Human Genetics to Identify Therapeutic Targets for NAFLD: Challenges and Opportunities

Abstract

Non-alcoholic fatty liver disease (NAFLD) is a continuous progression of pathophysiologic stages that is challenging to diagnose due to its inherent heterogeneity and poor standardization across a wide variety of diagnostic measures. NAFLD is heritable, and several loci have been robustly associated with various stages of disease. In the past few years, larger genetic association studies using new methodology have identified novel genes associated with NAFLD, some of which have shown therapeutic promise. This mini-review provides an overview of the heterogeneity in NAFLD phenotypes and diagnostic methods, discusses genetic associations in relation to the specific stages for which they were identified, and offers a perspective on the design of future genetic mapping studies to accelerate therapeutic target identification.

Keywords: GWAS; NAFLD; NASH; exomes; gene discovery; human genetics.

Figure 1

The spectrum of NAFLD and stage specific clinical measures. In NAFL, at least 5% of the hepatocytes have fat accumulation in the form of large lipid droplets in the cells that displace the nuclei (21) or many small lipid droplets (22). NAFL can also include inflammation. The transition to NASH occurs with hepatocellular injury in the form of ballooning and further inflammation. Fibrosis can develop in NASH and advance into cirrhosis, in which the liver shrinks and hardens. NAFL and NASH are reversible, as indicated by the rates regressing in severity. Both can also progress into cirrhosis. The rates of transition between each stage are broad ranges because they originate from studies with varying cohort sizes, time frames, treatments, and other variables (, , –26). Histology is the gold standard for diagnosing NAFLD and classifying the stage of disease. The components of NAS are listed above, with the area to the left of the white bars indicating scores of 0 for each feature. A selection of noninvasive diagnostic methods are shown below, with white boxes representing their range in effectively diagnosing different stages of NAFLD [ultrasound (27); FAST (28); MEFIB (29); SteatoTest (30); FibroMax (31); BARD (32); NAFLD fibrosis score (33); NIS4 (34); NASHTest (35)]. The lists are not comprehensive, and the modalities mentioned in the mini-review are emphasized in bold. CAP, controlled attenuation parameter; obtained from FibroScan (36, 37). MRE, magnetic resonance elastography (38). SWE, shear-wave elastography (38). ARFI, acoustic radiation force impulse (37). FIB-4, fibrosis-4 (39). FLI, fatty liver index (40). HSI, hepatic steatosis index (41). ELF, Enhanced Liver Fibrosis (42). APRI, AST-to-platelet ratio index (43). CK-18, cytokeratin 18 (44). CXCL10, C-X-C motif chemokine ligand 10 (45). FGF21, fibroblast growth factor 21 (44). PRO-C3, plasma collagen type III (46).

Figure 2

Heterogeneity and pleiotropy of genes associated with NAFLD phenotypes. Loci with their corresponding genes which have been shown to be associated with NAFLD through genome- and/or exome-wide association studies are displayed on the right side of the chord diagram, indicated by grey bars. On the left side of the diagram are the NAFLD states with their associated measurements, each represented by a different color. Each ribbon represents a significant association identified between each gene and the described state. * indicates histologically defined state (, , , –, –96).

Figure 3

Current state of statistical power to detect genetic associations with NAFLD. Statistical power across odds ratios per allele frequencies computed with a type 1 error rate set at α=0.05 shown for current sample size of NAFLD population studies (left) and for an idealized future NAFLD case/control study (right). Power for frequencies and odds ratios regions are shaded from red to white, in which red indicates regions where statistical power is not sufficient to detect an association, and white areas are within the detection limit. Gray areas highlight the section of the power curve for associating rare genetic variants (MAF < 0.025). Associated loci are overlaid according to their effect size and minor allele frequencies. For example, PNPLA3 rs62021874, which has been robusted associated with NAFLD across several studies, is well within the detectable region given its minor allele frequency of 0.28 and odds ratio of 1.8, whereas PYGO1 (frequency = 0.05, odds ratio = 1.3), which has been associated in a single study, is nominally powered to detect an association given the current sample size of NAFLD population studies (56, 82, 85, 93).