Polymorphisms in the GCK gene increase the risk of clopidogrel resistance in stable coronary artery disease (SCAD) patients

ABSTRACT

Background

Diabetes mellitus is a major factor in clopidogrel resistance (CR), and the glucokinase (GCK) gene plays a pivotal role in glucose homeostasis. This study investigated the contribution of GCK polymorphisms to CR risk.

Methods

Two hundred SCAD patients were recruited, and their platelet functions were detected by the Verify-Now P2Y12 assay. The polymorphisms of GCK were tested based on the methods of polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP). We investigated the associations of GCK polymorphisms and CR. Multivariate logistic regression was performed to analyse the correlations between GCK polymorphisms and clinical values.

Results

Our study found that the SNPs rs4607517 and rs6975024 were associated with CR. Additionally, patients with the G allele of rs4607517had a greater CR risk, but the C allele of rs6975024 might be a protective factor. Finally, logistic regression revealed that CC + TC (rs6975024) as well as the values of albumin were correlated with a decreased risk of CR, and higher levels of uric acid (UA) may be positively associated with CR.

Conclusion

The GCK gene polymorphisms might increase the CR risk in SCAD patients. Meanwhile, higher albumin levels and lower UA values might decrease the risk.

KEYWORDS:

• Polymorphisms
• coronary artery disease
• platelet aggregation inhibitors

1. Background

Clopidogrel, a novel antiplatelet medicine in the treatment of coronary artery disease (CAD) patients after PCI, could prevent major adverse cardiovascular events (MACE) incidence by inhibiting P2Y12 receptor and lowering platelet aggregation [1]. While the clopidogrel pharmacodynamic response varies greatly in different patients, the clinical phenomenon of the failure in platelet inhibition after clopidogrel administration is called clopidogrel resistance (CR) [2]. Although the pathogenesis of CR remains unclear, various extrinsic and genetic elements may result in the variability of clopidogrel antiplatelet effect [3]. Among them, diabetes mellitus (DM) was identified as a major predictive factor for CR [4].

It was reported that patients with DM suffer an impaired response to clopidogrel treatment, which is due to the attenuation of the clopidogrel PK profile and an altered functional status of the P2Y12 signalling pathway [5]. Hyperglycaemia, insulin deficiency, cellular abnormalities, and metabolic conditions had been suggested to be the mechanisms in DM patients with hyperreactive platelets [6]. A large number of researches have indicated that the glucokinase (GCK) gene plays a pivotal role in DM. GCK (located at 7p15.3-p15.1) encodes a glucose enzyme and catalyses the committed step of insulin secretion, balances glucose homeostasis and regulates glycogen synthesis [7]. Our former study revealed that the higher methylation of cg18492943 would lower the GCK mRNA expression, leading to high platelet activity in non- DM patients [8].

In addition to DNA methylation, gene polymorphisms might be another vital factor that influences biological function. There are many genetic variants of GCK, such as rs4607517 and rs6975024. Some studies have pointed out that the GCK rs4607517 polymorphism is associated with the consumption of sweets in gestational DM [9], and rs4607517 is significantly associated with the progression to a decline in insulin sensitivity [10]. However, the effects of the rs4607517 and rs6975024 polymorphisms in GCK concerning CR are poorly comprehended. Hence, in this research, we attempted to evaluate whether polymorphisms of rs4607517 and rs6975024 are involved in CR in SCAD patients administered with clopidogrel.

2. Material and methods

2.1. Study population

From September 2018 to December 2019, two hundred stable CAD patients treated at Ningbo No. 1 Hospital were selected in our investigation. The patients were enrolled according to the following criteria: (1) based on recent European Cardiovascular Society (ESC) guidelines, the SCAD patients were diagnosed by coronary angiography and underwent PCI by drug-eluting stent; and (2) the patients were taking 300 mg clopidogrel as well as 300 mg aspirin as loading-doses before stenting and 75 mg clopidogrel as well as 100 mg aspirin daily as maintenance-doses. The exclusion criteria were as follows: (1) hepatic or kidney malfunction; (2) chronic heart failure; (3) rheumatological disorders; (4) acute infection; (5) history of cancer; (6) concomitant treatment with IIb/IIIa inhibitors (tirofiban), warfarin or new oral anticoagulants (NOACs); and (7) platelets <150,000 μL or more than 500,000 μL.

