A meta-analysis of the endothelial protein C receptor rs867186 genotype in malaria

BACKGROUND : During P. falciparum infection, the binding of P. falciparum erythrocyte membrane protein 1 (PfEMP1) to endothelial cells (EC) results in the sequestration of pRBC. Several receptors located on the endothelial cells, including intercellular adhesion molecule 1 (ICAM-1), CD36, and endothelial protein C receptor (EPCR), contribute to PfEMP1 adhesion to the microvasculature. PfEMP1, expressed on the surface of parasitized red blood cells (pRBC), is composed of cysteine-rich interdomain regions (CIDR) and Duffy binding-like (DBL) domains. CIDRα1 competitively binds to EPCR with activated protein C (APC) and impairs cytoprotective and anticoagulant effects by APC, which plays important roles in severe malaria (SM) pathogenesis such as cerebral malaria (CM) and severe malaria anemia (SMA). The strategy to inhibit EPCR binding to pRBC while concomitantly strengthen its binding to APC may be crucial in restoring disrupted protein C (PC) system’s function. The purpose of this study is to evaluate the association between malaria severity and the EPCR genotypes as well as with soluble EPCR (sEPCR), and the study also addresses the physiological relevance of EPCR genetic polymorphism. RESULTS : In this study, we conducted a meta-analysis on the eligible studies by comparing the frequency of EPCR rs867186-GG versus rs867186-GA and -AA genotype in SM, mild malaria (MM) or uncomplicated malaria (UM) patients and healthy individuals from Thailand, Uganda, Benin, Tanzania, and Ghana. We also determined the relationship between rs867186 genotype and sEPCR levels. Our results showed that the genotype rs867186-GG is higher in MM/UM than in SM patients. SM patients carrying the rs867186-GG genotype have higher plasma soluble EPCR (sEPCR) levels than in rs867186-AG and rs867186-AA carriers. MM/UM patients carrying the rs867186-AG

carrying the rs867186-AG genotype have signi cantly higher level of sEPCR compared to those carrying rs867186-AA. Similarly, the rs867186-GG is associated with high sEPCR level in healthy individuals.
CONCLUSIONS : This meta-analysis demonstrates that pRBCs and EPCR interactions are associated with malaria severity, and treatments that block their binding via PfEMP1 CIDRα1 could be a potential therapy for SM.

Background
We de ne severe malaria (SM) as cerebral malaria (CM) or severe malarial anemia (SMA). SM also include malaria patients with high levels of parasitemia, hypoglycemia, and increased creatinine, which is indicative of high mortality among patients with P. falciparum infections [1,2]. The pathological feature of SM is the sequestration of pRBC to host microvasculature beds in a variety of organs and tissues. The sequestration occurs when P. falciparum binds to endothelial cells (known as cytoadherence) via the P. falciparum erythrocyte membrane protein 1 (PfEMP1) [3][4][5][6][7][8][9]. PfEMP1 protein is encoded by approximate 60 var genes. However, each parasite only expresses one var gene at a time [10]. The extracellular part of PfEMP1 is composed of cysteine-rich interdomain regions (CIDR) and Duffy binding-like (DBL) domains [11][12][13]. DBL-CIDR in tandem within N-terminus is the most conserved extracellular region [14]. The parasite sequestration, which is associated with disease severity, is determined by the binding preference and a nity of PfEMP1 to host endothelial cell receptors [15][16][17]. The most widely studied receptors are CD36 and intercellular adhesion molecule 1 (ICAM-1). CIDRα2-6 domains of PfEMP1 bind to CD36 [11,18], and DBLβ5 domains bind to ICAM-1 [13,18]. pRBCs usually bind to CD36 on endothelial cells (or platelets and monocytes) in uncomplicated or mild malaria (UM/MM), and bind to ICAM-1 in SM [15][16][17]. Recently, CIDRα1 domains of PfEMP1 was found to bind endothelial protein C receptor (EPCR) [11,18]. Turner et al. discovered that EPCR interacts with PfEMP1 domain cassette 8 (DC8) and DC13 (both of which contain CIDRα1 domain) and correlated with SM [19]. The binding of CIDRα1 to EPCR inhibits activation of protein C (APC)'s, induces apoptosis, in ammation, and coagulation. As a result, pRBCs-EPCR binding suppresses cytoprotection and barrier functions in hosts [20][21][22], which are part of mechanisms of SM of CM and SMA [23]. By examining different subsets of PfEMP1 var genes in CM samples from Ugandan children by qRT-PCR, Shabani et al. found that children expressed the highest number of EPCR-PfEMP1 binding transcripts during the onset of CM and SMA, as they recovered from the disease, the number of EPCR-binding PfEMP1 transcripts decreased [12]. Their ndings that increased EPCR-PfEMP1 binding (var CIDRα1 domain) transcript is associated with severity of malaria indicate that EPCR-binding PfEMP1 transcript could be a novel therapeutic target for SM. One interesting thing is that although the abundance of total transcripts increased with malaria severity, there is no individual transcript signi cantly increased, which suggests that all EPCR-binding CIDRα1 subtypes are required to be targeted to deplete P. falciparum sequestration in host blood circulation, in particular for sequestration in brain blood circulation [11]. Several lines of evidence support a correlation between the plasma EPCR (soluble EPCR, sEPCR) levels and the genetic polymorphisms of the EPCR gene, particularly the EPCR rs867186-A/G genotype [1,2,[24][25][26]. sEPCR can competitively bind to the CIDRα1 domain and therefore displaces EPCR binding sites for APC, which restores cellular EPCR function. Due to lack of systemic analysis of EPCR genotypes, sEPCR levels, and their relationship with malaria disease severity, in the present study, we performed a meta-analysis based on related literature to determine the clinical signi cance of the EPCR gene in SM pathogenesis.

