Phosphodiesterase-5 inhibition augments endogenous antitumor immunity by reducing myeloid-derived suppressor cell function

Abstract

Phosphodiesterase-5 (PDE5) inhibitors (sildenafil, tadalafil, and vardenafil) are agents currently in clinical use for nonmalignant conditions. We report the use of PDE5 inhibitors as modulators of the antitumor immune response. In several mouse tumor models, PDE5 inhibition reverses tumor-induced immunosuppressive mechanisms and enables a measurable antitumor immune response to be generated that substantially delays tumor progression. In particular, sildenafil, down-regulates arginase 1 and nitric oxide synthase-2 expression, thereby reducing the suppressive machinery of CD11b+/Gr-1+ myeloid-derived suppressor cells (MDSCs) recruited by growing tumors. By removing these tumor escape mechanisms, sildenafil enhances intratumoral T cell infiltration and activation, reduces tumor outgrowth, and improves the antitumor efficacy of adoptive T cell therapy. Sildenafil also restores in vitro T cell proliferation of peripheral blood mononuclear cells from multiple myeloma and head and neck cancer patients. In light of the recent data that enzymes mediating MDSC-dependent immunosuppression in mice are active also in humans, these findings demonstrate a potentially novel use of PDE5 inhibitors as adjuncts to tumor-specific immune therapy.

Figure 1.

PDE5 inhibition imparts a measurable immune-mediated antitumor effect. BALB/c or BALB/c Rag-2^−/− mice were challenged s.c. with 0.5 × 10⁶ CT26WT (A), C26GM colon carcinoma (B), or TS/A mammary adenocarcinoma (C) cells. (D) C57BL/6 mice were challenged with 0.5 × 10⁶ MCA203 s.c. PDE5 inhibitors were started on day 0. Sildenafil was added to the drinking water or given i.p. daily. Tadalafil was given i.p. daily where indicated in B. Error bar values are shown.

Figure 2.

Sildenafil enhances antitumor CTL activity in vivo. (A) BALB/c mice were either challenged with CT26WT cells, vaccinated with 10⁶ γ-irradiated CT26WT cells, or injected with PBS on day 0. Sildenafil was added to the drinking water where indicated. On day 12, all groups received 10⁷ AH-1–pulsed CFSE^high-labeled BALB/c splenocytes admixed with 10⁷ HA-pulsed CFSE^low-labeled BALB/c splenocytes. The mice were killed 40 h later. Single cell suspensions from spleens and tumor draining lymph nodes were analyzed by FACS. (B) BALB/c mice were either challenged with C26GM cells, vaccinated with 10⁶ γ-irradiated C26GM cells, or injected with PBS on day 0. Sildenafil was added to the drinking water on day 0. After 5 d, the mice received CFSE-labeled target splenocytes as described in A. Analysis was performed as described in the Materials and methods.

Figure 3.

Sildenafil improves the efficacy of ACT. BALB/c mice were challenged s.c. with either C26GM (A) or 4T1-HA (B) cells on day 0. Tumor-specific T cells were transferred on day 1. Sildenafil treatment was started on day 1. Tumors were surgically removed and weighed (right) either on day 10 (A) or 21 (B). *, P^A < 0.00001 by using one-way ANOVA; +, P^T < 0.001 by using a paired t test comparing sildenafil to sildenafil + ACT groups. Error bar values are shown.

Figure 4.

PDE5 inhibition increases intratumoral T cell infiltration and activation of tumor-specific CD8⁺ T cells. (A) BALB/c mice were challenged with CT26WT cells s.c. on day 0, and 20 mg/kg/d sildenafil was added to the drinking water where indicated. The mice were killed 15 d later, and the tumors were stained with hematoxylin-eosin. (B–D) BALB/c mice were challenged with C26GM on day 0. They were treated with sildenafil, received 20 × 10⁶ C26GM vaccine-primed splenocytes from H2^d pIL-2/GFP mice, or were given both treatments as indicated. (B) The tumors were surgically removed 9 d later, and the CD8⁺ T cell infiltrate was determined by flow cytometry. The percentage of CD8⁺ T cells was plotted against the tumor size at the time of tumor harvest. The software-provided tool (Sigma plot) was used to fit a three-parameter exponential decay curve (y = 36.13 + 92xe^−5.93x). Pearson bivariate correlation, P = 0.0000002. Data are derived from three independent experiments, with each containing five mice per group. (C) Intratumoral single cell suspensions were labeled with anti-CD25 or -CD69 antibodies. Data are expressed as the percentage of positive cells gated on the CD8⁺ T cell population. (D) IL-2 production in the ACT groups is reported as the percentage of GFP⁺ CD8⁺ T cells. Data are derived from two independent experiments. The paired t test p-value is reported. (E) BALB/c mice were challenged with 0.5 × 10⁶ C26GM cells on day 0. They were treated with sildenafil starting on day 0, the anti-CD8⁺ depleting antibody, or both. The one-way ANOVA p-values (P^A) are indicated. Data are reported from one of two similar experiments. Error bar values are shown.

