Association between ABCB1 C3435T polymorphism and drug-resistant epilepsy in Han Chinese

Transcription

1 Epilepsy & Behavior 11 (2007) Association between ABCB1 C3435T polymorphism and drug-resistant epilepsy in Han Chinese Patrick Kwan a, *, Larry Baum a, Virginia Wong b, Ping Wing Ng c, Colin HT Lui d, Ngai Chuen Sin e, Andrew C.F. Hui a, Evelyn Yu a, Lawrence K.S. Wong a a Division of Neurology, Department of Medicine and Therapeutics, Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China b Department of Paediatric and Adolescent Medicine, University of Hong Kong, Hong Kong, China c Division of Neurology, Department of Medicine and Geriatrics, United Christian Hospital, Hong Kong, China d Division of Neurology, Department of Medicine, Queen Elizabeth Hospital, Hong Kong, China e Department of Paediatrics, Prince of Wales Hospital, Hong Kong, China Received 22 November 2006; revised 26 March 2007; accepted 12 April 2007 Available online 22 May 2007 Abstract There is accumulating evidence to suggest that overexpression of efflux drug transporters at the blood brain barrier, by reducing antiepileptic drug (AED) accumulation in the seizure foci, contributes to drug resistance in epilepsy. P-glycoprotein, encoded by the ABCB1 gene, is the most studied drug transporter. There are conflicting data as to whether the CC genotype of the ABCB1 3435C > T polymorphism is associated with drug resistance in Caucasian patients with epilepsy. We investigated this association in ethnic Chinese. ABCB1 3435C > T was genotyped in 746 Han Chinese patients with epilepsy and 179 controls. Patients with drug-resistant epilepsy were more likely to have the TT genotype compared with those with drug-responsive epilepsy (16.7% vs 7.4%, odds ratio = 2.5, 95% confidence interval = , P = ). Our results contrast with those of studies of Caucasians, and highlight the complexity of the possible role of this polymorphism in AED response in different ethnic populations. Ó 2007 Elsevier Inc. All rights reserved. Keywords: Epilepsy; Drug resistance; ABCB1; P-glycoprotein; Pharmacogenetics; Genetic polymorphisms; Antiepileptic drugs; Chinese; MDR1; Drug transporters 1. Introduction Despite antiepileptic drug (AED) treatment, 30% of patients continue to have seizures [1]. There is accumulating evidence to suggest that overexpression of efflux drug transporters at the blood brain barrier, by reducing AED accumulation in the seizure foci, contributes to drug resistance in epilepsy [2 4]. P-glycoprotein (P-gp), encoded by the ABCB1 (or MDR1) gene, is the most studied drug transporter. A common single-nucleotide polymorphism (SNP), 3435C > T in exon 26 of ABCB1, although synonymous, has been found to be associated with altered P-gp * Corresponding author. Fax: address: patrickkwan@cuhk.edu.hk (P. Kwan). level and activity in duodenal enterocytes, affecting drug absorption [5]. It has been hypothesized that the 3435CC genotype, possibly by increasing P-gp expression and/or function in the brain, might be associated with drug resistance in epilepsy. As the demonstration of such an association could potentially influence drug choice in patients with different genotypes, a number of studies have been performed in both Caucasian [6 10] and non-caucasian [11 14] subjects to test this hypothesis. The results of these studies have been conflicting and inconclusive. It is important to investigate whether the same association (or lack thereof) may be observed in different ethnic groups due to considerable ethnic variation in the frequencies of the 3435C > T genotypes [15], and because genetic variants may have inconsistent effects across different ethnic groups /$ - see front matter Ó 2007 Elsevier Inc. All rights reserved. doi: /j.yebeh

