Pharmacogenomic study of anti-tuberculosis drugs-induced liver injury in Thais

概要

Anti-tuberculosis drugs-induced liver injury (ATDILI) is the common side effect leading to TB treatment disruption. The mechanism of disease is still poorly understood. I conducted comprehensive studies to identify all possible genetic factors and proposed the predictive system for ATDILI in Thais.

This study carried out in Thai tuberculosis patients including 79 ATDILI cases and 239 tolerant controls from our network hospitals in Thailand. In discovery state of a Genome Wide Association Study (GWAS), I identified the strong association signals from chromosome 8 originated from NAT2 region. The A allele of rs1495741, the top SNP at intergenic region of NAT2 and PSD3 (14 kb distance from NAT2), was strongly associated with ATDILI (recessive model: OR = 6.01, 95%CI = 3.42-10.57, P = 6.86E-11). The genotypes of this particular SNP were mostly found in agreement with NAT2 predicted phenotypes, such as the genotype of AA with the slow acetylators, the genotype of AG with the intermediate acetylators, and the genotype of GG with the fast acetylators. In this study, the results demonstrated 94.98% concordant rate between the NAT2 genotypes and NAT2 predicted phenotypes.

Since the replicated sample set was not available for Thai ATDILI GWAS, a GWAS meta-analysis seems to be another way to confirm the association and increase power to detect additional genetic factors. I performed GWAS meta-analysis using data from the Indonesian ATDILI GWAS, the Japanese ATDILI GWAS, and the Thai ATDILI GWAS. The GWAS was performed with the same quality control criteria. The association results were combined in meta-analysis. Data of 74 cases and 186 controls from the Indonesian study, 70 cases and 270 controls from the Japanese study were combined with the 79 cases and 239 controls Thai GWAS. A Total 221 cases and 659 controls with 481,313 shared SNPs were included in this meta-analysis. The results showed similar signal at the NAT2 region on chromosome 8. No other candidate was detected in this meta-analysis.

So far, only SNPs in NAT2 showed genome-wide association with ATDILI. There are numerous candidate gene studies targeting on this certain gene with the risk of ATDILI in many populations. Previous meta-analysis has already confirmed the remarkable role of the NAT2 slow acetylator and the risk of ATDILI. Additionally, a novel hypothesis was proposed a few years ago about the marked decreased in the metabolizing rate of the NAT2 ultra-slow acetylator within NAT2 slow acetylator group. Therefore, I collected the reported data from published literatures and re-classified them into the NAT2 ultra-slow acetylators. The systematic review and meta-analysis were conducted to evaluate the possible association of the NAT2 ultra-slow acetylators and the risk of ATDILI. As a result, the NAT2 ultra-slow acetylators conferred higher risk with ATDILI when compared with the fast acetylators (OR: 3.60; 95% CI: 2.30-5.63; P = 1.76E-08).

Based on my results, I built the predictive models which included both genetic and non-genetic factors which found to be significant in my study. The best predictive model for Thai ATDILI was constructed from the NAT2 slow-acetylators with Age. The predictive formula for Thai ATDILI was Log (odds) of ATDILI = -2.642 + (1.688NAT2 slow acetylator) + (0.018 Age).

In this study, I confirmed the strong association of NAT2 from the GWAS of ATDILI for the first time and confirmed that NAT2 was a shared genetic factor for ATDILI at least among three populations Indonesia, Japan, and Thailand. The novel ultra-slow acetylator subgroup tended to have higher impact to the risk of ATDILI. I also constructed the predictive system, which might be useful for clinical setting in Thailand.