The ethical approval was obtained from the Ethics Committee at Ningbo No. 1 Hospital, and all selected patients provided their informed consents. The whole research protocol was conformed to the principles of Helsinki Declaration.

2.2. Clinical data collection and platelet function measurements

The clinical information of the patients, such as age, gender, and comorbidity, was collected from the case history system. Their biochemical values, including HbA1c, hsCRP, GLU and so on, were measured based on a standard process. All raw data were entered into a computer database.

Since the patients were administered clopidogrel, their platelet reactivity was stable at the 30th day since PCI operation. Hence, platelet-function was evaluated at this time [11]. The samples of peripheral blood were collected by the double-syringe technique. To avoid spontaneous platelet activation, the first 2 to 4 mL of free-flowing blood was discarded. And then, platelet functions were measured through Verify-Now P2Y12 assay (Accumetrics Inc., San Diego, CA). The residual platelet reactivity (RPR) cutoff value greater than 240 reaction units indicated the existence of CR [12].

2.3. DNA extraction and polymorphism genotyping

By the DNA Blood Mini Kit (Qiagen, Hilden, Germany), the genomic DNA were extracted from 3 ml peripheral blood of patients, and these DNA samples were preserved at -20°C refrigerator until the next step. Based on PyroMark Assay Design software, the primers sequences for SNP genotyping were synthetized.

Polymerase chain reaction (PCR) amplification was performed by ABI-7000 (Applied Biosystems, U.S.A.) as follows: 50 μl reaction volume containing 1 μl of template DNA, 1.5 μl of 10 mM dNTP, 5 μl of Taq Buffer, 1.0 μl of 25 mM MgCl2, 1.5 μl of upstream and downstream primers, 1 μl of platinum Taq polymerase 1 U and 37.5 μl of water. The PCR cycling contains 2 cycles of denaturation at 95°C for 3 min, 32 cycles of denaturation at 95°C for 35 s, followed by annealing at 55°C for 48 s, extension at 65°C for 3 min, and 1 cycle of repair extension at 65°C lasting for 9 min. Then, the PCR products would be purified, measured and sequenced. After Msp I was digested, the alleles were detected by agarose gel electrophoresis, and the PCR-RFLP results were observed and recorded in UV transmittance apparatus.

2.4. Statistical analysis

A series of association analyses were performed between rs4607517 and rs6975024 polymorphisms and clinical characteristics and CR. Categorical variables are shown as mean ± standard deviation (m ± SD)and were compared by either Fisher’s exact test or χ² test. Moreover, if continuous variables were not normally distributed, they would be described as median with interquartile ranges (IQRs) and compared by non-parametric tests, and those that were normally distributed are expressed as m ± SD and were assessed with t test. Multiple logistic regression was performed to analyse correlations between polymorphisms in GCK and the clinical values of SCAD patients suffering from CR. All statistical analyses were operated by the software of PASW Statistics 20.0 (SPSS, Inc. Chicago, IL, U.S.A.). And if two-sided P values were less than 0.05, we considered the results were of statistical significance.

3. Results

3.1. Patients’ baseline characteristics

From September 2018 to November 2019, a sum of 200 SCAD patients who met the inclusion criteria was recruited for our research. The study population was all Chinese Han population. Among them, 96 were considered as CR according to the platelet function measurements. The demographic and baseline characteristics of these 200 patients were summarized in Table 1. Besides age, albumin, creatinine, and uric acid, other nongenetic values were well matched. The results manifested that older age, lower albumin levels, and higher creatinine and uric acid levels might increase the exposure of CR.

Table 1. Comparison of characteristics between CR and NCR patients.

