NKG2D CAR T-cell therapy inhibits the growth of NKG2D ligand heterogeneous tumors

Abstract

Tumor heterogeneity presents a substantial barrier to increasing clinical responses mediated by targeted therapies. Broadening the immune response elicited by treatments that target a single antigen is necessary for the elimination of tumor variants that fail to express the targeted antigen. In this study, it is shown that adoptive transfer of T cells bearing a chimeric antigen receptor (CAR) inhibited the growth of target-expressing and -deficient tumor cells within ovarian and lymphoma tumors. Mice bearing the ID8 ovarian or RMA lymphoma tumors were treated with T cells transduced with a NKG2D-based CAR (chNKG2D). NKG2D CAR T-cell therapy protected mice from heterogeneous RMA tumors. Moreover, adoptive transfer of chNKG2D T cells mediated tumor protection against highly heterogeneous ovarian tumors in which 50, 20 or only 7% of tumor cells expressed significant amounts of NKG2D ligands. CAR T cells did not mediate an in vivo response against tumor cells that did not express sufficient amounts of NKG2D ligands, and the number of ligand-expressing tumor cells correlated with therapeutic efficacy. In addition, tumor-free surviving mice were protected against a tumor re-challenge with NKG2D ligand-negative ovarian tumor cells. These data indicate that NKG2D CAR T-cell treatment can be an effective therapy against heterogeneous tumors and induce tumor-specific immunity against ligand-deficient tumor cells.

Figure 1. CAR T cell therapy treats heterogeneous lymphomas

(a) Rae-1 expression on RMA (black) and RMA-RG (gray) tumor cells was assessed by flow cytometry. (b) RMA tumor cells (1 × 10⁴ or 5 × 10⁴), RMA-RG cells (5 × 10⁴) or a 4:1 mixture of both tumor cells (RMA-RG:RMA) were injected s.c. into mice (n = 8). At the same time as tumor inoculation, wtNKG2D or chNKG2D T cells were injected together with the tumor cells s.c.. Tumor area was measured and data are presented in mm². (***, p < 0.001 vs. wtNKG2D). RMA tumor cells (2.5 × 10⁴), RMA-RG cells (2.5 × 10⁴) or a 1:1 mixture of both tumor cells (2.5 × 10⁴ RMA-RG: 2.5 × 10⁴ RMA) were injected s.c. into mice (n = 8). At the same time, wtNKG2D T cells or chNKG2D T cells were injected together with the tumor cells s.c.. (c) Tumor area and (d) survival were measured (n = 8). Data are presented in mm². (*, p < 0.05; **, p < 0.01; ***, p < 0.001 vs. RMA + chNKG2D). Differences between groups in a and b were analyzed by ANOVA and Newman-Keuls post test, and by log-rank test in c. SD is shown.

Figure 2. CAR T cell therapy treats heterogeneous ovarian tumors

(a) Expression of the NKG2D ligand Rae1 on sorted ID8/GFP-shRae1 cells (black unfilled; MFI 17), ID8/RFP cells (black filled; MFI 89), and isotype control cells (gray unfilled; MFI 10). (b) Lysis of target ID8/GFP cells (unfilled triangle) or ID8/GFP-shRae1 cells (filled square) by effector chNKG2D T cells or lysis of ID8/GFP cells by wtNKG2D T cells (unfilled square) is shown for three different effector to target ratios. (*, p < 0.05 vs. wtNKG2D + ID8/GFP; †, p < 0.05 vs. chNKG2D + ID8/GFP-shRae1). (c) IFN-γ production by wtNKG2D or chNKG2D T cells stimulated with media or irradiated RMA, ID8, or ID8/GFP-shRae1 tumor cells was measured. Differences between groups were analyzed by Student’s t-test (***, p < 0.001 vs. media). (d) Mice (10–12 per group) were injected with 10⁶ ID8/GFP-shRae1 cells, ID8/RFP cells, or a 1:1 mixture of both. The mice receiving the 1:1 tumor mixture received the same number (10⁶ cells) of each tumor type as each single tumor control, which was double the total tumor cell inoculum. On weeks 5, 7, and 9 tumor-bearing mice were treated with 5 × 10⁶ chNKG2D or wtNKG2D T cells. The number of solid tumors and tumor cells in the ascites was assessed on week 12 (***, p < 0.001 vs. RFP+WT; †, p < 0.001 vs. RFP+CH). (e) The number of RFP (Rae1⁺) and GFP (Rae1⁻) tumor cells in the ascites of mice bearing 1:1 mixed tumors was assessed. (f) Mice (6–12 per group) were injected with 2 × 10⁶ ID8/GFP-shRae1 cells, ID8/RFP cells, or a 4:1 or 13:1 (ID8/GFP-shRae1: ID8/RFP) mixture of both. Mice were treated with 5 × 10⁶ chNKG2D or wtNKG2D T cells on weeks 5, 7, and 9. Mice were sacrificed twelve weeks post-tumor inoculation and the number of solid tumors was assessed. Differences between groups were analyzed by ANOVA and Newman-Keuls post test. (***, p < 0.001; *, p < 0.05 vs. ID8/RFP (+ chNKG2D T cells); †, p < 0.001 vs. shID8/GFP (+ chNKG2D T cells) and ID8/RFP).

Figure 3. NKG2D CAR T cell therapy confers protection against a NKG2D ligand-negative tumor cell re-challenge

Mice injected with ID8/GFP tumor cells received three doses of chNKG2D T cells 5, 7, and 9 weeks post-tumor inoculation, and these mice became long-term tumor-free survivors. (a) Tumor-surviving mice were re-challenged with 2 × 10⁶ ID8/GFP-shRae1 cells i.p. 250 days after ID8/GFP injection. After 8 weeks, the number of solid tumors and tumor cells in ascites were assessed (***, p < 0.001 vs. naïve). (b) Splenocytes from tumor-surviving mice who did not receive a tumor re-challenge were isolated (at 250 days) and tested for production of IFN-γ in response to ID8, ID8/GFP-shRae1, RMA, and RMA-Rae1 (RMA-R) cells (***, p < 0.001 vs. naïve).