed that Smchd1 loss did not alter the differential methylation of the ICR that lies within exon 2 of Igf2r and that controls imprinting of the Igf2r imprinted cluster of genes (Additional file 12).
The SNP analysis also revealed exclusive expression of the FVB/n alleles of genes within a region of chromosome 17 surrounding Smchd1 (that is, from Tgif1 at position [Chr17:70844412 MM10] to Ttc27 at [Chr17:74770280 MM10]. This resulted from FVB/n alleles of genes linked to the mutant Smchd1 MommeD1 allele being retained in the backcross to C57Bl/6 J.
While this finding was not relevant to the genomic imprinting data for the Igf2r imprinted cluster, it indicated that we were successful in obtaining clean embryonic placental tissue for the analysis and defined the region of the FVB/n chromosome surrounding the Smchd1 MommeD1 that was retained in our backcrossed animals.
The clustered protocadherin genes show altered expression in Smchd1 mutants
Previously, we had shown that expression of members of the clustered protocadherin family was deregulated in Smchd1 mutant MEFs (normal and transformed) [34]. Since Pcdhβ22 also showed significantly altered transcript levels in the microarray analysis of E9.5 embryos, we considered the possibility that Smchd1 might function in the mechanism controlling the random combinatorial monoallelic expression of the clustered protocadherins ( Figure 6A). We screened the full transcript set from all three types of clustered protocadherin gene for deregulated expression levels in the adult male brain and found that all three clusters showed altered expression in Smchd1 mutants ( Figure 6B). The strongest changes occurred in the Pcdhα cluster, where Pcdhα1 was increased 22-fold and Pcdhα8 was increased approximately 7.5-fold. Three Pcdhα transcripts (Pcdhα9, PcdhαC1, and PcdhαC2) were significantly decreased. Almost all members of the Pcdhβ cluster (Pcdhβ1-6, 8, 10, 11, 14, 17 to 22) displayed increased transcript levels ranging from 2 to 4-fold, while the Pcdhγ cluster was least affected both in terms of fold change and number of alternative transcripts displaying some alteration.
Discussion
We initially identified Smchd1 as an epigenetic modifier in an ENU mutagenesis screen [30] and demonstrated a critical role for Smchd1 in either the completion or maintenance, but not the initiation, of X inactivation [31]. This study considerably extends the known genes failing X inactivation in Smchd1 mutants (7 previously published, 66 in the current study, for a nonredundant total list of 70 genes). While it is clear that some genes do not fail X inactivation in Smchd1 mutants [32], we believe that the current set of genes does not reflect the full extent of genes failing X inactivation since it is likely that many X-linked genes would not be expressed at sufficient levels at this developmental stage to be detected.
Our main aim was to identify autosomal genes whose correct expression was directly dependent upon Smchd1 function. We had reasoned that there would be many autosomal genes in female Smchd1 mutants that would display deregulated expression that was secondary to failure of X inactivation. To ensure that we could identify autosomal genes that were directly regulated by Smchd1, we analyzed males separately from females and concentrated on genes whose expression was altered to a similar degree in both males and females. Our analysis revealed a relatively short list of nine genes showing statistically significant altered expression in Smchd1 mutant males; all of these genes were similarly deregulated in females.
Amongst this list were two genes (Ndn and Mkrn3) from the Snrpn imprinted gene cluster with increased levels of expression. Specific analysis revealed further genes from the cluster (Magel2, Peg12/Frat3, DOKist4, AK045535, and AK086712) showing increased expression in Smchd1 mutants. The deregulated genes displayed epigenetic signatures consistent with biallelic expression and loss of imprinting. RNA-FISH showed that Ndn and Magel2 displayed biallelic expression but Snrpn retained monoallelic expression. From this result it is clear that the Snrpn imprinted cluster has two distinct parts in terms of its imprinting mechanism; (a) the region distal to the Snrpn start site containing genes, whose imprinted expression was dependent upon Smchd1, and (b) the region proximal to and including Snrpn, where genomic imprinting was maintained. This division of the cluster agrees with a previous model for the control of imprinting at the locus where genes from group (a) were controlled by DNA methylation and those from group (b) were regulated by the paternally expressed RNA transcript that initiated from the ICR [68]. To our knowledge, no other animal model has disassociated imprinting of the genes distal to Snrpn were quantified by qRT-PCR using RNA derived from whole adult male mouse brains. The synthesis of first-strand cDNA was primed with a cocktail of the reverse primers used for qRT-PCR. In each case, the qRT-PCR signal was normalized relative to that of Rala and plotted relative to the corresponding Smchd1 +/+ sample. Statistical analysis was performed using the t test. ** P < 0.01 and * P < 0.05 compared with wildtype. Error bars indicate standard error.
from the remainder of the cluster. The Snrpn cluster produces a poorly understood set of ncRNA transcript variants that are implicated in directing the imprinted expression of other genes within the cluster. Usually, these are expressed from the paternal allele and include the U-exon/Snrpn/Ube3a-ATS that has been proposed to control imprinted expression of Ube3a in the brain [69]. Recent reports have, however, identified yet other alternative Snrpn ncRNA transcripts emanating from upstream promoters and including upstream exons that are expressed from the maternal allele and required for silencing of genes on the maternal allele [70,71]. Our finding that Smchd1 is essential for some of the activities associated with the imprinting of the Snrpn cluster (that is, silencing the maternal allele of Ndn/Magel2/Mkrn3 and Peg12/Frat3), but apparently not others (that is, silencing Ube3a on the paternal allele in the brain), adds further complexity to the regulation of this imprinted locus. Another new feature that has been identified in this study is the identification of a gene flanking Ndn, Magel2, and Mkrn3, which is likely to be imprinted and encodes a hypothetical novel tyrosine-rich region profile/EGF-like domain-containing protein.
Our SNP analysis of placental RNA-seq data also showed that the Igf2r imprinted gene cluster was disrupted in Smchd1 mutants, with the Slc22a3 gene showing loss of imprinting resulting in biallelic expression but imprinted Igf2r expression being unaffected. Imprinting of this cluster is governed by an ICR within exon 2 of Igf2r, which coincides with the transcriptional start site of the Airn ncRNA. It has recently been demonstrated that it is the transcriptional overlap of the Airn ncRNA and not the ncRNA transcript itself that induces the imprinting of Igf2r [72]. This is not the case for Slc22a3, whose imprinting is driven by the Airn ncRNA transcript itself, which targets G9a in cis to the allele of Slc22a3 that is silenced [73]. Thus, we conclude that Smchd1 functions as part of the mechanism by which Airn ncRNA induces facultative heterochromatin of linked genes in the Igf2r imprinted gene cluster.
We specifically demonstrated that the differential DNA methylation marking the ICR associated with the Snrpn imprinted gene cluster was maintained. The sDMRs associated with the disrupted genes in this cluster were, however, almost completely unmethylated. Differential methylation of the ICR controlling the Igf2r imprinted cluster was also not altered in Smchd1 mutants. The CpG island associated with Slc22a3 gene does not appear to carry a differential methylation mark in the placenta [67] and was not assessed in this study.
It is possible that Smchd1 is involved in regulating the expression of further imprinted genes not identified in this study. A complete analysis would involve crossing the Smchd1 mutant allele onto a suitable mouse strain carrying many SNPs in imprinted genes for determining allelic expression and analysis of suitable tissues where their expression is imprinted. If the imprinting of other genes involves Smchd1, we would predict that they would be situated within clusters of imprinted genes where the imprinting was regulated by the expression of an imprinted ncRNA similar to the situation for the Snrpn and Igf2r clusters. Our finding that in some Smchd1 MommeD1/MommeD1 tissues the sDMR associated with Cdkn1c had altered levels of methylation may support this assertion, since regulation of the Kcnq1 imprinted gene cluster involves the imprinted Kcnq1ot1 ncRNA.
Our analysis has also highlighted altered expression of the clustered protocadherins in brain. These genes display an unusual form of random combinatorial monoallelic expression where each neuron expresses a unique combination of the protocadherin isoforms. This is proposed to confer individual neurons with a unique cell surface identity that may be important for neural circuit assembly [74]. The regulatory mechanism underlying the form of monoallelic expression displayed by protocadherins is poorly characterized but believed to involve the interaction of CTCF [75] with multiple enhancers and promoters [11,76]. We found that expression levels of some of the protocadherins, in particular that of Pcdhα1, was strongly deregulated in the Smchd1 mutant brains (approximately 22-fold upregulated compared with wildtype). This is considerably more than the approximately 2-fold seen for the genes in the Snrpn imprinted gene cluster showing loss of imprinting, but this is possibly due to the unique form of random combinatorial monoallelic expression displayed by the clustered protocadherins, where Pcdhα chooses between 12 different alternative first exons to be utilized in each cell.
The exact molecular mechanism by which the Smchd1 protein functions in these roles remains unknown. Smchd1 was named because the predicted encoded protein contains the hinge domain characteristic of SMC proteins [77]. The SMC proteins usually act as heterodimers (that is, SMC1 and 3 in cohesin, SMC2 and 4 in condensin, SMC5 and 6 in an unnamed complex involved in DNA repair) that interact through their hinge domains. The hinge domain of SMC proteins is also where the protein folds back on itself to bring together two nucleotide binding domains lying at the amino-and carboxy-terminuses of SMC proteins to form an ATPase domain, which hydrolyses ATP to provide the energy needed to manipulate chromosome-size molecules. The predicted Smchd1 protein also contains an ATPase domain, but it varies from the ABC-type ATPase contained in the SMC proteins and is similar to the ATPase domain from the GHKL phosphotransferase superfamily of type II DNA topoisomerases, Hsp90, and MutL, and bacterial and mitochondrial protein kinases [78]. From this, one would predict that Smchd1 might function as part of a protein complex that manipulates chromatin ultrastructure in an ATP-dependent manner. It is possible that Smchd1 is involved in physically sequestering the chromosomal region containing the allele being silenced to a nuclear territory that is remote from the active allele and where it can be stably maintained in its silent state. Such a function is consistent with the essential role we have previously shown for Smchd1 in X inactivation and with our current findings in the Snrpn and Igf2r imprinted gene clusters, and the clustered protocadherins.
Conclusions
This work extends the involvement of Smchd1 from X inactivation to two other forms of monoallelic expression. Since the Killer cell lectin-like receptors (Klra), which are also subject to monoallelic expression, have previously also been reported to be deregulated in transformed MEFs and tumors that are mutant for Smchd1 [34], it will be interesting to determine whether Smchd1 can be implicated in other forms of monoallelic expression (that is, the olfactory receptors, T cell receptors). Moreover, SMCHD1 mutation has recently been shown to contribute to an inherited muscular weakness syndrome, FSHD2, which requires digenic inheritance of a heterozygous SMCHD1 mutation and a 'permissive' D4Z4 microsatellite array haplotype [79]. Heterozygous mutation of SMCHD1 results in hypomethylation of the D4Z4 microsatellite array and, consequently, the DUX4 retrogene embedded in the array becomes expressed in skeletal muscle tissue in a variegated manner. A seemingly distinct role for Smchd1 has also been suggested with the identification GMI1 as the likely Arabidopsis ortholog of Smchd1 [80]. GMI1 encodes a γ-radiation inducible protein that is proposed to be involved in homologous recombination. If Smchd1 also proves to function in this role in mammals, its diversity of function may rival that of the SMC proteins.
Mice and genotyping
All experimental animals were treated in accordance with the Australian Government National Health and Medical Research Council guidelines for the care of experimental animals and the work was approved by the animal ethics committees of the Queensland Institute of Medical Research and the Walter and Eliza Hall Institute. Mice carrying the mutant Smchd1 MommeD1 allele were genotyped by allelic discrimination using a custom Taqman assay [31] (primers and Taqman probes are listed in Additional file 13). Mice carrying the ubiquitously expressed GFP transgene (UBI-GFP) were genotyped by observing GFP fluorescence.
Embryos and cell cultures
All embryos were produced by natural matings between heterozygous (Smchd1 MommeD1/+ ) males and females to produce wildtype, heterozygous, and homozygous Smchd1 mutant embryos. Embryos for microarray analysis were dissected at E9.5, with the yolk sacs of each embryo used for genotyping. Embryos for generation of MEFs were dissected at E14.5 and a small section of the tail was removed for Smchd1 genotyping. The remaining carcass of each embryo was individually treated for preparation of MEFs [81].
Microarray analysis
Total RNA was extracted from Smchd1 MommeD1/MommeD1 and Smchd1 +/+ E9.5 embryos with RNA yields quantified using a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific Australia, Scoresby, Vic, Australia) and RNA integrity assessed using a Bioanalyzer (Agilent 2100, Agilent Laboratories, Santa Clara, CA, USA). Only samples with an RNA integrity number [82] greater than eight were used.
RNA-seq analysis
Total RNA was prepared from E9.5 embryos or the embryonic portion of placentas from F1 (C57Bl/6 J sire cross FVB/n dam) E14.5 male embryos dissected with the aid of a paternally inherited ubiquitin-GFP transgene [66]. RNA yields and integrity were assessed as for the microarrays. RNA-seq was performed by the Australian Genome Research Facility. The sequence reads were mapped to MM10 using TopHat 2 and expression differences between genotypes analyzed by Cufflinks/Cuffdiff v 2.0.2. For allele-specific expression, SNPs were identified in the RNA-seq sequence reads and allelic expression ratios quantified bioinformatically: SNPs were called in the RNA-seq data using the Samtools/BCFtools programs mpileup, bcftools and vcfutils, and mapped to the mouse RefSeq genes. For each called SNP and each sample, the ratio of reads supporting either the variant or reference was calculated. Significantly different allelic expression was identified by testing whether the means of these ratios were different in the Smchd1 +/+ or Smchd1 MommeD1/MommeD1 male embryos (using a twotailed t test).
To confirm allelic bias, the placentas from reciprocal cross F1 mice (FVB/n sire cross C57Bl/6 J dam) were dissected and total RNA purified. RT-PCR was used to amplify the specific portions of transcripts containing the SNPs to be analyzed and PCR product used to generate barcoded libraries, which were pooled for Ion Torrent sequencing (Life Technologies, Grand Island, NY, USA). Allelic expression ratios were then determined from the mapped reads.
qPCR and qRT-PCR Quantitative PCR (qPCR and qRT-PCR) was performed using Platinum SYBR Green qPCR SuperMix-UDG (Life Technologies, Grand Island, NY, USA). Primers were designed using Primer3 and levels of specific PCR amplicons determined relative to a standard curve. For qRT-PCR, first-strand cDNA synthesis was primed either by oligo dT or by a cocktail of specific reverse primers. The sequences of all primers used are listed in Additional file 13.
MeDIP analysis
MeDIP was performed using a protocol adapted from a published method [87]. Genomic DNA was sheared by sonication (Bioruptor, Diagenode, Liège, Belgium) to between 300 and 500 bp. DNA (1 μg) was then denatured and MeDIP performed with 1 μg of anti-5MeC monoclonal antibody clone 33D3 (ab10805, Abcam, Cambridge, UK) plus 0.5 mg of Dynabeads Protein G (Life Technologies, Grand Island, NY, USA). MeDIP material was purified by magnetic separation, phenol extraction, and precipitation. The recovery of specific genomic regions was measured by qPCR relative to Rhox6/9, as we had previously shown that CpG methylation at these loci was not affected by Smchd1 loss [31].
ChIP analysis
ChIP for H3K4me2 was performed using an adaptation of a published protocol [88]. Chromatin was prepared from isolated nuclei of formaldehyde cross-linked MEFs and sonically sheared to between 200 and 500 bp. Sheared chromatin was immunoprecipitated with anti-H3K4me2 Ab (ab11946, Abcam, Cambridge, UK) plus 0.5 mg of Dynabeads ProteinG (Life Technologies, Grand Island, NY, USA). ChIP material was purified, as for MeDIP, and recovery of specific genomic regions in the ChIP samples was measured by qPCR relative to input DNA.
Bisulfite sequencing
Nested primers for amplification of bisulfite converted DNA (Additional file 13) were designed using MethPrimer [89] and the sequencing data analyzed using BiQ analyzer [90]. Genomic DNA (1 μg) was bisulfite treated using the Qiagen EpiTect Bisulfite Kit (Qiagen, Hilden, Germany). After PCR, the product was purified (Qiagen PCR cleanup kit), cloned into pGemT (Promega, Madison, WI, USA) and colonies picked for PCR amplification with SP6 and T7 primers. The PCR product was sequenced with either SP6 or T7 primer.
RNA-FISH
RNA-FISH was carried out as previously described [91] on Smchd1 +/+ and Smchd1 MommeD1/MommeD1 MEFs grown on gelatin-coated glass coverslips fixed with 3% paraformaldehyde. The probe for Snrpn was a BAC (RP23-97I5, RPCI-23 Female (C57BL/6 J) Mouse BAC Library), while the probes for Ndn and Magel2 consisted of most of each gene PCR amplified from mouse genomic DNA (Additional file 12) and cloned into pGemT. The probes were labeled with Orange or Green dUTP (Abbott Molecular, Abbott Park, Illinois, USA) by nick translation and cell nuclei were counterstained with DAPI (1 μg/ml) before visualization on Zeiss Axioplan epifluorescence microscope equipped with a SPOT RT3 CCD camera. For quantification, 100 randomly selected nuclei were scored for the signal from each gene (that is, Snrpn and Ndn or Magel2).
Data access
Microarray and RNA-seq data reported in this study have been submitted to the NCBI Gene Expression Omnibus under the accession numbers GSE44958 and GSE44669, respectively.
Additional file 4: Many X-linked genes may be upregulated, owing to X inactivation failure resulting from Smchd1-loss. The mean log2 fold change of all expressed autosomal and X-linked genes (A value > 7.0) in Smchd1 MommeD1/MommeD1 female embryos compared with Smchd1 +/+ embryos is plotted.
Additional file 5: Confirmation of differential expression identified in microarrays. Expression levels of the four most differentially expressed genes in the microarrays were quantified using qRT-PCR. RNA was derived from male and female Smchd1 +/+ and Smchd1 MommeD1/MommeD1 E9.5 embryos (samples included the four samples used for the microarrays but with two to four independent additional samples added, depending on genotype). The synthesis of first-strand cDNA was primed with oligo dT. In each case, the qRT-PCR signal was normalized relative to that of Rala and plotted relative to the corresponding Smchd1 +/+ sample. The genotype, sex, and number of replicates are indicated in each case. Statistical analysis | 2017-06-25T12:09:10.097Z | 2013-07-02T00:00:00.000 | {
"year": 2013,
"sha1": "ab132ce4e1c02c752f9a2f250d85b4a719759089",
"oa_license": "CCBY",
"oa_url": "https://epigeneticsandchromatin.biomedcentral.com/counter/pdf/10.1186/1756-8935-6-19",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "a0a560ee7b33d9b87be28944b265a0d3f9208f1b",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine",
"Biology"
]
} |
257568845 | pes2o/s2orc | v3-fos-license | Children Under 5 Years Old Mortality and Associated Factors
A Mortality of a child, also defined as under-5 years mortality or child death, indicates the death of children under five or the ages between one month to four years. Many deaths in developing countries go unreported since many low-income families cannot afford to register their babies in the official government registry. A cross-sectional study was conducted in this study for the period of study. This study was carried out from 24/12/2021 to 11/ 3/2021 in two governmental hospitals: AL Zahra teaching hospital and the Al Hakeem hospital in Al Najaf AL Ashraf province. The study sample was children under five years old who were diagnosed with death and living in Al-Najaf province. This study’s findings show that the highest numbers of respondents, 50%, were married at age 13-18 years. 64% of the total number were from rural areas, and according to family income, 60% of respondents had a family income range from 501- 1.000.000 IQD. According to obstetrical history, 93% of women had died fetus 59% of them died in the third trimester. 30% of women had died children under five years of age. 63% of these events happened for different reasons. In addition, 61% of mothers had a medical history. However, 55% of fathers had the same medical history except for liver diseases. Also, there was 16% of parents had an incompatible blood group. According to history using medications, 73% of mothers have a history of medication use
INTRODUCTION
Mortality of a child, also defined as under-5 years mortality or child death, indicates the death of children under five or the ages between one month to four years. Many deaths in developing countries go unreported since many low-income families cannot afford to register their babies in the official government registry. A child's death is emotionally hard on their parents and their relationship as a couple. 6.3 million live-born children were estimated in 2013 worldwide to die before five years (Liu et al., 2017) [1]. This estimated number decreased from 9. will possibly only be accomplished by a few nations (You et al., 2010) [2].
The reasons for children's mortality rate differed from one place to another; in 2016, the annual number of deaths from diarrhea reported among Chinese children <5 years from all-cause diarrhea was 42%. Of the 53.5 rotavirus diarrhea deaths, it is estimated that 92.7% (49.6) occurred in rural areas (Zhang et al., 2015) [3]. In East Delhi, one study was conducted on children aged six months to 5 years admitted to the pediatric emergency department during the study period from February 2012 to March 2013; this study found that 36% of children had clinical features of vitamin deficiency, and 51.1% of children were underweight (Sakshi Sachdeva et al. 2016) [4]. In 2015, among the 5.941 million children who did not live to age 5, 2.681 million (45.1%) died in the neonatal period (figure 1). The leading causes of death in children under 5 were preterm birth complications (Li Liu et al., 2015) [5]. This study aims to describe children's sociodemographic characteristics, determine the causes of mortality of children under five years old and determine the associated factors causes of mortality of children under five years old.
Study Settings
A cross-sectional study was conducted in this study for the period of study. This study was carried out from Mar. 2021 to Des. 2021; two governmental hospitals are AL Zahra Teaching Hospital and the Al Hakeem Hospital in Al Najaf AL Ashraf province. The study sample was women who died of children under five years old living in Al-Najaf province.
Study Sample
One hundred cases were selected by using simple random sampling. The p in the sample size calculation is taken from the previous study (Ibadi, 2019) [6], with a confidence interval of p = 0.05 and a power of study 1.645.
Data Collection
A structured questionnaire form was applied to gather data to elicit information from the study participants. The questionnaire included the following information.
1. Sociodemographic data: These include five information items on the children, mothers', fathers', and parents' educational levels and occupations.
2. Current baby status and background information: These include weight, height, type of feeding, weaning age, and information about helminth infections.
3. Medical Family history: Includes history of anemia among family members, anemia during the pregnancy period, taking supplementary drugs, and chronic diseases.
4. Obstetrical history: Include information about the mother if she had a bad obstetrical history.
Statistical Analysis
Statistical Package for the Social Sciences (SPSS) version 21 was used to analyze the data. Descriptive statistics were calculated for selected numerical and categorical variables for descriptive data.
Ethical Consideration:
Al-Najaf Health Directorate was permitted to do this work. Written consent was taken from all participants. This table shows a significant correlation between the death of children under five years with infected children by diseases, the medical history of the mother, taking medications during pregnancy, and the bad obstetric history of mothers. Table one shows that the greatest number of respondents 50%, were married at age 13-18 years. 64% of the total number were from rural areas, and according to family income, 60% of respondents had a family income range from 501-1.000.000 IQD. 28% of women had a secondary education level; however, 26% of husbands had diplomas. Finally, 84% of women were housewives, and 23% worked as plant and machine operators' assemblers. Paul P.
DISCUSSION
(2019) [7] confirmed this result in India; he found that the prevalence of morbidity and mortality were crucially higher among the births of women who married before 18 years of age, and there was a positive association (Paul, [11,12]. According to obstetrical history, 93% of women had died fetus 59% of them died in the third trimester. 30% of women had died children under five years of age. 63% of these events happened for different reasons, 63% were unknown, accidents caused 11%, infectious diseases caused 20%, and 6% caused deformities. The common bad obstetrical conditions were 39% abortions, 16% neonatal death, 5% stillbirth, and 44% unexpected fetus death. This result was very near to the finding by Singh and Sidhu in 2010; they found that 50-60% of all first-trimester pregnancy losses harbor a chromosomal abnormality, which leads to abnormal growth and development of the pregnancy. Most abnormal pregnancies fail in the first trimester (Singh and Sidhu, 2010) [13]. Many studies have also found multiple bad obstetric issues associated with under-5 child mortality ( [19]. This result is similar to the Stock & Norman study conducted in 2019; they found that using medication's potential impacts include increasing rates of miscarriage, stillbirth, fetal growth perturbations, and preterm birth. Thus, the evaluation of pregnancy treatments needs to be broad, including women whose pregnancies have not continued past the first trimester, and be of sufficient scale to recognize severe but rare events (Stock & Norman, 2019) [20].
CONCLUSIONS
The findings of this study show that 1. A high number of respondents 50% were married at age 13-18 years.
2. 64% of the total numbers were from rural areas.
4. According to obstetrical history, 93% of women had died fetus 59% died in the third trimester. 30% of women had died children under five years of age. 63% of these events happened for different reasons.
5. 61% of mothers had a medical history. However, 55% of fathers had the same medical history except for liver diseases. Also, there was 16% of parents had an incompatible blood group.
6. 73% of mothers have a history of medication use. | 2023-03-17T15:17:42.516Z | 2023-01-01T00:00:00.000 | {
"year": 2023,
"sha1": "a2b9f66b71312034979f8459be5bd7baf5618a88",
"oa_license": null,
"oa_url": "https://www.mathewsopenaccess.com/scholarly-articles/children-under-5-years-old-mortality-and-associated-factors.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "5a580140ec27af4f44777c2bf882e223665c4325",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
12946119 | pes2o/s2orc | v3-fos-license | On the utility of diagnostic instruments for pediatric delirium in critical illness: an evaluation of the Pediatric Anesthesia Emergence Delirium Scale, the Delirium Rating Scale 88, and the Delirium Rating Scale-Revised R-98
Purpose Delirium is a poor-prognosis neuropsychiatric disorder. Pediatric delirium (PD) remains understudied, particularly at pediatric intensive care units (PICU). Although the Pediatric Anesthesia Emergence Delirium (PAED) scale, the Delirium Rating Scale (DRS-88), and the Delirium Rating Scale-Revised (DRS-R-98) are available, none have been validated for use in PICU settings. The aim of the present study was to investigate the use of the DRS/PAED instruments as diagnostic tools for PD in the PICU. Methods A prospective panel study was conducted, under circumstances of routine clinical care, investigating the diagnostic properties of the PAED, DRS-88, and DRS-R-98 in PICU patients at a tertiary university medical center. A total of 182 non-electively admitted, critically ill pediatric patients, aged 1–17 years, were included between November 2006 and February 2010. Sensitivity, specificity, and receiver operating characteristic (ROC) curves were calculated. Three psychometric properties were analyzed: (1) internal consistency (2) proportion of items not rateable, and (3) discriminative ability. Results The PAED could be completed in 144 (93.5%) patients, much more frequently than either the DRS-88 (66.9%) or the DRS-R-98 (46.8%). Compared with the clinical gold standard diagnosis of delirium, the PAED had a sensitivity of 91% and a specificity of 98% (AUC 0.99). The optimal PAED cutoff score as a screening instrument in this PICU setting was 8. Cronbach’s alpha was 0.89; discriminative ability was high. Conclusions The PAED is a valid instrument for PD in critically ill children, given its reliance on routinely rateable observational signs and symptoms.
Introduction
Delirium is a neuropsychiatric disorder secondary to a general medical condition and represents a serious complication of the underlying disease or its treatment [1]. Three subtypes of delirium can be distinguished: (1) hypoactive delirium, which is characterized by apathy, decreased responsiveness, and withdrawal; (2) hyperactive delirium, which is characterized by restlessness, emotional lability, and agitation; (3) mixed delirium, which is a combination of the two other subtypes [2]. The condition is frequently seen in critically ill adult and geriatric patients and is associated with poor prognosis, reflected by worse functional and cognitive outcome, prolonged hospital stay, and a higher mortality rate after discharge from hospital [1,[3][4][5]. Although the clinical picture of delirium in adults is well known, comparatively scarce attention has been paid to its counterpart in children-pediatric delirium (PD)-even though data suggest that PD is a similarly serious condition [2, [6][7][8].
It is important to correctly diagnose delirium in children, in order to make early intervention possible. Adult delirium is often missed by clinicians, because it commonly manifests as the hypoactive subtype [9][10][11]; anecdotal evidence suggests that hypoactive forms, and misclassification thereof, may be even more prevalent in the pediatric population. As a consequence, there is an urgent need for a valid instrument which can be used by staff at pediatric intensive care units (PICU) [2, 12] to assess PD. The Pediatric Confusion Assessment Method for the Intensive Care Unit (pCAM-ICU), the Pediatric Anesthesia Emergence Delirium (PAED) scale, the Delirium Rating Scale (DRS-88), and the Delirium Rating Scale-Revised (DRS-R-98) are instruments that have been developed to diagnose delirium in clinical practice. The pCAM-ICU has recently been validated as a highly reliable instrument in pediatric critically ill children (8-15 years) [13].
The PAED, which only includes items covering behavioral symptoms, has been validated in children in the age range of 18 months-6 years in a post-anesthesia setting [14]. However, in that study, the diagnosis of delirium was only made if the anesthetist decided that treatment of the condition was necessary. Therefore, delirium could have been underdiagnosed. The DRS-88 and the DRS-R-98 focus more on cognitive functioning [15] and have only been validated in adults, and not in ICU settings [12]. Although Turkel and colleagues [16] reported that the DRS-88 could be applied to diagnose delirium in children (6 months-19 years), their study was carried out in retrospect and no control group was available. Thus, to our knowledge, only the pCAM-ICU has been tested in PICU settings in children, whereas the DRS/PAED instruments have not been tested in a PICU setting.
Therefore, the main objective of the present prospective panel study was to investigate, under circumstances of routine clinical care, the use of the DRS/PAED instruments as diagnostic tools for PD in a PICU setting, taking into account both item and sum score levels of the instruments. It was hypothesized that (1) the PAED would be more suitable than the DRS in this setting, as it contains only behavioral items, which are easier to assess in young critically ill children and, therefore, (2) a significant negative impact of younger age was expected with regard to the proportion of rateable items for both versions of the DRS (as it relies on cognitive items that cannot be assessed in the youngest or most ill children), but not for the PAED.
Method
Setting and patients Data pertained to a cohort of critically ill (defined as a threatening failure of the brain, heart, or lungs) pediatric patients, admitted from November 2006 until February 2010. The setting was a tertiary eight-bed PICU at Maastricht University Medical Centre (MUMC?) in the Netherlands. All non-electively admitted patients to the PICU between the ages of 1 and 17 years, ventilated and non-ventilated, were included. The sampling frame also included patients who had been admitted to the PICU after an elective surgical procedure and who were still at the PICU after 48 h. Patients were excluded if they were (1) non-Dutch speakers and (2) not in need of intensive care. As the information necessary to rate the DRS/PAED instruments is collected as part of routine clinical care at the PICU, under Dutch law the study did not fall under the remit of the medical ethics committee, provided data anonymity is preserved at all times.
Measurement procedures
The Pediatric Anesthesia Emergence Delirium (PAED) scale The PAED was developed by Sikich and Lerman [14] in order to trace emergence delirium (ED). ED is defined as a disorder occurring during awakening from anesthesia in the immediate postoperative period. It is characterized by confusion, hallucinations, delusions and can be accompanied by moaning, restlessness, and agitation. The scale consists of five items which are rated on a 5-point scale ranging from 'extremely' to 'not at all.' A score of 10 or higher indicates diagnosis of delirium [14]. For the purpose of the present analyses, a maximum of two missing items (2/5 = 40%) was allowed a priori; a weighted sum score was computed in the case of one or two missing items.
The Delirium Rating Scale
The DRS-88 was developed by Trzepacz and colleagues [17]. It consists of ten items encompassing a range of areas which can be affected by delirium such as hallucinations, psychomotor behavior, cognitive status, sleep-wake cycle disturbance, and lability of mood. A score of 10 or higher indicates a diagnosis of delirium [18,19]. For the purpose of the present analyses, a maximum of four missing items (4/10 = 40%) was allowed a priori; a weighted sum score was computed in the case of four or fewer missing items.
The Delirium Rating Scale-Revised (DRS-R-98)
The DRS-R-98 is a revised form of the original DRS and consists of 16 items, of which the last three are optional and scored only when there is suspicion of delirium (score of 15.25 or higher on the first 13 items); the three optional items serve to differentiate between delirium and other disorders. For the purpose of the present analyses, a maximum of five missing items was allowed a priori if only the first 13 items were scored (5/13 = 39%) and a weighted sum score was computed in the case of five or fewer missing items. If all 16 items were scored, a maximum of 6 missing items was allowed (6/16 = 38%) and a weighted sum score was computed in cases of 6 or fewer missing items [15].
Clinical/expert judgment of the gold standard
The PICU pediatric neuropsychiatrist (JS) routinely examined all children using the criteria of the revised fourth edition of the DSM-IV [1]. He reviewed and discussed all available information and opinions (from parents, nurses, and the medical charts) with the other members of the PICU team before making a final diagnosis [7,12]. He also made further categorizations into the hyperactive, hypoactive, and mixed states of PD.
Procedure
Patients at our PICU are routinely assessed for the presence of delirium under circumstances of quality improvement in clinical care since 2005, using a schedule of twice-daily examinations. The PAED, DRS-88, and DRS-R-98 are completed during the first assessment, which is carried out by the intern over the course of their attachment to the unit, having received on-site clinical training by the pediatric neuropsychiatrist of the unit. The second assessment was carried out-within 3 h-by the pediatric neuropsychiatrist, who was unaware of the screening results. PAED, DRS-88, and DRS-R-98 were completed once a day. Once PD was diagnosed, appropriate management was initiated in collaboration with the PICU team.
Statistical analysis
Data were analyzed using STATA, version 11.0 [20]. For each patient, a single assessment day during hospitalization was selected. For patients diagnosed with PD, PAED, DRS-88, and DRS-R-98 assessments on the first day of the PD gold standard diagnosis were selected; for patients without PD, PAED, DRS-88, and DRS-R-98 assessments were randomly selected in such a way that groups were frequency-matched on number of admission days until assessment. Exclusion criteria for analysis were unconsciousness or deep sedation according to the clinical judgment of the responsible intensivist treating the patient.
A patient was regarded non-rateable on a scale if more than the maximum allowed number of items was missing, precluding diagnosis. Non-rateability of items was examined as a function of age using one-way analysis of variance (ANOVA, one-tailed).
Three psychometric properties of the PAED, DRS-88, and DRS-R-98 (with exclusion of the three diagnostic items of the last scale, because scoring of these items is optional) were evaluated: (1) internal consistency, expressed as Cronbach's alpha; (2) number and proportion of items not rateable; (3) discriminative ability, expressed as the area under the receiver operating characteristic (ROC) curve, was calculated per instrument item and for instrument sum scores, with 95% confidence intervals (95% CI). ROC curves allow for exploration of the relationship between the sensitivity and specificity of a continuous predictor (in this case the scores for each item on a scale) and a dichotomous outcome (in this case PD) [21]. Instruments with the largest area under the curve (AUC) are most accurate in distinguishing between patients with and without delirium.
Given the fact that PD is a serious condition requiring immediate intervention, sensitivity was given more weight than specificity in determining the optimal cutoff scores for PD. Cutoff scores for optimal sensitivity can be inferred from the curve. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for the diagnosis of delirium were calculated, as well as the likelihood ratio for a positive test result (the ratio between true and false positive rate).
Population characteristics
From November 2006 to February 2010, 182 acute, nonelectively admitted patients were included. Twenty-eight patients without a diagnosis of delirium were excluded because of deep sedation or coma on the assessment day selected for the analysis. Of the remaining 154 patients, 26 (16.9%) were diagnosed with delirium. Mean age in PD patients was 7.2 years [SD (standard deviation) = 5.4] compared to 6.7 years (SD = 5.0) in non-PD patients. Two routinely used illness severity indicators are the pediatric index of mortality (PIM) and the pediatric risk of mortality (PRISM II). Mean PIM score in the current sample was 8.8 (SD = 16.7) in patients with and 3.6 (SD = 5.6) in patients without delirium. Mean PRISM II score was 8.8 (SD = 19.2) in patients with and 3.6 (SD = 5.8) in patients without delirium. Mean length of stay in the PD group was 7.4 days (SD = 4.9) compared to 3.9 days (SD = 3.3) in the non-PD group. In both groups, the most common primary PICU indication was a respiratory disorder, followed by neurological disorders (Table 1).
Of the 144 cases in which the PAED could be assessed, 23 were diagnosed with delirium by the gold standard, and 21 of these were correctly identified by the PAED ( Table 2). The PPV and NPV were 91.3 and 98.3%, respectively. The LR for a positive diagnosis was 55.2 and the AUC was 0.99 (Fig. 1). The optimum cutoff score was 8 (sensitivity = 100%; specificity = 92.6%). Secondly, out of 103 cases in which the DRS-88 could be assessed, 12 were diagnosed with delirium by the gold standard. Eleven of these patients were correctly identified by the DRS-88. The PPV and NPV were 98.9 and 100%, respectively (Table 2). Finally, out of 73 cases in which the DRS-R-98 could be assessed, 4 were diagnosed with delirium by the gold standard. The DRS-R-98 identified 3. The PPV and NPV were 100 and 98.6%, respectively (Table 2).
Per-item analysis
Cronbach's alpha was 0.89 for the PAED, 0.69 for the DRS-88, and 0.57 for the DRS-R-98. All items of the Tables 4 and 5. However, in cases in which these items could be rated, they demonstrated high discriminative ability, with the exception of two items: delusions (both scales) and longterm memory (DRS-R-98).
Discussion
To our knowledge, this is the first prospective study evaluating the utility of the DRS and the PAED as assessment tools for PD in a PICU setting. The PAED was rateable most often (93.5%) and proved to be excellent in discriminating between patients with and without PD (AUC = 0.99). This is in line with the findings of two recent studies which both found an AUC of 0.98 for the PAED in detecting ED in a post-anesthesia setting [22,23]. It seems, therefore, that the PAED can be used in both settings. Both versions of the DRS could not be rated as frequently as the PAED, as these scales contain numerous items that are difficult to assess in young children, especially regarding cognition. As the largest part of the PICU population consists of children under the age of 3, this is a severe drawback [7]. In addition, these tools require substantially more time to score. Taken together, in this study, the DRS scales were found to be far less suitable for routine PD assessment at a PICU. a Visuospatial ability was either scored as being normal or unrateable (due to sedation, severity of illness, and young age). Therefore, the AUC and CI of this item could not be estimated b Short-term and long-term memory were scored as being normal or unrateable (due to sedation, severity of illness, and young age) and rarely as affected. Therefore, their standard error (SE) and thus their CI could not be estimated However, the PAED is not flawless: there were two false negatives (1 hypoactive and 1 hyperactive) that were probably due to the fluctuating nature and/or the subtle presentation of delirium and two false positives (1 hyperactive and 1 mixed). In these cases, it subsequently appeared that the presentations (emotional and/or behavioral disturbances) were better explained by other somatic complications.
Taken together, the strengths of this study are (1) it was executed during daily regular clinical care, and (2) it also included the (near) majority of the PICU children, namely those in the age range greater than 1 and less than 5 years. This broad inclusion could explain the incidence of PD (16.9%) in our study compared to the 13% found by Smith and colleagues [13] who only included children above the age of 5 years. However, because the pCAM-ICU is relatively new we did not have sufficient data points to evaluate its use.
Some limitations are also apparent. Firstly, data for 40 of 182 patients (22%) were assessed in retrospect, because patients were not observed during weekends or holidays. In these cases, we had to rely on collateral information from parents, nurses, intensivists, and child neurologists to reconstruct ratings. However, there were no significant differences between the results with or without retrospective data (data not shown). Second, given the fluctuating nature of delirium, it could be that some cases were missed. Thirdly, only one observer rated the PAED each day. Because the PAED contains an element of subjectivity, depending on what a given observer defines as 'a little bit' or 'extreme' in rating a given symptom, inflated random error may have been the result. More objective anchors should be developed for the PAED to overcome this problem. Unfortunately, reliability analysis was not possible given that the PICU staff over the study period was not yet participating as rater of the different instruments-this is the topic of a current study. Fourth, this is one of the first psychometric studies conducted in a PICU setting [24] where a higher proportion of missing data can be expected because assessments are complicated by factors such as severity of critical illness, young age, mechanical ventilation, sedation, and limited amount of time available for assessments. Because listwise deletion would lead to too much data loss we used a conservative strategy and deleted subjects for which less than 60% of the items were scored for a specific scale. Missing data were corrected using a weighted sum procedure. Fifth, the findings suggest that the incidence of delirium is much lower than in adults, especially with respect to the hypoactive subtype. The difficulties with the assessment of delirium in young and/or non-verbal patients could be an explanation for the relatively low rate of PD; in addition, the exclusive reliance of the PAED on subtle behavioral alterations may explain the relatively low rate of hypoactive delirium. Finally, replications of the current findings are necessary as it remains to be seen whether results can be generalized to other PICU settings. A sufficient level of involvement of child psychiatrists and their trainees is therefore recommended.
Conclusion
Given (1) the potentially dangerous nature of delirium, (2) the high workload at the PICU, (3) a shortage of pediatric neuropsychiatrists, and (4) the fact that 75% of the children at the PICU are below the age of 3 years, an instrument is needed that is easy to use at the bedside, not time-consuming, and suitable for young children as well. The PAED seems to fulfill these requirements. It is therefore concluded that the PAED should be used for PD screening at the PICU, preferably multiple times per day. Whenever a score of 8 or higher (or other scores, depending on the patient mix of the PICU in question) is obtained for a patient, a pediatric neuropsychiatrist should be consulted to assess whether the child is delirious or not. In this way, the likelihood that PD will be missed may be reduced. Future studies could be directed towards the development of an improved screening instrument for PD, based on the PAED. | 2016-05-04T20:20:58.661Z | 2011-05-13T00:00:00.000 | {
"year": 2011,
"sha1": "b40c201b58cb01de913834131d710aa9eb4349c3",
"oa_license": "CCBYNC",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s00134-011-2244-y.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "b40c201b58cb01de913834131d710aa9eb4349c3",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
201631784 | pes2o/s2orc | v3-fos-license | Ferlin Overview: From Membrane to Cancer Biology
In mammal myocytes, endothelial cells and inner ear cells, ferlins are proteins involved in membrane processes such as fusion, recycling, endo- and exocytosis. They harbour several C2 domains allowing their interaction with phospholipids. The expression of several Ferlin genes was described as altered in several tumoural tissues. Intriguingly, beyond a simple alteration, myoferlin, otoferlin and Fer1L4 expressions were negatively correlated with patient survival in some cancer types. Therefore, it can be assumed that membrane biology is of extreme importance for cell survival and signalling, making Ferlin proteins core machinery indispensable for cancer cell adaptation to hostile environments. The evidences suggest that myoferlin, when overexpressed, enhances cancer cell proliferation, migration and metabolism by affecting various aspects of membrane biology. Targeting myoferlin using pharmacological compounds, gene transfer technology, or interfering RNA is now considered as an emerging therapeutic strategy.
Introduction
Ferlin is a family of proteins involved in vesicle fusions. To date, more than 760 articles in Pubmed refer to one of its members. Most of these publications are related to muscle biology, while less than 50 are directly related to cancer. However, the emerging idea of targeting plasma membranes [1] and the discovery of a significant correlation between Ferlin gene expression and cancer patient survival, brings attention to cancer. This review focused attention on the roles of these proteins, first in a healthy context, then in cancer.
During the maturation of spermatids to motile spermatozoa in Caenorhabditis elegans worm, large vesicles called membranous organelles fuse with the spermatid plasma membrane. This step requires a functional FER-1 protein encoded by the fer-1 gene (fertilization defective-1) [2]. When FER-1 was identified and sequenced, no other known proteins had strong resemblance to it. Subsequently, homologs were found by sequence similarity in mammals, forming a family of similar proteins now called ferlins. In humans, a first C. elegans fer-1 homolog gene was discovered and the protein encoded by this gene was named dysferlin [3]. Shortly after, a second human FER-1-Like gene was identified. The product of the gene was named otoferlin [4]. The human EST database mining revealed a dysferlin paralog called myoferlin [5,6]. Three new members joined the ferlin gene family: FER1L4, a pseudogene; FER1L5; and FER1L6. The main features of ferlins are summarized in Table 1. The dysferlin mutations were involved in Limb-Girdle muscular dystrophy 2B (LGMD2B), a autosomal recessive degenerative myopathy, and in Miyoshi muscular dystrophy 1 (MMD1), a late-onset muscular dystrophy [3,7]. The otoferlin mutations were described in the non-syndromic prelingual deafness (DFNB9) and in the auditory neuropathy autosomal recessive 1 (AUNB1) [4,8,9]. Nowadays, myoferlin and the 3 last members of the ferlin family are still not linked to human genetic diseases. However, myoferlin was proposed as a modifier protein for muscular dystrophy phenotype [5] and studies of myoferlin-null mice demonstrated impaired myoblast fusion and myofiber formation during muscle development and regeneration [10]. More recently, a truncated variant of myoferlin was associated with Limb-Girdle type muscular dystrophy and cardiomyopathy [11]. Here under, this review discusses that ferlins, mainly myoferlin, are involved in neoplastic diseases and are potential therapeutic targets.
