Rare Pathogenic Variants Predispose to Hepatocellular Carcinoma in Nonalcoholic Fatty Liver Disease

Abstract

Nonalcoholic fatty liver disease (NAFLD) is a rising cause of hepatocellular carcinoma (HCC). We examined whether inherited pathogenic variants in candidate genes (n = 181) were enriched in patients with NAFLD-HCC. To this end, we resequenced peripheral blood DNA of 142 NAFLD-HCC, 59 NAFLD with advanced fibrosis, and 50 controls, and considered 404 healthy individuals from 1000 G. Pathogenic variants were defined according to ClinVar, likely pathogenic as rare variants predicted to alter protein activity. In NAFLD-HCC patients, we detected an enrichment in pathogenic (p = 0.024), and likely pathogenic variants (p = 1.9*10^-6), particularly in APOB (p = 0.047). APOB variants were associated with lower circulating triglycerides and higher HDL cholesterol (p < 0.01). A genetic risk score predicted NAFLD-HCC (OR 4.96, 3.29-7.55; p = 5.1*10^-16), outperforming the diagnostic accuracy of common genetic risk variants, and of clinical risk factors (p < 0.05). In conclusion, rare pathogenic variants in genes involved in liver disease and cancer predisposition are associated with NAFLD-HCC development.

Figure 1

Study design. (a) Study cohorts composition and enrolment criteria. (b) Study flow-chart. NAFLD: non-alcoholic fatty liver disease. HCC: hepatocellular carcinoma.

Figure 2

Enrichment in pathogenic variants in patients with NAFLD-HCC. (a) Frequency of pathogenic variants (mutational rate %: sum of mutated/total alleles) in NAFLD-HCC cases vs. controls. (b) Frequency of likely pathogenic variants (rare variants with high likelihood of altering protein activity) in NAFLD-HCC cases vs. controls. *p < 0.05; **p < 0.01; ***p < 0.005 by Fisher’s exact test.

Figure 3

Genes enriched in pathogenic variants. Comutation plot showing the distribution of rare pathogenic variants (upper panel), as well as of common genetic variants (bottom panel) predisposing to hepatic fat accumulation and NAFLD-HCC in the 251 individuals of the EPIDEMIC project. Genes significantly enriched in variants in cases vs. controls are marked by and asterisk (*by Fisher’s exact test). Gene abbreviations are shown in Supplementary Material.

Figure 4

Genetic risk score. Diagnostic accuracy of the Genetic risk score (GRS) for NAFLD-HCC in the 655 individuals included in the study. (a) comparison of the diagnostic accuracy in the study cohorts; p < 0.05 for diagnostic accuracy in the overall vs. single cohorts. (b) Diagnostic accuracy of GRS vs. PNPLA3 I148M variant alone and a combination of PNPLA3 I148M and TM6SF2 E167K variants in determining NAFLD HCC risk; p < 2*10⁻⁸ for alternative genetic scores vs. the overall GRS. (c) Additive value of adding GRS to a diagnostic model based on clinical risk factors, in determining NAFLD-HCC risk in the 251 individuals of the EPIDEMIC study; p = 0.17. Comparison of diagnostic accuracy was performed by two-sided Venkatraman test.