Index	Non-CR (n = 114)	CR (n = 96)	χ^2/z/t	P value
Age (years)	59 (51.2, 68.15)	67 (55.1, 72.95)	11.098	0.002
BMI (kg/m^2)	23.71 (20.62, 26.91)	23.98 (22.87, 25.26)	0.275	0.623
Total cholesterol (mg/dL)	4.39 (3.79, 5.68)	4.18 (3.74, 4.98)	0.767	0.367
Triglyceride (mg/dL)	1.34 (1.07, 2.23)	1.42 (0.93, 1.63)	2.259	0.144
HDL (mg/dL)	0.93 (0.78, 1.11)	0.99 (0.76, 1.21)	1.337	0.256
LDL (mg/dL)	2.56 (2.03, 3.51)	2.36 (1.95, 2.99)	1.248	0.257
Fasting blood-glucose	5.16 (4.67, 5.81)	5.37 (4.61, 7.16)	1.797	0.185
ALT (umol/L)	26.1 (16.6, 47.2)	23.5 (15.1, 42.4)	0.956	0.337
AST (umol/L)	25.2 (18.7, 135.2)	26.1 (18.2, 77.5)	0.179	0.675
Albumin (g/L)	40.3 (36.7, 43.1)	38.2 (36.1, 40.1)	12.867	<0.001
BUN (mmol/L)	5.21 (4.06, 6.75)	5.52 (4.85, 6.64)	1.359	0.250
Cr (μmol/L)	67.3 (63.2, 73.8)	76.5 (64.5, 85.6)	10.499	0.012
UA (ummol/L)	278 (152, 333)	333 (279, 409)	19.812	0.003
hsCRP (mg/dL)	1.84 (0.99, 8.47)	3.32 (0.61, 10.31)	0.204	0.878
HbA1c (%)	5.9 (5.6, 6.3)	5.7 (5.5, 6.7)	0.079	0.782
Gender (male)	27	69	1.074	0.300
Hypertension, n	27	69	1.463	0.226
Diabetes mellitus, n	78	18	0.754	0.385
Concomitant medication
Pantoprazole	114	96	—	—
Statin	114	96	—	—

3.2. Rs4607517 and rs6975024 polymorphisms and clopidogrel resistance

rs4607517 and rs6975024 were common variances in GCK. We calculated the frequency of these two SNPs and ensured that both were in Hardy–Weinberg equilibrium. The allele frequencies of rs4607517 and rs6975024 were as followings: rs4607517, G vs. A: 0.754 vs. 0.246; rs6975024, C vs. T: 0.302 vs. 0.698.

Via PCR-RFLP, our group investigated the associations of CR with rs4607517 and rs6975024. As shown in Table 2, our exploration revealed that rs4607517 and rs6975024weresignificantly related to CR (P_rs4607517 = 0.010; P_rs6975024 = 0.004).

Table 2. Comparison of the SNPs in rs4607517 and rs6975024 between CR and NCR patients.

SNP ID	Non-CR (n = 114)	CR (n = 96)	χ^2	P value
rs4607517
GG	56	67	9.312	0.010
AG	48	23
AA	10	6
rs6975024
CC	10	10	11.226	0.004
TC	59	28
TT	45	58

Additionally, we assessed the relationship by comparing different SNP models (homozygous and dominant models, recessive and heterozygous models).As shown in Table 3, for rs4607517, patients carried with G allele were more likely to get a higher CR risk (P = 0.006), and for rs6975024, the C allele might be a protective factor against CR.

Table 3. Comparison of the different SNP models between CR and NCR patients.

	Non-CR	CR	P value		Non-CR	CR	P value
rs4607517				rs6975024
GG vs.	56	67	0.003	CC vs.	10	10	0.505
AG + AA	58	29		TC + TT	104	76
GG + AG vs.	104	90	0.492	CC + TC vs.	69	38	0.003
AA	10	6		TT	45	58
GG vs.	56	67	0.201	CC vs.	10	10	0.604
AA	10	6		TT	45	58
G vs.	160	157	0.006	C vs.	79	48	0.032
A	68	35		T	149	144

Note: Bold values represent that the results were of statistical significance.

3.3. Subgroup analysis

Due to the differences in CR and NCR patients’ baseline characteristics, our group implemented subgroup analyses in line with age, albumin, CREA, and UA. As shown in Table 4, for rs4607517, if patients were younger than 60, suffered from renal inadequacy, and did not have hyperuricaemia, the polymorphism of rs4607517 was correlated with the risk of CR. Meanwhile, for rs6975024, if patients were older than 60, suffered from hypoalbuminemia and renal inadequacy, and did not have hyperuricaemia, the polymorphism of rs6975024 would affect the clopidogrel antiplatelet effect (Table 5).

Table 4. Comparison of GCK rs4607517 in the subgroups between CR and NCR patients.

