1Department of Medical Biology, Faculty of Medicine, Ankara University, Ankara, .... Freq., frequency; Obs., observed; Exp., expected; HWE, HardyâWeinberg ...
DNA AND CELL BIOLOGY Volume 29, Number 2, 2010 ª Mary Ann Liebert, Inc. Pp. 83–90 DOI: 10.1089=dna.2009.0953
Haplotype-Based Analysis of MDR1/ABCB1 Gene Polymorphisms in a Turkish Population Gu¨vem Gu¨mu¨s¸-Akay,1 Aydin Ru¨stemog˘lu,2 Aynur Karadag˘,1 and Asuman Sungurog˘lu1
The three functional single-nucleotide polymorphisms (SNPs) of the MDR1 gene, C1236T, G2677T=A, and C3435T, exhibit an interpopulation difference. In this study, we analyzed the haplotype frequencies of these three SNPs in 107 unrelated healthy Turkish subjects and compared them with those of other reported populations. We found that C1236T, G2677T=A, and C3435T SNPs are expected to be structured in 10 different haplotypes, with 4 prominent haplotypes T-T-T (33.7%), C-G-C (25.0%), T-G-C (10.9%), and C-T-T (8.7%). There was a statistically significant linkage disequilibrium between all C1236T, G2677T=A, and C3435T SNPs ( p < 0.0001); however, our results indicated that only loci 2677 and 3435 show relatively strong linkage disequilibrium (Lewontin’s coefficient [D0 ] ¼ 0.74, Pearson’s correlation [r2] ¼ 0.47). The haplotype frequency distribution of our study group was found to be significantly different from that in Han Chinese, Uygur Chinese, Kazakh Chinese, Indian, Malay, Japanese, Caucasian, and Ashkenazi Jewish populations ( p < 0.0001). The results of this study may contribute to population-specific haplotype data on the MDR1 gene and may serve as a basis for studies on response to P-glycoprotein substrate drugs as well as for future association studies of certain diseases in Turkish population.
et al., 2000). Thereafter, a great number of studies on the effects of C1236T (rs1128503), C3435T, and G2677T=A (rs2032582) SNPs on MDR1 expression and function in tissues and also association with drug response in various disease conditions have been conducted repeatedly in different populations (Hoffmeyer et al., 2000; Kroetz et al., 2003; Sai et al., 2003; Jamroziak et al., 2004; Sakaeda, 2005; Xhemo et al., 2007; Kato et al., 2008). However, discordant results have been obtained (Hoffmeyer et al., 2000; Sakaeda et al., 2001; Goh et al., 2002; Illmer et al., 2002; Kim, 2002) and led to the suggestion that grouping these polymorphisms into haplotypes may provide a better understanding of the observed inconsistencies and serve as a useful predictor of the functional consequences of MDR1 polymorphisms. The frequencies of variant MDR1 alleles differ, depending on racial background (Tang et al., 2002; Lee et al., 2005; Kimchi-Sarfaty et al., 2007a), raising the need for generating the allelic, genotypic, and particularly haplotypic data for each individual population. The aims of this study were to scrutinize how C1236T (Gly412Gly) in exon 12, G2677T=A (Ala893Ser=Thr) in exon 21, and C3435T (Ilc1145Ile) in exon 26 SNPs are structured in haplotypes in a representative group of the Turkish population and how the haplotype frequencies vary from those of other reported populations.
Introduction
P
-glycoprotein (P-gp), also known as MDR1 or ABCB1, has been identified as a mediator of the multidrug resistance phenotype in tumor cells treated with a variety of antitumor agents (Riordan et al., 1985). This protein is a member of the ATP-binding cassette superfamily and is composed of 1280 amino acids with two homologous halves containing six putative hydrophobic transmembrane segments and an intracellular binding site for ATP (Tang et al., 2002; Sakaeda, 2005). In addition to tumor cells, P-gp is also widely expressed in normal tissues, including the liver, kidneys, small and large intestines, brain, testis, muscle tissue, placenta, and adrenals (Sakaeda, 2005), and has been shown to prevent intracellular accumulation of potentially toxic substances and metabolites (Schinkel, 1997). Human P-gp is encoded by the MDR1 (ABCB1) gene, which is located on chromosomal region 7q21.1 (Callen et al., 1987). To date, over 60 single-nucleotide polymorphisms (SNPs) have been identified in the coding region of MDR1 gene, of which more than 20 are known to be silent (www.ncbi.nlm.nih.gov= SNP=snp_ref.cgi?chooseRs¼coding&locusId¼5243&mrna, accessed January 5, 2009). Among the first SNPs to be identified is C3435T (rs1045642), which has been shown to correlate with expression levels and function of P-gp (Hoffmeyer
1
Department of Medical Biology, Faculty of Medicine, Ankara University, Ankara, Turkey. Department of Medical Biology, Faculty of Medicine, Gaziosmanpas¸ a University, Tokat, Turkey.
