Single Nucleus RNA Sequencing of Pre-Malignant Liver Reveals Disease-Associated Hepatocyte State with HCC Prognostic Potential

Current approaches to stage chronic liver diseases have limited utility to directly predict liver cancer risk. Here, we employed single nucleus RNA sequencing (snRNA-seq) to characterize the cellular microenvironment of healthy and chronically injured pre-malignant livers using two distinct mouse models. Analysis of 40,748 hepatic nuclei unraveled a previously uncharacterized disease-associated hepatocyte transcriptional state (daHep). These cells were absent in healthy livers, but were increasingly prevalent as chronic liver disease progressed towards hepatocarcinogenesis. Gene expression deconvolution of 1,439 human liver transcriptomes from publicly available datasets revealed that daHep frequencies highly correlate with current histopathological liver disease staging systems. Importantly, we show that high daHep levels precede carcinogenesis in mice and humans and predict a higher risk of hepatocellular carcinoma (HCC) development. This novel transcriptional signature with diagnostic and, more importantly, prognostic significance has the potential to change the way chronic liver disease patients are staged, surveilled and risk-stratified. Overall design: To identify and characterize cell states associated with the chronically injured pre-malignant liver, we employed a droplet-based (10x chromium) single nucleus transcriptomics approach. Hepatic nuclei were isolated and profiled from (a) healthy mice fed normal chow, (b) mice subjected to a choline-deficient, ethionine-supplemented (CDE) diet, and (c) mice provided with thioacetamide (TAA) in the drinking water. We obtained a total of 40,748 single nucleus transcriptomes (16,222 healthy; 14,507 CDE; and 10,019 TAA) from three mice per condition.