Index		Genotype	Non-CR	CR	X^2	P value
Age	<60	GG	27	20	6.435	0.040
		AG	24	4
		AA	6	3
	≥60	GG	29	47	3.880	0.144
		AG	24	19
		AA	4	3
Albumin	<40	GG	31	54	8.091	0.017
		AG	14	14
		AA	10	3
	≥40	GG	25	13	9.044	0.011
		AG	34	9
		AA	0	3
CREA	<97	GG	53	57	5.566	0.062
		AG	38	19
		AA	10	6
	≥97	GG	3	10	6.312	0.012
		AG	10	4
		AA	0	0
UA	<428	GG	48	63	13.996	0.001
		AG	45	20
		AA	10	3
	≥428	GG	8	4	4.295	0.117
		AG	3	3
		AA	0	3

Table 5. Comparison of GCK rs6975024 in the subgroups between CR and NCR patients.

Index		Genotype	Non-CR	CR	X^2	P value
Age	<60	GG	4	3	4.868b	0.088
		AG	27	6
		AA	26	18
	≥60	GG	6	7	8.336c	0.015
		AG	32	22
		AA	19	40
Albumin	<40	GG	8	7	6.062b	0.048
		AG	20	14
		AA	27	50
	≥40	GG	2	3	2.484c	0.289
		AG	39	14
		AA	18	8
CREA	<97	GG	10	10	5.093b	0.078
		AG	46	24
		AA	45	48
	≥97	GG	0	0	14.748c	0.000
		AG	13	4
		AA	0	10
UA	<428	GG	10	7	14.155b	0.001
		AG	56	25
		AA	37	54
	≥428	GG	0	3	4.295c	0.117
		AG	3	3
		AA	8	4

3.4. Multiple logistic regression

Multiple logistic regression analysis was used between clinical and polymorphism variables in consideration of the influence of confounding factors. The results were displayed that CC + TC (rs6975024) was a protective fact or (OR = −0.772, P = 0.025). Additionally, the level of albumin was inversely correlated with CR (OR = −0.145, P = 0.005), whereas the values of UA were associated with CR (OR = 0.007, P < 0.001) (Table 6).

Table 6. Multiple logistic regression analysis of nongenetic and genetic factors.

Index	B	Std. Error	Wald	P value	Exp (B) (95% C.I)
rs4607517 GG + AG/AA	.035	.793	.002	0.965	1.036 (0.219,4.901)
rs6975024 CC + TC/TT	−.772	.345	5.001	0.025	0.462 (0.235,0.909)
Age	.014	.018	.620	0.431	1.014 (0.980,1.050)
Albumin	−.145	.051	7.984	0.005	0.865 (0.782,0.956)
CREA	.015	.011	1.875	0.171	1.016 (0.993,1.038)
UA	.007	.002	15.891	0.000	1.007 (1.004,1.001)
Constant	2.786	2.632	1.121	0.290	16.213

Note: Bold values represent that the results were of statistical significance.

4. Discussion

Antiplatelet medicine could decrease the MACE risk of CAD patients; hence, it is the cornerstone of CAD treatment. By inhibiting the P2Y12 receptor, clopidogrel could reduce platelet aggregation, which was induced through adenosine diphosphate (ADP) [13]. However, approximately 10–40% of patients have suffered from clopidogrel resistance. With a higher Residual platelet reactivity (RPR) after clopidogrel loading presented larger intracoronary thrombus burden, worse post-PCI myocardial flow and perfusion [14]. With the development of new P2Y12 inhibitors, prasugrel and ticagrelor could inhibit platelet activity more rapidly, consistently, and effectively. But they cause more hemorrhagic events [15]. The COSTIC trial showed that for Chinese CAD patients, clopidogrel was more suitable compared with ticagrelor [16]. Therefore, diversiform responding to clopidogrel therapy and the underlying causes are increasingly being considered.

Numerous extrinsic elements might result in CR. First, the drug interactions between clopidogrel and PPIs [17] and between clopidogrel and traditional Chinese medicine [18] affect the pharmacokinetics of clopidogrel. In addition, smokers with clopidogrel therapy display lower platelet reactivity and have a reduced risk of CR, which was the opposite of our normal conception [19]. Moreover, comorbidities, such as renal failure and type 2 diabetes (T2D) [20], might be related to high platelet reactivity. Platelet hyper aggregation as well as hypercoagulation are interrelated with a higher thrombogenic risk, especially for T2D patients. In our present study, due to limited samples, we did not find diabetes or the value of glucose to be associated with CR. However, we discovered that the values of albumin were correlated with a decreased risk of CR, and higher levels of UA may be positively associated with CR. While the results of previous correlation analyses were not consistent, another investigation pointed out that the levels of uric acid were less likely to influence clopidogrel antiplatelet effects [21]. In the future, larger sample trials are needed to provide reliable conclusions.

