ABCC2 Polymorphisms and Haplotype are Associated with Drug Resistance in Chinese Epileptic Patients
The first three authors contributed equally to this work.
SUMMARY
Aims: Some study found that ATP-binding cassette (ABC) efflux transporters play an important role in antiepileptic drug resistance, especially ABCB1 and ABCC2. The aims of this study were to evaluate the relationship between the genetic polymorphisms of ABCC2 and ABCB1 and the therapeutic efficacy of antiepileptic drugs (AEDs) in Chinese epileptic patients. Methods:ABCB1 rs1045642 (3435C>T) and ABCC2 rs717620 (−24C>T), rs3740066 (3972C>T), and rs2273697 (1249G>A) polymorphisms loci in 537 Chinese epilepsy patients (217 drug resistant patients and 320 drug responders) were genotyped by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP). Results:ABCC2 rs717620 −24TT genotype was significantly associated with drug resistant epilepsy (odds ratio [OR]= 4.06 [1.79–9.20], P= 0.001). The OR values of ABCC2 rs717620 −24 CT+TT genotypes and ABCC2 rs3740066 (3972C>T) CT+TT genotypes were markedly higher in drug resistant patients (OR = 1.57 [1.08–2.29], P= 0.018; OR = 1.49 [1.02–2.18], P= 0.038, respectively) compared with responsive patients. ABCC2 rs2273697 (1249G>A) and ABCB1 rs1045642 (3435C>T) polymorphisms were not associated with drug resistant epilepsy. Linkage disequilibrium (LD) test showed that the ABCC2 rs717620 were in strong LD with rs2273697 (D’= 0.694) and rs3740066 (D’= 0.699). The frequencies of haplotypes TGT (ABCC2 −24C>T/ABCC2 1249G>A/ABCC2 3972C>T) in resistant patients was significantly higher than those in responsive patients (21.0% vs. 14.2%, P < 0.05). Conclusion:ABCC2−24C>T, 3972C>T polymorphisms and one ABCC2 haplotype is associated with AED resistance; ABCC2 1249G>A and ABCB1 3435C>T polymorphisms are not associated with AED resistance in our study. These data suggest that ABCC2 polymorphisms and haplotype may affect the response of antiepileptic drugs.
Introduction
Epilepsy is a chronic neurological illness, in which abnormal electrical activity in the brain causes involuntary changes in body movement, function, sensation, or behavior [1]. The World Health Organization estimates that epilepsy affects approximately 50 million people worldwide. Now-a-days, we recognize that genetic factors play an even more important role in the pathogenesis of epilepsy and drug efficacy than previously appreciated [1, 2]. Approximately, every second epileptic patient became resistant to the initial drugs [3]. Why so many epileptic patients show resistance against the antiepileptic drugs (AEDs) is not well-understood. Some potential mechanisms have been suggested with regard to the targets and transporters of the drugs. The voltage-gated ion channels are targets of some AEDs; alternations in the composition and functionality of them may lead to resistance in epileptic patients [4, 5]. The overexpression of ATP-binding cassette (ABC) efflux transporters in the blood-brain barrier (BBB), causing low-drug concentration at their target, may be a potential mechanism [6]. Seizures, exposure to some AEDs such as carbamazepine, and genetic factors may influence the expression of transporters [7–9]. Some ABC transporters such as multidrug resistance 1 (MDR1, p-glycoprotein, ABCB1), and a group of multidrug resistance proteins (MRPs, ABCCs), have been shown to mediate AEDs in brain [7, 9, 10]. These transporters are expressed in endothelial cells, astrocytes, and neurons of human brain tissues [11]. Therefore, these transporters may influence the efficacy of AEDs.
Initial observation of the relationship between ABCB1 3435C>T and drug resistant epilepsy was followed by a number of studies to evaluate the single nucleotide polymorphism (SNP) as a predictor of AED resistance [12, 13]. However, some studies have conversely suggested no association between ABCB1 polymorphisms and the response to AEDs [14, 15]. Leschziner et al. studied a cohort of 503 epilepsy patients and found no evidence to support that ABCB1 common variation influences drug withdrawal outcomes [16]. ABCC2 is one of the ABC efflux transporters, whose role in human brain tissues is not fully understood due to its low expression in normal brain tissues [17]. But some studies found that the expression of ABCC2 was upregulated in the brain tissues of epileptic patients [18, 19]. Moreover, in the ABCC2-deficient rat model, it is found that ABCC2 is involved in carbamazepine efficacy [20]. Recently, a study found a significant association between ABCC2 V417I polymorphism and reduced oral bioavailability of talinolol [21]. Kin et al. found that the ABCC2 1249G>A, and ABCB1 3435C>T, 1236C>T, 2677G>T/A polymorphisms were not associated with AED resistance in Japanese epileptic patients [22]. The study of Ufer et al. showed that a higher risk of AED failure in ABCC2−24T allele carriers is possible due to upregulation of ABCB1, but the mechanism is unknown [23].
