Spatial immune profiling to identify GPNMB as a mediator of immune suppression that limits response to PD-1 blockade in TNBC.

Abstract

2582 Background: Breast cancer (BC) is the leading cause of female cancer-related deaths. Triple-negative BC (TNBC), accounting for ~25% of BC-related deaths, carries the poorest prognosis due to limited targeted therapies and the significant, yet still modest overall, benefit achieved with immune checkpoint blockade (ICB). Identifying drivers of immune evasion is essential to expand patient benefit from ICB. Glycoprotein-NMB (GPNMB) is highly expressed in TNBC and associated with immune suppression, metastasis, and poor clinical outcomes. However, its role in shaping the TNBC immune landscape and contributing to ICB resistance remains unclear. We investigated the impact of tumor-derived GPNMB on the tumor immune microenvironment and whether its inhibition enhances ICB efficacy. Methods: Imaging mass cytometry, a high-dimensional spatial proteomics platform enabling simultaneous single-cell and spatial analysis of intact tissue, was used to map the immune landscape of GPNMB-proficient and -deficient EO771 tumors. Therapeutic relevance was evaluated by combining GPNMB loss with anti-PD-1 treatment. Associations between tumor GPNMB expression, immune infiltration, and clinical outcomes were evaluated in human TNBC tumors. Results: GPNMB loss in mammary tumor cells significantly impaired tumor growth in syngeneic, but not athymic (T cell-deficient) mice, suggesting a T cell-dependent mechanism. GPNMB KO tumors showed increased immune infiltration, notably GZMB + /PD-1 + CD8 + T cells, and enhanced formation of stimulatory CD8 + T cell/CD4 + T cell/CD86 + macrophage immune triads. GPNMB loss impaired tumor engraftment in an adoptive OT-1 T cell transfer model and significantly improved response to anti-PD-1 therapy in two independent TNBC models, reducing breast tumor growth and lung metastasis. Therapeutic benefit was accompanied by increased intratumoral T cell infiltration and functional reinvigoration, reflected by increased GZMB + CD8 + T cells. In human TNBC, high tumor GPNMB expression was associated with reduced intratumoral T cell infiltration and poorer clinical outcomes. GPNMB high tumors exhibited reduced infiltration of CD8 + , GZMB + CD8⁺, and CD4 + T cells and were enriched in immune-desert and margin-restricted phenotypes. In contrast, GPNMB low tumors showed increased CD8 + T cell/CD4 + T cell/macrophage triads, indicating enhanced immune activation. Together, these data show that elevated GPNMB expression in human TNBC is associated with T cell exclusion, and an immune-excluded tumor architecture. Conclusions: Our findings identify tumor-derived GPNMB as a clinically relevant mediator of T cell exclusion and suppression in TNBC. Therapeutic targeting of GPNMB in combination with PD-1 blockade may enhance the efficacy of ICB and expand the subset of TNBC patients who derive meaningful clinical benefit, ultimately improving clinical outcomes.