Robust tumor immunity to melanoma mediated by interleukin-9-producing T cells

Abstract

Interleukin-9 (IL-9) is a T cell cytokine that acts through a γC-family receptor on target cells and is associated with inflammation and allergy. We determined that T cells from mice deficient in the T helper type 17 (T(H)17) pathway genes encoding retinoid-related orphan receptor γ (ROR-γ) and IL-23 receptor (IL-23R) produced abundant IL-9, and we found substantial growth inhibition of B16F10 melanoma in these mice. IL-9-blocking antibodies reversed this tumor growth inhibition and enhanced tumor growth in wild-type (WT) mice. Il9r(-/-) mice showed accelerated tumor growth, and administration of recombinant IL-9 (rIL-9) to tumor-bearing WT and Rag1(-/-) mice inhibited melanoma as well as lung carcinoma growth. Adoptive transfer of tumor-antigen-specific T(H)9 cells into both WT and Rag1(-/-) mice suppressed melanoma growth; this effect was abrogated by treatment with neutralizing antibodies to IL-9. Exogenous rIL-9 inhibited tumor growth in Rag1(-/-) mice but not in mast-cell-deficient mice, suggesting that the targets of IL-9 in this setting include mast cells but not T or B cells. In addition, we found higher numbers of T(H)9 cells in normal human skin and blood compared to metastatic lesions of subjects with progressive stage IV melanoma. These results suggest a role for IL-9 in tumor immunity and offer insight into potential therapeutic strategies.

Figure 1. Deficiency of ROR-γ is associated with inhibited melanoma growth and increased tumor lymphocytic infiltration

B16F10 melanoma cells were injected subcutaneously into RORc^−/− ch and RORc^+/+ ch mice. Tumor growth (a) and mice survival (b) was monitored over time. (c) Tumor infiltrating lymphocytes (TILs) were recovered and counted as described in methods. (d) Expression of IL-17A, IFN-γ, and TNF-a was analyzed in CD4⁺ and CD8⁺T cells of tumor draining LNs by flow cytometry. Data is represented as Mean ± SEM in a–b (n= 8 mice per group (p<. 005: ***), in c (TILs from 4 mice were pooled and analyzed) and in d (n=8 mice per group, p<. 005: ***, p<. 025: **). Two-three additional independent experiments provided similar results.

Figure 2. Increased IL-9 expression in ROR-γ⁻T cells

Sorted naïve T_H cells (CD4⁺CD25⁻CD62Lhigh) from RORc^+/+ and RORc^−/− mice were differentiated under T_H17 polarizing conditions. After 4 days, cells were harvested for transcriptional profiling experiments. Gene expression analysis was performed and expression of a set of genes is depicted as fold change (RORc^−/− vs. RORc^+/+). Two additional microarray analyses provided similar results.

Figure 3. The role of IL-9 in melanoma immunity in IL23R^−/− mice

(a) B16F10 melanoma cells were injected subcutaneously into RORc^−/− ch, and control mice (RORc^+/+ ch). Anti-IL-9 neutralizing antibody was administered to RORc^−/−ch mice. (b) in vitro differentiated T_H17 cells from IL23R^+/+ and IL23R^−/− mice were restimulated and IL-9 and IL-17A were estimated by ELISA. (c) B16F10 melanoma cells were injected subcutaneously into IL23R^−/−and their controls. Neutralizing antibody to IL-9 was administered to both WT and IL23R^−/−mice. Melanoma growth was monitored over time. (d) TILs were isolated from tumor growing in WT mice. IL-9 was stained and quantified by flow cytometry. (e) IL-9 and IL-17 were estimated by ELISA in supernatants from tumor draining LNCs. Data is represented as Mean ± SEM and statistically significant differences were observed compared to controls (p<. 005: ***, p<. 025: **, p<.05: *). At least two additional independent experiments produced similar results.

Figure 4. T_H9 cells inhibit melanoma growth

Differentiated-T_H cells from OT-II mice were generated and transferred into WT mice (WT-C57BL/6: a) or Rag1^−/−C57BL/6 mice (b–c). On the same day, B16F10-ova cells were injected subcutaneously. Melanoma growth (a–c) was monitored overtime. Neutralizing antibodies to IL-9 or isotype was given to T_H9 treated mice (b). Data is represented as Mean ± SEM (a–c, n=6 mice per group) and statistically significant differences were observed compared to No-T cells group (a), and as depicted (b–c) (p<. 005: ***, p<. 025: **, p<.05: *).

Figure 5. Abrogation of IL-9/IL-9R signaling promotes melanoma development, and treatment with rIL-9 inhibits melanoma development

(a, b, d & e) B16F10 melanoma cells were injected subcutaneously into IL9R^−/− and controls (IL9R^−/+) mice (a), normal WT mice (b) and Rag1^−/− mice (d) and Kit W-sh (mast cell deficient) mice (e). (c & f) Lewis lung carcinoma cells were injected subcutaneously into normal WT mice (c) and Kit W-sh (mast cell deficient) mice (f). Where indicated, rIL-9 was administered. Control mice received PBS. Tumor growth was monitored over time. Data is represented as Mean ± SEM (n=4 mice per group, 2–4 additional independent experiment produced similar results, and statistically significant differences were observed compared to controls (p<. 025: **, p<.05: *, ns: not significant).

Figure 6. Presence of T_H9 cells in human skin and peripheral blood

(a–b) Memory T cells from peripheral blood and skin-resident T cells of healthy donors were stained for IL-9, IFN-γ, IL-17 and IL-4 and analyzed by flow cytometry. (c) IL-9 production was measured in cell free supernatants by luminex assay from memory T cells of peripheral blood from healthy donors, skin T cells from healthy donors and tumor-infiltrating lymphocytes of subjects with metastatic melanoma. Data is represented as Mean ± SEM (skin: n=8, PBMCs: n=3, MM: n=8).