The ZNF737-CXCL10 axis drives immune exclusion and resistance to anti-PD-1 therapy in bladder cancer.

[BACKGROUND] The efficacy of immune checkpoint blockade (ICB) in bladder cancer (BLCA) is limited to a minority of patients. This limitation is primarily due to a non-inflamed or "cold" tumor microenvironment (TME) that lacks T-cell infiltration. Identifying the tumor-intrinsic molecular drivers and the underlying repressive machinery of this phenotype is critical for developing superior biomarkers and novel "cold-to-hot" combination therapies.

[METHODS] We conducted a systematic multi-omics investigation using pan-cancer bioinformatic analyses, bulk and single-cell RNA sequencing (scRNA-seq), and validation across multiple independent patient cohorts. The function and mechanism of ZNF737 were interrogated through in vitro assays, including cell migration, invasion, T-cell chemotaxis, dual-luciferase reporter assays, and MHC-I-restricted cytotoxicity assays using HLA-matched, sorted CD8+ T cells. Crucially, the therapeutic potential was evaluated in an HLA-matched huPBMC-NOG humanized mouse model to exclude potential alloreactive interference. The clinical relevance and spatial context were established using multiplex immunofluorescence (mIF) on patient tissue microarrays (TMAs). Furthermore, we correlated ZNF737 expression with patient outcomes in our institutional and several public immunotherapy cohorts.

[RESULTS] We identified the transcription factor ZNF737 as a key driver of immune evasion specifically in BLCA. ScRNA-seq and mIF analyses revealed that ZNF737 is predominantly expressed by malignant epithelial cells. Mechanistically, tumor-cell ZNF737 was found to directly transcriptionally repress the key chemokine CXCL10. Functionally, this led to impaired CD8+ T cell recruitment and blunted antigen-specific cytotoxicity in vitro while simultaneously promoting cancer cell migration and invasion. In vivo, ZNF737 depletion in humanized mice remodeled the TME from an "immune-desert" to an "inflamed" state, significantly sensitizing tumors to anti-PD-1 therapy. In patient tissues, ZNF737 expression orchestrated a spatially immune-excluded TME, characterized by a physical segregation of CD8+ T cells from tumor nests. Critically, high ZNF737 expression served as a robust and independent predictor of primary resistance to ICB across multiple clinical cohorts.

[CONCLUSIONS] ZNF737 is a pivotal, tumor cell-intrinsic repressor that establishes an immune-excluded and cytotoxicity-resistant TME in BLCA. Our findings establish ZNF737 as a mechanistic driver of immune evasion and a powerful PD-L1-complementary biomarker for predicting ICB resistance. These results position ZNF737 as a high-priority therapeutic target for developing "cold-to-hot" combination strategies in bladder cancer.