Genomic Organization of Ferlin Gene Family
Ferlin genomic organization has not been extensively investigated. Nonetheless, valuable information was obtained from sequencing and subsequent gene annotation (www.ensembl.org). In C. elegans, fer-1 gene is approximately 8.6 kb in length and composed of 21 exons [2]. In humans, dysferlin gene (DYSF) is composed of 55 exons [12], and encodes 19 splice variant transcripts. Otoferlin gene (OTOF) contains 47 exons and encodes 7 splice variants. One of them is retaining an intronic sequence from other locus and is not coding for protein. An alternate splicing results in a neuronal-specific domain for otoferlin, regulated by the inclusion of exon 47 [8]. Myoferlin gene (MYOF), is composed of 54 exons and encodes for 9 splice variants. Four of them are not translated to protein and the shortest retains an intronic sequence. Myoferlin promoter includes several consensus-binding sites, such as for Myc, MEF2, CEBP, Sp1, AP1, and NFAT. The latter is able to bind endogenous NFATc1 and NFATc3 [13]. FER1L5 encodes 7 splice variants obtained by the arrangement of 53 exons. Five transcripts are known to encode proteins when the 2 shortest are retaining intronic sequences and do not encode protein. FER1L6 gene is composed of 41 exons and encodes a unique transcript. The main features of ferlin genes are summarized in Table 2.
Ferlin's Structure and Localization
Caenorrhabditis elegans FER-1 is a large protein rich in charged residues. Charged amino acids are distributed throughout the whole protein length such that no particularly acidic or basic domains are observed. The hydrophobicity plot described a 35 amino acid long hydrophobic region at the C-terminal end [2]. To the authors' knowledge, it has never been experimentally demonstrated. Similarity studies suggest that this region might be a transmembrane domain. FER-1 sequence analysis with Pfam protein families database [14] revealed the existence of 4 C2 domains and several other domains.
Ferlins are proteins harboring multi-C2 domains. These structural domains are~130 amino acid long independently folded modules found in several eukaryotic proteins. They were identified in classical Protein Kinase C (PKC) as the second conserved domain out of four. The typical C2 domain is composed of a beta-sandwich made of 8 beta-strands coordinating calcium ions, participating to their ability to bind phospholipids (for review [15]). However, some C2 domains have lost their capacity to bind calcium but still bind membranes [16]. A large variety of proteins containing C2 domains have been identified, and most of them are involved in membrane biology, such as vesicular transport (synaptotagmin), GTPase regulation (Ras GTPase activating protein) or lipid modification (phospholipase C) (for review [17]).
Human ferlin proteins harbour 5 to 7 C2 domains as described in the Pfam database ( Figure 1A). According to this database, in humans, 342 proteins harbour C2 domains. However, the occurrence of multiple tandem C2 domains is uncommon. Only three vertebrate protein families contain more than two C2 domains: The multiple C2 domain and transmembrane region proteins (MCTP) [18], the E-Syt (extended synaptotagmins) [19], and the ferlins. The typical feature of a C2 domain is its ability to interact with two or three calcium ions. The prototype of this domain is the C2A contained in PKC that binds phospholipids in a calcium-dependent manner. Several other distinct C2 domain subtypes, e.g. those found in PI3K and in PTEN, do not have calcium binding abilities and instead specialize in protein-protein interactions [16,17]. In classical Ca 2+ -binding C2 domains, 5 aspartate residues are involved in the ion binding [20]. Clustal omega alignment of ferlin C2 domains with PKC and synaptotagmin I C2 domains revealed that the 5 Ca 2+ -binding aspartic acids were conserved or substituted by a glutamic acid in the C2E and C2F domains of all human paralogs ( Figure 1B). The aspartic acid to glutamic acid substitution is considered as highly conservative and observed in some non-ferlin Ca 2+ -binding C2 domains [21]. Some ferlins showed more C2 domains with Ca 2+ -binding potential, e.g. dysferlin and myoferlin C2C and C2D, otoferlin C2D and fer1L6 C2D [22]. The phylogenic tree created by neighbour-joining of a Clustal omega alignment of C2 domain sequences shows that a C2 domain is more similar to others at a similar position in ortholog proteins than it is to the other C2 domains within the same protein [23]. A Clustal omega alignment reveals an evolutionary distribution of the ferlin proteins into two main subgroups ( Figure 1C): The type 1 ferlins containing a DysF domain and the type-2 ferlins without the DysF domain [22]. This domain is present in yeast peroxisomal proteins where its established function is to regulate the peroxisome size and number [24]. In mammals, despite the fact that its solution structure was resolved [25] and that many pathogenic point mutations occur in this region [26,27], the function of this domain remains unknown. ferlins containing a DysF domain and the type-2 ferlins without the DysF domain [22]. This domain is present in yeast peroxisomal proteins where its established function is to regulate the peroxisome size and number [24]. In mammals, despite the fact that its solution structure was resolved [25] and that many pathogenic point mutations occur in this region [26,27], the function of this domain remains unknown. Immunodetection of a myoferlin-haemagglutinin fusion protein in non-permeabilised COS-7 cells confirmed the presence of the C-terminal domain of the protein in the extracellular compartment [28], supporting the functionality of the putative trans-membrane region. The sublocalisation of ferlins was further studied, indicating robust membrane localisation for dysferlin, myoferlin and Fer1L6 while only low levels of otoferlin were at the plasma membrane and Fer1L5 was intracellular. Dysferlin and myoferlin were localised within the endo-lysosomal pathway accumulating in late endosomes and in recycling compartment. GFP-myoferlin fusion protein revealed that myoferlin was colocalized with lysosomal markers in NIH3T3 cells [29]. Otoferlin has been shown to move from the trans-Golgi network to the plasma membrane and inversely. Fer1L5 was cytosolic while Fer1L6 was detected in a specific sub-compartment of the trans-Golgi network compartment [30].
Ferlin's Interactions with Phospholipids
Ferlins are regarded as intrinsic membrane proteins through their putative transmembrane region. However, they can also interact with membranes by other domains. Experimentally, myoferlin C2A was the single C2 domain able to bind to phospholipid vesicles. A significant presence of the negatively charged phosphatidylserine (PS) was required for this interaction. Myoferlin C2A binding to PS-containing vesicles did not occur with calcium concentration similar to the one observed in the basal physiological condition (0.1 µM). Indeed, the half-maximal binding was observed at 1 µM [31], suggesting that the C2A domain is involved in specific processes inside the cell requiring Ca 2+ release from intracellular stock, like in Ca 2+ -regulated exocytosis. When cells are stimulated by various means, including depolarization and ligand binding, the cytosolic Ca 2+ concentration increases to the concentration up to 1 µM or more [32], similar to the one required by myoferlin C2A domain to bind lipids. It appears that dysferlin C2A domain has the same binding properties as myoferlin C2A domain. However, its half-maximal lipid binding was higher (4.5 µM) [31]. A recent publication confirmed that myoferlin and dysferlin C2A domains exhibit different Ca 2+ affinities. However, they describe myoferlin C2A domain with a lower Ca 2+ affinity than the dysferlin homolog C2 domain, and a marginal binding of myoferlin C2A domain to phospholipid mixture containing PS [33]. The binding of dysferlin C2A to PS was confirmed and extended to several phosphoinositide monophosphates in a Ca 2+ -dependent fashion. Therrien et al. observed that all remaining dysferlin C2 domains were able to bind to PS but independently of Ca 2+ [34]. The laurdan fluorescence emission experiments suggest that dysferlin and myoferlin contribute to increase the lipid order in lipid vesicles. The magnitude of this observation was calcium-enhanced and C2 domains within both N-and C-termini of ferlins influenced lipid packing. The experiments conducted with individual recombinant ferlin's C2A-C domains demonstrated that all of them are able to increase lipid order [35].
The authors described in the first part of this review the conservation of the 5 Ca 2+ -binding aspartate residues in the C2D-F domains of otoferlin making them putative Ca 2+ -binding sites. In addition to its C2D-F domains, otoferlin is also able to bind Ca 2+ via its C2B and C2C domains [36]. Despite the fact that C2A domain from otoferlin does not possess all five aspartate residues, its ability to bind Ca 2+ is still under debate. Therrien and colleagues showed that otoferlin C2A domain can bind PS in a Ca 2+ -dependent fashion, suggesting an interaction with this ion [34]. This interaction was confirmed by a direct measure of otoferlin-binding to liposomes in the presence of Ca 2+ (1 mM). Moreover, C2A-C domains seem to bind lipids also under calcium free conditions [36]. At the opposite, a spectroscopy analysis indicates that otoferlin C2A domain is unable to coordinate Ca 2+ ion [37].
Floatation assays were unable to confirm the interaction between otoferlin C2A and lipids. This may be due to the presence of a shorter membrane-interacting loop at the top of the domain [37]. As for dysferlin and myoferlin, otoferlin increases lipid order in vesicles. However, its C2A does not participate to the phenomenon [35].
Ferlin proteins contain also a FerA domain recently described as a four-helix bundle fold with its own Ca 2+ -dependent phospholipid-binding activity [38].
In Mammal Muscle Cells
Dysferlin and myoferlin have a specific temporal pattern of expression in an in vitro model of muscle development. Myoferlin was highly expressed in myoblasts that have elongated prior to fusion to syncytial myotubes. After fusion, myoferlin expression was decreased. The dysferlin expression increased concomitantly with the fusion and maturation of myotubes [31]. A proteomic analysis revealed the interacting partners of dysferlin during muscle differentiation [39]. It appeared that the number of partners decreases during the differentiation process, while the core-set of partners is large (115 proteins). Surprisingly, the dysferlin homolog myoferlin was consistently co-immunoprecipitated with dysferlin. The gene ontology analysis of the core-set proteins indicates that the highest ranked clusters are related to vesicle trafficking. In the C2C12 myoblast model, immunoprecipitation experiments showed that myoferlin interacts with the Eps15 Homology Domain 2 (EHD2) apparently through a NPF (asparagine-proline-phenylalanine) motif in its C2B domain [40]. EHD2 has been implicated in endocytic recycling. It was inferred that the interaction between EHD2 and myoferlin might indirectly regulate disassembly or reorganization of the cytoskeleton that accompanies myoblast fusion.
Dysferlin-null mice develop a slowly progressive muscular dystrophy with a loss of plasma membrane integrity. The presence of a stable and functional dystrophin-glycoprotein complex (DGC), involved in muscle injury-susceptibility when altered, suggests that dysferlin has a role in sarcolemma repair process. This was confirmed in dysferlin-null mice by a markedly delayed membrane resealing, even in the presence of Ca 2+ [41]. Pharmacological experiments conducted in skeletal muscles demonstrated that dysferlin modulates smooth reticulum Ca 2+ release and that in its absence injuries cause an increased ryanodine receptor (RyR1)-mediated Ca 2+ leak from the smooth reticulum into the cytoplasm [42]. In the SJL/J mice model of dysferlinopathy, annexin-1 and -2 co-precipitate with muscle dysferlin and co-localise at sarcolemma in an injury-dependent manner [43]. An immunofluorescence analysis of mitochondrial respiratory chain complexes in the muscles from the patients with dysferlinopathy revealed complex I-and complex IV-deficient myofibers [44]. This report is particularly interesting in light of the dysferlin_v1 alternate transcript discovered in skeletal muscle [45] and harboring a mitochondrial importation signal [39].
Intriguingly, at the site of membrane injury, only the C-terminal extremity of dysferlin was immunodetected. It was reported than dysferlin was cleaved by calpain [46], one of its interacting proteins [39]. The cleavage generate a C-terminal fragment called mini-dysferlin C72 bearing two cytoplasmic C2 domains anchored by a transmembrane domain [46]. Myoferlin expression is also up regulated in damaged myofibers and in surrounding mononuclear muscle and inflammatory cells [13]. As it was observed for dysferlin, myoferlin can be cleaved by calpain to produce a mini-myoferlin module composed of the C2E and C2F domains [47].
Membrane repair requires the accumulation and fusion of vesicles with each other and with plasma membrane at the disruption point. A role for dysferlin and myoferlin in these processes is consistent with the presence of several C2 domains and with their homology with FER-1 having a role in vesicle fusion. Moreover, mini-dysferlin and mini-myoferlin bear structural resemblance to synaptotagmin, a well-known actor in synaptic vesicle fusion with the presynaptic membrane [48].
In mouse skeletal muscle, myoferlin was found at the nuclear and plasma membrane [5]. It is highly expressed in myoblasts before their fusion to myotubes [10,31] and found to be highly concentrated at the site of apposed myoblast and myotube membranes, and at site of contact between two myotubes [10]. Myoblast fusion requires a Ca 2+ concentration increase to 1.4 µM [49], similar to the one reported for myoferlin C2A binding to phospholipids [31]. Myoferlin-null mice myoblasts show impaired fusion in vitro, producing mice with smaller muscles and smaller myofibers in vivo [10]. All together, these observations support a role for myoferlin in the maturation of myotubes and the formation of large myotubes that arise from the fusion of myoblasts to multinucleated myotubes.
Interestingly, myoferlin-null mice are unresponsive to IGF-1 for the myoblast fusion to the pre-existing myofibers. Mechanistic experiments indicate a defect in IGF-1 internalization and a redirection of the IGF1R to the lysosomal degradation pathway instead of recycling. As expected, myoferlin-null myoblasts lacked the IGF1-induced increase in AKT and MAPK activity downstream to IGFR [50].
The defects in myoblast fusion and muscle repair observed in myoferlin-null mice are reminiscent of what was reported in muscle lacking nuclear factor of activated T-cells (NFAT). Demonbreun and colleagues suggested that in injured myofibers, the membrane damages induce an intracellular increase of Ca 2+ concentration producing a calcineurin-dependent NFAT activation and subsequent translocation to the nucleus. The activated NFAT can therefore bind to its response element on the myoferlin promoter [13].
Using HeLa and HEK293T cell lines overexpressing ADAM-12, it was discovered that myoferlin was one of the ten most abundant interacting partners of ADAM-12 [51]. Though this was discovered in an artificial overexpressing model using cancer cells, it can be considered as pertinent in the context of muscle cell repair. Indeed, ADAM-12 is a marker of skeletal muscle regeneration interacting with the actin-binding protein α-actinin-2 in the context of myoblast fusion [52].
The differentiating myoblast C2C12 expressed Fer1L5 at the protein level with an expression pattern similar to dysferlin throughout myoblast differentiation. Fer1L5 shares with myoferlin a NPF motif in its C2B domain. As in myoferlin, this motif was described as interacting with EHD2, but also with EHD1 [53].
In mammal Inner Ear Cells
In adult mouse cochlea, otoferlin gene expression is limited to inner hair cells (IHC) [4]. In these cells, the strongest immunostaining of otoferlin was associated with the basolateral region, where the afferent synaptic contacts are located, suggesting that otoferlin is a component of the IHC presynaptic machinery. Ultrastructural observations confirmed the association of otoferlin with the synaptic vesicles. It appears that otoferlin is not necessary for the synapse formation [54], but rather regulates the Ca 2+ -induced synaptic vesicle exocytosis [36].
At molecular level, otoferlin interacts with plasma membrane t-SNARE (soluble N-ethylmaleimide-sensitive-factor attachment protein receptor) proteins (syntaxin 1 and SNAP-25) in a Ca 2+ -dependent manner [54]. Supporting this discovery, both t-SNARE proteins are known to interact with synaptotagmin I, a C2 domain harbouring protein, in the context of the classical synaptic vesicles docking [55,56]. It was reported that otoferlin relies on C2F domain for its Ca 2+ -dependent interaction with t-SNARE [57][58][59]. However, others suggest a Ca 2+ -dependent interaction through the C2C, C2D, C2E and C2F domains and a Ca 2+ -independent interaction via the C2A and C2B domains. The SNARE-mediated membrane fusion was reconstituted with proteoliposomes. This assay indicates that in presence of Ca 2+ , otoferlin accelerates the fusion process [36], suggesting that otoferlin operates as a calcium-sensor for SNARE-mediated membrane fusion.
In Mammal Endothelial Cells
Bernatchez and colleagues reported that dysferlin and myoferlin are abundant in caveolae-enriched membrane microdomains/lipid rafts (CEM/LR) isolated from human endothelial cells and are highly expressed in mouse blood vessels [28,60]. As observed for dysferlin in muscle cells, myoferlin regulates the endothelial cell membrane resealing after physical damage. In endothelial cells, myoferlin silencing reduced or abolished the ERK-1/2, JNK or PLCγ phosphorylation by VEGF, resulting from a loss of VEGFR-2 stabilization at the membrane. Indeed, myoferlin silencing caused an increase in VEGFR2 polyubiquitination, which leads to its degradation [28]. In contrast to what was observed in myoferlin-silenced endothelial cells, dysferlin gene silencing decrease neither VEGFR2 expression nor its downstream signalling. However, dysferlin-siRNA treated endothelial cells showed a near-complete inhibition of proliferation when they were sub-confluent. The proliferation decrease seems to be due to an impaired attachment rather than to cell death, as supported by adhesion assays and PECAM-1 poly-ubiquitination that leads to its degradation. Co-immunoprecipitation and co-localisation experiments support the formation of a molecular complex between dysferlin and PECAM-1. This PECAM-1 degradation leads, in dysferlin-null mice, to a blunted VEGF-induced angiogenesis [60]. Another angiogenic tyrosine kinase receptor Tie-2 (tyrosine kinase with Ig and epidermal growth factor homology domains-2) is significantly less expressed at the plasma membrane when myoferlin is silenced in endothelial cells [61]. In this case, it appears that proteasomal degradation plays a minor role in the down regulation of the receptor. Strikingly, G-protein coupled receptors (GCPR) were unaffected by the decrease of myoferlin expression, suggesting a selective effect on receptor tyrosine kinases (RTK).
It was also reported that in endothelial cells, myoferlin is required for an efficient clathrin and caveolae/raft-dependent endocytosis, is co-localized with Dynamin-2 protein [62] and that the FASL-induced lysosome fusion to plasma membrane is mediated by dysferlin C2A domain [63].
Other Mammal's Cells
Dysferlin and myoferlin are expressed in both basal and ciliated airway epithelial cells from healthy human lungs [64]. In the airway epithelial cell line (16HBE), dysferlin and myoferlin were immuno-detected at the plasma membrane, Golgi membrane and in cytoplasm but not in the nuclei. The silencing of myoferlin in these cells induces the loss of zonula occludens (ZO)-1, inducing apoptosis [64].
Myoferlin was also detected in exosomes from human eye trabecular meshwork cells [65] and in phagocytes where it participates to the fusion between lysosomes and the plasma membrane, thus promoting the release of lysosomal contents [29].
The Fer1L5 gene expression was largely restricted to the pancreas, where it was alternatively spliced by removing exon 51 [30].
Ferlins in Cancer, Potential Targets to Kill Cancer
It is clear from the data above that ferlins are consistently involved in membrane processes requiring membrane fusion, including endocytosis, exocytosis, membrane repair, recycling and remodelling. Membrane processes are of extreme importance for cell survival and signalling, making them core machinery for cancer cell adaptation to hostile environments.
Considering that ferlins have been only scarcely investigated in cancer, the authors next sought to mine publicly available databases and gain information regarding ferlin's expression or mutation in tumors. Using the FireBrowse gene expression viewer (firebrowse.org), The Cancer Genome Atlas (TCGA) RNAseq data of all ferlin's genes in neoplastic tissues were investigated in order to obtain a differential expression in comparison to their normal counterparts. It appears that all ferlin genes are modulated in several cancer types. Myoferlin and fer1l4 genes are more frequently up regulated than down regulated, while dysferlin, fer1l5, and fer1l6 are more frequently down regulated ( Figure 2).
Breast Cancer and Melanoma
A mathematical model was proposed to examine the role of myoferlin in cancer cell invasion. This model confirms the experimental observation of decreased invasion of the myoferlin-null breast MDA-MB-231 cell line, and predicts that the pro-invasion effect of myoferlin may be in large partly mediated by MMPs [98]. The model was further validated in vitro suggesting a mesenchymal to epithelial transition (MET) when myoferlin was knockdown [99,100]. Using the same cell model, Blackstone and colleagues showed that myoferlin depletion increased cell adhesion to PET substrate by enhancing focal adhesion kinase (FAK) and its associated protein paxillin (PAX) phosphorylation [101]. Interestingly, myoferlin was reported as regulating the cell migration through a TGF-β1 autocrine loop [102]. Recently, related results were reported in melanoma [103]. Myoferlin expression was first correlated with vasculogenic mimicry (VM) in patients, then its in vitro depletion in A375 cell line impaired VM, migration, and invasion by decreasing MMP-2 production.
Several evidences, obtained from normal endothelial cells, indicate that myoferlin is involved in RTKs recycling (see above). Our group showed that MDA-MB-231 and -468 cells depleted for myoferlin were unable to migrate and to undergo EMT upon EGF stimulation. The authors discovered that myoferlin depletion altered the EGFR fate after ligand binding, most probably by inhibiting the non-clathrin mediated endocytosis [104]. Unexpectedly, myoferlin seemed to be physically associated with lysosomal fraction in MCF-7 cells [105], supporting its involvement in the membrane receptor recycling.
The co-localisation of myoferlin with caveolin-1 [104], the main component of caveolae considered as a metabolic hub [106] prompted our group to investigate the implication of myoferlin in energy metabolism.
In this context, the authors showed in triple-negative breast cancer cells that myoferlin-silencing produces an accumulation of monounsaturated fatty acids (C16:1). Its depletion further decreased oxygen consumption switching the cell metabolism toward glycolysis [107]. This was the first report of the role of myoferlin in mitochondrial function and cell metabolism. A recent report describing the link between dysferlin mutations and mitochondrial respiratory complexes in muscular dysferlinopathy emerged (see above) [44]. It is also intriguing that dysferlin_v1 alternate transcript discovered in skeletal muscle [45] harbours a mitochondrial importation signal [39].
Several breast cancer cell lines and tissues showed a calpain-independent myoferlin cleavage, regardless of cell injuries and subsequent Ca 2+ influx [108]. The resulting cleaved myoferlin increases ERK phosphorylation in an overexpressing HEK293 system. It would be of interest to further study the link between mini-myoferlin and KRAS mutated cancers as ERK is a mid-pathway signalling protein in this context.
Pancreas and Colon Cancers
In pancreas adenocarcinoma (PAAD), myoferlin is overexpressed in high grade PAAD in comparison to low grade [73]. The patients with high myoferlin PAAD had a significantly worse prognosis than those with low myoferlin PAAD, with myoferlin appearing as an independent prognosis factor. The experiments undertaken with pancreatic cell lines and siRNA-mediated silencing demonstrated that myoferlin requested to maintain a high proliferation rate. The authors reported that myoferlin is a key element in VEGF exocytosis by PAAD cell lines, correlating with microvessel density in PAAD tissue [109]. Recently, it was demonstrated that myoferlin is critical to maintain mitochondrial structure and oxidative phosphorylation [110]. This discovery was extended to colon cancer where myoferlin seemed also to protect cells from p53-driven apoptosis [111]. The concept claiming that metastatic dissemination relies on oxidative phosphorylation is broadly accepted [112,113]. Based on these reports, the authors discovered that myoferlin was overexpressed in PAAD cells with a high metastatic potential, where it controls mitochondrial respiration [114].
Recently, FER1L4 methylated DNA marker in pancreatic juice has been strongly associated with pancreatic ductal adenocarcinoma suggesting its use as a biomarker for early detection [115].
Lung Cancer
In mice bearing solid LLC lung tumours, the intratumoral injection of myoferlin siRNA mixed with a lipidic vector reduced the tumour volume by 73%. The observed reduction was neither the consequence of a difference in blood vessel density nor of VEGF secretion. However, a significant reduction of the proportion of the Ki67-positive cells indicated a decrease in cell proliferation [75]. Myoferlin was reported as expressed in human non-small cell lung cancer tissues where it was correlated with VEGFR2, thyroid transcription factor (TTF)-1 and transformation-related protein (p63), especially in the low stage tumours [116].
Recently, it was suggested that long non-coding RNA Fer1L4 negatively controlled proliferation and migration of lung cancer cells, probably through the PI3K/AKT pathway [117]. The same observation was made in osteosarcoma cells [118], esophageal squamous cell carcinoma [119], and hepatocellular carcinoma [120].
Liver Cancer
In the hepatocellular carcinoma (HCC) cell line, the silencing of the transcriptional coactivator of the serum response factor (SRF), Megakaryoblastic Leukemia 1/2 (MKL1/2), induced a reduction of myoferlin gene expression. It was shown by chromatin immunoprecipitation that MKL1/2 binds effectively to the myoferlin promoter [76]. As in other cancer types, HCC required myoferlin to proliferate and perform invasion or anchorage-independent cell growth. Its depletion enhanced EGFR phosphorylation, in agreement with the concept of myoferlin being a regulator of RTK recycling.
Head and Neck Cancer
A myoferlin expression pattern was investigated in oropharyngeal squamous carcinoma (OPSCC). It was reported that myoferlin was overexpressed in 50% of the cases and significantly associated with worse survival. Moreover, HPV-negative patients had significantly higher expressions of myoferlin. A subgroup survival analysis indicates the interaction between these two parameters as HPV-negative has the worst prognosis when myoferlin is highly expressed. Nuclear myoferlin expression appeared to be highly predictive of the clinical outcome and associated with IL-6 and nanog overexpression [77]. Upon HNSCC cell line stimulation with IL-6, myoferlin dissociates from EHD2 and binds activated STAT3 to drive it in the nucleus. The observation was extended to breast cancer cell lines [69].
Gastric Cancer
Recently, a profiling study reported that FER1L4 was a long non-coding RNA (lncRNA) strongly downregulated in gastric cancer tissue [96], in plasma from gastric cancer patients [121] and in human gastric cancer cell lines [122]. In gastric cancer tissues, FER1L4 lncRNA was associated with the tumour diameter, differentiation state, tumour classification, invasion, metastasis, TNM stage and serum CA72-4. Interestingly, the abundance of this lncRNA decreases in plasma shortly after surgery [121]. The same team reported that the FER1L4 lncRNA is a target of miR-106a-5p [122,123]. The cell depletion in FER1L4 lncRNA resulted in an increase in miR-106a-5p and in a decrease of its endogenous target PTEN, suggesting a competing endogenous RNA (ceRNA) [124] role for FER1L4 lncRNA [122]. The control of miR-106a-5p by FER1L4 lncRNA was extended to colon cancer [125] and HCC [126], while it was described over miR-18a-5p in osteosarcoma [127].
Gynecological Cancers
Lnc Fer1L4 was briefly investigated in ovarian cancer where it was described as downregulated in cancer cells in comparison to normal ovarian epithelial cells [128]. Interestingly, the Fer1L4 expression correlates negatively with the paclitaxel resistance and its re-expression restore the paclitaxel sensitivity through the inhibition of a MAPK signalling pathway.
Conclusions
This review clearly shows that all ferlin proteins are membrane-based molecular actors sharing structural similarities. Far beyond their well-described involvement in physiological membrane fusion, several correlations apparently link ferlins, and most particularly myoferlin, to cancer prognosis. However, further investigations are still needed to discover the direct link between myoferlin and cancer biology. Encouragingly, there are many indications that myoferlin depletion interferes with growth factor exocytosis, surface receptor fate determination, exosome composition, and metabolism, indicating the future research axes.
Self-sufficiency in growth factor signalling is a hallmark of cancer cells. Cancer cells overproduce the growth factor to stimulate unregulated proliferation in an autocrine, juxtacrine or paracrine fashions. In this context, myoferlin could be considered as a cancer growth promoter as it helps the exocytosis of the growth factors, at least VEGF. In normal cells, myoferlin was described as involved in receptor tyrosine kinase (EGFR and VEGFR) recycling or expression, allowing as such, the cell response to the growth factors. Knowing that some cancer cells exhibit mutations in tyrosine kinase receptors, which lead to a constitutive receptor activation triggering the downstream pathways, it can be speculated that myoferlin depletion could impede cell proliferation in these cases. This role was indeed described in breast cancer cells [104].
Exosomes are small extracellular vesicles released on exocytosis of multivesicular bodies filled with intraluminal vesicles. They represent an important role in intercellular communication, serving as carrier for the transfer of miRNA and proteins between cells. The exosomes are increasingly described as cancer biomarkers [129] and involved in the preparation of the tumour microenvironment [130].
Interestingly, myoferlin was demonstrated to be present in exosomes isolated from several cancer cell types. However, the biological significance of this localization has still to be investigated.
Metabolism recently integrated the hallmarks of cancer [131], and mitochondria were recognised as key players in cancer metabolism [132]. The indications that myoferlin is necessary for optimal mitochondrial function is a promising avenue in the search for an innovative therapy.
Myoferlin, being overexpressed in several cancer types, offers very promising advantages for cancer diagnosis and targeting. Targeting myoferlin at the expression or functional levels remains, however, the next challenge. Interestingly, recent studies identified new small compounds interacting with the myoferlin C2D domain and demonstrating promising anti-tumoral/metastasis properties in breast and pancreas cancer [133,134].
Gene transfer strategies have undergone profound development in recent years and this is particularly applicable for recessive disorders. The adeno-associated virus (AAV) is a non-pathogenic vector used in a treatment strategy aiming at delivering full-length dysferlin or shorter variants to skeletal muscle in dysferlin-null mice. Several well documented reports demonstrate an improvement in the outcome measures after dysferlin gene therapy [135][136][137][138]. Similar AAV vectors were used as a gene delivery system in cancer [139,140], allowing the dream of myoferlin negative-dominant delivery to cancer cells. Moreover, the sleeping beauty transposon system [141] may overcome some of the limitations associated with viral gene transfer vectors and transient non-viral gene delivery approaches that are being used in the majority of ongoing clinical trials.
Statistical Methods
The multivariate Cox regressions (Table 3) were performed with the coxph function from the R survival library. For each cancer and data type, OncoLnc attempted to construct a model with gene expression, sex, age, and grade or histology as multivariates [90]. The clinical information was obtained from TCGA and only patients who contained all the necessary clinical information were included in the analysis. The patients were split into low and high expressing according to the median gene expression.
Funding: This research was funded by "Fonds Léon Fredericq" and by the "Patrimoine de l'Université de Liège". | 2019-08-25T13:04:21.701Z | 2019-08-22T00:00:00.000 | {
"year": 2019,
"sha1": "6d492e5e7fb2fb196a357620e4eb476e0415f4e4",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2073-4409/8/9/954/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "300afcf5b9a5630ef096061acdc0631aeb9ad2b9",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine",
"Biology"
]
} |
225965950 | pes2o/s2orc | v3-fos-license | Simulation Experiment on Drag Reduction Effect of the V-shaped Groove Structure of Bionic Water Strider Robot
The drag reduction of the groove surface is inspired by the shark skin structures, the groove structure is transplanted onto the water-repellent robot and subjected to drag reduction simulation experiments. The effect of the width, height and apex angle of the V-shaped groove on the drag reduction coefficient is studied. When the dimensionless parameters of the width, height and apex angle of the groove are 16, 11and 70°, the best drag reduction effect is obtained, the drag reduction rate is about 6.5%. The characteristic parameters obtained through simulation experiments can provide a design basis for the drag reduction research of bionic water hammer robot.
Introduction
The drag reduction of the groove surface is inspired by the shark skin structure, the shark's skin is not smooth, but has a certain shape of the groove texture. This non-smooth skin structure is generally considered to be an important reason why sharks can swim at high speed in the water. The NASA Research Center found that the micro-trench can effectively reduce the frictional resistance of the wall surface, which break through the traditional way of thinking that the smoother the surface is the smaller the resistance [1][2][3]. Bechert compared the grooves of triangles, rectangles and semicircles to show that the V-grooves are considered to be the most favorable structure for drag reduction [4][5][6]. This paper will study the characteristic dimensions of V-grooves. It is planned to transplant the groove structure to bionic water strider robot to reduce the resistance. Bionic water strider robot has a wide range of applications in water detection and environmental monitoring.
Experimental program
In this paper, the simulation experiment is carried out by Computational Fluid Dynamics software (CFD), which can obtain the drag coefficient of the bionic water strider robot. The general steps consist of four parts, which are to establish three-dimensional model, discrete grid, fluent simulation and analysis of experimental results [7,8].
Establish three-dimensional model
The foot of the bionic water strider robot is a rectangular parallelepiped with a length, width and height of 50 mm × 10 mm × 2 mm. The bottom surface of the rectangular parallelepiped has a V-shaped groove, and its characteristic dimensions are shown in Fig. 1. The V-shaped groove depth is h, the groove pitch is S, and the apex angle of the groove is α. Due to the small scale of the V-shaped groove structure, the simulation experiment encounter the problems of large calculation amount and difficulty in mesh division. Therefore, the parameters of the groove depth and the groove pitch are infinite according to the equations (1) (2) and (3).
Where h is the groove depth, h+ is the dimensionless depth of the groove, s is the groove width, s+ is the groove dimensionless width, v is the velocity of the fluid, and μτ is the wall friction velocity.
Discrete grid
When dividing the grid, create the part of the bottom surface groove, the upper part is part of smooth , and the part of wall on both sides, the minimum mesh size of the groove surface is set to 0.05, the other faces are set to 1, and the final mesh is divided into about 500,000, and the mesh quality is checked.
Fluent simulation
The divided mesh is imported into fluent for numerical simulation [9,10], and the turbulence intensity is set to 5%. The model is solved by K-WSST and second-order upwind calculation, and the control convergence precision is set to 10 -5 .
Effect of groove spacing on drag coefficient
For the V-shaped groove structure, firstly study the influence of the groove spacing on the drag reduction rate, the fluid velocity is 0.5 m/s, the dimensionless depth of the trench is h+=10, and the corresponding groove dimensionless spacing is 10 simulation experiments, the same groove depth and resistance of V-grooves at different groove spacing are known. The coefficient is shown in Fig 2. It can be seen that the V-shaped groove has a smaller drag reduction coefficient at s+=16.
Effect of groove height on drag coefficient
The effect of groove depth on the drag reduction rate is studied. The fluid velocity is 0.5 m/s, the dimensionless spacing of the groove is s+=15, and the corresponding groove dimensionless depth is 6, 7, 8, 9, 10, 11, 12, 13, 14, and 15. Through simulation experiments, It can be seen that the same groove spacing, the drag reduction ratio of the V-shaped groove at different groove depths, and the drag coefficient are as shown in Fig. 3. When h+=11, V-shaped groove has a smaller drag reduction coefficient.
Effect of apex angle on drag coefficient
After selecting the two feature sizes of the trench depth h and the trench pitch s, the influence of the top angle α of the V-shaped trench on the drag reduction coefficient is analyzed. Here we assume that the top of the trench is symmetrical at the bottom. We set the groove top angles to 10°, 20°, 30°, 40°, 50°, 60°, 70°, 80°, 90°, 100°, 110°, and 120°, respectively. Simulation experiment, h+=11, s+=16, v=0.5m/s. In the simulation results, the resistance coefficients corresponding to different top angle Fig. 4, in the V-shaped groove. When the apex angle of the groove d is 70°, the V-shaped groove has a small drag reduction coefficient.
Conclusion
The V-shaped groove achieves the best drag reduction effect when the depth dimensionless parameter is 11, the width dimensionless parameter is 16, and the apex angle is 70°, compared with a smooth surface, the drag reduction rate is about 6.5%. The characteristic parameters obtained through simulation experiments can provide a design basis for the drag reduction research of bionic water hammer robot. | 2020-06-25T09:06:51.934Z | 2020-04-01T00:00:00.000 | {
"year": 2020,
"sha1": "1d8189fd8aef45f405c14e742fa937b8df0febea",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1742-6596/1533/3/032091",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "0e14fc783075ad7ab2e8d733714d0f9f42ef2644",
"s2fieldsofstudy": [
"Engineering"
],
"extfieldsofstudy": [
"Materials Science"
]
} |
263694844 | pes2o/s2orc | v3-fos-license | Implementing Nutritional Strategies in the Workplace
The lifestyle of a health professional is mired with long hours, undue stress, shift work, poor sleep, inactivity, and unhealthy eating, despite their knowledge and expertise in what habits promote or preserve health. Healthcare workplaces are largely environments with underfunding, poor infrastructure, overcrowding
Introduction
Why promote healthy eating?Promoting healthy eating and physical activity in the workplace enhances productivity -a balanced diet helps employees maintain focus, energy, and efficiency throughout their workday. 3hen employees feel physically and mentally healthy, they are more likely to have a positive attitude towards their work and colleagues. 4A nutritious diet can strengthen the immune system, making employees less prone to illnesses.In addition to physical exercise, a healthy diet may lead to decreased absenteeism and fewer disruptions in work schedules. 5rganising activities centred around healthy eating, such as group lunches or potlucks featuring nutritious options, can encourage camaraderie and strengthen team bonds.
Healthy dietary patterns were defined in the 2015 Dietary Guidelines Advisory Committee Scientific Report as diets that are high in fruits, vegetables, whole grains, low and non-fat dairy, and lean protein.
Other characteristics of healthy dietary patterns are that they are low in saturated fat, trans fat, sodium and added sugars.Encouraging healthy eating habits can help prevent chronic diseases like obesity, diabetes, and heart disease. 6By supporting employees' well-being, employers demonstrate their care for staff's long-term health and happiness.Practical benchmarks of lost work and performance may help employers assess the financial impact of suboptimal health.Healthier employees tend to have fewer medical expenses, which can lead to lower healthcare costs for both employees and employers. 7An organisation that promotes healthy habits demonstrates social responsibility and commitment to employee well-being.This can enhance an organisational reputation and attract, as well as retain top talent. 8The UK Parliamentary Hospital Food Standards Panel recommended that 'every hospital has a responsibility to provide the highest level of care possible for their patients and this, without question, includes the quality and nutritional value of the food that is served and eaten.' • have some dairy or dairy alternatives (such as calciumfortified soya drinks), and choose lower fat and lower sugar options.• eat some beans, pulses, fish, eggs, meat, and other proteins.This includes -two portions of fish every week, one of which should be oily.If consuming more than 90g of red or processed meat per day, try to cut down to no more than 70g on average.
• choose unsaturated oils and spreads and eat them in small amounts.
Food Labelling
Food labelling is an effective tool for promoting healthier food choices among employees.Research shows that providing clear and easy-to-understand nutrition information on food items can help employees make informed decisions about their food choices, such as provided by NHS Inform. 11,12There is evidence that food labelling decreases consumer intake of energy by 6.6%, total fat by 10.6%, and other unhealthy dietary options by while increasing vegetable consumption by 13.5%.For the food industry labelling decreased product contents of sodium by 8.9% and artificial trans-fat by 64.3%.No significant impact was identified by label placement or type, duration, labelled product, region, population, voluntary or legislative approaches, combined intervention components, study design, or quality. 13However, the effectiveness of food labelling relies on employee engagement and understanding of the nutrition information provided.While nutrition knowledge is a prerequisite for label reading behaviour, trust and attitude are also important.Organisations should ensure that employees are educated on how to read and interpret food labels and incorporate this information into their daily food choices.
Catering Initiatives
Hospital catering has two-pronged challenges -on the one hand, most elderly patients suffer from undernutrition or patients with a range of conditions lose their appetite as a result of their illness, hence hospitals have to cater to different content and texture of food served to their patients (multisensory flavour experiences) 14 ; while healthy staff have limited access to food from external sources and need to be presented with options which are nutritious but healthy, in a competitive marketplace.Thus, workplace catering services focussed on staff can play a pivotal role in promoting healthier food options.Studies have shown that implementing healthy catering guidelines, such as offering more fruits, vegetables, whole grains, lean proteins, and low-fat dairy products, can improve the overall nutritional quality of meals consumed by employees. 15
Healthier Vending Machines
With most people in employment spending 60% of their waking hours in a work environment -where they will typically consume drinks, snacks and at least one meal -the workplace provides an opportunity to influence positive eating habits.The BMA Facilities charter recommends that between 11 pm and 7 am hospital staff should be offered access to healthy vending machines.Vending machines provided 922.2 million products, including snacks and cold drinks, in the UK in 2015.Almost 85% of vending machines are found in the workplace, and commonly in NHS working environments.With the NHS being the UK's biggest employer, healthier vending can offer a significant opportunity to encourage healthy consumption habits amongst NHS staff, as well as patients and visitors.In 2015, only 39% of vended cold drinks in the UK were reported to be "low sugar, diet or water", and 14% of snacks were labelled as "healthier". 16Thus, replacing traditional vending machine offerings with healthier snack options can improve employees' dietary choices.Research suggests that employees are more likely to choose healthier snacks when they are readily available in vending machines. 17However, healthier vending machine initiatives may face challenges in terms of product availability, cost, and consumer preferences.
Prompts, Educational & Cost Incentives
A simple, theory-based point-of-purchase system of prompts can produce small but significant reductions in the energy content of snack purchases from hospital shops. 18Providing educational resources, such as healthy cookbooks and sandwich containers, can encourage employees to adopt healthier eating habits.These resources offer practical guidance on preparing nutritious meals and snacks at home, which can contribute to improved overall diet quality. 19The effectiveness of these resources relies on employees' motivation and willingness to engage with the information provided and there is evidence that cost and healthy incentives can be sustained beyond the period of the offer. 20
Messaging
Communication strategies, such as posters, newsletters, and digital messaging platforms, can raise awareness about the importance of healthy eating and promote positive behaviour change among employees.Research indicates that consistent and engaging messaging can lead to improvements in employees' dietary habits. 21Organisations should tailor their messaging tools to their specific workplace culture and employee population to maximize effectiveness.
Access to Drinking Water
Providing employees with reusable water bottles or easy access to drinking water can encourage increased water consumption and reduce the intake of sugar-sweetened beverages.Studies have shown that increasing water consumption is positively associated with better weight management and overall health. 22
Engaging Champions
Role modelling health behaviours through engaging champions can foster a supportive environment for healthy eating in the workplace.Champions can be influential employees who advocate for and demonstrate healthy behaviours, inspiring others to follow suit. 23commendations for Organisations 1. Implement a combination of interventions to address multiple aspects of healthy eating in the workplace.2. Ensure that interventions are tailored to the specific needs and preferences of employees.Provide ongoing education and support to help employees engage with and maintain healthy eating behaviours.3. Monitor and evaluate the success of interventions to identify areas for improvement and adapt strategies accordingly.4. Share guidelines and details about nutritional values, ingredients, and cooking methods when relevant. 5. Participate in community healthy eating initiatives and promote them through posters, leaflets, intranets, screen savers, payslips, and employee meetings.6. Offer healthy packed lunch recipes and organize 'taste and try' events.7. Organise themed healthy eating days or weeks, such as focusing on breakfast, pasta, or fruit.8. Conduct educational sessions or workshops on nutrition, healthy eating, and cooking for employees.9. Support employees striving to maintain a healthy diet or lose weight by encouraging the formation of support groups or enrolment in programs like Weight Watchers.10.Collaborate with external nutrition or dietetics experts to provide seminars on healthy eating and nutrition.
