The team’s first clue came from combination experiments in living animals. When the researchers paired anti-platelet agents with an antibody that blocks PD-L1—a cornerstone immune checkpoint target exploited by tumors to disarm T cells—they observed a synergistic anti-tumor effect that neither treatment achieved alone. That synergy hinted that platelets were actively contributing to the tumor’s defenses rather than merely bystanders in the tumor microenvironment. When the investigators looked more closely at tumors, they found that activated platelets were adhering directly to tumor-infiltrating CD8-positive T cells. Critically, when platelet activity was suppressed alongside PD-L1 blockade, this adhesion was reduced, and the T cells showed enhanced activation. In other words, stripping platelets away from the immune cells appeared to unleash at least some of the anti-tumor firepower that checkpoint inhibitors are designed to restore.

To determine whether this was a direct cause-and-effect relationship rather than an indirect consequence of the tumor environment, the team built a controlled laboratory system that isolated the players. They used cytotoxic T lymphocytes derived from induced pluripotent stem cells—a line known as CTL3-3—engineered to specifically recognize the WT1 antigen, a protein expressed on many cancer cells and a well-established tumor-associated target. When intact platelets were added to the co-culture of these antigen-specific killer cells and cancer cells, the platelets suppressed the T cells’ cytotoxic activity. The suppression was unmistakable: T cells that would otherwise have efficiently killed their WT1-expressing targets became measurably less lethal in the presence of platelets.

One obvious question was whether platelets were exerting this effect through soluble factors—chemical signals released into the surrounding fluid—rather than through physical contact. To test this, the researchers collected supernatants, the liquid fractions from cultures of activated platelets, and applied them to the T cells. These supernatants did partially suppress cytotoxicity, suggesting that platelet-derived soluble mediators contribute something to the inhibitory effect. But the suppression achieved by soluble factors alone was consistently weaker than the suppression caused by intact platelets. That difference was telling. It implied that the dominant mechanism required the platelets themselves to be physically present, in contact with the T cells, rather than merely whispering chemical instructions from a distance.

Flow cytometry analysis then revealed the nature of that physical contact. Activated platelets express P-selectin, also known as CD62P, a transmembrane adhesion molecule stored in platelet granules and rapidly displayed on the cell surface when platelets are activated. The analysis showed that these P-selectin-bearing platelets directly interact with the CTL3-3 T cells. The team confirmed the interaction with two independent imaging approaches: immunofluorescence staining, which visualized the molecular partners at the interface between the cells, and electron microscopy, which captured the ultrastructural detail of platelets clinging to T cell surfaces. Together, these techniques established that the platelet-T cell adhesion was not an artifact of the assay but a reproducible, physically verifiable phenomenon.

The molecular identity of the T cell-side partner proved to be PSGL-1, the canonical ligand for P-selectin. PSGL-1 is a glycoprotein best known for its role in guiding leukocytes through the bloodstream and into tissues during inflammation, where it engages selectins on vascular endothelial cells. The researchers used neutralizing antibodies to block either P-selectin on the platelets or PSGL-1 on the T cells. Either intervention inhibited the adhesion between the two cell types and, crucially, restored the cytotoxic activity of the T cells in the presence of platelets. This bidirectional blockade experiment is the strongest evidence in the study that the P-selectin–PSGL-1 axis is the functional conduit through which platelets suppress tumor killing. When the handshake is interrupted, the T cells recover their ability to destroy their targets.

Equally important is what the platelets did not do. The researchers found that the interaction suppressed the antigen-specific killing activity of the T cells without affecting their proliferation. The T cells continued to divide and expand normally; they simply became less effective executioners while in contact with platelets. This distinction matters for therapeutic thinking. A mechanism that throttles the cytotoxic function of T cells while leaving their numbers intact suggests a reversible, contact-dependent form of immune regulation—one that could potentially be switched off pharmacologically without causing the broader immune dysfunction that might accompany agents that deplete or globally suppress T cells. It also distinguishes this mechanism from classical checkpoint pathways such as PD-1/PD-L1, which alter T cell activation states through signaling rather than through physical sequestration.