2 P. Kwan et al. / Epilepsy & Behavior 11 (2007) [16]. It is also important that, to avoid bias and chance finding, genetic association studies be repeated in different patient cohorts [17]. The relationship between the ABCB1 3435C > T polymorphism and drug response of patients with epilepsy remains controversial. Therefore, we investigated whether the 3435C > T SNP of ABCB1 might be associated with drug response among patients with epilepsy of Chinese ethnicity. 2. Methods 2.1. Subjects This multicenter study was performed in Hong Kong. The study was approved by the ethics committees of the participating hospitals. Patients were recruited from the neurology/epilepsy clinics and inpatient wards of four major regional hospitals, serving a total population of 2.5 million (representing 35% of the whole population of the city). Patients with epilepsy at least 15 years of age and of Han Chinese ethnicity (based on selfreported ancestry) were eligible for inclusion if they had been receiving AED treatment for at least a year. Exclusion criteria included poor compliance with AED therapy, unreliable record of seizure frequency, significant psychiatric comorbidity, history of pseudoseizures, alcohol or illicit drug abuse, and presence of progressive or degenerative neurological or systemic disorders. After written informed consent was obtained from the patients or their guardians, a standardized questionnaire was administered to collect demographic details and information on seizure types and frequency, past medical history, AED history, concomitant drug history, and relevant family history. Seizures and epilepsy syndromes were classified according to international guidelines [18,19]. A 5-mL venous blood sample was taken for DNA extraction and genotyping. One hundred seventy-nine DNA samples extracted from cord bloods of unidentified neonates from a DNA repository at the Chinese University of Hong Kong served as controls. Subject information and genotype data were identified by a coded ID to protect privacy and to ensure that the genotyping was done blind Phenotyping Response to treatment was assessed by the recruiting neurologists in each center who were blind to the genotype results. Three phenotypic groups were defined: drug-responsive, drug-resistant, and active. Patients who had not experienced any seizure for at least a year up to the date of recruitment, and received a stable dose of an AED, were considered to have drug-responsive epilepsy [1]. Patients who had had an average of one seizure or more per month over the previous year despite treatment with two or more AEDs at therapeutic dosages and/or serum drug concentrations were considered to have drug-resistant epilepsy. In this context, AEDs must have failed primarily because of inadequate efficacy instead of adverse effects. To address what is frequently encountered in clinical practice, patients whose seizure status did not fulfill these criteria, that is, who had failed only one AED and/or had one or more seizures per year but less than one per month, were considered to have active epilepsy to represent an intermediate group in terms of drug response. Patients who underwent epilepsy surgery for refractory epilepsy were classified as having drug-resistant epilepsy, regardless of their postoperative seizure control status Genotyping Patients were genotyped using a PCR-RFLP assay to analyze the 3435C > T polymorphism of the ABCB1 gene as described previously [20]. To confirm accuracy, genotyping of all the drug-resistant and drugresponsive patients was repeated by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF) on a Sequenom Mass Array system. Any discrepancies between the two methods were resolved by direct sequencing using an Applied Biosystems 3100 capillary sequencer Population stratification To detect any unsuspected population stratification, randomly selected polymorphisms on eight different chromosomes were selected from Hap- Map and typed using MALDI-TOF Statistical analysis For the primary analysis, we compared genotype and allelic frequencies between patients with drug-resistant and drug-responsive epilepsy with the v 2 test. Where deviation from Hardy Weinberg equilibrium existed, we applied the method proposed by Schaid and Jacobsen to correct for any effect on the Type 1 error rate and power when comparing allele frequencies [21]. Logistic regression was used to adjust for age, sex, and epilepsy syndrome. For the secondary analysis, v 2 for trend was used to assess whether a dose phenotype effect existed for genotype and level of drug resistance. P values (two-sided) were considered statistically significant. 