Search strategy
Pubmed, Medline, Web of Science, Scopus and Embase were searched until April 2019 using the following terms: "endothelial protein C receptor," "EPCR," "malaria," "severe malaria/complicated malaria/uncomplicated malaria," "cerebral malaria," "severe malaria anemia," "polymorphism," and "clinical studies." Articles screened were obtained by titles rst and then by article abstracts. After excluding nonrelevant publications, the remaining papers were evaluated in the full-text version based on inclusion/exclusion criteria. All searched data were retrieved, and the language of publication was restricted to English.

Selection criteria
In this meta-analysis, we collected all eligible publications regarding the correlation between EPCR and clinical outcomes and clinicopathological features in malaria patients. The inclusion criteria includes: (1) studies that evaluate and analyze rs867186-A/G genotype using DNA from peripheral blood mononuclear cells published during 2014 to 2016; (2) studies that determine the correlation between rs867186-AG genotype and malaria severity; (3) studies that de ne CM as the presence of one of the following signs: high P. falciparum parasitemia (>100,000 parasites/μl), hypoglycaemia (glucose level <22 nmol/l), severe anemia (hematocrit <20% or hemoglobin <7g/dl), increased serum creatinine (>3 mg/dl), and unrousable coma (or a Blantyre Coma Score below 3) caused by malaria infection regardless of other signs (cerebral malaria); and (4) studies that have enough information to estimate and analyze hazard ratio (HR) and 95% con dence interval (CI). The exclusion criteria includes: (1) letters, case reports, reviews, editorials, conference abstracts, and expert opinion; (2) studies without information on the genotype of the AA and AG alleles of rs867186 (rs867186-AG genotype); and (3) all in vitro/ex vivo studies including cell culture and animal studies.

Data extraction
Two investigators independently extracted the data from the eligible studies. Disagreements were resolved through full discussion until consensus was reached. The following data was identi ed and recorded for each study: year of publication, the rst author's name, number of cases, sample source, rs867186-AG genotype, and clinicopathological parameters. Data for study characteristics were collected and summarized, and the heterogeneity of each study was evaluated as well.

