How a Cellular Stress Switch Helps Tumours Hide From the Immune System

Deep inside every tumour, a quiet molecular negotiation is under way. Cancer cells and the immune cells surrounding them exchange a constant stream of signals — surface molecules, cytokines, chemokines and tiny membrane-wrapped vesicles — and the outcome of that conversation can determine whether a tumour is eliminated or escapes. A new review published in the Journal of Cellular and Molecular Medicine argues that one of the most influential voices in this dialogue has long been overlooked: the unfolded protein response, a stress-signalling network best known for keeping the endoplasmic reticulum running smoothly. The authors, led by Yew Hwang Chee and colleagues working with Afshin Samali and Adrienne M. Gorman, synthesise evidence showing that this ancient quality-control system can be hijacked by tumours to cloak themselves from immune detection, blunt cytotoxic attack and rewire the tumour microenvironment in their favour.

The unfolded protein response, or UPR, exists because the endoplasmic reticulum is the cell’s protein-folding factory. When oncogenic stress, oxidative damage or metabolic strain disrupts folding, misfolded proteins pile up in the ER lumen, a condition known as ER stress. Three transmembrane sensors then sound the alarm: IRE1α, PERK and ATF6. IRE1α, a dual kinase and ribonuclease, splices the mRNA of X-box binding protein 1 to produce the transcription factor XBP1s, which ramps up production of chaperones and folding enzymes, while also degrading selected mRNAs through a process called regulated IRE1-dependent decay. PERK phosphorylates the translation initiation factor eIF2α, throttling global protein synthesis to lighten the folding load while permitting translation of ATF4, which steers the cell toward redox balance and metabolic adaptation. ATF6 travels to the Golgi apparatus, where it is cleaved into a cytosolic transcription factor that boosts folding capacity. Together these branches either restore order or, if the damage is beyond repair, push the cell into apoptosis.

What makes the UPR so consequential for cancer immunology is that the same secretory pathway it governs is the one tumours and immune cells use to talk to each other. Death receptors, immune checkpoint ligands such as PD-L1, chemokines and cytokines all depend on ER function for their production and trafficking. The review lays out three arenas in which tumour-intrinsic UPR signalling shapes the immune battlefield: immune recognition of tumour cells, immune-mediated tumour cell killing, and the broader chemical and vesicular communication between tumour and immune cells. In each arena, the evidence points to a system that can be bent toward immune evasion.

Immune recognition begins with antigen presentation. Tumour-derived mutant proteins are chopped by the proteasome into peptides, ferried into the ER by the transporter associated with antigen processing, loaded onto MHC class I molecules with the help of chaperones such as Tapasin, and displayed on the cell surface for scrutiny by CD8-positive cytotoxic T cells. NK cells, by contrast, monitor MHC-I expression itself, becoming activated when it drops or when stress ligands appear. The review details how UPR signalling can sabotage this display at multiple points. In lung and cervical cancer cells, ER stress raises XBP1s, which induces the microRNA miR-346; this microRNA targets TAP1 mRNA and cuts off the peptide supply to the ER. XBP1 inhibition restores surface MHC-I in tongue squamous cell carcinoma by suppressing TPP2, a protease that devours antigenic peptides. IRE1-dependent decay can even degrade the mRNAs encoding MHC-I heavy chains and Tapasin in dendritic cells, limiting T cell activation from the other side of the synapse. PERK contributes too: by attenuating translation through eIF2α phosphorylation, it reduces the pool of peptides available for loading onto MHC-I molecules.

A second recognition route is immunogenic cell death, a form of dying that broadcasts danger signals. When chemotherapy or radiation triggers ER stress, the chaperone calreticulin flips to the cell surface as an ‘eat me’ signal for dendritic cells via the receptor LRP1, while ATP and HMGB1 spill out of the dying cell to activate immune receptors such as P2X7R and TLR4. Here the UPR can work in the immune system’s favour. PERK-eIF2α signalling has emerged as a major driver of calreticulin exposure, apparently through caspase-8-dependent trafficking of the chaperone from the ER to the plasma membrane via the Golgi. In murine melanoma, doxorubicin-induced ER stress promotes calreticulin exposure, and its engagement of the activating receptor NKp46 strengthens the drug’s anti-tumour effect. Intriguingly, the review notes that ICD can also occur when PERK is inhibited in ER-stressed melanoma cells, with dying cells exposing calreticulin and releasing HMGB1 and ATP independently of caspase activation — a reminder that PERK is important but not universally required, and that the rules depend on the type of stress and the tumour involved. Blocking the ER chaperone BiP, which normally holds the three UPR sensors in check, similarly enhances radiation-induced ICD in human glioma stem cells.

