Profiling 43 normal human livers using single-cell multiome, spatial transcriptomics, and CODEX revealed cell-type-specific senescence programs associated with age, fibrosis, and cancer, with chemotherapy intensifying hepatocyte senescence 5-fold relative to aging and senescent cells sharing AP-1 activation, pro-inflammatory cytokines, and apoptosis resistance across conditions.
Key Findings
Results
CDKN1A+ senescent cells were identified across multiple liver cell types including hepatocytes, fibroblasts, cholangiocytes, and endothelial cells.
Senescent populations were characterized by CDKN1A expression across distinct cell types in the liver
These populations were associated with age, liver disease, or cancer depending on cell type
The study profiled 43 normal human livers spanning multiple ages and fibrosis stages
Complementary platforms included single-cell multiome, Xenium spatial transcriptomics, and CODEX
Results
Senescence programs differed between aged and fibrotic livers, with similar patterns recapitulated in fibrotic mouse models.
Aged and fibrotic livers showed distinct senescence signatures despite both conditions exhibiting CDKN1A+ populations
Fibrotic mouse models were used to validate and complement findings from human liver profiling
The distinction between aging-associated and fibrosis-associated senescence suggests condition-specific mechanisms
Results
CDKN1A+ senescent hepatocytes localized periportally in spatial analyses.
Xenium spatial transcriptomics was used to determine the spatial organization of senescent hepatocytes
Periportal localization of CDKN1A+ hepatocytes suggests a zonation-dependent senescence pattern
Spatial organization was assessed across 43 normal human livers
Results
SERPINE1+ aging-associated hepatocytes formed spatial clusters potentially mediated by Claudins and THBS1.
SERPINE1 (also known as PAI-1) marked a distinct aging-associated hepatocyte population
These cells formed spatial clusters detectable by Xenium spatial transcriptomics
Claudins and THBS1 were identified as potential mediators of this spatial clustering
Results
Fibrotic regions contained CXCL12+ senescent fibroblasts that interacted with CXCR4+ immune cells.
CXCL12 expression marked senescent fibroblasts specifically in fibrotic regions of the liver
CXCR4+ immune cells were identified as interaction partners of CXCL12+ fibroblasts
This ligand-receptor interaction suggests a mechanism by which senescent fibroblasts recruit or communicate with immune cells in fibrotic liver
Results
Chemotherapy intensified hepatocyte senescence by 5-fold relative to aging and induced unique CDKN2A+ populations.
24 colorectal cancer liver metastases were analyzed alongside normal liver samples
Chemotherapy-associated senescence was 5-fold greater in hepatocytes compared to aging-associated senescence
Chemotherapy uniquely induced CDKN2A+ senescent populations not prominently observed under aging or fibrosis conditions
CDKN2A (encoding p16INK4a) expression distinguished chemotherapy-induced from other forms of senescence
Results
Across all conditions, senescent cells shared activation of AP-1, pro-inflammatory cytokines, and apoptosis resistance.
AP-1 transcription factor activation was a common feature of senescent cells regardless of the inducing condition
Pro-inflammatory cytokine expression was shared across senescent cell populations, consistent with a senescence-associated secretory phenotype (SASP)
Apoptosis resistance was identified as a shared characteristic across senescent cell populations
These shared features were identified across aging, fibrosis, and cancer-associated senescence contexts
The authors suggest these shared features represent 'therapeutic opportunities'
Methods
The study employed a multi-platform approach profiling 43 normal human livers across ages and fibrosis stages.
Single-cell multiome (simultaneous RNA and ATAC sequencing) was used for transcriptomic and chromatin accessibility profiling
Xenium spatial transcriptomics provided in situ spatial resolution of gene expression
CODEX (co-detection by indexing) provided spatial protein-level data
24 colorectal cancer liver metastases were included to assess cancer-associated senescence
Fibrotic mouse models provided additional validation of human findings
Karpova A, Li X, Peng C, Gallant K, Rapp D, Alligood D, et al.. (2026). Cellular senescence in human liver under normal aging and cancer.. Cell genomics. https://doi.org/10.1016/j.xgen.2025.101133