Genetic polymorphisms of the human PNPLA3 gene are strongly associated with severity of non-alcoholic fatty liver disease in Japanese

Abstract

Background: Nonalcoholic fatty liver disease (NAFLD) includes a broad range of liver pathologies from simple steatosis to cirrhosis and fibrosis, in which a subtype accompanying hepatocyte degeneration and fibrosis is classified as nonalcoholic steatohepatitis (NASH). NASH accounts for approximately 10-30% of NAFLD and causes a higher frequency of liver-related death, and its progression of NASH has been considered to be complex involving multiple genetic factors interacting with the environment and lifestyle.

Principal findings: To identify genetic factors related to NAFLD in the Japanese, we performed a genome-wide association study recruiting 529 histologically diagnosed NAFLD patients and 932 population controls. A significant association was observed for a cluster of SNPs in PNPLA3 on chromosome 22q13 with the strongest p-value of 1.4 × 10(-10) (OR = 1.66, 95%CI: 1.43-1.94) for rs738409. Rs738409 also showed the strongest association (p = 3.6 × 10(-6)) with the histological classifications proposed by Matteoni and colleagues based on the degree of inflammation, ballooning degeneration, fibrosis and Mallory-Denk body. In addition, there were marked differences in rs738409 genotype distributions between type4 subgroup corresponding to NASH and the other three subgroups (p = 4.8 × 10(-6), OR = 1.96, 95%CI: 1.47-2.62). Moreover, a subgroup analysis of NAFLD patients against controls showed a significant association of rs738409 with type4 (p = 1.7 × 10(-16), OR = 2.18, 95%CI: 1.81-2.63) whereas no association was obtained for type1 to type3 (p = 0.41). Rs738409 also showed strong associations with three clinical traits related to the prognosis of NAFLD, namely, levels of hyaluronic acid (p = 4.6 × 10(-4)), HbA1c (p = 0.0011) and iron deposition in the liver (p = 5.6 × 10(-4)).

Conclusions: With these results we clearly demonstrated that Matteoni type4 NAFLD is both a genetically and clinically different subset from the other spectrums of the disease and that the PNPLA3 gene is strongly associated with the progression of NASH in Japanese population.

Figure 1. Manhattan plot of the GWA study.

Association p-values are calculated by exact trend test and plotted along the chromosome in −log₁₀ scale. The horizontal line indicates Bonferroni-adjusted significance threshold (p = 1.03×10⁻⁷).

Figure 2. A schematic organization of the human PNPLA3 locus at 22q13.31 with the genome scan results.

P-values calculated by the exact trend test were plotted in –log₁₀ scale. Red and blue dots indicate the p-values of genotyped and imputed SNPs, respectively. Local recombination rate obtained from HAPMAP release 22 is indicated by a red line plotted in cm/Mb scale. The structure and orientation of four genes in the region are shown below the plots with their transcriptional orientations according to NCBI Reference Sequence Build 36.3. LD blocks were generated according to pairwise LD estimates of the SNPs located within the region using the genome scan results. The LD block showing the strongest association is highlighted with the triangle, and the corresponding chromosomal region is represented by the dotted lines.

Figure 3. Histogram of odds ratios for genotype distribution of rs738409 between Matteoni types.

Each box denotes the odds ratio (OR) comparing the corresponding Matteoni types on the horizontal axes. N represents the number of samples. Odds ratios and p-values are calculated for the higher Matteoni type per risk allele (G) on additive model by multivariable logistic regression adjusted for age, sex and BMI, and are shown with 95% CI above each box.