In Guangzhou, inside a colorectal cancer research center, scientists appear to have caught tumor cells doing the biological equivalent of peeling off their name tags before security walks by.
The name tag, in this case, is MHC-I. If that sounds like a government form, fair. But MHC-I is basically a little display tray on the surface of most cells. It shows tiny protein snippets to CD8+ T cells, your immune system's suspicious bouncers, so they can ask: "Are you normal, or are you trying to start a malignant side hustle?"
In a new Gut paper, Cai and colleagues studied microsatellite stable colorectal cancer, or MSS CRC, the kind that usually gives immunotherapy the cold shoulder. Their big finding: when tumors lose or reduce PTEN, they can trigger an NRF2-dependent autophagy pathway that chews up MHC-I. No display tray, no tumor mugshot, no CD8+ T-cell recognition. The immune system is still on patrol, but the suspect has ducked into a fake mustache store.
And here is where it gets interesting: PTEN loss is reported in about 19-36% of MSS colorectal cancers in this study's framing. That is not a rounding error. That is a meaningful slice of patients sitting in the "why did checkpoint blockade do basically nothing?" column.
The Cold Tumor Problem, With Fewer Icicles
Checkpoint inhibitors such as anti-PD-1 drugs work best when T cells are already near the tumor and can recognize what they are looking at. MSI-high colorectal cancers often generate lots of abnormal protein fragments, which makes them easier for the immune system to spot. MSS tumors tend to be quieter. Fewer obvious flags. Less immune chaos. More "nothing to see here, officer."
That clinical gap is well known. A 2024 systematic review of immune checkpoint inhibitors in MSS colorectal cancer found that ICI monotherapy or dual ICI approaches usually produced response rates below 10%, while some combinations did better, especially in patients without active liver metastases (Guven et al., 2024). Translation: the denominator is large, the numerator is rude, and the confidence interval is not bringing snacks.
So researchers have been hunting for the tumor tricks that keep MSS CRC immune-cold. Low antigen load is one piece. Suppressive immune cells are another. Bad antigen presentation is a very big piece, because even a perfectly caffeinated T cell cannot attack what it cannot see.
PTEN Was Supposed to Be the Tumor Suppressor. Then It Got Weird.
PTEN is famous as a tumor suppressor, mostly through its role in restraining the PI3K-AKT pathway. In the usual textbook version, PTEN is a brake pedal. Lose it, and cancer growth signaling gets more enthusiastic than anyone asked for.
But Cai and colleagues found a less expected job. PTEN physically helps stabilize KEAP1, a protein that normally keeps NRF2 under control. NRF2 is a stress-response transcription factor. In normal cells, that can be useful. It helps cells respond to oxidative stress, inflammation, and other molecular nonsense life throws at them.
In PTEN-deficient MSS CRC, though, KEAP1 gets degraded through p62-related machinery. NRF2 then becomes overactive. That overactive NRF2 turns up selective autophagy machinery. Autophagy is the cell's recycling system, which is great when it is clearing junk and less great when it starts tossing out the immune evidence.
The target in this story: MHC-I.
When the Evidence Goes to the Cellular Dumpster
MHC-I downregulation is a classic immune escape move. Reviews have described multiple ways tumors reduce antigen presentation, from genetic defects to signaling changes to trafficking problems (Taylor and Balko, 2022). More recent work has sharpened the autophagy angle too. A 2024 Cell study showed that IRGQ can route MHC-I molecules toward lysosomal degradation, reducing immune visibility (Herhaus et al., 2024). Apparently, tumors read the same playbook and keep dog-earing the "hide the ID badge" chapter.
Cai's study puts PTEN-deficient MSS CRC into that same conceptual neighborhood. Low PTEN correlated with fewer CD8+ T cells and worse outcomes among patients receiving immune checkpoint blockade. The abstract does not hand us a neat hazard ratio or p-value to lovingly frame on the wall, which pains the spreadsheet goblin inside every data person. But the biological chain is unusually tidy: PTEN loss -> KEAP1 loss -> NRF2 activation -> autophagy machinery -> MHC-I degradation -> weaker CD8+ T-cell recognition.
Then the team tested a potential fix. In models, the NRF2 inhibitor ML385 restored surface MHC-I and made PTEN-deficient tumors more sensitive to anti-PD-1 therapy (Cai et al., 2026).
That does not mean ML385 is ready to become tomorrow's standard treatment. Preclinical wins are like promising first dates: encouraging, but you do not merge bank accounts yet. Still, the logic is attractive. Instead of only pressing harder on the T-cell gas pedal, maybe we also force the tumor to put its license plate back on.
Why This Could Matter
If this result holds up in larger cohorts and clinical testing, PTEN status could help identify a subgroup of MSS CRC patients whose tumors resist immunotherapy by destroying MHC-I. That would be useful because MSS CRC is common, checkpoint blockade alone usually disappoints, and oncologists need better ways to separate "unlikely to benefit" from "might benefit if we fix the right upstream problem."
The broader lesson is elegant in a mildly infuriating way: cancer immunity is not only about whether immune cells are strong enough. It is also about whether tumor cells keep the evidence visible. PTEN-deficient MSS colorectal cancer may be hiding that evidence through NRF2-powered autophagy. The immune system cannot convict what the tumor has already shredded.
References
-
Cai R, Zhan W, Lyu X, et al. PTEN deficiency impairs MHC-I-mediated tumour immunity via NRF2-dependent autophagy in microsatellite stable colorectal cancer. Gut. 2026. DOI: 10.1136/gutjnl-2026-338195. PMID: 42173678.
-
Guven DC, Kavgaci G, Erul E, et al. The efficacy of immune checkpoint inhibitors in microsatellite stable colorectal cancer: a systematic review. The Oncologist. 2024;29(5):e580-e600. DOI: 10.1093/oncolo/oyae013. PMCID: PMC11067816.
-
Herhaus L, Gestal-Mato U, Eapen VV, et al. IRGQ-mediated autophagy in MHC class I quality control promotes tumor immune evasion. Cell. 2024;187(25):7285-7302.e29. DOI: 10.1016/j.cell.2024.09.048.
-
Taylor BC, Balko JM. Mechanisms of MHC-I downregulation and role in immunotherapy response. Frontiers in Immunology. 2022;13:844866. DOI: 10.3389/fimmu.2022.844866. PMCID: PMC8920040.
Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.