Tumors alter the extracellular matrix (ECM) and glycosylation patterns, producing epitopes that are limited or absent on normal adult tissues. The O-glycosylated type III connecting segment (IIICS) domain of fibronectin, referred to here as Tn-FN, is one such tumor-associated glycoform. The authors evaluated Tn-FN as a candidate antigen for chimeric antigen receptor (CAR T cell) therapy based on its tumor-restricted presentation and potential as an extracellular-matrix–derived target.
The study generated a CAR using the FDC6 antigen recognition element fused to a BBζ signaling cassette (FDC6-BBζ) to target Tn-FN. This construct was benchmarked against two comparator CARs: L19-BBζ, targeting the extra domain B of fibronectin (EDB-FN), and 5E5-BBζ, targeting the Tn-glycosylated form of MUC1 (Tn-MUC1). The head-to-head comparison was intended to assess relative activation, cytotoxicity, and in vivo efficacy across ECM- and glycoform-directed CAR strategies.
FDC6-BBζ CAR T cells mediated robust, antigen-dependent activation and cytotoxicity in vitro. In these assays FDC6-BBζ outperformed L19-BBζ and achieved levels of activity comparable to the Tn-MUC1–directed 5E5-BBζ. Cytotoxicity was shown to require intact tumor IFNγ receptor 1 (IFNγR1), indicating a dependence on tumor responsiveness to interferon-gamma for optimal killing by these CARs.
In NOD/SCID gamma xenograft models of prostate cancer, FDC6-BBζ demonstrated potent antitumor activity, matching the efficacy of 5E5-BBζ and exceeding that of L19-BBζ in the reported experiments. Both FDC6-BBζ and 5E5-BBζ achieved durable tumor control in these xenograft studies according to the abstract. Specific dosing, cohort sizes, and survival curves were not reported in the abstract and require consultation of the full article for quantitative details.
Tumors treated with FDC6-BBζ and 5E5-BBζ showed increased intratumoral CD3+ T-cell infiltration. The treated tumors also exhibited reduced overlap between tumor tissue and collagen, suggesting that targeting ECM-associated glycoforms can alter tumor matrix architecture and immune cell access.
Effective CAR-mediated cytotoxicity required an intact tumor IFNγR1. The EDB-directed L19-BBζ construct showed an additional dependence on Fas, whereas FDC6-BBζ and 5E5-BBζ were reported to be less dependent on Fas-mediated pathways. These observations indicate that distinct CAR targets and antigen contexts can engage different tumor death pathways and that IFNγ signaling is a common requirement for full activity in this set of constructs.
Pharmacologic activation of tumoral innate sensing via TLR2/6 or TLR4 agonists restored FDC6-BBζ killing of tumor targets lacking IFNγR1. TLR agonism induced broad inflammatory and stress-response transcriptional programs within tumor cells, reprogramming tumor state and enabling CAR T–mediated cytotoxicity despite impaired IFNγ signaling.
Follow-up perturbation studies implicated caspase-dependent cell-death mechanisms and suggested a contribution from inflammasome-linked signaling in the restored cytotoxicity after TLR agonism. Notably, blockade of ferroptosis did not prevent the restored killing, indicating that ferroptotic pathways were not the primary mechanism of death in this context according to the abstract.
The data support Tn-FN as a glycoform-restricted, ECM-derived CAR target suitable for therapeutic development. The findings also demonstrate that combining CAR T cells with innate immune agonists targeting TLR2/6 or TLR4 can reprogram tumor cell state and overcome resistance linked to defective IFNγ signaling. This suggests a potential therapeutic strategy of pairing ECM- or glycoform-targeted CARs with tumor-directed innate agonists to address tumor-intrinsic resistance mechanisms.
The authors disclose that two investigators (T.R.K-P. and A.D.P.) are inventors on patents related to CAR T therapies targeting Tn-FN; all other authors reported no potential conflicts. The abstract does not provide detailed quantitative methods, exact agonists or doses used for TLR stimulation, cohort sizes, or full statistical outcomes; those specifics were not reported in the PubMed abstract and require review of the full text (PMCID PMC13488855) for complete experimental parameters and data.