Hypothesis

Senescent CD8+ T cells (CD28‑CD57+) accumulate in the liver during prediabetes and secrete the chemokine CXCL9, which directly suppresses Wnt/β‑catenin signaling in zone 3 hepatocytes. This suppression blunts the pericentral transcriptional program (e.g., glutamine synthetase, Cyp2e1) and promotes a bi‑zonal hepatocyte phenotype, precipitating mitochondrial dysfunction, lipid accumulation, and steatosis before fibrosis sets in.

Rationale

• Senescent CD8+ T cells are sufficient to induce systemic inflammaging and tissue senescence【1](https://doi.org/10.1126/science.aax0860)【2](https://doi.org/10.1038/s41419-019-1494-4).
• CXCL9 is a known senescence‑associated secretory phenotype (SASP) factor upregulated in exhausted CD8+ T cells and can act on CXCR3‑expressing non‑immune cells.
• Zone 3 hepatocytes rely on Wnt/β‑catenin signaling for their metabolic identity; loss of this pathway correlates with loss of zonation and steatosis【4](https://pmc.ncbi.nlm.nih.gov/articles/PMC8319533/)【5](https://doi.org/10.1038/ncomms15691).
• Aging disrupts hepatic zonation, creating aberrant bi‑zonal hepatocytes that precede fibrosis【6](https://pubmed.ncbi.nlm.nih.gov/40622856/).

Mechanistic Insight

We propose that CXCL9, via CXCR3 on hepatocytes, activates a non‑canonical pathway that leads to GSK‑3β activation and subsequent β‑catenin degradation, independent of inflammatory cytokine milieus. This connects immune senescence directly to the metabolic reprogramming that underlies the Zone 3 vulnerability.

Testable Predictions

1. Correlation – CXCL9 levels in liver sinusoidal blood will positively correlate with the frequency of senescent CD8+ T cells and inversely with nuclear β‑catenin in zone 3 hepatocytes across human NAFLD spectra (from simple steatosis to NASH).
2. Causality – Liver‑specific overexpression of CXCL9 in mice will recapitulate zonation loss (loss of glutamine synthetase pericentral staining) and steatosis without altering overall inflammation.
3. Rescue – Genetic deletion of Cxcr3 in hepatocytes or pharmacological blockade of CXCR3 will preserve zonation signaling and prevent lipid accumulation despite the presence of senescent CD8+ T cells.
4. Senolytic Specificity – Selective clearance of senescent CD8+ T cells (using a CD8‑targeted senolytic) will reduce hepatic CXCL9, restore β‑catenin activity, and improve zonation before fibrosis develops, whereas non‑selective senolytics will show a smaller effect.

Experimental Design

• Human validation: Obtain liver biopsies and matched serum from patients stratified by NAFLD activity score. Perform immunohistochemistry for CD28‑CD57+ CD8+ T cells, CXCL9, and glutamine synthetase; quantify serum CXCL9 by ELISA. Perform regression analysis.
• Mouse models: a. Gain‑of‑function: Alb‑CreERT2;Rosa26‑LSL‑CXCL9 mice induced with tamoxifen; assess zonation (GS, Cyp2e1) and lipid accumulation (Oil Red O) at 4, 8, 12 weeks. b. Loss‑of‑function: Cxcr3^fl/fl; Alb‑CreERT2 mice crossed with a pro‑senescent CD8+ T‑cell model (e.g., p16^3MR‑driven senescence) to test rescue. c. Selective senolytic: Use a CD8‑targeted CAR‑NK cell or a senolytic pro‑drug activated by CD8‑associated protease (e.g., granzyme B) to deplete senescent CD8+ T cells; compare to dasatinib+quercetin.
• Readouts: Zonation markers (immunofluorescence, qPCR), mitochondrial respiration (Seahorse), hepatic triglyceride content, insulin tolerance test, fibrosis scoring (Sirius Red).

Falsifiability

If CXCL9 neutralization or hepatocyte CXCR3 deletion fails to preserve zonation or prevent steatosis in the presence of senescent CD8+ T cells, the hypothesis is refuted. Likewise, if senescent CD8+ T‑cell depletion does not lower hepatic CXCL9 or improve zonation, the proposed causal link is invalid.

Implications

Confirming this mechanism would reposition early immunomodulation—specifically targeting the CXCL9/CXCR3 axis—as a preventive strategy for NAFLD, distinct from broad senolytic approaches that act after architectural damage has occurred.