Statistical analysis
The data analysis was performed using the software Review Manager 5.3 (Cochrane Collaboration, Oxford, UK) and Stata 12.0 (Stata Corporation, TX, USA). Comparisons of dichotomous measures were conducted by pooled odds ratios (ORs). We determined heterogeneity by a Chi-square test with signi cance set at P <0.10; We determined the total variation among studies by I square. I square value is an estimate of variance due to between-study heterogeneity rather than chance (the Cochran Q statistics). P-value of <0.05 was considered to be statistically signi cant. If there were heterogeneity among studies (when I square exceeded 50%), we used a random-effect model to pool the ORs; otherwise, a xed-effect model was selected.

Identi cation of relevant studies and study characteristics
Twenty-eight articles were identi ed, and twenty-three of those were excluded due to non-original articles (review), laboratory studies, and studies irrelevant to the current analysis. Finally, ve studies from 2014 to 2016 [1, 2, 24-26] met the selection criteria and were eligible and included in the nal meta-analysis ( Fig. 1).
A total of 2995 SM, 487 uncomplicated malaria, and 2105 healthy controls from Thailand, Uganda, Benin, Tanzania, and Ghana were enrolled. The characteristics are listed in Table 1. The de nition criteria for the ve manuscripts' SM are presented in Table 2.

rs867186 A>G and the risk of malaria
We compared the genotype frequencies of GG (rs867186-GG) versus AA (rs867186-AA) of rs867186 in malaria patients and healthy individuals. The prevalence of rs867186-GG is not signi cantly higher in healthy controls than in malaria patients. The pooled OR from 3 studies, including 2532 malaria patients and 2105 healthy individuals, is shown in Fig. 2  4. Association of rs867186-GG genotype with the protection of SM Among SM and uncomplicated/mild malaria (UM/MM) patients, we analyzed genotype frequencies of the GG alleles (rs867186-GG) and compared with AA and AG alleles (rs867186-AA+AG) of rs867186. We observed that the prevalence of rs867186-GG is higher in mild malaria than in SM patients. The pooled OR from 2 studies, including 892 patients with SM and 437 mild malaria is shown in Fig. 3 (OR=0.36, 95% CI=0.14-0.91, P=0.03). The possible reason that the rs867186 genotypes distinguish between mild malaria and SM but not healthy controls and malaria patients (Fig. 3) may be due to healthy controls lacking the pressure exerted by parasites of the genus Plasmodium, which causes malaria infection.

Sensitivity analyses and publication bias
A sensitivity analysis is an important part of a meta-analysis as it aims to determine the robustness of the observed outcomes. A sensitivity analysis was performed to assess the results' stability. We perform the sensitivity analysis by excluding one study at a time to see whether or not the results remained consistent.
The results showed that the pooled ORs are not signi cantly changed, indicating the stability of analytic results. The funnel plots are largely symmetric, indicating that there are no publication biases in the metaanalysis of the role of rs867186-GG genotype in SM (Fig. 4).