3. Results Seven hundred forty-six (221 drug-resistant, 297 drugresponsive, and 228 active) patients were recruited. Their mean age was 36.2 years (SD = 16.2); 49.9% were male. Epilepsy was classified as idiopathic in 18.4% of patients, cryptogenic in 31.4%, and symptomatic in 46.6%, and was unclassifiable in 3.6%. Consistent with previous observational data [1], compared with the active and drugresponsive groups, the drug-resistant group had a larger proportion of patients with symptomatic epilepsy syndromes and a smaller proportion of patients with idiopathic syndromes (Table 1). The mean duration of epilepsy was 16.5 years (SD = 12). The median number of AEDs that failed among drug-resistant patients was 3 (range: 2 9), and their mean monthly seizure frequency was 11 (SD = 47.3). The distributions of 3435C > T genotypes in the controls and drug-resistant patients were consistent with Hardy Weinberg equilibrium (both P > 0.05), but deviation from equilibrium was observed for the active and drug-responsive groups and for the patient cohort as a whole (all P < 0.05) (Table 2). To examine the possibility that drug-responsive and drug-resistant groups were covertly stratified, eight random polymorphisms across the genome were genotyped (Table 3). None were significantly associated with drug responsiveness. All were in Hardy Weinberg equilibrium in both the drug-resistant and drug-responsive groups. Patients with drug-resistant epilepsy were more likely to have the TT genotype at ABCB compared with those with drug-responsive epilepsy (16.7% vs 7.4%, odds ratio [OR] = 2.5, 95% confidence interval [CI] = , P = ). To account for deviation from Hardy Weinberg equilibrium in the drug-responsive group, we applied the method proposed by Schaid and Jacobsen to compare

3 114 P. Kwan et al. / Epilepsy & Behavior 11 (2007) Table 1 Epilepsy syndromes among different patient groups and the whole cohort (v 2 (4) = 43.6, P < 0.001) a Epilepsy syndrome Drug-resistant patients Active patients Drug-responsive patients All patients Idiopathic 15 (7.0) b 52 (23.8) 70 (24.4) 137 (19.1) Symptomatic 125 (58.1) 115 (52.8) 108 (37.8) 348 (48.4) Cryptogenic 75 (34.9) 51 (23.4) 108 (37.8) 234 (32.5) Total 215 (100) 218 (100) 286 (100) 719 (100) a Epilepsy was unclassifiable in 3.6% of patients. They were not included in this v 2 analysis. b Number (%). Table 2 Observed genotype frequencies in patient and control groups Genotype Nonepileptic controls Patients with epilepsy Drug-resistant Active Drug-responsive Total CC 68 (38.0) a 80 (36.2) 80 (35.1) 114 (38.4) 274 (36.8) CT 92 (51.4) 104 (47.1) 122 (53.5) 161 (54.3) 387 (51.9) TT 19 (10.6) 37 (16.7) 26 (11.4) 22 (7.4) 85 (11.4) Total 179 (100) 221 (100) 228 (100) 297 (100) 746 (100) a Number (%). Table 3 Distribution of randomly selected polymorphisms among the patients with drug-resistant and drug-responsive epilepsy Polymorphism and group Chromosome no. Genotypes (no. of patients) P value a rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive rs Drug-resistant Drug-responsive a The P values reflect the results of v 2 tests with one degree of freedom for differences in genotype frequencies between patients with drug-resistant epilepsy and those with drug-responsive epilepsy. There were no significant deviations from Hardy Weinberg equilibrium for the eight markers within either the drug-resistant or drug-responsive group. the allele frequencies [21]. The z statistic for the T allele obtained was , which was equivalent to a P value (two-sided) of 0.046, suggesting the homozygous T allele was associated with drug resistance, after correction of deviations from Hardy Weinberg equilibrium. Without correction for deviation from Hardy Weinberg equilibrium, the P value for association of the T allele with drug resistance was

