Defective NKT cell activation by CD1d+ TRAMP prostate tumor cells is corrected by interleukin-12 with α-galactosylceramide

Abstract

Numerical and functional defects of invariant natural killer T cells (iNKT) have been documented in human and mouse cancers, resulting in a defect in IFN production in several malignancies. iNKT cells recognize glycolipids presented on CD1d molecules by dendritic and related cells, leading to their activation and thereby regulating immune reactions. Activated iNKT cells cytokine secretion and cytotoxicity can inhibit existing and spontaneous tumor growth, progression, and metastasis. We have identified functional iNKT cell defects in the murine TRAMP prostate cancer model. We found that iNKT cells show the ability to migrate into TRAMP prostate tumors. This infiltration was mediated through CCL2: CCR5 chemokine: receptor interaction. Prostate tumor cells expressing CD1d partially activated iNKT cells, as appreciated by up-regulation of CD25, PD-1 and the IL-12R. However, despite inducing up-regulation of these activation markers and, hence, delivering positive signals, prostate tumor cells inhibited the IL-12-induced STAT4 phosphorylation in a cell-cell contact dependent but CD1d-independent manner. Consequently, tumor cells did not induce secretion of IFNgamma by iNKT cells. Blocking the inhibitory Ly49 receptor on iNKT cells in the presence of alpha-GalCer restored their IFNgamma production in vivo and in vitro. However, Ly49 blockade alone was not sufficient. Importantly, this defect could be also be reversed into vigorous secretion of IFNgamma by the addition of both IL-12 and the exogenous CD1d ligand alpha-galactosylceramide, but not by IL-12 alone, both in vivo and in vitro. These data underscore the potential to optimize iNKT-based therapeutic approaches.

Figure 1. Tumor-bearing TRAMP mice exhibit defective cytokine production in vivo by iNKT cells and DC, albeit harboring normal iNKT cell frequencies.

A. Cell suspensions of spleens and livers of WT and TRAMP mice were stained with mAbs against CD3, NK1.1 and PBS-57-loaded CD1d tetramers and analyzed by FACS, identifying comparable numbers of iNKT cells. (n = 5). B. TRAMP mice or age-matched C57Bl/6 mice were injected with α-GalCer. Mice were sacrificed 90 min (IL-4, IL-10) and 300 min later (IFNγ, IL-12 p70) and cytokines determined by ELISA. (*, p,0.05, ***, p<0.001, n.d, not detectable, n = 3).

Figure 2. iNKT cells migrate into murine prostate tumors.

A. WT splenocytes (row 1), TRAMP splenocytes (row 2), and TRAMP prostate-TIL (row 3) were stained with indicated mAbs. CD45⁺MHC class II⁻ cells (left) were gated and analyzed for the expression of CD69. B. Left: CCL2 content in culture medium of TRAMP-C2 cells cultured overnight as measured by ELISA. Histograms: CCL2 expression in TRAMP-C2 cells (left) and primary TRAMP tumor cells (right), (dashed lines: isotype controls). C. CCR5 expression on gated iNKT cells shown in A (dashed lines: isotype controls). D. Liver IHL were incubated with medium, anti-CCR5 or isotype controls and allowed to migrate towards TRAMP-C2 or medium. Migrated iNKT cells were quantified by flow cytometry. Data shown in A–C, and D are representative of 3 and 2 experiments, respectively.

Figure 3. Prostate tumors, primary prostate epithelium, and prostate tumor cell lines express CD1d.

A. Row 1: CD1d expression of TRAMP-C2 cells (bold line) and WT splenocytes (thin line)(dashed line: isotype control), as analyzed by FACS. Top right: CD1d protein expression of TRAMP-C2 cells as determined by Western Blot using rat-anti-mouse CD1d mAb (3C11). Row 2: Fluorescence micrographs show TRAMP-C2 cells stained against mouse CD1d (left) and DAPI (center). B. CD1d expression of primary mouse prostate tumors. Contour plot: Gating of Gr-1⁻, MHC-classII⁻ cells from CD45⁻ tumor cells. Histogram: CD1d expression on gated tumor cells stained against mouse CD1d (solid) or isotype control (dashed line). C. CD1d expression in human CaP cell lines and primary prostate epithelium stained with human CD1d or isotype controls (dashed lines). D. TRAMP-C2 cells express functional CD1d. TRAMP-C2 were pulsed with α-GalCer or vehicle, washed, loaded with anti-CD1d or isotype control mAbs and incubated with DN32 iNKT hybridoma cells overnight. IL-2 in cell culture supernatants was measured by ELISA. Data shown are representative of 3 experiments (**, P<0.005; ***, p<0.001).

