Scientists Uncover Why Esophageal Cancer Shrugs Off PD-1 Blockade

Immunotherapy has transformed the treatment landscape for many cancers, yet esophageal cancer remains one of the stubborn exceptions. While checkpoint inhibitors that block the programmed death protein 1, or PD-1, have produced dramatic responses in melanoma and lung cancer, a large fraction of patients with esophageal cancer either fail to respond at all or relapse after an initial period of control. A new study published in BMC Medicine by a team led by researchers at Tangdu Hospital of the Fourth Military Medical University in Xi’an, China, offers a detailed mechanistic explanation for this failure. The researchers report that a protein called signal regulatory protein alpha, or SIRPα, produced by the esophageal cancer cells themselves, acts as a double-edged driver of resistance: it simultaneously pushes tumor cells into a proliferative dominance that lets them outgrow therapy and exhausts the T cells that would otherwise destroy them.

The investigation began with an unusually direct experimental design. The team collected twenty-one esophageal cancer samples from patients who had either received anti-PD-1 treatment or had not, and subjected them to single-cell RNA sequencing, a technique that profiles gene expression in individual cells rather than in bulk tissue. This allowed the researchers to see, cell by cell, what changed inside tumors when PD-1 blockade was applied. What emerged was striking: after anti-PD-1 therapy, the tumor cells showed a marked decrease in their intrinsic expression of SIRPα. The finding suggested that SIRPα-high tumor cells might be the ones being eliminated by successful immunotherapy, while cells that retained or re-established the protein could survive and dominate the tumor after treatment.

To test whether SIRPα levels actually matter for patients, the researchers assembled a much larger clinical cohort: 322 esophageal cancer samples and 161 adjacent paracancerous tissue samples. Across this collection, high SIRPα expression correlated with poor prognosis and worse outcomes from therapy. In other words, tumors rich in SIRPα were not only more aggressive but also less responsive to the immunotherapies designed to defeat them. This epidemiological signal, combined with the single-cell data, positioned SIRPα as a candidate mediator of anti-PD-1 resistance rather than a mere bystander protein whose expression happens to change during treatment.

Having established the correlation, the team turned to the machinery behind it. Using chromatin immunoprecipitation followed by quantitative polymerase chain reaction, dual luciferase reporter assays, co-immunoprecipitation, western blotting, and mass spectrometry, they traced the regulatory circuitry that controls SIRPα in esophageal cancer cells. Their experiments identified the transcription factor Yin-Yang 1, known as YY1, as a direct activator of SIRPα expression. YY1 binds to the SIRPα gene and switches it on, placing SIRPα under the control of a transcriptional program that the tumor itself orchestrates. This was a critical clarification, because SIRPα had previously been studied mainly as an immune checkpoint protein on macrophages, where it recognizes the don’t-eat-me signal CD47 and prevents phagocytosis of cancer cells. The new work shifts attention to SIRPα as a molecule the tumor cells manufacture for their own benefit.

The first mechanism of resistance that the researchers uncovered concerns tumor growth. Through what they describe as the YY1–SIRPα–PI3K/AKT pathway, SIRPα regulates the level of caspase, the family of enzymes that executes programmed cell death. By modulating caspase signaling, SIRPα inhibits apoptosis, the suicide program that damaged or stressed cells are supposed to activate. The consequence is that SIRPα-high tumor cells survive where they should die, continue dividing, and gradually form dominant clones that take over the tumor. When anti-PD-1 therapy arrives, it faces a tumor already organized around a population of hard-to-kill, rapidly proliferating cells. Genetic experiments in which the researchers knocked down SIRPα confirmed the point: lowering SIRPα levels prevented tumor proliferation and reduced recurrence, demonstrating that the protein is not merely associated with aggressive disease but functionally drives it.

The second mechanism is immunological, and it is perhaps the more surprising of the two. The researchers showed that SIRPα present on the surface of tumor cells directly inhibits the killing activity of T cells and drives them into a state of exhaustion, in which the immune cells lose their ability to mount an effective attack. Exhausted T cells express inhibitory receptors, produce less interferon-gamma, and generally behave as if they have been worn down by chronic stimulation. Crucially, the team found that this T-cell suppression operates through a CD47-independent pathway. That distinction matters enormously for drug development, because most existing efforts to target the CD47–SIRPα axis, such as antibodies that block CD47 to unleash macrophages, assume that SIRPα’s effects depend on its interaction with CD47. A CD47-independent mechanism means that simply blocking CD47 would not neutralize what SIRPα is doing to T cells, and that SIRPα itself must be targeted directly.

