The Neuropilin-2 Axis Regulates Carcinogenesis and Immune Surveillance in the Oral Cavity

Abstract

Introduction: Secreted Semaphorin-3F (SEMA3F) is a chemorepulsive protein in Neuropilin-2 (NRP2)-expressing cells. It guides nervous and circulatory system patterning during embryogenesis. SEMA3F is frequently deleted in human small cell lung cancer and other cancer types including head and neck squamous cell carcinoma, as such it is a postulated as a potential tumor suppressor gene. Recent data from our group has identified NRP2 in CD4+ T cells. Aims: Our goal is to investigate the role of endogenous Sema3F and NRP2 in oral carcinogenesis, tumor progression, and tumor immunity. Methods: Different stages of oral carcinogenesis were evaluated for NRP2 expression. Sema3F expression was evaluated in OSCC cell lines and compared to normal mucosa. Transgenic mice with Sema3F deletion in adult keratinocytes were used to analyze cancer initiation and progression before or after exposure to carcinogen compared to controls. Tongues of Nrp2-global knockout (KO) or wildtype mice were injected with syngeneic oral cancer xenografts and tumors were evaluated for CD4+ and CD8+ T cell infiltration. Antigen-induced hypersensitivity reactions in Sema3F-KO, K14-Sema3F-iKO, Nrp2-KO, and CD4-Nrp2-KO mice were used to evaluate for ear swelling and CD4+ T cell infiltration in experimental mice compared to controls. Results: Nrp2 expression was absent in normal oral and skin epithelium but was upregulated during the late stages of dysplasia in both human and mouse carcinogenesis. Conversely, Sema3F was present in normal mouse tongue epithelium but decreased in OSCC cell lines. When Sema3F was deleted prior to exposure to carcinogen, the majority of control mice (Sema3F-intact) (25 out of 30) developed carcinoma in situ (CIS) or invasive oral squamous cell carcinoma (OSCC), while only 17% of K14-Sema3F-KO mice (4 out of 23) developed CIS, with none progressing to OSCC. When Sema3F knockout was induced post-carcinogen exposure, mice that lacked Sema3F in their epithelium exhibited more infiltrative tumors compared to controls. Oral cancer grafts displayed increased infiltration of CD4+ and CD8+ lymphocytes in Nrp2-/- mice compared to Nrp2+/+ mice. Mice lacking epithelial Sema3F or Nrp2-expressing T cells exhibited prolonged inflammation, tissue swelling, and CD4+ T cell infiltration, whereas control mice resolved quickly. Conclusion: Sema3F is not a tumor suppressor in OSCC, primarily because normal oral epithelium lacks Nrp2 expression and Sema3F cannot function in an autocrine fashion to inhibit carcinogenesis. Furthermore, mice that lacked Sema3F in their epithelium exhibited reduced carcinogenesis. Cancerous and inflamed tissues lacking Sema3F in the epithelium or Nrp2 on CD4+ T cells exhibited increased immune surveillance and T lymphocyte recruitment abrogating cancer initiation and preventing oral cancer progression. We conclude that the SEMA3F/NRP2 pathway suppresses host immunity and provides new potential targets for immunotherapy in oral cancer.