Association of GCK gene DNA methylation with the risk of clopidogrel resistance in acute coronary syndrome patients

Abstract Backgrounds Clopidogrel resistance (CR), which was manifested as the failure of platelet inhibition in clopidogrel treatment, was likely to lead to cardiovascular events. Our study was aimed to explore the contribution of DNA methylation in glucokinase (GCK) to the CR risk. Methods Among 36 CR and 36 non‐CR acute coronary syndrome (ACS) patients, the platelet functions were evaluated by VerifyNow P2Y12 assay (turbidimetric‐based optical detection) and DNA methylation levels on two fragments of the CGI from the GCK were investigated through bisulfite pyrosequencing methods. In addition, the GCK mRNA expression was analyzed via quantitative real‐time PCR. Lastly, the logistic regression was employed to test the interaction between GCK methylation and nongenetic variables in CR patients. Results Subunit analysis showed that in male patients without DM but suffering from dyslipidemia, the increased methylation of cg18492943 indicated a risk of poor clopidogrel response (male, NCR vs CR(%): 84.86 ± 6.29 vs 88.16 ± 4.32, P = .032; without DM, NCR vs CR (%): 84.66 ± 6.18 vs 88.16 ± 4.17, P = .029; and dyslipidemia, NCR vs CR (%): 83.81 ± 6.96 vs 88.39 ± 4.74, P = .042).In addition, GCK mRNA expression was reduced in CR patients without DM. Moreover, regression analysis indicated that the values of platelet distribution width (PDW), total cholesterol (TC), and uric acid (UA) were correlated with the incidence of CR, and hypertension lowered the CR risk. Conclusions A higher methylation of cg18492943 in GCK gene would lower the expression of GCK mRNA, which might contribute to CR in patients without DM. Meanwhile, PDW and TC might be risk factors in CR.


| INTRODUC TI ON
The dual-antiplatelet therapy (taking orally aspirin together with ticagrelor or clopidogrel) is used to inhibit platelet activating and aggregating and has been the cornerstone treatment in acute coronary syndrome (ACS) patients after percutaneous coronary intervention (PCI). Although ticagrelor showed fast, consistent, and effective results among ACS patients, 1 the bleeding incidence was higher than clopidogrel. 2 Compared with treatment by aspirin and clopidogrel, a latest KAMIR-NIH research showed that aspirin together with ticagrelor or prasugrel displayed a similar risk of all-cause mortality but an elevating bleeding risk. 3 Hence, clopidogrel might be a better treatment in East Asian ACS patients.
Clopidogrel, by inhibiting P2Y12 receptor, could prevent adenosine diphosphate-induced platelets aggregation and lower the incidence of major adverse cardiovascular events (MACE) in ACS sufferers after PCI. 4 While different patients responded greatly to clopidogrel treatment. 5 Approximately 10%-40% of patients continue to manifest adverse cardiovascular risk. 6 This effect may be due to the failure of platelet inhibition responding to clopidogrel, a clinical phenomenon named clopidogrel resistance (CR). 7 Moreover, by VerifyNow P2Y12 assay, some researchers have noted that P2Y12 reaction units (PRU) higher than 240 could indicate the existence of CR. 8 But, the CR pathogenesis remains not clearly understood.
Many extrinsic factors could influence platelet activity after taking clopidogrel, such as comorbidities, drug interactions, and smoking. 9 Among these factors, diabetes mellitus (DM) may be an important clinical factor attributed to platelet dysfunction. 10 The mechanisms involved in hyperreactive platelets in DM patients have been suggested to be insulin deficiency, hyperglycemia, metabolic conditions, and cellular abnormalities. 11 The glucokinase gene (GCK), sited on 7p15.3-p15.1, codes the glucose enzyme. 12 Through catalyzing the rate-limiting step of insulin secretion, GCK participates in balancing glucose homeostasis and glycogen synthesis. 13 Many studies have indicated that GCK plays a vital role in diabetes. Through systemic analysis of gene expression profiles, one study revealed that the GCK gene played a role in obese type 2 diabetes. 14 Another study investigated the effect of DNA methylation on obesity and type 2 diabetes mellitus and found that DNA hypermethylation could lead to a decline in hepatic GCK expression and was correlated with susceptibility to diabetes. 15 Additionally, if GCK DNA methylation changes, then glucose homeostasis in F1 offspring would be disrupted after maternal bisphenol A exposure. 16 Although our former studies showed that the DNA methylation of P2Y12 17 and PON1 18 might be associated with CR, the effect of GCK DNA methylation on CR is not clearly understood.
In this study, in ACS patients with clopidogrel treatment, we attempted to explore whether the selected CpG islands methylation in GCK is associated with CR.

