Abstract Despite advances in immunotherapy, most solid tumors remain resistant to treatment. Immune cell engagers redirect cytotoxic lymphocytes against cancer, but limited tumor access, immunosuppressive microenvironments and systemic immune activation limit efficacy. Here we develop live immune modulating engagers (LIME), a modular platform where non-pathogenic, tumor-tropic Escherichia coli display tandem single-chain variable fragments targeting a tumor-associated antigen and an activating receptor on T or natural killer cells. LIME bridged effector and tumor cells, induced transcriptional programs of T cell activation, metabolism and proliferation, and enhanced cytotoxicity across cancer cell lines and patient-derived organoids. In mouse models, LIME safely accumulated in tumors, outperformed tarlatamab in small cell lung cancer, and induced durable immunity in lymphoma. RAS inhibition and PD-L1 blockade enhanced LIME activity in pancreatic cancer and induced humoral responses. Multi-lineage immune modulation remained tumor-confined, without organ toxicity. These findings establish LIME as a versatile living therapeutic platform for programmable, tumor-restricted immune orchestration.

Patients with metastatic high-grade serous ovarian carcinoma are often unresponsive to immunotherapies; here we identify salt-inducible kinases (SIKs) as key drivers of immunosuppression. Human T cells in the presence of patient ascites express high levels of SIK and the upstream kinase LKB1, whereas SIK inhibition reprograms human T cells and strongly activates antitumor responses. In syngeneic mice with resistant high-grade serous ovarian carcinoma, genetic ablation and pharmaceutical inhibition of SIK consistently demonstrated therapeutic efficacy and survival advantages, and combination of PD-1 blockade with SIK inhibition further extended survival. We identified a major role of T cell-intrinsic SIK2 and -3 signaling in driving immunosuppression in part by TXNIP induction and LYST suppression. Multi-omics analyses on SIK inhibitor therapy revealed reduced disease progression, increased T cell infiltration with enhanced cytotoxicity and effector cytokine IFN-γ, and a shift from immunosuppressive to immunostimulatory cellular niche. We propose SIK inhibitors as a new immunotherapy.

Cyclin D-CDK4/6 is a component of mammalian cell-cycle machinery that drives cell proliferation. Small-molecule inhibitors of CDK4/6 have been approved for treatment of breast cancer patients. In addition to halting cell-cycle progression, inhibition of CDK4/6 can affect other tumor cell-intrinsic and-extrinsic functions. Here, we examined the impact of CDK4/6 inhibition on the CD155-CD226-TIGIT pathway that operates at the interface of tumor cells and the immune environment. We demonstrate that inhibition of CDK4/6 upregulates the expression of surface CD155 protein in cancer cells and downregulates an immuno-inhibitory receptor TIGIT in tumor-infiltrating lymphocytes. We observed these effects in human breast cancer cell lines, in mouse mammary carcinoma allograft models, in freshly resected human breast tumors and in paired pre-/on-treatment biopsies of breast cancers from patients undergoing monotherapy with a CDK4/6 inhibitor. We propose that inhibition of CDK4/6, through its tumor cell-intrinsic and-extrinsic effects, may shift the balance from the immunoinhibitory CD155-TIGIT to the immunostimulatory CD155-CD226 interaction, and through this mechanism may augment the antitumor immunity. Our results suggest that coadministration of CDK4/6 inhibitors and anti-TIGIT antibodies may further promote CD155-CD226-signaling and may have a strong synergistic antitumor effect.

Catalog of qRT-PCR primers, CRISPR guide sequences, and antibodies used in this study

differential gene expression in HupT3 cells following TFAP4 knockout and overexpression

genome-scale CRISPR-Cas9 screens

The cyclin E-cyclin-dependent kinase 2 (CDK2) complex is a component of mammalian cell-cycle machinery that drives cell division. Hyperactivation of cyclin E-CDK2 is frequent in human cancers. Small-molecule CDK2 inhibitors are tested in clinical trials for cancer patients. Here, we report that cyclin E-CDK2 has a cell-cycle-independent function in regulating the global transcriptional program of cancer cells. CDK2 phosphorylates bromodomain-containing protein-4 (BRD4) and regulates its chromatin association. Overexpression of cyclin E and the resulting activation of CDK2 in cancer cells alter the cancer cell transcriptome, repress the expression of interferon-stimulated genes, and confer resistance to immunotherapy. Conversely, CDK2 inhibition has the opposite effect and augments the efficacy of immune checkpoint blockade. CDK2 inhibition also increases tumor infiltration by dendritic cells (DCs) and enhances antigen cross-presentation to CD8 T cells. These studies reveal an additional function of cyclin E-CDK2 in tumorigenesis and identify inhibition of CDK2 with clinically available compounds as a strategy for enhancing immune checkpoint blockade.

