Association between Platelet-Specific Collagen Receptor Glycoprotein 6 Gene Variants, Selected Biomarkers, and Recurrent Pregnancy Loss in Korean Women

This paper investigates whether glycoprotein 6 (GP6) gene polymorphisms are a risk factor for recurrent pregnancy loss (RPL) in Korean women. Genotypes were determined by polymerase chain reaction-restriction fragment length polymorphism and real-time polymerase chain reaction amplification. We identified five polymorphisms in the GP6 gene: rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A. GP6 rs1654410 CC was associated with decreased RPL risk (adjusted odds ratio = 0.292, 95% confidence interval = 0.105–0.815, p = 0.019), and recessive genotypes were also significantly associated with decreased RPL risk (adjusted odds ratio = 0.348, 95% confidence interval = 0.128−0.944, p = 0.038). GP6 rs1654419 GA was associated with decreased RPL risk (adjusted odds ratio = 0.607, 95% confidence interval = 0.375-0.982, p = 0.042), and dominant genotypes were significantly associated with decreased RPL risk (adjusted odds ratio = 0.563, 95% confidence interval = 0.358−0.885, p = 0.013). Altogether, the genotype frequencies of GP6 rs1654410 T>C and GP6 rs1654419 G>A were significantly different between RPL patients and control participants. Therefore, although GP6 polymorphisms may be useful as biomarkers of RPL, additional studies with heterogeneous cohorts are required to better understand the influence of GP6 and assess its performance as a biomarker.


Introduction
Recurrent pregnancy loss (RPL) is defined as two or more consecutive pregnancy losses [1]. Factors that contribute to the etiology of RPL include advanced maternal age, maternal anatomic anomalies, placental anomalies, chromosome abnormalities, endocrine dysfunction, antiphospholipid syndrome, hereditary thrombophilia, psychological trauma, and environmental factors, such as smoking, excessive alcohol consumption, and stress [2]. Moreover, the likelihood of pregnancy loss is 5% higher for women
For RFLP analysis of the SNPs, the PCR products for GP6T>C rs1654410, GP6T>G, rs1671153, GP6G>Ars1654419 were digested with the restriction enzymes Mbo II, Hph I, and Cse I, respectively. To confirm the three SNPs and validate the RFLP results, 10-20% of the samples were randomly selected, used for a second round of PCR, and analyzed by DNA sequencing using an automatic ABI3730xL DNA analyzer (Applied Biosystems, Forster City, CA, USA). Samples were genotyped by real-time polymerase chain reaction (PCR) using a qPCR kit with the primers listed in Supplementary  Table S1. The conditions for the five polymorphisms (rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A) are also listed in Supplementary Table S1.
2.3. Assessment of Plasma Plasminogen Activator Inhibitor-1 (PAI-1), Homocysteine, Total Cholesterol, Uric Acid, and Blood Coagulation Status Plasma PAI-1, homocysteine, total cholesterol, and uric acid were measured in blood collected from RPL patients. Plasma was separated by centrifuging whole blood samples at 1000 × g for 15 min. PAI-1 levels were determined using a human serpin E1/PAI-1 immunoassay (R&D Systems, Minneapolis, MN, USA). Uric acid and total cholesterol were measured using commercially available enzymatic colorimetric tests (Roche Diagnostics, GmbH, Mannheim, Germany). Homocysteine was measured using a fluorescence polarization immunoassay using the Abbott IMx analyzer (Abbott Laboratories, Abbott Park, IL, USA).

