Vaccine-boosted CAR T crosstalk with host immunity to reject tumors with antigen heterogeneity

Abstract

Chimeric antigen receptor (CAR) T cell therapy effectively treats human cancer, but the loss of the antigen recognized by the CAR poses a major obstacle. We found that in vivo vaccine boosting of CAR T cells triggers the engagement of the endogenous immune system to circumvent antigen-negative tumor escape. Vaccine-boosted CAR T promoted dendritic cell (DC) recruitment to tumors, increased tumor antigen uptake by DCs, and elicited the priming of endogenous anti-tumor T cells. This process was accompanied by shifts in CAR T metabolism toward oxidative phosphorylation (OXPHOS) and was critically dependent on CAR-T-derived IFN-γ. Antigen spreading (AS) induced by vaccine-boosted CAR T enabled a proportion of complete responses even when the initial tumor was 50% CAR antigen negative, and heterogeneous tumor control was further enhanced by the genetic amplification of CAR T IFN-γ expression. Thus, CAR-T-cell-derived IFN-γ plays a critical role in promoting AS, and vaccine boosting provides a clinically translatable strategy to drive such responses against solid tumors.

Keywords: CAR T; IFN-γ; antigen loss; antigen spreading; antigenic heterogeneity; cancer; chimeric antigen receptor T cell; immunotherapy; solid tumor; tumor heterogeneity; vaccine.

Figure 1.. Vaccine boosting enables CAR T-cells to elicit endogenous T-cell responses in multiple tumor models.

(A) Schematic of CAR T-vax therapy. Created with BioRender.com. (B) IFN-γ ELISPOT. Mice bearing EGFRvIII⁺CT-2A tumors (n=5) treated with or without CAR T + various combinations of vaccine components. (C) Priming of endogenous CD8⁺ and CD4⁺ T-cells in EGFRvIII⁺CT-2A tumor-bearing mice (n=5–6) following CAR-T ± vax as measured by IFN-γ ELISPOT. (D-G) Mice (n=5–6) bearing OVA⁺ B16F10 tumors received FITC/TA99 CAR T-vax. (D) IFN-γ ELISPOT measuring OVA-specific endogenous T-cell responses. (F) IFN-γ ELISPOT measuring endogenous T-cell responses against Trp1^−/− B16F10 cells. (F-G)Tetramer-staining showing representative flow cytometry staining (F) and mean percentages of SIINFEKL tetramer⁺ endogenous T cells (G). (H) IFN-γ ELISPOT. Mice (n=4–5) bearing B16F10 tumors were treated with vax only, FITC-CAR T, or FITC/TA99 CAR T ± vax. Error bars show mean ± 95% CI. ***, p<0.0001; **, p<0.01; *, p<0.05; n.s., not significant by one-way ANOVA with Tukey’s post-test.

Figure 2.. Endogenous tumor-infiltrating T cells show transcriptional changes associated with enhanced anti-tumor activity in response to CAR T-vax therapy.

(A) Enumeration of intratumoral host T-cells in tumor-bearing mice (n=5) post CAR T ± vax treatment. (B-G) Tumor-bearing mice were treated with CAR T ± vax, TILs were isolated for scRNA-seq. (B) Experimental setup/timeline. Created with BioRender.com. (C) UMAP of endogenous T-cells obtained from tumors. (D) Curated clusters based on signature gene expression. (E) Stacked charts showing proportions of each T-cell cluster. (F) Dot plots showing differential expression of signature genes in endogenous CD8⁺ CTLs or CD4⁺ Th cells. (G) Cytotoxicity score of endogenous p15E-specific TILs and TILs of unknown specificity. All mice bear EGFRvIII⁺CT-2A tumors. Error bars are mean ± 95% CI, ****p<0.0001; **, p<0.01; n.s., not significant by Student’s t-test for A, by two-sided Wilcoxon rank-sum test for G.

Figure 3.. Vaccine-driven antigen spreading is required for long-tumor tumor control in immunocompetent mice.