Overall, whether the polymorphisms of these genes are associated with AED resistance is still not clear. To clarify whether ABCC2 and ABCB1 genes are involved in drug resistance epilepsy, we investigated the effects of ABCB1 rs1045642 and ABCC2 rs717620, rs3740066, and rs2273697 polymorphisms on AED resistance in Chinese epileptic patients.
Methods and Materials
Subjects
A total of 537 Chinese epileptic patients treated with AEDs (326 males, 211 female, mean age: 16.7 ± 11.7 years) from Xiangya hospital, the Second Xiangya Hospital of Central South University, and Hunan Provincial People's Hospital, were recruited in this study. The patients were diagnosed and classified according to guidelines of the International League against Epilepsy. Exclusion criteria included severe adverse drug reactions, unreliable record of seizure frequency, poor compliance with AEDs, history of alcohol or drug abuse, presence of progressive or degenerative neurological or systemic disorders, and hepatic or renal failure. A standardized questionnaire was administered to collect demographic details and clinical data such as seizure types and frequency, past medical history, AED history, concomitant drug history, and relevant family history. All patients or their parents gave their written consent to participate in the study. The study protocol was approved by the ethics committee of Xiangya School of Medicine and ethics committee of Institute of Clinical Pharmacology of Central South University. A clinical study admission (the registration number: ChiCTR-TCH-0000813) was approved by Chinese Clinical Trail Register. The patients were considered to be drug-responsive if they had not experienced any type of seizures for a minimum of 1 year after receiving AEDs. Drug resistance was defined as having at least four seizures during the previous year while trying at least three antiepileptic medications at maximal tolerated doses [16, 22].
Genotyping
Blood samples (5 mL) for genotyping were obtained with EDTA through the venipuncture used for pharmacokinetic monitoring and frozen at −80% for 24 h. DNA was isolated using phenol-chloroform extraction method. Genotyping was performed using the polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) assay. The designed primer pairs for ABCB1 rs1045642 and ABCC2 rs717620, rs3740066, and rs2273697 loci were summarized in Table 1. Each genotyping was performed in a blinded manner. Automatic sequencing was performed for randomly selected DNA samples from patients.
| Gene | SNPs | Primer sequence | Restriction enzyme |
|---|---|---|---|
| ABCC2 | rs 717620 | 5′-taaatggttgggatgaaagg-3′ | Bpi I |
| (−24C>T) | 5′-gctttagaccaattgcacatc-3′ | ||
| rs2273697 | 5′-gggcaaagaagtgtgtggat-3′ | Nco I | |
| (1249G>A) | 5′-acatcaggttcactgtttctccca-3′ | ||
| rs 3740066 | 5′-atccttggctttgtccttgg-3′ | BsiE I | |
| (3972C>T) | 5′-aataatcaggttcactgtttctcccac-3′ | ||
| ABCB1 | rs1045642 | 5′-tgttttcagctgcttfatgg-3′ | Bsp143 I |
| (3435C>T) | 5′-aaggcatgtatgttggcctc-3′ |
Statistical Analysis
The SPSS software package (version 13.0 for Windows; SPSS, Chicago, IL, USA) was used for statistical analysis. Hardy–Weinberg equilibrium was analyzed with χ2 test or Fisher's Exact test as applicable in the studied sample. Age was compared between responsive and nonresponsive patients with the Student's t-test. Sex and other characteristics about patients are compared with χ2 test. The relationship between various genotypes and responsiveness was examined by binary logistic regression, after adjustment for age, sex, and seizure type. Linkage disequilibrium and haplotypes were analyzed using SHEsis (http://analysis.bio-x.cn/myAnalysis.php) [24, 25]. Statistical significance was accepted when P < 0.05.