Conclusion
Calder, P. C. Defining a Healthy Diet: Evidence for the Role of Contemporary Dietary Patterns in Recommendations for Individuals 1. Engage with workplace interventions and educational resources to improve knowledge and skills related to healthy eating.2.Seek support from colleagues, friends, and family to maintain motivation and accountability for healthy eating habits.3. Set realistic goals for improving dietary habits and track progress over time.4. Prioritise self-care and stress management as these factors can influence dietary choices and overall health. | 2023-10-06T15:20:37.670Z | 2023-10-03T00:00:00.000 | {
"year": 2023,
"sha1": "032db6461551226513bcc7ee180c78151257aa77",
"oa_license": "CCBY",
"oa_url": "http://sushrutajnl.net/index.php/sushruta/article/download/165/310",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "1385af79945140e1b56ad3903a012980b3f61eab",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
11074761 | pes2o/s2orc | v3-fos-license | Palmitoleate inhibits insulin transcription by activating the ERK1/2 pathway in rat pancreatic β-cells
The aim of the present study was to evaluate the effects of palmitoleate on insulin secretion and insulin mRNA levels, and to investigate the transcriptional regulation of insulin. INS-1 rat insulinoma cells were treated with palmitoleate in the presence of high glucose, and the amount of secreted insulin was measured via radioimmunoassay. Reverse transcription-quantitative polymerase chain reaction was performed to evaluate the mRNA levels of insulin and pancreatic and duodenal homeobox 1 (PDX1) under palmitoleate treatment. The levels of PDX1, peroxisome proliferator-activated receptor gamma (PPARG), extracellular signal-regulated kinase (ERK)1/2 and phosphorylated ERK1/2 were measured using western blot analysis. Low concentrations of palmitoleate significantly induced insulin secretion (P=0.024), whereas the mRNA levels of insulin and PDX1 were markedly reduced. However, the inhibitory effects were reversed with the addition of U0126, suggesting that the ERK1/2-mediated pathway may be the underlying mechanism responsible for palmitoleate-induced downregulation of insulin mRNA. Exposure of INS-1 cells to high glucose significantly increased the phosphorylation of ERK1/2 (P=0.039), which was further enhanced by palmitoleate (P=0.025). Exposure of INS-1 cells to high glucose significantly decreased PPARG (P=0.001), which was further decreased by the addition of palmitoleate. U0126 was able to reverse the palmitoleate-induced effects. In conclusion, the present study suggested that palmitoleate may induce insulin secretion and inhibit insulin mRNA expression in pancreatic β-cells.
Introduction
The incidence rates of metabolic dysfunctions such as diabetes, obesity, fatty liver disease and atherosclerosis are increasing, typically occurring together in the same individual and hence referred to as metabolic syndrome (1). Previous studies investigating these dysfunctions indicate that adipose tissue serves a key role in regulating the body's metabolism (2). Evidence suggests that adipose tissues in metabolic disorders produce peptides and hormones, which alter the normal homeostasis of lipids and carbohydrates. For example, in obese patients, fatty acids produced within adipose tissue may negatively impact peripheral tissues, contributing to hepatic steatosis or insulin resistance in muscles (2,3). Various research models have revealed that lipid-mediated metabolic dysfunction is largely due to exposure to the lipids in adipose tissue (2,3). The specific physiological and pathological roles of fatty acids vary depending on their chain length and degree of saturation, leading to differing effects on pancreatic β-cells (4).
Palmitate appears to be the most cytotoxic long-chain saturated fatty acid, whereas its monounsaturated equivalent, palmitoleate, is innocuous and protects β-cells against the toxicity of saturated fatty acids (5,6). Dhayal et al (7) demonstrated that the cytoprotective effect of palmitoleate may be due to mediation by a fatty acid receptor rather than direct metabolic activation. Dimopoulos et al (8) reported that palmitoleate was able to promote the uptake and utilization of glucose in normal rat L6 skeletal muscle cells and those treated with the insulin resistance-inducing palmitate. Furthermore, levels of circulating palmitoleate exhibited a positive correlation with insulin sensitivity in humans (9).
Palmitoleate is considered to be a major signaling lipid hormone, regulating metabolic activities in multiple tissues, and serum fatty acid composition in response to changes in lipid metabolism in adipose tissue (10). However, the impact of palmitoleate on insulin secretion and intracellular expression in pancreatic β-cells, and therefore its role in insulin biology, remains to be elucidated. In the present study, the rat insulinoma cell line INS-1was used to examine the direct effects of palmitoleate on insulin secretion and intracellular insulin expression. It was revealed that palmitoleate-regulated insulin gene expression is associated with the extracellular signal-regulated kinase-1/2 (ERK1/2), peroxisome proliferator-activated receptor gamma (PPARG) and pancreatic and duodenal homeobox 1 (PDX1) signaling pathways. The results of the present study provided evidence that palmitoleate modulates insulin secretion and inhibits insulin mRNA expression. INS-1 cells were maintained in RPMI-1640 medium containing 11 mM glucose supplemented with 10% (v/v) heat-inactivated FBS, 100 U/ml penicillin, 100 U/ml streptomycin, 10 mM HEPES, 1 mM sodium pyruvate and 50 µM β-mercaptoethanol (all Takara Biotechnology Co., Ltd., Dalian, China). Culture medium was replaced daily until cell confluency reached 80-90%.
Measurement of insulin secretion and intracellular insulin content.
INS-1 cells were seeded in 24-well plates and maintained in medium containing 5.5 mM of glucose for 16 h one day prior to the experiment. Cells were subsequently serum-starved in medium containing 0.5% BSA for 2 h, prior to being maintained in KRB buffer containing 115 mM NaCl, 5 mM KCl, 24 mM NaHCO 3 , 2.5 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES and 0.1% BSA for 24 h. Cells were incubated in KRB in the presence of various concentrations (0, 0.05, 0.1, 0.2, 0.3, 0.4 and 0.5 mM) of palmitoleate and insulin concentrations in the KRB buffer were measured using a rat insulin RIA kit. Plates were then washed once with PBS, intracellular insulin was extracted with 150 µl acid ethanol [a 140:57:3 (v/v/v) solution of ethanol/water/concentrated HCl] and the insulin content was determined via RIA, using rat insulin as the standard. For pharmacological inhibition of MEK1/2, cells were treated with 10 µM of U0126 for 30 min prior to palmitoleate treatment.
RNA extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR). The acid guanidinium thiocyanate-phenol-chloroform extraction procedure (10) was used for the extraction of total RNA from INS-1 cells. The quantity and quality of the extracted RNA was confirmed by absorption measurement at 260 and 280 nm using a U-2000A spectrophotometer (Hitachi, Ltd., Tokyo, Japan). The prepared RNA (10 µg) was mixed with 500 pmol of the oligo dT primer for 15 min at 68˚C. The RNA sample was reverse-transcribed at 37˚C for 15 min into first-strand cDNA in RT solution performed with 5 µl 5X PrimeScript buffer, 1 µl PrimeScript RT Enzyme Mix I, 1 ul Oligo dT Primer (50 µM), 1.5 µl random 6-mers (100 µM) and 1 µl total RNA (PrimeScript RT reagent kit; Takara Biotechnology Co., Ltd), with a total volume of 25 µl. RT-qPCR was performed using an ABI 7000 sequence detection system (Applied Biosystems; Thermo Fisher Scientific, Inc., Waltham, MA, USA). TaqMan gene expression assays were performed in accordance with the manufacturer's instructions (SYBR Premix Ex Taq; Takara Biotechnology Co., Ltd.). Primers and probes (Takara Biotechnology Co., Ltd.) used for RT-qPCR are listed in Table I. qPCR, which was optimized for a final reaction volume of 20 µl, was performed with 10 µl SYBR Premix Ex Taq, 7 µl nuclease-free water, 1 µl forward primer (10 µM), 1 µl reverse primer (10 µM) and 1 µl cDNA template. Amplification was performed under the following conditions: INS-1, 94˚C for 2 min, followed by 30 cycles of 94˚C for 15 sec and 63˚C for 1 min; PDX-1, 94˚C for 2 min, followed by 40 cycles of 94˚C for 15 sec and 63.5˚C for 1 min; actin, 94˚C for 2 min, followed by 25 cycles of 94˚C for 15 sec and 59˚C for 1 min. PCR amplification of GAPDH mRNA was conducted according to the manufacturer's instructions for the TaqMan GAPDH control reagent. Experiments were performed in triplicate and the results of the RT-qPCR were calculated using the 2 -∆∆Cq method with minor revisions (12).
Statistical analysis.
Values are expressed as the mean ± standard error of the mean. Differences were analyzed using one-way analysis of variance. Where appropriate, Student's t-test was used for the comparison of two treatments. P<0.05 was considered to indicate a statistically significant difference, as calculated using SPSS 13.0 (SPSS, Inc., Chicago, IL, USA).
Palmitoleate impacts insulin secretion and intracellular content.
In the present study, 2x10 5 normalized INS-1 cells were treated in the presence of 2.8 or 16.7 mM glucose with or without palmitoleate, and the amount of insulin secreted into the culture medium and the intracellular insulin content were measured. Under low-glucose conditions (2.8 mM), 0.2 mM palmitoleate caused no significant changes to insulin secretion into the medium or the cellular insulin content (Fig. 1A and B). However, in the presence of high glucose (16.7 mM), the effect of palmitoleate on insulin secretion in the medium appeared to be dependent on its concentration, with the maximal effect at 0.20 mM to achieve a statistically significant increase by 34% vs. the control (i.e., 0 mM palmitoleate; P=0.024; Fig. 1A). Under the same conditions, the intracellular insulin content in INS-1 cells decreased significantly by 40% vs. the control (P=0.007; Fig. 1B). These results indicated that under high-glucose conditions, the expression of insulin may be inhibited by palmitoleate treatment, although palmitoleate within a certain concentration range (around 0.20 mM) can promote insulin secretion. The palmitoleate-treated total insulin (insulin release and insulin content) was significantly decreased compared with that observed with no palmitoleate-treatment in high glucose (P=0.039; data not shown).
Palmitoleate decreases insulin mRNA levels in a dose-dependent manner. The observation that palmitoleate inhibited the intracellular insulin content in a dose-dependent manner suggested that palmitoleate may modulate insulin mRNA levels. To investigate this, insulin mRNA levels in rat INS-1 cells were measured in response to different final concentrations of palmitoleate (0.05-0.5 mM) using RT-qPCR. The results showed that palmitoleate significantly decreased insulin mRNA levels in a dose-dependent manner (P=0.011; Fig. 2).
Palmitoleate elevates high glucose-induced ERK1/2 phosphorylation in INS-1 cells. The phosphorylation level of ERK1/2 (pERK1/2), which is an intracellular kinase activated by various cytokines and chemokines, was evaluated to elucidate the mechanism by which insulin mRNA levels are downregulated in response to high glucose and palmitoleate. As demonstrated in Fig. 3, exposure of INS-1 cells to high glucose (16.7 mM) significantly increased ERK1/2 phosphorylation in comparison with low glucose (2.8 mM; P=0.039). Furthermore, with the addition of 0.2 mM palmitoleate, pERK1/2 was significantly increased (P=0.025) in cells exposed to high glucose. However, when the specific ERK1/2 inhibitor U0126 was introduced, the effects of palmitoleate were significantly reduced (P=0.002). This further supported the hypothesis that palmitoleate elevates ERK1/2 activation and suggested that the activation of ERK1/2 may be the mechanism underlying the response elicited by palmitoleate.
Palmitoleate reduces increases in high glucose-induced PDX1 mRNA in INS-1 cells.
Palmitate has been previously shown to inhibit insulin gene expression by decreasing PDX1 nuclear exclusion in MIN6 cells, and PDX1 serves a crucial role in the control of insulin gene expression in pancreatic β-cells (13,14).
To further explore the possible molecular mechanisms by which PDX1 is associated with palmitoleate-mediated inhibition of insulin expression in INS-1 cells, PDX1 mRNA levels following glucose and palmitoleate treatments in the absence or presence of U0126 were assessed. Although a 24-h exposure to high glucose (16.7 mM) induced a significant increase in PDX1and insulin mRNA levels (P=0.001), the addition of palmitoleate significantly reduced these effects (P= 0.001; Fig. 4A and B). Furthermore, when U0126 was introduced, the inhibition of PDX1 and insulin mRNA levels was reversed ( Fig. 4A and B).
Palmitoleate reduces increases in high glucose-induced PPARG and PDX1 expression in INS-1 cells. PPARG
has been demonstrated to serve a role in PDX1 mRNA expression and the transcriptional activity of PPARG is inhibited by MAPK-mediated phosphorylation of PPARG at Ser84 (15)(16)(17). This suggests that PPARG activation is an upstream event that regulates the status of PDX1 upon treatment with palmitoleate. In the present study, exposure of INS-1 cells to high glucose (16.7 mM) significantly decreased PPARG and increased PDX1 expression (P<0.005 for each), while addition of palmitoleate caused significant decreases (P<0.005; Fig. 5). As PPARG increased, PDX1 expression increased correspondingly.
Palmitoleate activates the ERK1/2 pathway in INS-1 cells under high-glucose treatment.
The observation that 0.2 mM palmitoleate increased insulin secretion (Fig. 2) and activated ERK1/2 phosphorylation in INS-1 cells under high glucose conditions (Fig. 3) indicated that phosphorylation of ERK1/2 may be the mechanism underlying palmitoleate-mediated insulin secretion. To investigate this further, insulin secretion into the medium and cellular insulin content in the normalized INS-1 cells was measured following the addition of U0126. Consistent with the previous results, under high glucose conditions, 0.2 mM palmitoleate significantly promoted insulin secretion (P=0.001), whereas intracellular insulin levels were markedly reduced. When ERK1/2 phosphorylation was blocked via the introduction of U0126, the secretion of insulin was significantly elevated (P= 0.001; Fig. 6A) and the decrease in intracellular content of insulin was significantly reversed (P=0.001; Fig. 6B). These results suggested that the activation of the ERK1/2 pathway is associated with palmitoleate-mediated cellular processes and functions as a mediator of the effect of palmitoleate treatment on insulin. Furthermore, in high glucose conditions, the total insulin (insulin released and cellular insulin content) was significantly increased following palmitoleate and U0126 treatment compared with that observed following treatment with only palmitoleate (P=0.027; data not shown).
Discussion
Palmitoleate is a newly identified adipokine that has various insulin-associated effects, including the stimulation of uptake and utilization of glucose in normal and insulin-resistant skeletal muscle cells, improvement of insulin resistance (8,9), suppression of lipid cytotoxicity and apoptosis induced by saturated fatty acids in pancreatic β-cells (10), inhibition of liver steatosis (10) and depression of adipose inflammation (13). However, the signaling pathways associated with the regulation of insulin mRNA levels in response to palmitoleate are yet to be elucidated. It has been demonstrated that free fatty acid-induced insulin secretion is glucose-dependent, and that at elevated concentrations of extracellular glucose, free fatty acids will further amplify glucose-mediated insulin secretion (14)(15)(16)(17). The present study demonstrated that the free fatty acid palmitoleate differentially regulates insulin secretion and insulin mRNA levels in the rat-derived insulin-producing cell line INS-1and rat pancreatic islets (18). Palmitoleate treatment induced an increase in insulin secretion under high glucose conditions, while its effect under low glucose conditions was minor, indicating that the glucose concentration influences insulin expression mediated by palmitoleate treatment. Although the underlying mechanisms remain to be elucidated, changes in the ERK1/PPAR/PDX1 signaling pathway were observed in the present study. A 24-h exposure of INS-1 cells to 0.05-0.5 mM palmitoleate induced a dose-biphasic increase in glucose-stimulated insulin secretion with a maximum effect at 0.2 mM. Conversely, 0.2 mM palmitoleate inhibited insulin gene expression at a high glucose concentration by impairing the glucose-induced increase in insulin mRNA levels. The 0.2 mM concentration was subsequently used to investigate the underlying mechanisms of the insulin secretion response to palmitoleate stimulation.
It is well known that G proteins dissociate into the free subunits G α and G βγ upon receptor activation, with the two components leading to different cellular signaling events (19). In the example of G protein-coupled mitogen, G βγ mediates the intracellular signaling pathway, which facilitates cell proliferation by activating p21 ras , which in turn activates p42 and MAPKs (20). The present study revealed that palmitoleate increases glucose-induced ERK1/2 phosphorylation and decreases glucose-induced PDX1 expression as well as PPARG expression. Furthermore, inhibition of ERK1/2 potentiated the expression of PDX1, insulin mRNA and PPARG.
Previous studies have demonstrated that the transcriptional activity of PPARG is suppressed by phosphorylation of Ser82 mediated by MAPK (21,22), and the results of the present study are consistent with this conclusion. It has been previously shown that in INS-1 cells, elevated PPARG promotes PDX1 transcription and expression (23,24), suggesting that the activation of PPARG is an upstream event regulating PDX1. In the present study, it was also demonstrated that the inhibition of PPARG was associated with decreased PDX1 mRNA levels and protein expression and that these effects may also be reversed by the ERK1/2 inhibitor U0126. Together, these results suggested that the ERK1/2, PPARG and PDX1 signaling pathways are implicated in the inhibitory effects of palmitoleate on insulin gene expression.
In the present study, it was demonstrated that the adipose tissue-derived hormone palmitoleate acts directly on pancreatic β-cells. The strongest induction of insulin release in INS-1 cells in the presence of high glucose was observed with a 0.2 mM palmitoleate treatment for 24 h. The results of the present study revealed the association of the ERK1/2, PPARG and PDX1 pathways in palmitoleate-mediated inhibition of glucose-induced insulin gene expression in INS-1 cells. While it remains to be determined whether the mechanisms in humans are similar, the present study may be clinically significant: If elevation of the palmitoleate concentration is able to improve insulin resistance and stimulate insulin release, the inhibitory effects on intracellular insulin expression should be taken into account. In vivo experiments to evaluate the effects of palmitoleate on insulin secretion and expression in pancreatic β-cells are currently underway.
In conclusion, the present study revealed that high glucose-induced phosphorylation of ERK1/2 is able to facilitate insulin transcription, which may be inhibited by palmitoleate. However, palmitoleate may exert its effects via multiple pathways. | 2018-04-03T01:50:27.246Z | 2017-04-18T00:00:00.000 | {
"year": 2017,
"sha1": "d829f4ed1c82e85dd16fdcb855e96c67c7257f15",
"oa_license": "CCBYNCND",
"oa_url": "https://www.spandidos-publications.com/10.3892/etm.2017.4344/download",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "d829f4ed1c82e85dd16fdcb855e96c67c7257f15",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
15191799 | pes2o/s2orc | v3-fos-license | Impact of prior statin therapy on the outcome of patients with suspected ventilator-associated pneumonia: an observational study
Introduction Ventilator-associated pneumonia (VAP) is the most commonly acquired infection in intensive care units (ICU). Its outcome is related, at least in part, to the host’s response. Statins have anti-inflammatory effects and may thus improve the outcome. We aimed to assess the impact of prior statin use in the setting of VAP. Methods A six-year cohort study was conducted in a French ICU at a teaching hospital. All of the patients with suspected VAP were included. Baseline characteristics, outcomes, statin exposure, and the description of suspected episodes were collected prospectively. The primary endpoint was 30-day mortality. Patients who were taking statins before admission to the ICU whether or not treatment was continued thereafter (‘previous users’ group) were compared to those without prior statin therapy (‘statin-naive’ group). A survival analysis using a Cox model was conducted in the whole cohort and in the subgroup of prior statin users. Results Among the 349 patients included, 93 (26.6%) had taken statins. At baseline, these patients were at higher risk of complications than statin-naive ones (for example, older, more likely to be men and to have underlying diseases, greater simplified acute physiology score II (SAPS II)). There was, however, no difference regarding severity at the time VAP was suspected (sequential organ failure assessment (SOFA): 9.0 (4.0 to 16.0) versus 8.0 (4.0 to 17.0); P = 0.11). Nonetheless, 30-day mortality in statin users was not different from that in statin-naive patients (35.5% versus 26.2%, respectively; adjusted hazard ratio (HR) = 1.23 (0.79 to 1.90) 95% confidence interval (CI); P = 0.36). In contrast, after limiting analysis to prior statin users and adjusting for potential confounders, those who continued the treatment had better survival than those who did not (HR = 0.47; (0.22 to 0.97) 95% CI; P = 0.04). Conclusions Statin continuation in prior users could provide protective effects in patients with suspected VAP.
Introduction
Ventilator-associated pneumonia (VAP) occurs in almost 30% of patients on prolonged mechanical ventilation (MV) [1]. It is therefore the most common nosocomial infection in mechanically ventilated patients and is associated with high morbidity and mortality despite appropriate initial antibiotics [2]. Since MV could promote lung inflammation (that is, biotrauma) leading to ventilatorinduced lung injury (VILI), especially if applied on infected lungs, drugs with anti-inflammatory properties are of potential interest in the setting of VAP [3,4]. Moreover, it has been shown that the outcome of patients with VAP was closely related, at least in part, to the degree of both pulmonary and systemic inflammation [5,6].
Statins are lipid-lowering agents that reduce the risk of cardiovascular events by inhibiting 3-hydroxy-3 methylglutaryl coenzyme A reductase [7,8]. Immunomodulatory properties have also been described [9]. Although controversial, a number of publications have raised the possibility that these drugs could exert protective anti-inflammatory effects in the context of sepsis, especially pneumonia [10][11][12][13][14][15]. Interestingly, some experimental studies have demonstrated a lung-protective effect through VILI attenuation [16,17]. However, it is worth noting that the clinical relevance of these findings has not been yet demonstrated [18][19][20]. The main drawback of the studies published so far, in which the protective effect of statins against sepsis was evaluated, is their retrospective design, which makes it difficult to control for the many clinically relevant and especially 'protective' confounders (that is, better health follow-up, greater degree of treatment compliance), since patients with prior statin therapy were always compared with those without. In addition, since sepsis was generally community-acquired in these studies, some significant covariates such as treatment with antibiotics prior to hospital admission may not have been reliably recorded. Moreover, both the type and severity of the infection varied considerably making the case mix somewhat different from one report to another.
In contrast, interesting results were obtained in certain randomized controlled trials (RCTs) that targeted one type of infection and/or one population of patients (for example, intensive care unit (ICU) patients) [20,21]. Thus, Makris et al. showed that prior treatment with statins could decrease mortality and ICU length of stay in selected patients with VAP. In contrast, a recent RCT that compared atorvastatin to placebo in patients with suspected VAP has failed to demonstrate any benefit [22].
As a result, statins clinical interest in patients with VAP is uncertain. Moreover, whether statins should be stopped or continued in ICU patients remains an unsolved issue. Since VAP occurs a certain time after ICU admission in therefore closely followed patients, relevant data related to episodes of pneumonia can be considered reliable if collected prospectively, as was the case in our medical ICU [23]. We therefore decided to assess the impact of prior statin therapy on the outcome (that is, ICU all-cause 30-day mortality) of patients with suspected VAP included in our cohort. To this purpose, two distinct sets of analysis were performed. First, we compared the patients with prior statin exposure to those without. Second, we considered only the patients with prior exposure and evaluated mortality according to continuation or discontinuation of the drug in the ICU.
Study population
The database used in this study has already been described elsewhere [23]. Briefly, every patient admitted to our ICU between January, 2006 and January, 2013 was considered if subjected to MV for more than 48 hours. Each patient with suspected VAP according to the physician's clinical judgment was eligible. However, only those with a 'modified' clinical pulmonary infection score (CPIS) of five or greater, provided empirical antibiotics were delivered promptly (that is, day 1), were kept for analysis in the present work, as in the recently published 'STATIN-VAP' study, in order to include the patients with a sufficiently high level of suspicion regarding VAP diagnosis [22,24,25].
In accordance with French law, no informed consent was required since all measurements were part of routine management, as confirmed by our local ethics committee (Comité de Protection des Personnes Nord-Est), which gave us its approval for conducting our study.
Definitions
The patients with suspected VAP were classified into two distinct groups according to prior statin therapy. Thus, the 'previous users' group included all of the patients treated with statins prior to ICU admission whether or not they were continued thereafter. The 'statin-naive' group included patients without prior statins. For secondary analyses (see below), statin previous users were separated into two groups according to drug continuation until VAP episode (that is, statin continuation or discontinuation in the ICU).
Since tracheal aspirate quantitative cultures were performed, the 10 6 colony-forming units (CFU)/mL cutoff value was applied for differentiating between positive and negative results.
Bacteria were considered multidrug resistant (MDR) in the following cases according to recent recommendations: (i) Pseudomonas aeruginosa resistant to imipenem and/or antipseudomonal penicillins and/or one aminoside and/or ciprofloxacine; (ii) Enterobacteriacae if resistant to third-generation cephalosporins and/or fluoroquinolone and/or an aminoside; (iii) Staphylococcus aureus if resistant to oxacillin [26]. Patients with negative tracheal aspirate cultures were considered free of MDR bacteria.
Immunosuppression was defined as neutropenia (polymorphonuclear cells counts less than 1500/mm 3 ), any immunosuppressive treatment prior to ICU admission including steroids if given for more than one month.
Data collection
Using a recording form, 'modified' CPIS value, demographic data and usually reported risk factors for MDR bacteria were prospectively recorded (that is, time between VAP suspicion and ICU admission, previous hospitalization, exposure to antibiotics defined as the administration of at least one two-day course of antibiotics within the past 30 days, residence in a nursing home, underlying chronic obstructive pulmonary disease). The clinical course of VAP was also assessed through day-1 and day-3 sequential organ failure assessment (SOFA) scores, the duration of mechanical ventilation and the number of ventilator-free days. The primary outcome was 30-day all-cause mortality.
Prior statin therapy was assessed through retrospectively collected data after the prospective cohort had been established. All the available records (both in and out of hospital) were used to determine whether patients had received statins before and after ICU admission. They were considered as previous users if they used to take statins prior to ICU admission regardless of how long they did. In addition, procalcitonin (PCT) measurement was usually performed in every patient with suspected sepsis as a reliable tool to improve diagnosis and antimicrobial management [27]. Tracheal aspirate samples were taken in every patient within a 24-hour period following the clinical suspicion. The results of the bacterial cultures were used to calculate the 'day-3 CPIS' since one point was added to the value obtained at day 1 if at least 10 6 CFU /mL were recovered. One point was then added if the direct examination showed the same germ.
VAP management
The guidelines for the antibiotic therapy management were based on the knowledge of local susceptibility patterns of the most frequently isolated bacteria, as well on the clinical judgment of the attending physician. The first-line treatment (that is, the one delivered within the first 24 hours following the clinical suspicion of VAP) was considered appropriate if the isolated pathogen(s) was (were) susceptible to at least one drug administered at the onset of sepsis according to the corresponding susceptibility testing report. When no antibiotic was given within the first 24 hours of management, the treatment was considered inappropriate regardless of the subsequently isolated pathogen.
Statistical analysis
Values are expressed as median (range) unless otherwise stated.
Comparative analysis
First, patients with VAP were compared according to prior statin use as defined above. Continuous variables were compared using the Mann-Whitney U test, and categorical variables were compared using the chi-squared test.
Survival analysis
Then, 30-day survival analyses were conducted.
First, the whole cohort was considered. The survival of patients regarding statin therapy was analyzed through the construction of the corresponding Kaplan-Meier curves compared by the log-rank test. Then, in addition to statin exposure (that is, 'previous users' vs. 'statinnaive'), every variable associated with 30-day death in the ICU according to the univariate analysis was entered into a multivariate Cox model if the P value was less than 0.20.
In the second set of analyses restricted to prior statin users, 30-day survival was evaluated by univariate analysis as described above. Again, the corresponding Kaplan-Meier curve was constructed and a multivariate analysis based on a Cox model was performed in an attempt to withdraw potential confounders. In addition to statin use after admission to the ICU (that is, 'statin continued' vs. 'statin discontinued'), covariates were selected according to the results of the univariate analysis. Only those associated with death with a P value less than 0.20 were included into the model. For all Cox model analyses, scaled Schoenfeld residuals (graphical inspection and formal testing for a nonzero slope in a regression of the residuals on functions of time) were used to check the Cox model, for each variable.
The functional form of continuous variables (V1, V2…) was checked with martingale residual analysis and by means of fractional polynomials. No a priori interactions were clinically indicated or tested.
The goodness-of-fit was assessed by the Cox-Snell residuals, bias-corrected Akaike's information criterion.
All tests were two-tailed. A P value of <0.05 was considered statistically significant. STATA software was used for all analyses (StataCorp, College Station, TX, USA).
Patients' characteristics
Between 1 January 2006 and 31 January 2013, 631 episodes of suspected VAP were recorded. Seventy-four episodes were excluded because of missing data (n = 22), inclusion in a RCT comparing atorvastatin to placebo for the management of VAP (n = 52), CPIS less than five or the absence of antibiotics delivered on day 1 (n = 208). Among the 349 remaining cases, 93 (26.6%) had been treated with statins, while 256 had not (73.4%) (Figure 1). Among the former patients (that is, 'previous users'), statins were continued until VAP in 52 (55.9%) for a duration of 9.2 (2.0 to 71.0) days.
The main baseline characteristics of the included patients are presented in Table 1. It is worth noting that statin previous users were sicker than statin-naive patients at ICU admission. Actually, the statin patients were significantly older, had higher SAPS II scores on admission, and were more likely to harbor underlying diseases. In contrast, main admission diagnosis was similar.
VAP episodes description
Among the 349 episodes of VAP analyzed, a positive bacterial culture was obtained in around 80% of the cases in each group. For the whole population, Enterobacteriaceae were the most frequently isolated pathogen (29.9%). This proportion reached 33.5% in the patients with statins.
Overall, MDR germs, as defined above, were isolated in 33.1% of patients with statins and 28.6% of those without (P = 0.33).
The first-line antibiotic therapy was appropriate in most of the cases (76.9% and 76.2% in the 'statin previous users' and in the 'statin-naive' groups, respectively; P = 0.89).
Regarding the severity of VAP episodes, the occurrence of septic shock on day 1 was similar in the two groups (38.0% versus 32.0%, P = 0.29), as were the CPIS values and the SOFA scores on day 1 and day 3 (Table 2). Similarly, the same proportions of patients in both groups were given steroids (45.1% vs. 45.0%, respectively; P = 0.99).
No significant difference was found regarding PCT measurements.
Survival analysis
The overall mortality rate for the whole cohort reached 28.6%. Prior statin therapy did not modify the outcome (Tables 2 and 3) (Figure 2).
To eliminate potential confounders, a time-dependent multivariate analysis based on a Cox model was conducted. Statin therapy was found not to have any protective effect (Table 4). Conversely, as expected, immunosuppression, age, cirrhosis, high SAPS II, SOFA day-1 values and the use of steroids during the VAP episode were independently associated with decreased survival in the ICU.
We conducted additional survival analyses within the subset of patients in whom statins previous users (n = 93). The patients in whom statins were continued once admitted to the ICU (n = 52) were distinguished from the remaining ones (n = 41).
First, we compared those patients with respect to their baseline characteristics and main admission diagnosis as well in an attempt to identify potential factors likely to have influence on the decision to stop statin therapy in the ICU (Table 5). No difference was found between groups, except hospitalization prior to ICU admission that was found to be more frequent in the patients in whom statins were discontinued.
The mortality rate tended to be lower in patients who continued statin therapy than in the others (28.8% vs. 43.9%; P = 0.13) ( Table 6). Moreover, after adjusting for age and SOFA day-1 values, statin continuation in prior users was found to be significantly associated with an improved outcome (adjusted hazard ratio (HR) = 0.47; (0.22 to 0.97) 95% confidence interval (CI); P = 0.04) ( Table 7) (Figure 3). Similar findings were obtained when the analysis was restricted to the only patients with positive quantitative cultures (n = 73), that is those with confirmed VAP (Table 8).
Discussion
In addition to their lipid-lowering properties, statins have pleiotropic effects, which are likely to improve the outcome of patients with coronary heart disease by reducing the risk of new cardiovascular events. Among these effects, statins are able to modulate the host immune response [9]. This probably results from a wide range of interferences with proinflammatory pathways. Using statin therapy to protect patients with sepsis could therefore be a reasonable strategy, as these drugs could dampen an overwhelming host inflammatory response to infection, especially within the lung [28]. Many studies, mainly observational, have addressed this issue [15]. However, contrasting findings have been reported. This may result from both study design and the widely varying characteristics of the populations included. All in all, the protective effect of statins during sepsis remains an unsolved issue. Ventilator-associated pneumonia remains one of the major complications of protracted stays in the ICU. Despite early and appropriate antibiotics, the response rates to antibacterial therapy remain low. This could result from deep alterations of the host immune response in this setting, as a shift toward a proinflammatory state occurs within the lung if subjected to MV. This proinflammatory state is likely to weaken the host's ability to clear pathogens [29]. As suggested in previous studies [20], statin therapy may therefore have a beneficial effect in patients with VAP.
We report herein the findings from a single-center cohort study that investigated, retrospectively, the effects of statin therapy on the outcome of patients with suspected VAP. In the first set of analyses, we compared statin-naive patients with those who received statins, whether or not treatment was discontinued upon admission to the ICU. Baseline characteristics according to statin continuation. VAP, ventilator-associated pneumonia; SAPS II, simplified acute physiology score II; ICU, intensive care unit; COPD, chronic obstructive pulmonary disease.
We found that mortality was not different between the groups, and this despite the fact that previous users were older and sicker. A multivariate survival analysis was then conducted but failed to demonstrate any difference in survival despite multiple adjustments. Moreover, various relevant variables may have been missing from our analysis, especially because of the retrospective design of our study. Thus, additional atherosclerosis risk factors such as smoking or obesity, which are frequently associated with statin intake, could act as masking confounders. Unfortunately, we were unable to adjust for such factors, thus making it more difficult to demonstrate any beneficial effect. We therefore conducted a second set of analyses restricted to patients with prior statin therapy, in an Table 6 Baseline characteristics and VAP episode description in the statin previous users according to the 30-day mortality attempt to overcome this issue. This approach was also supported by the fact that it remains unknown whether statins should be continued or not when patients on statin therapy present with a critical illness requiring treatment in an ICU, notably because of toxicity concerns. Similar to our first findings, we found no statistically significant difference between patients who continued statin therapy and those who did not after admission to the ICU. However, after adjusting for potential confounders, continued statin therapy was found to be independently related to survival in the ICU. Several explanations could be provided. First, this subset of patients (that is, previous statin users) was probably more homogeneous regarding cardiovascular risk factors and associated illnesses, thereby making the existence of uncontrolled confounding variables less likely. Second, the protective effect of statins during sepsis may decrease with the duration of discontinuation. Obviously, the most consistent published data were obtained in patients hospitalized with severe communityacquired pneumonia, that is to say patients actually exposed to statins at least until the onset of sepsis [13,14]. If statins were discontinued on admission, it is likely that the anti-inflammatory effects of statins were diminished if not abolished when the VAP occurred, that is to say several days, if not weeks, later. This is especially true in our cohort since most of the VAP episodes recorded were lateonset. Conversely, a previously published RCT showed that in previous statin users hospitalized with signs of infection, drug continuation was not associated with a better outcome [30]. It is worth noting that illness severity was mild in this study. Indeed, the same authors have reported more recently the results of another RCT comparing atorvastatin 20 mg daily to placebo given upon ICU admission to critically ill patients with sepsis [21]. These authors showed that in the predefined subgroup of patients who received statins prior to ICU admission there was a significant decrease in 28-day mortality in the treatment arm when compared to placebo, whereas no effect was measured if the new users (that is, statin-naive patients) were also considered. It is worth noting that in our cohort, we did not find any statin new users. We cannot therefore conclude about the putative protective effect of statin de novo treatment in the ICU, an issue addressed by several ongoing or recently completed RCTs. Among them, the 'STATIN VAP' study showed that no beneficial effect could be expected from simvastatin when given in patients with suspected VAP [22]. Although speculative, one can hypothesize that statin anti-inflammatory effects, if any, could be obtained and protect the host as well only if given prior to infection. Several limitations should, however, be mentioned. First, although most of the data were collected prospectively, information regarding statin use was acquired retrospectively and could therefore be a matter of concern. However, we systematically excluded from the analysis patients with unreliable or missing data. Second, because of the retrospective design of the study, the risk of missing confounding data is real. Our findings should thus be interpreted with caution. Third, the assessment of renal function was based on serum creatinine alone, which is subject to criticism, especially in the ICU setting. Fourth, liver test abnormalities on admission were not recorded in our database. We cannot exclude that statin discontinuation was thus influenced in some patients, accounting thereby for differences of outcome. In addition, we were unable to demonstrate the direct detrimental effects of statin in our patients since, for instance, creatine kinase levels were not available. Some of our conclusions are therefore only speculative. Moreover, we did not record which type of statin was used in the included patients, despite the fact that anti-inflammatory properties may differ according to the considered subclass (that is, lipophilic vs. hydrophilic).
Similarly, patients were classified as previous users if any statin therapy has been given prior to ICU admission regardless of treatment duration, making unlikely any conclusion about the minimal required time of exposure. In addition, inflammation assessment relied on the sole PCT measurement. Inflammatory cytokines assessment would have maybe provided different insights regarding host response. However, PCT monitoring has been shown to be clinically relevant in patients with VAP [31]. Moreover, we cannot exclude that some patients without confirmed VAP were included in the study since we used diagnosis criteria known to lack of specificity [24]. As a result, any possible protective effect of statins in our cohort may be independent from VAP occurrence. However, since we considered the only episodes in which the physician in charge promptly delivered antibiotics, we believe that it reflects real-life practice. In addition, the CPIS value on day 3 was rather high (6.5 (2) points within the whole cohort), strongly suggesting thereby VAP diagnosis in our patients. Moreover, it is worth noting that such a beneficial effect on survival of statin exposure was also noticed in the subset of patients of microbiologically proven VAP. Finally, the sample size in our subgroup analysis was quite small, raising the risk of a type-2 error because of a lack of statistical power.
Conclusions
In our cohort, continuing statins after admission to an ICU may have a protective effect in patients with suspected VAP.
Key messages
Statins protective effects in patients with pneumonia as well as drug continuation after ICU admission are unsolved issues. In our cohort of patients with suspected ventilatorassociated pneumonia, statin continuation in previous users was shown to be an independent predictor of survival. | 2016-05-16T03:57:40.036Z | 2014-04-28T00:00:00.000 | {
"year": 2014,
"sha1": "e63e4faf1d5f6130e35f9488d344deb30ce1dffd",
"oa_license": "CCBY",
"oa_url": "https://ccforum.biomedcentral.com/track/pdf/10.1186/cc13845",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "4985c9af97579b03d37c94985b12be7a96329525",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
267542093 | pes2o/s2orc | v3-fos-license | Pre-Service Teachers’ Educational Video Mediated Experience vis-a-vis Learning Engagement in the New Normal
. Technology-mediated instruction using online instructional materials, particularly educational videos, reinforces teaching-learning. However, students’ engagement becomes an issue in this learning environment. This paper ascertained the relationship between educational video utilization and learning engagement in technology-mediated learning. This study also investigated the perceived impacts of educational video utilization and the extent of respondents’ learning engagement in technology-mediated learning. This study employed a quantitative non-experimental method through a descriptive correlational research method. The data were obtained from the survey questionnaire administered to 284 students in one University in Bicol Region, Philippines. It was analyzed and interpreted using different statistical treatments such as frequency count, percentage technique, weighted mean, PPMC, and t-test. Findings revealed that educational video utilization impacted the three aspects (cognitive, emotional, and behavioral) as perceived by the respondents. There was a large extent of learning engagement of the respondents in three areas: knowledge development, social interaction, and motivation. Furthermore, the result showed a significant correlation between the two variables; hence, it is recommended that language teachers utilize educational videos during the teaching-learning process to increase learning engagement in technology-mediated learning.
Introduction
Nowadays, the urge for an accelerated digital transformation of academic institutions has led to the employment of technology-mediated learning [1].Technology-mediated learning resulted in using technology in students' learning processes in different disciplines, as it is the only avenue to teach, communicate, and collaborate [2].It is one of the actions taken in response to the harmful effects of the COVID-19 pandemic on different aspects of society, such as dramatic loss of lives and economic and social disruption [3].It also caused the world's greatest disruption in education systems, affecting roughly 1.6 billion students in more than 200 nations.More than 94 percent of the world's student population has been affected by school, institution, and other learning facility closures [4].
The rise of technology-mediated learning has been visible, and the utilization of online instructional materials has dominated.Additionally, educational videos have become a common online instructional material increasingly used in virtual education worldwide.Teachers predominantly use video in their lessons, with various aims, and the most common is to raise students' conceptual knowledge or interest [5].Teachers and students benefit greatly from educational videos, which boost course performance and significantly affect student motivation, confidence, and attitudes.It also provides benefits for engagement in some specific aspects, such as broadening involvement, emotional engagement, and overall course engagement [6].
However, despite the numerous benefits of videos, some barriers to student engagement have been identified.Since instructors are unavailable to learn through videos, there is a greater need for self-discipline on the part of students.Also, self-regulation, a significant predictor of emotional, behavioral, and cognitive engagement, is difficult to monitor remotely.These are key challenges since the recent study of Giray (2022) revealed that many Filipino students are still struggling with learner control amidst a shift to online learning due to the pandemic [7].Our previous study showed the learning and academic engagement of the students in technology-mediated learning became a challenge due to access and internet connectivity [5].
Furthermore, the transition to online instruction was more challenging for ESL students than for teachers [8].Given the significance of English as a Second Language (ESL) programs in achieving proficiency in English as it is the language of commerce and law and the primary medium of instruction in education [9], the role of the teachers is to create and innovate ways for students during the teaching-learning process, such as using videos, podcasts, and other online resources.
Based on the given context, educational videos are integral to teaching-learning.However, issues are presented regarding the student's engagement in technology-mediated learning.Therefore, the researchers were determined to investigate further the perceived impact of educational video utilization covering the emotional, behavioral, and cognitive dimensions and the extent of respondents' learning engagement in knowledge development, social interaction, and motivation.Moreover, this paper ascertained if there is a significant relationship between the perceived impact of educational video utilization and the extent of respondents' learning engagement in technology-mediated learning.
Educational Videos
Multimedia learning tools have become critical parts of classrooms.They are common virtual tools that aid in understanding concepts.They may be films or clips to educate the audience on a topic or set of topics.These media are widely available and accessible on different platforms on the Internet 10].
These reasons and the improvements in digital technology and modern internet connections have aided in making videos a standard component of most learning environments.In fact, videos are used in classrooms more than ever.It was found that 70% of educators use videos in their classes many times each week [1]].There is also a growing strategic teaching approach in modern classrooms called video-assisted learning (VAL) [12].It pertains to strategically using audio-visual aids, such as videos, to help students acquire knowledge, competence, and skills.It is also defined as a strategic teaching approach to improve students' comprehension, cognitive ability, and socio-emotional skills [13].Other than that, the emergence of the pandemic made educational videos more crucial in the instructional process.The situation forced educators to integrate video lessons into blended learning.They are now utilized in synchronous and asynchronous sessions to teach students [14].
Educational videos have numerous benefits in the teaching-learning process, especially for students.They are known for providing the audience with a rich sensory experience.As a result, learning becomes a more engaging experience for students [15].Particularly, concise, conversational, and interactive videos are remarkably effective in facilitating learning 16].They also improve cognitive and affective learning outcomes, increase knowledge retention, stimulate understanding and aptitude, accommodate different learning styles, and foster creative and critical thinking skills [17].
Moreover, educational videos also increase students' interest, given their proficiency with technology and appetite for online video consumption.It was supported by research that approximately 83% of individuals would rather watch videos than read instructive or informational content [18].Further, it was revealed that students enjoy classes more when videos and other media are displayed, especially in real-time [19].Similarly, the results obtained from the recent survey revealed that 82% of the teachers polled thought that using videos in teaching boosts students' achievement [20].
Animated teaching videos improve students' learning experiences by increasing interest and engagement, improving students' self-assessed understanding of the materials by simplifying technical concepts and providing visual aids, enabling flexible and self-paced learning, and providing a refreshing change from traditional teaching [21].Videos supported the students' understanding of the concepts applied in the sessions.They also stated that students liked the ICT-favored multimedia classroom where videos could be shown for teaching and learning [22].
Other than that, the use of educational videos also benefits educators.They make the delivery of instruction easier and more time-efficient [23].Technology and videos also allow educators to experiment with digital tools.Consequently, according to a comparable study, teachers' desire to develop their skills and create informative videos appears to be growing [24].It was also suggested to use storytelling techniques, engaging voiceover, practical examples, good visuals, particularity in repeating information, adding humor, and asking questions and comments for feedback [25].
Furthermore, it was recommended that making educational videos more valuable to students through animated videos, whiteboard videos, live videos, videos mixing voice and pictures, and Khan Academy-style videos be considered while developing videos because these sorts of videos were the most popular and short videos should be prepared because long movies might generate boredom and distraction [26].Videos are also recommended to be compatible with mobile, laptops, and other up-to-date devices.This compilation of published literature and conducted studies aided in grasping the concept of educational videos.It provided an overview of how educational videos were used and their implications.Other than that, it presented the facets of the topic that had already been studied.
Learning Engagement
The learning process outcomes are greatly influenced by how active a learner is in its duration.This activeness, also called learning engagement, can aid in achieving targeted course outcomes as it promotes meaningful learning experiences and higher-level critical thinking [27].Learning engagement is defined as the degree of attention, curiosity, interest, optimism, and passion that students show when they are learning or being taught, which extends to the level of motivation they have to learn and progress in their education.Engaged learners take responsibility for their decisions, use feedback, evaluate personal conduct, and analyze appropriate responses to engage with learning opportunities and act autonomously for improvement [28].