Once a tumour cell is recognised, it must still be killed, and here the UPR again tilts the scales. PD-L1, the co-inhibitory ligand that engages PD-1 on T cells and drives their exhaustion, is upregulated by ER stressors such as thapsigargin or the BiP inhibitor HA15 through IRE1α’s RNase activity, and BiP appears to stabilise PD-L1 protein directly in triple-negative breast, head and neck and pancreatic cancer cell lines. Strikingly, combining HA15 with anti-PD-1 therapy boosts CD8-positive T cell infiltration and cytotoxicity in murine melanoma, showing that UPR-driven checkpoint modulation can be therapeutically exploited. NK cell attack is blunted through a parallel route: IRE1α-XBP1s signalling represses the stress ligand MICA in melanoma cells by inhibiting the transcription factor E2F1, while ER stress lowers MICA/B post-transcriptionally in hepatocellular carcinoma, and PERK-eIF2α-ATF4 signalling regulates additional NK-activating ligands including ULBP1 and B7-H6. Even the extrinsic death receptor pathway is under UPR control, with XBP1s promoting CD95 expression while RIDD suppresses it, and PERK knockdown independently altering CD95 protein levels.

Beyond direct contact, tumour cells flood their surroundings with soluble signals and extracellular vesicles, and the UPR shapes both. ER stress in prostate cancer, melanoma and lung carcinoma cells promotes release of IL-6, IL-23 and TNF-α, which polarise macrophages toward a pro-inflammatory phenotype via TLR4 while simultaneously impairing dendritic cell cross-presentation and CD8-positive T cell activation. By contrast, HA15-induced ER stress in melanoma enhances secretion of the chemokines CXCL9-11 along with IL-6 and TNF-α, drawing cytotoxic T cells into the tumour — an example of the UPR’s double-edged nature. In triple-negative breast cancer and glioblastoma, IRE1α activity remodels the secretome and myeloid cell recruitment, while loss of tumour IRE1α in lung cancer increases dendritic cell infiltration partly by reducing the immunosuppressive lipid prostaglandin E2. Exosomes add another layer: PERK and IRE1α regulate multivesicular body formation, ATF4 limits lysosomal acidification by inhibiting vesicular-ATPase assembly, and ER stress boosts release of exosomes carrying PD-L1 or miR-27a-3p, which push macrophages toward pro-tumourigenic, PD-L1-high states.

The immune cells themselves are not passive bystanders. Tumour-derived factors activate PERK-ATF4 signalling in macrophages within glioblastoma, driving GLUT1 expression, glucose uptake and lactate production that support an immunosuppressive phenotype; in melanoma, the same pathway promotes serine biosynthesis and lipid oxidation through ATF4-PSAT1. PERK also governs myeloid-derived suppressor cell function through NRF2, and its loss triggers mitochondrial dysfunction and STING-dependent immune activation that reinvigorates CD8-positive T cell responses. Pharmacologically, the IRE1α kinase inhibitor Compound 18 reduced tumour-associated fibroblasts and myeloid-derived suppressor cells in breast tumour xenografts, while the RNase inhibitor MKC8866 — now in Phase 2 clinical trials as ORIN1001 — synergised with docetaxel and with anti-PD-1 therapy to increase T cell and NK infiltration in mouse models. PERK deletion in melanoma cells shrank tumours only in mice with functional T cells, underscoring the immune dependence of the effect, and the PERK inhibitor HC-5404 is being evaluated in a Phase 1 trial across multiple solid tumours.

The review’s central message is that UPR signalling is neither friend nor foe to anti-tumour immunity but a context-dependent dial. Acute ER stress can render tumours more immunogenic and vulnerable, whereas chronic stress fosters adaptation and immune suppression, with the outcome shaped by which branch is engaged, the intensity and duration of stress, tumour lineage and the surrounding immune landscape. The authors highlight the tantalising possibility that IRE1α inhibition can convert immunologically cold tumours into hot ones, restoring sensitivity to checkpoint inhibitors, as seen when docetaxel and MKC8866 were combined in triple-negative breast cancer models. Yet because the UPR is essential for the survival and function of secretory immune cells, the therapeutic window may be narrow, demanding careful target selection, dosing and combination strategies. Open questions abound — whether UPR modulation can enhance CAR-T and CAR-NK cell therapies, and which tumour types with high basal UPR activity stand to benefit most — but the answer to the last may already be taking shape: tumours with pre-existing immune infiltration, where the conversation between cancer and immunity is still alive enough to be redirected.

Subject of Research: The role of the unfolded protein response in regulating tumour-immune cell communication in the tumour microenvironment

Article Title: The Unfolded Protein Response as a Modulator of Cancer‐Immune Cell Communication in the Tumour Microenvironment

Article References: Chee, Y. H., Moncan, M., Reidy, E., Gorman, A. M., & Samali, A. (2026). The Unfolded Protein Response as a Modulator of Cancer‐Immune Cell Communication in the Tumour Microenvironment. Journal of Cellular and Molecular Medicine, 30(19), Article e71380. https://doi.org/10.1111/jcmm.71380

Image Credits: AI Generated

DOI: 10.1111/jcmm.71380

Keywords: unfolded protein response, ER stress, tumour microenvironment, cancer immunotherapy, IRE1α, PERK, PD-L1, antigen presentation, immunogenic cell death, NK cells, T cells, exosomes