The SERPENTINE investigators reported a stark clinical divergence by metastatic site in patients with MSS mCRC treated with dual PD-L1 and CTLA-4 blockade. Patients without liver involvement derived clinical benefit, the authors noted. Every patient with liver metastases progressed on therapy, according to the study team. The trial is a phase II study (Clinical Trial NCT06680739), and the article frames liver involvement as a dominant determinant of resistance to immune checkpoint inhibitors.
The study team integrated paired pre-treatment and on-treatment single-cell transcriptomic and TCR profiling with deep peripheral blood TCR sequencing. They analyzed 388,196 single cells and 95,617 intratumoral clonotypes, plus peripheral repertoires comprising 57,902,418 clonotypes. These data indicated dual ICI mobilized systemic T-cell priming irrespective of metastatic site; therapy-expanded clones appeared in circulation and trafficked to both lung and liver lesions. However, investigators reported the eventual anti-tumor outcome was determined locally within each metastatic microenvironment, not by the absence of systemic activation.
In lung metastases, the study team observed that therapy-expanded T-cell clones concentrated into tissue-resident, proliferative, tumor-reactive CD8+ T-cell states. Those intratumoral states exhibited features the authors associated with productive effector programs. Short bursts of proliferation and localized enrichment of tumor-reactive CD8+ compartments were reported in the lung lesions, suggesting focused intratumoral engagement after systemic priming.
By contrast, liver metastases showed a different cellular trajectory. The same therapy-driven clones that reached the liver failed to consolidate into the tissue-resident, proliferative CD8+ states seen in lung lesions. Instead, investigators reported a dispersed activation pattern characterized by other immune shifts described below.
The cohort showed expansion of regulatory T cells and suppressive myeloid populations specifically within liver lesions. The authors reported diversification of intratumoral TCR repertoires in these hepatic tumors and fragmentation of exhaustion-associated T-cell gene networks. In plain terms, therapy-induced activation in the liver did not coalesce into a few dominant, tumor-reactive clonotypes; it scattered across more diverse repertoires while immunoregulatory elements increased.
Cross-lineage co-expression analyses performed by the study team indicated coordinated multicellular changes: suppressive myeloid and regulatory T-cell programs rose in tandem with the dispersed TCR signal. The investigators interpreted these findings as evidence that the liver microenvironment misdirects otherwise therapy-expanded T cells toward non-productive or suppressed states.
The analysis revealed an unexpected coupling: tumor proliferation increased alongside cytotoxic CD8+ T-cell programs. The study authors reported this co-occurrence after dual checkpoint blockade, meaning that intratumoral cytotoxic signatures were present without corresponding tumor control. Investigators framed this as effector engagement that nevertheless fails to restrain tumor growth in liver metastases. The data therefore separate systemic priming from effective local tumor suppression and place emphasis on intratumoral coordination as the point of therapeutic failure.
The authors reported that raw and processed single-cell sequencing data will be made publicly available upon final publication. The source documents trial registration (NCT06680739) and provides extensive cohort-level single-cell and TCR sequencing counts in the methods and abstract. The study clarifies that resistance in liver metastases stems from tissue-specific misdirection of immune activation rather than a global failure to prime T cells.
Unanswered questions remain about whether local microenvironmental interventions can redirect systemic, therapy-expanded T cells toward productive antitumor states in hepatic lesions. The source did not report interventional rescue experiments or specific therapeutic strategies to overcome liver-associated resistance. Further mechanistic and interventional work will be required to determine whether modifying the hepatic microenvironment—or selectively targeting suppressive myeloid and regulatory compartments—can convert dispersed activation into focused, tumor-restraining immunity.