Statistical Analysis
Differences in the frequency of GP6 polymorphisms between the control and patient groups were assessed using Fisher's exact test and a logistic regression model. Odds ratios (ORs), adjusted odds ratios (AORs), and 95% confidence intervals (CIs) were used to examine the association between GP6 polymorphisms and RPL risk. The data are presented as mean ± standard deviation (SD) for continuous variables or as percentages for categorical variables. Statistical analyses were carried out using MedCalc version 12.1.4 (MedCalc Software bvba, Mariakerke, Belgium) or GraphPad Prism 4.0 (GraphPad Software, Inc., San Diego, CA, USA). The HAPSTAT program (v.3.0, www.bios.unc.edu/~{}lin/hapstat/) was used with a strong synergistic effect to estimate the frequency of polymorphic haplotypes, p-values of <0.05 were considered statistically significant. The false discovery rate (FDR) was also used to adjust for multiple comparisons; associations with an FDR-corrected p-value of <0.05 were considered statistically significant [16]. Genetic interaction analysis was performed with the open-source multifactor dimensionality reduction (MDR) software package (v.2.0) available from www.epistasis.org. The MDR method consists of two main steps [17][18][19]. During ANOVA analysis, the Kruskal-Wallis test was used for small sample sizes and/or when the P-value of Levene's test was less than 0.05 ( Figure 1).

Results
The clinical profiles and demographic characteristics of RPL patients and control participants are presented in Table 1. The mean ages of RPL and control patients were 32.75 ± 3.84 and 33.20 ± 4.54 years, respectively. When comparing the RPL patients with the control participants, the RPL patients had significantly higher platelet counts (PLT), prothrombin times (PT), activated partial thromboplastin times (aPTT), luteinizing hormone (LH) levels (mIU/mL, mean ±SD), and estradiol (E2) levels (pg/mL, mean ±SD) (all p < 0.05; Table 1). However, many of the factors measured here undergo changes during the course of pregnancy.
Since environmental factors are known to contribute to RPL, we also investigated the interaction of clinical characteristics and GP6 genotypes. Logistic regression analysis was used for the association of each SNP with RPL risk adjusted by environmental factors. Table 4 shows that the GP6 rs1654419 GG + GA genotype increased RPL risk when the following characteristics were present: platelet count of ≥242.11 × 10 3 (AOR = 4.461; 95% CI = 2.234-8.905), follicle-stimulating hormone (FSH) level < 8.13 mIU/mL (AOR = 2.716; 95% CI = 1.458-5.057), and E2 level < 26.00 pg/mL (AOR = 2.322; 95% CI = 1.134-4.755). Thus, the GP6 rs1654419 genotype may present an increased risk of RPL. However, these factors measured here undergo changes in the course of pregnancy.  Haplotype analyses were conducted to further assess the association of RPL occurrence with the five polymorphisms (Supplementary Table S2). Several haplotypes were found to be significantly associated with the incidence of RPL after FDR correction. Furthermore, the results from genotype combination analyses were significant for each allele, and these results were related to RPL incidence (Supplementary Table S3).
When the haplotype analysis was limited to four polymorphisms, we found that RPL was significantly associated with each allele (Supplementary Table S4). Limiting the haplotype analysis to three polymorphisms revealed that three variants (GP6 rs1654410, GP6 rs1671153, GP6 rs1654419) were the leading genetic risk factors related to RPL (Supplementary Table S5). Since clinical factors, including aPTT, PAI-1, PT, and homocysteine, have been associated with RPL, we investigated the association of these clinical variables with GP6 gene variants in RPL patients who had suffered three or more pregnancy losses. We found that aPTT, PAI-1, PT, homocysteine, and PAI-1 levels were significantly associated with these RPL patients (Supplementary Tables S6-S8). Furthermore, Figure 1 depicts the analysis of variance for aPTT, PAI-1, PT, and homocysteine levels according to GP6 polymorphisms (RPL ≥ 3). Our association data are statistically quite weak when comparing to the normal genome-wide association study (GWAS) significance threshold of 5 × 10 −8 .

Discussion
In this study, we selected five GP6 polymorphisms (rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A) that were candidate biomarkers for RPL risk.
