Identification and characterization of platinum agent-induced chemoimmunomodulation in triple-negative breast cancer

Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype. Although early-stage outcomes have improved, advanced TNBC (aTNBC) continues to have poor outcomes, with a median overall survival < 18 months and heterogeneous benefit from first-line therapy. Although there have been advances in aTNBC therapy, chemotherapy remains central to curative regimens. Among the most critical dimensions of chemotherapeutic efficacy are immunomodulatory effects elicited by chemotherapy (chemoimmunomodulation; (CIM)). Platinum agents are commonly used in aTNBC and induce CIM through ROS generation, immunogenic cell death, DAMP release, and cGAS-STING/interferon signaling. However, platinum-induced CIM induction is heterogeneous and cell type-dependent, thereby limiting its efforts to optimize its use. The goal of this study was to identify and characterize platinum agent-induced CIM to address this challenge. To achieve this, we applied our Chemoimmunomodulation Induction Classifier (CIMIC) pipeline to delta gene expression values (ΔGE; Δlog 2 (TPM+1)) derived from human TNBC cell lines (N = 8; 3 biological replicates each) profiled via bulk RNA-sequencing pre- and post-48-hour IC 30 cisplatin exposure. CIMIC is an iterative unsupervised clustering pipeline that classifies samples into distinct groups based on their induction of 3,100 genes spanning 19 CIM pathways. Two distinct CIM states emerged: a functional (Fun-CIM) and a dysfunctional (Dys-CIM) state. The Fun-CIM state exhibited significantly elevated induction of antitumoral inflammatory markers ( TNFAIP8L1 and TNSF12 ) and reactive oxygen species potentiators (ABL1 ) (Fun-CIM vs Dys-CIM all fold change (FC) > 1.50; p < 0.05; FDR < 0.10), whereas the Dys-CIM state exhibited elevated induction of DNA-damage repair molecules ( XRCC4 and XPA ), unfolded protein response genes ( SEC61G ) , and proteostasis mediators ( RPS3A and RPS ) (Dys-CIM vs Fun-CIM all fold change (FC) > 1.12 ; p < 0.05; FDR < 0.10) Overrepresentation analysis revealed an enrichment for cytokine production, reactive oxygen species response, and phagocytic potentiation for the genes induced within the Fun-CIM group, and proteostasis programs within the Dys-CIM group (p < 0.05 FDR < 0.15) corroborating our findings. Together, these data show that CIMIC hold promise in resolving platinating agent-induced immunomodulation into functional and dysfunctional induction states from paired pre-/post- transcriptomes. Our results highlight CIM heterogeneity in response to platinating agent administered at the same potency and highlight tumor‑intrinsic stress‑adaptation as a potential key determinant of CIM trajectory. To our knowledge, CIMIC is the first unsupervised pipeline built to classify CIM trajectories. Future work will define baseline predictors of CIM trajectories with respect to platinum agents.