Results
The study population consisted of 320 AED responders (16.4 ± 12.3 years) and 217 resistant patients (17.0 ± 10.7 years). The most commonly used AED is carbamazepine (65.4%), followed by valproic acid (64.6%), oxcarbazepine (27.7%), phenobarbital (20.9%), lamotrigine (17.7%), phenytoinum natricum (15.8%), topiramate (14.0%), and levetiracetam (11.0%; Table 2).
| Parameters | Drug-responsive (320) | Drug-resistant (217) | Total (537) |
|---|---|---|---|
| Sex (%) | |||
| Male | 187 (58.4%) | 139 (64.1%) | 326 (60.7%) |
| Female | 133 (41.6%) | 78 (35.9%) | 211 (39.3%) |
| Age (years) | 16.4 ± 12.3 | 17.0 ± 10.7 | 16.7 ± 11.7 |
| AEDs used in patients (%) | |||
| Carbamazepine | 215 (68.1%) | 136 (62.7%) | 351 (65.4%) |
| Valproic acid | 168 (52.5%) | 159 (73.3%) | 347 (64.6%) |
| Phenytoinum | 46 (14.4%) | 39 (18.0%) | 85 (15.8%) |
| Phenobarbital | 54 (16.8%) | 68 (31.3%) | 112 (20.9%) |
| Lamotrigine | 39 (12.2%) | 56 (25.8%) | 95 (17.7%) |
| Levetiracetam | 23 (7.20%) | 36 (16.6%) | 59 (11.0%) |
| Topiramate | 43 (13.4%) | 32 (14.7%) | 75 (14.0%) |
| Oxcarbazepine | 70 (21.9%) | 78 (36%) | 148 (27.7%) |
| Seizure type (%) | |||
| Partial | 125 (39.0%) | 79 (36.4%) | 204 (38.0%) |
| Generalized | 155 (48.4%) | 115 (53.0%) | 270 (50.3%) |
| Nonclassifiable | 40 (12.6%) | 23 (10.6%) | 63 (11.7%) |
| Etiology (%) | |||
| Idiopathic | 121 (37.8%) | 75 (34.6%) | 196 (35.1%) |
| Symptomatic | 85 (26.6%) | 64 (29.5%) | 149 (31.0%) |
| Cryptogenetic | 114 (35.6%) | 78 (35.9%) | 192 (33.9%) |
We determined three SNPs in ABCC2 in 217 drug resistant patients and 320 drug responders (Table 3). The investigated SNPs were all in Hardy–Weinberg equilibrium. The OR values in patients with ABCC2−24TT genotype or CT+TT genotypes were significantly higher in drug-resistant patients than those in drug responders (OR = 4.06 [1.79–9.20], P= 0.001; OR = 1.57 [1.08–2.29], P < 0.01). The ABCC2 rs2273697 (1249G>A) polymorphism was not significantly associated with drug resistance after subjects were adjusted for sex, age, and seizure type. The OR value in patients with ABCC2 3972 CT+TT genotypes were also higher in drug-resistant patients than that in drug responders (OR = 1.49 [1.02–2.18], P= 0.038). In an independent cohort of young responders and nonresponders, the CT or TT genotypes were overrepresented in nonresponders group (OR = 1.95 [1.11–3.43], P= 0.020; OR = 3.30 [1.29–8.40], P= 0.013; data not shown). We did not observe any significant differences in ABCB1 rs1045642 (3435C>T) SNP between drug resistant patients and responders. After adjusted with age, sex, seizure types, the OR values in ABCB1 3435TT homozygous and 3435 CT+TT genotypes are 1.04 (0.587–1.85, P= 0.888) and 1.02 (0.696–1.48, P= 0.936), respectively.