2
83
¨ MU ¨ S¸ -AKAY ET AL. GU
84 Materials and Methods Study population One hundred and seven unrelated healthy Turkish subjects living in Turkey for at least three generations were included in this study. DNA samples were obtained from the DNA bank of the Department of Medical Biology, Faculty of Medicine, Ankara University. All the samples had been provided by the healthy volunteers who previously participated in research projects conducted by the department after their written informed consent was obtained, indicating permission for anonymous use of their DNA samples in other studies. The study was performed in accordance with the Declaration of Helsinki and approved by the Research Ethics Committee of Ankara University Faculty of Medicine. Genotyping Polymerase chain reaction–restriction fragment length polymorphism was used for detection of C1236T, G2677T=A, and C3435T SNPs as previously described with slight modifications (Wasilewska et al., 2007; Kato et al., 2008). Details with respect to primers, polymerase chain reaction conditions, restriction enzymes, and digestion conditions are given in Table 1. Genotypes were determined by agarose gel electrophoresis of restriction digests on 3% Nusieve GTG agarose (BioWhittaker Molecular Applications [BMA], Rockland, ME) gel containing 0.5 mg=mL ethidium bromide. Data analysis All of the genetic data analyses were performed using Arlequin software, version 3.11. Fisher’s exact test was used to assess the genotype frequencies for each SNP in the study group for deviation from Hardy–Weinberg equilibrium. The haplotype frequencies were estimated based on the expectation-maximization algorithm (Excoffier and Slatkin, 1995). Linkage disequilibrium (LD) between SNP pairs was estimated by Lewontin’s coefficient (D0 ) and Pearson’s correlation (r2) (Lewontin and Kojima, 1960; Lewontin, 1964). Gene ˆ ) was calculated using the following formula: diversity (H ˆ ¼ n (1 Pk pi2 ), where n is the number of gene copies H i¼1 n1 in the sample, k is the number of haplotypes, and pi is the sample frequency of the ith haplotype. Standard deviation of qffiffiffiffiffiffiffiffiffiffiffi ˆ ¼ V(H) ˆ , where V is ˆ was computed as follows: SD (H) the H
ˆ (Nei, 1987). Pairwise comparisons the sampling variance of H of the allele and haplotype frequencies obtained from this study with the previously published data for different populations in the literature were done using Fisher’s exact test. These comparisons were made only for the populations for whom all haplotypic data were available. A p-value of less than 0.05 was considered statistically significant. Results The allelic, genotypic, and haplotypic frequency distributions of C1236T, G2677T=A, and C3435T genetic variants in the MDR1 gene were analyzed in 107 Turkish subjects. Allele and genotype frequencies of these three SNPs are given in Table 2. Frequencies of the wild-type C and mutant T alleles of 1236 locus were found as 45.8% and 54.2%, consecutively. Twenty-two (20.6%), 54 (50.5%), and 31 (28.9%) individuals had the genotypes homozygous wild-type CC, heterozygous CT, and homozygous mutant TT for exon 12 C1236T SNP, respectively. The allelic distribution of G2677T=A variant was calculated as 45.8% for G, 48.6% for T, and 5.6% for A. When the genotypes of locus 2677 were surveyed, the most frequently observed genotype was GT (48 individuals, 44.9%), followed by TT (24 individuals, 22.4%), GG (23 individuals, 21.5%), TA (8 individuals, 7.5%), and GA (4 individuals, 3.7%), respectively. No individual showed an AA genotype for this locus. The occurrences of the C and T alleles for the C3435T SNP were 47.7% and 52.3%, respectively. The genotype distribution of C3435T variant was as follows: 27.1% (29 individuals) for CC, 41.1% (44 individuals) for CT, and 31.8% (34 individuals) for TT. Genotype frequencies of