Gene-wide characterization of common quantitative trait loci for ABCB1 mRNA expression in normal liver tissues in the Chinese population

Abstract

In order to comprehensively screen genetic variants leading to differential expression of the important human ABCB1 gene in the primary drug-metabolizing organ, ABCB1 mRNA expression levels were measured in 73 normal liver tissue samples from Chinese subjects. A set of Tag SNPs. were genotyped. In addition, imputation was performed within a 500 kb region around the ABCB1 gene using the reference panels of 1,000 Genome project and HapMap III. Bayesian regression was used to assess the strength of associations by compute Bayes Factors for imputed SNPs. Through imputation and linkage disequilibrium analysis, the imputed loci rs28373093, rs1002205, rs1029421, rs2285647, and rs10235835, may represent independent and strong association signals. rs28373093, a polymorphism 1.5 kb upstream from the ABCB1 transcription start site, has the strongest association. 2677 G>A/T and 3435C>T confer a clear gene-dosage effect on ABCB1 mRNA expression. The systematic characterization of gene-wide common quantitative trait loci associated with ABCB1 mRNA expression in normal liver tissues would provide the candidate markers to ABCB1-relevant clinical phenotypes in Chinese population.

Figure 1. Association results of imputed loci using 1000 Genome datasets as reference panel.

The upper panel illustrates log-transformed BFs of single SNPs with Bayesian regression. The black dots are the imputed loci showing the associations with ABCB1 gene expression. The red dots represent Tag SNPs with significant associations. The lower panel indicates the infor measures of imputed loci. Physical coordinates in the figure are based on Human Reference Genome Sequence Build 37.

Figure 2. LD plot of SNPs with top-ranked BFs in CHB of 1000 Genome Phase I.

The colors indicate the strength of pairwise LD according to r ² metrics. The SNPs marked with asterisks represent independent strong associations. Tag SNPs are shadowed in pink.

Figure 3. Multi-SNP results of Bayesian analyses.

(A) BF values of pairwise loci combinations in Multi-SNP analyses. The colors are corresponding to BFs. (B) BF values and posterior probabilities for different multi-SNP combinations. Means of BFs corresponding to different SNP combinations are charted in line, and means of posterior probabilities in bars.

Figure 4. Allele specific ABCB1 mRNA expression in normal liver samples and the phases of Tag SNPs with significant associations.

In the upper panel, allelic expression ratios of samples heterozygous for at least one marker SNP are presented by bar plots. Each allelic expression ratio was independently measured in triplicate using SNaPshot. In the lower panel, each pair of columns represents the two phases of associated Tag SNPs in each sample heterozygous for marker SNP. Alleles associated with higher gene expression are denoted in pink, and alleles associated with reduced expression in cyan.

Figure 5. Relationship between ABCB1 mRNA expression and the genotype combinations of 2677G>A/T and 3435C>T.

The increase of T alleles at both loci is associated with reduced ABCB1 gene expression.

Figure 6. Allelic expression ratios resulted from collaborative effects of different alleles of multiple functional SNPs.

Allele M1 and M2 of the marker SNP indicating AEI correspond to the alleles of single cis-acting SNP or the haplotypes consisting of multiple cis-acting SNPs which are in the same phase with M1 and M2. Positive-effect alleles associated with high gene expression are denoted in pink, and negative-effect alleles associated with low expression are denoted in cyan. In the left panel, allelic expression ratios clearly indicate the effect of single cis-acting SNP. In the right panel, allelic expression ratios are blurred by multiple cis-acting effects.