Combined tumor-directed recruitment and protection from immune suppression enable CAR T cell efficacy in solid tumors

Abstract

CAR T cell therapy remains ineffective in solid tumors, due largely to poor infiltration and T cell suppression at the tumor site. T regulatory (T_reg) cells suppress the immune response via inhibitory factors such as transforming growth factor-β (TGF-β). T_reg cells expressing the C-C chemokine receptor 8 (CCR8) have been associated with poor prognosis in solid tumors. We postulated that CCR8 could be exploited to redirect effector T cells to the tumor site while a dominant-negative TGF-β receptor 2 (DNR) can simultaneously shield them from TGF-β. We identified that CCL1 from activated T cells potentiates a feedback loop for CCR8⁺ T cell recruitment to the tumor site. This sustained and improved infiltration of engineered T cells synergized with TGF-β shielding for improved therapeutic efficacy. Our results demonstrate that addition of CCR8 and DNR into CAR T cells can render them effective in solid tumors.

Fig. 1. CCR8 is used by a subset of transferred antigen-specific T cells to traffic to a murine pancreatic cancer in a CCL1-dependent manner.

(A) Schematic experimental layout to determine the gene expression profile of OT-I T cells after adoptive transfer. (B) Volcano plot for gene expression profile of tumor-infiltrating OT-I T cell compared to a splenocyte benchmark. Gene expression is depicted on the x axis versus the P value depicted on the y axis. (n = 3 mice). (C) RT-PCR gene analysis of explanted Panc02-OVA tumors (n = 4). HPRT, hypoxanthine-guanine phosphoribosyltransferase. (D) Enzyme-linked immunosorbent assay (ELISA) for murine CCL1 on organ lysates of control animals (n = 3) or Panc02-OVA tumor-bearing animals (n = 10). LNIL, lymph node ipsilateral to the tumor; LNCL, lymph node contralateral to the tumor. (E) Fourteen-day Panc02-OVA tumor explants, embedded in OCT medium and frozen for cryosectioning and stained with CCL1 and 4′,6-diamidino-2-phenylindole (DAPI) (representative of n = 3 mice). (F) ELISA for murine CCL1 on CD4⁺ or CD8⁺ magnetic-activated cell-sorted T cells after 24-hour stimulation with anti-CD3 and anti-CD28 antibodies or vehicle solutions only (representative of n = 3 independent experiments). (G) ELISA for murine CCL1 of supernatants generated upon 24- or 48-hour culture of different effector to target ratios, namely, 0, 1, 5, or 10 OT-I T cells to 1 tumor target cell (a constant number of 20.000 Panc02-OVA tumor cells was used) (representative of n = 3 independent experiments). nd stands for nondetectable. ns stands for nonsignificant P > 0.05. Experiments show mean values ± SEM of at least triplicates and are representative of at least three independent experiments. P values are based on two-sided unpaired t test. Data shown in (D) are pooled from three independent experiments.

Fig. 2. CCR8 transduction in OT-I T cells improves ACT efficacy through CCL1-dependent tumor trafficking.

(A) CCR8 staining of live, CD45⁺CD3⁺CD4⁺FoxP3⁺ CD44^high CD62L⁻ eT_reg and CD44^low CD62L⁺ cT_reg Panc02-OVA–infiltrating cells. CCR8 expression on CD4⁺, CD8⁺, eT_reg, and cT_reg T cells. (n = 3 mice). APC, Allophycocyanin. (B) CCR8-GFP transduction efficiency of murine T cells. FITC, fluorescein isothiocyanate. (C) In vitro migration of murine T cells to CCL1. (D) Experimental layout for (E) to (H). (E) Tumor growth curves of mice treated with a single intravenous (i.v.) injection of phosphate-buffered saline (PBS) or 10⁷ GFP-transduced or CCR8-transduced OT-I T cells (n = 5 mice per group). (F) Tumor survival curves of (E). (G) Panc02-OVA-CCL1 tumor growth curves of mice treated with a single intravenous injection of PBS or 10⁷ GFP-transduced or CCR8-transduced OT-I T cells (n = 5 mice per group). (H) Tumor survival curves of (G). (I) ACT tracking experiments in Panc02 or Panc02-CCL1 tumors by flow cytometry. s.c., subcutaneous. (J) Live, CD45.1⁺ tumor-infiltrating T cells (n = 3 mice). (K and L) ACT tracking in mice with tumors in a dorsal skinfold chamber to enable multiphoton intravital imaging (L) and speed quantification of tumor-infiltrating T cells (M) (n = 8 mice). FOV, field of view. Experiments show mean values ± SEM and are representative of three independent experiments, except (G) and (H) that are representative of two independent experiments. P values for (C), (J), (L), and (M) are based on a two-sided unpaired t test. Analyses of differences between groups for (E) and (G) were performed using two-way analysis of variance (ANOVA) with correction for multiple testing by the Bonferroni method. Comparison of survival rates for (F) and (H) was performed with the log-rank (Mantel-Cox) test.

Fig. 3. DNR-transduced T cells retain proliferative capacity despite TGF-β.

