The investigative team began with a clinically grounded question: which genetic alterations predict failure of immunotherapy in gastric cancer? To answer it, they performed whole-exome sequencing on tumor specimens from patients who had received neoadjuvant chemotherapy combined with PD-1 blockade, the current backbone of perioperative treatment for many locally advanced cases. Among the recurrently mutated genes associated with poor treatment outcome, KRAS emerged as a standout. The signal was particularly striking because it persisted even in tumors that carried the two established biomarkers used to select patients for immune checkpoint inhibitors: mismatch repair deficiency, known as dMMR status, and a combined positive score, or CPS, above ten. Patients whose tumors harbored both favorable biomarkers and a KRAS mutation nonetheless derived very limited benefit, a finding that challenges the assumption that biomarker-positive gastric cancers respond uniformly to immunotherapy.

Multiplex immunohistochemistry on patient samples added a spatial dimension to the genetic data. KRAS-mutant gastric tumors proved to be poorly infiltrated by lymphocytes, and the deficit was most pronounced among intratumoral CD8-positive T cells, the cytotoxic warriors that checkpoint inhibitors are designed to unleash. In their place, the tumors showed increased infiltration of CD68-positive myeloid cells, a population frequently associated with pro-tumor inflammation and immune suppression. The picture that emerged was of a tumor microenvironment stripped of its anti-cancer firepower and stocked instead with cells that dampen immune attack, a configuration that would predict resistance to PD-1-directed therapy regardless of biomarker status.

To move from correlation to causation, the researchers needed a model that faithfully reproduced both the genetics and the immunology of human disease. They generated a stomach-specific, genetically faithful and immunocompetent mouse model of KRASG12D-driven gastric adenocarcinoma, designated the CPPK model, built on Anxa10-CreERT2, KrasG12D and Trp53-floxed alleles. Because the model develops in animals with intact immune systems, it captures the full interplay between oncogenic signaling and the immune microenvironment, something conventional cell-line xenografts cannot do. The model recapitulated the stepwise development of human poorly differentiated gastric adenocarcinoma with remarkable fidelity, progressing from early antralization of gastric glands to advanced tubular and intestinal-type adenocarcinoma, and ultimately to liver and peritoneal metastasis. Mirroring the clinical observations, CPPK mice were intrinsically resistant to PD-1 blockade, providing a platform in which the biology of resistance could be dissected and therapeutic combinations tested.

Single-cell RNA sequencing of tumor tissues from the model exposed the architecture of immune suppression in unprecedented detail. Regulatory T cells, the immunosuppressive subset that normally keeps immune responses in check, comprised 41.7 percent of all tumor-infiltrating T cells, an extraordinary proportion that dwarfs what is seen in immunologically hot tumors. Within the NKT cell compartment, 55.3 percent of cells displayed an exhausted phenotype, functionally crippled by chronic antigen stimulation. Perhaps most tellingly, TGF-beta signaling was markedly activated in the tumor Tregs, and spatial analysis revealed that this pathway co-localized with the CTLA-4 molecule within the same cells. Together, these data identified TGF-beta-positive regulatory T cells as the central hub maintaining immune suppression, and they pointed to CTLA-4 as the molecular handle by which those cells might be targeted.

The therapeutic arm of the study tested MRTX1133, a recently developed small-molecule inhibitor of KRASG12D, in the mouse model. The drug effectively suppressed tumor growth and prolonged survival, and it inhibited KRASG12D gastric cancer cell lines and organoids with half-maximal inhibitory concentrations ranging from 271 to 485 nanomolar. On its face, this seemed like a victory for precision oncology. Yet functional immune analyses told a more sobering story: effective antitumor immunity was not restored by KRAS inhibition alone. The immune cells expanding after treatment were predominantly immunosuppressive Tregs rather than cytotoxic lymphocytes, and the function of cytotoxic T cells remained essentially unchanged. Mechanistically, TGF-beta signaling proved necessary to maintain Treg dominance and suppress effector T cell function, meaning that removing the oncogenic driver did not, by itself, dismantle the immunosuppressive apparatus the tumor had built.

This insight set up the study’s central therapeutic experiment. Because CTLA-4 blockade is known to act directly on regulatory T cells, depleting or disabling them within tumors, the researchers combined KRAS inhibition with an anti-CTLA-4 antibody. The combination markedly attenuated TGF-beta activity, impaired Treg function and synergistically enhanced antitumor immune responses. In other words, the targeted therapy shrank the tumor and the immunotherapy removed the suppressive cells that had been shielding it, addressing two halves of the same problem simultaneously. The doublet transformed an immunologically cold, Treg-dominated tumor into one in which effector immunity could operate.

The most dramatic results came when a third agent was layered on. Adding PD-1 blockade to the KRAS-CTLA-4 doublet led to complete eradication of three subcutaneous tumors in the preclinical models, and the combination significantly prolonged survival in the spontaneous CPPK model, where tumors arise in their native anatomical location and metastasize naturally. These findings identify oncogenic KRAS as a key regulator of TGF-beta-dependent immune suppression in gastric adenocarcinoma and provide a strong preclinical rationale for CTLA-4-based combination immunotherapy in this molecularly defined subset of patients. For a cancer subtype in which even biomarker-positive cases fail PD-1 therapy, the prospect of a rational triple combination is significant.

Beyond its immediate therapeutic implications, the study carries broader lessons for the field of cancer immunology. It demonstrates that oncogene addiction and immune evasion are not parallel phenomena but deeply intertwined ones, with mutant KRAS actively recruiting and maintaining a regulatory T cell niche through TGF-beta. It also illustrates why single-agent approaches so often disappoint: dismantling the driver does not automatically dismantle the microenvironment the driver created. The work underscores the value of genetically faithful, immunocompetent models for dissecting such interactions, and it suggests that biomarker-guided immunotherapy strategies should account for oncogenic drivers like KRAS even when canonical predictive markers appear favorable.

Clinical translation will require careful evaluation, as the findings remain preclinical and the safety of combining KRAS inhibitors with dual checkpoint blockade in patients with gastric cancer has yet to be established. Nevertheless, the study, carried out by researchers from Peking University Cancer Hospital & Institute and collaborating institutions and supported by grants from the National Science and Technology Major Project of China, the Beijing Natural Science Foundation and the National Natural Science Foundation of China, charts a clear path forward. For the roughly one in ten patients with gastric adenocarcinoma whose tumors harbor KRAS mutations, a group currently left behind by the immunotherapy revolution, the combination of KRAS inhibition with CTLA-4 blockade offers something they have not had before: a biologically grounded reason for hope.

Subject of Research: KRAS-mutant gastric adenocarcinoma and combination immunotherapy targeting the TGF-beta-dependent regulatory T cell microenvironment

Article Title: KRAS inhibition combined with CTLA-4 blockade overcomes immunotherapy resistance in KRAS-mutant gastric cancer

Article References: KRAS inhibition combined with CTLA-4 blockade overcomes immunotherapy resistance in KRAS-mutant gastric cancer. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: KRAS, gastric adenocarcinoma, CTLA-4, PD-1, TGF-beta, regulatory T cells, immunotherapy resistance, MRTX1133, tumor microenvironment, single-cell RNA sequencing, immune checkpoint inhibitors, Peking University Cancer Hospital