Then, we investigated the association between these polymorphisms and RPL prevalence. We found that GP6 polymorphisms were associated with an increased risk of RPL and that the GP6 haplotypes rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A occurred significantly more frequently in women with RPL. The haplotype analysis of the five genetic markers revealed that five haplotypes (GP6 rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A; [p < 0.001; FDR-corrected p < 0.001 OR, 19.13; 95% CI, 1.142-320.200; p = 0.0002; FDR-corrected p = 0.001] were significantly associated with increased RPL susceptibility. Many GP6 polymorphisms have been identified, and previous studies have demonstrated that genetic changes in GP6 have an effect on platelet function. Platelets are small, acellular cells that aggregate to form blood clots on blood vessels and are essential for hemostasis. As a platelet membrane protein, GP6 is generally accepted to be produced by platelet activation, adhesion, and aggregation. Interestingly, the importance of the genetic diversity of the GP6 gene for platelet aggregation was highlighted by a genome-wide meta-analysis conducted by Johnson et al. [20]. Moreover, a previous study identified several SNPs (rs1671153, rs1654419, and rs1613662) in patients with sticky platelet syndrome (SPS) [21] and a history of miscarriage (rs1671152, rs1654433, rs1613662, rs1654416, and rs2304167) [22]. In addition to their normal role in hemostasis, platelets can cause arterial thrombosis by blocking the collateral arteries after a collagen-rich atherosclerotic plaque has ruptured, resulting in heart attack and stroke [23]. After an atherosclerotic plaque rupture, the relationship between GP6 and thrombus GP6-specific signaling plays a major role in platelet adhesion to activated endothelium [24,25] and thrombus formation [26], whereas platelet GP6 mediates adherence to the endothelial matrix [27].
GP6 is a receptor that contains an immunoglobin-like domain, is structurally and functionally similar to an immunoreceptor, and is exclusively expressed in platelets and platelet precursors called megakaryocytes. Although it is unknown how thrombotic formation through GP6 can occur so rapidly, it is known that various agonists (collagen, thrombin, and ADP) activate GP6 on the platelet membrane to induce binding to fibrinogen and consequently promote the aggregation of platelets and granule release [28,29].
In addition, collagen activates platelets by promoting the phosphorylation of phospholipase C-γ2 (PLC-γ2) through GP6 [30]. Platelets that circulate along the bloodstream respond to collagen that is exposed at the wound site on the vessel wall via GP6 recognition and signaling. Collagen is a typical extracellular matrix that is used to form a thrombus, which is exposed to the bloodstream when a wound on the blood vessel wall develops, thus forming a platelet-attachable surface in the artery and acting as the main stimulant for platelet activation. Therefore, achieving a better understanding of this process may lead to the development of treatment modalities that inhibit or prevent thrombus formation. We predict that these mechanisms will be important in RPL patients based on the association with GP6.
In the present study, polymorphic GP6 genotype frequencies rose significantly as the frequency of pregnancy loss increased. Moreover, GP6 genotypes and haplotypes were associated with known contributors to increased blood coagulation, including elevated plasma PAI-1 and BMI, lower PT and aPTT, and plasma concentrations of vascular risk factors (homocysteine, FSH, and total cholesterol) in RPL. We anticipate that these GP6 genotype combinations contribute to the incidence of RPL in Korean women. Alterations in the fibrinolysis cascade that cause hypo-fibrinization or hyperfibrinolysis may interfere with the placenta and result in poor pregnancy outcomes. Defects in this process can negatively affect trophoblast transplantation as well as the placenta and may result in RPL. Moreover, higher PLT indices have been reported to increase the risk of thrombosis. Previous reports also suggest that hormones, including FSH, LH, and E2 are involved in RPL, since FSH, LH, and E2 levels were elevated in the RPL group compared with the control group. The control group consisted of women who had abortions with known causes, such as uterine septum and parental chromosomal abnormalities. However, in hemostasis, factor V Leiden mutation, prothrombin G20210A variant, the decreased activity of antithrombin, protein C, protein S, the increased coagulation factor VIII activity, dysfibrinogenaemia, fibrinolysis abnormalities, antiphospholipid syndrome, and the detection of plasma levels of substances released from platelets factors related to thrombophilia should also be investigated.
Reports on the relationship between GP6 and pregnancy loss are very limited. In addition, the biological function of GP6 pregnancy is not well known. Genetic mutations in the blood coagulation factor cause a prethrombotic condition through deficiencies of blood coagulation inhibitors, the overproduction of pre-coagulation proteins, abnormalities in fibrinolysis, and damage to the vascular endothelium. As the diagnosis of neonatal thromboembolism increases, the role of genetic risk factors becomes more important.