There are three dimensions of learning engagement.These are cognitive engagementwhich draws on the idea of investment and willingness to exert the effort necessary to comprehend complex ideas and master difficult skills; behavioral engagement-which draws on the idea of participation; and emotional engagement-which encompasses positive and negative reactions to teachers, classmates, academics, and school and is presumed to create ties to an institution and influence willingness to do the work [29].
The type of emotion experienced by learners is important in successful learning [30].Autonomous motivation generates better learning than controlled motivation, whereby feelings of pride and guilt drive the desire to meet internalized social expectations.It was supported by [31], who suggested that emotionally stimulating events are remembered more clearly, accurately, and for longer periods than emotionally neutral events.The evidence highlights that engaging students on an emotional level will likely lead to better memory recall in the long run.Furthermore, the learners' motivation and engagement were intertwined as they associated with the learners' positive academic behaviors, which are salient in achieving academic success [32].
Behavioral engagement displayed behaviors associated with a student's effort toward learning and the learning process for a single activity or their overall learning experience [33].Academic achievement and school retention positively correlate with behavioral engagement [34].Behavioral, emotional, and cognitive are the three dimensions of engagement.However, the simplest term to define is behavioral engagement because the simplest indicators of behavioral engagement are attention, raising hands, finishing tasks, turning on cameras, and unmuting microphones.
Significant educational outcomes, such as academic success and satisfaction, are linked with learner engagement.Behavioral, emotional, and cognitive engagements are positively correlated to the student's academic performance [35].Moreover, a study at the University of Mindanao suggested that student engagement positively influences course satisfaction [36].A study conducted in selected universities in Manila, Philippines, also highlighted how social engagement in school positively affects students' life satisfaction [37].
Teachers may also utilize varied instructional strategies appropriate to the subject matter and the background knowledge of the students and may introduce more autonomy-supportive learning environments as an alternative to improve their instructional practices and students' engagement [38].In addition, for students to maximize their academic involvement, there should be a significant collaboration between the instructor, the school, and the parents.
Students' academic engagement is intertwined with their learning success, mainly because students with a high level of emotional, behavioral, and cognitive engagement typically put more effort into learning the resources and materials [39].This collated data provided an overview of learning engagement.The findings and conclusions were stated to guide the definition of learning engagement in the study.They also served as part of the guiding framework of the present study.
Research Design
This study used quantitative non-experimental methods through a descriptive correlational method of research.The descriptive method was employed to determine the perceived impact of educational video utilization on respondents regarding cognitive, behavioral, and emotional aspects and the extent of respondents' learning engagement in knowledge development, social interaction, and motivation.The correlation method was utilized to determine the significant relationship between the perceived impact of educational video utilization and the extent of respondents' learning engagement in technology-mediated learning.
Research Locale and Respondents
This study was conducted in one state university in Bicol Region.The researchers purposely chose this school to determine the relationship between the perceived impact of educational video utilization and the extent of respondents' engagement in technology-mediated learning because the university adopted the digital way of disseminating knowledge to students.The university made a virtual learning portal (VLP) that serves as the university's learning method.Consequently, they require their teaching personnel to utilize online instructional materials like educational videos to deliver their lessons.The university adopts flexible learning methods and limited face-to-face classes for all programs.The students from the College of Education from 1 st year to 4 th year were the respondents of this study.The total population is 974, and Yamane's Formula was utilized to determine the sample size using a 5% margin of error, composed of 284 students.The distribution of respondents was selected through a random sampling technique utilizing a lottery method.The Bachelor of Secondary Education comprised 201 respondents from the 1st and 4th year of each specialization: BSEd English (50), BSEd Science (64), BSEd Math (47), and BSEd Filipino (40).The Bachelor of Elementary Education comprised 42 respondents, and the Bachelor of Technology and Livelihood Education comprised 41 respondents.
Research Instruments
The student perception survey questionnaire assessed respondents' engagement using educational videos in technology-mediated learning.It was divided into Part I, Perceived Impact of Educational Video Utilization on the Learning Engagement of ESL Students, and Part II, Extent of Learning Engagement of ESL Students.It comprised 60 items, to which respondents were asked to respond for 30 minutes.The first part is categorized into the three dimensions of student engagement: cognitive, behavioral, and emotional.It determined the perceived impact of educational video utilization on the learning engagement of respondents in terms of the aforementioned dimensions.Each category consists of 10 question items, for a total of 30.The results were determined through the Likert scale in five levels: (5) strongly agree, (4) agree, (3) neutral, (2) disagree, and (1) strongly disagree.
Meanwhile, survey questions about the extent of respondents' learning engagement comprised 30 items divided into different aspects such as knowledge development, social interaction, and motivation.The results of the survey were ascertained through the Likert scale in five levels: (5) to a very large extent, (4) to a large extent, (3) to a moderate extent, (2) to a small extent, and (1) to a very small extent.The survey questionnaire crafted by the researchers and the modified standardized questionnaire were used to ascertain the perceived impact of educational video utilization in terms of cognitive, behavioral, and emotional aspects and the extent of respondents' learning engagement regarding knowledge development, social interaction, and motivation.It was validated by three (3) language professors and one (1) statistician/mathematics teacher to ensure its accuracy and reliability.
Data Analysis
The data were analyzed and interpreted using Frequency Count, Percentage Technique, Mean Used, PPMC, and T-test.These statistical treatments were utilized to explain the distribution of respondents and the relationship between the variables.The frequency Count and Percentage Technique was used to identify the distribution of respondents, the number of occurrences in the researchers' survey questionnaire, and its percentage.Weighted Mean and Ranking were used to measure and interpret the data gathered.It was also used to determine the perceived impact of educational video utilization on respondents regarding cognitive, behavioral, and emotional aspects and the extent of their engagement in knowledge development, social interaction, and motivation in technology-mediated learning.The Pearson Product-Moment Correlation (PPMC) was used to determine the extent of the relationship between the perceived impact of educational video utilization and the extent of respondents' learning engagement in technology-mediated learning.A T-test was used to determine if the null hypothesis was rejected or accepted.
Results and Discussion
This section covers the result of interpreted data from the 284 respondents from the College of Education.The gathered data is analyzed and interpreted using the Weighted Mean and Ranking Technique.Presented in Table 1 is the summary of the perceived impact of educational video utilization on respondents based on the result of the survey test conducted.The data disclosed that cognitive engagement has (WM: 4.48-strongly agree), behavioral engagement has (WM: 4.28-strongly agree), and emotional engagement has (WM: 4.20strongly agree).These data obtained an average weighted mean of 4.32, described as strongly agree.As gleaned in the table, based on the respondents' perception, educational video utilization had identified impacts among the three aspects: cognitive, behavioral, and emotional.
The data implied that educational video utilization aids the students in comprehending the lessons and learning new concepts, fuels the imaginations and interests of the students, considers watching educational videos as an intellectually and emotionally stimulating activity, and brings satisfaction to the learning process.
Findings revealed that the educational videos helped the students comprehend the lessons thoroughly, discover things about the subjects, and help in their learning process.Moreover, it facilitates reflection, analysis, and critical thinking, increases the student's productivity, and even ignites their eagerness to study more.It was supported by the study of Ocampo (2018), which stated that videos supported students' understanding of the concepts used in the sessions.Also, integrating educational videos in lessons facilitates the learning process.It is supported by a current study that concise, conversational, and interactive videos are remarkably effective in facilitating learning [23].Furthermore, it showed that educational video utilization fosters reflection, analysis, and critical thinking.
Parallel to the above finding, educational videos help increase social interaction, have unlimited accessibility, provide a learning environment, pause and rewind, and allow a reasonable training period [40].It enables self-paced learning and helps the students to access the lessons conveniently.Also, it allows students to simplify technical concepts, provides visual aids, enables flexible and self-paced learning, and provides a refreshing change from traditional teaching.As to the perceived impact of educational video utilization on respondents in terms of emotional engagement, the data revealed that it aids the students to look after creativity, fuels their imaginations, contributes ideas to the lessons, is intellectually stimulating for learning, and considers watching educational videos as an emotionally stimulating activity that helps the students remember the information more clearly and accurately.
Meanwhile, Table 2 presents the result of the summary of the extent of respondents' learning engagement based on the result of the survey test conducted.The data disclosed that knowledge development had (WM: 4.08-large extent), social interaction (WM: 4.02-large extent), and motivation (WM: 4.04-large extent).These data obtained an average weighted mean of 4.05, which was interpreted as a large extent.As gleaned from the table, the extent of learning engagement among three areas, namely, knowledge development, social interaction, and motivation, was described as large.Hence, it implied that the respondents use sophisticated learning strategies in planning, monitoring, evaluating, and accomplishing their school tasks and activities.They are engaged in attending their class willingly, actively participate in class discussions, and aid their colleagues in explaining course materials.They even exert efforts in performing their responsibilities, particularly in group work.
These findings parallel the study of Fletcher (2016), who postulated that engaged learners take responsibility for their decisions, use feedback, evaluate personal conduct, and analyze appropriate responses to engage with learning opportunities and act autonomously for improvement [41].The data revealed that the respondents enjoy learning new things in class, behave accordingly, and listen attentively when other students and teachers are talking in class sessions.These data were supported by the study of Torto (2020), who stated that motivation was associated with emotional engagement that encompasses positive and negative reactions to teachers, classmates, academics, and school and is presumed to create ties to an institution and influence willingness to do work [29].In addition, the results of this study implied that the respondents exemplified positive behaviors in class as they listened carefully when their classmates and teachers were talking in class.It was apparent in Martin et al.'s (2017) findings that learners' motivation and engagement were intertwined as they associated with the learners' positive academic behaviors, which are salient in achieving academic success [42].
Table 3 presents the relationship of the perceived impact of educational videos in terms of cognitive aspect among the three variables of learning engagement: knowledge development, social interaction, and motivation.Findings revealed that the cognitive aspect is significantly correlated among the three variables of learning engagement (1.knowledge development, r=0.43;TV=8.11>CT=1.975,2. Social Interaction, r=0.39;TV=7.16>CV=1.975,3. Motivation, r=0.38;TV=6.89>CV=1.75)at a5% level of significance which resulted in the rejection of the null hypothesis.Moreover, there was a moderately positive correlation in the cognitive aspect among the three variables of respondents' learning engagement: knowledge development, social interaction, and motivation.
Based on the results, it can be inferred that there is a significant relationship between the perceived impact of educational video utilization in terms of cognitive aspect and the extent of respondents' learning engagement in technology-mediated learning.The data implied that educational video utilization impacted students' cognitive aspect, which aids in fostering learning engagement during the teaching and learning process.The educational video leads to a high student engagement rate since the cognitive process is a prerequisite for comprehending the complex information presented.The findings were parallel with the study of Priyakanth et al. (2021), who stated that making interactive videos assures learners' acquisition and advancement of new skills [43].These interactions assist the students in driving deeper engagement and changing how they think and process the information they have learned.
Presented in Table 4 is the relationship of the perceived impact of educational videos in terms of behavioral aspects among the three variables of learning engagement: knowledge development, social interaction, and motivation.The findings revealed that the behavioral aspect is significantly correlated among the three variables of learning engagement (1.knowledge development, r=0.58;TV=11.92>CV=1.975,2. Social Interaction, r=0.51;TV=9.84>CV=1.975,3. Motivation, r=0.43;TV=8.06>CV=1.975)at a 5% level of significance which resulted in the rejection of the null hypothesis.
Moreover, there was a high positive correlation in the behavioral aspect among the two variables of learning engagement, knowledge development, and social interaction.In contrast, there was a moderately positive correlation in motivation.Based on the results, it can be inferred that there is a significant relationship between the perceived impact of educational video utilization in terms of behavioral aspects and the extent of respondents' learning engagement in technology-mediated learning.The data implied that educational video utilization impacted the students' behavioral aspect, which contributed to an increase in their positive academic behaviors and the feelings they invested during the teaching and learning process.
Educational videos increased student engagement, as shown by lower dropout rates, more regular school attendance, improved performance at national assessments, and increased motivation [44].Furthermore, the findings coalesced with another report that not only were the educational videos effective in helping students, but they also evoked their engagement as they found these teacher-made videos enjoyable [45].
Presented in Table 5 is the relationship of the perceived impact of educational videos in terms of emotional aspects among the three variables of learning engagement: knowledge development, social interaction, and motivation.Findings revealed that the emotional aspect is significantly correlated among the three variables of learning engagement ( It can be inferred that there was a highly positive correlation between the emotional aspect and knowledge development.In contrast, there was a moderately positive correlation between the two variables of learning engagement: social interaction and motivation.Based on the results, there is a relationship between the perceived impact of educational video utilization in terms of emotional aspect and the extent of respondents' learning engagement in technology-mediated learning. The data implied that educational video utilization impacted the students' emotional aspect, which fuels their drive to engage in the teaching and learning process positively.The findings were in parallel with the study of Cole et al. (2021), which revealed that the students' motivations to interact with the course information can be influenced by including video in the course design to increase learning and learning engagement [46].Moreover, the educational videos were found to positively impact students' cognitive load, satisfaction, and academic achievement.
Conclusion
It can be concluded that educational video utilization impacts the cognitive, behavioral, and emotional aspects.The respondents indicated that educational videos aid them in comprehending the lesson, increase their engagement in classroom settings and building off others' ideas, fuel their imagination, and contribute ideas to the lessons.Furthermore, respondents were pleased with their ability to learn new things in class and demonstrated this by paying close attention when other students and teachers spoke in class.It should be noted that they believe learning is boring to some extent.It was also discovered that there is a significant relationship between the cognitive, behavioral, and emotional aspects of educational video use and learning engagement in technology-mediated learning.With these findings, it is suggested that teachers use educational videos in the teaching-learning process more frequently.In doing so, it may respond to the needs of every student as it may feed the curiosity and imagination of the students, and it is ideal for creating intellectual and emotionally stimulating activities.They may employ educational videos in the discussion and create activities that captivate and ignite the students' interest and willingness to participate actively during the teaching-learning process.The schools may maintain and improve the pedagogy in teaching by sending the teaching personnel to seminars and workshops that disseminate information about the effective utilization of educational videos in the teaching-learning process.
Table 1 .
Perceived Impact of Educational Video Utilization on Respondents
Table 2 .
Extent of Learning Engagement of ESL Students
Table 3 .
Relationship between Cognitive Aspect and Learning Engagement
Table 4 .
Relationship between Behavioral Aspect and Learning Engagement
Table 5 .
Relationship Between Emotional Aspect and the Three Variables of Learning Engagement | 2024-02-08T16:14:15.658Z | 2024-01-01T00:00:00.000 | {
"year": 2024,
"sha1": "887f3711d07e393d569d32b1dc22d403cab0b266",
"oa_license": "CCBY",
"oa_url": "https://www.shs-conferences.org/articles/shsconf/pdf/2024/02/shsconf_access2024_02005.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "1e3965de0ca9524544fb3fcb79cba993e1a853bf",
"s2fieldsofstudy": [
"Education",
"Computer Science"
],
"extfieldsofstudy": []
} |
268341700 | pes2o/s2orc | v3-fos-license | Simulation study on capacity planning and allocation of island microgrid
. In this paper, the energy storage capacity planning problem of a real island microgrid is deeply simulated. In the beginning, the overview and basic data of the island microgrid are described in detail, including the composition of the grid and its key parameters. Then, with the help of modeling and simulation technology, the operation of microgrid in various scenarios is comprehensively simulated and analyzed, focusing on how to optimize energy storage planning in the case of wind energy removal to ensure the smooth operation of the grid. The research deeply reveals the core role of stable power system in ensuring the reliability of power supply and the normal operation of the system. Among them, frequency stability is regarded as the cornerstone of power system operation, which is indispensable for maintaining the overall stability of the power grid and ensuring the continuous and stable power supply of users.Through this simulation study, we not only provide feasible schemes and suggestions for the energy storage capacity planning and allocation of island microgrids, but also verify that reasonable energy storage schemes have a positive effect on the overall stability and reliability of the power system.
Introduction
With the transformation of global energy structure and the continuous development of distributed energy technology, microgrid, as a new energy supply mode, has been widely used in remote areas, islands and other off-grid environments.[1][2][3] The application of microgrid can combine traditional power sources, new wind energy and energy storage equipment and load for coordinated control and operation.Effectively alleviate the impact of various distributed power sources directly connected to the grid [4].As a special type of island-type microgrid, due to its unique geographical location and energy demand, it puts forward higher requirements for energy storage capacity planning, energy scheduling and system stability.Therefore, it is of great significance to carry out the modeling and simulation research of islay-type microgrid for improving the reliability, economy and sustainability of power supply in the island area.Island-type microgrids usually contain a variety of distributed energy sources, such as renewable energy sources such as wind and solar, as well as traditional energy sources such as energy storage systems and diesel generators [5][6][7].The complementarity and synergy of these energy sources are essential for the stable operation of the microgrid [8][9].However, due to the complexity of the island environment, it may cause fluctuations in wind speed and instability in load demand, which brings great challenges to the operation and control of island-type microgrid [10].The microgrid model in this paper is derived from a real operating island microgrid, and is modeled and analyzed based on PSCAD platform to observe the influence of wind energy removal on the frequency variation of the grid.By adding a certain capacity of energy storage, it is verified that a certain amount of energy storage in the microgrid has a positive effect on the frequency stability of the grid [11][12][13].Through multiple sets of simulation experiments, the optimal energy storage capacity in line with the grid operation is explored.
Construction of micro-grid simulation model
Considering the electromagnetic transient state of f winding of synchronous generator set and the rotor motion process, the mathematical model of synchronous generator is established.The stator voltage equation is as follows: Stator current equation of synchronous generator: The rotor excitation winding voltage equation of synchronous generator: Rotor motion equation of synchronous generator: Where ra is the stator resistance; Xq is Q-axis reactance and q ' E Q-axis transient electromotive force; ' d 0 T Is the time constant of the excitation winding; Ef is the stator excitation electromotive force; J is the rotor's moment of inertia; D is the damping coefficient of the generator; Pm is the mechanical power; Pe is electromagnetic power; Vs is the system voltage; Xl is transformer and line impedance; Delta is the electrical Angle between the q and x axes.As known by formula (3)(4), the speed of the synchronous generator is determined by the rotor motion equation, and the power balance relationship determines the speed of the synchronous generator, and the speed of the synchronous generator determines the frequency of the microgrid.In order to study its speed characteristics, a mathematical model of the excitation system of the gas turbine is established, as shown in Figure 3 1 Ts 1 Ts
Fig. 1. Gas turbine excitation system model
Where, T1 and T2 are the time constant of the circuit of the integrated amplifier unit; Kp is the amplification factor of the proportional amplification circuit; Ta is the time constant of the adaptive unit.Gas turbine generator set has the characteristics of high adjustment capacity, so the offshore oil and gas field platform is basically used gas turbine generator set.The gas turbine speed control system is the main part of the gas turbine control system, which realizes the purpose of controlling the gas turbine speed by controlling the fuel output signal.The simplified model of gas turbine and speed control system is as follows: Where, K* is the per unit value of the unit regulating power of the offshore oil and gas field power system; PI regulator is for secondary frequency modulation, so that the frequency back to 50Hz; K is the proportion of fuel flow required by the gas turbine to maintain self-support under no-load conditions; a, b and c are parameters that reflect the characteristics of the fuel system; WF is the fuel flow output ratio.The included gas turbine-generator set model is shown as follows: 1 Ts
Fig. 3. gas turbine generator set model
The output of the fan is the output power of the wind turbine.Its ideal mathematical model is: Where, v is wind speed; ρ is air density; Rw is the radius of the wind turbine blade; Represents the wind energy utilization factor.For variable pitch systems, it is related to tip speed ratio and pitch Angle, for fixed pitch systems, it is only related to tip speed ratio.Set the active power of the constant output of the wind turbine to 12.5MW.
Simulation model of island microgrid
As shown in Figure 5.The simplified model of the island microgrid is shown below.In the figure below, there are five power sources to power the load.Generators Gen-1 to Gen-4 generate electricity for traditional gas turbines to power the load, while the wind turbine connected to bus 3 can output 12.5MW of power to the island microgrid.The active power generated by gas turbines and wind turbines provides the required electric energy for various loads in the grid.
Simulation analysis
This chapter will run the simulation models of each component of the microgrid in Section 3, establish the mathematical model of the sea island microgrid as shown in Figure 3-1, set the total load capacity to 27MW, and conduct the following sets of experiments.1.When the fan is not connected to the power grid, the running frequency of the power grid under normal operation is obtained by simulation 2.When the 12.5MW fan is connected to the power grid, the fan is cut out immediately during normal operation, and the frequency fluctuation of the power grid is observed.3.When a certain amount of energy storage is connected to the power grid, the fan is cut out instantly, whether the frequency of the power grid is within the normal allowable range, and multiple sets of simulation experiments are set to verify the minimum energy storage capacity to ensure the safe and stable operation of the power grid.
When the grid without wind turbine new energy to join, the grid model as shown in Figure 3 at this time the grid operation results as shown in the following waveform.The above simulation waveforms for each generator set in the wind turbine generator set is not connected to the voltage and power waveforms, as well as at this time the grid frequency waveforms, it can be seen in the initial operation of the grid, when the start of operation, this time the generator is in the initialization stage, this time the fluctuations are larger, when the generator is running stably, this time the grid frequency tends to be stable, and at this time, each generator set power is also relatively stable, the system to provide the power is equal to the load of the system.On the basis of Scene 1, wind energy is connected to the grid, the power of the turbine is set to 12.5MW, and the wind energy is completely removed at the 8s.At this time, the grid is powered by gas turbines and wind turbines simultaneously on the load side of the grid, and when the turbines in the grid are suddenly cut off at the 8th s, the frequency and power changes in the grid are shown in the figure below.From the simulation waveform, when the grid is in normal operation, the load side of the power provided by the gas turbine and wind turbine generator at the same time, and the grid power and frequency is more stable, when the 8th second instantaneous excision of the wind turbine generator, the power required by the load is completely synchronous generator to provide the generator, the generating units out of power instantly the grid frequency decreases, due to the lack of storage capacity, the simulation results show that at this point in time the grid The simulation results show that the grid frequency at this time will have a large drop and fall below 59.2Hz, the lowest frequency of normal grid operation, that is, at this time, the grid frequency exceeds the normal range of fluctuations allowed, the power grid anomaly.In addition to Scene 2, the turbine is cut off at the 8th second when energy storage of different capacities is added.When the instantaneous removal of 12.5MW wind energy supply, such as if there is a certain amount of energy storage as a backup power, simulation to test whether it will reduce the range of fluctuations in the frequency of the power grid, and a number of simulation experiments to test the optimal energy storage capacity in line with this power grid.At this point in time, the energy storage facility provides a certain amount of power supply to the grid, and the grid frequency waveform is shown in the figure below.After comparing the frequency fluctuation of the grid and the change of synchronous generator set output under the conditions of adding four different energy storage capacity configurations of 2MW, 5MW, the following conclusions can be obtained: (1) The inclusion of a certain amount of energy storage devices in the grid can improve the immunity of frequencies in the grid.(2) Once the wind turbine is removed from the grid, the grid is supplied with power from four synchronous generators and energy storage devices at the same time, and the fluctuation of the sudden increase in the active output of the synchronous generator sets in the system is reduced with the increase in the energy storage capacity.(3) The lower limit of the permitted grid frequency in this grid is 59.2 Hz, and after the comparison of the four energy storage capacity configurations, the energy storage capacity in the grid is not less than 5 MW, considering the reliability and economic factors of the grid operation.
Conclusions
In this paper, PSCAD is used to construct the overall model of the power grid and carry out simulation experiments to simulate the adverse impact of wind power removal on the grid frequency and select the optimal energy storage capacity.
From the perspective of stability, the overall cut off of wind power has a greater impact on the stability of the power grid, especially the frequency stability.In the case of 12.5MW wind power access, the overall cut off of all wind power will make the frequency in the power grid drop below 59.2Hz, seriously destroying the security and stability of the power grid.So that the actual frequency fluctuation of the power grid is smaller than the fluctuation range of the power grid operation.The designed simulation experiment verifies that the addition of 5.5MW energy storage in the grid is appropriate.
Governing system and gas turbine block diagram ICECAE 2024 , 01010 (2024) E3S Web of Conferences https://doi.org/10.1051/e3sconf/202449701010497 Governor system and gas turbine model Fig. 2. governor system and gas turbine model Is the sweep area of the blade;Pw-M is the mechanical power output of the wind turbine;
Fig. 10 .
Fig. 10.Frequency waveform of the grid when no new wind energy is connected | 2024-03-12T15:49:42.632Z | 2024-01-01T00:00:00.000 | {
"year": 2024,
"sha1": "502593277ceba0168416bb5b816826fb1b4e2eea",
"oa_license": "CCBY",
"oa_url": "https://www.e3s-conferences.org/articles/e3sconf/pdf/2024/27/e3sconf_icecae2024_01010.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "6921f4c7a01e14f82dc775c84810ecef3b69b304",
"s2fieldsofstudy": [
"Engineering",
"Environmental Science"
],
"extfieldsofstudy": []
} |
251294888 | pes2o/s2orc | v3-fos-license | Enchancement of resource-saving technology of corn growing on the basis of energy resources assessment
. The article deals with the issues of rational use of the farming system under the conditions of anthropogenic impact. In this regard, a new scientific direction is considered to improve the modern technology of growing crops that contribute to the rational use of the indicator of energy resources by 2-3 times . This approach necessitates a revision of the structure of crop placement depending on the solar radiation coefficient (K R ). In the long term, due to the irrational use of energy resources, the degradation of soil potential has intensified, which has dramatically affected the gross yield and the sustainability of agricultural production. Therefore, there is a need to move to a new level of assessment of the basic principles and methods of the farming system. Suggested comparative analysis of two technologies for the cultivation of corn for grain in the Zhambyl region is proposed: the existing and resource-saving. The article was carried out as part of the implementation of applied scientific research in the field of the agro-industrial complex under the scientific and technical program (IRNBR10764920) "Technologies and technical means of irrigation when introducing new irrigated lands, reconstructing and modernizing existing irrigation systems" for 2021-2023.
Introduction
The history of the development of civilization is inextricably linked with the development and improvement of technologies that require large expenditures of fuel, energy and raw materials. The more developed a society, the more it produces a complex energy-intensive and science-intensive product. In modern conditions, the development of agricultural production is hindered by the processes of degradation of the soil cover, the remaining low level of material and technical support for farms, as well as changes in natural resources. Under these conditions, scientific substantiation of changes in the structure of sown areas and diversification of agricultural crops is of great importance.
The degradation of natural resources occurred as a result of the transition period to market relations, ill-conceived and wasteful management of water and land potential for the development and placement of the agro-industrial sector. Therefore, there is an urgent need to develop and implement a set of measures aimed at the balanced use of a combination of natural and industrial resources.
Plants use nutrients and soil moisture as a "building material" from which biomass is formed as a result of the process of photosynthesis. Thus, the energy of solar radiation is embodied in plants in the form of the exchange energy of nutrients.Any human impact on the soil and plants should be aimed at intensifying these processes.
Conditions and methods of research
Many scientists [1][2][3][4][5][6][7][8][9][10][11][12][13][14] testify that climatic zones are formed under the influence of thermal energy resources and atmospheric precipitation, and some evaporation functions were used as energy resources in determining the moisture supply of geographical zones. Undoubtedly, the functions of air humidity and temperature deficit, evaporation to some extent reflect the energy resources of the climate, however, the radiation balance of the day surface (R) used [4,5] to determine the indicator of the effect of radiation on evaporation (P), and the radiation "dryness index » (Ř) [10] fully characterize the essence of the open V.V. Dokuchaev of the law of natural zonality.
The principles of conjugated and systemic approaches, as well as methods accepted in agrochemistry, soil science and hydrochemistry, were used in the research. The research results were processed using statistical methods. Accounting for the total costs for the production of agricultural products was carried out according to the method of G.A. Bulatkin [14].
Results of the study and their discussion
The energy resources of landscapes, as a process of heat exchange at a specific geographical point in space over a known period of time, is characterized by a balance of energy input and output [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. Therefore, the radiation balance of the day surface (R) used by A.A. Grigoriev to determine the indicator of the effect of radiation on evaporation (P1=P/LE), and the radiation "dryness index" (Ř=R/LOc) fully characterizes the essence of the open V.V. Dokuchaev of the law of natural zonality. However, the hydrothermal indicator (Ř) does not take into account the indicator of the absolute height of the terrain (H). Therefore, we have made attempts to fill this gap and clarify the actual energy indicators of the productivity of the natural system of Kazakhstan, taking into account the geography of the area. The correlation dependencies established by us indicate that the relationship between the terrain mark (H) and the dryness index (Ř) for certain regions of Kazakhstan has specific features. So, for the Turkestan region, it is described by an equation of the form: Ř=4.0-Н/250; for the North Kazakhstan region Ř=1.6-Н/794; for the West Kazakhstan region Ř=2.2-Н/113.6.
Further, taking into account that the dryness index (Ř=R/LOc) and equating it with the calculated one, it is possible to establish a relationship between the actual indicator of the radiation balance (R) depending on the absolute elevation of the terrain: An analysis of the solar energy input per unit area allowed us to identify the indicator of the radiation index of the absolute height of the area as the ratio of the radiation balance to the absolute mark of the area and describe it as Rн =R/H [11,12].
According to our calculations, in the Republic of Kazakhstan, the indicators (Rн)fluctuate within 2.0<Rн<0.3;from which it follows that the farming system can be divided into three zones: -neutral (Rн) >2.0 andmore; -optimal (Rн) = 0.3 -2.0; -dicey (Rн) <0.3 The data show that for many years the issues of zoning agricultural crops, despite sufficient heat supply, have not been successful, since cultivated crops have significant differences in relation to Rn.This circumstance allowed us to establish the following conceptual methodology for substantiating the classification of the farming system, based on the indicators of the radiation index of the absolute height of the area (Table 1). In addition, we have identified a relationship between the index of moisture coefficient (km) and the radiation index of the absolute elevation of the terrain (Rн), which is correlated in the following way: km = 0.281· RH +0.098; from here substituting togetherkm =Ос/ЕоиRн = R/Нafter transformation we get that: Hence, as can be seen from the data in Table 2, the radiation index of the absolute elevation of the terrain (Rн) in the objects of study under consideration is characterized in the range from 0.41 to 1.0. Hence, Pavlodar region can be considered the most favorable, where the indicator was 1.0. In other areas, it ranged from 0.41 to 0.65, i.e. in these areas, issues of the efficient use of energy resources should be worked out. On the other hand, the indicator of the radiation dryness index (Ř) in the irrigated areas of Kazakhstan represents a slight increase from the center of the country to the south and ranges from 2.05 to 4.26. At the same time, in relation to Pavlodar region, other studied regions were slightly higher, so Akmola region by 0.16; Zhambyl region by 1.64; Turkestan region by 2.21, which makes it clear about the sufficiency of thermal resources in these regions.
In relation to the solar radiation coefficient (KR), the indicators ranged from 0.79 to 3.31, however, in the Akmola region, the decrease in the solar radiation coefficient is 0.21, while in other studied areas it ranged from 0.65 to 2.31. At the same time, the highest indicator is noted in the Turkestan region (2.31). This allows us to conclude that, in relation to other regions, such as Zhambyl and Pavlodar in the Turkestan region, in order to obtain a sufficiently high yield of cultivated crops, large expenditures of irrigation water and technical resources are required.
It follows that the indicator of the solar radiation coefficient gives full grounds for the fact that in order to stabilize the ecological, reclamation and energy situation, it is necessary to develop improved technological maps for the cultivation of crops based on accounting for energy costs and fuel and energy resources (FER). Thus, for the irrigated zone of Kazakhstan, the optimal value of the radiation balance can be considered 136 kJ/cm2, which corresponds to the average indicator of the Pavlodar region, and in the Zhambyl and Turkestan regions this indicator exceeds by 40.5-52.1 kJ/cm2. This gives us reason that in these areas there are opportunities to get two crops a year by improving the technology for the revival of forest reclamation work aimed at optimizing the energy resources of the foothill zone of Kazakhstan.
Therefore, taking into account the current state of irrigated lands and the environmental and energy circumstances occurring in the arid zone, we propose the introduction of a landscape-restoration system of agriculture, which is based on the following approaches: -organization of water-saving technologies for the cultivation of crops based on the assessment of energy resources; -restoration of degraded lands through the use of bioameliorants from local raw materials; -high-quality improvement of drainage and waste water through the use of modern resource-saving technologies; -development of theoretical foundations for water consumption of agricultural crops based on energy resources; -development of the scientific foundations of the farming system, aimed at the qualitative improvement of the water-physical parameters of the soil; -implementation horticulture and viticulture in all farms on an area of at least 15-20% of the total area; -implementation resource-saving cultivation technologies aimed at maximizing the use of bioameliorants, biostimulators of plant growth and development made on the basis of local raw materials (phosphogypsum, glauconite clay, bentonite, etc.); -implementation less costly technologies for irrigating crops aimed at reducing soil salinity, improving and increasing soil fertility (sharbat method, discrete method, irrigation through a furrow, etc.).
In this article, we provide a comparative analysis of two technologies for the cultivation of corn for grain in the Zhambyl region -the existing and resource-saving.The studied technologies of cultivation of corn for grain have fundamental differences.The existing technology aims to obtain the highest possible yields by using an unlimited amount of water resources, fertilizers and plant protection products.As a result, it led to the deterioration of the reclamation and ecological condition of irrigated lands.
Resource-saving technology, in comparison with the existing one, implies a reduction in the use of water resources, fertilizers, plant protection products and their use, taking into account the needs of agricultural crops.The use of bio stimulators and bio ameliorants, prepared on the basis of local raw materials, provides an increase in soil fertility and desalinization of the soil in the rhizosphere zone, which contributes to the normal growth and development of the cultivated crop.
In our studies, the total energy costs per 1 ha of corn crops for grain are the sum of the total energy costs for machinery and equipment, working capital and labor resources (Table 3). The total energy costs per 1 ha of corn crops for grain when it is cultivated according to the existing technology amounted to 48095.9 MJ/ha, according to the resource-saving technology -31173.7 MJ/ha. Thus, the resource-saving technology is also energy-saving, allowing saving 16922.2 MJ/ha or 35.2% of energy compared to the existing one.
As follows from the data in Table 1, energy savings occurred mainly due to a reduction in energy costs for working capital, including fuel, fertilizers and pesticides.
In technologies, there are high fuel costs (44.5 and 59.4%), which requires the need to select vehicles, save fuel, etc.
In addition, it is of great interest to determine the energy content in the economically valuable part of the crop per 1 ha [13,14].
The energy content in the economically valuable part of the crop is calculated by the formula: The content of energy stored in corn for grain, taking into account by-products, is determined by the formula: Ео = Ес·Кo; where: Ес-energy content in the economically valuable part of the crop without byproducts, MJ/ha; Ео -energy content in the main and by-products, MJ/ha; У -economically valuable part of the crop (without by-products), kg; Кc -coefficient of conversion of a unit of production into dry matter; Е -total energy content in 1 kg of dry matter, MJ/ha; Кo -overall product yield ratio, defined as the ratio of main and by-products to the main one.
Using the above formulas, we determine the energy content in the main product with the existing technology at a yield of 3.65 t/ha. Ес. = У·Кc·Е = 3650·0.25·16.39 = 149558.75 MJ/ha. Given that the currently used system of machines during harvesting allows you to collect an insignificant part of economically valuable by-products, which is about 75% of the main products.Thus, the indicator of the overall output coefficient (Ko) will be 1 + 0.75 = 1.75.Then the total amount of energy in the main product and the collected side product will be: In the studied technologies, the accumulated energy of corn biomass significantly exceeds the costs of the total energy spent on corn cultivation (Eo>Qn).
Considering that the reduction of energy costs for the production of a unit of production is currently the main task, we tried to analyze the accumulated energy in the aboveground phytomass of agricultural crops of grain-grass crop rotation in the Zhambyl region (tables 4,5,6). Note: Eof -energy in the main product; Epf -energy in by-products; Efp is the energy of the total aboveground phytomass; EU -energy content of mineral fertilizers; Eya -is the energy contained in pesticides; Ec -energy contained in the seeds; An is the sum of the energy of materialized costs; Eananthropogenic energy spent on the cultivation of agricultural crops; Rn is the energy efficiency of agricultural technology.
From the data in Table 6 it follows that in the conditions of the Zhambyl region, the most energy efficient is the cultivation of alfalfa (38.8) and sugar beet (13.4). The rest of the studied crops are also considered energy efficient, however, in order to achieve higher rates, it is necessary to improve the technologies for growing agricultural crops in the direction of reducing the use of fuel and energy resources, the use of mineral fertilizers and minimizing the work of agricultural machinery.
Thus, the energy efficiency of grain corn cultivation in the conditions of the Zhambyl region is about (3.1), and the resource-saving technologies used, due to a more advanced irrigation system and the introduction of a biomeliorant, made it possible to save energy resources by 35.2%.
Therefore, further improvement of resource-saving technology should be aimed at reducing the total energy for working capital.
Conclusion
The proposed technology, based on the fundamental laws of nature, makes it possible to enchancement the farming system and intensify agriculture, aimed at accelerating the process of increasing the production of grain, industrial, fodder and vegetable crops. This provision necessitates a revision of the structure of crop placement depending on the solar radiation coefficient (KR). In the long term, due to the irrational use of energy resources, the degradation of soil potential has intensified, which has dramatically affected the gross yield and the sustainability of agricultural production. Therefore, there is a need to move to a new level of assessment of the basic principles and methods of the farming system. | 2022-08-04T15:05:09.147Z | 2022-01-01T00:00:00.000 | {
"year": 2022,
"sha1": "7ac26e8dd409ca0ed7cb1d96d43403b4bb2724e2",
"oa_license": "CCBY",
"oa_url": "https://www.bio-conferences.org/articles/bioconf/pdf/2022/10/bioconf_idsisa2022_05004.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "fbd81300346fc13dbdb62e8dfdaac7cdd04598ab",
"s2fieldsofstudy": [
"Agricultural And Food Sciences"
],
"extfieldsofstudy": []
} |
236420260 | pes2o/s2orc | v3-fos-license | PHARMACEUTICAL CARE PRACTICE IN THE COMMUNITY PHARMACY BY UTILIZING PROBLEM-BASED LEARNING IN REFLECTIVE PEDAGOGY PARADIGM METHOD
This study follows the previous study entitle Problem-based Learning (PBL) in reflective pedagogy paradigm (RPP): Innovative learning in pharmaceutical care, which identified the learning material as complicated. This study aimed to provide an appropriate method for performing pharmacy management and pharmaceutical care in Community Pharmacy. Data collection was done using an assessment instrument to identify student’s achievement. The previous study encompasses preceptors and students in developing learning material, which cause it more valid and reliable to be implemented. The study was conducted in Yogyakarta, Surakarta, and Semarang city. The effectiveness of the learning material was shown by the grade of student’s achievement in learning outcome and the clear state with confidence in the expression of reflection and action-plan. Most of the students in the 3 cities achieved an excellent grade both in the problem-solving field, and presentation of the assignment. The students reflected that the learning material is simple and suitable in practicing pharmaceutical care and pharmacy management, moreover, they can state their plan to work as a Community Pharmacist with confidence. Therefore, PBL in the RPP method is ready to be used in practicing pharmaceutical care in the Community Pharmacy.
INTRODUCTION
In Indonesia, pharmaceutical care should be provided by pharmacists based on the Indonesian pharmacist standard of competence according to the regulation of Indonesian Health Minister No.73/2016(Minister of Health, 2016. In achieving the competencies, Pharmacists should have good knowledge, attitude, and practice in caring for the patients, and they also should have good communication skills to educate the customers (Aurelia, 2013;Handayani et al., 2009;Hartayu et al., 2013). However, based on the result of some studies regarding customer satisfaction towards pharmaceutical care performance, most of the customers stated that they were not satisfied, and pharmaceutical care performance should be improved (Hartayu et al., 2013). Therefore, the School of Pharmacy, especially the Pharmacist profession program should improve the learning process in achieving the Pharmacists' competence in performing a good pharmaceutical care practice. According to those facts, an innovation of the learning process in pharmaceutical care has been done by infusing problem-based learning (PBL) in the reflective pedagogy paradigm (RPP).
Problem-based learning (PBL) is a common method used in the learning process to conduct pharmaceutical care and pharmacy management in the community pharmacy (Hartayu et al., 2018). PBL embraces the principles of good learning and teaching. It is 40 Titien Siwi Hartayu et al. student-directed, which encourages selfsufficiency and is a preparation for life-long learning, and also promotes active and deep learning (Tais et al., 2014). PBL often includes or requires peer teaching, which encourages students to digest information, so that they can present, explain, and utilize it confidently. Reusing knowledge reinforces the processes of remembering and digesting (Hmelo-Silver, 2004). In the end, PBL encourages, and indeed requires, reflection on the learning process: how did it go this time, and next time will the team do it more efficiently and more effectively? On the other hand, RPP was chosen due to the characteristic of the learning process that is used, i.e. RPP has been proven as an effective strategy to improve the quality of personality, including personality character building in humanity, such as compassion and to be a responsible person. RPP is appropriate to use in higher education such as the Faculty of Pharmacy (Defeo, 2009;Hayes, 2006), to improve the students' ability to reflect (Avoy et al., 2012;Crable et al., 2010;Hise, 2012), and RPP is also can be used to improve the understanding of Pharmacotherapy and communication skill (Defeo, 2009;Hayes, 2006;Wijoyo et al., 2016). In the point of view about pharmaceutical care practice, the innovation of the learning process is aimed specially to achieve learning outcomes such as 1. the ability to propose business plan; 2. the ability to evaluate medicines management; 3. the ability of dispensing and educating patient in rational used of medicines; and 4. the ability of proposing Home Pharmacy care (Hartayu et al., 2018;Setiawan et al., 2010).
RPP is known as a paradigm and the way to improve the quality of personality and character building with activity cycle: Context-experience-reflection-actionevaluation (JSEA, 1993;Metts, 1995). Therefore, PBL in RPP was designed by developing an Initiating-trigger case based on the real case as a problem to solve (PBL) and also as a context in the activity cycle, which will experience students in performing pharmaceutical care. Based on the experience in problem-solving through the real case, the students were then should write some reflection, and then make some plan to act, and finally, they should evaluate what have they done (Cycle of RPP). The preliminary study regarding PBL in RPP that has been done in 3 cities, i.e. Yogyakarta, Surakarta, and Semarang city in Indonesia, showed the achievement of the students' ability in understanding both the theory and performing in pharmaceutical care (Hartayu et al., 2018), and indicate that the strengthening of the learning process of PBL in RPP is the comprehension of reflection, which result in the ability to integrate competence, conscience, and compassion. Moreover, the key to succeed in pharmaceutical care performance is the appropriate initiating-trigger case. Unfortunately, the preliminary study also found that the initiating-trigger case was indicated unreal and complicated (Hartayu et al., 2018), therefore, revising and re-testing the Initiating-trigger case are needed. Based on the above facts regarding the inappropriately learning material that has been used to conduct PBL in RPP, the study is aimed to provide an appropriate method for performing pharmacy management and pharmaceutical care in Community Pharmacy.
METHODS
The steps of the study are developing an initiating-trigger case, study semester plan, and assessment instrument. The study starts by inviting preceptors who will be involved in the study as a field supervisor, to conduct a focus group discussion (FGD). FGD among preceptors and academic supervisors results in approval of prescription criteria which will be used as the Initiating-trigger case. Criteria of the prescription are the prescription that is usually served in the community pharmacy which has a common problem regarding medicines used. Based on the approval of the prescription's criteria, a scenario as the Initiating-trigger case was then be developed. Before it was used, the scenario was distributed to the students for the language tests.
Regarding the study semester plan, this study used the same study semester plan as the previous study, based on the reflective pedagogy paradigm cycle, i.e. contextexperience-reflection-action-evaluation.
Context is fulfilled by the scenario based on the common prescription from the community pharmacy as the Initiating-trigger case. Experience is fulfilled by using problembased learning to construct students' knowledge and experiences in problemsolving, and then the reflection is fulfilled by Wijoyo's reflection list (Wijoyo et al., 2016).