Abstract Background: Recurrent glioblastoma (rGBM) remains refractory to immunotherapy due to sparse and suppressed T cell infiltration. We recently reported that survival correlated with immune activation signatures in rGBM patients receiving the oncolytic HSV-1 (oHSV) rQNestin34.5v.2 (CAN-3110). Here, we provide in-situ and molecular evidence that a single oncolytic virus injection can induce durable, tumor-reactive T cell immunity in rGBM. Methods: We integrated highly multiplexed spatial proteomics (CODEX), spatial transcriptomics (Xenium) with custom probes for viral, immune and TCR targets, and bulk TCR-sequencing on paired pre- and post-treatment specimens from a phase 1 clinical trial (NCT03152318). Results: T cell densities strongly increased after treatment, with deep infiltration into viable tumor regions persisting up to two years after a single intratumoral oHSV injection. Cytotoxic GZMB+ T cells were located in close proximity with cleaved-caspase 3+ apoptotic tumor cells, and shorter T cell-tumor distance correlated with longer-progression free survival, demonstrating ongoing anti-tumor immunity. Spatial transcriptomics identified CD8+ T cells states expressing early TCR activation (NR4A1, CD69) and tissue residency (ZNF683 [HOBIT], ITGAE [CD103]) programs enriched in the tumor bed while stem-like T cells localized within lymphoid aggregates. Bulk and spatial TCR analyses revealed in-situ expansion of pre-existing tumoral T cell clones whose amplification correlated with survival. Expanded clones featured tissue resident phenotypes and were positioned closer to tumor cells than non-expanded T cells. Viral remnants were limited to necrotic regions and did not co-localize with T cells, suggesting persistent tumor recognition rather than viral antigen. Conclusion: These results provide in-situ evidence that a single intratumoral oncolytic virus injection can amplify pre-existing T cells clones and induce sustained T cell mediated tumor cytotoxicity even after virus clearance. This suggests that oncolytic virotherapy is as potent T cell activating strategy in rGBM. Citation Format: Maxime Meylan, Ye Tian, Lijian Wu, Alexander L. Ling, Daniel Kovarsky, Graham L. Barlow, Linh D. Nguyen, Jason Pyrdol, Lucas Westphal, Michel Julius, Nicolas L. Gonzalez Castro, Sydney D. Dumont, Andres Santos, Itay Tirosh, Mario L. Suva, E Antonio Chiocca, Kai W. Wucherpfennig. Pre-existing T cells drive durable anti-tumor immunity after oncolytic virus therapy in glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7743.

expression of Hallmark IFNγ/IFNα gene sets in HupT3 cells following TFAP4 knockout and overexpression

T cell states are prognostic in different cancer types. Recent technologies enable joint profiling of T cell RNA and T cell receptor (TCR) sequences at single-cell resolution. Here we present the TCR-RNA Integrating Model (TRIM), a multi-modal variational autoencoder framework that integrates RNA-TCR data and predicts T cell clonality and transcriptional states. TRIM learns a shared representation of the data conditioned on patient, tissue source, and treatment timepoint. We applied TRIM to three independent datasets that included T cells collected before and after checkpoint inhibitor treatment, sourced either from blood and tumor biopsies in patients with head and neck squamous cell carcinoma and colorectal cancer, or from tumor and adjacent tissue in a pan-cancer dataset. In all settings, TRIM accurately predicted intra-tumor T cell clonal expansion and transcriptional status based on T cells from blood or normal tissue before treatment, demonstrating its utility in modeling multimodal T cell data and predicting T cell response to treatment and disease progression.

enriched gene sets in HupT3 cells following TFAP4 knockout and overexpression, generated by the Enrichr search engine

This multi-institutional randomized phase II clinical trial (NCT03414684) investigated the efficacy and safety of carboplatin plus nivolumab compared to carboplatin in metastatic triple-negative breast cancer (mTNBC). The primary endpoint was progression-free survival (PFS) in a modified intention-to-treat (mITT) population of patients with chemotherapy-naïve (i.e., first-line) mTNBC. Secondary endpoints included overall survival (OS), confirmed objective response rate, confirmed clinical benefit rate, time to and duration of confirmed objective response, safety, and tolerability. Clinical outcomes were evaluated in the PD-L1-positive subgroup ( ≥ 1% immune cells; SP142 clone) per central review. Biospecimens were collected for correlative analyses. 75 patients were enrolled and treated between 2/2018-9/2020. Among the mITT population (n = 62), median PFS was 4.2 months with carboplatin plus nivolumab vs 5.5 months with carboplatin. Median OS did not significantly differ between arms (16.8 vs 11.1 months, respectively). In PD-L1-positive mTNBC patients (n = 24), median PFS was 8.3 vs 4.7 months; median OS was 17.6 vs 10.7 months. Grade ≥3 adverse events occurred in 56.8% of patients in the combination arm and 65.8% in the carboplatin arm. Carboplatin plus nivolumab did not significantly improve PFS compared to carboplatin in patients overall; however, a trend toward improved outcomes was observed in PD-L1-positive mTNBC patients. High mutational burden, interferon-gamma signaling, early circulating tumor DNA reduction, low baseline serum thymidine kinase activity, and urea cycle dysregulation in the microbiome emerged as potential predictors of response to immune checkpoint inhibitors with platinum.

Supplementary Figure 1. CODEX panel and experimental procedure. CODEX panel used to identify tumor cell subpopulations, immune cell populations and key cellular markers. (B) Experimental workflow for 5 pre- and post-treatment FFPE samples. ROIs were selected for multiplexed imaging and downstream analysis. (C) U Map clusters show major cell populations. Immune populations showed similar clustering in pre- and post- treatment samples.

Supplementary Table S3: ## Raw count matrix of the RNA-seq data. ## Atg5-KO vs Ctrl-KO (Veh).

Supplementary Fig. S6 shows functional links between autophagy and cytokine signaling pathways.

Supplementary Table S2: ## Raw count matrix of mini-pool CRISPR screens. ## Normalized gene scores in mini-pool screens.

Supplementary Figure 3. Digital quantification from immunohistochemistry. Digital quantification from immunohistochemistry for T cell subsets and marker analysis from paired tumor samples for five patients enrolled to cohort A.

Supplementary Figure 2. Patient enrollment, treatment and disposition per cohort.