Background: Although immunotherapy is effective in improving the clinical outcomes of patients with bladder cancer (BC), it is only effective in a small percentage of patients. Intercellular crosstalk in the tumor microenvironment strongly influences patient response to immunotherapy, while the crosstalk patterns of plasma cells (PCs) as endogenous antibody-producing cells remain unknown. Here, we aimed to explore the heterogeneity of PCs and their potential crosstalk patterns with BC tumor cells. Methods: Crosstalk patterns between PCs and tumor cells were revealed by performing integrated bulk and single-cell RNA sequencing (RNA-seq) and spatial transcriptome data analysis. A risk model was constructed based on ligand/receptor to quantify crosstalk patterns by stepwise regression Cox analysis. Results: Based on cell infiltration scores inferred from bulk RNA-seq data (n = 728), we found that high infiltration of PCs was associated with better overall survival (OS) and response to immunotherapy in BC. Further single-cell transcriptome analysis (n = 8; 41,894 filtered cells) identified two dominant types of PCs, IgG1 and IgA1 PCs. Signal transduction from tumor cells of specific states (stress-like and hypoxia-like tumor cells) to PCs, for example, via the LAMB3/CD44 and ANGPTL4/SDC1 ligand/receptor pairs, was validated by spatial transcriptome analysis and associated with poorer OS as well as nonresponse to immunotherapy. More importantly, a ligand/receptor pair-based risk model was constructed and showed excellent performance in predicting patient survival and immunotherapy response. Conclusions: PCs are an important component of the tumor microenvironment, and their crosstalk with tumor cells influences clinical outcomes and response to immunotherapies in BC patients. Graphical

Background/Objectives: Diffuse large B cell lymphoma (DLBCL) is a highly heterogeneous and aggressive lymphoma with a high incidence rate. Although modern therapeutic approaches have significantly improved patient survival rates, treatment relapse and drug resistance remain major clinical challenges. Programmed cell death (PCD) promotes tumorigenesis and regulates the tumor microenvironment (TME) and drug sensitivity. Exploring the application potential of PCD in DLBCL could pave the way for new treatment strategies for this malignancy. Methods: We systematically analyzed 13 types of PCD pathways and integrated transcriptomic and clinical data from 832 DLBCL patients (GSE10846, GSE11318, and GSE87371). A PCD-based prognostic signature, termed the Programmed Cell Death Score (PCDS), was constructed using 20 key PCD-related genes. Its clinical relevance was evaluated through survival analysis, drug response profiling, and tumor immune infiltration assessment using CIBERSORT, ESTIMATE, and ssGSEA algorithms. Results: The PCDS robustly stratified patients by survival and outperformed conventional clinical indicators such as age, stage, Eastern Cooperative Oncology Group (ECOG), and lactate dehydrogenase (LDH) in prognostic prediction. High-PCDS tumors were associated with immune suppression, characterized by reduced CD8⁺ T cell infiltration, elevated M2 macrophages, and increased programmed cell death protein 1 (PD-1)/programmed cell death ligand 1 (PD-L1) expression. Drug sensitivity analysis revealed that high-PCDS patients may benefit more from agents like sorafenib and fulvestrant, while low-PCDS patients responded better to NU7441. Functional validation using DLBCL cell lines and xenografts confirmed the oncogenic role of a representative gene (CTH) within the model. Conclusions: This study presents a novel prognostic scoring system derived from multiple PCD pathways that effectively stratifies DLBCL patients by risk and therapeutic responsiveness. Notably, the PCDS is closely associated with key immunological characteristics of the TME. These findings advance personalized treatment strategies and support clinically relevant decision-making in DLBCL.