Every student has to propose a community pharmacy business and home pharmacy care plan as an action in Pharmacist's managerial and clinical role. The evaluation was done by using an assessment instrument. The scopes of the learning process are developed based on the Competence Standard of Indonesian Pharmacist according to the MOH No.73/2016 and are described in Table 1 as follow: Assessment instrument was developed based on the activity in the PBL in RPP method which contains 7 steps such as 1. identify the problems; 2. hypothesis development; 3. identify the needs of data and provided data; 4. looking for data needed; 5. active self-learning; 6. sharing and pooling; and, 7. evaluating. Reflection was done after the 7 steps have been done. Preceptors and students filled up the reflection list. During 2 months internship, the students conducted an assignment regarding the PBL in RPP activities in the first month, and then in the second month, the students evaluated what have they done in the first month. The evaluation was done by presented the business plan, inventory control, dispensing, patient education, and home pharmacy care plan. The assessment instrument which was used, has been tested using a peer simulation method, and the assessment is done based on the ability to explain the result of a problem which was identified and the solution problem, including the appropriateness and systematically analysis; relevance of references; effective communication and teamwork; and, should be on time in submitting those assignments. The achievement grade is divided into 4 categories, i.e. excellent (>75); good (70-75); average (65-69); and poor (<65). The highest contribution to the grade is the content of the paperwork (60%). The steps of the studies summarize is described in Figure 1 entitle fishbone of the study as follow:
ETHICAL CLEARANCE
Ethical clearance of the study is stated by certificate no. 454/C.16/ FK/2017 of the Ethics Committee of the Faculty of Medicines Duta Wacana Christian University, Yogyakarta, Indonesia.
RESULTS AND DISCUSSION
The results of the study are divided into 4 assessments, i.e. 1. Implementation of the PBL in RPP method; 2. Assessment of Students 'reflection; 3. Reflection on the research implementation; and, 4. Evaluation of PBL in RPP method.
Implementation of the PBL in RPP Method
The difference of stage 2 (2018) compare with stage 1 (2017) regarding the implementation of the PBL in the RPP method is the learning material in stage 2 has been directed linked to the real problem they faced in the field, through the direct involvement of the preceptors in proposing the learning material. Consequently, the learning process in stage 2 is better due to the balance of the difficulty level that should be solved and the time of internship availability. The learning outcomes evaluation regarding the achievement of grade point towards students' paperwork are presented on Table 2, as follow: Supervisor/preceptor rate is 0-100 scale grade point. This grading point was based on a rating rubric toward student paperwork. Students' grade point 72/70 means the first supervisor gave 72 points and the second gave 70 toward student paperwork.
Assessment of Students' Reflection on the PBL in RPP First Cycle Implementation
In the reflection stage, the students are given some questions as a guideline to write self-reflection after they practiced in the pharmacy management and pharmaceutical care aspect. The questions are: 1. Put forward the experience that you have experienced (write down things related to the goal learning process, and relevance to your profession) 2. Knowledge: What knowledge you have gained after implement the learning material during the professional practice program, is it useful to support your profession in the future? 3. Life values: What values of life have you gained after practice the PBL in the RPP method?
4. Write down the action plan as a followup to the knowledge and values of life, which you have gained after practicing PBL in the RPP method. Answering the four questions, most students in Yogyakarta, Surakarta, and Semarang City are stated that the method experienced them in problem-solving, and relevant to the Pharmacist's responsibility. They experienced how to propose the business plan, inventory control, compounding and dispensing, improving communication skills, home pharmacy care, improving selfconfidence and networking, performing pharmaceutical care with confidence, and more responsible for their task. They have a willingness to be a responsible and proactive person. Based on the evaluation of the paperwork and students' reflection, students showed excellent performance in identifying problems, problem-solving strategy, and inventory control in the management cycle. The students' reflection was quantified based on the grading rubric in Wijoyo's (2016) experiment. The result of students' reflection assessment is described in Table 3 as follow: Titien Siwi Hartayu et al.
Reflection on the Research Implementation
After conducting the first cycle of PBL in RPP activities, academic supervisors and preceptors meet up to reflect the outcomes of the learning process. The reflection result can be described as follows: Preceptors in Surakarta City stated that students have been able to solve the problem by utilizing the Initiating-trigger case in the Home Pharmacy Care and providing a creative and realistic solution in Pharmaceutical Care and also in Pharmacy management. However, there are still groups of students who provide less realistic solutions, especially in a feasibility study to run the business of Community Pharmacy. On the other hand, academic supervisors observed that based on the achievement of scores, students have demonstrated their ability to extract data and information; determine and solve the problems which they have found.
In Yogyakarta City preceptors expressed that students have been able to solve the problem by exploring data and information thoroughly during conducting field observation, reviewing problems from the sight, and provide solutions fairly well. It is necessary to improve the sharpness of the analysis, especially in solving the problem of the pharmacy business feasibility study, that has not good yet and still in the average category; from the point view of performing pharmaceutical care and Home pharmacy care, students have been adequate and successful. Academic supervisors observed that in general, students who were practice in Yogyakarta City showed their effort and ability to solve the problems and be able to relate the problem with practicing in pharmaceutical care by using the PBL in RPP method. However, there were still groups of students who had not to provide information in detail about the problem. Therefore, the report of the solution has not accurate yet (regarding the Initiating-trigger case of the business plan feasibility study). Appreciation was given to the students who demonstrated mastery of the problem and were able to perform home pharmacy care and pharmaceutical care properly.
Regarding students who were practice in Semarang City, the preceptors opinionated that students demonstrated a strong willingness to learn, good pharmacological knowledge, and adaptability to the field conditions. However, there were still weaknesses in the mastery of pharmacy management material and the ability to conduct pharmaceutical care. Management capabilities were still weak in aspects of the business plan feasibility study and pharmacy management. On the other hand, academic supervisors noticed that by gaining new experience during the internship in the community pharmacy, students were able to determine some problems that often arise in performing pharmaceutical care. They were not only hearing, but also directly observed and at the same time tried to practice the theories acquired.
Evaluation PBL in RPP Method
Finally, The PBL in RPP method was evaluated by the preceptors based on the students' learning outcomes, and the evaluation result is described in Table 4 as follow: Table 4. Evaluation of PBL in RPP based on the student's learning outcomes
No Preceptors in Evaluation Results
1 Yogyakarta City a. The Initiating-trigger case is based on the daily activities which were carried out by the Pharmacist at the Community Pharmacy. b. The existence of an Initiating-trigger case makes it easy for all preceptors to choose relevant material to be taught and practiced at the Community Pharmacy by the students during the Pharmacists profession practice program. c. It can be utilized as Information about the composition of portfolio assessment in assessing student activity.
2 Surakarta City a. The existence of an Initiating-trigger case facilitates the depth of the mentoring process to the material learning about pharmacy management. b. All of the Initiating-trigger cases are ready to be implemented during the Pharmacists profession practice program.
3 Semarang City a. The existence of an Initiating-trigger case becomes a guideline in the process of the Pharmacists profession practice program. b. Highly appreciate the result of learning material ravishment. c. The Initiating-trigger case is more realistic and easier to be implemented.
In terms of supporting the realization of plenary health services in the field of Pharmacy, the government issued pharmaceutical care standards (MOH No. 1027/2004 which has been updated to the regulation of the MOH no. 35/2014) (MOH Regulation, 2004;MOH Regulation, 2014) where the role of pharmacists is very clearly regulated. Unfortunately, until nowadays, even though the quantity of pharmacists has increased significantly every year, the expected role has not been recognized yet. The role of pharmacists in health care services is needed to support the achievement of therapeutic goals. In supporting the achievement of therapeutic goals, a Pharmacist is expected to have good knowledge and skills in explaining the purpose of treatment, ways of storage, possible side effects, and how to cope with the adverse drug reaction (Aurelia, 2013;Hartayu et al., 2013). The result of the study showed that almost all of the students in Semarang city achieved an excellent categorize in proposing the business plan, performing pharmaceutical care, conducting inventory control, and proposing the home pharmacy care plan. On the other hand, students in Yogyakarta city only achieved a good categorize in proposing the business plan however, they also achieved an excellent grade in the 3 others, such as performing pharmaceutical care, inventory control, and home pharmacy care. Comparing with students in those two cities, students in Surakarta city achieved an excellent grade in performing pharmaceutical care and inventory control. Unfortunately, the community pharmacy in Surakarta city did not conduct the business plan and home pharmacy care due to the human resources constraint.
The student's reflection stated that the learning material is simple and realistic, because they used the learning based on the Titien Siwi Hartayu et al. real case, therefore, it would be very useful in the future job. The reflection was written in a very clear statement and structure. The action plan was stated in a very confident by 70% of the students that they are willing to work as a Community Pharmacist, and the rest of the amount wrote still in a normative manner, i.e. "I want to work based on my passion" without mentioning what is their passion.
One of the studies stated that PBL has not always been successful (Glew, 2003), and the other study mentioned that educational research has found it difficult to demonstrate positive effects of PBL on outcomes such as knowledge, critical thinking, reflective practice, and teamwork (Hounsell et al., 2002). However, based on the result of this study including the method evaluation by the preceptors, showed, that PBL in the RPP method is effective in the learning process to improve student's knowledge and skills in performing pharmacy management and pharmaceutical care, and also effective in improving character building such as humanity, having a sense of compassion and be a responsible person (Hartayu et al., 2018;Hise, 2012). Regarding the unsuccessful of conducting PBL, the reason is probably as a result of its not being implemented properly rather than there is a basic flaw in the method (Barrows, 1986;Hughes et al., 2003). In addition, the result showed that the strength of the method is the modifying of PBL that was infused in the Reflective pedagogy paradigm. The result of the study is also indicating that these findings for pharmaceutical education are consistent with recent research in nursing (Shin et al., 2013), medical (Tayyeb, 2013), dental education (Huang et al., 2013), and in the Universities learning process (Defeo, 2009;Hayes, 2006). The way of working of the RPP itself is to shape the student's personality by providing an experience of humanitarian values, then proceed with reflecting on the experiences. Therefore, the PBL in RPP method can increase student's reflective abilities (Avoy et al., 2012;Crable et al., 2010;Hise, 2012). The PBL in RPP method is also suitable as a model to make the students can integrate competence, conscience, and compassion in the field of their profession include improving an understanding of pharmacotherapy and communication skills of pharmacist profession students (Wijoyo et al., 2016).
In this study, the PBL in RPP method is implemented in the learning process of pharmacist profession practice program at the community pharmacy, with a focus on the 4 fields, such as the preparation of a business plan feasibility study to run the business of a community pharmacy; inventory control of medicines and medical devices; pharmaceutical care; and home pharmacy care plan. Due to the PBL in RPP method encourages students to think about and solve real problems, it is likely to aid their use of knowledge in clinical situations, helps develop their clinical reasoning, and encourage selfdirected learning throughout their professional careers (Handayani et al., 2009;Jones, 2006). In healthcare fields, these skills are highly desirable and valuable.
Based on these findings, the PBL in RPP method shows the ability to make prospective pharmacists as a learner to learn thoroughly in implementing the theory into practice, and reinforced by a comprehensive reflection process, and lead the Pharmacist Graduates to be able in integrating competence, conscience and compassion (Glew, 2003;Hise et al., 2010;Mann et al., 2009) these findings can be an answered of the previous study which demonstrated that by utilizing PBL alone, students perform better academically; even though there was no data that can infer this improvement results in better professional (Tais et al., 2014).
This makes PBL in the RPP method the preferred education method for healthcare courses. Unfortunately, the study did not measure the possibility of increasing cost regarding the implementation of the PBL in the RPP method, however, other study found that the main barriers for implementing PBL are a teacher and staff training and a necessary reduction in class size, which could increase the cost of pharmacy education (Finucane et al., 1998;Jones, 2006). On the other hand, other study found that PBL and | 2021-07-27T00:05:06.673Z | 2021-05-31T00:00:00.000 | {
"year": 2021,
"sha1": "d6a6e1e94c90a8c6935b4911ab0556b08d19db35",
"oa_license": "CCBY",
"oa_url": "https://e-journal.usd.ac.id/index.php/JFSK/article/download/2769/pdf",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "238c6d98e33fe56fc3ad9a9bfb8bb2252fce6cc9",
"s2fieldsofstudy": [
"Education",
"Medicine"
],
"extfieldsofstudy": [
"Psychology"
]
} |
196471332 | pes2o/s2orc | v3-fos-license | Modularization of Research Software for Collaborative Open Source Development
Software systems evolve over their lifetime. Changing conditions, such as requirements or customer requests make it inevitable for developers to perform adjustments to the underlying code base. Especially in the context of open source software where everybody can contribute, requirements can change over time and new user groups may be addressed. In particular, research software is often not structured with a maintainable and extensible architecture. In combination with obsolescent technologies, this is a challenging task for new developers, especially, when students are involved. In this paper, we report on the modularization process and architecture of our open source research project ExplorViz towards a microservice architecture. The new architecture facilitates a collaborative development process for both researchers and students. We describe the modularization measures and present how we solved occurring issues and enhanced our development process. Afterwards, we illustrate our modularization approach with our modernized, extensible software system architecture and highlight the improved collaborative development process. Finally, we present a proof-of-concept implementation featuring several developed extensions in terms of architecture and extensibility.
I. INTRODUCTION
Software systems are continuously evolving during their lifetime. Changing contexts, legal, or requirement changes such as customer requests make it inevitable for developers to perform modifications of existing software systems. Open source software is based on the open source model, which addresses a decentralized and collaborative software development. Open research software [1] is available to the public and enables anyone to copy, modify, and redistribute the underlying source code. In this context, where anyone can contribute code or feature requests, requirements can change over time and new user groups may appear. Although this development approach features a lot of collaboration and freedom, the resulting software does not necessarily constitute a maintainable and extensible underlying architecture. Additionally, employed technologies and frameworks can become obsolescent or are not updated anymore. In particular, research software is often not structured with a maintainable and extensible architecture [2]. This causes a challenging task for developers during the development, especially when inexperienced collaborators like students are involved. Based on several drivers, like technical issues or occurring organization problems, many research and industrial projects need to move their applications to other programming languages, frameworks, or even architectures. Currently, a tremendous movement in research and industry constitutes a migration or even modernization towards a microservice architecture, caused by promised benefits like scalability, agility, and reliability [3]. Unfortunately, the process of moving towards a microservice-based architecture is difficult, because there a several challenges to address from both technical and organizational perspectives [4]. In this paper, we report on the modularization process of our open source research project ExplorViz towards a more collaboration-oriented development process on the basis of a microservice architecture. We later call the outdated version ExplorViz Legacy, and the new version just ExplorViz.
The remainder of this paper is organized as follows. In Section II, we illustrate our problems and drivers for a modularization and architectural modernization. Afterwards, we illustrate our software system and underlying architecture of ExplorViz Legacy in Section III. The following modularization and modernization process as well as the target architecture of ExplorViz are described in Section IV. Section V introduces our proof of concept in detail, including an evaluation based on several developed extensions. Our ongoing work in terms of achieving an entire microservice architecture is presented in Section VI. Section VII discusses related work on modularization and modernization towards microservice architectures. Finally, the conclusions are drawn and an outlook is given.
II. PROBLEM STATEMENT
The open source research project ExplorViz started in 2012 as part of a PhD thesis and is further developed and maintained until today. ExplorViz enables a live monitoring and visualization of large software landscapes [5], [6]. The tool has the objective to aid the process of system and program comprehension for developers and operators. We successfully employed the software in several collaboration projects [7], [8] and experiments [9], [10]. The project is developed from the beginning on GitHub with a small set of core developers and many collaborators (more than 30 students) over the time. Several extensions have been implemented since the first version, which enhanced the tool's feature set. Unfortunately, this led to an unstructured architecture due to an unsuitable collaboration and integration process. In combination with technical debt and issues of our employed software framework and underlying architecture, we had to perform a technical and process-oriented modularization. Since 2012, several researchers, student assistants, and a total of 25 student theses as well as multiple projects contributed to ExplorViz. We initially chose the Java-based Google Web Toolkit (GWT) [11], which seemed to be a good fit in 2012, since Java is the most used language in our lectures. GWT provides different wrappers for Hypertext Markup Language (HTML) and compiles a set of Java classes to JavaScript (JS) to enable the execution of applications in web browsers. Employing GWT in our project resulted in a monolithic application (hereinafter referred to as ExplorViz Legacy), which introduced certain problems over the course of time.
1) Extensibility & Integrability: ExplorViz Legacy's concerns are divided in core logic (core), e.g., predefined software visualizations, and extensions. When ExplorViz Legacy was developed, students created new git branches to implement their given task, e.g., a new feature. However, there was no extension mechanism that allowed the integration of features without rupturing the core's code base. Therefore, most students created different, but necessary features in varying classes for the same functionality. Furthermore, completely new technologies were utilized, which introduced new, sometimes even unnecessary (due to the lack of knowledge), dependencies. Eventually, most of the developed features could not be easily integrated into the master branch and thus remained isolated in their feature branch.
2) Code Quality & Comprehensibility: After a short period of time, modern JS web frameworks became increasingly mature. Therefore, we started to use GWT's JavaScript Native Interface (JSNI) to embed JS functionality in client-related Java methods. Unfortunately, JSNI was overused and the result was a partitioning of the code base. Developers were now starting to write Java source code, only to access JS, HTML, and Cascading Style Sheets (CSS). Furthermore, the integration of modern JS libraries in order to improve the user experience in the frontend was problematic. Additionally, Google announced that JSNI would be removed with the upcoming release of Version 3, which required the migration of a majority of clientrelated code. Google also released a new web development programming language, named DART, which seemed to be the unofficial successor of GWT. Thus, we identified a potential risk, if we would perform a version update. Eventually, JSNI reduced our code quality. Our remaining Java classes further suffered from ignoring some of the most common Java conventions and resulting bugs. Students of our university know and use supporting software for code quality, e.g., static analysis tools such as Checkstyle [12] or PMD [13]. However, we did not define a common code style supported by these tools in ExplorViz Legacy. Therefore, a vast amount of extensions required a lot of refactoring, especially when we planned to integrate a feature into the core.
3) Software Configuration & Delivery: In ExplorViz Legacy, integrated features were deeply coupled with the core and could not be easily taken out. Often, users did not need all features, but only a certain subset of the overall functionality. Therefore, we introduced new branches with different configurations for several use cases, e.g., a live demo. Afterwards, users could download resulting artifacts, but the maintenance of related branches was cumbersome. Summarized, the stated problems worsened the extensibility, maintainability, and comprehension for developers of our software. Therefore, we were in need of modularizing and modernizing ExplorViz.
III. ExplorViz Legacy
The overall architecture and the employed software stack of ExplorViz Legacy is shown in Figure 1. We are instrumenting applications, regardless whether they are native applications or deployed artifacts in an application server like Apache Tomcat. The instrumentation is realized by our monitoring component, which employs in the case of Java AspectJ, an aspect-oriented programming extension for Java [14]. AspectJ allows us to intercept an application by bytecode-weaving in order to gather necessary monitoring information for analysis and visualization purposes. Subsequently, this information is transported via (Transmission Control Protocol (TCP) towards a server, which hosts our GWT application. This part represents the two major components of our architecture, namely analysis and visualization. The analysis component receives the monitoring information and reconstructs traces. These traces are stored in the file system and describe a software landscape consisting of monitored applications and communication in-between. Our user-management employs a H2 database [15] to store related data. The software landscape visualization is provided via Hypertext Transfer Protocol (HTTP) and is accessible by clients with a web browser. GWT is an open source framework, which allows to develop JS front-end applications in Java. It facilitates the usage of Java code for server (backend) and client (frontend) logic in a single web project. Client-related components are compiled to respective JS code. The communication between frontend and backend is handled through asynchronous remote procedure calls based on HTTP. In ExplorViz Legacy, the advantages of GWT proved to be a drawback, because every change affects the whole project due to its single code base. New developed features were hard-wired into the software system. Thus, a feature could not be maintained, extended, or replaced by another component with reasonable effort. This situation was a leading motivation for us to look for an up-todate framework replacement. We intended to take advantage of this situation and modularize our software system in order to move from a monolithic, to a distributed (web) application divided into separately maintainable and deployable backend and frontend components.
IV. MODULARIZATION PROCESS AND ARCHITECTURE OF ExplorViz
The previously mentioned drawbacks in ExplorViz Legacy and recent experience reports in literature about successful applications of alternative technologies, e.g., Representational State Transfer (REST or RESTful) Application Programming Interfaces (API) [16], [17], were triggers for a modularization and modernization. In [18], we gave a very brief description on the modernization process of ExplorViz towards a microservice architecture. During the modularization planning phase, we started with a requirement analysis for our modernized software system and identified technical and development process related impediments in the project. We kept in mind that our focus was to provide a collaborative development process, which encourages developers to participate in our research project [18]. Furthermore, developers, especially inexperienced ones, tend to have potential biases during the development of software, e.g., they make decisions on their existing knowledge instead of exploring unknown solutions [19]. A more detailed description of decision triggers and the decision making process will be published in a technical report [20]. In general, there exist many drivers and barriers for microservice adoption [21]. Typical barriers and challenges are the required additional governance of distributed, networked systems and the decentralized persistence of data.
As a result of this process, we agreed on building upon an architecture based on microservices as shown in Figure 2. This architectural style offers the ability do divide monolithic applications into small, lightweight, and independent services, which are also separately deployable [3], [22]- [24]. However, the obtained benefits of a microservice architecture can bring along some drawbacks, such as increased overall complexity and data consistency [25].
1) Extensibility & Integrability:
In a first step, we modularized our GWT project into two separated projects, i.e., backend and frontend, which are now two self-contained microservices. Thus, they can be developed technologically independent and deployed on different server nodes. This allows us to exchange the microservices, as long as we take our specified APIs into account. The backend is implemented as a Java-based web service based on the Jersey Project [26], which provides a RESTful API via HTTP for clients. Furthermore, we replaced our custom-made monitoring component by the monitoring framework Kieker [27]. This framework provides an extensible approach for monitoring and analyzing the runtime behavior of distributed software systems. Monitored information is sent via TCP to our backend, which employs the filesystem and H2 database for storage. The frontend uses the JS framework Ember.js, which enables us to offer visualizations of software landscapes to clients with a web browser [28]. Since Ember is based on the model-view-viewmodel architectural pattern, developers do not need to manually access the Document Object Model and thus need to write less source code. Ember uses Node.js as execution environment and emphasizes the use of components in web sites, i.e., self-contained, reusable, and exchangeable user interface fragments [29]. We build upon these components to encapsulate distinct visualization modes, especially for extensions. Communication, like a request of a software landscape from the backend, is abstracted by socalled Ember adapters. These adapters make it easy to request or send data by using the convention-over-configuration pattern. The introduced microservices, namely backend and frontend, represent the core of ExplorViz. As for future extensions, we implemented well-defined extension interfaces for both microservices, that allow their integration into the core.
2) Code Quality & Comprehensibility: New project developers, e.g., students, do not have to understand the complete project from the beginning. They can now extend the core by implementing new mechanics on the basis of a plug-in extension. Extensions can access the core functionality only by a well-defined read-only API, which is implemented by the backend, respectively frontend. This high level of encapsulation and modularization allows us to improve the project, while not breaking extension support. Additionally, we do no longer have a conglomeration between backend and frontend source code, especially the mix of Java and JS, in single components. This eased the development process and thus reduced the number of bugs, which previously occurred in ExplorViz Legacy. Another simplification was the use of json:api [30] as data exchange format specification between backend and frontend, which introduced a well-defined JavaScript Object Notation (JSON) format with attributes and relations for data objects.
A. Software Configuration & Delivery
One of our goals was the ability to easily exchange the microservices. We fulfill this task by employing frameworks, which are exchangeable with respect to their language domain, i.e., Java and JS. We anticipate that substituting these frameworks could be done with reasonable effort, if necessary. Furthermore, we offer pre-configured artifacts of our software for several use cases by employing Docker images. Thus, we are able to provide containers for the backend and frontend or special purposes, e.g., a fully functional live demo. Additionally, we implemented the capability to plug-in developed extensions in the backend, by providing a packagescanning mechanism. The mechanism scans a specific folder for compiled extensions and integrates them at runtime.
V. PROOF-OF-CONCEPT IMPLEMENTATION
We realized a proof-of-concept implementation and split our project as planned into two separate projects -a backend project based on Jersey, and a frontend project employing the JS framework Ember. Both frameworks have a large and active community and offer sufficient documentation, which is important for new developers. As shown in Figure 2, we strive for an easily maintainable, extensible, and plug-in-oriented microservice architecture. Since the end of our modularization and modernization process in early 2018, we were able to successfully develop several extensions both for the backend and the frontend. Two of them are described in the following.
1) Application Discovery: Although we employ a monitoring framework, it lacks a user-friendly, automated setup configuration due to its framework characteristics. Thus, users of ExplorViz experienced problems with instrumenting their applications for monitoring. In [31], we reported on our application discovery and monitoring management system to circumvent this drawback. The key concept is to utilize a software agent that simplifies the discovery of running applications within operating systems. Furthermore, this extension properly configures and manages the monitoring framework. The extension is divided in a frontend extension providing a configuration interface for the user, and a backend extension, which applies this configuration to the respective software agent lying on a software system.
Finally, we were able to conduct a first pilot study to evaluate the usability of our approach with respect to an easyto-use application monitoring. The improvement regarding the usability of the monitoring procedure of this extension was a great success. Thus, we recommend this extension for every user of ExplorViz.
2) Virtual Reality Support: An established way to understand the complexity of a software system is to employ visualizations of software landscapes. However, with the help of visualization alone, exploring unknown software is still a potentially challenging and time-consuming task. For this extension, three students followed a new approach using Virtual Reality (VR) for exploring software landscapes collaboratively. They employed head mounted displays (HTC Vive and Oculus Rift) to allow the collaborative exploration of software in VR. They built upon our microservice architecture and employed WebSocket connections to exchange data to achieve modular extensibility and high performance for this real-time user environment. As a proof of concept, they conducted a first usability evaluation with 22 probands. The results of this evaluation revealed a good usability and thus constituted a valuable extension to ExplorViz.
VI. RESTRUCTURED ARCHITECTURE AND NEW PROCESS
Our modularization approach started by dividing the old monolith into separated frontend and backend projects [18]. Since then, we further decomposed our backend into several microservices to address the problems stated in Section II. The resulting, restructured architecture is illustrated in Figure 3 and the new collaborative development process is described below. As reported in Section V, the new architecture already improved the collaboration with new developers who realized new features as modular extensions.
1) Extensibility & Integrability: Frontend extensions are based on Ember's addon mechanism. The backend, however, used the package scanning feature of Jersey to include extensions. The result of this procedure was again an unhandy configuration of a monolithic application with high coupling of its modules. Therefore, we once again restructured the approach for our backend plug-in extensions. The extensions are now decoupled and represent separated microservices. As a result, each extension is responsible for its own data persistence and error handling. Due to the decomposition of the backend, we are left with multiple Uniform Resource Identifiers (URI). Furthermore, new extensions will introduce additional endpoints, therefore more URIs again. To simplify the data exchange handling based on those endpoints, we employ a common approach for microservice-based backends. The frontend communicates with an API gateway instead of several single servers, thus only a single base Uniform Resource Locator (URL) with well-defined, multiple URIs. This gateway, a Nginx reverse proxy [32], passes requests based on their URI to the respective proxied microservices, e.g., the landscape service. Furthermore, the gateway acts as a single interface for extensions and offers additional features like caching and load balancing. Extension developers, who require a backend component, extend the gateway's configuration file, such that their frontend extension can access their complement. The inter-service communication is now realized with the help of Apache Kafka [33]. Kafka is a distributed streaming platform with fault-tolerance for loosely coupled systems. The decomposition into several independent microservices and the new inter-service communication approach both facilitate low coupling in our system.
2) Code Quality & Comprehensibility: The improvements for code quality and accessibility, which were introduced in our first modularization approach, showed a perceptible impact on contributor's work. For example, recurring students approved the easier access to ExplorViz and especially the obligatory exchange format json:api. However, we still lacked a common code style in terms of conventions and best practices. To achieve this and therefore facilitate maintainability, we defined compulsory rule sets for the quality assurance tools Checkstyle and PMD. In addition with SpotBugs [34], we impose their usage on contributors for Java code. For JS, we employ ESLint [35], i.e., a static analysis linter, with an Ember community-driven rule set. All tools are integrated into our continuous integration pipeline configured in TravisCI [36].
A. Software Configuration & Delivery
One major problem of ExplorViz Legacy was the necessary provision of software configurations for different use cases. The first iteration of modularization did not entirely solve this problem. The backend introduced a first approach for an integration of extensions, but their delivery was cumbersome. Due to the tight coupling at source code level we had to provide the compiled Java files of all extensions for download. Users had to copy these files to a specific folder in their already deployed ExplorViz backend. Therefore, configuration alterations were troublesome. With the architecture depicted in Figure 3 we can now provide a jar file for each service with an embedded web server. This modern approach for Java web applications facilitates delivery and configuration of ExplorViz's backend components. In the future, we are going to ship ready-to-use Docker images for each part of our software. The build of these images will be integrated into the continuous integration pipeline. Users are then able to employ docker-compose files to achieve their custom Ex-plorViz configuration or use a provided docker-compose file that fits their needs. As a result, we can provide an alternative, easy to use, and exchangeable configuration approach that essentially only requires a single command line instruction. The frontend requires another approach, since (to the best of our knowledge) it is not possible to install an Ember addon inside of a deployed Ember application. We are currently developing a build service for users that ships ready-to-use, pre-built configurations of our frontend. Users can download and deploy these packages. Alternatively, these configurations will also be usable as Docker containers.
VII. RELATED WORK
In the area of software engineering, there are many papers that perform a software modernization in other contexts. Thus, we restrict our related work to approaches, which focus on the modernization of monolithic applications towards a microservice architecture. [25] present a survey of architectural smells during the modernization towards a microservice architecture. They identified nine common pitfalls in terms of bad smells and provided potential solutions for them. ExplorViz Legacy was also covered by this survey and categorized by the "Single DevOps toolchain" pitfall. This pitfall concerns the usage of a single toolchain for all microservices. Fortunately, we addressed this pitfall since their observation during their survey by employing independent toolchains by means of pipelines within our continuous integration system for the backend and frontend microservices. [ Fig. 3: Architectural overview and software stack of the restructured ExplorViz.
process to decompose an existing software system into several microservices. Additionally, they report from their gained experiences towards applying their presented approach in a legacy modernization project. Although their modernization drivers and goals are similar to our procedure, their approach features a more abstract point of view on the modernization process. Furthermore, they focus on programming language modernization and transaction systems. In [3], the authors present an industrial case study concerning the evolution of a long-living software system, namely a large e-commerce application. The addressed monolithic legacy software system was replaced by a microservice-based system. Compared to our approach, this system was completely re-build without retaining code from the (commercial) legacy software system. Our focus is to facilitate the collaborative development of open source software and also addresses the development process. We are further planning to develop our pipeline towards continuous delivery for all microservices mentioned in Section VI to minimize the release cycles and offer development snapshots.
VIII. CONCLUSION
In this paper, we report on our modularization and modernization process of the open source research software ExplorViz, moving from a monolithic architecture towards a microservice architecture with the primary goal to ease the collaborative development, especially with students. We describe technical and development process related drawbacks of our initial project state until 2016 in ExplorViz Legacy and illustrate our modularization process and architecture. The process included not only a decomposition of our web-based application into several components, but also technical modernization of applied frameworks and libraries. Driven by the goal to easily extend our project in the future and facilitate a contribution by inexperienced collaborators, we offer a plug-in extension mechanism for our core project, both for backend and frontend. We realized our modularization process and architecture in terms of a proof-of-concept implementation and evaluated it afterwards by the development of several extensions of ExplorViz. However, the modularization process is not fully completed, as yet. We are still improving the project in order to achieve a fully decoupled microservice architecture, consisting of a set of self-contained systems and well-defined interfaces in-between. In the future, we are planning to evaluate our finalized project, especially in terms of developer collaboration. Additionally, we plan to move from our continuous-integration pipeline towards a continuous-delivery environment. Thus, we expect to decrease the interval between two releases and allow users to try out new versions, even development snapshots, as soon as possible. Furthermore, we plan to use architecture recovery tools like [37] for refactoring or documentation purposes in upcoming versions of ExplorViz. | 2019-07-12T10:27:42.000Z | 2019-06-30T00:00:00.000 | {
"year": 2019,
"sha1": "261240537905ef99cb8b5256131f7da2887fd353",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "e960fb0246c73eb7088b481716c80421fec25d38",
"s2fieldsofstudy": [
"Computer Science",
"Engineering"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
264307429 | pes2o/s2orc | v3-fos-license | Copy number variation and population-specific immune genes in the model vertebrate zebrafish
Copy number variation in large gene families is well characterized for plant resistance genes, but similar studies are rare in animals. The zebrafish (Danio rerio) has hundreds of NLR immune genes, making this species ideal for studying this phenomenon. By sequencing 93 zebrafish from multiple wild and laboratory populations, we identified a total of 1513 NLRs, many more than the previously known 400. Approximately half of those are present in all wild populations, but only 4% were found in 80% or more of the individual fish. Wild fish have up to two times as many NLRs per individual and up to four times as many NLRs per population than laboratory strains. In contrast to the massive variability of gene copies, nucleotide diversity in zebrafish NLR genes is very low: around half of the copies are monomorphic and the remaining ones have very few polymorphisms, likely a signature of purifying selection.
Introduction
The innate immune system of an organism provides the first defense line against pathogens.Immune genes tend to evolve quickly and are often associated with a high degree of genetic variability.Many genes and proteins of the immune system are lineage-specific (limited to specific groups of animals, plants, or other taxa), while others have defense roles in a wide range of species.In particular, proteins containing a large nucleotide-binding domain followed by smaller repeats have an immune function in animals, plants, fungi, and bacteria alike (Ting et al., 2008;Jones and Dangl, 2006;Uehling et al., 2017;Gao et al., 2022).In animals, these repeats are usually leucine-rich repeats (LRRs) and the proteins themselves are classified as NLRs (nucleotide binding domain leucine-rich repeat containing, also known as NOD-like receptors).They have a multitude of functions: some act as pathogen sensors or transcription factors (Almeida-da-Silva et al., 2023), others are components or modulators of inflammasomes, large protein complexes that are assembled within cells as part of the response to biological or chemical danger (Almeida-da-Silva et al., 2023).
Plants have their own NLRs that are structurally similar to the ones from animals and also carry out central functions in the immune response (Urbach and Ausubel, 2017;Yue et al., 2012).Their diversity has been extensively characterized in several species, including the thale cress (Arabidopsis thaliana), and vastly different repertoires have been found from different strains or individuals (Van de Weyer et al., 2019b).NLR repertoires can also be referred to as NLRomes, and a species-wide repertoire is called the 'pan-NLRome'.
The majority of fish NLRs represent a fish-specific subtype that was originally labeled NLR-C (Laing et al., 2008), although they can be further divided into at least six groups based on structural similarities and sequence of conserved exons (Howe et al., 2016;Adrian-Kalchhauser et al., 2020).A schematic structure of proteins encoded by zebrafish NLR-C genes is presented in Figure 1A.All of them possess a FISNA domain (fish-specific NACHT-associated domain), which precedes the nucleotidebinding domain NACHT and is encoded by the same large exon near the N-terminus of the protein (Howe et al., 2016).FISNA-NACHT is in some cases preceded by the effector domain PYD, but this is encoded by a separate exon (Howe et al., 2016).Additionally, many NLR-C proteins have a B30.2 eLife digest Humans and other animals have immune systems that protect them from bacteria, viruses and other potentially harmful microbes.Members of a family of genes known as the NLR family play various roles in helping to recognize and destroy these microbes.Different species have varying numbers of NLR genes, for example, humans have 22 NLRs, but fish can have hundreds.400 have been found in the small tropical zebrafish, also known as zebra danios.
Zebrafish are commonly used as model animals in research studies because they reproduce quickly and are easy to keep in fish tanks.Much of what we know about fish biology comes from studying strains of those laboratory zebrafish, including the 400 NLRs found in a specific laboratory strain.Many NLRs in zebrafish are extremely similar, suggesting that they have only evolved fairly recently through gene duplication.It remains unclear why laboratory zebrafish have so many almost identical NLRs, or if wild zebrafish also have lots of these genes.
To find out more, Schäfer et al. sequenced the DNA of NLRs from almost 100 zebrafish from multiple wild and laboratory populations.The approach identified over 1,500 different NLR genes, most of which, were previously unknown.Computational modelling suggested that each wild population of zebrafish may harbour up to around 2,000 NLR genes, but laboratory strains had much fewer NLRs.The numbers of NLR genes in individual zebrafish varied greatly -only 4% of the genes were present in 80% or more of the fish.Many genes were only found in specific populations or single individuals.
Together, these findings suggest that the NLR family has expanded in zebrafish as part of an ongoing evolutionary process that benefits the immune system of the fish.Similar trends have also been observed in the NLR genes of plants, indicating there may be an evolutionary strategy across all living things to continuously diversify large families of genes.Additionally, this work highlights the lack of diversity in the genes of laboratory animals compared with those of their wild relatives, which may impact how results from laboratory studies are used to inform conservation efforts or are interpreted in the context of human health.domain (also known as PRY/SPRY) at the C-terminal end, separated from FISNA-NACHT by multiple introns and exons containing the LRRs (Figure 1A; Howe et al., 2016).The B30.2 domain functions through protein-protein interaction (Woo et al., 2006) and is also found in a variety of other genes such as the large family of TRIM ubiquitin ligases (van der Aa et al., 2009;Howe et al., 2016;Suurväli et al., 2022) that are often also involved in immunity.
It is not known why fishes possess so many NLRs, how they evolve, and how much within-species genetic variability they have.The previously observed repeated expansions and contractions of this family suggest it to have a high rate of gene birth and death (Suurväli et al., 2022).Studies have shown that viral and bacterial infections can induce the expression of specific fish NLRs (reviewed in Chuphal et al., 2022).Some of these have PYD or CARD domains and can even form inflammasomes similar to mammalian NLRs (Kuri et al., 2017;Li et al., 2018b).A species-wide inventory of major NLR exons in a model species such as zebrafish would provide valuable insights into the evolution and diversity of this large immune gene family.
Results
By adapting and modifying a protocol that combines bait-based exon capture with PacBio SMRT technology (Witek et al., 2016), we successfully generated circular consensus sequence (CCS) data for targeted parts of the immune repertoire from 93 zebrafish (of initial 96), representing four wild populations (Figure 1B) and two laboratory strains.With this approach, we aimed to sequence all exons in zebrafish that encode the nucleotide-binding FISNA-NACHT domains and all exons that encode B30.2 domains.Samples of one wild population (DP) suffered from poor sequence coverage and had to be excluded from downstream analyses in order to avoid bias in interpretation.Results involving this population are only shown in figure supplements and not in the main figures.
Our protocol used PCR with primers targeting ligated adapters to amplify the below-nanogram amounts of genomic DNA obtained from exon capture.This limited our fragment sizes to the lengths of what the polymerase was able to amplify.Zebrafish NLRs can have their exons spread out across tens of kilobases, so that we cannot know which exons belong to the same gene.However, we were able to use captured sequence surrounding the targeted exons to distinguish among near-identical coding sequences and separate NLR-associated B30.2 domains from B30.2 elsewhere in the genome.
The zebrafish pan-NLRome
We used an orthology clustering approach on NLR sequences assembled from all populations to create a reference set of NLRs (a pan-NLRome).This resulted in the identification of 1513 unique FISNA-NACHT containing sequences and 567 for NLR-associated B30.2 (NLR-B30.2).Nearly 10% of the sequences (145 FISNA-NACHT and 64 NLR-B30.2) contained pre-mature stop codons that were at least 10 amino acids from the end and led to early truncation of the protein.In total, 101 of the 1513 FISNA-NACHT were preceded by an exon containing the N-terminal effector domain PYD.Nearly all of those (97 out of 101) were found in group 1 NLR-C genes identified by the presence of the characteristic sequence motif GIAGVGKT (Howe et al., 2016).Since the combination of FISNA and NACHT is only present in NLR-C, its count of 1513 can be considered equal to the total number of NLR-C genes in the data.We found each individual zebrafish to have 100-550 NLR genes from the pan-NLRome in at least one copy (Figures 2 and 3), and only 50-75% of these have a high-quality match in the GRCz11 reference genome (Figure 2-figure supplement 2).In general, laboratory zebrafish had less NLRs than wild samples (Figure 2).The number and length of CCS reads and assembled contigs (both prior to orthology clustering) are presented in Whereas FISNA-NACHT is only found in NLRs, B30.2 domains are also found in other gene families.In addition to the 567 NLR-B30.2 domains, we also found 732 B30.2 domains not associated with NLRs.We were able to distinguish between them by utilizing the sequence of a short highly conserved 47 bp exon that appears to precede B30.2 in NLRs, but not in other genes (Figure 2 Copy number variation of NLR genes.(A) Sequence data from each individual zebrafish (vertical axis) was aligned to FISNA-NACHT exon sequences of the pan-NLRome (horizontal axis).Grayscale intensity shows, for each NLR, the proportion of NLR-aligning data in each given fish that matches this specific gene.Darker gray indicates a higher likelihood of this NLR being represented in multiple copies in the particular individual.Light gray indicates a single copy, white indicates absence.For clarity, only the 1235 FISNA-NACHT exons for which at least one fish had a minimum of 10 reads mapped to it are shown.(B) Numbers of pan-NLRome sequences (based on FISNACHT diagnosis) found in all three, two, or only one wild population.(C) Relative numbers of fish in which pan-NLRome sequences were found in wild populations.'Core' pan-NLRome: genes which are found in at least 80% of the sample (from a total of 57 wild fish); 'shell': genes in at least 20%; 'cloud': rare genes found in less than 20% of the sample.(D) Observed and estimated sizes of population-specific pan-NLRomes.Data points (filled circles and squares) show the average number of totally discovered NLR genes (as identified via their FISNA-NACHT domain) when investigating x fish.The dashed line is obtained by non-linear fit of the data to the function given in Equation 2. For all populations, the hypothetical pan-NLRome size -when extrapolating x → ∞ -is finite (see Table 1).
The online version of this article includes the following source data and figure supplement(s) for figure 3: Source data 1.Source tables for Figure 3 and its supplements.
Copy number variation in the pan-NLRome
Aligning CCS reads to the pan-NLRome revealed a considerable amount of variability in the proportion of reads mapping to them, both between and within populations (Figure 3A).This can be interpreted as the gene being present in different copy numbers.Furthermore, each NLR had its own distinct pattern of copy number variation, although generally the highest copy numbers were observed for the wild populations KG, SN, and CHT (Figure 3A).We also observed some sequencing batch-related differences, but the copy numbers differed even between individuals sequenced in the same batch.
-figure supplement 1).
There were also NLR sequences shared between just two wild populations, and some were restricted to a single population (Figure 3B).Moreover, we observed a lot of variability in the distributions of gene copies among fish within populations (Figure 3C).Only around 4% of the genes in the pan-NLRome were found in 80%, or more, of the wild fish.They constitute the core NLRome (Van de Weyer et al., 2019a).Most genes (51%) were found in the so-called shell of the pan-NLRome (20-80% of fish).Almost as many (45%) are found in a few fish (less than 20% of the sample) only.Although 60% of NLR genes occur in all wild populations, only 4% are omnipresent, that is, are in the core pan-NLRome.Thus, there is considerable variation in the NLR repertoires of individuals from the same population.
The total number of NLRs identified in a number x of individual fish can be fitted to a harmonic function (Medini et al., 2020).Using this function (see 'Materials and methods'), we estimated the sizes of the NLRomes of the populations (Figure 3D) and found a total of 520 and 570 NLRs in the laboratory strains TU and CGN, respectively (Table 1).For the wild populations, we estimated four times as many: 2283 in KG, ,896 in SN, and 2452 in CHT.
Differences from the reference genome
NLRs sequenced in this study were often different from those present in the reference genome GRCz11.Even NLRs sequenced from the strain that the reference genome itself is based on (TU) did not always align well to it.When the exon itself did align, the intronic sequences surrounding it could often be very different from the reference.In numbers, only around 75% of NLRs occurring in TU fish aligned to the reference genome GRCz11 with high mapping qualities (Figure 2-figure supplement 2A).This number dropped even lower elsewhere -from 60-65% of NLRs in CGN which aligned well to the reference, down to only around 50% for the wild populations.The majority of NLRs that did not map well had a very poor mapping quality of 1 (Figure 2-figure supplement 2B).Moreover, there were 9 FISNA-NACHT and 10 NLR-associated B30.2 in the pan-NLRome which did not map anywhere in the reference genome.
