Abstract Introduction: There is a strong body of evidence indicating that the expression of Neuropilin-2 (NRP2) is enriched in breast cancer stem cells (CSCs) and that NRP2 signaling is critical for breast CSC function and resistance development. For this reason, the rationale for targeting NRP2 as a therapeutic strategy is compelling and timely. A major limitation that has hampered the development of this therapy, however, has been the lack of availability of high-quality anti-human NRP2 monoclonal antibodies (mAbs) that block NRP2 signalling. aTyr Pharma has generated a panel of high-quality, anti-human NRP2 mAbs that have the potential to be developed for the clinical management of breast cancer. A significant advance made by aTyr is that through specific domain reactivity, they have demonstrated differential effects on ligand blocking, receptor homo and heterodimerization and functional activity. Importantly a subset of such antibodies show differential activity in the mammosphere assay of triple negative breast cancer. Results: Flow cytometry was used to assess the specificity of the in-house anti-NRP2 mAbs to NRP2 using A549 wild type versus NRP2 knockout clonal cells. The in-house anti-NRP2 mAbs bound to A549 wild-type cells while showing little or no binding to the NRP2 knockout clonal cells, and exhibited significantly superior specificity, and sensitivity compared to existing commercially available antibodies. Displacement studies demonstrated that the tested anti-NRP2 mAbs showed different capabilities in blocking of VEGF-C or SEMA-3F binding to Expi293-hNRP2 cells, and were categorized as blockers (>90% inhibition), partial blockers (30-90% inhibition), or non-blockers (no obvious inhibition). To further extend the assessment of the biological activity of the anti-NRP2 antibodies, their activity was assessed in a receptor dimerization assay. Vectors encoding a split luciferase pBiT1.1 and pBiT2.1, and a cell permeable substrate, were obtained from Promega corporation. The complete extracellular domain and transmembrane helices of NRP2, FLT4 (VEGFR3), KDR (VEGFR2) and plexin A1 (PLXNA1) were cloned into the vectors and screened for optimal orientation, following established methods. Tested antibodies were able to impair respective VEGF and SEMA3 induced dimerization of receptor pairs NRP2/FLT4, NRP2/KDR and NRP2/PLXNA1. Surprisingly select antibodies show extremely specific and non-obvious functional differentiation. Direct functional assessment of a subset of these antibodies on breast CSCs revealed that the VEGF blocking, but not the SEMA3 blocking anti-NRP2 amAbs had the ability to inhibit serial passage mammosphere formation, an indicator of self-renewal potential. Conclusions: A series of domain specific antibodies to NRP2 have been developed and characterized. These antibodies show differential binding to specific domains of NRP2, can inhibit either VEGF-C or Sema3F binding to NRP2, and differentially effect receptor dimerization. The use of these antibodies enabled us to implicate VEGF-C/NRP2 signaling but not SEMA-3F/NRP2 signaling in the function of breast CSCs. Citation Format: Leslie A. Nangle, Luke Burman, Hira L. Goel, Zhiwen Xu, Kristina Hamel, Nathaniel Bloom, Arthur M. Mercurio. Domain-specific antibodies to Neuropilin 2 implicate VEGF-C and not Semaphorin 3F in breast cancer stem cell function [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1785.