The broader context of this work sits at the intersection of hematology and tumor immunology, a crossroads that has attracted growing attention in recent years. Platelets are known to interact with cancer cells in ways that facilitate tumor progression and enhance malignancy—they shield circulating tumor cells from immune surveillance in the bloodstream, promote metastatic seeding, and release a cargo of growth factors and immunomodulatory molecules. Cancer patients frequently exhibit platelet activation and thrombocytosis, and high platelet counts have been associated with poorer prognosis across multiple tumor types. What has remained murky is the direct effect of platelets on the anti-tumor activity of CD8-positive T cells, the population that immunotherapies from checkpoint inhibitors to engineered T cell therapies ultimately depend upon. By demonstrating a direct, receptor-defined inhibitory interaction, this study fills a conceptual gap and provides a mechanistic explanation for why anti-platelet strategies might synergize with immunotherapy.

The therapeutic implications are tantalizing but must be tempered with caution. The combination of anti-platelet agents and PD-L1 blockade produced synergistic anti-tumor effects in the preclinical setting, and the mechanistic work identifies P-selectin and PSGL-1 as druggable nodes—both are accessible to neutralizing antibodies, and agents targeting the selectin pathway already exist in other clinical contexts. Yet translating these findings into patients will require careful navigation of platelet biology’s essential role in hemostasis. Systemic platelet inhibition carries bleeding risks, and any strategy would need to selectively disrupt the platelet-T cell interaction without compromising clotting. Blocking the P-selectin–PSGL-1 axis specifically, rather than platelet function broadly, may offer a narrower therapeutic window, but this remains to be tested in clinical studies. The current work was conducted in animal models and in vitro systems, and tumor immunology is notorious for promising mechanisms that fail to survive translation.

Nevertheless, the study reframes platelets from passive clotting agents into active participants in the immune battlefield of cancer, and it does so with unusual mechanistic precision—identifying the cell types, the contact dependence, the receptor-ligand pair, and the functional consequence, all within a single coherent experimental arc. For a field racing to explain why immunotherapies work spectacularly in some patients and fail in others, the discovery that activated platelets can handcuff CD8-positive T cells through P-selectin binding offers a new variable to measure, a new biomarker to explore, and a new combination to test. If subsequent studies confirm that the same interaction operates in human tumors and that disrupting it safely enhances checkpoint inhibitor responses, the humble platelet—long celebrated for saving lives one clot at a time—may find itself recast as an accomplice of cancer, and its inhibition as an unexpected weapon in the immunotherapy arsenal.

Subject of Research: The inhibition of CD8-positive T cell anti-tumor activity by platelet P-selectin binding to PSGL-1

Article Title: P-Selectin on activated platelets inhibits the anti-tumor activity of CD8+ T cells through PSGL-1 binding

Article References: Nishiguchi, S., Yokomura, M., Gomibuchi, Y., Nagano, S., Yasunaga, T., Hirohashi, Y., Kawamoto, H., Takagi, S., & Katayama, R. (2026). P-Selectin on activated platelets inhibits the anti-tumor activity of CD8+ T cells through PSGL-1 binding. Cancer Immunology, Immunotherapy. https://doi.org/10.1007/s00262-026-04578-7

Image Credits: AI Generated

DOI: 10.1007/s00262-026-04578-7

Keywords: platelets, P-selectin, PSGL-1, CD8-positive T cells, cancer immunotherapy, PD-L1 blockade, WT1 antigen, cytotoxic T lymphocytes, tumor microenvironment, immune checkpoint inhibitors, iPSC-derived T cells, tumor immunology

The finding rests on a deceptively simple biological insight. Cancer cells are addicted to nucleotides, the raw material of DNA and RNA, and they ramp up ribonucleotide production by 6 to 11-fold compared with normal proliferating cells to fuel their relentless division, metastasis, and drug resistance. But this hyperactive metabolism creates a vulnerability: when the balance between ribonucleotides and deoxyribonucleotides is thrown off, mitochondria become stressed, and their DNA — normally sealed inside these energy-producing organelles — leaks into the cytosol. The innate immune sensor cGAS recognizes this misplaced DNA and activates the STING pathway, prompting the cancer cell to pump out type I interferons and recruit the immune system against itself.

Fluoropyrimidines already exploit this logic in part. As antimetabolites, they disrupt nucleotide synthesis, damage DNA, and can ignite cGAS–STING signaling. Yet their clinical power has long been blunted by two problems: they are metabolically unstable, and they must compete with the supraphysiological pools of endogenous nucleotides that cancer cells maintain. Despite decades of optimization — enzyme inhibitors, the prodrug capecitabine, the S-1 regimen, trifluridine-tipiracil — outcomes in advanced colorectal cancer, which kills roughly 700,000 people each year, remain stubbornly poor. The Nanjing University-led team hypothesized that rather than fighting the tumor’s nucleotide surplus, they could weaponize it.