4 P. Kwan et al. / Epilepsy & Behavior 11 (2007) Table 4 Comparison of genotype or allele frequencies in patients with drugresistant vs drug-responsive epilepsy a Comparison Odds ratio (95% CI) P value CC vs TT 0.45 ( ) 0.01 CC vs non-cc 0.98 ( ) 0.92 TT vs non-tt 2.4 ( ) C vs T 0.64 ( ) 0.12 a All values are adjusted for age, sex, and epilepsy syndrome. The association of the TT genotype with drug resistance remained significant after adjusting for age, sex, and epilepsy syndrome (OR = 2.4, 95% CI = , P = 0.003) (Table 4). Similarly, the probability of having the CC rather than the TT genotype was lower in drug-resistant than in drug-responsive patients (OR = 0.42, 95% CI = , P = 0.004). The frequency of the TT genotype in the active group (11.4%) was between those of the resistant and responsive groups, and there was a positive linear relationship between level of drug resistance and TT genotype proportion (P = 0.004). The frequency of the CC genotype tended to be lower in drug-resistant (36.2%) than in drug-responsive (38.4%) patients, but the difference was not significant. Genotype frequencies in the control subjects were intermediate between those of resistant and responsive epilepsy groups. The differences in genotype frequencies between the control subjects and the patient groups were not significant. 4. Discussion Whether an association exists between ABCB1 C3435T and drug-resistant epilepsy remains hotly debated. A positive association between 3435CC genotype and poor response to AEDs was first reported among Caucasian subjects in a genotype study by Siddiqui et al. in 2003 [6]. Subsequent to this report, results from at least four more adequately powered studies including chronic patients of similar ethnic backgrounds have been published, and their results are conflicting [7 10]. The association was not found in the studies conducted in Australia [7] and Glasgow [8], but haplotype analysis in an Austrian study suggested that this genotype might be associated with the degree of drug resistance as measured by seizure frequency [9]. A recently published study enrolling newly diagnosed patients in a randomized controlled drug trial did not find any association between 3435C > T genotype and time to first seizure after starting drug therapy or time to 12-month remission [10]. In contrast, our results suggest that the TT genotype is associated with drug resistance in Chinese patients with epilepsy. Hardy Weinberg equilibrium is an indicator of the genetically representative nature of any given population and a lack of selection on the basis of genetic or environmental factors. It is recognized that deviation from Hardy Weinberg equilibrium could inflate the possibility of a false-positive association (Type 1 error) [21]. Our drug resistant group and nonepileptic control group was in Hardy Weinberg equilibrium, but the drug-responsive and active patient groups were not. Applying the correction when such deviation exists for the primary analysis of the study, the T allele displayed significant association with drug resistance (compared with drug-responsive patients). Common reasons for deviation from equilibrium include genotype error, hidden population stratification, and selection bias. However, genotype accuracy was confirmed by three independent methods. Genotypes of randomly selected SNPs throughout the genome did not detect any evidence of population stratification. The drug-responsive group (n = 297) was larger than the drug-resistant group (n = 221); thus, the patient pool was not enriched with drug-resistant cases (selection bias). Having considered the aforementioned possibilities, we submit it is plausible that ethnicity contributed to the discrepant findings between the present study and those conducted in Caucasian cohorts. Other potential explanations for the discrepant findings include phenotype definition, epilepsy syndromes, the drugs that were used, and chance. Whereas CC was less common than TT in Caucasian patients (16 19% vs 27 33% in drug-responsive patients) [6 8], the reverse was observed in our Han Chinese cohort (38% vs 7%), similar to other Asian populations [15]. Of note, whereas the TT genotype was associated with lower P-gp expression in enterocytes and lower activity in Caucasian volunteers [5], the opposite has been found in some Japanese subjects [22,23], although others failed to find any association [24]. Consistent with this, in a recent Japanese study, the TT genotype was associated with drug-resistant epilepsy in univariate analysis, although the association became insignificant after correcting for multiple comparisons [11]. Only 210 patients were included in that study. Two recent Korean studies with relatively small sample sizes (each less than 200) failed to find an association between 3435C > T and drug response in patients with epilepsy [13,14]. However, a study of 331 Chinese patients with epilepsy