the C1236T, G2677T=A, and C3435T SNPs were all in Hardy–Weinberg equilibrium ( p > 0.05; Table 2). The frequencies of linkages and degree of LD between different pairs of SNPs are tabulated in Table 3. When the linkage between C1236T and G2677T=A was analyzed, the most frequent haplotype was estimated as T-T with a frequency of 37.4%, followed by C-G (30.4%), T-G (15.4%), C-T (11.2%), C-A (4.7%), and T-A (0.9%), respectively. The D0 and r2 coefficients were calculated as 0.50 and 0.20, respectively, indicating moderate LD between loci 1236 and 2677 ( p < 0.0001). As to linkage between C1236T and C3435T SNPs, the frequencies of the estimated haplotypes were as follows: T-T (38.3%), C-C (32.2%), T-C (15.4%), and C-T (14.0%), in decreasing frequency. Somewhat moderate LD was found between loci 1236 and 3435 (D0 ¼ 0.42, r2 ¼ 0.17, p < 0.0001). The
Table 1. Primers, Restriction Enzymes, and Digestion Conditions Used for Genotyping Analysis of MDR1 Single-Nucleotide Polymorphisms SNPs C1236T G2677T G2677A C3435T
Primers
Restriction enzymes
Digestion conditions
F: 50 -TATCCTGTGTCTGTGAATTGCC-30 R: 50 -CCTGACTCACCACACCAATG-30 F: 50 -TGCAGGCTATAGGTTCCAGG-30 R: 50 -TTTAGTTTGACTCACCTTCCCG-30 F: 50 -TGCAGGCTATAGGTTCCAGG-30 R: 50 -GTTTGACTCACCTTCCCAG-30 F: 50 -TGTTTTCAGCTGCTTGATGG-30 R: 50 -AAGGCATGTATGTTGGCCTC-30
HaeIII
378C for 16 h
BanI
378C for 16 h, 658C for 16 h
BsrI
658C for 16 h, 808C for 20 min
Sau3AI
378C for 16 h
Polymerase chain reaction (PCR) conditions: 948C for 2 min (initial denaturation); 948C for 30 s, 608C for 30 s, 728C for 30 s (35 cycles); 728C for 7 min (final extension). SNP, single nucleotide polymorphism.
MDR1 GENE POLYMORPHISMS IN A TURKISH POPULATION
85
Table 2. Allele and Genotype Frequencies of MDR1 Single-Nucleotide Polymorphisms in a Turkish Population
SNP
Amino acid change
Position Exon
n n (Total ¼ 214 Geno- (Total ¼ 107 Allele chromosomes) Freq. type individuals)
rs1128503
1236
12
G412G
C T
98 116
0.458 0.542
rs2032582
2677
21
S893A=T
G T A
98 104 12
0.458 0.486 0.056
rs1045642
3435
26
I1145I
C T
102 112
0.477 0.523
CC CT TT GG GT GA TT TA AA CC CT TT
22 54 31 23 48 4 24 8 0 29 44 34
Frequency
Heterozygosity
Obs.
Exp.
Obs.
Exp.
P for HWE
0.206 0.505 0.289 0.215 0.449 0.037 0.224 0.075 0 0.271 0.411 0.318
0.210 0.496 0.294 0.210 0.445 0.051 0.236 0.055 0.003 0.228 0.498 0.274
0.505
0.496
1.000
0.561
0.551
0.716
0.411
0.498
0.501
Freq., frequency; Obs., observed; Exp., expected; HWE, Hardy–Weinberg equilibrium.
of the C1236T locus were found to be different from those in Ashkenazi Jewish ( p ¼ 0.015), Han Chinese ( p ¼ 0.030), Kazakh Chinese ( p ¼ 0.008), Japanese ( p ¼ 0.009), and Malay ( p ¼ 0.004) populations, but similar to Caucasian, Uygur Chinese, and Indian populations ( p > 0.05). When the allele frequency distribution of G2677T=A polymorphism in the Turkish population was compared with that in other populations, significant differences were found between the Turkish population and Ashkenazi Jewish ( p < 0.001), Han Chinese ( p ¼ 0.012), Japanese ( p ¼ 0.001), and Malay ( p ¼ 0.039) populations. However, allele frequency distribution of the G2677T=A SNP in our study group did not show any significant differences from that in Caucasian, Kazakh Chinese, Uygur Chinese, or Indian ( p > 0.05) populations. Pairwise comparisons with other populations indicated that the allele frequencies of the Turkish population in this study is significantly different at C3435T SNP from Han Chinese ( p ¼ 0.002), Kazakh Chinese ( p ¼ 0.012), Japanese ( p ¼ 0.010), and Malay ( p ¼ 0.002) populations, but similar to Ashkenazi Jewish, Caucasian, Uygur Chinese, and Indian ( p > 0.05) populations. As can be seen from the lower section of Table 5, the haplotype frequency distribution of our study group was found to be significantly different from that in all the other populations ( p < 0.0001).