Through systematic screening of nucleotide combinations with floxuridine, they discovered that pyrimidine ribonucleotides — especially CMP — transformed the drug’s activity. The synergy was strikingly specific: CMP alone did nothing, and it barely enhanced the widely used drug 5-fluorouracil or its metabolite fluorouridine. Only floxuridine and its active metabolite FdUMP, which feed directly into deoxyribonucleotide metabolism, responded. This selectivity traces to metabolic partitioning: 5-fluorouracil preferentially enters RNA metabolism, while floxuridine is channeled toward FdUMP, the molecule that covalently inhibits thymidylate synthase, the enzyme that makes the essential DNA building block TMP.

CRISPR–Cas9 library screening revealed exactly which enzymes were holding floxuridine back. Five pyrimidine metabolism genes — CMPK1, CMPK2, TYMP, UPP1, and UPP2 — emerged as the top hits that, when lost, made cells more sensitive to the drug. Analyzing over 4,700 patient samples from The Cancer Genome Atlas, the team showed that patients whose tumors expressed low levels of these five genes survived significantly longer, underscoring their clinical relevance. Remarkably, CMP and its intracellular metabolites, including uridine and 2′-deoxyuridine, were able to engage all five enzymes simultaneously, as demonstrated by drug affinity responsive target stability assays. No existing small-molecule inhibitor achieves this multi-target blockade, and developing one would be costly and unpredictable.

The metabolic consequences were dramatic. Liquid chromatography–tandem mass spectrometry showed that CMP boosted peak intracellular FdUMP levels nearly 500-fold and extended the metabolite’s half-life from 0.14 hours to 3.45 hours — a more than 20-fold increase in persistence. Cellular thermal shift assays confirmed that the accumulated FdUMP bound and stabilized thymidylate synthase more effectively, deepening the blockade of TMP synthesis. When the team supplemented cells with TMP itself, the synergy vanished entirely, proving the mechanism runs through thymineless death rather than direct interference with DNA replication. Purine nucleotides, by contrast, were cytotoxic at high concentrations and offered no benefit.

Depleted TMP synthesis triggered exactly the mitochondrial crisis the researchers predicted. Quantitative PCR along the mitochondrial genome revealed a position-dependent gradient of copy-number loss spreading from the replication origin — a molecular fingerprint of stalled replication forks. Damaged mitochondrial DNA then escaped into the cytosol, largely through VDAC1 channels: blocking VDAC1 oligomerization with the inhibitor VBIT-4 suppressed the interferon response just as effectively as depleting mitochondrial DNA altogether. Pharmacological inhibition of cGAS with RU.521 or of STING with H-151 likewise shut down the signal, and tumors grew poorly in response to the combination only in mice with functional STING.

Translating the chemistry into a clinic-ready formulation required solving a delivery dilemma. CMP and FdUMP are both highly hydrophilic, poorly permeable molecules, and the synergy collapsed when dosing intervals exceeded two hours — evidence that the two must be metabolized inside the same cell at the same time. The team exploited metal coordination chemistry, discovering that ferric iron could bridge the phosphate groups of CMP and FdUMP into self-assembling nanoscale coordination polymers. Wrapped in a PEGylated lipid shell, the resulting FC-NPs measured roughly 107 nanometers, remained stable in the bloodstream, and released their payload only in the acidic environment of endosomes, achieving true synchronized co-delivery.

In mouse models of colorectal cancer, the nanoparticles delivered on their promise. Three intratumoral injections of FC-NPs inhibited MC38 tumor growth by 82 percent, compared with just 31 percent for the free drug mixture, while cutting the required dosing frequency from thrice daily to once daily. Treated tumors flooded with IFN-β, CXCL10, TNF-α, and IFN-γ, and produced markedly more tumor antigen-specific CD8-positive T cells. The efficacy vanished in mice lacking cross-presenting dendritic cells or depleted of CD8-positive T cells, confirming that adaptive antitumor immunity — the downstream product of STING activation — was the engine of the response. Most strikingly, adding anti-PD-L1 antibody to the nanoparticles produced complete tumor rejection in all treated mice, a result neither therapy achieved alone. Because floxuridine is already the standard agent for hepatic artery infusion in colorectal liver metastases, the authors argue that co-delivering CMP in that regional setting offers a direct, actionable path to the clinic, with CMP showing good biosafety in vivo.

