Deep metabolic phenotyping of humans with protein-altering variants in TM6SF2 using a genome-first approach

Abstract

Background & Aim

An unbiased genome-first approach can expand the molecular understanding of specific genes in disease-agnostic biobanks for deeper phenotyping. TM6SF2 represents a good candidate for this approach due to its known association with steatotic liver disease (SLD).

Methods

We screened participants with whole-exome sequences in the Penn Medicine Biobank (PMBB, n >40,000) and the UK Biobank (UKB, n >200,000) for protein-altering variants in TM6SF2 and evaluated their association with liver phenotypes and clinical outcomes.

Results

Missense variants in TM6SF2 (E167K, L156P, P216L) were associated with an increased risk of clinically diagnosed and imaging-proven steatosis, independent of the PNPLA3 I48M risk allele and hepatitis B/C (p <0.001). E167K homozygotes had significantly increased risk of SLD (odds ratio [OR] 5.38, p <0.001), steatohepatitis (OR 5.76, p <0.05) and hepatocellular carcinoma (OR 11.22, p <0.0001), while heterozygous carriers of L156P and P216L were also at an increased risk of steatohepatitis. In addition, carriers of E167K are at a 3-fold increased risk of at-risk MASH (OR 2.75, p <0.001). CT-derived liver fat scores were higher in E167K and L156P in an allele-dose manner (p <0.05). This corresponded with the UKB nuclear magnetic resonance-derived lipidomic analyses (n = 105,348), revealing all carriers to exhibit lower total cholesterol, triglycerides and total choline. In silico predictions suggested that these missense variants cause structural disruptions in the EXPERA domain, leading to reduced protein function. This hypothesis was supported by the association of rare loss-of-function variants in TM6SF2 with an increased risk of SLD (OR 4.9, p <0.05), primarily driven by a novel rare stop-gain variant (W35X) with the same directionality.

Conclusion

The functional genetic study of protein-altering variants provides insights on the association between loss of TM6SF2 function and SLD and provides the basis for future mechanistic studies.

Impact and implications:

The genome-first approach expands insights into genetic risk factors for steatotic liver disease with TM6SF2 being a focal point due to its known association with plasma lipid traits. Our findings validated the association of two missense variants (E167K and L156P) with increased risk of hepatic steatosis on CT and MRI scans, as well as the risk of clinically diagnosed hepatocellular carcinoma independent of the common PNPLA3 I48M risk variant. Notably, we also identified a predicted deleterious missense variant (P216L) linked to steatotic risk and demonstrated that an aggregated gene burden of rare putative loss-of-function variants was associated with the risk of hepatic steatosis. Combined, this study sets the stage for future mechanistic investigations into the functional consequences of TM6SF2 variants in metabolic dysfunction-associated steatotic liver disease.