Therapy-imprinted fibroblast memory in cholangiocarcinoma: rewiring CAF niches for immune evasion and treatment-resistant relapse.

Abstract

Cholangiocarcinoma (CCA) is a highly desmoplastic biliary malignancy in which durable benefit from chemotherapy, immune checkpoint blockade, targeted therapy and local treatment remains limited for many patients. Cancer-associated fibroblasts (CAFs) are increasingly recognized as dynamic regulators of immune escape and therapeutic resistance rather than passive matrix-producing bystanders. Here, we propose therapy-imprinted fibroblast memory as a framework for understanding how treatment-conditioned CAF states may persist after therapeutic or tissue-injury pressure and contribute to immune exclusion, residual tumor-cell survival and relapse in CCA. This concept is distinguished from baseline CAF heterogeneity and transient stromal activation by prior therapeutic exposure, persistence beyond the acute injury phase, measurable molecular or spatial signatures, functional effects on tumor or immune behavior, and potential reversibility or targetability. Throughout the review, CCA-specific studies are prioritized to define disease-relevant CAF states, hepatobiliary injury contexts and therapeutic pressures, whereas evidence from pancreatic cancer, breast cancer, colorectal cancer, lung cancer, fibrosis and wound-healing models is used as mechanistic support for conserved stromal biology and hypothesis generation rather than as direct proof of therapy-imprinted CAF memory in CCA. We first summarize the established baseline CAF landscape in treatment-naive CCA, including myofibroblastic, inflammatory, antigen-presentation-like, perivascular and FAP-positive immunoregulatory CAF states. We then discuss how cytotoxic therapy, immune pressure, targeted therapy, radiotherapy, hypoxia, bile-acid stress, cholestasis and wound-healing signals may reshape these baseline programs into persistent post-treatment CAF niches. We further describe how therapy-conditioned CAF programs may reorganize spatial niches at invasive fronts, fibrotic septa, perivascular regions, perineural compartments and immune-excluded tumor borders, thereby supporting residual tumor cells, restricting cytotoxic lymphocyte access and promoting treatment-resistant relapse. Finally, we outline biomarker and therapeutic strategies for identifying and rewiring pathogenic CAF niches using single-cell and spatial multiomics, multiplex imaging, patient-derived models, FAP imaging, senescence-associated interventions, ECM normalization and rational stromal-immunotherapy combinations. This framework emphasizes that post-treatment CAF evolution should be interpreted against the baseline CCA CAF landscape rather than treated as an extension of static CAF taxonomy alone.