A Quantitative Imaging Toolkit to Investigate the Spatial Regulation of Liver Metabolism in Health and Disease

The liver is the central metabolic hub of the body, responsible for processing nutrients and microbial metabolites arriving from the gut while maintaining systemic energy homeostasis by catabolizing stored nutrients during fasting. A critical component of this function is the storage and mobilization of lipid in the form of lipid droplets, and when this process goes awry, the consequences are severe. Hepatic steatosis, or the accumulation of lipid in the liver,  affects approximately 30% of the Western world and is generally clinically benign, but a subset of patients progress to the more severe MASH, which can advance to cirrhosis, liver failure, and HCC. While much has been established about how lipid accumulation drives lipotoxicity and hepatocyte death, why some individuals progress from hepatic steatosis, or MASLD, to MASH while others do not, remains poorly understood. Compounding this, a distinct subset of HCC arises in the context of steatotic liver without fibrosis or cirrhosis, and these tumors frequently display a lipidated phenotype. The mechanisms governing this fibrosis-independent path to malignancy are largely unknown. Together, these unresolved questions necessitate a deeper understanding of how the liver processes and stores lipid at the cellular level, and what factors are disrupted when these mechanisms fail.

Our laboratory has previously shown that BNIP3, canonically known as a mitochondrial cargo receptor localized to the outer mitochondrial membrane and strongly upregulated in the liver during fasting, plays a role in hepatic lipid homeostasis. Loss of BNIP3 in the liver causes spontaneous lipid accumulation with hallmarks of hepatic steatosis. In work presented here, we demonstrate that DEN-induced liver tumors in bnip3-null mice are more lipid-laden and more aggressive than their wild-type counterparts. We further show that in human HCC, tumors expressing high levels of lipogenic genes have a significantly worse prognosis when BNIP3 expression is low, whereas tumors with high lipogenic gene expression but accompanying high levels of BNIP3 are less steatotic and associated with improved outcomes. Using a primary mouse HCC cell line, we demonstrate that BNIP3 is specifically required for mediating lipid clearance rather than for regulating lipogenesis or fatty acid oxidation. To investigate this phenomenon in a non-tumorigenic context and at higher resolution, we turn to primary mouse hepatocyte culture combined with quantitative high-resolution imaging.

However, the liver is a metabolically heterogeneous organ that is spatially organized across the hepatic lobule such that periportal and pericentral hepatocytes perform fundamentally distinct programs governing processes such as lipid synthesis and fatty acid oxidation, resulting in functional specialization between the zones. Thus far the tools available for studying intracellular differences in cultured primary hepatocytes while accounting for zone of origin remain severely limited. This thesis addresses this limitation through the development and validation of a quantitative imaging pipeline that assigns zone of origin identity to individual primary mouse hepatocytes using endogenous zonal marker proteins retained during the first 24 hours of culture, combined with machine learning-based segmentation of lipid droplets, mitochondria, and lysosomes at single-cell resolution. Deploying this pipeline, we uncover zone-specific differences in lipid droplet size distributions, giant LD burden, and mitochondrial morphology that cannot be resolved by bulk or tissue-level approaches. We further demonstrate that BNIP3-mediated lipid clearance is spatially enriched in Zones 2 and 3, the zones most vulnerable to steatosis, and that this function operates independently of its canonical autophagic activity, as we observe no corresponding increase in lipophagic flux. These findings implicate a thus far uncharacterized mechanism by which BNIP3 governs hepatic lipid homeostasis.