Purifying selection on single-nucleotide variants
We used the pan-NLRome as a reference for identifying single-nucleotide polymorphisms in the data.NLR sequence diversity was rare, with a large fraction of exons not having any variants in any of the populations.If variants were present, nucleotide diversity ( θπ ) was up to 0.016 and Watterson's estimator ( θw ) up to 0.021 (Figure 4A and B).In laboratory strains, genetic variability of FISNA-NACHT exceeded that of B30.2, but no such pattern was observed for wild populations.B30.2 exons of laboratory strains were also less variable than B30.2 from wild zebrafish (Figure 4B).The proportion of exons without any polymorphisms was much higher among FISNA-NACHT than among B30.2 (Figure 4C).The majority of variable NLR exons had θπ/θw ratios of less than 1 (Figure 4D), indicating an excess of rare alleles.
Discussion
We sequenced and assembled the FISNA-NACHT and B30.2 exons of hundreds of NLRs from 93 zebrafish.We were able to capture the diversity of this gene family in three wild populations and two laboratory strains, and produced lower coverage NLR data for an additional wild population (DP).
Analyzing the 73 zebrafish from populations other than DP, we found evidence that each genome from a wild individual contains only a fraction of more than 1500 identified NLR copies.The number of NLRs found per individual, each with one or more copies, ranged from around 100-550.Some of the lower counts were likely underestimated due to low sequencing depths in specific samples.Since all samples from population DP suffered from low read depth, their analysis is only shown in figure supplements.As targeted sequencing based on bait-capture requires sufficient homology between bait and target, diverged NLR exons may have been missed in our approach.This affects B30.2 exons even more than FISNA-NACHT exons because they are much shorter.However, the observed slow increase in newfound NLR gene copies per sequenced individual after the first few individuals indicates that not many NLRs were missed.The sizes of NLR repertoires differ between zebrafish individuals in the three wild populations.
Nonlinear fitting of NLR counts to Equation 2 suggested that the investigated populations all possess closed pan-NLRomes with roughly 500-600 NLRs in the laboratory strains and around 2000 NLRs in the wild populations.The total numbers of NLRs with a B30.2 exon are about 170 in the laboratory strains and between 1100 and 1500 for the wild populations (Table 1).To explore the entire NLRome of wild populations, large samples are needed: based on the curve-fitting results, we estimate that capturing 90% of the NLRome may require up to several hundred thousand fish (Table 1).Orthogroup clustering with the data from DP resulted in 47 FISNA-NACHT exons which did not occur in any other population.Our results suggest that the pan-NLRome of the entire species must be vastly larger than what we have been able to detect with our limited sample sizes from a limited number of populations.Geographically distant populations -for example, in Nepal or the Western Ghats (Whiteley et al., 2011) -likely harbor many more NLRs which are not present in the populations we sequenced.
Although a few zebrafish assemblies are available in addition to the reference genome, for instance, the fDanRer4.1 assembly from the Tree of Life Initiative (GCA_944039275.1),none of those provide a suitable framework for mapping and analyzing NLRs on their own.One of the hindrances is the fact that the majority of NLR genes are located on the notoriously difficult to assemble long arm of chromosome 4, which harbors plenty of structural variation (McConnell et al., 2023).Furthermore, large multi-copy gene families are difficult to analyze.Read mapping and counting of copies in a particular genome is not trivial.Any downstream analysis which relies on clearly distinguishing paralogous and orthologous comparisons becomes fuzzy, if not impossible.Still, improving sequencing technology and the rising interest in pan-genomic studies Bayer et al., 2020;Sherman and Salzberg, 2020;Liao et al., 2023 have already started to transform the data structures in which genomes are stored, away from a single-reference genome-based view, toward graph-based genome networks.Whether the promise of a thereby improved inventory of structural variation of a species holds up remains still to be seen.Anyway, as shown for the zebrafish NLRs, the availability of a single high-quality reference genome is certainly not sufficient neither to identify nor to understand the diversity of large gene families.
Properties of the zebrafish NLRome
We have previously demonstrated a substantial reduction in single-nucleotide variation in zebrafish laboratory strains compared to wild populations (Suurväli et al., 2020).Here, we showed that the copy numbers of the NLRome and their variation are also heavily reduced.The most obvious explanation for this observation is the recent population bottleneck which marks the establishment of laboratory strains.The reduction in copy number variation in the major histocompatibility complex (MHC) locus in a population of greater prairie-chicken was attributed to a recent bottleneck as well (Eimes et al., 2011).Additionally, the reduced amount of pathogenic challenges in a laboratory environment could lead to a steady loss of expendable genes.For these reasons, one has to exercise caution when extending conclusions from immune-related studies on laboratory zebrafish to wild zebrafish.The same caution should also be exercised when extending results from laboratory organisms to other species, including human.
Studies have shown that even mammals have hundreds of genes with diverse molecular functions that are affected by copy number variation, even though it rarely involves full genes (Kooverjee et al., 2023;Zarrei et al., 2015).One example of the latter is the MHC locus, which harbors varying numbers of gene copies between closely related species of ruminants (He et al., 2024) and has haplotype-specific copies in mice (Lilue et al., 2018).However, the vast number of NLRs in zebrafish combined with presence/absence variation (McConnell et al., 2023) and high rates of duplication exceeds what has been found in other animals so far.A comparable situation can be found in the NLR genes of the thale cress (A.thaliana).Our predicted number of NLRs in a zebrafish population is on the same scale as the 2127 NLRs found in the thale cress NLRome (Van de Weyer et al., 2019b).Moreover, copy numbers also vary greatly between A. thaliana accessions (Lee and Chae, 2020).A total of 464 conserved, high-confidence orthogroups were identified in A. thaliana, 106-143 of which were defined as the core NLRome because they were found in a subset comprising at least 80% of the accessions (Van de Weyer et al., 2019b).In wild zebrafish, we found a set of 880 NLR genes which were detected in at least one individual from three wild populations, but only 58 NLRs were found in the vast majority (more than 80%) of wild individuals.Although structural similarities of NLRs in plants and animals are thought to be the result of convergent evolution (Yue et al., 2012), it appears that the similarities extend to their evolutionary trajectories.However, the overall number of gene copies as well as the variation in copy numbers within populations and in individual gene repertoires are more extreme in zebrafish than in A. thaliana.
We postulate that as immune genes, many NLR genes are likely shared between populations because they provide a fitness advantage in the defense against common pathogens.The additional NLRs shared among only some of the wild populations and the population-specific NLRs may represent local adaptations to ecological niches.Additionally, there could be functional redundancy within the NLRome, so that different individuals have different NLRs with the same functional role.In general, the fact that hundreds of NLR gene copies are maintained in zebrafish, together with a signature of purifying selection, suggests that the evolution of these genes is far from neutral.Although the expression of fish NLRs is often induced by pathogen exposure (reviewed in Chuphal et al., 2022), the exact function of most zebrafish NLR-C genes remains unclear.It is possible that some of them participate in the formation and activity of inflammasomes (Li, 2018a;Valera-Pérez et al., 2019;Lozano-Gil et al., 2022, Kuri et al., 2017), but we only found the N-terminal effector domains (CARD or PYD) that are typically involved in this function (Petrilli et al., 2005) in a small subset of NLR-C genes.
Although we mainly used the counts of FISNA-NACHT orthogroups to estimate total numbers of NLRs, we also analyzed the B30.2 exons of NLR-C genes.In general, NLR-associated B30.2 exons exhibit patterns of copy number variation that are similar to those seen for FISNA-NACHT.For example, about half of the B30.2 sequences are found in all wild populations, similar to the set of 880 FISNA-NACHT exon sequences conserved among populations.
What drives the copy number differences?
There are at least two mechanisms which could contribute to the extensive copy number variation seen among zebrafish populations: first, it could be attributed to a high degree of haplotypic variation.Large DNA fragments contain different sets of genes and gene copies, similar to the zebrafish MHC loci (McConnell et al., 2014).Extensive haplotypic variation occurs on the long arm of chromosome 4, the location containing over two-thirds of all NLRs in zebrafish (McConnell et al., 2023).Such segregating haplotype blocks would explain the existence of the core NLRome, but not the frequent presence of genes that occur only in a single individual.
Alternatively or additionally, the evolution of NLR-C genes could be driven by duplication events (Cannon et al., 2004) and gene conversion (Laing et al., 2008).Gene duplications can be caused by unequal recombination, transposon activity, or whole genome/chromosome duplications (Magadum et al., 2013;Kapitonov and Jurka, 2007).The arrangement of NLR-B30.2 genes in clusters on the long arm of chromosome 4 suggests that tandem duplication via unequal crossing-over (Otto et al., 2022) played the most important role in the expansion.Since there are many transposable elements on the long arm of chromosome 4 (Howe et al., 2013), it would be reasonable to assume that at least some of them have assisted in the local expansion and transfer of NLR exons and genes to chromosomes other than chromosome 4. Since our targeted sequencing approach does not elucidate the genomic arrangement of the NLR gene copies and many of them do not have recognizable orthologs in the reference genome, we cannot draw further conclusions about the role of tandem arrays in their evolutionary trajectory.
It is tempting to speculate that chromosome 4 could be a source of NLRs which continuously generates new copies.However, gene gains must be balanced by gene loss to maintain a stable genome size.NLR-C genes may be lost via accumulation of random mutations due to a lack of selective pressure and loss-of-function mutations, but they may also be lost through unequal recombination.This mechanism would allow only NLR genes contributing to the functionality of the immune system to be kept, while others would disappear.
In the similarly evolving plant NLRs, tandem duplication is thought to be the primary driver of NLR gene expansion (Cannon et al., 2004), but they are also often associated with transposable elements.If the diversity of unrelated NLR genes in such distantly related species is driven by common molecular mechanisms, then the same mechanisms might also act on NLRs of other phylogenetic clades and even on unrelated large gene families, such as odorant receptors (Mombaerts, 1999).
Conclusion
This study showcases an example of the evolutionary dynamics affecting very large gene families.The sheer amount of copy number variation that appears to be present in a single gene family of zebrafish is staggering, with different individuals each having numerous genes that are not present in all others.This can only be caused by diversity-generating mechanisms that are active even now.In this study, we have laid the groundwork for future studies investigating the molecular basis and evolutionary mechanisms contributing to the diversity of large, vertebrate gene families.
Samples
Wild zebrafish from four sites in India and Bangladesh (Figure 1B) had been collected in the frame of other projects (e.g., Whiteley et al., 2011;Shelton et al., 2020).Laboratory zebrafish from the Tübingen (TU) and Cologne (CGN) strains were provided by Dr. Cornelia Stein from the Hammerschmidt laboratory (Institute for Zoology, University of Cologne).All samples were stored in 95% ethanol until use.Tail fins from 20 fish per wild population and 8 fish per laboratory strain were used as starting material for the subsequent steps.
DNA extraction, exon capture, and sequencing Genomic DNA was extracted with kits from QIAGEN (MagAttract HMW kit) and MACHEREY-NAGEL (Nucleospin Tissue Kit), followed by shearing with red miniTUBEs on the Covaris M220 ultrasonicator.Nicks in the DNA were repaired with PreCR Repair Mix (New England Biolabs).Samples were barcoded with the NEBNext Ultra II DNA Library Prep Kit, then pooled together in batches of four or eight (details provided in Appendix 1).RNA baits for the exon capture (Daicel Arbor Biosciences) were custom-designed to target immune genes of interest (mainly NLRs, but also some others) based on version GRCz10 of the reference genome.Bait sequences and target locations are available in Figure 2-source data 2. Exon capture and PacBio library preparation were both done according to a protocol adapted from Witek et al., 2016.Libraries were sequenced at the Max Planck-Genome-Centre Cologne, with PacBio Sequel and Sequel II.Additional details are provided in Appendix 1.
Read processing, mapping, and clustering
Raw sequences were de-multiplexed with lima.Consensus sequences of DNA fragments with at least three passes (CCS reads) were inferred with ccs, followed by PCR duplicate removal with pbmarkdup.
Modeling
To estimate the full size of each population's NLR repertoire, we calculated the increment in the total number of identified NLR exon sequences when adding sequence data from one additional individual of a population to a set of already surveyed individuals.As noted earlier (Medini et al., 2020), these increments are well approximated by a power-law decay.
Briefly, given a sample of n individuals, there are ways to choose x − 1 individuals from the entire sample and add another -not yet chosen -one.For each x , we calculated the increment in the number of identified exon sequences and averaged over all possible choices of individuals.Summation of the average increments yields the total number of exons identified with x individuals, as plotted in Figure 3D.Then, we fitted the nonlinear function where H(x, β) is the generalized harmonic number with parameter β , that is, It represents the sum of increments, decaying according to a power-law, with parameters α (intercept) and β (decay rate).Importantly, if β > 1 , the series in Equation 3 converges and its limit may be interpreted as the size of a closed NLRome.The NLRome is open, if β ≤ 1 .Values of the fitted parameters and saturation limits are presented in Table 1.
Genetic diversity
Single-nucleotide genotypes in each fish were identified from the. bam output of pbmm2 by using deepvariant (r1.0) (Poplin et al., 2018) with the PacBio model.Joint genotyping of the individual samples was done with glnexus (v1.2.7-0-g0e74fc4) (Yun et al., 2021) with its deepvariant-specific setting.Per-site θπ of the NLR exons was calculated with vcftools (v0.1.16)(Danecek et al., 2011).Watterson's estimator of the scaled mutation rate is where S is the number of segregating sides seen in a sample of n aligned sequences, each of size l (here, 1761 bp for the FISNA-NACHT exons and 540 bp for the B30.2 exons).
Under neutrality (all alleles confer the same fitness to an individual) and constant population size over time, one expects equality θπ = θw .
Data availability
NLR reads are available in the NCBI Sequence Read Archive (BioProject PRJNA966920).Scripts are available on GitHub (https://github.com/YSchaefer/pacbio_zebrafish,copy archived at Schaefer, 2024).Sequences of the hybridization baits are provided as a source dataset.
The following dataset was generated: The following previously published dataset was used: Library construction and sequencing The final libraries were prepared with the SMRTbell Template Prep Kit 1.0-SPv3 (Pacific Biosciences).At the ligation step, the recommended amount of PacBio adapters was increased from 1 to 5 µl per 40 µl total reaction volume and the reaction was incubated overnight at room temperature.For the SN and CHT libraries in pools of eight (see Appendix 1-table 1), barcoded PacBio adapters were used instead of regular ones.The product codes for barcodes were BC1001 and BC1002 for CHT, BC 1003 and BC1004 for SN.The first libraries (TU, CGN, 4 DP and 4 KG samples) were size selected to 2-8 kb with the BluePippin pulsed field electrophoresis system (Sage Science).The following libraries were size selected to 1.5-8 kb.
All sequencing was done at the Max Planck-Genome-Centre Cologne.All TU, CGN, DP, and KG zebrafish, as well as four CHT and four SN samples were sequenced with 1M v2 SMRT Cells of the Sequel instrument (Pacific Biosciences).The rest of the samples (all with barcoded adapters) were multiplexed together and sequenced with an 8M SMRT Cell of the much higher throughput Sequel 2 instrument (Pacific Biosciences).One of the already sequenced SN samples (SN24) was also resequenced in this run as it yielded no data in the first one.Furthermore, Pacific Biosciences upgraded their kits with a superior polymerase after we had sequenced TU, CGN and the first four samples of each wild population; all samples other than those were sequenced with their LR (long run) polymerase.
An overview of the sequencing is presented in Appendix 1-table 1.
Read processing and assembly
Raw data were de-multiplexed and stripped of primer/adapter sequences with lima from Pacific Biosciences.For the samples sequenced with the PacBio Sequel I, the parameters -enforcefirst-barcode -split-bam-named -W 100 were used with lima v1.0.0 for the runs without the LR polymerase.For Sequel runs with the LR polymerase, lima v1.8.0 and v1.9.0 were used with the same parameters.To remove PacBio barcodes from the data produced on Sequel II, lima v1.11.0 was used with parameters -split-bam-named -peek-guess and for the subsequent removal of NEBNext barcodes, the parameters were changed to -enforce-first-barcode -split-bam-named -peek-guess.
Consensus sequences of all DNA fragments with a minimum of three passes (henceforth referred to as CCS reads) were calculated using ccs (v4.2.0,Pacific Biosciences) with default parameters.PCR duplicates were identified and flagged with pbmarkdup (v1.0.0,Pacific Biosciences) with default parameters, then excluded from downstream analyses.Any chimeric reads containing a primer sequence in the middle were identified with blastn (v2.11.0+) (Altschul et al., 1990) and removed.The filtered CCS reads were assembled into contigs for each fish separately using hifiasm (v0.15.4-r347) (Cheng et al., 2021) with default parameters.
NLR identification
To obtain a list of NLR gene positions in the reference genome, we first extracted known NLR locations from Ensembl.In addition, the reference genome was translated in all frames using transeq (from EMBOSS:6.6.0.0) (Rice et al., 2000) and searched for further NLRs using hmmsearch from hmmer (v3.2.1) (Wheeler and Eddy, 2013), without bias correction and with the hidden Markov model (HMM) profiles for zf_FISNA-NACHT and zf_B30.2from Adrian- Kalchhauser et al., 2020.Each position in which the zf_FISNA-NACHT model found a hit with a maximum i-Evalue of 1e − 200 and a minimum alignment length of 500 aa was considered a FISNA-NACHT exon.The filtering thresholds for B30.2 exons were an i-Evalue of 1e − 5 and a minimum alignment length of 150 aa.This approach was used both during bait design and as a preparatory step for the first round of read filtering.
To distinguish CCS reads of NLR genes from other CCS reads, the CCS reads of each fish were mapped against the reference genome GRCz11 using pbmm2 (v1.3.0) with preset ccs.CCS reads which mapped within a known NLR gene or one found with our HMM-based approach with any mapping quality were considered potential NLR reads and used as input for subsequent steps.
De novo assembled contigs containing NLR exon sequences were identified by translating all contigs of each fish in all frames with transeq (from EMBOSS:6.6.0.0) and subsequently searching for FISNA-NACHT and B30.2 domains using hmmsearch from hmmer (v3.2.1) without bias correction and the HMMs zf_FISNA-NACHT and zf_B30.2again.The HMM-based approach was chosen for the contigs in particular because we assumed that there would be NLR sequences in the data which
Figure 1 .
Figure 1.Structure of zebrafish NLRs and a map showing the origin of wild zebrafish samples.(A) Generalized, schematic representation of the domain architecture of an NLR-C protein.Each box represents a translated exon.The N-terminal repeats, the death-fold domain, as well as the B30.2 domain only occur in subsets of NLR-C genes.The number of N-terminal repeats and leucine-rich repeats can vary.Domains that can be either present or absent in different NLRs are surrounded by square brackets.(B) Sampling sites for wild zebrafish.All sites are located near the Bay of Bengal.Final sequenced sample sizes are indicated in parentheses.The map is based on geographic data collected and published by AQUASTAT from the Food and Agriculture Organization of the United Nations (FAO, 2021).The population DP is marked with an asterisk because its analysis and results are presented only in figure supplements.
Figure 2 .
Figure 2. Total counts of NLRs found per individual, shown for each population.Black diamonds on the box plots denote means, horizontal lines denote medians.Left side: two laboratory strains; right side: three wild populations.The online version of this article includes the following source data and figure supplement(s) for figure 2: Source data 1.Source tables for Figure 2 and its supplements.Source data 2. Sequences and target locations of RNA baits.
Figure supplement 1 .
Figure supplement 1.Sequencing and assembly statistics of circular consensus sequence (CCS) reads from NLR exons.
Figure 3 .
Figure3.Copy number variation of NLR genes.(A) Sequence data from each individual zebrafish (vertical axis) was aligned to FISNA-NACHT exon sequences of the pan-NLRome (horizontal axis).Grayscale intensity shows, for each NLR, the proportion of NLR-aligning data in each given fish that matches this specific gene.Darker gray indicates a higher likelihood of this NLR being represented in multiple copies in the particular individual.Light gray indicates a single copy, white indicates absence.For clarity, only the 1235 FISNA-NACHT exons for which at least one fish had a minimum of 10 reads mapped to it are shown.(B) Numbers of pan-NLRome sequences (based on FISNACHT diagnosis) found in all three, two, or only one wild population.(C) Relative numbers of fish in which pan-NLRome sequences were found in wild populations.'Core' pan-NLRome: genes which are found in at least 80% of the sample (from a total of 57 wild fish); 'shell': genes in at least 20%; 'cloud': rare genes found in less than 20% of the sample.(D) Observed and estimated sizes of population-specific pan-NLRomes.Data points (filled circles and squares) show the average number of totally
Figure supplement 2 .
Figure supplement 2. Copy number variation of NLR genes, including the DP population.
Figure 4 .
Figure 4. Single-nucleotide variation in NLR exons.Pairwise nucleotide diversity ( θπ ) and Watterson's estimator of the scaled mutation rate ( θw ) for FISNA-NACHT (A) and NLR-associated B30.2 (B) exons.(C) Proportion of exons without any single nucleotide polymorphisms.(D) Ratio of θπ/θw .Only exons with at least one single-nucleotide polymorphism are shown.The dotted, horizontal line marks a ratio of 1, the expected value under neutrality and constant population size.The black diamonds on box plots denote means, horizontal lines denote medians.The online version of this article includes the following source data and figure supplement(s) for figure 4: Source data 1.Source tables for Figure 4 and its supplement.
Figure supplement 1 .
Figure supplement 1. Single-nucleotide polymorphisms of different NLR exons shown by population, including DP.
Table 1 .
Values of fitted parameters and saturation limits for FISNA-NACHT and NLR-B30.2 exons, by population. | 2023-08-29T13:11:10.831Z | 2023-10-16T00:00:00.000 | {
"year": 2024,
"sha1": "2c21ea6925185271a29963e3bdf0911313f4e5c6",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.7554/elife.98058",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "4f90d7383e579486e1ac305c3298663e19adc5c2",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology"
]
} |
56474930 | pes2o/s2orc | v3-fos-license | Panic Indicator for Measurements of Pessimistic Sentiments from Business News
Computational semantic analysis was applied for measurements of pessimistic sentiments from business news. Using semantic tree method, the number of pessimistic and optimistic news was estimated for the analysis of the bearish and bullish stock markets. It was found that the number of pessimistic and optimistic macroeconomic news is very sensitive to the sharp changes of the market indices at the extreme market conditions. A new sentiment indicator was constructed for quantitative measurements of the pessimistic investor sentiments. The proposed sentiment indicator is called a Panic Indicator. We found that the Panic Indicator is appropriate for the explanation of the relationship between the negative public information and stock index declines (S&P-TSX), as well as the sharp changes of VIX index. The proposed Panic Indicator would be useful for stock price modeling, the quantified description of the pessimistic opinions, and computational trading algorithms.
Introduction
Since the financial crisis on the US market in 2007-2008, there has been an increasing interest in the study of the role of the investor sentiments in the stock market. The pessimistic or optimistic sentiments can be defined as the herd behaviour of investors, who follow the economic, political, or social factors with respect to the future stock price returns. The basic concept of herd behavior was stated by Tarde in 1903 in the idea of the imitation process, for which people learn from one another through the process of imitation. The market crowd behavior was studied by Soros (1987) using the concept of reflexivity of the market participants. The concept of reflexivity may explain the violation of the Random Walk Hypothesis during boom and bust market through disequilibrium process in the financial market. The negative expectations or pessimistic sentiments can lead to the downward direction of the stock market.
The difficulty for measuring herding effects lies in the lack of the effective model of the disequilibrium market as well as the methods, which could measure the activity of rational and irrational investors in the speculative market. There are different models for the analysis of the investor sentiments. Existing quantitative approaches of the sentiment measurements have mainly focused on the empirical investigation of the time-series of the price returns, trading volume, the dividend premium, IPO volume, volatility (see, for example, Baker & Urgler, 2007;Brown & Cliff, 2004). DeLong, Shliefer, Summers and Waldmann (1990) proposed the behavioural model to explain the risk created by the unpredictability of the unsophisticated investor's opinions. The model is based on the abstractive qualitative analysis of two types of investors: rational arbitrageurs (sentiment-free) and irrational investors. They showed that the market risk could be generated by the activity of the unsophisticated investors. The trading activity of the unsophisticated investors could limit the applications of the various arbitrage strategies applied by rational arbitrageurs. Remorov (2014) showed that the disequilibrium market can be modelled through cash-flow exchange between market-makers at the stock market crashes through behavioural process. Graham (1999) investigated the impact of public and private information on the herd behaviour among investment analysts. Graham proposed the qualitative reputation herding model to describe the relation between information signals with herd behaviour of smart analysts. Hong et al. (2000) developed a behavioural model, which explains the sharp decline of profitability of momentum strategies with the firm size. Most interesting research works are related to the information-based herding model (Welch, 1992), compensation based herding model (Brennan, 1993), and reputation based herding model (Scharfstein & Stein, 1990).
The study of the information-based herding requires the advanced semantic methods for the computational analysis of the public information impact. Beth (1954) invented the method of semantic tableaux (semantic tree method). A "Bag of Words" method was developed by Harris (1954) and is commonly used methods of the document classification, which consist of the distributional methodology for linguistics. Using the concept of the similarity of the members for each basis classes, Harris (1954) grouped the members according to the distributional relations between members of each basis classes. Some approaches were developed using the "Bag of Words" methods (see, for example, a Noun Phrase approach reported by Tolle & Chen, 2000). The goal of the automated application the textual analysis is closely related to the machine learning algorithms, which are intensively developed with the web-semantic analysis. The machine learning algorithm uses different statistical techniques to optimize the model parameters with forecasting the stock price movement. The Genetic Algorithm was applied by Thomas and Sycara (2002) to investigate the stock price movement with the number of postings and the number of posted words on a daily basis. The Naïve Bayesian technique uses the weighted vector of keywords for the stock rice prediction applying textual classification; see Seo, Giampapa et al. (2002).
Public business news is primary information source for the both rational and irrational investors. There is a limited range of semantic techniques for the analysis of the public business news with respect to the pessimistic/optimistic investor sentiments. Significant work was undertaken to develop the computational linguistic technique for stock price forecast. Mitchell and Mulherin (1994) studied the relation between the number of news announcements reported daily by Dow Jones and Company and aggregate measures of securities market activity including trading volume and market returns. Tetlock et al. (2008) showed that the linguistic media content may capture the information of the company fundamentals, and; therefore, investors may quickly incorporate this information in the different strategies. Ormos and Vazsonyi (2011) investigated the relation of the nouns and adjectives within business sentences with positive and negative returns of the S&P 500 stocks. Xie et al. (2012) developed the semantic approach for the analysis the business news with respect to the stock price movement and the company performance. They showed that their developed algorithm of the semantic analysis gives the significantly better prediction performance of the price return. Birza and Lott (2011) empirically showed that business news about GDP and unemployment does affect the stock returns. Zhang et al. (2011) investigated the textual information from Twitter and found that emotional tweet percentage significantly negatively correlated with Dow Jones, NASDAQ and S&P 500, but displayed significant positive correlation to VIX.
The range of the potential applications of the semantic analysis is extremely broad. In the current report, the new methodology of the pessimistic sentiment measurements was set out for computational linguistic analysis with the focus on the extreme market conditions. Our motivation of the sentiments measurement methodology is related to the construction of the computational linguistic analysis of the impact of public news on equity market for replication of the typical behaviour of financial analysts. The new sentiment indicator is presented for the measurement of the panic herd behaviour during the extreme market conditions. We investigate how shocking market conditions are related to the issues of business news and investor sentiments. The current article demonstrates first results of the study of the relations between stock price and investor sentiment during market crashes.
Measurements of the Pessimistic Sentiments from Business News
Our sentiment technique is mainly based on the semantic tree method with subjective classification of the selected sentences. Some principles of the subjective classification and semantic tree methods are described in the academic literature (Beth, 1955;Bondecka-Krzykowska, 2005;Liu, 2010). A large group of the sentences were selected from three types of the textual information sources: business sentences from the economic analytical articles, titles from analytical articles, sentences from educational courses, (for example, CFA courses) with positive or negative impact on the equity market. The textual database exceeded 65,000 sentences (connectives). The textual information was used as a reference benchmark database for the application for the semantic analysis of new business news. The selected sentences included the information from economics, accounting, derivative instruments, technical analysis for the stock market, political news, social news, and so on. Although some topics, such as the technical analysis from CFA courses, don't have scientific support, we used the application of technical analysis for the stock market as an important part of the herd behavioral processes in the financial industry.
Our developed algorithm monitors business analytical news on a daily basis directly from web-publishers, such as Thomson Reuters, Bloomberg, The Wall Street Journal, and The New York Times Company. The number of www.ccsenet.org/ibr International Business Research Vol. 7, No. 5;2014 issued negative and positive news was calculated for the weekly period from different news agencies, however, in the current report we present the results of our analysis obtained from Thomson Reuters. The articles from Thomson Reuters were collected since 2007 year; the total number of articles exceeded 6 mln articles. Textual information from newspaper titles and news abstracts was used for the semantic analysis; because the titles and abstracts better describe the news announcement in general. Similar approach was used by Birza and Lott (2011).
Initially, new business news was separated by 6 topic categories with the following hierarchy: macroeconomic news, government news, stock market news, real estate news, company specific news, and commodity news using the key-word classification (for example, the key-words "oil" and "NYSE" define commodity news and stock market news, respectively).
The logic tree method was applied for the selection of the documents as optimistic, pessimistic, and neutral information with respect to the growth of the US stock market. The method of the verification consists of finding the best coincidences of the benchmark sentences with the tested sentences. Each sentence from the benchmark sentence database (65,000 sentences) was compared with the title or abstract of the article from the web-publisher. Then, the number of optimistic and pessimistic news was counted for each day. The results of the sentiment analysis are published and updated weekly in web-site www.newsreturn.com.
Results
In the current article, the sentiments measurements rely upon on the extreme panic market conditions with two key objectives. The first is the development of the application of the semantic approach for the computational analysis of public business news to determine the analyst opinions. The second is the application of the sentiment indicator for the measurement of the investor sentiments with respect to their expectations of the future stock price return. We consider the usefulness of the application of the news sentiment indicator from two perspectives: as a quantitative measure of the pessimistic investor sentiments from business news and the impact of negative news on the equity market. Many of the study in the literature rely on the relationship of public information and stock price return. Nevertheless, the impact of public news can be strictly criticized for number reasons. For example, the issue of public news can cause the stock price changes and vice versa. As pointed out by Mitchell and Mulherin (1994), some newspaper articles might be written after the closing of the stock market on the day of the economic releases. For this reason, we also investigated the relation between the issued articles with the past events in stock market. We found that issued news, which is specified as stock market news, had the highest correlation with past events (stock price movement) and, therefore, did not have the predictive power for modeling. In contrast to stock market news, macroeconomic news demonstrated short and long-term correlation with future stock price indices. For this reason, macroeconomic news was taken for our investigation for the detection of the news impact on the investor sentiments.
We also tried to use the statistical methods to count the bad and good words from the articles, but we found that the justification of optimistic or pessimistic documents from the statistical methods (bad-good word counting) did not work.
The sample tests of randomly selected news showed that the accuracy of the sentiment news prediction exceeded ~90%. The subjective sentiment classification is most critical part in our computational linguistic technique, because this approach replicates the behaviour of the equity analyst for the short/long term price return perspective.
Obviously, the proposed method is very sensitive to the subjective classification of the original benchmark sentences for affective judgment of the optimistic or pessimistic documents. However, we consider that the proposed method is simple and robust for the semantic analysis without contribution of the possible modeling risk from the statistical methods. Moreover, the proposed method can be easily extended and modified for optimization with the maximum news impact for forecasting purposes.
We analyzed the time-series of various indicators, which include optimistic and pessimistic news from different selected topics: macroeconomic news, political news, monetary policy, stock market, real estate, company specific news, and commodities. Since the number of combinations of the possible indicators is extremely broad, we focused on the sensitivity of the indicators to the extreme market conditions only. We observed from the analysis of the pessimistic sentiments that the most appropriate textual information can be found from macroeconomic news. The textual information from stock market news, commodities news, or company specific news reflected the past event in the stock market, and, therefore, we didn't use these articles for our analysis.
To define whether the number of macroeconomic news is a good measure of the investor sentiments, we examine the historical dependencies of different variables for conformance with our assumptions about the www.ccsenet.org/ibr Vol. 7, No. 5;2014 relationship of negative news with market indices. Our empirical analysis showed that the ratio of pessimistic to optimistic macroeconomic news is the most sensitive to the negative stock price declines during stock market crisis. This ratio is called a pessimistic rate, Equation (1): (1) where P k is the pessimistic rate (PR). Figure 1 summarizes the calculated pessimistic rates for a period from 1/1/2007 to 1/1/2014. As can be seen, the pessimistic rates demonstrate a positive trend with jumps during bearish market period (2007)(2008).
Discussion
It is interesting to observe how negative information was issued during the financial crisis. The sharp changes of PR are characterized by the sentiment overreaction to the market through negative news. The pessimistic rates exhibit the considerable correlation with market index movement during the shocking market conditions. Figures 2 and 3 show the comparison of the stock index (S&P-TSX Composite index) and VIX index with the pessimistic rates, respectively. As can be seen, the density of negative news reaches maximum during the shocking market conditions. The Pearson correlation (6-month period) of the pessimistic rate with weekly return (S&P-TSX) reached 60% and statistically significant during the market crashes, Figure 3. The PR values differentiate the period of a stable market and shocking market conditions. Moreover, for the high volatile market, the PR values deviate considerably. Our result supports the study of Zhang et al. (2011), who showed that the emotional textual information from Twitter significantly negatively correlated with Dow Jones, NASDAQ and S&P 500, but displayed significant positive correlation to VIX. The similar sentiment indicator, which can be constructed from the pessimistic rate, is calculated from the difference of optimistic and pessimistic news to the total number of macroeconomic news. The proposed indicator is called a Panic Indicator (PI), Equation (2): (2) The panic indicator can be expressed through the pessimistic rate: (3) The range of the Panic Indicator is-1 to +1, i.e., the minimum value (-1) corresponds to the extreme case, when all published macroeconomic articles are negative; neutral market corresponds to the PI values close to zero, bullish market conditions correspond to the high positive PI. Since the pessimistic rates change from the large range from 0.5 to 3.14, see the historical distribution of PR, Figure 4, we consider that the panic indicator is more suitable for the quantitative analysis of stock price return.
The historical distribution of the panic indicator is shown in Figure 5. It is interesting to observe that the distribution of the Panic Indicator exhibits two picks near the neutral PI and negative skewness. Comparison of PI values with S&PTSX index shows that if PI values is lower than (-0.2), the market reaches the bearish market conditions. Vol. 7, No. 5;2014 The proposed panic indicator is very straightforward, using the total number of positive and negative news, we may classify the bearish or bullish market. Moreover, the extreme events are clearly observed in both in analytical news and market indices, since the significant investor pessimism can be observed from business news, which is clearly confirmed by the current report. Although we may construct many different sentiment indicators from business news, we consider that the proposed sentiment indicator can be easily explained and replicated for the stock price forecast.
Obviously, the proposed indicator can be easily applied for the regression analysis of the stock price return. The application of the panic indicator could work well for the various herding strategies, when the traditional macroeconomic and fundamental factors don't have predictive power. For example, from the observed correlation between sentiment indicator and S&P-TSX index returns (Figure 3) during financial crisis, we can apply a simple regression formula for the analysis of the stock price forecast: The results of the sentiment measurements demonstrated a good performance of the panic indicator during the financial crisis. The proposed panic indicator would be useful for stock price modeling, description of the pessimistic opinions, and computational trading algorithms. The proposed sentiment indicator (Panic Indicator) has following advantages.
(1) The application of the Panic Indicator can provide alternative information for the analysis of the bullish or bearish market conditions. (2) The Panic Indicator can be an additional source of the market and liquidity risk, when the stock price decline is controlled by the negative information flows. (3) The proposed indicator exhibits the predictive power for the extreme market conditions. | 2018-12-18T01:50:08.728Z | 2014-07-21T00:00:00.000 | {
"year": 2014,
"sha1": "50e8d4c7c2b794566249e4ce1f6f03eb5be22d93",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.5539/ibr.v7n5p103",
"oa_status": "GOLD",
"pdf_src": "ElsevierPush",
"pdf_hash": "c2c2df3886bf29eaa8d7c37ece6f598bd253f2e8",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Economics"
]
} |
201211447 | pes2o/s2orc | v3-fos-license | Bifidobacterium Is Enriched in Gut Microbiome of Kashmiri Women with Polycystic Ovary Syndrome
Polycystic ovary syndrome (PCOS) is a very common endocrine condition in women in India. Gut microbiome alterations were shown to be involved in PCOS, yet it is remarkably understudied in Indian women who have a higher incidence of PCOS as compared to other ethnic populations. During the regional PCOS screening program among young women, we recruited 19 drug naive women with PCOS and 20 control women at the Sher-i-Kashmir Institute of Medical Sciences, Kashmir, North India. We profiled the gut microbiome in faecal samples by 16S rRNA sequencing and included 40/58 operational taxonomic units (OTUs) detected in at least 1/3 of the subjects with relative abundance (RA) ≥ 0.1%. We compared the RAs at a family/genus level in PCOS/non-PCOS groups and their correlation with 33 metabolic and hormonal factors, and corrected for multiple testing, while taking the variation in day of menstrual cycle at sample collection, age and BMI into account. Five genera were significantly enriched in PCOS cases: Sarcina, Megasphaera, and previously reported for PCOS Bifidobacterium, Collinsella and Paraprevotella confirmed by different statistical models. At the family level, the relative abundance of Bifidobacteriaceae was enriched, whereas Peptococcaceae was decreased among cases. We observed increased relative abundance of Collinsella and Paraprevotella with higher fasting blood glucose levels, and Paraprevotella and Alkalibacterium with larger hip, waist circumference, weight, and Peptococcaceae with lower prolactin levels. We also detected a novel association between Eubacterium and follicle-stimulating hormone levels and between Bifidobacterium and alkaline phosphatase, independently of the BMI of the participants. Our report supports that there is a relationship between gut microbiome composition and PCOS with links to specific reproductive health metabolic and hormonal predictors in Indian women.
Introduction
PCOS is a common endocrine condition affecting women of reproductive age, characterized by hyperandrogenism, oligo-or amenorrhea, and polycystic ovaries on transabdominal ultrasonography. The worldwide prevalence of PCOS among women of fertile age is about 6-18%, and it varies with the use of different diagnostic criteria, such as Rotterdam criteria, 2003, National Institutes of Health (NIH), and the Androgen Excess Society (AES) criteria, 2006 [1][2][3]. In India, PCOS affects the life of an estimated 6.5 to 19.4 million women of reproductive age, and the prevalence is increasing in parallel with the obesity epidemic [4], while the studies addressing the importance of gut dysbiosis/health in this index population are completely lacking. Indian women are reported to have a high prevalence of PCOS [5] and Indian women with PCOS have higher fasting insulin levels and greater insulin resistance compared with white women with PCOS [6]. Moreover, the prevalence of PCOS is even higher among Kashmiri women and is among the highest in a published series globally [7]. In this regard, at many places in India including Kashmir, screening studies are being carried out on adolescent girls and young women for early diagnosis of PCOS and treatment [8][9][10].
Initially defined as a gynecological disorder [11], PCOS is associated with a constellation of metabolic conditions, such as obesity, dyslipidaemia, metabolic syndrome, endothelial dysfunction, inflammation, insulin resistance, hypertension and other cardiovascular risks [12][13][14][15][16][17][18] in addition to infertility, pregnancy complications and depression [19][20][21]. The development of PCOS is multifaceted and involves genetic [22,23], gestational, environment [24] and lifestyle aspects [25]. The precise factors responsible for these key biochemical and metabolic derangements affecting external environment and internal ecosystems, such as gut microbiota, remain largely unexplored. From a genetic point of view, the human gut microbiome is a collection of microbial genomes of microorganisms that inhabit the human gut [26]. The human gut microbiome is a complex ecosystem harbouring numerous microbes taking part in essential functions of the host organism [27]. Earlier case-control studies in Europeans [28] and Chinese [29] showed differences in diversity and relative abundance of particular taxa and their correlation to hormonal levels and body mass index (BMI). The relative abundances of Bacteroides, Escherichia/Shigella and Streptococcus were inversely correlated with ghrelin, and positively correlated with testosterone and BMI in Chinese women with PCOS. Overall, relative abundances of Akkermansia and Ruminococcaceae are inversely correlated with body weight, sex-hormone, and brain-gut peptides, and are decreased in PCOS [29]. The gut microbiota alterations in humans are associated with obesity [30,31], which may be the driving factor for some of the links with PCOS. Additionally, the peripheral insulin sensitivity in metabolic syndrome subjects improves upon transfer of stool from healthy donors, suggesting a causal effect of gut microbiome on metabolism [32]. Moreover, gut microbiota and its metabolites can control inflammatory processes, brain gut peptide secretion as well as islet b-cell proliferation. Hence, changes in microbiota composition may lead to excessive accumulation of fat ultimately causing insulin resistance and compensatory hyperinsulinemia [32,33]. Gut microbiota also associate with hormonal outcomes. In a female mouse model, the testosterone levels increased upon infecting it with male faecal microbiota compared to unmanipulated females [34]. A host's estrous cycles, sex hormones and morphological changes in ovaries are affected by gut microbial composition of the host [35]. In rats, the pre-natal exposure to high androgen levels in daughters from mothers with PCOS led to dysbiosis in gut microbiome and impairment of the cardiometabolic functions [36]. The association between gut microbiome, obesity and host genetics suggests that certain bacteria predisposing to a healthy or unhealthy metabolic state may be partially heritable [37][38][39]. It is also hypothesized that the diet might induce bacterial dysbiosis leading to inflammation, insulin resistance and hyperandrogenemia in PCOS [40]. Gut microbiota also play a prominent role in thyroid dysfunction of the host [41], which is a comorbid condition to PCOS [42]. As PCOS is a complex multi-faceted disease involving multiple systems of a woman's body and it also has a strong environmental component, we questioned to what extent the gut microbiome is involved and whether such involvement can be explained by, e.g., obesity, hormonal levels, and insulin resistance.
We profiled the gut microbiome of young women with PCOS and healthy controls from Kashmir, Northern India, in order to address this understudied population with high PCOS incidence. We used 16S rRNA profiling and further tested whether the gut microbiome composition associates with blood biochemistry and hormonal levels.
Study Cohort and Recruitment
Twenty women with PCOS (drug naive) and 20 control women without PCOS, both groups in age ranging 16-25 years, were identified during regional PCOS screening of college and school students in different schools and colleges in Kashmir, North India, from January to May 2017. Those women with regular cyclicity in their menstrual cycles (21-35 days), no signs of hyperandrogenism and that had normal ovarian morphology as evidenced through transabdominal ultrasonography were invited to take part as controls. The case group consisted of participants with menstrual disturbances including oligomenorrhea (menstrual interval > 35 days or <8 cycles/year) or amenorrhea (no menstrual cycle in last >6 months), hyperandrogenism (male pattern hair growth, androgenic alopecia), polycystic ovaries on transabdominal ultrasonography and qualified Rotterdam criteria, 2003 [1] for PCOS diagnosis. According to the Rotterdam criteria 2003, a clinical diagnosis of PCOS requires that a patient presents with two of the following symptoms: (i) oligoovulation or anovulation, (ii) Hyperandrogenism-(including signs such as hirsutism), and (iii) polycystic ovaries visible on ultrasound. Hirsutism was assessed by Ferriman-Gallwey [43]. In total, nine body parts, including the upper lip, chin, chest, upper and lower abdomen, thighs, upper and lower back, and upper arms, were examined for the presence of terminal hair. Terminal hairs at these sites were graded in the range of 0-4. The Ferriman-Gallwey score was determined as the cumulative score of all the sites. Hirsutism was defined by the presence of a modified Ferriman-Gallwey score of 8 or higher.
Initially, among the individuals who gave informed consent during the interview, 32 were selected as non-PCOS and 53 selected as PCOS and were invited to donate blood and stool samples on days 3-7 (early follicular phase) of spontaneous cycle to maximise hormonal homogeneity. Tests for follicle stimulating hormone (FSH), luteinising hormone (LH) and testosterone by standard are best performed during the follicular phase (3 to 7 days of the menstrual cycle.) The blood and stool donation took place within the same day at the Endocrine clinic of the Sher-i-Kashmir Institute of Medical Sciences (SKIMS, a tertiary care hospital). Eventually, 23 individuals without and 20 individuals with PCOS donated both blood and stool samples. Three samples from the control group were excluded later as they did not fit in the required follicular phase.