Discussion
The cell surface protein of EPCR was originally cloned in 1994 [27]. Multiple ligands of EPCR have been found, including PfEMP1, protein C (PC)/activated protein C (APC), factor VIIa, tissue factor, and a speci c variant of the T-cell receptor. Under physiological conditions, Protein C is activated by the thrombinthrombomodulin complex, and activated protein C (APC) then cleaves the protease-activated receptor (PAR1) at Arg46, which triggers an anti-apoptotic and anti-in ammatory reaction to inhibit thrombin To assess the effect of rs867186-GG polymorphism for the risk of SM, we analyzed genotyped patients with mild malaria and SM on rs867186-GG of previous publications. We found that the rs867186-GG genotype appears signi cantly more frequent in patients with mild malaria than those with SM (P=0.03 in Fig. 3). However, when we compared genotypes rs867186-GG versus rs867186-AA in malaria patients and healthy individuals, the frequency of rs867186-GG in malaria patients and healthy controls is very similar (p=0.90 in Fig. 2). This was not caused by studies bias, as the heterogeneity is calculated as I 2 =0 (Fig. 2), and this implies that we can take a xed-effect model to pool the ORs rather than a random effect model. The genotype difference of the GG and AA alleles of rs867186 is likely due to the pressure exerted by parasites of the genus Plasmodium that cause malaria. Thus, comparing healthy controls that lack these pressure with any disease status is unnecessary. This is a possible reason that these genotypes distinguish between mild and SM but not between healthy controls and malaria patients. The studies from Hansson et al. [1] and Schuldt et al. [25] reported that the genotype rs867186-GG does not have a protective role in malaria patients compared to healthy controls. However, they did not compare the difference in rs867186-GG genotype between SM and MM.
The metalloprotease cleaves the surface cellular EPCR, which releases a soluble EPCR (sEPCR) that circulates in the plasma [19]. The relationship between the polymorphisms in EPCR and plasma sEPCR levels has been conducted in patients with thrombosis to evaluate the risk of venous thrombosis by Medina et al. [39]. Their data indicate that individuals carrying some speci c genotypes have high sEPCR and APC levels, thus having a lower risk of venous thromboembolism [39]. However, no systemic analysis has been made in malaria patients regarding the EPCR genotype and plasma EPCR levels. When we examined the relationship between rs867186 polymorphisms and plasma sEPCR evaluating risk of SM, we found that the carriers of the rs867186-GG genotype have signi cantly higher sEPCR levels than those with the AG and AA genotypes in SM; carriers of the rs867186-AG genotype have signi cantly higher sEPCR levels than those with AA genotype in uncomplicated malaria, and carriers of the rs867186-GG genotype have signi cantly higher sEPCR levels than those with AG genotype in healthy individuals (Table 3). These results support that the rs867186 GG genotype is associated with elevated sEPCR levels in SM (

Conclusions
This meta-analysis summarizes previous observations and indicates the EPCR rs867186-GG genotypes are associated with increased soluble EPCR and may have an important role in SM. pRBCs bind to EPCR on endothelial cells and block APC's access to EPCR, inducing pro-apoptotic, pro-in ammatory, and loss of local vascular barrier integrity. Targeting EPCR-PfEMP1 binding can potentially reverse P. falciparum sequestration in host blood circulation, and in particular for sequestration in brain blood circulation, which could be an effective intervention to prevent SM. In the future, more studies are needed to assess the adhesive abilities of PfEMP1 expressed in isolation of pRBC from SM, CM patients, and even pregnant women with malaria. Furthermore, the mechanism of EPCR genotype protecting individuals against SM leads to an increased risk of thrombotic disease is also worthy of exploration.
Although we only found ve publications with a total of 2995 SM and 487 uncomplicated malaria, performing a meta-analysis is justi ed since we are working with a xed-effect. A summary based on two or more studies yields a more precise estimate of the true effect than either study alone [41].

Declarations
Authors' contribution LY and ML contributed substantially to the study design, data acquisition, analysis, and interpretation of data. RX and SG contributed substantially to the acquisition, analysis, interpretation of data, and performed the statistical analysis. LY and ML have been involved in the drafting process and critical revision of the article for important intellectual content. The corresponding authors had full access to all data and the nal responsibility for the decision to submit the article for publication. All authors read and approved the nal manuscript.

Funding
This study was supported by NIH grants U54MD007595 from NIMHD.

Availability of data and materials
The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.    Figure 1 Flow chart of the study. Among twenty-eight publications identi ed by the search method, twenty-three of those were excluded due to exclusion criteria of laboratory studies, non-original articles, lack of matched controls, or studies irrelevant to the current analysis. Five studies were included in this meta-analysis.  Forest plot for the protection of SM and EPCR polymorphism in the genotype of rs867186 GG vs. AG and AA. The frequency of rs867186-GG versus rs867186-AA and AG was analyzed in SM and uncomplicated/mild malaria patients. The frequency of rs867186-GG is higher in mild malaria than in SM patients. The pooled OR from 2 studies, including 892 patients with SM and 437 mild malaria, is shown (OR=0.36, 95% CI=0.14-0.91, P=0.03).

Figure 4
Funnel plot of publication bias in the meta-analysis of rs867186 genotype. The funnel plots are largely symmetrical, suggesting there are no publication biases in the meta-analysis of the role of rs867186-GG genotype in SM.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. PRISMAChecklist.pdf