in Taiwan identified several haplotype combinations that were associated with drug responsiveness or resistance [12]. By recruiting a larger sample, the present study found a statistically significant association between 3435TT genotype and drug-resistant epilepsy in Han Chinese patients. The 3435C > T SNP is in strong linkage disequilibrium with G2677T/A (nonsynonymous), but sitespecific mutagenesis experiments substituting at these sites found no effect on P-gp function in vivo [25]. The actual causal SNP remains to be determined. Haplotype frequencies may differ between ethnic groups. A haplotype containing 3435C > T and a putative causal SNP allele might be common in Chinese but rare in Caucasians, or vice versa. Thus, association studies could produce opposite results in different ethnic groups. There is also a high likelihood that discrepancies in results across the studies have arisen as a result of differ-

5 116 P. Kwan et al. / Epilepsy & Behavior 11 (2007) ences in phenotype definitions used. Drug responsiveness is generally defined as freedom from seizures for at least 1 year. For drug resistance, most studies required failure of two or three AEDs, but with seizure frequency varying from P4 [6,7,13] to only P1 [8,11] per year in different studies. Recently, Leschziner et al. studied initial drug response in a clinical trial instead of recruiting medically intractable patients [10]. It is arguable, at least theoretically, whether patients with only one seizure per year can be considered at a biological level to be pharmacoresistant. To maximize phenotypic differences between the comparison groups, a higher seizure frequency was adopted in defining drug resistance in the present study, with a mean seizure frequency of 11 per month and some patients having failed up to nine AEDs. Similar to the approach of the Austrian study [9], we included the active epilepsy group to represent patients with an intermediate phenotype in drug response. We recognize that this was a heterogeneous patient group with patients who might eventually fulfill the classifications of the drug-responsive or drug-resistant group, and the extent of resistance cannot be objectively assessed, especially in chronic epilepsy cases, without detailed knowledge of pre- and posttreatment seizure frequencies. Nonetheless, inclusion of this intermediate group aimed to reflect what is commonly encountered in clinical practice, and the frequency of the TT genotype in the active group was intermediate between those of the resistant and responsive groups. Like other studies, we did not consider whether individual AEDs that failed might be a confounding factor. This is because, despite results from animal studies, the substrate status of many AEDs for P-gp (and its clinical significance) remains ill-defined [3]. Our sample size did not provide sufficient statistical power for subgroup analyses based on epilepsy syndromes or etiologies. As ABCB1 is highly inducible [3], it might be speculated that subtle differences among studies in epilepsy types, ages of patients, duration of treatment, and AEDs used might induce large variations in P-gp activity, contributing to the discrepant genetic association results. More detailed comparison using individual patient data will be needed to address this possibility. In summary, our results suggest that ABCB1 3435TT is associated with drug resistance in Chinese patients with epilepsy. This association is in contrast to results of some studies on Caucasians, but consistent with evidence obtained in Japanese subjects. As with other genetic association studies, it is important to replicate the study in another population of Chinese patients to rule out a chance finding. Further studies using haplotype blocks including SNPs in linkage disequilibrium with 3435C > T (including 2677G > T or A, and 1236C > T) [9] should also be conducted. With sufficient individual studies performed, ethnic group-specific meta-analysis of ABCB1 polymorphisms and drug resistance in epilepsy should be considered, preferably using individual patient data to take into account the inconsistency in phenotype definitions between the studies. Our findings highlight the complexity of the possible role of this SNP in AED response in different ethnic