| Gene | SNP rs# | Drug-resistant (n = 217) | Drug-responder (n = 320) | Odds ratio (95%CI) | P-value |
|---|---|---|---|---|---|
| ABCC2 | rs717620 | CC 118 (54.4%) | CC 207 (64.5%) | – | – |
| (−24C>T) | CT 77 (35.5%) | CT 103 (32.2%) | 1.32 (0.889–1.97) | 0.167 | |
| TT 22 (10.1%) | TT 10 (3.10%) | 4.06 (1.79–9.20) | 0.001 | ||
| CT+TT 99 (45.6%) | CT+TT 113 (35.3%) | 1.57 (1.08–2.29) | 0.018 | ||
| rs227369 | GG 181 (83.4%) | GG 262 (81.9%) | – | – | |
| (1249G>A) | GA 35 (16.1%) | GA 56 (17.5%) | 0.915 (0. 559–1.50) | 0.725 | |
| AA 1 (0.500%) | AA 2 (0.600%) | 1.68 (0.147–19.0) | 0.679 | ||
| GA+AA 36 (16.7%) | GA+AA 62 (18.1%) | 0.932 (0.573–1.52) | 0.777 | ||
| rs3740066 | CC 123(56.7%) | CC 200 (62.5%) | – | – | |
| (3972C>T) | CT 75 (34.6%) | CT 102 (31.9%) | 1.41 (0.943–2.10) | 0.094 | |
| TT 19 (8.70%) | TT 18 (5.60%) | 1.94 (0.939–4.00) | 0.073 | ||
| CT+TT 94 (43.3%) | CT+TT 120 (37.5%) | 1.49 (1.02–2.18) | 0.038 | ||
| ABCB1 | rs1045642 | CC 81 (37.3%) | CC 116 (36.3%) | – | – |
| (3435C>T) | CT 105 (48.4%) | CT 161(50.3%) | 1.01 (0.683–1.50) | 0.967 | |
| TT 31 (14.3%) | TT 43 (13.4%) | 1.04 (0.587–1.85) | 0.888 | ||
| CT+TT 136 (62.7%) | CT+TT 204 (63.9%) | 1.02 (0.696–1.48) | 0.936 |
- Data has been adjusted for age, sex, and seizure type. Homozygous wild-type patients served as the reference group. OR, odds ratio; CI, confidence interval.
The linkage disequilibrium test of three SNPs in ABCC2 was shown in Figure 1. Rs717620 and rs2273697 were in strong linkage disequilibrium (D-prime = 0.694) and rs717620 and rs3740066 were also in strong linkage disequilibrium (D-prime = 0.699). The haplotypes (ABCC2−24C>T/ABCC2 1249G>A/ABCC2 3972C>T) test showed that one haplotype frequency in drug resistant group was different from those in drug responsive group (Table 4). The frequency of haplotype TGT (ABCC2−24C>T/ABCC2 1249G>A/ABCC2 3972C>T) in resistant patients was significantly higher than those in responsive patients (21.0% vs. 14.2%, P < 0.05).
The linkage disequilibrium of three SNPs in ABCC2 in all patients. Linkage disequilibrium between pairs of polymorphisms is shown in diamonds (D-prime), with darker shading indicating greater D-prime.
| Haplotype | Frequency | P-value | ||
|---|---|---|---|---|
| ABCC2 | resistant | responsive | Fisher's P value | Pearson's P value |
| CGC | 0.600 | 0.657 | 0.055 | 0.055 |
| CGT | 0.040 | 0.065 | 0.086 | 0.086 |
| CAC | 0.074 | 0.083 | 0.591 | 0.591 |
| TGC | 0.064 | 0.042 | 0.111 | 0.111 |
| TGT | 0.210 | 0.142 | 0.003* | 0.003* |
- Polymorphisms are in the same order as in Table 3. All those frequencies <0.01 will be ignored in analysis.
- *P < 0.05.
Discussion
Despite more and more research on AED resistance, we still know little about the mechanism. One possible reason lays on overexpression of multidrug transporter proteins [26]. As described in one review [27], ABC transporters are considered as one of the hottest topics in epileptology. ABCB1 was regarded as the first factor influencing AEDs in BBB. Similar to ABCB1, ABCC2 expression was also increased in brain-derived endothelial cells from patients with AED resistance [28].
So in the present study, we investigated three SNPs in ABCC2 and one SNP in ABCB1 in 537 responsive or resistant epilepsy patients and found that ABCC2−24C>T and 3972C>T polymorphisms are associated with AED resistance (Table 3). But we found no association between polymorphisms of ABCC2 1249G>A, ABCB1 3435C>T and AED resistance. These data may have some conflicts due to reduced activity of ABCC2−24T gene product ABCC2 protein. In one study, the ABCC2−24T construct showed an 18.7% reduced activity in vitro because of its lower mRNA levels in normal tissues [29]. Ufer et al. also found the association between −24T allele carriers and AED failure in Caucasian. Moreover, they further found that −24C>T genotype did not affect hippocampal ABCC2 expression but increased ABCB1 expression [23]. Despite the expression and functional relevance of ABCC2 in kidney tissue and other cell lines, its condition in BBB or brain tissue still remains controversial. Increased ABCC2 mRNA and ABCC2 protein expression in brain-derived endothelial cells from AED resistant patients have been reported [30, 31]. Therefore, it is possible that tissue dependent differential regulation of ABCC2 expression induces different conditions in hepatic, BBB, or brain tissue [23].