Women on antibiotic treatment or those taking contraceptives, steroids, anti-epileptics, insulin sensitizers, proton-pump inhibitors, non-steroid anti-inflammatory drugs, or with any previous history of systemic sickness such as diabetes mellitus, coronary artery disease, non-classical congenital adrenal hyperplasia (NCAH), Cushing syndrome, hyperprolactinemia, thyroid dysfunction, gastrointestinal disease and appendectomy were excluded from the study. All participants were non-vegetarian. One person from the PCOS group was later excluded due to low sequencing read depth. The study protocol was approved by the Institutional Ethics Committee (IEC) of SKIMS Kashmir and written informed consent was obtained from all the participants involved in this study.
Blood Biochemistry, Hormonal Tests and Anthropometry
Fasting blood samples were collected in dry and EDTA coated tubes after 10-12 h fasting. Sera were separated to be used in measuring glucose, lipids, alkaline phosphatase, aspartate aminotransferase, albumin and creatinine, in addition to hormones (prolactin, thyroid stimulating hormone (TSH), thyroxine (T4), triiodothyronine (T3), cortisol, luteinising hormone (LH), follicle stimulating hormone (FSH), and total testosterone. The EDTA-containing aliquot was immediately placed on ice and centrifuged within 30 min; plasma was collected and was stored at −80 • C for further analysis. Alkaline phosphatase (ALP), aspartate aminotransferase (AST), creatinine, albumin and lipid measurements were performed using a fully automated chemistry analyser (Hitachi 912 at SKIMS). Estimation of serum T3, T4, cortisol, TSH, prolactin, LH, FSH and total testosterone was done by Radioimmunoassay (RIA) using commercial kits and according to the supplier protocol at SKIMS. Plasma glucose was measured by the glucose oxidase peroxidase method. Body weight (light clothing without shoes), height and waist circumference were measured with standard calibrated instruments (SECA 213, Hamburg, Germany), followed by a detailed systemic examination that included blood pressure measurement (Omron HEM7120).
Stool Samples
Each participant was asked to provide fresh stool sample (approx. 5 g) at the same day with blood collection, by using the stool collection and stabilization kit (OMNIgene ® •GUT OMR-200, DNA Genotek, Canada). All samples were collected in the morning and were stored at ambient temperature until further processing which took place within the next 60 days. Consistency of each sample was recorded on the following scale: 1 = Sausage shaped with cracks on the surface, 2 = Sausage shaped and smooth soft stool, 3 = Solid stool with clumps, 4 = Watery stool.
Metagenomic DNA Extraction and 16S rRNA Sequencing Data Generation
DNA extraction was performed by as per the instruction manual using ZymoBIOMICS TM DNA kit by Zymo Research USA. The DNA concentrations were estimated by Qubit Fluorometer (Thermo Fisher) and checked by Agilent TapeStation 2200. The microbiota characterization was performed by targeting the hypervariable regions V3-V4 of 16S rRNA gene using a paired-end approach using the specific primers published earlier [44] and according to the manufacturer's instructions [45]. The amplified regions were combined with dual-index barcodes, Nextera ® XT Index Kit v2 Set A, B and C, Illumina (Illumina Inc., San Diego, CA, USA) The sequencing run was performed with MiSeq 600 cycle Reagent Kit v3, Illumina USA and sequenced on a MiSeq Illumina system by the Stremble Ventures LTD, Cyprus.
Bioinformatic Analysis and Quality Control
The bioinformatic analysis including read QC and operational taxonomic unit (OTU) classification was performed from FASTQ files with paired-end reads. Specifically, the Basespace platform and the 16S Metagenomics from Basespace (Illumina recommended software) were used to deal with quality filtering and denoising of the raw reads. This also classified OTUs using Ribosomal Database Project Classifier [46] against the Illuminacurated version of the GreenGenes reference taxonomy database [46,47]. The number of reads classified at the genus and family level are given in Table 1. Resulting Krona plots of all taxonomic entities detected are provided in Figure S1A,B. Among those, genera and families that were detected in more than 1/3 of the samples were considered prevalent OTUs and, among those, the ones with >0.1% average relative abundance across all subjects were included in analysis ( Figure S2A,B). Table 1. Characteristics of the study sample.
Mean (SD) or n [%] Mean (SD) or n [%]
Anthropometrics and blood measurements We performed all the analyses at Family and Genus levels for which the proportion of successfully classified reads across samples was over 95%. We used the non-parametric Mann-Whitney U-test [48] to compare median abundancies of each species between PCOS cases and controls. For these analyses, we considered results with a p-value < 0.05 as statis-tically significant. To account for potential confounding factors, we used the Multivariate Association with Linear Model, MaAsLin, in R [49], which allows for the use of covariates in the model. In brief, MaAsLin fits a linear model for each species and the variable of interest, here PCOS, after arcsin-square-root transformation of the proportional values of the species. This transformation has been shown to stabilise variance and normalise proportional data well [49]. The method can also include a boosting step to select factors among a large set of variables to be associated with the species. We turned this feature off since we focused only on one variable of interest, PCOS. We used the default settings of the programme, i.e., minimum relative abundance of 0.01%; minimum percentage of samples in which a feature must have the minimum relative abundance in order not to be removed of 10%; outlier removal by a Grubbs test with the significance cut-off used to indicate an outlier at 0.05; multiple test correction by the Benjamini-Hochberg (BH) [50] method; and the threshold to use for significance for the generated q-values (BH FDR) of 0.25. As potential confounders in the model, we used stool consistency, read depth, age, day of the menstrual cycle during sample collection and BMI. The linear model analysis with MaAsLin was performed also for LH, FSH, and testosterone levels in cases and controls together, adjusting for stool consistency, read depth, age, day of the menstrual cycle during sample collection and BMI. For bacteria abundances that were different between cases and controls, we performed follow-up analyses with a range of quantitative traits measured from the participants, such as blood glucose and lipid levels (Table S1), to dissect the underlying mechanistic links between the identified OTUs and PCOS. We used linear regression by MaAsLin and the same adjustments as before. The analyses with hormonal variables and other quantitative traits for those family and genera that showed a statistically significant association were also performed separately in PCOS cases and controls.
Microbial Distance and α Diversity
Inter-individual microbial distance (Bray-Curtis dissimilarity) and α diversity (Shannon's diversity index) were calculated using functions vegdist and diversity of the R package vegan [51]. We analysed the association between PCOS and (i) Bray-Curtis dissimilarity and (ii) Shannon's diversity index using Spearman correlation in R. Both analyses were additionally adjusted for stool consistency, day of menstrual cycle at sample collection and sequencing read depth.
16 SrRNA Gut Microbiome in PCOS
In this study, we initially profiled the microbiome in DNA isolated from faecal samples by 16S rRNA sequencing in 20/20 women with/without PCOS from Kashmir, India. After the quality control of 16S rRNA sequencing data, and removing one person from the case group due to low number of reads, we analysed the 39 microbiomes of Kashmiri women with/without PCOS (Tables 1 and S1). The Krona plots ( Figure S1A,B) and the bar plots, (Figures 1 and S2) show that the core phyla of gut microbiomes are mainly dominated by Firmicutes and Bacteroidetes followed by Proteobacteria in controls ( Figure S1A) and Actinobacteria in patients ( Figure S1B). We defined 40 OTUs at the family level and 58 at the genus level ( Figure S2A,B). These were detected in at least 30% of the samples with relative abundance > 0.1%.
Associations with PCOS at the Genus and Family Level
The analyses of the taxa at genus level showed statistically significant (p < 0.05) differences when compared between individuals with and without PCOS for 12 genera either by the Mann-Whitney U-test or by linear modelling with MaAsLin ( Table 2). Ten of these bacterial genera reached false discovery rate (FDR) corrected statistical significance (Q < 0.25) in the linear modelling with arcsin-square root transformation and outlier exclusions ( Table 2). The most striking differences of relative abundance were observed for Sarcina (cases vs. controls: 0.28% vs. 0.06%), Megasphaera (3.62% vs. 1.17%) and Bifidobacterium (7.7% vs. 3.1%) (Figure 2). After adjustment by stool consistency, day of menstrual cycle at data collection point, sequencing read depth, age and BMI (Model2) the relative abundance of Sarcina, Megasphaera, Collinsella, Paraprevotella, and Bifidobacterium associated with PCOS, consistently in all three statistical models performed (Tables 2 and S2, Figure 2). It is worth noting that, although Alkalibacterium associated with case/control status in all of the three models, the direction of association by the Mann-Whitney U-test was the opposite of what was estimated by MaAslin. Therefore, we considered it as a false positive.
At the family level, increased relative abundance of Bifidobacteriaceae associated with case status (7.63% vs. 3.07%), whereas relative abundance of Aerococcaceae (0.08% vs. 0.31%) and Peptococcaceae (0.15% vs. 0.27%) were decreased among individuals with PCOS (Figures 2, S1 and S2) when analysed by our Model 2, taking the BMI differences into account. However, only Peptococcaceae and Bifidobacteriaceae showed consistency between the Mann-Whitney U-test and all three MaAsLin models and were considered further. Figure 3 shows the Shannon diversity index between case and control groups. Table 3. Statistically significant differences between PCOS cases and controls in the relative abundancies at the family level.
Hormonal Profiles at the Genus and Family Level
Upon linear modelling of the FSH, LH, LH to FSH and testosterone hormonal profiles for all study participants, Eubacterium reached the FDR-corrected significance level for association with FSH (coeff = 0.15, p = 0.0003, Q = 0.023, Table S8, Figure 4a). This association remained significant after adjustment for the covariates (Table S8). We further checked the association separately in cases and controls and observed statistically significant associations in both groups, in the same direction (Cases: coeff = 0.016, p = 0.022; Controls: coeff = 0.017, p = 0.017). For the other hormones, we did not observe statistically significant associations for any bacterial species (Tables S5-S7). Linear regression analysis of hormonal profiles at family level relative abundancies did not yield any statistically significant associations (Tables S8-S11).
Follow-Up of the Identified OTUs at the Genus and Family Level
The study participants were assessed for a number of quantitative traits (Tables 1 and S1) enabling us to test the supporting hypothesis of involvement of gut microbiome with continuous biochemical and hormonal markers of PCOS constellation. At genus level (Table 2), the quantitative trait analyses showed an association between higher Bifidobacterium and higher alkaline-phosphatase (ALP) level, which remained significantly associated after further adjustment by BMI (coeff = 0.0011, p = 0.00019, Q = 0.214), (Table S12, Figure 3b). We were statistically underpowered to detect any associations separately in cases and controls. Among the remaining genera from Table 2 Table S13). None of these associations were significant after BMI adjustment (Table S12), except the one between Bifidobacterium and ALP.
Discussion
This is the first investigation of PCOS gut microbiome in women from Kashmir, India where incidence of PCOS is high and comorbid with greater insulin resistance. Here, we compared 39 individuals with/without PCOS by their gut microbiome composition and dissected the latter in relation to 33 quantitative endophenotypes. We performed a step-wise association modelling moving from simplistic towards more complex models with covariates. We identified robust signals both at genus and family level taxonomy. Seven taxa were significantly different in PCOS cases, including enrichment at both genus and family levels, i.e., relative abundance of Bifidobacterium and Bifidobacteriaceae. When considering all samples together, we observed the domination by Firmicutes followed by Bacteroidetes, Actinobacteria and Proteobacteria, the same enterotype reported by Das et al. [52] for samples collected from two distinct regions of India, including Leh, located in proximity of Kashmir. Overall, the gut microbiome of women with PCOS had a higher bacterial diversity compared to that of women without PCOS, when measured by Shannon's index. This is a unique finding as previous studies showed decreased diversity among individuals with PCOS as compared to without [53,54] and our observation may be specific to the Kashmiri population. At the genus level, the relative abundances of five genera including Lactobacillus, Bifidobacterium, Sarcina, Megasphaera, Collinsella and Paraprevotella, were increased in gut microbiome of women with PCOS. Association of Lactobacillus seems to be dependent on BMI as tested by our model 2, whereas the other genera remained significantly associated after adjustment for BMI. Among them, Bifidobacterium also associated with higher ALP, independent of BMI. We also show that Collinsella and Paraprevotella are associated with higher fasting blood glucose levels. Our study is the first to report Sarcina and Megasphaera in relation to PCOS. At the hormonal level analysis of all taxa, we show that increased Eubacterium is associated with increased FSH.
At the family level, the gut microbiomes of women with PCOS are enriched with Bifidobacteriaceae, whereas they harbour lower relative abundance of the Peptococcaceae family when the technical covariates are accounted for in statistical models. The cases had higher mean BMI; however, there was no difference in WHR between two study groups, thus not suggesting higher adiposity as the driver for anthropometric differences (Table 1). We additionally adjusted our tests for variation in BMI, so the associations we report are not likely to be confounded by the BMI of the participants, unless reported otherwise. The individuals with PCOS also had higher blood glucose, prolactin and TSH levels (Tables 1 and S1) consistent with earlier reports [55]. It is of note that both fasting glucose and prolactin were increased among our case group and at the same time we saw linear associations between decreased abundance of family Peptococcaceae and lower prolactin, and increased abundance of genera Collinsela and Paraprevotella and fasting glucose, when looking at the whole population. However, these were likely to be driven by BMI. Previous studies have indicated that BMI has an important role in the diversity of the gut microbiome, and that having a higher BMI is strongly linked to gut dysbiosis [56][57][58].
To date, a few observational studies on the gut microbiome of individuals with PCOS exist [26,29,35,54,59]. It has been reported that the probiotic Bifidobacterium lactis V9 may increase sex hormone levels in PCOS patients [59]; however, we cannot test this at the cross-sectional design as we had limited the taxonomic resolution to genus and family levels. Relative Bifidobacterium abundance in human gut is known to be driven by lactose intolerance. In lactose-intolerant individuals, lactose is not metabolized in the small intestine and proceeds to the colon where it is fermented by members of the gut microbiome. This fermentation leads to gas production, a major symptom associated with lactose intol-erance [60]. Thus, genetic variants that reduce lactase activity can promote the growth of lactose-fermenting bacteria in the colon, but only if the individual consumes dairy products. Taken together with our findings, there could be a relationship between the adult type hypolactasia and PCOS. There has been only one small study looking at such an association; however, it has not been replicated [61]. The Bifidobacterium abundance in gut is known to associate with favorable metabolic outcomes, but a recent report showed that not all strains of Bifidobacterium are functional [62]. These strain level differences can only be resolved by metagenomic sequencing, which requires further investigation. We earlier reported that rs182549 (human LCT gene)*C allele, predisposing to lactose intolerance, and thus to increased Bifidobacterium in gut, was negatively associated with height and obesity and positively associated with several nutritional phenotypes, type 2 diabetes (T2DM) risk and family history of T2DM in the UK Biobank cohort [39]. Moreover, the functional LCT SNP rs4988235 variant was in strong linkage disequilibrium associated with 1,5-anhydroglucitol (p = 4.23 × 10 −28 ), which is an indicator of glycemic variability [39]. In our study, Bifidobacterium also associated with ALP in the overall sample; however, the ALP values from individuals with PCOS were higher than of control group, 250.7 vs. 210 mg/dL (p = 0.016, by t-test). ALP is a main indicator of liver function; however, it is also secreted in the intestine where it plays a vital role in maintaining gut homeostasis [63]. Thus, the link between ALP and Bifidobacterium could be mediated through liver function, or could be attributed to direct influence of colonic ALP on the gut homeostasis. It is of note that increased prevalence of non-alcoholic fatty liver disease (NAFLD) and hepatic steatosis were reported in patients with PCOS [64][65][66]. It is hypothesised that alcohol producing bacteria such as Bifidobacterium may contribute to the pathogenesis of NAFLD in PCOS [67,68]. Another possible link could be via primary bile acids like glyco-and tauro-conjugated which were found to be elevated in women with PCOS compared to controls and were positively associated with HA [69]. Bile acid metabolism begins in the gastrointestinal tract involving microbiota possessing bile salt hydrolase activity [70], which is common in Bifidobacterium and Lactobacillus [71]. Bifidobacterium is an important genus for host health, associated with favorable outcomes, especially in the early years of life. Taken together with shared genetics and association to glycemic traits, as well as liver functioning, the association of Bifidobacterium may have multiple (age-dependent) facades, including unfavorable effects.
Collinsella and Paraprevotella, which we observed to be associated with fasting glucose, were earlier shown to be enriched in the gut of obese women without PCOS, as compared to non-obese PCOS cases, indicating that these genera could be more specific to obesity than the PCOS cluster of phenotypes [29]. On the other hand, the fact that many PCOS patients also have hyperandrogenism may explain the relationship between PCOS and Paraprevotella. The genus has been detected in prenatal androgenized rat models, where Paraprevotella was significantly enriched in androgenized rodents [72], thus suggesting a mechanistic link between androgens and Paraprevotella.
However, as for Collinsella, we see a shared genetic link with genus in larger population based studies; in our recent report on genetics of random glucose in humans, we suggest that ABO/FUT2 genetic locus potentially orchestrates the correlation between Collinsella and the host glucose level [73]. Another piece of supporting evidence from a shared genetic background concerns family Peptococcaceae; rs7574352 associated with the family Peptococcaceae [39] is located in the intergenic region in the proximity (220kb apart) of IRF1, which is involved in insulin resistance and susceptibility to T2DM, also suggested as a risk loci for PCOS [74,75].
Contrary to our findings highlighting Lactobacillus enrichment in PCOS cases as driven by BMI, Lactobacillus was earlier found enriched in non-obese controls when compared to obese PCOS individuals [29], suggesting the involvement of BMI as a confounder or mediator of the association. A randomized controlled study focusing on metabolic benefits of synbiotics in PCOS reported that Lactobacillus intake resulted in lower insulin concentration and insulin resistance and higher insulin sensitivity [59], suggesting a causal role for Lactobacillus for PCOS accompanied with high BMI. The genera Megasphaera, enriched in PCOS gut microbiome samples, are also involved in lactic acid fermentation but have no established role in the PCOS human gut. Sarcina is a member of the family Clostridiaceae; the genus Sarcina ventriculi is increasingly common Gram-positive coccus recognized in gastric biopsies from patients with delayed gastric emptying [76], one of the clinical symptoms in PCOS.
At the hormonal level analysis of all taxa, we show that increased Eubacterium is associated with increased FSH; however, the association was similar both in individuals with and without PCOS and does not provide any insight for the main objective of our study and will not be discussed further.
One recent study on a Westernised (Finnish) population conducted by Lüll et al. [77] has observed statistically significant associations between diversity measures and PCOSrelated hormonal and metabolic parameters, such as BMI, sex-hormone binding globulin (SHBG) levels and insulin resistance. In the same study, bacterial diversity indices did not differ significantly between PCOS and controls unless they were stratified by prediabetes status; then, women with PCOS showed a lower Shannon diversity index. On the contrary to what has been shown by Lüll et al., we did not observe differences in Clostridiales, Ruminococcaceae UCG-002, and Clostridiales Family XIII AD3011 group and genus Dorea. We cannot exclude the fact that differences could well be due to the dietary and regional lifestyle differences, i.e., Finland vs. Kashmir.
The main limitation of our study is the relatively small sample size, however, comparable to the size of other pilot studies in the field. In order to deal with limited sample size, we used the non-parametric Mann-Whitney U-test as an exploratory analysis. Although this method does not allow to account for covariates, the associations with families Peptococcaceae, Bifidobacteriaceae and associations with genera Sarcina, Megasphaera, Bifidobacterium, Paraprevotella and Collinsella turned out to be robust and confirmed by the MaAsLin approach. In addition, our study might be affected by the lack of resolution, which can be compensated by using shotgun metagenomics sequencing in further research. Our study is focused on gut microbiome composition in PCOS, but we had no information on the saliva microbiome phylogenetic profile, which has lower diversity in PCOS and is altered in relation to clinical symptoms of the disease [78].
The microbiome research in PCOS is a fast-expanding field; however, many pilot studies published for PCOS gut microbiome composition require replication of reported association with specific taxa. The emerging field of faecal microbiota transplantation might become promising in PCOS management, given its first encouraging results in rats, which led to decreased androgen biosynthesis and normalisation of ovarian morphology [35].
In this study, we report an increased abundance of genera involved in lactic acid fermentation in the stool samples of individuals with PCOS. These bacterial genera highlight specific effects of gut microbiota beneficial for human metabolism while the links also point out a shared genetic background between hist metabolism and gut microbiota. However, the data available to us do not allow conclusive inferences about these microorganisms being a cause or consequence of metabolic disturbances in PCOS. Further studies are needed for evaluation, whether abundance in these bacterial genera is specific to Indian women. Our findings show importance of gut microbiota in maintenance of hormonal levels in PCOS and warrant further well-powered research.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13020379/s1, Figure S1A,B: Separate Krona plots for case and control groups. title, Figure S2: Bar plots depicting the relative abundancies of the OTUs at genus and family level for each individual. Table S1: Characteristics of the study sample. Table S2: Association results from MaAsLin between all the bacteria at family level and PCOS case-control status, Table S3: Association results from MaAsLin between all the bacteria at genus level and PCOS case-control status, Table S4: Association results from MaAsLin between all the bacteria at family level and follicle stimulating hormone (FSH) levels, Table S5: Association results from MaAsLin between all the bacteria at family level and lutenising hormone (LH) levels, Table S6: Association results from MaAsLin between all the bacteria at family level and lutenising hormone (LH) to follicle stimulating hormone (FSH) ratio (LH/FSH), Table S7: Association results from MaAsLin between all the bacteria at family level and testosterone (TESTO) levels, Table S8: Association results from MaAsLin between all the bacteria at genus level and follicle stimulating hormone (FSH) levels, Table S9: Association results from MaAsLin between all the bacteria at genus level and lutenising hormone (LH) to follicle stimulating hormone (FSH) ratio (LH/FSH), Table S10: Association results from MaAsLin between all the bacteria at genus level and luteinising hormone (LH) levels, Table S11: Association results from MaAsLin between all the bacteria at genus level and testosterone (TESTO) levels. Table S12: Association results from MaAsLin between PCOS associated bacteria at genus level and quantitative traits measured in the case control population, Table S13: Association results from MaAsLin between PCOS associated bacteria at family level and quantitative traits measured in the case control population. Data Availability Statement: The datasets generated during and/or analysed during the current study are not publicly available for reasons related to privacy and participant consent but are available from the corresponding author on reasonable request. Summary-statistics level results are included in the Supplementary Files attached to the manuscript. | 2019-08-23T06:03:56.924Z | 2019-07-30T00:00:00.000 | {
"year": 2022,
"sha1": "7514272b4727e3685f93d361b227f94eea76349a",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2073-4425/13/2/379/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "a2e921facdbe73abe85f28c09e06f5569b0dca55",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
218959726 | pes2o/s2orc | v3-fos-license | Christmas Allowance as a Non ‐ System Tool for Sustainability of Quality of Life of Slovak Seniors
: Creating conditions for a sustainable level of quality of life for older people is considered the dominant priority when setting up the pension system with an emphasis on the income situation of Slovak seniors for old ‐ age pensions. An old ‐ age pension as a systemic benefit is an important element in maintaining the quality of life of older people. The amount of old ‐ age pensions is currently at the center of discussions between institutions and the Government of the Slovak Republic. A major social but also economic problem in the coming decade will be how to maintain the income of older people, pensioners, needed to ensure their sustainable quality of life. In particular, it is necessary to ensure that pensioners receive sufficient income so that they do not become a population at risk of poverty as the groups at risk of poverty or social exclusion also includes people over the age of 65. The paper focuses on the analysis of the socio ‐ economic situation of Slovak pensioners in the Slovak Republic in connection with material deprivation. At the same time we analyze the Christmas contribution as a non ‐ systemic benefit which, on the one hand, has a positive social impact on the recipients of pension benefits who are entitled to the Christmas benefit but, on the other hand, has a negative impact on the general government budget.
Introduction
Research on quality of life in recent years is extensive in many countries, proving that it has a place in the social sciences [1][2][3]. It includes economic, social, and above all living conditions, as well as material security, political freedom, independence, and social justice [4][5][6][7][8][9][10][11][12]. Pigou (1924), as one of the first scientists in the context of discussions on economics and social well-being, mentioned the term ʺquality of lifeʺ [4]. Most of the quality of life research has been conducted in the United States and focused on exploring satisfaction, happiness, and well-being. The theoretical model of quality of life as ʺgood lifeʺ was created by Lawton [5]. Holková's research into quality of life was based on the theory of consumption. Among other things it focused on examining the relationship of savings and spending over the life cycle with a view to maintaining an adequate standard of living during retirement [1]. The increase in the number of older people is an urgent problem in todayʹs society, as it also entails increased costs for the health system and social security. Politicians in the Slovak Republic also target retirement households as potential voters and are therefore willing to support non-systemic measures. The Government of the Slovak Republic has already taken this path in 2006 and in its program statement it has committed itself to provide Slovak seniors with a Christmas contribution differentiated according to their pension according to the financial possibilities of the state budget. Already in 2006, the Government of the Slovak Republic undertook in its Program Statement to provide Slovak seniors with a Christmas allowance differentiated according to the amount of their pension and according to the financial possibilities of the state budget. The aim of the government was to improve the social situation of recipients of old-age and disability pensions. Our goal is to analyze in detail the legislative and technical process of paying the Christmas allowance in the Slovak social system and at the same time to point out its non-systematic nature.
Theoretical Background
The high rates of economic growth, the development of new technologies, and the growth of industrial production have led to quality of life becoming associated with the negative impacts of economic development on the environment. The issue of quality of life gradually came into the environment as a broadly conceived approach, oriented not only toward the state of the environment but also on the overall state of society. In this way, critics of unlimited economic growth have expressed concern about the future conditions for human life. In the 1960s, the Rome Club linked the issue of quality of life to the global problems of a deteriorating environment. Galbraith [13], in the context of increasing consumption, used the term quality of life in response to mass consumption. Rostow [14] examined the phases of economic growth and assumed that after a massive mass consumption stage, society would move to the search for a new quality of life, but this was not fulfilled.
In the 1960s and 1970s, the concept of quality of life also began to be used in sociology to highlight the need to seek new ways of assessing quality of life by using indicators that take into account the social context and assess not only objective but also subjective perceptions of oneʹs own quality of individual life [15]. At that time, sociology introduced the concept of social indicators describing the objective conditions of life of society.
Psychology has been gradually introduced into the study of quality of life. Psychology is interested in the overall subjective evaluation of life from a position of an individual, analyzing factors influencing their subjective feeling. It monitors the quality of life through cognitive and emotional dimensions. The cognitive dimension consists of conscious and rational evaluation of life while the emotional dimension affects emotional experience. Thus, quality of life began to be monitored in the context of well-being, which is a subjective approach to quality of life, i.e., subjective evaluation. Discussions on quality of life have been conducted both in the UK and in the US where the first research was conducted. Scientists focused mainly on the content and definition of this term. Many contradictory opinions and concepts have emerged. Most of them were futuristic, focused on models of future perspectives of way of life. The trend of the 1970s and 1980s became the so-called post-materialism, which emphasized the value of quality of life, where priority was given to civil liberties, personality development, leisure time development, and tourism.
In the 1990s, research into quality of life further expanded with a view to a thorough definition of the theoretical foundations of quality of life and the development of indicators that would allow quantifying the quality of life in its complexity. Discussions on defining the importance of quality of life in the context of sustainable development have been renewed. The basis of this theory is quality of life is determined primarily by the highest number of possible ways of satisfying the individualʹs desires, achieving maximum yield and usefulness [16]. Laluha understands quality of life as a socioeconomic category that is tied to the satisfaction of needs, and quality of life depends on the height, nature, and structure of the needs being met [17]. According to Lehman, quality of life is a subjective factor manifested by a feeling of well-being depending on personal characteristics (age and sex), objective indicators (e.g., level of financial income), and subjective indicators (e.g., satisfaction with financial income) [18]. An important part of quality of life is from an economic point of view: the standard of living which is measured by means of significant economic indicators such as income, expenditure, consumption, and household equipment. Income comes from economic activity during an active working life. In retired families, these are savings, but especially old-age pensions. In addressing the quality of life of Slovak pensioners, we focused on the phenomenon of standard of living defined in by the Economic Encyclopedia [19] as ʺthe degree of satisfaction of peopleʹs life needs and a summary of conditions arising from social relationships in which these needs are satisfiedʺ. Standard of living is used as an analytical tool to monitor the living conditions of the population to study its socio-economic status and to measure the satisfaction of life needs. It is necessary to realize that the practical assurance of quality of life is a social and political problem and, therefore, the low level of pension benefits is subsidized by the Government of the Slovak Republic, mainly from the means of general taxation in the form of a Christmas allowance to pensions. It follows that quality of life does not have a uniform definition and research focuses on several areas of life, housing conditions, living needs, environmental conditions, and rest and recreation conditions, as well as social comfort. The quality of research into retirement age is an important area. Assessing the quality of life of older people is increasingly important not only because of the increasing number of older adults but also because of the increase in life expectancy and technological advances in medicine [20].
We are of the opinion that the standard of living plays an important role in assessing the quality of life, as well as the practical assurance of the quality of life is above all a social and political problem. Standard of living was used from the outset as an analytical tool for monitoring the living conditions of the population. The indicators of the living standard describe basic aspects of poverty and social exclusion and allowed the main risk areas and their impact on poverty and social exclusion to be identified, which we will discuss in the next section.
The cost of retirement households, covering their living needs as well as the cost of state social security systems, are increasing. In the next part of the research we focused on the quality of life of retired households in the context of retirement incomes. Adequacy of pensions is closely linked to the well-being of people and the social sustainability of the pension system [21].
Socio-Economic Situation of Slovak Pensionersʹ Households
Solidarity can be defined as the voluntary renunciation of part of oneʹs own surplus or advantage in favor of one who is deficient or disadvantaged [22,23].
According to Keller [24], the functioning of society would not have been possible without some degree of solidarity; the author even argues that solidarity is a prerequisite for the very existence of society. Solidarity between different population groups (especially between the rich and the poor) is a manifestation of the social model applied in each country. At present, help through charitable organizations is not enough but the solidarity of the rich and the poor must also be reflected in the social model as one of the foundations of the construction of society. The welfare state is a unique phenomenon in each country as it reflects certain specific historical, cultural, economic, political, and other conditions. The welfare state has grown up as a means of economic and political stabilization and job creation. As the Danish social theorist G. Esping-Andersen [25] writes, it is not possible to imagine a welfare state that does not, in one way or another, distribute income and resources. Today, however, it is highly costly and confronted with changes such as aging and is, therefore, under pressure to be rebuilt. The welfare state in Slovakia has profiled itself as a system orientated by its social policy primarily to dampen social impacts on the living conditions of people and to fulfill the protective function. In building its system, Slovakia followed the Bismarck system and built social protection mainly on the insurance model.
The welfare state in Slovakia has become a frequently used and debated term in recent years, especially in connection with the direction of reforms, also the left-wing of governmental power which sought at least partial decomodification by strengthening solidarity and introducing various social benefits where we clearly include the Christmas contribution. The welfare state is not only a professional but political term, too.
The welfare state focuses on social policy areas aimed at responding to the social risks of a citizen (e.g., old age) or to the possible negative consequences a citizen faces (e.g., disability). At the same time, its aim is to eliminate the social consequences that may result in the poverty of vulnerable groups [26].
The concept of social exclusion is closely linked to the concept of poverty. Both of these concepts are closely related yet, even if they partially overlap, they are not identical phenomena. They differ in the causes and individual dimensions, as well as how they are manifested [27].
The problem of poverty and social exclusion has not been the focus of EU member states for quite a long time. It was not until the 1970s and 1980s that, in times of economic crisis, joint, coordinated action on combating poverty appeared. The European Commission (EC) considers poor people to be families and groups of people whose material, cultural, and social resources are so limited that they exclude them from the minimum accepted standard of living of the member state in which they live. In the Slovak Republic, the term "poverty" was defined in the National Action Plan of Social Inclusion 2004-2006 of the Slovak Republic and was based on the approach of the European Commission [28].
The share of people at risk of poverty or social exclusion in the Slovak Republic decreased from 2005 to 2009 (from 32.0% to 19.6%). The economic crisis saw a slight increase (to 20.6% in 2010 and 2011). Since 2012, indicators in Slovakia have been improving. Since 2014, the trend of decreasing the percentage of people at risk of poverty or social exclusion has continued. In 2017, 16.3% of the population in Slovakia was at risk of poverty or social exclusion, which is below the EU average (22.5%) [29,30].
Since 2008, 255,000 people have been lifted out of poverty or social exclusion, thus fulfilling Slovakiaʹs Europe 2020 target to reduce the number of people at risk by at least 170,000 by 2020 ( Figure 1). However, there are population groups that are more at risk, not only in Slovakia but in other EU countries, as well. These are mainly children from socially disadvantaged backgrounds and the Roma minority. Children were three times more at risk of poverty compared to those aged 65 and over. Groups at risk of poverty or social exclusion also include single-parent families, people with disabilities, and homeless people. In general, the proportion of people at risk of poverty is gradually decreasing with increasing age. This is mainly due to the setting up of the social protection system and expenditure, e.g., social transfers, including old-age and survivorsʹ pension benefits. The impact of social transfers, including old-age and survivorsʹ pension benefits, on reducing poverty risk was up to 66.5% [30]. The amount of old-age retirement pensions is lower than incomes from individuals' economically active period. The decrease in incomes of households of pensioners decreases the possibility of meeting the needs of pensionersʹ households, which is also negatively influenced by the growth of consumer prices, which in June 2019 increased by 2.6% in total compared to 2018.
The Government of the Slovak Republic also responds to this fact with the valorization of pension benefits. For 2018 to 2021, a minimum rate of valorization of pension benefits was introduced. Pensioners increase their pension benefits by the percentage of the year-on-year growth in consumer prices of pensionersʹ households, at least by a fixed amount determined as 2% of the average amount of the respective type of pension. Starting in January 2020, approximately 88% of Slovak Republic European Union seniors will valorize their pension benefit for pension inflation, which is 2.9%. The remaining pensioners will increase their pension by a fixed amount corresponding to 2% of the average amount of the pension in question, which is €9 for 2020. Thanks to this measure, below-average retirement pension benefits will grow faster and may dampen the decline in the income of pensionersʹ households as up to 6.9% of people aged 65 and over live at risk of poverty ( Figure 2). The 55-75 years old age group has experienced the greatest increase in incomes over the last 20 years and pensioner poverty has declined very rapidly in many countries, so it is now below the average in the OECD population [33]. Over the past 20 years, average old-age pensions in Slovakia have been in the order of 43 to 48 percent of the average wage in a given year. For comparison, while in 1991 the average old-age pension amounted to €67, the average gross monthly wage in the economy was €125. Currently, retirement pension beneficiaries have an average pension of €444 and an average wage of €1013 ( Figure 3). set one of its goals to raise their standard of living through a Christmas allowance [35]. On the contrary, child poverty has deepened and is currently above average in the general population. The results of a survey in 2016 show that 21.4% of children aged 0-15 were at risk of poverty and compared to the survey in 2015, the indicator increased by 0.6%. In the case of children 0-17 years, based on the results of the Labor Force Sample Survey, their share in households without employment decreased from 9.3% to 7.7%. In this age group, according to The European Union Statistics on Income and Living Conditions (EU-SILC) 2016 results, the percentage of those who live in households with very low work intensity (occasional work) rose from 8% to 8.2%. This is despite the clear evidence that the well-being of children is a key determinant of the quality of their adult life-how much they will earn, how healthy they are, and so on. The increase in child poverty deserves more political attention than it currently receives in many countries.
In the following part of the paper, we will discuss the genesis of legislative changes in the Christmas pension as well as its merits. We will also focus on the financial impact of paying Christmas pensions on public finances and, in particular, on the income situation of pension beneficiaries.
Materials and Methods
Through a theoretical analysis of domestic as well as foreign sources we defined dimensions of quality of life in the context of a fiscal instrument for more than a ten-year observation period. By analyzing the documents, we examined the legislative definition of the Christmas contribution. Furthermore, in the empirical part, we used a secondary analysis on the basis of which we analyzed the data on poverty of selected population groups. By using abstraction, we ignored some of the parts and features of the subject matter in question so as to address only those characteristics that were essential to our focus. We formulated the conclusions by generalizing on the basis of the examined indicators.
The Mechanism for Determining the Amount of the Christmas Allowance
The legislative conditions for the provision of the Christmas allowance are regulated by Act No. 592/2006 on the provision of a Christmas allowance to certain pension beneficiaries and on the amendment of certain acts, as amended [36]. The Christmas allowance is paid to the recipient of a pension (retirement pension, early retirement pension, disability pension, social pension, widowʹs pension, widowerʹs pension, orphanʹs pension, pre-qualified retirement pension, pre-qualified retirement pension, retrained retirement pension, pre-qualified orphanʹs retirement pension) resident in the territory of the Slovak Republic, provided that the amount of the pension payable in December of a calendar year does not exceed 60% of the average monthly wage in the economy of the Slovak Republic reported by the Statistical Office of the Slovak Republic for the calendar year preceding.
In order to financially help more people in the Christmas period, which brings with it a higher financial burden, in 2019, the ceiling for entitlement to the Christmas allowance was increased from 60% to 65% of the average monthly wage in the Slovak economy reported by the Statistical Office of the Slovak Republic in the calendar year preceding the calendar year in which the Christmas allowance is paid. The average monthly wage in 2018 was €1013.
In if the amount of the pension or the sum of the amounts of the pensions is more than twice the subsistence minimum for one adult natural person; the amount of the Christmas allowance is at least €10.
According to the Ministry of Labor, Social Affairs and Family of the Slovak Republic (2019), the total number of persons concerned, including the Christmas allowances paid, was estimated at approximately 1,30 million people in 2019. Compared to 2018, the number of Christmas subscribers was expected to increase by approximately 72,000 people due to an increase in the entitlement to the Christmas subsidy from €572.4 to €658.5. Total expenditure on the Christmas allowance was estimated at around 154.3 million euros in 2019.
The Christmas allowance is paid by the pension payer (Social Insurance Agency) or one of the authorities (Military Social Security Office, Customs Directorate of the Slovak Republic, Directorate General of the Prison and Judicial Guard Corps, Directorate General of Railway Police, Ministry of Interior, Slovak Information Service, and National Security Authority) to which the Christmas post belongs. If the pension to which the Christmas allowance is paid is assigned to both the Social Insurance Agency and the Office, the Christmas allowance is paid by the payer of the pension.
According to Article 1 § 1 par. 1 of Act No. 592/2006 Coll. as amended, a state social benefit does not belong to a recipient of a pension residing in the territory of one of the states of the European Economic Area and Switzerland except for a citizen of the Slovak Republic. Furthermore, the Christmas allowance does not belong to a recipient of a pension resident in the territory of a contracting state, as no Slovak social security agreement applies to this state social benefit.
In relation to the determination of the condition of residence in the territory of the Slovak Republic, the Slovak Republic faced an action brought by the Commission before the Court of Justice. The Commission based its legal opinion on the Regulation on the Coordination of the Social Security Systems of the Member States of the Union Regulation (EC) No. 883/2004 (Regulation of the European Parliament, 2004), which prohibits, in principle, discrimination on the basis of the criterion of the state of residence [37]. In its judgment on 16 September 2015, the Court ruled that the granting of the Christmas allowance is subject to precise and objective conditions which leave the competent authorities no discretion as to the personal needs of the applicant. The Court also held that, although the purpose of the Christmas allowance is to supplement the means of subsistence of persons who have reached a certain age, it is also intended to improve the difficult social situation of other lowincome recipients. In those circumstances, the Court ruled that the Commission had not shown that the Christmas allowance could be classified as an old-age benefit which, as a result, fell within the scope of the regulation and, therefore, dismissed the action (Court of Justice of the European Union, 2015) [38]. In 2010, the Christmas allowance payment mechanism was modified, which was more effective than the Christmas allowance payment mechanism in previous years, as a more efficient escalation of borders and amounts of the Christmas allowance were used, while the new concept took greater account of lower income retirees. The total number of pensioners who received the Christmas allowance in 2010 was 1,131,102 pensioners. In the Regulation of the Government of the Slovak Republic No. 409 of October 2010, the amounts of the Christmas allowance for 2010 were fixed at: The year 2014 brought some changes in the payment of the Christmas allowance to pension beneficiaries under Act No. 240/2014 Coll. In particular, the following changes were made:
Discussion
• the maximum amount of the Christmas allowance increased to €87.26; • the coefficient of 0.10 in the Christmas allowance formula was changed to 0.18, thus achieving increased solidarity with low-income pension recipients; • pensions were taken into account: received from the special system of social security for police and soldiers, received from abroad, paid out of the old-age pension savings scheme; and • the amount of the Christmas allowance was increased by €12.74 for recipients of the Christmas allowance with retirement pensions up to €396.18-twice the subsistence minimum for one adult natural person-qualified.
The Christmas allowance in 2014 was arranged created by for pensioners who had a pension equal or lower than €494. 40 In 2015, the amount of the Christmas allowance increased once again by €12.74 for recipients of the Christmas allowance with the total amount of retirement pensions up to and including €396.18 qualified. The entitlement to the Christmas allowance in 2015 included pensioners who had a pension less than or equal to €514.80.
As the subsistence level did not increase in 2016, the individual amounts necessary to calculate the amount of the Christmas allowance did not change. Expenditure for the payment of the Christmas allowance to some pension beneficiaries in 2016 (excluding supplements) was 77.4 million euros, while in 2015 it was 77.2 million euros.
In order to improve the status of low-income pensioners, the Act No. 266/2017 Coll. was created. A one-off increase in the amount of the Christmas allowance by €12.74 was granted in 2017 to those Christmas benefit recipients whose sums of several pensions were up to twice the amount of the subsistence minimum for one adult natural person. The amount of the Christmas allowance was increased by €12.74 for recipients of the Christmas allowance with the total amount of pensions up to and including €398.96. In December 2018, the Social Insurance Agency paid a Christmas allowance to 1,179,494 pensioners in the total amount of €74,188,318.09 (Figure 4). The Christmas allowance was calculated in 2018 according to the formula: VP = 87.26 − 0.18 × (D − ZM). (Note: In 2018, the subsistence minimum amount for one adult natural person was €198.09.) The Christmas allowance was increased by a single € 2.74 if the amount of the pension or the sum of the pension amounts was €410.14 per month or less. Based on the detailed legislative and technical process of the Christmas allowance as a nonsystemic benefit in the Slovak social system, we were able to clearly prove that seniors were not the main group at risk of poverty. Furthermore, through the analysis we managed to show that some Slovak politicians say that they are thinking of improving the quality of life of Slovak pensioners through the transformation of so-called Christmas benefit for the 13th pension, but this is not true, as its benefit for the elderly decreases from year to year, as evidenced by the values calculated in Figure 5. Calculated figures show that Christmas pensions accounted for only 57.89% of the monthly pension for people with the lowest pension in 2006 and only 14.47% of the monthly pension for people with pension at the upper limit of the qualifying pension. In 2018, this ratio represented only 42.55% of the monthly pension for people with the lowest pension level (or 48.76% with a one-off increase of €12.74). For pension beneficiaries at the maximum level, it was only 3.70%, which is clearly not the thirteenth pension.
Conclusions
In the paper, we discussed the Christmas allowance as a new form of income increase in retirement age, which should serve to maintain an adequate standard of living for Slovak seniors. The Christmas allowance as a non-systemic benefit has a positive social impact on the recipients of pension benefits who qualify for the Christmas benefit, but has a negative impact on the general government budget (Figure 4) and at the same time it is not paid in a sufficiently targeted way since it does not take into account the actual property situation of pensioners but only the amount of the old-age pension. In particular, the government should provide financial support to those groups that are more vulnerable to poverty, especially families with children.
The situation of children in Slovakia is not favorable. This is despite the economic situation in 2018 improving and the number of people at risk of poverty or social exclusion declining. Among the demographic groups, it is children who are most at risk of income poverty in Slovakia. Persistent poverty is twice as high in children up to 18 years as in the general population [48].
The higher risk of child poverty and in particular poverty and deprivation of single-parent households and multi-child households is also related to the lower efficiency of social transfers compared to the EU average. Social transfers reduce the poverty of children under 18 in the EU by 13.3 percentage points on average, in Slovakia only 8.4 percentage points [49].
