The problem with cancer immunotherapy is that your T cells can show up to the tumor and still behave like interns on their first day - badge clipped on, eyes wide, not entirely sure which button starts the machinery.
That is basically the mess this new npj Precision Oncology paper tries to sort out. The authors looked at head and neck tumors treated with neoadjuvant immunotherapy, meaning treatment given before surgery, and asked a very practical question: can we tell ahead of time which patients already have T cells capable of turning into effective tumor-fighting cells? Their answer points to a molecule called GZMK, short for granzyme K, as a sort of biological "recently promoted" sticker on activated T cells.[1]
Your Immune System's Security Team, but Make It Complicated
Cytotoxic T cells are the immune system's professional bouncers. They inspect suspicious cells, recognize bad actors, and if needed, remove them with extreme prejudice. They do a lot of this using granzymes, which are proteins stored like tiny biochemical switchblades.[1]
But tumors are not polite opponents. They warp the neighborhood, exhaust immune cells, throw up suppressive signals, and generally turn the tumor microenvironment into the cellular equivalent of a workplace where every meeting should have been an email. In head and neck cancer, neoadjuvant checkpoint immunotherapy can wake up anti-tumor T cells, but not everybody benefits, and recurrence is still common.[2,3]
That gap matters. If you could tell which tumors already contain T cells that are ready to mature into effective fighters, you might predict response better, spare some people a detour through ineffective therapy, and design smarter combinations for the ones whose immune cells are stuck in neutral.[1,2]
What the Researchers Actually Did
The team used single-cell RNA sequencing and T cell receptor tracking on tumor samples taken before and after immunotherapy. Translation: they did the molecular equivalent of putting name tags on thousands of individual T cells and then following who changed, who expanded, and who mostly loitered near the snacks.[1]
They found that GZMK expression inside activated intratumoral T cell subsets marked cells that had recently differentiated - in other words, cells that were not just present, but actively moving along the path toward a more useful anti-tumor state. Patients whose tumors already showed efficient baseline differentiation of these GZMK-positive progenitor-like T cells were more likely to have stronger treatment-induced tumor regression.[1]
That is the punchline. Not all tumors fail because they lack T cells. Some fail because the T cells are there but cannot complete the journey from "aware something is wrong" to "actively ruining the tumor's week."
Why GZMK Is Sneakily Interesting
GZMK is not as famous as granzyme B, the bigger celebrity in T-cell killing. Granzyme K has been a bit of an oddball in the literature, sometimes linked with cytotoxicity, sometimes inflammation, sometimes weird in-between states that make immunologists stare at UMAP plots like they are trying to read tea leaves.[1,4]
Here, GZMK seems less like a generic kill switch and more like a timestamp. It reflects a T cell in motion, recently differentiated, not fully stalled out. That fits with other recent studies showing that response to neoadjuvant immunotherapy depends on the presence of tumor-resident or early-responsive T cell populations that can be reactivated rather than manufactured from scratch on command.[5-7]
Which, frankly, makes sense. Anyone who has spent time around actual experiments knows biology loves momentum. Cells that are already halfway up the hill are easier to push than cells face-down in the mud.
Why This Could Matter Outside the Figure Legend
If these findings hold up in larger cohorts, GZMK-related T-cell profiling could become a biomarker for who is likely to benefit from pre-surgical immunotherapy.[1] That would be useful on its own. More interestingly, it suggests a therapeutic strategy: do not just "add immunotherapy" louder. Figure out what blocks T-cell differentiation inside the tumor and remove that roadblock.
That idea lines up with the broader direction of the field. Recent work in head and neck cancer has tied successful neoadjuvant immunotherapy to specific T-cell states, tissue-resident programs, and dynamic remodeling of the tumor immune environment.[2,5-8] Clinically, this is not abstract bench gossip anymore. Perioperative pembrolizumab moved forward enough that head and neck cancer immunotherapy is now edging into routine earlier-stage treatment decisions, not just salvage-mode last stands.[8]
Of course, this paper is not magic. It does not prove GZMK causes response. It identifies a marker of immune competence, and biomarkers have a long history of looking amazing right up until a larger validation cohort arrives carrying a baseball bat. Also, single-cell studies are powerful, but they are built on expensive, finicky workflows that tend to conceal several months of optimization, one existential crisis, and at least two libraries that absolutely refused to behave.
Still, the core idea is strong: what matters may not be whether a tumor has T cells, but whether those T cells can still grow up on schedule. In cancer biology, that counts as a very good plot twist.
References
-
Sievers CM, Chang TG, Robbins Y, et al. GZMK expression within activated intratumoral T-cell subsets reflects differentiation efficiency and predicts response to cancer immunotherapy. npj Precision Oncology. 2026. DOI: 10.1038/s41698-026-01437-7
-
Zhao M, Schoenfeld JD, Egloff AM, et al. T cell dynamics with neoadjuvant immunotherapy in head and neck cancer. Nature Reviews Clinical Oncology. 2025;22:83-94. DOI: 10.1038/s41571-024-00969-w
-
Goetz JW, Rabinowits G, Kalman N, Villa A. A Review of Immunotherapy for Head and Neck Cancer. Journal of Dental Research. 2024;103(12):1185-1196. DOI: 10.1177/00220345241271992. PMCID: PMC11653306
-
Bouwman AC, van Daalen KR, Crnko S, ten Broeke T, Bovenschen N. Intracellular and Extracellular Roles of Granzyme K. Frontiers in Immunology. 2021;12:677707. DOI: 10.3389/fimmu.2021.677707. PMCID: PMC8129556
-
Oliveira G, Egloff AM, Afeyan AB, et al. Preexisting tumor-resident T cells with cytotoxic potential associate with response to neoadjuvant anti-PD-1 in head and neck cancer. Science Immunology. 2023;8(87):eadf4968. DOI: 10.1126/sciimmunol.adf4968
-
Luoma AM, Suo S, Wang Y, et al. Tissue-resident memory and circulating T cells are early responders to pre-surgical cancer immunotherapy. Cell. 2022;185(16):2918-2935.e29. DOI: 10.1016/j.cell.2022.07.006
-
Li H, Zandberg DP, Kulkarni A, et al. Distinct CD8+ T cell dynamics associate with response to neoadjuvant cancer immunotherapies. Cancer Cell. 2025. DOI: 10.1016/j.ccell.2025.02.026
-
Uppaluri R, Haddad RI, Tao Y, et al. Neoadjuvant and Adjuvant Pembrolizumab in Locally Advanced Head and Neck Cancer. New England Journal of Medicine. 2025;393:37-50. DOI: 10.1056/NEJMoa2415434
Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.