LAUSR

Most common human diseases develop because of a combination of endogenous risk combined with exposures. Said another way, these diseases are attributed to genetic risk combined with environmental triggers. Nonalcoholic fatty liver disease (NAFLD) is no exception. There have been both genetic and environmental variables that have been associated with NAFLD. One recent article, however, goes beyond these associations to understand how environmental exposures in the setting of genetic risk variants can exacerbate the development of NAFLD. Genetic associations studies usually assume an additive genetic model to assess for association with disease. This translates to a model where having 0, 1, or 2 alleles of a genetic variant increases the risk of the disease linearly. However, there are situations where those same alleles instead of increasing risk additively increase risk multiplicatively. That is, having 0, 1, or 2 alleles increases risk under one set of conditions (body mass index [BMI] <25 kg/m) from 1 to 2 to 4 and another set of conditions (BMI >35 kg/m) from 1 to 8 to 16 (Fig. 1A). The condition under which the effects of the same genotype changes is called a modifier. Identifying environmental modifiers of genetic risk can be done by looking for gene environment interactions. It is important to identify modifiers of disease because if the modifier can be changed, the increased genetic risk for development of disease can be preferentially and dramatically mitigated. A recent article by Stender et al. examined the effect of adiposity on the risk of developing NAFLD in individuals that carried genetic variants encoding patatin-like phospholipase domain containing 3 (PNPLA3) p.I148M (isoleucine to methionine at position 148), TM6SF2 (transmembrane 6 superfamily member 2) p.E167K (glutamic acid to lysine at position 167), and GCKR (glucokinase regulator) p.P446L (proline to leucine). Associations of these variants with hepatic steatosis, nonalcoholic steatohepatitis, and fibrosis has been reported in many cohorts and ancestries. This article examined whether the effects of these variants on development of NAFLD were modified by BMI. They showed that as BMI increases from less than 25 kg/m to 25-30 kg/m to 30-35 kg/m and to >35 kg/m, the effect of all three genes on hepatic triglyceride content percent as measured using magnetic resonance spectroscopy in the Dallas Heart Study also increased. The difference in hepatic triglyceride percent median increased from 0.98% in the low-risk PNPLA3-II and PNPLA3-MM, BMI <25 kg/m individuals to up to 9.4% in PNPLA3-II and PNPLA3-MM, BMI >35 kg/m individuals (Fig. 1B for PNPLA3). Similar interactions were noted by another group using visceral adipose tissue and computed tomography measured fatty liver. They repeated analyses using alanine aminotransferase (ALT) as the outcome to test whether it also affected hepatic injury. They found a significant interaction for PNPLA3-M with serum ALT levels where the difference in ALT between PNPLA3-II and PNPLA3-MM, BMI <25 kg/m individuals was 1.0 IU/L up to 7 IU/L between PNPLA3-II and PNPLA3-MM, BMI >35 kg/m individuals in the Copenhagen Cohort (Fig. 1C). These data corroborate a similar result reported for PNPLA3-M with liver function tests in an Italian cohort. Stender et al. also noted that the difference in odds ratio (OR) between PNPLA3-II and PNPLA3-MM, BMI <25 kg/m individuals of 1.5-4.8 between PNPLA3-II and PNPLA3-MM, BMI >35 kg/m individuals suggesting an interaction with this liver disease measure also (Fig. 1D). They observed statistically significant, but smaller, interactions for the GCKR and TM6SF2 variants in affecting hepatic triglyceride content, but not with ALT and cirrhosis. This may be attributed to biology or to low power to detect an effect for these variants for ALT and cirrhosis. Even though the percentage difference in liver fat and ALT in Stender et al. across obesity groups seems small, these estimates come from cross-sectional measures that can sum to larger effects over a lifetime.