Subject of Research: A metabolic strategy in which ribonucleotides amplify fluoropyrimidine-induced mitochondrial DNA release and cGAS–STING activation to potentiate colorectal cancer chemoimmunotherapy

Article Title: Ribonucleotides Amplify Mitochondrial DNA‐Driven cGAS–STING Activation via FdUMP‐ribonucleotide Hybrid Particles to Potentiate Chemoimmunotherapy

Article References: Wang, C., Zhao, C., Yao, Y., Zou, Z., Shi, A., Feng, Z., Yang, M., Yu, Y., Liu, Y., Rui, X., & Wu, J. (2026). Ribonucleotides Amplify Mitochondrial DNA‐Driven cGAS–STING Activation via FdUMP‐ribonucleotide Hybrid Particles to Potentiate Chemoimmunotherapy. Advanced Science, Article e77614. https://doi.org/10.1002/advs.77614

Image Credits: AI Generated

DOI: 10.1002/advs.77614

Keywords: cGAS–STING, mitochondrial DNA, ribonucleotides, floxuridine, colorectal cancer, chemoimmunotherapy, lipid nanoparticles, nucleotide metabolism, FdUMP, thymidylate synthase, PD-L1 blockade, CRISPR screen

Oncolytic viruses occupy a distinctive niche in cancer immunotherapy. Unlike checkpoint inhibitors or cytokine drugs, they combine direct tumor killing with in situ immunization, releasing tumor-associated antigens and damage-associated molecular patterns as they replicate within malignant cells. Vaccinia virus has long been considered one of the most attractive platforms for this approach because it replicates efficiently in tumor tissue, carries a large genome that tolerates the insertion of therapeutic transgenes, and has an extensive clinical safety record from its use as a smallpox vaccine. Yet the immune contexture of many tumors, particularly notoriously immunologically cold malignancies such as pancreatic ductal adenocarcinoma, resists this strategy. Tumors are dominated by suppressive cells, including regulatory T cells and M2-like tumor-associated macrophages, and the cytokine signals that would expand cytotoxic lymphocytes are weak or absent.

Interleukin-15 has emerged as one of the most compelling cytokine candidates to fill that gap. Unlike its structural relative interleukin-2, which expands both effector and regulatory T cell populations and can provoke systemic toxicity, IL-15 selectively promotes the activation, survival, and persistence of natural killer cells and CD8-positive T cells. Critically, IL-15 biology depends on trans-presentation: the cytokine must be produced by one cell, bound by the high-affinity interleukin-15 receptor alpha chain on that same cell, and then presented in membrane form to neighboring immune cells expressing the beta and gamma receptor chains. Soluble IL-15 delivered as a drug has struggled clinically because it lacks this spatially restricted, membrane-associated presentation, leading to short half-life and off-target effects. The Nagoya team reasoned that the ideal way to deliver IL-15 trans-presentation would be to force the tumor cells themselves, as they are being lysed by the virus, to display IL-15Rα loaded with IL-15 on their surface.

To achieve this, the researchers exploited an unusual feature of vaccinia virology. Vaccinia virus encodes A56, a fusion regulatory protein that works with its partner K2 to control cell-to-cell spread and superinfection exclusion, and the related A56 extracellular domain had previously been used in fusogenic oncolytic designs. The team engineered a novel virus, designated IL15Rα-VV, in which the extracellular domain of A56 was replaced with the extracellular domain of IL-15Rα, while the virus simultaneously expressed IL-15. The result is a chimeric membrane protein that anchors IL-15Rα to the surface of infected cells through the viral fusion machinery, coupling localized IL-15 trans-presentation directly to the fusogenic oncolysis program of the virus. In essence, every tumor cell the virus infects becomes a display platform for the cytokine-receptor complex that stimulates nearby killer lymphocytes.