populations, and the need for unified phenotypic definitions of drug response and resistance to facilitate comparison between results obtained in different studies. Acknowledgment This study was partially funded by a Hong Kong University Small Project Grant. References [1] Kwan P, Brodie MJ. Early identification of refractory epilepsy. N Engl J Med 2000;342: [2] Sisodiya SM. Mechanism of antiepileptic drug resistance. Curr Opin Neurol 2003;16: [3] Kwan P, Brodie MJ. Potential role of drug transporters in the pathogenesis of medically intractable epilepsy. Epilepsia 2005;46: [4] Löscher W, Potschka H. Drug resistance in brain diseases and the role of drug efflux transporters. Nat Neurosci Rev 2005;6: [5] Hoffmeyer S, Burk O, von Richter O, et al. Functional polymorphisms of the human multidrug-resistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo. Proc Natl Acad Sci USA 2000;97: [6] Siddiqui A, Kerb R, Weale ME, et al. Association of multidrug resistance in epilepsy with a polymorphism in the drug-transporter gene ABCB1. N Engl J Med 2003;348: [7] Tan NCK, Heron SE, Scheffer IE, et al. Failure to confirm association of a polymorphism in ABCA1 with multidrug-resistant epilepsy. Neurology 2004;63: [8] Sills GJ, Mohanraj R, Butler E, et al. Lack of association between the C3435T polymorphism in the human multidrug resistance (ABCB1) gene and response to antiepileptic drug treatment. Epilepsia 2005;46: [9] Zimprich F, Sunder-Plassmann R, Stogmann E, et al. Association of an ABCB1 gene haplotype with pharmacoresistance in temporal lobe epilepsy. Neurology 2004;63: [10] Leschziner G, Jorgensen AL, Andrew T, et al. Clinical factors and ABCB1 polymorphisms in prediction of antiepileptic drug response: a prospective cohort study. Lancet Neurol 2006;5: [11] Seo T, Ishitsu T, Ueda N, et al. ABCB1 polymorphisms influence the response to antiepileptic drugs in Japanese epilepsy patients. Pharmacogenomics 2006;7: [12] Hung CC, Tai JJ, Lin CJ, Lee MJ, Liou HH. Complex haplotypic effects of the ABCB1 gene on epilepsy treatment response. Pharmacogenomics 2005;6: [13] Kim YO, Kim MK, Woo YJ, et al. Single nucleotide polymorphisms in the multidrug resistance 1 gene in Korean epileptics. Seizure 2006;15: [14] Kim DW, Kim M, Lee SK, Kang R, Lee SY. Lack of association between C3435T nucleotide MDR1 genetic polymorphism and multidrug-resistant epilepsy. Seizure 2006;15: [15] Ameyaw MM, Regateiro F, Li T, et al. MDR1 pharmacogenetics: frequency of the C3435T mutation in exon 26 is significantly influenced by ethnicity. Pharmacogenetics 2001;11:7 21. [16] Goldstein DB, Hirschhorn JN. In genetic control of disease, does race matter? Nat Genet 2004;36: [17] Tan NC, Mulley JC, Berkovic SF. Genetic association studies in epilepsy: the truth is out there. Epilepsia 2004;45: [18] Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised clinical and electroencephalographic classification of epileptic seizures. Epilepsia 1981;22:

6 P. Kwan et al. / Epilepsy & Behavior 11 (2007) [19] Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised clinical classification of epilepsies and epileptic syndromes. Epilepsia 1989;30: [20] Hamdy SI, Hiratsuka M, Narahara K, et al. Genotype and allele frequencies of TPMT, NAT2, GST, SULT1A1 and MDR-1 in the Egyptian population. Br J Clin Pharmacol 2003;55: [21] Schaid DJ, Jacobsen SJ. Biased tests of association: comparisons of allele frequencies when departing from Hardy Weinberg proportions. Am J Epidemiol 1999;149: [22] Nakamura T, Sakaeda T, Horinouchi M, et al. Effect of the mutation (C3435T) at exon 26 of the MDR1 gene on expression level of MDR1 messenger ribonucleic acid in duodenal enterocytes of healthy Japanese subjects. Clin Pharmacol Ther 2002;71: [23] Sakaeda T, Nakkamura T, Horinouchi M, et al. MDR1 genotyperelated pharmacokinetics of digoxin after single oral administration in healthy Japanese subjects. Pharm Res 2001;18: [24] Goto M, Masuda S, Saito H, et al. C3435T polymorphism in the MDR1 gene affects the enterocyte expression level of CYP3A4 rather than Pgp in recipients of living-donor liver transplantation. Pharmacogenetics 2002;12: [25] Morita N, Yasumori T, Nakayama K. Human MDR1 polymorphism: G2677T/A and C3435T have no effect on MDR1 transport activities. Biochem Pharmacol 2003;65: