Tumor-derived semaphorin 4A improves PD-1-blocking antibody efficacy by enhancing CD8+ T cell cytotoxicity and proliferation

Abstract

Immune checkpoint inhibitors (ICIs) have caused revolutionary changes in cancer treatment, but low response rates remain a challenge. Semaphorin 4A (Sema4A) modulates the immune system through multiple mechanisms in mice, although the role of human Sema4A in the tumor microenvironment remains unclear. This study demonstrates that histologically Sema4A-positive non-small cell lung cancer (NSCLC) responded significantly better to anti-programmed cell death 1 (PD-1) antibody than Sema4A-negative NSCLC. Intriguingly, SEMA4A expression in human NSCLC was mainly derived from tumor cells and was associated with T cell activation. Sema4A promoted cytotoxicity and proliferation of tumor-specific CD8⁺ T cells without terminal exhaustion by enhancing mammalian target of rapamycin complex 1 and polyamine synthesis, which led to improved efficacy of PD-1 inhibitors in murine models. Improved T cell activation by recombinant Sema4A was also confirmed using isolated tumor-infiltrating T cells from patients with cancer. Thus, Sema4A might be a promising therapeutic target and biomarker for predicting and promoting ICI efficacy.

Fig. 1.. Sema4A expression in tumors is linked to preferable prognosis and T cellactivation gene signatures in patients with NSCLC and HNSCC.

(A) OS curves of patients in the TCGA dataset who were followed up for 5 years were compared between the SEMA4A-high (red, upper 50%) and SEMA4A-low (black, lower 50%) groups for LUAD (n = 253 versus 252, **P < 0.01, log-rank test) and HNSCC (n = 259 versus 260, *P < 0.05, log-rank test). HR, hazard ratio; CI, confidence interval. (B) GSEA results of the TCGA dataset showing representative gene signatures that are significantly enriched in SEMA4A-high tumors compared to SEMA4A-low tumors. NES, normalized enrichment score; FDR, false discovery rate. (C) IHC staining for Sema4A in NSCLC specimens collected before the initiation of PD-1–blocking treatment. Sema4A staining (brown). Magnification, ×20. Scale bar, 100 μm. (D) Number of cases who responded or did not respond to the anti–PD-1 antibody pembrolizumab in the Sema4A-positive group (34 responders of 42 total cases) versus the Sema4A-negative group (7 responders of 20 total cases) (***P < 0.001, Fisher’s exact test). Red represents responders, who completed six or more courses, and black represents nonresponders, who completed fewer than six courses. (E) PFS of patients with NSCLC treated with pembrolizumab was compared between Sema4A-positive (red, n = 42) and Sema4A-negative (black, n = 20) groups. ***P < 0.001, log-rank test.

Fig. 2.. Sema4A expression in cancer cells impedes tumor progression in murine models.

(A) Tumor volume and (B) tumor weight in a syngeneic murine model are shown. Mice were administered KP^OVA with or without Sema4A overexpression [OVA-mock (O), black; OVA-Sema4A (O4A), red]. n = 4 per group, *P < 0.05, Student’s t test. The experiment was performed three times with similar results. (C) Tumor volume and (D) tumor weight in a syngeneic murine model are shown. Mice were administered KP^O or Sema4A overexpression with different expression levels [O, black; OVA-Sema4A-low (O4A-low), blue or OVA-Sema4A-high (O4A-high), red]. n = 6 per group, *P < 0.05, **P < 0.01, ***P < 0.001, one-way analysis of variance (ANOVA). The experiment was performed three times with similar results. (E) CD8 staining (brown) of tumors in a subcutaneous tumor model. Magnification, ×20. Scale bar, 100 μm. (F) CD8⁺ cell count in tumors were compared between KP^O and KP^O4A. n = 8 per group, ***P < 0.001, Student’s t test. (G) Granzyme B staining (brown) of tumors in the subcutaneous tumor model. Magnification, ×20. Scale bar, 100 μm. (H) Granzyme B⁺ cell count in tumors were compared between KP^O and KP^O4A. n = 8 per group, ***P < 0.001, Student’s t test. (I) Kaplan-Meier survival curves of mice in an intratracheal injection model administered KP^O or KP^O4A. n = 7 per group, P = 0.0079, log-rank test. The experiment was performed three times with similar results. Results are expressed as the mean ± SEM.

Fig. 3.. Sema4A expression in LUAD cells improve antitumor immunity in murine models.

(A) Tumor weight, (B) immune cell population profiling, (C) differentiation and activation, and (D) exhaustion markers of tumor-infiltrating lymphocytes (TILs) in KP^OVA-mock (KP^O) versus KP^OVA-Sema4A (KP^O4A) at day 8 after injection were analyzed by flow cytometry. n = 8 per group, NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, Student’s t test. (E) Tumor growth curves of mice subcutaneously injected with KP^O or KP^O4A cells (3 × 10⁶ cells) in Rag2 KO mice. n = 5 per group. (F) Tumor growth curves and tumor weight of mice subcutaneously injected with KP^O or KP^O4A cells (3 × 10⁶ cells) and treated with anti-CD8 depletion antibody (Anti-CD8) or isotype control. n = 5 or 6 per group, ****P < 0.0001, one-way ANOVA. The experiments were performed at least three times with similar results. Results are expressed as the mean ± SEM.

Fig. 4.. rSema4A increases the sensitivity to antiPD-1 antibody treatment and the cytotoxic activity of CD8⁺ T cells.