(A) CD45⁺CD3⁺CD4⁺FoxP3⁺ CD44^high CD62L⁻ eT_reg and CD44^low CD62L⁺ cT_reg Panc02-OVA infiltrating cells (n = 5 mice). PE, Phycoerythrin. (B) Percentage of T_reg/CD4⁺ in (A) (n = 5 mice). T_reg cells: lymph nodes 15406 ± 5351, spleen 6992 ± 9702, and tumor 893 ± 244 (mean absolute cells ± SD). (C) Percentage eT_reg cells in T_reg cells. (D) Percentage TGF-β⁺ cells in eT_reg (n = 10). (E) Murine TGF-β ELISA on 20.000 tumor cells supernatant (n = 3). (F) DNR downstream signaling. AA, amino acid. (G) Transduction efficiency of murine T cells with DNR. (H) T cell proliferation measured through flow cytometry (n = 3). (I) Experimental layout for tumor challenge with Panc02-OVA. (J) Tumor growth curve in mice treated with a single intravenous injection of PBS or 10⁷ GFP-transduced or DNR-transduced OT-I T cells (n = 5). (K) Tumor survival curves of (J). (L) Flow cytometry ACT tracking in Panc02 tumor–bearing mice treated with an equal mix of DNR-mCherry T cells and GFP T cells (n = 3). Experiments show mean values ± SEM and are representative of three independent experiments, except (A) to (C), which was performed for a total of 5 mice, and (D) a total of 10 mice. P values for (B), (C), (D), (E), and (L) are based on two-sided unpaired t test. Analyses of differences between groups for (J) were performed using two-way ANOVA with correction for multiple testing by the Bonferroni method. Comparison of survival rates for (K) was performed with the log-rank (Mantel-Cox) test.

Fig. 4. CCR8-DNR–transduced CAR T cells mediate superior in vivo efficacy in pancreatic solid tumors.

(A) Schematic representation of murine retroviral construct for T cell engineering. 5’ LTR, long terminal repeat; Amp R, Ampicilin resistance. (B) Transduction efficiency of primary murine T cells with CCR8-DNR-CAR. (C) Boyden chamber assay to assess in vitro migration of primary murine T cells to recombinant murine CCL1 measured by flow cytometry. Migrated cells were normalized for CCR8 expression. n = 3. (D) T cell fold expansion over 48 hours with or without TGF-β on the culture media, measured through flow cytometry (representative of n = 3 independent experiments). (E) ELISA for murine IFN-γ was performed on supernatants generated after 24-hour culture of T cells with Panc02-EpCAM tumor cells, in a ratio of 10 effector cells to 1 target cell (representative of n = 3 independent experiments). (F) Panc02-EpCAM tumor cell killing by different T cell conditions, in a 10-fold concentration, measured over time through xCELLigence. Representative of n = 3 independent experiments. (G) Experimental layout for the ACT experiments depicted in (H) and (I). (H) Growth curves of Panc02-OVA-EpCAM tumors in C57Bl/6 mice that were treated with a single intravenous injection of PBS (n = 5 mice) or 10⁷ CAR-transduced, DNR-CAR–transduced, CCR8-CAR–transduced (n = 8 mice) or CCR8-DNR-CAR–transduced T cells (n = 7 mice). (I) Tumor survival curves of a tumor challenge experiment with Panc02-OVA-EpCAM tumors. Mice were treated with a single intravenous injection of PBS (n = 5 mice) or 10⁷ CAR-transduced, DNR-CAR–transduced, CCR8-CAR–transduced (n = 7 mice), or CCR8-DNR-CAR–transduced T cells (n = 8 mice). Experiments show mean values ± SEM and are representative of three independent experiments. P values for (C) to (E) are based on two-sided unpaired t test. Analyses of differences between groups for (F) and (H) were performed using two-way ANOVA with correction for multiple testing by the Bonferroni method. Comparison of survival rates for (I) was performed with the log-rank (Mantel-Cox) test.

Fig. 5. CCR8-DNR-CAR T cells are effective in human tumor models.

(A) Schematic representation of construct. (B) Transduction efficiency of human T cells with CCR8-DNR-CAR. (C) Human IFN-γ ELISA on 24-hour coculture supernatants. (D) SUIT-2-MSLN killing measured by xCELLigence. (E) Human CCL1 ELISA on 24-hour coculture supernatants. (C to E) 10:1 effector to target cells. (F) Forty-eight–hour T cell expansion, flow cytometry. (G) In vitro migration of T cells to CCL1, flow cytometry. (H) Experimental layout for (I) to (M). Tumor growth and survival curves of SUIT-2-MSLN (I and J), MIA PaCa-2-MSLN (K and L), and SUIT-2-MSLN-CCL1 (M) treated with a single intravenous injection of PBS or 10⁷ CAR-, DNR-CAR–, CCR8-CAR–, or CCR8-DNR-CAR–transduced T cells (n = 5 mice per group). (N) Flow cytometry quantification of CAR T cells per bead per milligram on day 27 after tumor implantation (n = 5 mice). (O) Experimental layout for (P). (P) SUIT-2-MSLN tumor–bearing mice were administered with 10⁷ CAR T cells that were either wild-type (wt) or CCL1 knockout (ko). This was followed by a treatment with 10⁷ CCR8-transduced T cells (n = 7 mice). Experiments show mean values ± SEM of n = 7 healthy donors for (B) to (G), one experiment for (I) to (L) and (P), and two independent experiments for (M) and (N). P values for (C), (E), (F), (G), (N), and (P) are based on a two-sided unpaired t test. (D), (I), (K), and (M) were assessed through two-way ANOVA with correction for multiple testing by the Bonferroni method. (J and L) Survival rate comparison through the log-rank (Mantel-Cox) test.