We also find that ABCC2 3972 CT+TT genotypes are associated with drug resistance. In another study, it is found that the silent SNP 3972C>T is strongly linked to −24C>T, which is associated with a decreased function of the transporter [32]. In our study, we also investigated the linkage disequilibrium and found strong link between 3972C>T and −24C>T (D’= 0.699, Figure 1). Although 3972C>T is a synonymous SNP, we think that it might still have significant function to affect the absorption or excretion of its substrates. One study reported the in vivo association of c.3972C> T variant with increased area under the concentration-time curve of irinotecan and its metabolites [33], revealed the functional relevance of this variant for bioavailability of drugs and toxins.
Haenisch et al. observed a significant association of the 1249G>A polymorphism with reduced oral bioavailability of talinolol, presumably as a result of increased ABCC2 activity [21]. But another study showed that 417I-ABCC2 (c.1249A) variant indicated reduced carbamazepine transporter activity compared with 417V-ABCC2 (c.1249G). Interestingly, they found that the 1249G>A showed an association with adverse drug reaction but not with drug efficacy in central nervous system [34]. In this study, we found that there is no association between ABCC2 1249G>A polymorphism and AED resistance.
Currently, there are various studies focusing on the relationship between ABCB1 3435C>T polymorphism and AED efficacy, but the results have been inconclusive. In some studies, the results were positive [12, 13], but in others, results were negative [15, 16]. Moreover, in some meta-analyses about ABCB1 3435C>T, the results were also negative [15, 35]. In a meta-analysis of 23 studies (7,067 patients), no significant association of ABCB1 alleles, genotypes or haplotypes with the response to treatment was found in the overall population or in each ethnicity subgroup [35]. In a functional study on the polymorphism in the cortex and hippocampus of drug-resistant patients, no association was detected between 3435C>T and 2677G>T/A polymorphisms and ABCB1 mRNA expression levels [36]. In our study, we did not find any association between 3435C>T polymorphism and AED efficacy either (Table 3).
Now-a-days, more and more studies focused not only on the simple one SNP but also on the haplotypes. We therefore analyzed the haplotypes of the three SNPs. For the first time, we found that there were five haplotypes (>1%), and one haplotype was associated with AED resistance (Table 4). ABCC2 TGT haplotype frequency is higher in resistant group than in responsive group. Think about of three SNPs’ function, 417I-ABCC2 (c.1249A) variant indicated reduced carbamazepine transporter activity compared with 417V-ABCC2 (c.1249G); the silent SNP 3972C>T is strongly linked to −24C>T, which is associated with a decreased function of the transporter. We can see that the TGT haplotype really further reduce the transporters activity.
As far as we know, this is the first report about some association between ABCC2 haplotype and AED efficacy. Patrick Kwan et al. studied a total of 25 tagging SNPs from ABCC2, ABCC5, and ABCG2 genes. However, no haplotypes were over 1% frequency, which included all SNPs of the genes associated with drug resistance [37]. Haerian et al. studied ABCB1 haplotypes on five loci and there were no positive results either [38]. Although so many studies have been launched on the relationship between ABC transporters and AED resistance, the overall results are inconsistent.
Epilepsy is a complicated disease and most patients were treated with AEDs of varying doses and kinds, making the drug-resistance mechanism extremely complex. Moreover, there is no precise phenotype or definition of drug resistance. Another unfavorable factor for our study is that the study population is limited and modest effects of other SNPs on drug resistance can not be excluded. All these factors contributed to the limitation of our results. Further studies with larger sample size on more related SNPs in different candidate genes are needed to confirm the positive results.
In conclusion, our data revealed that ABCC2−24TT or CT+TT genotypes, ABCC2 3972 CT+TT genotypes, and one ABCC2 haplotype is obviously associated with drug resistance in epileptic patients. However, the ABCC2 1249G>A polymorphism and ABCB1 3435C>T polymorphism are not associated with this process.
Acknowledgments
We thank the study participants. This work was supported by the National High-Tech R&D Program of China (863 Program; 2009AA022704), National Natural Science Foundation of China (81173129), Program for Changjiang Scholars and Innovative Research Team in University of Ministry of Education of China (IRT0946), and the Open Foundation of Innovative Platform in University of Hunan Province of China (10K078).
Conflict of Interest
The authors declare no conflict of interest.