| Samples
Seventy-two ACS patients were picked up in Ningbo No. 1 Hospital from September 2012 to December 2018. These patients, who complained of chest pain, were diagnosed by physical examination, laboratory tests, and coronary angiography. The followings were our inclusion criteria: (a) Based on recent ESC guideline, these ACS patients underwent PCI through drug-eluting stent; (b) before PCI, the patients received 300 mg clopidogrel and 300 mg aspirin as loading doses, and they were subsequently administered 75 mg clopidogrel and 100 mg aspirin daily as a maintenance doses; and (c) the patients were without aspirin low response (aspirin reaction units <550). And our exclusion criteria were shown as below: (a) chronic heart failure; (b) acute infection; (c) rheumatological disorders; (d) history of malignancy; (e) history of active bleeding; (f) abnormal hepatic or kidney function; (g) concomitant treatment using warfarin or IIb/ IIIa inhibitors (Tirofiban); and (h) platelets <150 000 μL or more than 500 000 μL.
Ethical approval was obtained for human sample collection from the Ethics Committees at Ningbo No. 1 Hospital, and all selected patients offered their written informed consent. The study protocol conformed to the principles outlined in the Declaration of Helsinki.

| Biochemical analyses and platelet function measurements
Blood samples from 36 CR and 36 NCR patients were collected, and their biochemical values, such as LDL, GLU, HbA1c, and BUN, were tested based on a standard process according to the manufacturer, and the raw data were collected.
The 2016 ACC/AHA guideline 19 pointed out that the ACS patients, who would more likely to suffer the risk of CVD after PCI, were suggested to access platelet function. Thus, in this study, the patient platelet function was measured after 1 month since PCI.
Through VerifyNow P2Y12 assay (Accumetrics Inc), platelet function was assessed to evaluate the responsiveness to P2Y12 antagonists. 20 This turbidimetric-based optical detection system evaluated platelet aggregation as an increase in light transmittance from whole blood. The result was informed as PRU. And the PRU greater than 240 indicated the existence of clopidogrel resistance. 8

| Level of GCK mRNA determination
Using a RNeasy Plus Universal Kit (QIAGEN), RNA was separated from blood samples. One microgram of RNA along with the PrimeScript ™ RT Reagent Kit and gDNA Eraser (Takara Bio) was used to generate cDNA. Then, after diluting the template cDNA, the relative expression of GCK mRNA was measured by the ABI 7500 qRT-PCR System (Applied Biosystems). GAPDH was applied to normalize GCK expression. The relative quantitative was implemented via the comparative CT method (ΔΔCt). The amplification primers involving in qRT-PCR were schemed by Prime 5 software, and the sequences are shown in Table 2. Multiple linear regression was used to determine the effect between GCK DNA methylation and clinical features. Logistic regression was utilized for analyzing correlations between GCK methylation and biochemical elements among ACS patients suffered from clopidogrel resistance. A two-sided P value, which was < .05, was regarded as statistically significant.

| Patient characteristics
From May 2010 to December 2018, a total of 72 CAD patients who met the above requirements were recruited in this study. Among these patients, 36 were considered CR. The clinical baselines of CR and NCR patients are presented in Table 3. Except for uric acid (UA), the other clinical indexes were well matched. This result indicated that higher uric acid might increase the CR risk.

