Cold tumors earn their name through a combination of cell-intrinsic and cell-extrinsic defenses. During oncogenesis, cancers are sculpted by immune surveillance and emerge only after disabling antigen processing and presentation, often through loss of β2-microglobulin or downregulation of major histocompatibility antigens, impairing interferon-γ signaling, and blunting cytotoxic T-cell activity. Beyond these internal evasions, immune-desert and immune-excluded tumors also lack the external cues an adaptive response requires: a microenvironment free of suppressive cancer-associated fibroblasts, tumor-associated macrophages, and myeloid-derived suppressor cells, and sufficient adjuvanticity — the danger signals, or damage-associated molecular patterns, needed to activate dendritic cells and drive cross-priming of CD8-positive T cells. Overcoming both layers of resistance, the authors contend, is precisely what ZBP1-driven signaling is built to do.

ZBP1, short for Z-form nucleic acid binding protein 1 and formerly known as DAI, is an interferon-stimulated gene product equipped with two N-terminal Zα domains that selectively bind nucleic acids in the left-handed Z conformation — unusual helical forms of both RNA and DNA. When ZBP1 encounters Z-form nucleic acids, likely oligomerizing through a process of Z-NA bridging followed by RHIM-domain self-association, liquid-liquid phase separation, and condensate formation, it recruits the kinase RIPK3 through homotypic RHIM interactions. Activated RIPK3 then phosphorylates MLKL, which oligomerizes, migrates to cellular membranes, and punches pores, triggering necroptosis — a lytic, highly inflammatory form of programmed cell death. ZBP1 can also engage RIPK1 through the same RHIM-dependent mechanism, coupling Z-NA sensing directly to NF-κB-driven inflammatory transcription, so the pathway inflames tissue even when cells survive.

The pathway’s virological credentials are central to its therapeutic appeal. During influenza A virus and certain herpesvirus infections, replicating viruses generate nuclear Z-RNAs that ZBP1 detects, initiating an inside-out necroptotic program in which RIPK3 and MLKL are activated within the nucleus, rupturing the nuclear envelope, spilling genomic DNA into the cytosol, and culminating in cell death. Post-necroptotic zombie cells can continue releasing cytokines and danger signals long after membrane rupture, prolonging the inflammatory consequences. Even sub-lethal necroptosis can be durably inflammatory: MLKL pores can leak inflammatory mediators without killing the cell, and ESCRT-III-driven membrane repair can close those pores for at least 16 hours in cultured fibroblasts, with cells eventually dephosphorylating MLKL, reversing phosphatidylserine exposure, and resuming robust growth. Influenza-surviving cells can maintain a type I interferon-driven inflammatory state for days or weeks, and RIPK3 can independently activate NF-κB- and IRF-3-dependent gene expression programs even in cell types refractory to necroptotic rupture, such as cortical neurons.

Translated to oncology, these mechanisms attack every major barrier to checkpoint inhibitor responsiveness at once. Frank necroptosis releases a rich repertoire of damage-associated molecular patterns and cytokines that activate innate sensing pathways, and in tandem with RIPK1- and RIPK3-mediated transcriptional programming it boosts the mobilization, activation, and antigen loading of antigen-presenting cells, driving cross-priming of cytotoxic T lymphocytes. The immunogenicity of the pathway is rooted in evolution: the adaptive immune system evolved to respond to pathogens, not tumors, and ZBP1-triggered programs are de facto antiviral defenses that both eliminate the infected or transformed cell and supply the adjuvant cues for full adaptive immune activation. The catch is that tumors know this too. Many cancers epigenetically silence RIPK3, removing the signaling conduit that ZBP1 needs, and thereby escaping necroptotic elimination during their evolution.

The Fox Chase team highlights an exploitable loophole: while tumor cells frequently lose the ZBP1-RIPK3 machinery, stromal cells within the tumor microenvironment often retain it intact. Inflammatory cancer-associated fibroblasts, or iCAFs, are a particularly attractive target. Although iCAF-rich tumors are characterized by elevated regulatory T cells, impaired cytotoxic CD8 T-cell activity, and extensive fibrotic remodeling, the iCAFs themselves remain competent for ZBP1 and RIPK3 signaling. Preclinical studies show that activating ZBP1 within iCAFs can potently reverse checkpoint inhibitor unresponsiveness, meaning the pathway can be reignited inside the tumor mass even when the malignant cells themselves are genetically deaf to the signal. This reframing turns a suppressive stromal compartment into a therapeutic vulnerability.