Despite the fact that the group of pensioners does not belong to the category most at risk of poverty, the Government of the Slovak Republic is gradually considering transforming the Christmas contribution into the so-called 13th pension for recipients of pension benefits. However, the current Christmas allowance cannot be considered the thirteenth pension, since it amounted to only 57.89% of the monthly pension for people with the lowest pensions in 2006 and only 14.47% of the monthly pension for people with a pension at the upper limit the pension is still paid. In 2018, this ratio represented only 42.55% of the monthly pension for people with the lowest pensions (or 48.76% with a one-off increase of €12.74). For pension beneficiaries at the maximum level, it was only 3.70%, which is clearly not the thirteenth pension ( Figure 5). However, it is not yet known whether there would be so-called the thirteenth full pension paid to all pension beneficiaries in future, or if it would depend, for example, on the amount of the subsistence minimum. The first option would have a significant positive impact on the incomes of beneficiaries of all types of pensions, as well as on increasing the rate of replacement of pensions to the average wage, while avoiding discrimination, as it would be granted to all pension beneficiaries. The disadvantage of this alternative is, in particular, a considerable widening of the deficit of the Social Insurance Agency, which is already an issue. At the same time, this would be a non-systemic measure, as the provision does not result from the actuarial principle applied in the pension system and at the same time the element of solidarity is disappearing. The second alternative, which is based on solidarity, also shows positive aspects in the form of a positive impact on the incomes of the lowest pensioners, as well as an increase in the replacement rate of this group of pensioners. The negative side of this alternative is the clear dissatisfaction of higher income pensioners, the deepening of the deficit of the Social Insurance Agency, as well as the unclear change in the construction of the subsistence minimum for the future. It should be borne in mind that, according to a survey by the Statistical Office, recipients of pension benefits are not among the most vulnerable groups at risk of poverty. If the Government of the Slovak Republic is interested in systematic assistance to pension benefit recipients, it may proceed with a systematic increase in pensions. However, this measure does not affect the marketing of approximately 1.2 million pensioners who qualify for the Christmas contribution as if the same amount of resources were distributed over a regular monthly pension. If the financial amount allocated to pay the Christmas allowance was divided into a monthly pension benefit, the means would be used to improve the standard of living of pensioners and not to purchase Christmas gifts for their family members. Defending the merits of the Christmas allowance through the "payout history" is not a sufficient argument for its merits [50]. Quite the contrary, it addresses a political dimension rather than professional. The horizontal dimension clearly points to the need for a systematic increase in pension benefits, which will also ensure an overall improvement in the quality of life of seniors, as Slovak seniors ranked 25th in the Global Retirement Index [51]. | 2020-05-21T00:13:07.978Z | 2020-05-07T00:00:00.000 | {
"year": 2020,
"sha1": "dc43630c9c6266304b7aa1d290cfffd860e5d724",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2071-1050/12/9/3810/pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "b3d2017090d6ba69d373e1b040c4fb32baedb9c1",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Business"
]
} |
9513440 | pes2o/s2orc | v3-fos-license | Adolescent Ethanol Exposure Leads to Stimulus-Specific Changes in Cytokine Reactivity and Hypothalamic-Pituitary-Adrenal Axis Sensitivity in Adulthood
Adolescent alcohol use comprises a significant public health concern and is often characterized by binge-like consumption patterns. While ethanol exposure in adulthood has been shown to alter the stress response, including the Hypothalamic–Pituitary–Adrenal (HPA) axis, few studies have examined whether binge-like ethanol exposure during adolescence results in enduring changes in HPA axis sensitivity in adulthood. In the present studies, adolescent Sprague-Dawley rats were given intragastric (i.g.) intubations of ethanol (4 g/kg) or vehicle once per day for three consecutive days, beginning on postnatal day (P) 30 (±1). This exposure was followed by a 2-day period of rest/withdrawal. Rats received a total of either two (Experiments 1, 2 and 3) or four (Experiment 4) cycles of ethanol exposure and were subsequently allowed to age normally until adulthood. In Experiment 1, adult, (P71–75), ethanol- or vehicle-exposed rats received a 60 min restraint stress challenge. In Experiment 2, rats received a 50 μg/kg injection of lipopolysaccharide (LPS). In Experiment 3, rats received a challenge of 2.5 g/kg ethanol (intraperitoneally; i.p.). In Experiment 4, male and female ethanol- or vehicle- exposed rats received a 50 μg/kg injection of LPS. In all experiments, blood samples were collected for later assessment of corticosterone (CORT), blood ethanol concentrations (BECs), and the cellular fraction of blood was analyzed for cytokine gene expression. As expected, all three challenges led to a time-dependent surge in CORT. Gene expression analyses of cytokines (Interleukin [IL]-6, IL-1β, and Tumor necrosis factor alpha [TNFα]) from the cellular fraction of blood revealed unique, time-dependent patterns of cytokine expression depending upon the nature of the adult challenge incurred (restraint, LPS, or EtOH). Importantly, adolescent ethanol exposure led to attenuated restraint and LPS-induced cytokine expression in males, whereas female rats displayed an absence of cytokine alterations, and a tendency toward heightened HPA axis reactivity. These findings suggest that adolescent ethanol exposure may cause lasting alterations in cytokine regulation and HPA axis sensitivity that (a) persist into adulthood; (b) may vary depending on the nature of the challenge incurred during adulthood; and that (c) are sex-specific.
INTRODUCTION
Alcohol use and abuse disorders comprise a substantial public health and financial burden, resulting in an estimated 3.3 million deaths per year globally and $223.5 billion in financial burden to the United States alone (Centers for Disease Control and Prevention, 2014;World Health Organization, 2014). It has been estimated that almost 75% of this impact can be attributed to binge drinking, defined as ethanol consumption that results in Blood Ethanol Concentrations (BECs) of 0.08 mg/dL or higher (Centers for Disease Control and Prevention, 2014). Alcohol consumption is prevalent in adolescence and a large percentage of adolescent individuals that consume alcohol exhibit binge-like consumption patterns. These patterns of intake typically begin in early to late adolescence, peaking in early adulthood (21-25 years of age), with consumption tapering across adulthood in individuals that exhibit normative alcohol use. Binge-like alcohol consumption can be particularly harmful, and early, high frequency binge drinking has been shown to correlate with later alcohol use disorder (AUD) development (Grant and Dawson, 1997;Crews et al., 2016).
Adolescence is a time of unique sensitivity to the consequences of ethanol exposure. Mirroring what is seen in humans, adolescent rodents consume notably higher quantities of ethanol than adult counterparts (Doremus et al., 2005;Spear, 2016). Adolescent rats show altered reactivity to ethanol exposure with increased sensitivity to the positive effects of ethanol exposure and decreased sensitivity to the negative effects (Doremus et al., 2003;Varlinskaya and Spear, 2004a). Adolescent animals have also exhibited increased locomotor sensitization to ethanol and higher sensitivity to the motivational effects of ethanol (Pautassi et al., 2008) compared to adult counterparts that received similar quantities of ethanol. This differential reaction to a spectrum of consequences of ethanol exposure likely contributes to adolescent animals consuming higher quantities of ethanol, since insensitivity to negative consequences of high dose ethanol use (sedation, motor impairment, aversion) and hypersensitivity to various positive elements would seem to promote future ethanol consumption (Varlinskaya and Spear, 2004b).
A more concerning element of adolescent ethanol exposure is its ability to cause potentially lifelong changes in development. Adolescence is a critical time of neural refinement during which the brain undergoes a series of structural and functional changes. Increases in myelination, decreases in gray matter volumes, and a reduced total number of synapses indicate that adolescence is a time during which novel connections are being made and prior connections are being refined (Giedd, 2008). Underlying neurodevelopmental changes during adolescence may confer added potential for insult from substances such as ethanol (Guerri and Pascual, 2010). A common functional consequence of adolescent ethanol exposure is the retention of an adolescent-like phenotype into adulthood. This ''locking in'' of the adolescent phenotype into adulthood has been shown to occur in a variety of phenomena and is not seen in adult rats following exposure to similar quantities of ethanol. Behavioral phenomenon such as resilience to withdrawalinduced anxiety (Doremus et al., 2003), reduced sensitivity to the sedative effects of ethanol (Matthews et al., 2008), as well as reduced sensitivity to conditioned taste aversion (CTA) induction (Diaz-Granados and Graham, 2007) have all demonstrated this locking in-like effect following adolescent ethanol exposure when challenged during adulthood. This phenomenon has also been demonstrated to extend beyond just ethanol challenge effects. Observed baseline differences in impulsivity, novelty seeking, and many other phenomena have also shown to linger into adulthood following adolescent ethanol exposure (see Spear and Swartzwelder, 2014 for a recent review).
Ethanol alone is a potent stimulator of the hypothalamicpituitary-adrenal (HPA) axis as demonstrated by dose-dependent increases in plasma adrenocorticotropic hormone (ACTH) and corticosterone (CORT; Rivier et al., 1984;Doremus-Fitzwater et al., 2014). Importantly, cross sectional studies comparing adolescent and adult rats have indicated that stressors such as hypoxia and restraint stress result in prolonged ACTH and CORT responses when compared to adult rodents (Romeo et al., 2016), though differences in peak CORT responses were not observed. A similar difference occurs following ethanol challenge as well, where adolescents exhibit delayed recovery (or shutoff) of the axis relative to adults. While this effect was more pronounced in female rats, both male and female adolescent rats have shown increased CORT at several points after ethanol when compared to adult animals challenged with the same dose of ethanol (Willey et al., 2012). In contrast, prior exposure to ethanol in adolescence has been shown to blunt the HPA-axis response to subsequent ethanol exposure up to 24 days later (Lee and Rivier, 2003). Thus, intrinsic differences in the HPA axis response to stress between adolescents and adults are now well-precedented in the literature, though the mechanisms underlying these differences remain obscure.
Beyond effects on the HPA axis, ethanol also has substantial effects on cytokine release and the immune system. Cytokines are a class of proteins that play a key role in propagation of the inflammatory response evoked by immunological threats, tissue damage, and other challenges such as stress (see Deak et al., 2015 for a recent review). In particular, the pro-inflammatory cytokines IL-1, TNF-α and IL-6 are rapidly induced by immune challenges or psychological stress, and often act synergistically with one another to influence host defense and tissue repair. These cytokines also potently induce HPA axis activation, with glucocorticoids in turn influencing subsequent cytokine expression and activity. In this way, HPA-immune interactions serve as an exemplar of bi-directional communication between the brain and the immune system.
From a longitudinal perspective, recent studies have begun to explore whether the adolescent period also represents a critical period during which ethanol exposure can produce life-long changes in HPA axis sensitivity and immune function. For instance, adolescent ethanol exposure (6 exposures to 3 g/kg intraperitoneal (i.p.)) in male rats resulted in reduced baseline CORT and sensitized CORT release following acute or repeated ethanol challenge in adulthood (Przybycien-Szymanska et al., 2010. These changes in CORT patterns correlated with decreased vasopressin (AVP) and increased corticotropinreleasing hormone (CRH) within the paraventricular nucleus of the hypothalamus (PVN) of adolescent ethanol exposed rats following acute and binge ethanol challenge, suggesting alterations in central regulation of HPA axis regulation. In contrast, a more recent study showed that adolescent ethanol exposure (daily exposure to ethanol vapor for 6 h maintaining blood alcohol levels, BALs of 200 mg/dL) blunted the HPA axis response to adult ethanol challenge (3.2-4.5 g/kg intragastric (i.g.); Allen et al., 2016). Though dose, route and frequency of ethanol exposure during adolescence are likely explanations for the differences between these studies, a more fundamental question regarding the potential for stressor-specific outcomes in adults with a history of adolescent ethanol exposure still remains.
With this in mind, our primary goal was to establish how binge-like ethanol exposure during adolescence would alter cytokine and HPA axis responses to stress challenges incurred during adulthood. In doing so, we adopted an adolescent alcohol exposure procedure that has previously been shown to produce changes lasting to adulthood in HPA axis regulation, including alterations in CORT release as well as changes in AVP and CRH following binge and acute ethanol exposure (Przybycien-Szymanska et al., 2010. Following intermittent binge ethanol exposure in adolescence, three distinct studies were executed in which rats were given different challenges in adulthood to test whether axis sensitivity may depend upon the nature/modality of the stress challenge experienced as adults. Thus, adult rats with a history of adolescent ethanol exposure were challenged by restraint, lipopolysaccharide (LPS) administration, or adult ethanol exposure (Experiments 1-3 respectively). Though our primary interest was HPA axis sensitivity indicated by plasma CORT responses, we also had a secondary interest in how adolescent intermittent ethanol (AIE) would influence the expression of cytokines within the cellular fraction of whole blood as a secondary index of stress sensitivity, and due to the potential mechanistic role that plasma cytokines might play in HPA axis responses (Rivest et al., 2000;Szelényi, 2001). In human research, analysis of peripheral cytokine levels has been shown to be useful in monitoring general immune activation and to correlate well with expected levels of neuroimmune activation (Sullivan et al., 2006). In addition, elevated levels of peripheral pro-inflammatory cytokines have been shown to correlate with psychiatric illness in a specific subset of individuals (Fillman et al., 2016). Finally, due to the literature demonstrating that male animals display profound HPA axis blunting following adolescent ethanol exposure, we examined potential sex differences in HPA axis reactivity following adolescent alcohol exposure.
Subjects
In the first three studies, male Sprague-Dawley rats (N = 50) were purchased from Harlan Laboratories and shipped at postnatal day (P) 22 ± 1. In the fourth study, male and female Sprague-Dawley rats (N = 30) were bred in-house using breeders originally derived from Harlan (Envigo). With day of birth being deemed postnatal day (P) 0, rats were allowed to develop normally until P21 at which point animals were weaned and pair-housed with a same-sex partner from a different litter in standard Plexiglas bins. In all cases, rats were given 1 week to acclimate to colony conditions and were then handled (P29) for 2-3 min to acclimate them to human contact prior to experimentation. Colony conditions were maintained at 22 ± 1 • C on a 12:12 light:dark cycle. Animals were pair-housed in standard Plexiglas bins with ad libitum access to food and water. Rats housed in pairs were always assigned to the same experimental group. In all experiments, rats were treated in accordance with Public Health Service (PHS) policy and all experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) at Binghamton University.
Adolescent Alcohol Exposure Procedure
Starting at early adolescence (P30-32) rats received 4.0 g/kg intragastric (i.g.) intubations of ethanol or an equivolume dose of vehicle (tap water) once per day at approximately 1000 h. When administered i.g., ethanol solution was prepared daily using 95% ethanol stock diluted to a final working concentration of 20% in tap water. Weights were taken daily at least 90 min prior to intubations. Animals received three consecutive days of intubations followed by a 2-day period of rest/withdrawal during which animals remained undisturbed in their home cage. This cycle of intubations (3 days on, 2 days off) was then repeated a second time (Experiments 1-3), reaching its conclusion on P40-42. In Experiment 4, this cycle was administered four times, spanning ages of approximately P30-P50. Rats were then reared to adulthood without further manipulation until weights were again collected on P73-75.
At this time, rats were randomly divided into cohorts for independent experiments.
Drug Preparations
When injected i.p., ethanol was prepared in a 20% solution mixed fresh daily using 95% ethanol stock with sterile physiological saline used as the vehicle. LPS (from serotype E0111:B4; Sigma Chemical Co.) solution was initially diluted 1.0 mg/mL using sterile (pyrogen-free) physiological saline and stored in frozen aliquots at −20 • C until needed. On the day of experimentation, an aliquot of LPS was thawed and mixed fresh daily to the working concentration of 50 µg/mL LPS and delivered on a 1 mL/kg basis, also in pyrogen-free physiological saline.
Blood Sampling and Processing Procedure
In each experiment, rats were removed from their home cage and briefly placed into restraint tubes for the acquisition of blood samples. Blood samples were collected using the tail clip method in which the last 0.5-1.0 mm of the tail was transected. The tail was then stroked until whole blood was obtained as previously described (Hueston and Deak, 2014b). Rats were immediately returned to their home cage following each time point except in conditions where rats were assigned to the restraint stress condition. Animals were returned to their home cage within 2 min after first intrusion. Following collection, plasma was immediately separated through refrigerated centrifugation and stored at −20 • C until analysis. RNA was extracted from the remaining fractionated layers of leukocytes and erythrocytes. Prior studies have shown that RNA extraction of whole blood is possible (Schwochow et al., 2012). Due to elimination of the plasma fraction and the lack of nuclei or cell organelles such as mitochondria in erythrocytes, it was reasoned that extraction on the combined buffy layer and erythrocyte fractions would largely reflect gene expression changes in leukocytes. The combined layers were stored at −80 • C until analysis of expression of blood cytokine levels using Real Time Reverse Transcription Polymerase Chain Reaction (RT-PCR).
Corticosterone Measures
Quantitative determinations of plasma CORT levels were assessed using a CORT EIA kit (Cat No: ADI-901-097; Enzo Life Sciences, Farmingdale, NY, USA) as described in Hueston and Deak (2014a) with an inter-assay variability of 7.616% and an assay sensitivity of 27.0 pg/mL. Manufacturer's instructions were followed for all steps except that samples were heat inactivated by immersion in 75 • C water for a period of 60 min to denature endogenous CBG.
Reverse-Transcription Polymerase Chain Reaction
All RT-PCR was conducted using procedures described in previous work (Hueston and Deak, 2014a). Blood pellets were stored at −80 • C until the time of RNA extraction. Total RNA in the sample was homogenized in Trizol RNA reagent (Invitrogen) using a TissueLyser and 5 mm stainless steel beads (Qiagen). Total RNA was then extracted through the use of RNeasy mini columns (Qiagen) following manufacturer's instructions. RNA yield and purity were evaluated using a Nanodrop system (ThermoScientific) and cDNA was synthesized using a QuantiTect reverse transcription kit (Qiagen). All RT-PCR was run using a CFX384 real-time PCR detection system (Bio-Rad). All results were normalized to β-Actin as a housekeeper gene. Primer sequences for all targets run can be found in Table 1.
Blood Ethanol Concentrations
BECs were determined from 5 µL samples of plasma using an Analox AM-1 alcohol analyzer (Analox Instruments, Lunenburg, MA, USA; as described in Doremus-Fitzwater et al., 2015). The machine was calibrated using a 200 mg/dL standard and tested every 15 samples using an alcohol quality control standard of known concentration purchased from Analox Instruments. Samples below 10-15 mg/dL were within a range considered to be noise and as such were interpreted as 0 values.
Statistical Analyses
All analyses of CORT levels, BECs and cytokine expression were analyzed using a mixed factorial ANOVA (p < 0.05). Fisher's Least Significant Difference (p < 0.05) was used as a post hoc examination in instances where two way interactions were noted to identify loci of significant difference. Outliers were defined as data points that were more extreme than ±2 standard deviations from a given experimental group's mean. If an outlier was identified within a given PCR reaction, it was excluded from that specific analysis but not across the remainder of the gene targets. Before analyzing individual cytokine data, β-Actin was examined to determine differences that may exist between groups in expression of this target gene of reference. All gene expression was reported quantified relative to expression of this reference gene. In the first experiment, restraint stress was chosen as a challenge due to its ability to reliably produce activation of the HPA axis and for reported differences between adolescent and adult CORT responses to restraint stress (Romeo et al., 2016). Male rats (n = 6-8 per group; N = 14) were given two cycles of intermittent ethanol exposure or vehicle intubations as described above. At adulthood (P73-75), rats from both groups were placed into clear Plexiglas restraint tubes and a baseline blood sample was collected immediately. Rats remained in restraint and subsequent blood samples were collected at 15, 30 and 60 min time points. Immediately after the 60 min time point, rats were returned to their home cage and a final recovery sample was collected 60 min later (i.e., 120 min following stress onset; see Figure 1A).
Experiment 2: Axis Reactivity to Lipopolysaccharide Challenge Following Adolescent Ethanol Exposure
In the second experiment, LPS challenge was chosen as a representative, immune-based challenge that is known to activate the HPA axis through cytokine-mediated mechanisms. Male rats (n = 8 per group; N = 16) were given two cycles of intermittent ethanol exposure or vehicle intubations as described above.
Experiment 3: Axis Reactivity to Ethanol Challenge Following Adolescent Ethanol Exposure
In the third experiment, rats were given an acute ethanol challenge during adulthood to test for durable changes in HPA axis sensitivity to the same challenge(s) utilized during adolescence, and to provide a referent to other published studies (Przybycien-Szymanska et al., 2010Allen et al., 2016). Male rats (n = 8-10 per group; N = 18) were given two cycles of intermittent ethanol exposure or vehicle intubations as described above. During adulthood (P73-75), the test day began with collection of a baseline blood sample, after which rats were given a 2.5 g/kg i.p. ethanol or equivolume vehicle. Subsequent blood samples were collected 30, 60, 120 and 240 min post injection (see Figure 1A).
Experiment 4: Sex-Differences in the Consequences of Adult LPS Challenge Following Adolescent Ethanol Exposure
The goals of the fourth experiment were to: (a) replicate the initial effect of reduced CORT and cytokine responses observed in adults with a history of adolescent alcohol exposure; (b) extend the period of alcohol exposure more broadly across the adolescent period; and (c) determine whether the impact of adolescent alcohol exposure on adult stress sensitivity would be sex-specific. Because early life shipping can produce variations in stress history and thereby influence stress reactivity (Laroche et al., 2009a,b;Vargas et al., 2016), rats for Experiment 4 were bred in house. Thus, in Experiment 4, rats (n = 6-8 per group; N = 30) were given four cycles of vehicle or ethanol intubations utilizing the procedure described above. After reaching adulthood, all rats received a 50 µg/kg i.p. LPS injection. Blood was collected at baseline, 45, 90, 180 and 360 min post injection.
Experiments 1-3
Plasma Corticosterone Response to Challenges Data were analyzed using a mixed between-subjects (adolescent Ethanol or Vehicle exposure) and repeated-measures (time course) analysis of variance (ANOVA). Restraint challenge showed significant CORT elevation peaking around 30 min and CORT levels returning to baseline by 120 min (F (4,48) = 62.07, p < 0.001; Figure 1B). LPS challenge produced comparable increases in CORT across the 45 and 90 min time points, which then decreased by the 180 min time point (F (3,42) = 22.64, p < 0.001; Figure 1C). Adult ethanol challenge increased CORT within 30 min and sustained high CORT levels through the 120 min time point (F (4,64) = 70.65, p < 0.001). CORT still showed notable elevation at 240 min ( Figure 1D), though the response was clearly resolving at this time point. Although no statistically significant effects of adolescent ethanol history on CORT response were observed on in response to the adult challenges, visual inspection of the data suggested a moderate attenuation of CORT in response to LPS challenge.
Blood Ethanol Concentrations Following Ethanol Challenge
As expected, rats injected with ethanol exhibited a marked and time-dependent elevation in BECs that peaked at 30-60 min and tapered over the remainder of the time course (F (4,64) = 200.15, p < 0.001; Figure 1E). Rats with a history of adolescent alcohol exposure showed trended higher mean BECs at both 30 and 60 min after ethanol, though this effect was not statistically significant. It should be noted that mean peak BECs were approximately 25 mg/dL higher at both time points relative to rats that received vehicle during adolescence. Potential differences in alcohol pharmacokinetics in adolescent-exposed rats could therefore be an area worth examining in more targeted studies.
Blood Pellet mRNA Expression Levels Following Restraint Challenge
While restraint challenge produced no significant alterations in IL-6 expression (Figure 2A), significantly increased IL-1β expression was noted at 30 min (F (4,48) = 10.16, p < 0.001) and IL-1β remained elevated throughout the remainder of the time course (Figure 2D). No alterations in TNF-α expression were noted (Figure 2G) however; significant increases in Nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha (IκBα) expression at 30 and 60 min (F (4,48) = 13.53, p < 0.001) were seen that had returned to baseline levels by the 120 min time point (Figure 2J). A pattern showing that adolescent exposure to ethanol attenuated response to adult restraint challenge was seen in IL-6, IL-1β, TNF-α, and IκBα expression. A trend showing attenuation of IL-6 expression in the adolescent ethanol exposed rats was seen at the 60-min time point (F (4,48) = 2.55, p = 0.051; Figure 2A). A similar trend showed that IL-1β expression was attenuated in the adolescent ethanol exposed rats at the 30-min time point (F (4,48) = 2.44, p = 0.06; Figure 2D). Finally, a similar reduction was seen at the 60-min time point in both TNF-α (F (4,48) = 5.49, p < 0.005; Figure 2G) and IκBα (F (4,48) = 4.77, p < 0.005; Figure 2J) expression.
Blood Pellet mRNA Expression Levels Following Ethanol Challenge
Data were analyzed using a mixed between subjects (adolescent Ethanol or Vehicle exposure) and repeated measures (0, 30, 60, 120 and 240 min post-injection) ANOVA. Analysis revealed significant changes in overall IL-6, IL-1β, TNF-α and IκBα expression across time. A significant decrease in overall IL-6 expression (F (4,64) = 21.89, p < 0.001) was noted from 60 min post-challenge to 240 min ( Figure 2C). Significant increases in IL-1β expression did not begin to occur until after 120 min showing continued escalation at the 240 min time point (F (4,64) = 27.40, p < 0.001; Figure 2F). TNF-α also showed a significant reduction in expression (F (4,64) = 39.58, p < 0.001) during the 30, 60 and 120 min time points and began to show rebound elevation at 240 min ( Figure 2I). IκBα showed an increase in expression through the 60 min time point that had begun to taper by the 240 min time point (F (4,64) = 17.27, p < 0.001; Figure 2L). Rats that received ethanol exposure during adolescence demonstrated significantly reduced IL-6 expression in contrast to vehicle-exposed rats (F (4,64) = 3.03, p < 0.05; Figure 2C). No effects of adolescent ethanol exposure on IL-1β, TNF-α, or IκBα were observed.
Corticosterone Response to Challenge
Data were analyzed using a mixed between-subjects (adolescent Ethanol or Vehicle exposure; male or female) and repeatedmeasures (time course) ANOVA. Sex-specific analyses were then run using a mixed between subjects (adolescent Ethanol or Vehicle exposure) and repeated measures (time course) ANOVA. When ANOVAs for each sex were run separately, LPS challenge produced significant increases in both male and female CORT release with female animals showing overall higher levels of CORT than males. A significant main effect of time point was found in both females (F (4,48) = 40.86, p < 0.001; Figure 3C) as well as in males (F (4,56) = 22.58, p < 0.001; Figure 3B). CORT increased at the 45-min time point and remained at peak through the 360-min time point in both males (F (4,56) = 22.58, p < 0.001; Figure 3B) and females (F (4,48) = 40.86, p < 0.001; Figure 3C). Although there was a trend indicating that adolescent ethanol exposure increased CORT release following LPS challenge in female rats (F (4,48) = 1.68, p = 0.17), no such effect was seen in male rats (Figures 3B,C). A planned comparison examining CORT in female adolescent ethanol and vehicle exposed animals at the 90 min timepoint revealed a significant sensitization of LPS induced CORT in adolescent ethanol exposed animals when compared to vehicle exposed animals at the same timepoint (t (12) = 2.34, p < 0.05).
Blood Pellet mRNA Expression Levels Following LPS Challenge
When the data were analyzed using an omnibus ANOVA including sex as a variable, a significant sex by adolescent interaction was noted (F (1,26) = 7.20, p < 0.05; Figures 4A,B). When ANOVAs for each sex were run separately, IL-6 showed a trend in the male adolescent ethanol exposed rats showing attenuation of expression in contrast to vehicle exposed animals (F (4,56) = 2.04, p = 0.10; Figure 4A). No such effects were noted in female animals ( Figure 4B). While adolescent ethanol exposure visibily reduced IL-6 expression in male rats it had little recognizable effect on IL-6 expression in the females. When data were analyzed using an omnibus ANOVA including sex as a variable, rats that received adolescent ethanol exposure showed significantly reduced levels of IL-1β expression when compared to vehicle exposed counterparts (F (4,104) = 2.95, p < 0.05) however, this appears to be largely a result of the differences noted in the male animals (Figures 4C,D). When ANOVAs for each sex were run separately, a significant attenuation of IL-1β expression was replicated in the male, adolescent ethanol exposed rats at the 180 min time point (F (4,56) = 3.44, p < 0.05; Figure 4C). Interestingly, no such interaction was noted for the female animals although a significant increase in overall expression was noted at the 90 and 180 min time points (F (4,48) = 14.63, p < 0.001; Figure 4D).
When data were analyzed using an omnibus ANOVA including sex as a variable, males showed much greater TNF-α increase in expression over the time course in contrast to female counterparts, regardless of adolescent exposure (F (4,104) = 2.71, p < 0.05; Figures 4E,F). Regardless of sex, TNF-α expression increased with peak levels being reached by the 360 min time point (F (4,104) = 11.06, p < 0.001; Figures 4E,F). When ANOVAs for each sex were run separately, TNF-α expression in males showed a similar pattern of expression as in the previous experiment with a noticeable but not significant attenuation of expression developing in the adolescent ethanol exposed animals. Males showed much higher overall levels of expression than females (F (1,26) = 5.07, p < 0.05; Figures 4E,F). Females also showed a significant increase in overall TNF-α expression at the 360 min time point (F (4,48) = 4.90765, p < 0.05; Figure 4F).
IκBα closely mirrored the pattern of expression observed in Experiment 1. When data were analyzed using an omnibus ANOVA including sex as a variable, a significant three way interaction between adolescent exposure, sex and treatment (F (4,104) = 3.60, p < 0.01) revealed that again adolescent ethanol exposure resulted in significant attenuation of IκBα exposure but exclusively in the male animals (Figures 4G,H). Female animals showed little to no consequence of adolescent ethanol exposure. Interestingly while in both female and male rats IκBα expression peaked at approximately the 1500% mark in the vehicle exposed male rats expression continued to escalate until the 360 min time point reaching almost double the magnitude of expression seen in the remaining three groups (Figures 4G,H).
DISCUSSION
Our primary goal in these studies was to examine how AIE would alter the HPA axis response to different types of challenge in adulthood. This set of studies shows that not only does adolescent ethanol exposure produce lasting changes in HPA axis sensitivity into adulthood, but that these changes are both sexually dimorphic and stimulus specific. Whereas adolescent ethanol exposed male rats showed blunting of peripheral cytokine expression in several key pro-inflammatory cytokines in response to restraint and LPS administered during adulthood, minimal differences were observed in response to ethanol challenge. These findings suggest that the HPA axis is not debilitated by adolescent ethanol challenge and might suggest adaptations that occur extrinsic to the axis. The three challenges used in the study each represent unique modalities for stimulating the HPA axis, restraint being psychological, LPS being immune, and ethanol being the same stimulus used to invoke the changes. Each stimulus produced a unique profile of cytokine expression in blood and the effects of AIE were different between challenge modality.
Alcohol is known to produce immune dysfunction. However, unlike other illnesses, the immune dysfunction that occurs is rarely at a level that would be considered clinical, rather it becomes notable following exposure to a subsequent challenge (Szabo and Saha, 2015). One example of this is the increased susceptibility to bacterial infections such as pneumonia and tuberculosis that occurs in heavy drinkers and alcohol dependent individuals (Cook, 1998;Ogunsakin et al., 2016). In addition, evidence from both human and animal studies has shown that prenatal ethanol exposure can result in increased vulnerability to infection and disease that may persist across the lifespan (Gauthier, 2015). In this study, a pronounced immunosuppressive effect of AIE in response to subsequent LPS challenge was seen in male rats while no difference was noted in females. In contrast, no effect of AIE on the CORT response in male rats was noted, yet a sensitization was observed in females. Although the functional significance for males and females with a history of adolescent ethanol exposure remains unclear, these divergent outcomes suggest that males may be more prone to infection, whereas females may be more prone to stress-related disorders (depression, anxiety) that are often co-morbid with HPA axis dysregulation (Weinberg et al., 2008). In this way, adolescent ethanol exposure may contribute to sex differences in specific disease vulnerabilities that emerge later in life.
Prenatal Alcohol Exposure (PAE) has also been shown to alter HPA axis reactivity in adulthood, suggesting there may be multiple developmental windows during which ethanol can impact later stress sensitivity. For instance, PAE-exposed animals show hyperactivity of the CORT response to restraint stress imposed later in life (Weinberg, 1992;Wieczorek et al., 2015). This hyperactivity has also been shown in response to LPS challenge, IL-1β challenge, local inflammatory challenge probed via turpentine injection, and other types of stressors (Lee and Rivier, 1996;Raineki et al., 2014). The consequences of PAE on adult challenge also appear to be sexually dimorphic as well as stimulus specific. PAE has led to an enhanced HPA axis response in adults to stressors such as acute restraint and forced swim in female rats but not in males (Weinberg et al., 2008). In contrast, solely PAE-exposed male rodents have shown a hyper-reactive HPA response to prolonged restraint and exposure to cold (Kim et al., 1999;Weinberg et al., 2008). Both sexes have shown hyper-reactivity of the axis to immune challenges such as LPS and IL-1β challenge (Weinberg et al., 2008). The results of the present studies clearly demonstrate that adolescent ethanol exposure is capable of causing changes in HPA axis reactivity lasting into adulthood. Based on the results of this study, the effects of adolescent ethanol exposure on later adult challenge appear to be stimulus specific and show sex specific effects, much like has been shown in studies involving PAE.
Prior work from our lab demonstrated that male adolescent rats displayed a blunted cytokine response in the CNS following either LPS challenge or ethanol challenge relative to adult comparators (Doremus-Fitzwater et al., 2015). One potential interpretation of the blunted cytokine responses observed in adolescent-exposed males in the present studies is that this may represent a ''locking-in-like'' effect. It is noteworthy, however, that, circulating concentrations of endotoxin following LPS challenge were substantially reduced in adolescents compared to adults, suggesting that adolescents may not process LPS in the same way as their adult counterparts. Whether this effect is through differential uptake and transit of LPS into the bloodstream, enhanced clearance of LPS from blood, or other immunological differences between adolescents and adults remains unclear. Nevertheless, plasma endotoxin concentrations were predictive of the HPA axis response (Doremus-Fitzwater et al., 2015). Unfortunately, it was not possible to measure plasma endotoxin concentrations in the present studies due to the limited available sample. Future studies will be necessary to test whether adolescent alcohol exposure impacts subsequent immune processing of LPS in a manner consistent with a ''locking in-like'' effect. The lack of AIE effects on cytokine expression in female rodents could suggest that female adolescent and adults would not demonstrate the previously noted differences in response to LPS challenge or that females displayed some form of resiliency to the changes that resulted in the ''locking-in-like'' effect.
Future studies should further probe the mechanisms by which adolescent alcohol exposure influences adult cytokine reactivity and stress sensitivity. LPS invokes an inflammatory response at least in part through its effects on Toll-like receptor 4 (TLR4), activating the NFKB pathway, and ultimately stimulating inflammatory cytokine expression and release (Doremus-Fitzwater et al., 2015). Female rodents have shown an enhanced ethanol-induced inflammatory response during adolescence, as exhibited by increased inflammatory cytokine levels in serum and in brain, however, this effect was not observed in male rats (Pascual et al., 2016). This difference may be mediated by up-regulated TLR4 expression during intoxication that are more pronounced in female rodents than in males, as supported by the fact that TLR4 KO mice do not exhibit these cytokine differences (Pascual et al., 2016). Binge ethanol expression has also been shown to interfere with LPS induced NFKB activity. Ethanol induces Heat shock factor protein 1 (HSF1) and 70 kilodalton heat shock protein (hsp70), two enzymes that ultimately inhibit TLR4/Myeloid differentiation primary response gene (MYD88) signaling via inhibition of the NFKB pathway (Muralidharan et al., 2014). This results in something akin to endotoxin tolerance in monocytes and macrophages pretreated with ethanol (Muralidharan et al., 2014). Lingering inhibition of TLR4/MYD88 signaling resulting from binge ethanol exposure could lead to a reduced cytokine response in male rats that could be normalized in females through sensitized TLR4 expression in female rats following AIE. While this is just one potential explanation for the effects seen in this study, future studies should address the mechanism of these changes and TLR4 signaling pathways might be a good place to start.
Although the results of this study showed few consequences of adolescent ethanol exposure on the response to adult ethanol challenge, the existing literature examining similar manipulations is conflicting. The work of Pak using a similar model of ethanol administration has shown sensitization of the axis response in males following adult ethanol challenge (Przybycien-Szymanska et al., 2010. In contrast, a different study in which male adolescent rats received AIE and were challenged with ethanol in adulthood showed blunting of the axis response to later ethanol challenge (Allen et al., 2016). There are a number of factors that could account for this disparity. In the first set of studies, the number of ethanol cycles as well as the challenge dose was different, and the experiments were conducted using Wistar rats. In the second set of studies the method of adolescent administration was via daily vapor inhalation, whereas our studies utilized an intermittent, intragastric administration procedure. This could suggest that the dose, route, and schedule of ethanol exposure during adolescence may be critical determinants of long-term changes in HPA axis reactivity and cytokine gene induction.
One important question to address in future studies will be the mechanisms underlying the apparent HPA axis sensitization observed in females as a result of adolescent ethanol exposure. In particular, it will be important to identify whether these changes reflect altered function of neural circuits that govern HPA axis regulation, or intrinsic differences in sensitivity of endocrine glands comprising the HPA axis. This can be readily tested through a systematic series of hormonal challenges (CRH or ACTH injection) as we have done under other circumstances (Hueston and Deak, 2014b; also see Spencer and Deak, 2016 for an overview). Such studies are planned for the near future. It will also be important to extend these physiological findings to behavioral circumstances as well. For instance, the suppressed cytokine response evoked by LPS would predict that males with a history of adolescent ethanol exposure should display reduced sickness behavior under conditions of acute illness as adults. Given the HPA dysregulation observed in females with an adolescent history of ethanol, one might predict heightened anxiety and/or depression in adulthood. Given the established association between age of first alcohol use and subsequent vulnerability to addiction (Ystrom et al., 2014), it will also be important to assess whether adolescent ethanol exposure might increase vulnerability to addiction, particularly in females since stress dysregulation has been posited as one pathway towards addiction (Miller and Spear, 2006;Koob, 2008). These and many other issues will need to be explored in future studies.
As with any pre-clinical model of ethanol exposure, the mode of ethanol delivery and subsequent BECs achieved require careful consideration. We elected to utilize intragastric intubation for these studies because it mimics the normal route of administration for ethanol, and allows for binge-like doses to be achieved consistently and reliably. Such BECs would be difficult to achieve through voluntary consumption in rat, though that may be a path for future studies. BECs during the adolescent intubation regimen were not assessed due to concerns that the blood sampling procedure might have an enduring influence on the study outcomes. However, prior studies from our lab reported peak BECs of approximately 180 mg/dl after delivery of 4 g/kg (i.g.) of ethanol, which is consistent with a binge-like range of BECs (Buck et al., 2011;Doremus-Fitzwater et al., 2015). Because BECs were not assessed in the present studies, we cannot rule out the possibility that the sex-specific outcomes observed here might be due to intrinsic differences in ethanol metabolism (or other pharmacokinetic differences) between males and females (Truxell et al., 2007). Finally, our studies did not include a group of non-intubated controls to assess the influence of the intubation procedure itself. We occasionally use non-intubated controls and have not seen any evidence emerge to demonstrate long-term effects of the intubation procedure itself thus far. For practical purposes, however, we cannot do so in every study design. Nevertheless, these issues raise a more general question regarding whether any stress challenge (not just ethanol) imposed during adolescence might lead to similar outcomes. This is an interesting question that will have to be resolved in future studies. Overall, any ethanol delivery approach has both strengths and limitations and the study outcomes should be considered within an appropriate framework.
While this study yielded several interesting results, there are several procedural differences that must be considered in the interpretation of study outcomes. For instance, Experiments 1-3 utilize rats shipped from a vendor post weaning, whereas Experiment 4 utilized rats bred in house. The use of shipped animals for developmental studies is not ideal (see Laroche et al., 2009a,b;Spencer and Deak, 2016;Vargas et al., 2016), though it remains common practice for many investigators who do not have access to a breeding facility. Despite this limitation, Experiments 2 and 4 both yielded similar, durable influences of adolescent ethanol on LPS-induced cytokines as adults, even though the overall magnitude of cytokine responses across studies varied widely. A second disparity in study design was the number of ethanol cycles that the animals received during adolescence. Two cycles of intermittent ethanol exposure were utilized in Experiments 1-3, whereas four cycles were utilized in Experiment 4. Ultimately, the increase in number of cycles did not alter the directionality of the effects seen although it did appear to increase their magnitude. The model of ethanol administration is an important consideration because other studies have suggested that it is not only the amount of ethanol that is the dominant factor in creating durable effects resulting from AIE, but also the number of periods of abstinence between ethanol exposures (Lopez and Becker, 2005;Diaz-Granados and Graham, 2007). Indeed, periods of acute withdrawal that occur throughout intermittent ethanol exposure may play a critical role in causing some of the lasting changes of ethanol (Spear, 2016). Finally, to truly attribute the results of this study to an adolescent critical period, follow up studies that incorporate an adult control (that receive the same quantity of ethanol over the same time frame) and that control for the time difference between AIE and challenge between the two studies would be needed. Despite these minor limitations, the present study yielded important and highly consistent results that warrant further investigation.
A final element of the study that affects interpretation of the results is the use of the cellular fraction of blood to study peripheral cytokine expression (Wang et al., 2016;Opstad et al., 2017). Although it is more common in contemporary studies to utilize plasma protein measures to detect cytokines, cytokines in plasma have very short half-lives and the assays to measure them require comparatively large samples. Thus, a viable alternative for procuring cytokine signals is to utilize gene expression in the cellular fraction, which can be done in very small samples, thereby allowing for the use of a within-subjects experimental design. Although there is always a question as to how such gene expression changes might align with functional protein measures, the detailed time course information gathered in the present study can be used to inform selection of more discrete time points for protein measures. Such studies are already planned in the near future. Furthermore, while the functional relationship between peripheral cytokines and central cytokines is unclear, prior studies have suggested that peripheral cytokine release may play a role in activating subsequent central action. LPS induced TNF-α in serum has been shown to play a role in triggering a proinflammatory response in the brain (Qin et al., 2007(Qin et al., , 2008. Further research probing the relationship between these elements of the immune response would be worthwhile. In conclusion, the effects of ethanol during adolescence appear to be mitigated by many factors. The results of this study as well as other work suggest that how the ethanol is administered, the nature/modality of the subsequent challenge stimulus, and the sex of the animals all may interact with AIE. While the rates of binge drinking in both adolescents and adults is very high, men are almost twice as likely to engage in binge drinking as women (Centers for Disease Control and Prevention, 2012). Differences in how men and women may be affected by ethanol, particularly during periods of developmental vulnerability, may contribute to the difference in AUD prevalence that exists between men and women (Nolen-Hoeksema, 2004). Despite all of this, it has been reported that women suffer more negative consequences of heavy ethanol use, including increased susceptibility to physical illness at lower levels of exposure than has been seen in men (Nolen-Hoeksema, 2004). Several drugs of abuse have been linked to weakened natural immune system function during use and at times lingering afterwards. Ethanol abuse specifically has been shown to produce a pronounced immunosuppressive effect and can increase the chances of individuals with ethanol being vulnerable to infection (Szabo, 1997;Szabo and Saha, 2015). Alterations in HPA axis reactivity and cytokine function could help to explain some of ethanol's effects on the immune system and the potential for adolescent ethanol use to cause lasting alterations in immune function. This could lead to a lifetime of heightened immune vulnerability to an array of types of challenge stimuli. Better understanding of how adolescent ethanol exposure affects the immune system and the longevity of such changes could help to better understand the risk of individuals exposed to future insults and help identify targets to reverse negative alterations.
AUTHOR CONTRIBUTIONS
ASV, TD-F, AG and TD made substantial contributions to the conception or design of the work as well as participated in the acquisition, analysis, or interpretation of data for the work. In addition, they played a significant role in drafting the work or revising it critically for important intellectual content and submitted for final approval of the version to be published. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
FUNDING
Research reported in this publication was supported by the National Institute on Alcohol Abuse and Alcoholism of the National Institute of Health under Award Number P50AA017823 to TD and the Center for Development and Behavioral Neuroscience at Binghamton University. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the above stated funding agencies. | 2017-05-17T19:31:57.934Z | 2017-05-04T00:00:00.000 | {
"year": 2017,
"sha1": "edb46f650a6bb6b663cba7e8463561df7a2b9754",
"oa_license": "CCBY",
"oa_url": "https://www.frontiersin.org/articles/10.3389/fnbeh.2017.00078/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "edb46f650a6bb6b663cba7e8463561df7a2b9754",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.