GP6 is a platelet transmembrane glycoprotein that plays a significant role in collagen-initiated signal transduction and platelet pro-coagulant activity; therefore, the observed variation in the GP6 gene region may influence risk for thromboembolic disorders [44].
A previous report shows that a significantly higher occurrence of mutant genotypes of GP6 SNPs, namely rs1671153, rs1654410, rs1654419, and rs1613662, in recurrent miscarriages (RM) cases, suggesting a risk association for pregnancy loss [45]. Another study has shown risk associations of mutant genotypes at rs1671153, rs1654419, and rs1613662 SNPs with thrombotic disorders [46].
Patients of pregnancy loss demonstrate reduced platelet function [47]. The GP6 plays a critical role in platelet adhesion and activation, and GP6 has a major role in collagen-induced platelet signaling [45]. It is thought that changes in platelet adhesion and activation by GP6 polymorphism (rs1654410T>C, rs1654419 G>A, and rs1654431 G>A) would have influenced platelet reactivity toward collagen and therefore influence platelet function and the maintenance of pregnancy.
Based on the present study, we suggest that the genotype combinations of GP6 polymorphisms rs1654410/GP6 rs1671153 (CC/TT) may contribute to the diagnosis of RPL in Korean women. Therefore, additional studies are needed to clarify the association between GP6 polymorphisms and RPL.
MicroRNAs (miRNAs) are associated with platelet reactivity; however, there is a lack of information regarding the role of miRNAs in megakaryocyte signaling cascades, and miRNAs are not known to regulate collagen-induced GP6 signaling. One previous report bioinformatically predicted that miR-15a-5p targets were expressed in megakaryocytes and subsequently enriched target genes that are also known to be downstream targets of platelet GP6 signaling [48]. These findings indicate that miR-15a-5p regulates the potential master-miRNA GP6-mediated megakaryocyte signaling and platelet activation. Therefore, the present study provides a basis for future research associated with miRNA in RPL patients.
It is not sufficient to make recommendations for RPL patient management based on our research results alone. Our results indicate that several polymorphisms are associated with clinical variables in RPL patients: GP6 rs1671153 T>G, rs1654419 G>A, and rs12610286 A>G were associated with higher homocysteine levels, elevated creatinine levels, and PAI-1, respectively. Therefore, GP6 polymorphisms may contribute to RPL and are potential biomarkers for assessing RPL risk. However, our study does have various limitations, which are outlined in the Discussion, and efforts to overcome these problems are necessary. Our chosen method of liquid biopsy has been recently investigated in depth, and we therefore plan to continue research to prevent platelet-related diseases by analyzing GP6-related microRNA, non-coding RNAs, and exosomes; in this regard, the present study provides a basis for future research associated with miRNA in RPL patients.
Although we found an association between GP6 polymorphisms and RPL, there are some limitations to our study: lack of a survey for vascular risk factors, clinical insufficiency from control participants, and the relatively small sample size of the control and RPL group, which highlights the need for additional studies in larger patient populations. Mechanism by which GP6 polymorphisms affect the onset of RPL remains unclear. Additional information regarding risk factors for RPL patients is lacking, and additional research is needed. The population of this study was restricted to Korean patients.

Conclusions
In conclusion, we identified associations between GP6 gene polymorphisms (rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A) and RPL prevalence in Korean women. Our findings suggest that GP6 polymorphisms may contribute to RPL and are potential biomarkers for assessing RPL risk. For example, several polymorphisms were associated with clinical variables in RPL patients: GP6 rs1671153 T>G, rs1654419 G>A, and rs12610286 A>G were associated with higher homocysteine levels, elevated creatinine levels, and PAI-1, respectively. Moreover, the haplotype frequencies of GP6 rs1654410 T>C, rs1671153 T>G, rs1654419 G>A, rs12610286 A>G, and rs1654431 G>A were significantly different between RPL patients and control participants. Together, these results highlight the need for large and heterogeneous genetic studies to confirm the present findings and to validate potential biomarkers of RPL for prevention and prognosis.