Damaging mutations in LXRα uncouple lipogenesis from hepatotoxicity and implicate hepatic cholesterol sensing in human liver health

The nuclear receptor Liver X Receptor-α (LXRα) activates lipogenic gene expression in hepatocytes. Its inhibition has therefore been proposed as a strategy to treat metabolic-dysfunction-associated steatotic liver disease (MASLD). In order to understand the impact of reducing LXRα activity on human health we first examined the association between the carriage of rare loss of function mutations in NR1H3 (encoding LXRα) and metabolic and hepatic phenotypes. We identified 63 rare predicted damaging variants in the ligand binding domain of LXRα in 454,787 participants in UK Biobank. On functional characterisation, 42 of these were found to be severely impaired. Consistent with loss of the lipogenic actions of LXRα, carriers of damaging mutations in LXRα had reduced serum triglycerides (ß=-0.13 s.d. ±0.03, P=2.7x10 ^-5 , N(carriers)=971). Surprisingly, these carriers also had elevated concentrations of serum liver enzymes (e.g. ALT: ß=0.17s.d. ±0.03, P=1.1x10 ^-8 , N(carriers)=972) with a 35% increased risk of clinically significant elevations in ALT (OR=1.32, 95%CI:1.15-1.53, P=1.2x10 ^-4 , N(carriers)=972), suggestive of hepatotoxicity. We generated a knock-in mouse carrying one of the most severely damaging mutations ( Nr1h3 p.W441R) which we demonstrated to have dominant negative properties. Homozygous knock-in mice rapidly developed severe hepatitis and fibrotic liver injury following exposure to western diet despite markedly reduced steatosis, liver triglycerides and lipogenic gene expression. This phenotype was completely rescued by viral over-expression of wildtype LXRα specifically in hepatocytes, indicating a cell-autonomous effect of the mutant on hepatocyte health. While homozygous LXRα knockout mice showed some evidence of hepatocyte injury under similar dietary conditions, the phenotype of the LXRα ^W441R/W441R mouse was much more severe, suggesting that dominant negative mutations that actively co-repress target genes can result in pathological impacts significantly more severe than those seen with simple absence of the receptor. In summary, our results show that loss of function mutations in LXRα occur in at least 1/450 people and are associated with evidence of liver dysfunction. These findings implicate LXRα in the maintenance of human liver health, identify a new murine model of rapidly progressive fibrotic liver disease and caution against LXR antagonism as a therapeutic strategy for MASLD.