Figure 5.

Sildenafil down-regulates IL-4Rα expression on CD11b⁺ cells. CD4⁺, CD8⁺, CD11b⁺, or CD11c⁺ cells were magnetically purified from spleens of C26GM-bearing mice challenged 9 d earlier. CD4⁺ or CD8⁺ T cells were stimulated with ConA for 3 d, with or without 50 μg/ml sildenafil. (A) Purified cells were cultured separately for 2 d with or without sildenafil, harvested, and used to determine cGMP levels, as described in Materials and methods. (B) T cell proliferation was determined by [³H]thymidine incorporation. (C) CD11c⁺ cells were analyzed for B7.2 and MHC class II expression. (D) CD11b⁺ cells were analyzed for IL-4Rα expression by flow cytometry. The t test p-value is reported. Error bar values are shown. MFI, multiplicity of infection.

Figure 6.

In vivo PDE5 inhibition down-regulates tumor-associated MDSC suppressive pathways. BALB/c mice were challenged with 0.5 × 10⁶ C26GM cells and treated with sildenafil starting on day 0 or left untreated for 9 d. CD11b⁺ cells were obtained from the tumors and used to measure (A) intracellular cGMP and (B) IL-4Rα surface expression by flow cytometry. (C) Western blot analysis was performed for NOS2, ARG1, and β-actin expression on purified tumor-associated CD11b⁺ cells. (D) NO production was determined as the concentration of NO₃-NO₂ in the supernatant, and arginase activity was determined on cell lysates and normalized for the number of cells. (E) BALB/c mice were challenged with C26GM, treated with sildenafil or the anti–Gr-1 depleting antibody, both treatments, or left untreated. Best fit of the data was obtained by four-parameter sigmoid curves. ANOVA p-values (P^A) are reported. Error bar values are shown. MFI, multiplicity of infection.

Figure 7.

PDE5 inhibition reverts MDSC suppressive pathways. Splenic CD11b⁺ cells from C26GM tumor-bearing mice were isolated and added to CFSE-labeled splenocytes containing either (A) naive HA-specific CD8⁺ (CL4) cells or (B) naive HA-specific CD4⁺ (6.5) cells. The cultures were stimulated for 4 d with the relevant peptide in the presence or absence of 50 μg/ml sildenafil. Proliferation was evaluated as CFSE dilution by FACS analysis. (C) Splenic CD11b⁺ cells were magnetically purified from B16GM tumor-bearing C57BL/6-NOS^+/+ or C57BL/6-NOS^−/− mice and added to CFSE-labeled splenocytes containing naive OVA-specific CD4⁺ T cells. The cultures were stimulated for 4 d with the relevant peptide in the presence or absence of 50 μg/ml sildenafil. Proliferation was evaluated as CFSE dilution by flow cytometry. Data derived from one of two independent experiments with similar results are reported. Error bar values are shown.

Figure 8.

PDE5 inhibition restores proliferation of head and neck and myeloma lymphocytes. (A) Unfractionated or CD14-depleted PBMCs from MM patients were stimulated with anti-CD3/CD28 antibody–coated beads in the presence of NorNOHA, l-NMMA, both NorNOHA and l-NMMA, sildenafil, or no inhibitor. The CD3⁺ T cell expansion was measured 5 d later by flow cytometry. (B) Ficoll-purified PBMCs from healthy donors (n = 4), head and neck cancer patients (H &N; n = 7), or MM patients (n = 7) were stimulated as described in A in the presence or absence of sildenafil. CD4⁺ and CD8⁺ T cell expansion was measured by flow cytometry 5 d later. Data are reported as fold change. t test p-values are reported. Horizontal lines represent the median, the 10th and 90th percentile.