– (A) Tumor growth curves in the subcutaneous tumor model. Anti–PD-1 antibody or isotype control were injected into mice, and therapeutic effects of KP^OVA-mock (KP^O) and KP^OVA-Sema4A (KP^O4A) were compared (n = 7 per group). The experiment was performed three times with similar results. (B) Representative flow cytometry data of the frequency of OVA-tetramer⁺ and IFN-γ production in CD8⁺ T cells from TILs. T cells from TILs were cultured with no stimulation, anti-CD3 antibody, and both anti-CD3 antibody and rSema4A. (C) Percentage of OVA-tetramer⁺ and IFN-γ⁺ in CD8⁺ T cells after anti-CD3 antibody stimulation versus anti-CD3 antibody and rSema4A stimulation. n = 5 per group, *P < 0.05, **P < 0.01, paired t test. The experiment was performed three times with similar results. (D) mRNA expression of Ifng, Prf1, and Gzmb in CD8⁺ T cells was compared after anti-CD3 antibody stimulation and anti-CD3 antibody plus rSema4A stimulation (n = 3). (E) Killing rate of cancer cells: KP^O and KP^O4A cells with OT-1 CD8⁺ T cells. n = 4 to 6 per group, *P < 0.05, **P < 0.01, one-way ANOVA. The experiment was performed three times with similar results. Results are expressed as the mean ± SEM.

Fig. 5.. The Sema4A–plexin B axis promotes CD8⁺ T cell effector function by activating the mTOR-S6K signaling pathway.

(A) CD4⁺ T cells and CD8⁺ T cells from TILs of KP^OVA-mock (KP^O) tumors were incubated with Sema4A-Fc or control human IgG1 (hIgG). After washing, the amount of Sema4A-Fc binding to the cell surface was evaluated by flow cytometry using secondary antibody. The experiment was performed three times with similar results. (B) The expression levels and positivity of Sema4A receptors (plexin B1, plexin B2, and Neuropilin-1) in CD8⁺ T cells were evaluated by flow cytometry. Cells were also stained with an isotype-matched control Ab (gray-filled histogram). n = 5 per group, **P < 0.01, one-way ANOVA. The experiment was performed three times with similar results. (C) Intracellular phosphorylation levels of S6K, AKT (Ser⁴⁷³), and AKT (Thr³⁰⁸) were determined by intracellular staining. The positive rates in CD8⁺ T cells were compared between KP^O and KP^OVA-Sema4A (KP^O4A). n = 7 per group, **P < 0.01, ****P < 0.0001, Student’s t test. The experiment was performed three times with similar results. (D) Percentage of OVA-tetramer⁺ and IFN-γ⁺ CD8⁺ T cells were compared after anti-CD3 antibody stimulation versus anti-CD3 antibody plus rSema4A stimulation, with or without rapamycin (20 ng/ml). n = 3 per group, *P < 0.05, one-way ANOVA. The experiment was performed three times with similar results. Results are expressed as the mean ± SEM.

Fig. 6.. The Sema4A–plexin B axis induces CD8⁺ T cell proliferation by promoting the polyamine biosynthesis pathway.

(A) mRNA expression levels of polyamine biosynthesis pathway components in CD8⁺ T cells were compared after anti-CD3 antibody stimulation versus anti-CD3 antibody plus rSema4A stimulation (n = 3). (B) Summary of an analysis of metabolites involved in the polyamine biosynthesis pathway of CD8⁺ T cells simulated with anti-CD3 antibody (n = 3) versus anti-CD3 antibody plus rSema4A stimulation (n = 4). ND, not detected, *P < 0.05, **P < 0.01, ***P < 0.001, Student’s t test. (C) Percentage of IFN-γ⁺ and mean fluorescence intensity (MFI) of Ki-67 in CD8⁺ T cells after anti-CD3 antibody stimulation versus anti-CD3 antibody plus rSema4A stimulation with or without sardomozide dihydrochloride (10 μM) and eflornithine hydrochloride hydrate (500 μM). n = 3 per group, *P < 0.05, **P < 0.01, one-way ANOVA. SAM, S-adenosylmethionine; dcSAM, decarboxylated S-adenosylmethionine; MTA, methylthioadenosine. Results are expressed as the mean ± SEM.

Fig. 7.. rSema4A improves the efficacy of antiPD-1 antibody treatment in murine models.

– (A) Schematic image of the treatment schedule in the subcutaneous tumor model. (B) Tumor growth curves of mice subcutaneously injected with KP^OVA-mock cells (KP^O) (3 × 10⁶ cells) and treated with anti–PD-1 antibodies or isotype control, with or without rSema4A. n = 9 to 12 per group, *P < 0.05, one-way ANOVA. The experiment was performed three times with similar results. (C) Tumor growth curves of CD8-depleted mice subcutaneously injected with KP^O (3 × 10⁶ cells) and treated with anti–PD-1 antibodies or isotype control, with or without rSema4A. n = 6 to 7 per group. The experiment was performed three times with similar results. (D) Percentage of OVA-tetramer⁺ and IFN-γ⁺ CD8⁺ T cells and T_reg cell count from TILs were compared between the anti–PD-1 antibody group and the anti–PD-1 antibody plus rSema4A group. n = 6 to 7 per group, *P < 0.05, Student’s t test. The experiment was performed three times with similar results. Results are expressed as the mean ± SEM.

Fig. 8.. rSema4A increases the effector function and proliferation of CD8⁺ T cells from human tumors.

(A) Representative flow cytometry data of the frequencies of IFN-γ production and Ki-67 positivity in CD8⁺ T cells from TILs of human cancers. T cells from TILs were cultured with anti-CD3 antibody or both anti-CD3 antibody and rSema4A. FSC, forward scatter. (B) Percentages of CD8⁺ T cells positive for IFN-γ (n = 7) and Ki-67 (n = 5) after anti-CD3 antibody stimulation versus anti-CD3 antibody plus rSema4A stimulation. *P < 0.05, paired t test. LC, lung cancer; LC-PE, pleural effusion of patients with lung cancer.