haplotype frequency distribution regarding loci 2677 and 3435 was estimated as T-T 42.5%, G-C 36.0%, G-T 9.8%, T-C 6.1%, A-C 5.6%, and A-T 0%. The D0 and r2 were 0.74 and 0.47, respectively, representing relatively strong LD between these two loci ( p < 0.0001). In this study, 24 different genotype combinations composed of three SNPs (C1236T-G2677T=A-C3435T) were determined (Table 4). The most frequently observed genotype combinations were CT-GT-CT (21 individuals, 19.6%), TTTT-TT (16 individuals, 15.0%), CT-GG-CC (8 individuals, 7.5%), CC-GG-CC (7 individuals, 6.5%), CT-GG-CT (7 individuals, 6.5%), and CT-GT-TT (7 individuals, 6.5%), respectively. The haplotype frequencies concerning all C1236T, G2677T=A, and C3435T SNPs are presented in Table 4. We found that these three SNPs were expected to be structured in 10 different haplotypes with four prominent haplotypes, T-T-T (33.7%), C-G-C (25.0%), T-G-C (10.9%), and C-T-T (8.7%). The other two possible haplotypes, C-A-T and T-A-T, ˆ was calculated were missing in this survey. In this study, H as 0.798 0.017 for the MDR1 gene (Table 4). A comparison between the allele and haplotype frequencies of the Turkish population and other ethnic groups from previous studies are summarized in Table 5. The allele frequencies
Table 3. Frequencies of Linkages and Degree of Linkage Disequilibriums Between Different Pairs of Single-Nucleotide Polymorphisms C1236T-G2677T=A Haplotype C-A C-G C-T T-A T-G T-T Total
a
n
C1236T-C3435T
Frequency
10 65 24 2 33 80 214 D0 ¼ 0.501, r2 ¼ 0.204
0.047 0.304 0.112 0.009 0.154 0.374 1.000
Haplotype C-C C-T T-C T-T
Total
Data are given in alphabetical order. a Number of gametes=chromosomes. D0 , Lewontin’s coefficient; r2, Pearson’s correlation.
G2677T=A-C3435T
a
Frequency
Haplotype
69 30 33 82
0.322 0.140 0.154 0.383
A-C A-T G-C G-T T-C T-T Total
n
214 1.000 D0 ¼ 0.421, r2 ¼ 0.168
na
Frequency
12 0 77 21 13 91 214 D0 ¼ 0.738, r2 ¼ 0.469
0.056 0.000 0.360 0.098 0.061 0.425 1.000
¨ MU ¨ S¸ -AKAY ET AL. GU
86
Table 4. Frequencies of Genotype Combinations and Haplotypes of the MDR1 Single-Nucleotide Polymorphisms in a Turkish Population Genotype combinations
Haplotypes
C1236T-G2677T=A-C3435T
na
Frequency
CC-GA-CC CC-GG-CC CC-GT-CT CC-GT-TT CC-TA-CC CC-TT-CC CC-TT-TT CT-GG-CC CT-GG-CT CT-GG-TT CT-GT-CC CT-GT-CT CT-GT-TT CT-TA-CT CT-TT-CT CT-TT-TT TT-GA-CC TT-GG-CC TT-GT-CC TT-GT-CT TT-GT-TT TT-TA-CC TT-TT-CC TT-TT-TT Total
3 7 4 2 1 1 3 8 7 1 3 21 7 6 1 2 1 1 2 5 3 1 1 16 107
0.028 0.065 0.037 0.019 0.009 0.009 0.028 0.075 0.065 0.009 0.028 0.196 0.065 0.056 0.009 0.019 0.009 0.009 0.019 0.047 0.028 0.009 0.009 0.150 1.000
ˆ SD H
C1236T-G2677T=A-C3435T
nb
Frequency
C-A-C C-G-C C-G-T C-T-C C-T-T T-A-C T-G-C T-G-T T-T-C T-T-T
10 59 10 4 15 2 19 10 8 77
0.045 0.250 0.054 0.022 0.087 0.011 0.109 0.045 0.039 0.337
214
1.000
Total 0.798 0.017
Data are given in alphabetical order. a Number of individuals. b Number of gametes=chromosomes. ˆ , gene diversity; SD, standard deviation. H
Discussion In the recent years, there has been increasing evidence that SNPs in the MDR1 gene may affect P-gp function both as a potential determinant of interindividual variability in drug disposition (Sai et al., 2003; Wasilewska et al., 2007; Xu et al., 2008) and also in susceptibility to certain diseases including several types of cancer ( Jamroziak et al., 2004; Kurzawski et al., 2005; Turgut et al., 2007). Despite this, there is a substantial controversy about the functional importance of these SNPs, particularly in the cases of C3435T and G2677T=A, on P-gp expression=function. For