When Cancer Moves to the Liver, It Builds an Entirely Different Support Structure

Pancreatic ductal adenocarcinoma (PDAC) kills most of the patients it touches, and it does so overwhelmingly through metastatic disease, with the liver as its most frequent destination [1]. A new study from the University of California, Irvine, published in Cellular and Molecular Gastroenterology and Hepatology, reports that pancreatic cancer cells growing in the liver rely on a signaling relationship with local support cells that is distinct from the one driving tumors in the pancreas itself, and that disrupting it with a drug already approved for other cancers slows liver tumor growth in preclinical models [1].

Key Takeaways

• Fibroblasts surrounding liver-metastatic PDAC lesions differ molecularly from fibroblasts surrounding primary pancreatic tumors, and produce elevated levels of hepatocyte growth factor (HGF) [1].
• PDAC cells residing in the liver express correspondingly high levels of MET, the receptor for HGF, establishing a paracrine signaling loop specific to that metastatic site [1].
• In immune-competent mouse models, genetic disruption of MET signaling and treatment with cabozantinib, an existing MET-targeting drug, each reduced liver tumor burden more than they affected primary pancreatic tumors [1].

A Tumor Microenvironment That Changes With Its Address

Most knowledge of the PDAC tumor microenvironment comes from studying the primary tumor in the pancreas, where cancer-associated fibroblasts support tumor growth, remodel surrounding tissue, suppress immune activity, and contribute to resistance to chemotherapy [1]. Far less is known about how pancreatic cancer cells interact with stromal cells after they leave the pancreas and colonize a new organ, even though the liver is the most common site of PDAC spread.

The research team, led by Christopher James Halbrook, Ph.D., associate professor at UC Irvine, together with first author Rima Singh, Ph.D., and collaborator Nina G. Steele, Ph.D., Assistant Professor of Internal Medicine at the University of Cincinnati and the Pancreatic Cancer Center at Henry Ford Health, set out to compare the stromal environment of primary pancreatic tumors directly against that of liver metastases arising from the same cancer cells [1]. Using single-cell transcriptomic sequencing, the investigators profiled individual cells from both tumor sites in parallel mouse models, then applied a ligand-receptor interaction analysis to identify which signaling pathways were selectively active in the liver niche [1]. The comparison was extended into human tissue, using laboratory-grown fibroblasts isolated from pancreatic cancer patients at UC Irvine to confirm that the signaling relationship identified in mice was also present in human cells [1].

Identifying the HGF-MET Axis

The analysis showed that fibroblasts in the liver metastatic niche produce markedly higher levels of hepatocyte growth factor than fibroblasts surrounding primary pancreatic tumors, and that pancreatic cancer cells residing in the liver carry correspondingly elevated levels of MET, the receptor through which HGF signals [1]. Functional assays confirmed this was not merely a correlation: liver-derived fibroblasts directly activated mitogenic MET signaling in PDAC cells, a growth-promoting circuit that was substantially weaker in the primary tumor setting [1].

The HGF-MET relationship identified in this study fits within a broader, previously described pattern in cancer biology, in which HGF-MET signaling in the liver microenvironment has been associated with tumor cell metabolic reprogramming and reduced sensitivity to chemotherapy. This study adds a direct, side-by-side comparison showing that this axis is disproportionately active at the liver metastatic site, rather than a general feature of PDAC biology present equally at both locations [1].

Genetic and Pharmacologic MET Inhibition

To test whether the HGF-MET axis was not only present but also functionally important to tumor growth, the researchers pursued two complementary approaches in immune-competent mouse models: genetic disruption of MET in cancer cells and pharmacologic inhibition with cabozantinib, a small-molecule MET inhibitor already in clinical use for other malignancies [1]. Both interventions reduced pancreatic tumor growth, and the effect was more pronounced in the liver than in the pancreas, with most treated animals showing either no visible liver tumors or only small residual tumor masses [1].

Establishing this metastatic-site-specific model required substantial technical development. Singh, whose doctoral thesis centered on the project, described mastering the liver surgical technique used to generate the metastatic model as the central methodological challenge, refined over repeated attempts with guidance from Oliver Gene McDonald, MD, Ph.D., associate professor and a collaborator at the University of Miami [1].

Why Site-Specific Biology Matters, and Where Single-Agent MET Inhibition Has Struggled

The finding is consistent with a broader challenge that has faced MET-targeted therapy in gastrointestinal cancers to date. Single-agent MET kinase inhibitors and anti-MET monoclonal antibodies have generally produced limited clinical responses in digestive tract malignancies, partly due to adaptive resistance mechanisms, including MET-independent autophagy pathways and compensatory HGF secretion by cancer-associated fibroblasts that can bypass MET blockade [2].

