Metabolic-associated steatotic liver disease (MASLD) stands as one of the most pervasive liver disorders worldwide, impacting approximately a quarter of the global population. This condition, marked by abnormal lipid accumulation within hepatocytes, follows a pathological trajectory from simple steatosis to inflammation, fibrosis, and in severe cases, cirrhosis and hepatic failure. Despite its prevalence, the molecular and cellular mechanisms underpinning MASLD’s early stages remain incompletely understood. Groundbreaking research published in Nature now unravels the spatial and molecular dynamics of lipid buildup in human liver tissue, shedding unprecedented light on hepatocyte adaptation during steatosis.
The research team pioneered the use of spatial transcriptomics combined with machine learning to quantify and map lipid content directly within human liver biopsies. Their cohort, comprising four patients exhibiting a spectrum of steatosis severity, revealed a striking zonal pattern: lipid accumulation originated predominantly in the pericentral zones of the hepatic lobule. This zonation aligns with the liver’s known metabolic heterogeneity, but importantly, the study delineates how gene expression within pericentral hepatocytes dynamically changes as lipid content increases, a hallmark of MASLD progression.
By leveraging advanced algorithms, the investigators segmented and quantified lipid droplets from high-resolution histological images stained with hematoxylin and eosin. They categorized liver tissue microenvironments into lipid content bins, allowing them to average gene expression profiles across these bins. This approach uncovered a coordinated transcriptional response: hepatocytes reduced the expression of genes promoting lipid influx and synthesis, including SLC27A3, DGAT1, and NPC1L1. Simultaneously, there was a compensatory increase in the expression of genes facilitating lipid breakdown and export, such as those involved in mitochondrial beta-oxidation (DIO1, HSD17B4) and lipid secretion (ABCB4).
One of the most compelling findings lies in the upregulation of transcriptional regulators KLF10 and ONECUT2, which may orchestrate the hepatocyte adaptation to lipid overload. Intriguingly, the study also reported a downregulation of nuclear-encoded mitochondrial respiratory genes (NDUFA9, NDUFS4, NDUFS8), indicating potential mitochondrial dysfunction in early MASLD stages. Conversely, mitochondrial genome-encoded transcripts such as MT-ND1 and MT-ND4L were elevated, suggesting a complex compensatory mitochondrial response.
Further scrutiny into genes linked with lipid droplet stability revealed that HSD17B13, an enzyme stabilizing lipid droplets, exhibited decreased expression with increasing lipid load. Conversely, ELOVL5, a gene previously associated with reduced triglyceride accumulation, was upregulated. These molecular signatures collectively highlight hepatocytes’ adaptive attempts to mitigate excessive lipid storage and forestall lipotoxicity in early disease phases.
To validate these transcriptomic insights at single-droplet resolution, the team employed Visium HD technology complemented by artificial intelligence-driven segmentation of individual lipid droplets. This innovative approach utilized 2-micrometer spots to capture RNA signals within isolated droplets. Differential gene expression analysis between hepatocytes containing large versus small lipid droplets corroborated the broader spatial transcriptomics findings, underscoring the robustness of the transcriptional reprogramming observed in steatotic hepatocytes.
Comparative analyses with murine models of steatosis induced by choline-deficient, amino acid-defined high-fat diets revealed only weak correlations in gene expression changes with early human steatosis. Notably, a stronger correlation was observed with a patient exhibiting severe steatosis, illuminating species-specific divergences in the pathophysiological response to lipid overload. This highlights the necessity of human-centric studies to accurately capture the complexity of MASLD.
The data generated form a comprehensive spatial atlas that offers unprecedented detail of the hepatic transcriptional landscape associated with lipid accumulation. By parsing these complex datasets, the study provides a vital resource to understand how hepatocytes modulate lipid metabolism, mitochondrial function, and transcriptional regulation during early MASLD. Such insights have significant implications for the development of targeted therapeutics aimed at halting or reversing steatotic progression before irreversible damage ensues.
Notably, the study elucidates early compensatory mechanisms in the human liver that might be masked or absent in advanced disease stages or animal models. Understanding these adaptive responses could lead to the identification of novel biomarkers predictive of disease trajectory and therapeutic response. The research underscores the importance of spatial context in interpreting transcriptomic data, as conventional bulk RNA sequencing might obscure zonal nuances critical for disease pathology.
The integration of histological imaging, spatial transcriptomics, machine learning, and high-resolution spatial RNA capture delivers a powerful methodological framework. This multidisciplinary approach facilitates a granular dissection of liver microenvironments, linking lipid droplet morphology to precise gene expression signatures. The technological innovations presented could be extended to other organs and diseases where spatial heterogeneity plays a crucial role.
Importantly, the research team has provided an accessible online web application to explore this expansive dataset, allowing investigators worldwide to query gene expression patterns linked with human hepatic steatosis. This open-access tool democratizes data exploration, fostering collaborative discovery and accelerating translational research efforts focused on metabolic liver disease.
In conclusion, this work profoundly advances our understanding of MASLD pathogenesis, offering a spatially resolved molecular blueprint of hepatocyte adaptations to lipid accumulation. By mapping dynamic gene expression changes aligned with lipid burden, it reveals potential cellular strategies aimed at minimizing lipotoxicity. This foundational knowledge is pivotal for guiding future interventions and underscores the power of spatially resolved multi-omics in deciphering complex human diseases.
Subject of Research: Metabolic-associated steatotic liver disease (MASLD) and spatial transcriptomic analysis of lipid accumulation in human hepatocytes.
Article Title: A spatial atlas of the healthy human liver from live donors.
Article References:
Yakubovsky, O., Bahar Halpern, K., Shir, S. et al. A spatial atlas of the healthy human liver from live donors. Nature (2026). https://doi.org/10.1038/s41586-026-10377-y
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41586-026-10377-y
Keywords: MASLD, steatosis, hepatocytes, spatial transcriptomics, lipid droplets, machine learning, mitochondrial dysfunction, gene expression, liver zonation, human liver atlas