In vitro characterization confirmed that the engineering did not compromise the virus’s core fitness. IL15Rα-VV maintained replication kinetics comparable to the parental vaccinia strain while driving robust surface expression of IL-15Rα on infected tumor cells. Infected cultures showed enhanced cell fusion and increased cytotoxicity, indicating that the A56-anchored design preserved, and in some respects amplified, the fusogenic lytic phenotype. These experiments established that the immunomodulatory payload and the oncolytic engine could be integrated without sacrificing either function, a balance that has frustrated many previous attempts to arm oncolytic viruses with cytokines.

The decisive tests came in bilateral subcutaneous syngeneic pancreatic tumor models, a stringent setting that measures not only local tumor control but also the generation of systemic immunity against untreated, anatomically distant lesions. When IL15Rα-VV was injected intratumorally into one tumor, growth of both the injected and the non-injected contralateral tumors was significantly suppressed compared with animals treated with the parental virus or with an IL-15-secreting vaccinia virus lacking the anchored receptor. This abscopal effect is the hallmark of a therapy that has converted tumor lysis into productive immunity, and it outperformed the strategy of simply flooding the tumor with soluble IL-15, validating the trans-presentation hypothesis at the heart of the design.

Deep immunophenotyping of the treated tumors revealed the mechanism. IL15Rα-VV remodeled the tumor microenvironment by increasing infiltration of natural killer cells and activated CD8-positive T cells while simultaneously reducing the populations that normally shield the tumor: regulatory T cells and M2-like macrophages. The shift is significant because it reverses the suppressive myeloid and regulatory architecture that characterizes pancreatic tumors rather than merely adding effector cells to a hostile milieu. The cytokine signal, delivered in membrane-bound form at the site of viral lysis, appears to act locally where it is needed while avoiding the systemic expansion of undesirable populations that has limited systemic IL-15 and IL-2 therapies.

Immune effects extended well beyond the tumor itself. Tumor-draining lymph nodes from treated animals showed expansion of activated, effector, central memory, and progenitor exhausted CD8-positive T cells, along with increased dendritic cells. The expansion of progenitor exhausted T cells, or Tpex populations, is particularly intriguing, as these self-renewing cells are considered the reservoir from which durable antitumor responses can be reactivated. Consistent with this, although PD-L1 expression increased within treated tumors, a predictable consequence of inflammatory cytokine signaling, IL15Rα-VV reduced the overall frequency of exhausted CD8-positive T cells, and the therapy showed enhanced antitumor efficacy when combined with PD-L1 blockade. This positions the virus as a rational partner for checkpoint inhibitors, with the virus generating and replenishing the T cell pool that checkpoint blockade reinvigorates.

The study demonstrates that A56-mediated surface display of IL-15Rα is a viable strategy to couple fusogenic oncolysis with localized IL-15 trans-presentation, amplifying oncolytic virus-mediated antitumor immunity in models of a cancer type that has proven remarkably resistant to immunotherapy. By engineering the virus to commandeer the trans-presentation axis rather than simply secreting the cytokine, the researchers have created a system in which tumor cell death and immune activation are mechanistically linked at the same membrane. The work, funded in part by the Japan Grant-in-Aid for Scientific Research and KM Biologics, and performed under approved animal protocols at Nagoya University, suggests a path toward oncolytic vaccines that not only kill tumors but actively reprogram their immunological surroundings into an environment in which cytotoxic lymphocytes can be primed, expanded, and sustained.

Subject of Research: An A56-anchored IL-15Rα fusogenic oncolytic vaccinia virus engineered to reprogram the tumor microenvironment and enhance antitumor immunity.

Article Title: A56-anchored IL-15Rα arms a fusogenic oncolytic vaccinia virus to reprogram the tumor microenvironment and enhance antitumor immunity

Article References: Aboalela, M. A., Abdelmoneim, M., Orikono, Y., Matsumura, S., Bustos-Villalobos, I., Takido, Y., Muramatsu, Y., Sibal, P. A., Saito, R., & Kasuya, H. (2026). A56-anchored IL-15Rα arms a fusogenic oncolytic vaccinia virus to reprogram the tumor microenvironment and enhance antitumor immunity. Cancer Immunology, Immunotherapy. https://doi.org/10.1007/s00262-026-04558-x

Image Credits: AI Generated

DOI: 10.1007/s00262-026-04558-x

Keywords: oncolytic virus, vaccinia virus, IL-15, IL-15Rα, cancer immunotherapy, pancreatic cancer, tumor microenvironment, CD8 T cells, natural killer cells, PD-L1 blockade, trans-presentation, fusogenic oncolysis