To test these mechanisms in a living system, the researchers turned to NCG mice, an immunodeficient strain that can be humanized to carry a functional human immune system. Using an adeno-associated virus of serotype 9, or AAV9, carrying a short hairpin RNA against SIRPα, they delivered a gene-silencing therapy to established esophageal tumors and then combined it with pembrolizumab, a clinically approved anti-PD-1 antibody. The results were decisive. Coadministration of the SIRPα-targeted therapy and anti-PD-1 treatment produced a significantly greater antitumor effect than either monotherapy alone. The combination essentially removed the two legs on which resistance stands: suppressing SIRPα weakened the tumor’s proliferative dominance and relieved T-cell exhaustion, allowing the PD-1 blockade to work on a reinvigorated immune response rather than against a wall of suppression.

The technical breadth of the study deserves emphasis, because each method answered a different question. Single-cell RNA sequencing revealed which cells expressed SIRPα and how expression shifted after therapy. The large clinical cohort connected expression to patient outcomes. ChIP-qPCR and reporter assays established YY1 as the upstream transcriptional driver. Co-immunoprecipitation and mass spectrometry mapped the protein interactions downstream of SIRPα. Co-culture experiments, flow cytometry, and tumor-killing assays demonstrated the functional impact of SIRPα on T-cell behavior, measuring how effectively immune cells destroyed tumor cells when SIRPα was present or removed. Finally, the humanized mouse experiments provided the translational proof of concept, showing that a targeted intervention against SIRPα could be delivered in vivo and could convert a nonresponsive tumor into one that responds to checkpoint blockade.

The implications for patients with esophageal cancer are considerable, but the authors are careful to frame them as a starting point rather than a finished therapy. As they note in their conclusions, while the findings suggest SIRPα as a potential therapeutic target, the clinical relevance and safety of intervening against it require rigorous evaluation in future human trials. Several questions remain open. It is not yet known which patients would benefit most from SIRPα-directed treatment, whether SIRPα levels in tumor biopsies could serve as a predictive biomarker for anti-PD-1 response, or what side effects might arise from interfering with a protein that also functions on immune cells. The AAV9 delivery approach used in mice would also need substantial development before it could be applied in humans. Nevertheless, the study provides something that has been missing in esophageal oncology: a coherent, mechanistically grounded account of why PD-1 blockade so often fails, and a concrete strategy, pairing SIRPα suppression with existing immunotherapy, for overcoming that failure. If the findings translate to the clinic, the YY1–SIRPα axis could become a focal point for the next generation of combinatorial immunotherapy in one of the world’s most lethal cancers.

Subject of Research: Tumor-intrinsic SIRPα-mediated resistance to anti-PD-1 immunotherapy in esophageal cancer

Article Title: Tumor-intrinsic SIRPα drives anti-PD-1 resistance in esophageal cancer by coordinating proliferative dominance and T-cell exhaustion: A target for combinatorial immunotherapy

Article References: Tang, X.-Y., Pan, J.-Y., Zhang, X.-X., Du, W.-G., Zhao, X.-J., Dong, H., Zhou, Y.-L., Li, M.-C., Zhang, R.-Z., Shu, C., Shen, Y., Feng, Z.-B., Xu, X.-L., Gao, Y., Yan, X.-L., Yan, B., Ma, N., & Zhao, J.-B. (2026). Tumor-intrinsic SIRPα drives anti-PD-1 resistance in esophageal cancer by coordinating proliferative dominance and T-cell exhaustion: A target for combinatorial immunotherapy. BMC Medicine. https://doi.org/10.1186/s12916-026-05213-z

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05213-z

Keywords: esophageal cancer, SIRPα, anti-PD-1 resistance, immunotherapy, T-cell exhaustion, YY1, PI3K/AKT, CD47-independent pathway, pembrolizumab, tumor microenvironment, single-cell RNA sequencing, combinatorial therapy