MASLD risk variants reveal cell- and stimulus-specific gene regulation
Background: Connecting risk variants to disease mechanisms
A study found that noncoding variants associated with the risk of metabolic dysfunction-associated steatotic liver disease (MASLD) differ in gene-regulatory activity depending on cell type and stimulus conditions. Combining analyses of human liver tissue with cell experiments, the researchers identified hundreds of differentially active variants and examined their links to target genes and processes including lipid metabolism. The study was published in Nature Genetics on June 2, 2026.
Genome-wide association studies (GWAS) have identified more than 100 MASLD-associated loci. However, statistical associations alone cannot explain which genes noncoding variants regulate or how they alter disease risk. In which cells and conditions does a variant’s function become apparent? This study addressed that question, moving beyond locating risk variants.
Key findings: Tracking regulatory activity and target genes
The researchers analyzed chromatin accessibility in nuclei from human MASLD liver tissue. Risk variants were enriched in cell-type-specific regulatory elements, where transcription factors that determine cell identity bound. They then used massively parallel reporter assays (MPRA) to assess differences in regulatory activity between alleles. Chromatin accessibility and reporter activity measure distinct properties.
The public extended-data description lists 5,369 MASLD-related variants assessed by MPRA in HepG2 cells. Across the study, the researchers identified hundreds of differentially active variants (DAVs) with regulatory activity that differed between alleles; their effects varied by cell type and stimulus condition. Comparisons between HepG2 and LX-2 cells and analyses across stimulus conditions support considering the cellular environment when interpreting variant function.
To connect variants with target genes, the team integrated liver expression quantitative trait loci (eQTLs), chromatin-looping information, and single-cell CRISPR interference (CRISPRi) screens. This approach examined the relationship between genetic variants and expression levels, spatial contacts between genomic regions, and responses to suppressing regulatory regions. DAVs at several loci, including SLC22A3 and the triglyceride-metabolism regulatory genes APOA5, ANGPTL3, and LPL, were linked to the regulation of gene expression, lipid metabolism, and hepatic stellate-cell activity. These findings should not be read to mean that every listed gene regulates each of those three processes.
The study also included risk prediction analyses. In the All of Us cohort, a polygenic risk score (PRS) based on 114 DAVs considered mutually independent according to linkage disequilibrium (LD) was compared with a score based on nine missense variants and with existing scores. The analysis included 37,153 people with MASLD and 40,048 controls, with adjustment for age, sex, sequencing center, and the first 10 genetic principal components. The researchers reported improved risk prediction using DAVs. However, the available public description does not allow the magnitude of that improvement or clinical discrimination to be established numerically.
Implications and outlook: From functional validation to clinical use
This study presents a framework for connecting noncoding variants to gene regulation and MASLD pathophysiology. The finding that the same variant can have different activity depending on cell type and stimulus conditions provides a basis for choosing conditions in future functional experiments.
Cell-based functional validation and cohort risk-score analyses alone do not establish an individual diagnostic or prevention tool. Assessing practical use will require further evaluation of predictive performance, intended populations, and clinical utility. This article’s review was limited to the public abstract and extended-data description; the full experimental design and quantitative results in the paywalled article and supplementary materials could not be checked.
Source: Nature Genetics article abstract and public extended-data description
Nature Genetics, Published online: 02 June 2026. DOI: 10.1038/s41588-026-02617-8
Summary: Resolving the historical linkage disequilibrium confounding and variant-to-gene mapping deficits that bottlenecked metabolic liver disease GWAS data, this investigation combines massively parallel reporter assays (MPRA) and CRISPR screens. By programmatically interrogating high-throughput chromatin accessibility parameters across thousands of non-coding loci, the computing platform isolates functional causal variants altering transcription factor binding affinity metrics. This molecular profiling charts non-linear transcription-factor-to-target-gene expression fluxes regulating hepatic lipid metabolism, delivering a generalizable computational baseline for validated target discovery, precise multivariate risk prediction, and companion diagnostic (CDx) platform engineering.
Researchers can use these findings to prioritize variants and target genes for functional validation. For example, they could select SLC22A3 or candidates involved in triglyceride metabolism, then design studies that alter regulatory-region activity in relevant cells and under relevant stimulus conditions and re-examine gene expression and lipid-accumulation responses. This is a potential application drawn from the findings; it does not mean that effectiveness as a therapeutic target has been established. DAV-based risk scores suggest that functional information could be incorporated into risk prediction. Before clinical use, predictive performance, intended populations, and clinical utility require further evaluation.