Back

Cirrhosis severity modulates proteomic and immune landscapes in hepatocellular carcinoma

Shastry, R. K. R.; Kale, A.; Tee, S. S.

2025-11-28 cancer biology
10.1101/2025.11.24.690286 bioRxiv
Show abstract

Hepatocellular carcinoma (HCC) most often arises in cirrhotic livers, yet the biological impact of cirrhosis severity on the tumor proteome remains poorly understood. Here, we performed a comprehensive quantitative proteomic analysis of 141 HCC tumors stratified by the degree of cirrhosis to delineate fibrosis-related molecular heterogeneity. Principal component analysis revealed substantial overlap between groups, indicating that increasing cirrhosis severity produces incremental rather than global proteomic shifts. Nevertheless, 295 proteins showed significant differential abundance (Hedges g > 0.3, p < 0.05), including upregulation of HMOX1 and LRSAM1 and downregulation of CYP27A1 in Medium/High cirrhosis. Pathway enrichment highlighted alterations in organelle assembly, heme metabolism, and extracellular-matrix disassembly, alongside reduced cytoskeletal and focal-adhesion integrity. Proteome-based immune deconvolution suggested proportional remodeling of natural-killer and B-cell subsets with advancing cirrhosis. A five-gene signature derived from the most upregulated proteins (HMOX1, LRSAM1, MAP2, EDEM3, RFC3) was associated with significantly worse overall survival in the TCGA-LIHC cohort (Hazard ratio 1.7, p = 0.0033). Together, these findings demonstrate that cirrhosis severity subtly but measurably reshapes the tumor proteome and immune landscape in HCC, reflecting a continuum of biological remodeling rather than discrete molecular states. Incorporating cirrhosis gradation into molecular classification may improve risk stratification, biomarker discovery, and the design of precision therapies for HCC.

Matching journals

The top 12 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.