Preclinical work in other contexts has suggested that combining a MET kinase inhibitor with an HGF-neutralizing antibody, an inhibitor of HGF activation, or an autophagy-blocking agent may more durably shut down the pathway than MET inhibition alone [2]. The UC Irvine findings do not directly resolve this resistance question, since the mouse studies reported here used MET disruption on its own rather than in combination, but they reinforce the rationale for pursuing combination strategies specifically in the liver metastatic setting rather than assuming a single approach will work equally across all disease sites.

What the Investigators Say It Means

Steele described the finding’s direct translational relevance as what distinguishes it from purely descriptive tumor microenvironment studies, noting that an existing, clinically available drug meaningfully suppressed liver tumor growth by targeting the fibroblast-cancer cell signaling axis identified through the transcriptomic analysis [1].

Halbrook framed the broader implication as a case for studying metastatic disease at its actual site of growth rather than inferring its biology from the primary tumor, noting that the vulnerabilities identified in this study would not have been apparent from primary pancreatic tumor tissue alone [1]. He further suggested that the underlying principle, that different organs may impose different, exploitable signaling requirements on the same cancer cells, could extend to other malignancies that commonly metastasize to the liver, beyond pancreatic cancer specifically [1].

Study Limitations and Next Steps

The findings are preclinical, generated in mouse models and validated in laboratory-cultured human fibroblasts, and have not yet been tested in patients [1]. The manuscript itself, framed by the authors’ own stated future directions, notes that additional work will be needed to determine which features of the liver-specific signaling network are driven by the tissue of origin of the cancer versus the identity of the organ it has colonized, a distinction relevant to whether this approach would generalize to metastases from other primary tumor types [1]. The research team has stated it is now planning clinical trials at the Chao Family Comprehensive Cancer Center to begin testing this organ-specific therapeutic strategy in patients, though no trial has yet been registered [1].

Editorial Transparency Note

This article is based on the peer-reviewed manuscript by Singh et al., published online ahead of print in Cellular and Molecular Gastroenterology and Hepatology on September 4, 2026, and on a UC Irvine institutional press release issued the same week [1]. Background context on HGF-MET signaling and resistance to MET-targeted therapy in gastrointestinal cancers is drawn from prior published reviews of the pathway [2].

Pancreatic cancer incidence and survival figures are sourced from the National Cancer Institute’s SEER Cancer Stat Facts program. This article is intended for informational purposes for healthcare professionals and does not constitute medical advice. This article discusses cabozantinib (Cabometyx®; Exelixis) in the context of an investigational preclinical application; it is not an FDA-approved indication for pancreatic cancer liver metastasis.

References

[1] Singh R, Yousefian N, Allen WM, Silva AD, Anaraki C, Michels AK, Calderon S, Loveless IM, Rastelli D, George A, Aldakkak M, McDonald OG, Lytle NK, Morton JP, Ayabe RI, Imagawa D, Jutric Z, Martinez TF, Steele NG, Halbrook CJ. Comparative Analysis of Primary and Liver Fibroblasts Reveals MET as a Potent Target in Pancreatic Cancer Metastasis. Cell Mol Gastroenterol Hepatol. 2026 Sep 4:101872. doi: 10.1016/j.jcmgh.2026.101872. Epub ahead of print. PMID: 42697413.
[2] Papaccio F, Della Corte CM, Viscardi G, Di Liello R, Esposito G, Sparano F, Ciardiello F, Morgillo F. HGF/MET and the Immune System: Relevance for Cancer Immunotherapy. Int J Mol Sci. 2018 Nov 14;19(11):3595. doi: 10.3390/ijms19113595. PMID: 30441809; PMCID: PMC6274701.
[3] National Cancer Institute. Cancer Stat Facts: Pancreatic Cancer. Surveillance, Epidemiology, and End Results (SEER) Program. Online. Last accessed on September 7, 2026.
[4] University of California, Irvine, Charlie Dunlop School of Biological Sciences. UCI-Led Study Identifies Promising Target for Pancreatic Cancer That Spreads to The Liver. Press release, September 8, 2026.

This article is intended for informational purposes for healthcare professionals and does not constitute medical advice.

Featured image:© 2026 CHA/JCO ONC. Used with permission.

DOI