instance, the highest P-gp activity has been reported in 3435CC homozygotes, intermediate activity in 3435CT heterozygotes, and lowest activity in 3435TT homozygotes ( Jamroziak et al., 2006). In contrast, Illmer et al. have shown significant association of the 2677GG and 3435CC genotypes with lower MDR1 expression (Illmer et al., 2002). Additionally, Calado et al. (2002) implicated similar P-gp function among different G2677T and C3435T genotypes. In fact, three most common SNPs, namely C1236T, G2677T=A, and C3435T, have been shown to be part of a common haplotype, and contradictory results suggested that the main determinant of functional differences in P-gp resides not in SNP differences, but rather in the features of MDR1 haplotypes (Kroetz et al., 2003). To better clarify the interindividual=population differences in many drug responses and
in disease susceptibilities, it is especially important to consider the variability and to analyze the diversity in haplotype structure within and between different ethnic populations. Therefore, this study analyzed the three most common C1236T, G2677T=A, and C3435T SNPs of MDR1 gene in a haplotypic manner and demonstrated the variations from other populations, if any. To the best of our knowledge, this is the first study reporting haplotype-based analysis of common functional polymorphisms of MDR1 gene in a representative group of the Turkish population. In this study, the frequency of the variant T allele for C1236T SNP has been calculated as 54.2%, which is consistent with our previous report with a smaller population, that is, 54.5% (Gu¨mu¨s¸ -Akay et al., 2008). In this study, the frequencies of the T and A alleles for G2677T=A polymorphism have been found as 48.6% and 5.6%, respectively. However, Sapmaz et al. (2008) have reported the frequencies of variant T and A alleles as 47.0% and 0%, respectively, in 70 healthy Turkish subjects. Failure to observe the A allele might have been due to a smaller population size in that study. As for the C3435T SNP, the occurrence of the variant T allele has been found as 52.3%, which is in accordance with other studies previously conducted in a Turkish population (Turgut et al., 2006; Tufan et al., 2007). Synonymous C3435T SNP has been reported to be associated with altered P-gp activity (Hoffmeyer et al., 2000;
MDR1 GENE POLYMORPHISMS IN A TURKISH POPULATION
87
Table 5. Pairwise Allele and Haplotype Frequency Comparisons of C1236T, G2677T=A, and C3435T Single-Nucleotide Polymorphisms Between Turkish Population and Previously Published Populations for Whom All the Haplotypic Data Were Available Allele frequency SNP C1236T
G2677T=A
C3435T
a
Population
N
Turkish (this study) Ashkenazi Jewish Caucasian Chinese (Han) Chinese (Kazakh) Chinese (Uygur) Indian Japanese Malay
107 101 97 165 108 161 62 154 85
Turkish (this study) Ashkenazi Jewish Caucasian Chinese (Han) Chinese (Kazakh) Chinese (Uygur) Indian Japanese Malay
Turkish (this study) Ashkenazi Jewish Caucasian Chinese (Han) Chinese (Kazakh) Chinese (Uygur) Indian Japanese Malay
107 101 98 165 108 161 68 154 93
107 101 99 165 108 161 68 154 92
PDb Turkish
C
T
0.458 0.579 0.541 0.364 0.329 0.407 0.395 0.344 0.306
0.542 0.421 0.459 0.636 0.671 0.593 0.605 0.656 0.694
G
T
A
0.458 0.594 0.500 0.446 0.509 0.450 0.338 0.429 0.575
0.486 0.406 0.464 0.421 0.389 0.506 0.518 0.406 0.360
0.056 0 0.036 0.133 0.102 0.044 0.044 0.165 0.065
C
T
0.477 0.500 0.434 0.621 0.602 0.472 0.397 0.594 0.630
0.523 0.500 0.566 0.379 0.398 0.528 0.603 0.406 0.370
– 0.015 0.108 0.030 0.008 0.235 0.307 0.009 0.004
–