| The association of CR and GCK DNA methylation
In the present research, according to the reference of our 850 k methylation chip, we chosecg01700200 (chr7:44227705-44227975) and cg18492943 (chr7:44203220-44203435) on GCK for bisulfite pyrosequencing. All of the genes were located on the gene body. We evaluated the association of the GCK DNA methylation between CR and NCR patients. As displayed in Figure 1 and Table 4, in targeted fragments, the methylation levels in cg01700200 and cg18492943 were not significantly interrelated with clopidogrel poor response.
Then, by different clinical baseline characteristics, we carried out a subunit analysis to estimate whether the GCK DNA methylation (cg01700200 and cg18492943) were associated with CR.
We observed that in male patients without DM but suffering from  Table 5).

| The association of GCK mRNA expression and clopidogrel resistance
Through qRT-PCR, we assessed the relative mRNA expression in GCK to determine that the various GCK expression levels would influence on clopidogrel response. Unfortunately, the result was insignificant. While in a subgroup of patients without DM, we observed that GCK mRNA expression was decreased when patients were suffering from clopidogrel poor response (Figure 3).

| Regression analysis
For various clinical factors that might influence DNA methylation, we investigated the effect of these variables on GCK gene methylation through multiple linear regression, and we discovered that uric acid might be related to the methylation of cg18492943 (F = 1.092, P TA B L E 1 Primers for cg01700200 and cg18492943 CpG island loci analysis Moreover, in consideration of genetic and nongenetic factors influenced on clopidogrel response, we implemented a logistic regression analysis to assess the association between CR and these confounding factors. The results revealed that these variables (such as cg18492943, age, AST, TC, and PDW) were correlated with CR, while hypertension was protective factors of CR (Table 6).
Furthermore, since we performed logistic regression analysis in patients without DM, male, and dyslipidemia subgroups, the results indicated that in patients without DM, cg18492943, TG, TC, BUN, and PDW showed the risk of poor clopidogrel response, while cg01700200 was a protective factor (Table 7). However, the logistic regression analysis in other subgroups did not show any significant outcomes. Except for the GCK gene, other aspects might also influence glucose regulation and the response to clopidogrel. 11 Among these aspects, deficient insulin action is one of the most cardinal factors contributing to platelet dysfunction. 35 Platelets express insulin receptors and insulin-like growth factor-1 (IGF-1) receptors. 36   suffer from a risk of ischemic events. 39 IRS-independent pathways contribute to platelet hyperreactivity due to impaired responses to nitric oxide (NO) and prostacyclin, which are related to CR. 40 In our study, GCK, influenced by DNA methylation, could affect platelet activity in response to clopidogrel, which was also involved in insulin deficiency. However, the exact mechanisms of insulin action and CR

| D ISCUSS I ON
have not been well established, and future discoveries will help us to understand the possible molecular factors and signaling pathways.
In addition to genetic factors, various extrinsic elements (drug interactions, comorbidities, and environment) may also result in CR. Our subgroup analysis and logistic regression analysis in present study discovered that hypertension decreased the CR risk, and cg18492943, PDW, TC, and UA may be positively associated with clopidogrel resistance. First, platelet distribution width (PDW) could manifest morphological changes in platelets, and platelet size was suggested as an indicator of enhanced reactivity; thus, the higher PDW might potentially increase the risk of complications after stenting. 41 Another study demonstrated that the ratio of platelet to red cell distribution width (P-RDW) was lower in high on-treatment platelet reactivity (HPR) patients. 42 As a result, an assessment of the role of PDW in planning antiplatelet therapy is warranted.
Second, we discovered that total cholesterol was a protective factor. This finding was similar to a recent study, which also reported that a poorer clopidogrel response was significantly associated with To sum up, our study suggested that the higher methylation of cg18492943 in GCK gene could lead to a decline in GCK mRNA expression and potentially cause CR in patients without DM. Additionally, subgroup analysis and logistic regression analysis showed that the values of PDW, TC, and UA were interrelated with the incidence of CR, and hypertension lowered the CR risk. However, other advanced and effective planning studies with larger samples would facilitate the validity of our findings and the elucidation of CR pathogenesis.

ACK N OWLED G M ENT
We thank Dr Jin Yang for excellent technical assistance and editing the article.