Figure 4. TRAMP-C2 cells partially activate iNKT cells by up-regulating activation markers and cell-contact dependently block STAT4 phosphorylation in iNKT cells.

A. TRAMP-C2 and BM-DC were pulsed with α-GalCer or vehicle and used as stimulators for liver MNC as responder cells. Cell culture supernatants were measured 24 hrs (IL-4) and 48 hrs thereafter (IFNγ) by ELISA. B. WT hepatic MNC were exposed overnight to TRAMP-C2 cells, BM-DC, or left alone. Non-adherent cells were stained with indicated mAbs and analyzed by FACS. Histograms show expression of gated iNKT cells. C. Liver MNC were cultured overnight alone or in the presence of TRAMP-C2 cells, pretreated with saturating amounts of anti-mouse CD1d blocking mAbs (3C11) or separated by 0.4 µm transwell membranes. Cells were then stimulated for 30 mins. with 10 ng/ml rIL-12, immediately fixed and stained with mAbs against αβTCR and NK1.1. Cells were then permeabilized using methanol and intracellularly stained against phospho-STAT4 (Ser721). Histograms show phospho-STAT4 staining of electronically gated NKT cells (αβTCR ⁺NK1.1⁺) shown in left plot. Data shown are representative results of two experiments.

Figure 5. IL-12 combined with α-GalCer restores the ability to activate iNKT cells in response to prostate tumor cells.

A. IL12Rβ1 expression on TIL-iNKT and T cells from primary TRAMP prostate tumors, TRAMP spleen and WT spleen ex vivo. Contour plot shows presence of iNKT (CD3⁺CD1d tetramer⁺) and T cells (CD3⁺CD1d tetramer⁻) from TRAMP prostate tumors, gated on CD45⁺ TIL. Data shown are representative of 2 experiments. B. Liver MNC were cultured in the presence or absence of IL-12 with TRAMP-C2 or DC pulsed with α-GalCer or vehicle for 24 or 48 hrs. Cell culture supernatants were tested for cytokine secretion by ELISA. Data shown are representative of 3 experiments (*, P<0.05; ***, p<0.001; n.s., not significant). C. WT C57BL/6 mice were injected with TRAMP-C2 cells s.c. or left untreated. Two weeks thereafter, mice were injected with 150 µg Brefeldin-A and 1 µg α-GalCer with or without IL-12. Five hours later, spleen cells were stained with mAbs against αβTCR and NK1.1, followed by permeablization and subsequent intracellular staining with mAbs against IL-4, IFNγ, or rat IgG isotype controls. Histograms show total cytokine expression of gated NK1.1⁺ cells. D. Tumor-bearing TRAMP mice were i.p. injected with α-GalCer with or without recombinant IL-12. IFNγ levels from sera taken 90 mins later were determined by ELISA (n = 3, *, p<0.05).

Figure 6. Ly49 blockade concomitant with α-GalCer restores IFNγ production in iNKT cells.

A. Liver MNC were cultured with α-GalCer loaded or unloaded TRAMP-C2 cells in the presence or absence of exogenous IL-12 and upon blocking with Ly49C/F/H/I molecules on iNKT cells as indicated. Cell culture supernatants were tested for cytokine secretion by ELISA. Histogram shows MHC- (H2-Kb) expression on TRAMP-C2 cells (solid line) or WT C57Bl/6 DC (bold line), isotype controls (dashed line). B. Tumor-bearing TRAMP mice were i.p. injected with each 50 uG anti-Ly49C/F/H/I mAbs 18 h and 2 hours before 2 µG α-GalCer was injected i.p. Graphs shows IFNγ serum levels (n = 3, p<0.05).

Figure 7. Proposed working model of tumor cell-induced functional iNKT cell defects in TRAMP prostate cancer.

A. Tumor cells deliver 2 signals to iNKT cells: CD1d:TCR-mediated signal is sufficient for a partial activation, i.e. the up-regulation of activation markers and secretion of IL-4 by iNKT cells. MHC-I:Ly49 interaction inhibits enhancement of IFNγ production by iNKT cells. B. Stimulation with high-affinity ligand α-GalCer and IL-12 bypasses tumor cell-induced block of IFNγ production.