How, then, to switch ZBP1 on in the clinic? One strategy targets ADAR1, the p150 isoform of which quench Z-RNAs by binding, editing, and sequestering them before ZBP1 can see them. No clinically viable ADAR1 inhibitors exist yet, but several routes are advancing: all-trans retinoic acid induces proteasomal degradation of ADAR1 and synergizes with PD-1 blockade in pancreatic cancer models; a structurally specific PROTAC called Z-PROTAC selectively degrades ADAR1; and the nuclear export inhibitor selinexor traps ADAR1 p150 in the nucleus, relieving cytoplasmic suppression of ZBP1 and promoting beneficial ZBP1-dependent signaling during antitumor immune responses.

A second family of approaches forces Z-form nucleic acids to accumulate directly. The curaxin compound CBL0137 induces the B-to-Z transition in genomic DNA; ZBP1, which shuttles between cytoplasm and nucleus, detects the resulting nuclear Z-DNA, activates RIPK3, and triggers necroptosis that boosts checkpoint responses in preclinical cold tumor models. CBL0137 is now being tested as an adjuvant to nivolumab and ipilimumab in a Phase I trial for refractory or recurrent melanoma — a rational setting given that 30 to 40 percent of melanoma patients show primary resistance and another 20 to 25 percent eventually acquire it. On the RNA side, spliceosome inhibition with agents such as pladienolide B causes intron retention and the accrual of endogenous retroelement-derived Z-RNAs that ZBP1 directly senses; in small cell lung cancer models this induced ZBP1-dependent death in necroptosis-competent tumor cells and fibroblasts and enhanced immunotherapy responsiveness. Disrupting mRNA 3-end processing works similarly: inhibiting the CPSF3 nuclease with JTE-607, or deleting the exoribonuclease XRN1, causes aberrantly elongated readthrough transcripts laden with Z-forming retroelements to pile up in the nucleus, mirroring what happens during influenza and herpesvirus infections when viral proteins block CPSF activity.

Oncolytic viruses constitute a third and especially actionable class. Talimogene laherparepvec, the FDA-approved oncolytic herpes simplex virus for advanced melanoma, generates copious Z-RNA — mostly from disrupted termination of host pre-mRNAs caused by the viral ICP27 protein blocking CPSF. But HSV-1 also encodes a RHIM-containing protein, ICP6, that blocks ZBP1-RIPK3 signaling in human cells, so the authors argue that engineered HSV-1 variants with a non-functional ICP6 RHIM could be even more powerful checkpoint adjuvants than the approved virus. Pelareorep, an oncolytic reovirus, has shown promise with checkpoint inhibitors in early trials of pancreatic adenocarcinoma and hormone receptor-positive metastatic breast cancer and induces necroptotic death, though whether it activates ZBP1 through Z-RNA remains unproven. Oncolytic influenza A viruses, whose NS1 protein potently disrupts CPSF, represent a still untapped way of triggering the pathway.

Toxicity is the sobering counterweight, because ZBP1-driven necroptosis is anything but benign — during severe influenza it drives the pulmonary immunopathology that makes the disease lethal, and systemically induced ZBP1 activity risks necroinflammatory injury in vulnerable tissues such as lung and gastrointestinal epithelium. The authors therefore propose a prime-trigger-release paradigm: first prime the tumor microenvironment to express ZBP1, for instance by intra-tumoral delivery of a replication-deficient adenovirus encoding interferon-γ; then trigger activation with a pharmacologic Z-NA inducer such as CBL0137 or a CPSF3 inhibitor; finally, let the released antigens and danger signals propagate systemically under checkpoint blockade. Localized activation could achieve an abscopal-like systemic tumor control while sparing healthy tissue. Immune-excluded cancers with dense fibrotic stroma — pancreatic adenocarcinoma, cholangiocarcinoma, desmoplastic melanoma — are flagged as the most compelling initial targets, where ZBP1 ignition in fibroblasts may dismantle stromal walls and grant T cells access to tumor nests. If the strategy survives those tests, a sensor forged by millennia of antiviral warfare may become one of immunotherapy’s most versatile ignition keys.

Subject of Research: ZBP1-driven necroptotic and inflammatory signaling as a strategy to overcome resistance to immune checkpoint inhibitors in immunologically cold tumors

Article Title: Leveraging ZBP1 signaling for cancer immunotherapy

Article References: Desai, P., Wen, C., & Balachandran, S. (2026). Leveraging ZBP1 signaling for cancer immunotherapy. Frontiers in Cell Death, 5, Article 1841599. https://doi.org/10.3389/fceld.2026.1841599

Image Credits: AI Generated

DOI: 10.3389/fceld.2026.1841599

Keywords: ZBP1, necroptosis, RIPK3, MLKL, immune checkpoint inhibitors, cold tumors, Z-RNA, Z-DNA, ADAR1, cancer-associated fibroblasts, oncolytic viruses, cancer immunotherapy