Additionally, many studies have suggested the genetic variants in clopidogrel metabolic genes, such as ABCB1, CYP enzymes, paraoxonase-1, and P2Y12, are responsible for clopidogrel nonresponse [22]; other gene polymorphisms might also affect the clopidogrel antiplatelet effect. One recent study identified a statistically significant correlation between the rs2244613 polymorphism in the CES1 gene and residual platelet aggregation in patients undergoing clopidogrel therapy [23]. In this study, we investigated the associations of CR with rs4607517 as well as rs6975024, and found that SNPs in rs4607517 and rs6975024 were associated with CR, and patients with the G allele of rs4607517 had a higher risk of CR, but for rs6975024, the C allele might be a protective factor. These might be due to the change of GCK gene expression resulting from above variations. The GCK plays a vital role in DM, and changes in glucose balance induce platelet activity. Polymorphisms might influence the expression of GCK to alter the response to clopidogrel. Hence, our findings might be novel molecular markers for the identification of CR. Limited by sample size, additional studies with more advanced empirical approaches including gene expression analysis could further validate our findings in the future.

In addition to SNPs, epigenetic modifications in DNA could also influence platelet activation after clopidogrel treatment. Usually, the DNA hypermethylation leads to transcriptional silencing and influences its expression. And aberrant CpG methylation usually participate in various diseases, such as coronary heart disease [24]. Yang et al. found that in ischaemic stroke patients, ABCB1 promoter hypomethylation was associated with a decreased response to the treatment of clopidogrel [25], and another study indicated that P2Y12 promoter hypomethylation was correlate to higher platelet-activity and a poor risk for ischaemic events [26]. Former studies also indicated that methylation levels of P2Y12 [27], PON1 [28], and GCK [29] promoter might be relative to a decline in clopidogrel responding. Additionally, clinical pharmacokinetics research revealed that hypomethylation of the PON1 promoter may be a risk factor for bleeding after dual antiplatelet therapy [30]. Therefore, DNA methylation might be another novel molecular marker for the identification of higher platelet reactivity.

In addition to the changes in DNA sequences and their epigenetic modifications in clopidogrel resistance, some studies have focused on RNAs. A recent study performed lncRNA microarray and real-time RT-PCR analyses and indicated that in leukocytes from the clopidogrel-resistant group, lncRNAs (NONHSAT083775.2 and NONHSAT107804.2) were upregulated and one lncRNA (NONHSAT133455.2) was down regulated [31]. Moreover, another investigation focused on T2D patients, and through lncRNA array analysis in megakaryocytes, they screened differentially expressed lncRNAs and confirmed that lncRNA metallothionein 1 pseudogene 3 (MT1P3) was upregulated; moreover, they identified that through sponging miR-126, MT1P3 was a crucial regulator in increasing P2Y12 expression, which would result in platelet activation [32]. Interestingly, by suppressing the mediator lncRNA HIF1A-AS1, clopidogrel rescued apoptosis and promoted the proliferation of palmitic acid-induced damage [33], which provided new insight into clopidogrel for CAD treatment. Hence, lncRNAs might be discovered as a key molecular mechanism of CR in the future.

5. Conclusions

In summary, our study found that the SNPs rs4607517 and rs6975024 were associated with CR. Additionally, patients with the G allele of rs4607517 had a poorer risk for clopidogrel resistance, but for rs6975024, the C allele might be a protective factor. Finally, logistic regression revealed that CC + TC (rs6975024) as well as the values of albumin were correlated with a decreased risk of CR, and higher levels of UA may be positively associated with CR. The above findings might provide freshly insights for managing platelet hyperactivity-related diseases. However, with advanced approach and effective planning, we intend to perform larger sample studies to validate these discoveries in further research.

Acknowledgements

We thank Lindan Ji for excellent technical assistance. The authors also thank Dr Jin Xu for editing the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study was supported by the Zhejiang Provincial Natural Science Foundation of China [grant numbers LY19H020003, LY19H310002], Ningbo Natural Science Foundation [grant number 2017A610198], Ningbo social development project [grant number 2014C50071], the Plan of Science and Technology on Medicine and Health in Zhejiang Province [grant number 2019KY650], and the Ningbo Health Branding Subject Fund [grant number PPXK2018-01].