(A-E) Treatment of tumor-bearing WT or Rag1^−/− mice with WT CAR-T ± vax. (A) Tumor growth in individual mice. Untreated, n = 5; CAR-T in WT mice, n = 10; CAR T-vax, n = 15 and 10 in WT and Rag1^−/− mice, respectively. (B) Percentage of mice that completely rejected tumors or experienced tumor relapse. (C) Overall survival. (D-E) Surface EGFRvIII expression (D) and mean expression normalized to untreated tumors (E) on parental or representative relapsed tumors from WT and Rag1^−/− mice following CAR T-vax treatment. (F-G) Tumor-bearing WT or CD8α^−/− mice (n=5–8) ± CAR T-vax treatment. (F) Tumor growth. (G) Overall survival. (H) Individual tumor growth and overall survival of WT (n=10) or Rag1^−/− mice (n=5) bearing heterogeneous CT-2A tumors upon CAR T-vax treatment. EGFRvIII⁺:EGFRvIII⁻ cells were pre-mixed at the indicated ratios. All mice in A-G bear EGFRvIII⁺CT-2A tumors. Error bars are mean ± 95% CI, ***, p<0.0001; **, p<0.01; *, p<0.05 by Student’s t-test for E, by Log-rank (Mantel-Cox) test for C,G-H, by two-way ANOVA with Tukey’s post-test for F.

Figure. 4.. Vaccine boosting induces cell-intrinsic enhancements in CAR T-cell function that include metabolic reprogramming.

(A) Tumor growth (left) and overall survival (right) of tumor-bearing mice (n=8) after receiving vaccine-boosted or non-boosted CAR T cells. Created with BioRender.com. (B-C) Tumor-bearing mice received WT CAR T ± vax treatment, and CAR T-cells were isolated from spleens and tumors for RNA-seq. (B) Volcano plot showing differential gene expression in splenic CAR T-cells. (C) GSEA showing enriched pathways in intratumoral CAR T-cells. (D-E) Intracellular PGC-1α expression (D) and mitochondrial mass (E) in intratumoral CAR T cells from mice (n=5) 7 days post treatment with WT CAR T ± vax. (F) IFN-γ ELISPOT. Tumor-bearing mice (n=5) treated with WT CAR T ± vax or PGC-1α^−/− CAR T-vax. All mice bear EGFRvIII⁺CT-2A tumors. Error bars are mean ± 95% CI, **, p<0.01; *, p<0.05 by Student’s t-test for D-E, and one-way ANOVA with Tukey’s post-test for F.

Figure. 5.. Enhanced IFN-γ production by vaccine-boosted CAR T-cells is critical for antigen spreading.

(A) IFN-γ and TNF-α expression in intratumoral CAR T-cells from mice (n=5) treated with WT CAR T ± vax. (B) IFN-γ expression in intratumoral CAR T-cells from mice (n=5) 7 days post treatment with WT or PGC-1α^−/− CAR T-vax. (C) IFN-γ ELISPOT. Tumor-bearing mice (n=5) treated with WT CAR T or WT CAR T-vax + isotype control antibody (IgG), anti-TNF-α or anti-IFN-γ. (D-E) OVA⁺EGFRvIII⁺CT-2A tumor-bearing mice(n=5–10) treated by WT CAR T or WT CAR T-vax ± anti-IFN-γ. Endogenous OVA-specific T-cell responses detected by SIINFEKL-tetramer staining (D) and IFN-γ ELISPOT (E). (F) IFN-γ ELISPOT. Tumor-bearing mice (n=5) treated with WT CAR T-vax ± anti-IFN-γ at indicated time points. (G-H) Tumor growth (G) and overall survival (H) of mice left untreated (n=5) or treated (n=10) with WT CAR T-vax ± anti-IFN-γ. (I) IFN-γ ELISPOT. Tumor-bearing mice (n=5) treated with WT or IFN-γ^−/− CAR T ± vax. (J) Tumor growth in mice (n=5) left untreated or treated with WT or IFN-γ^−/− CAR T-vax. (K-L) Tumor growth (K) and overall survival (L) of WT or IFN-γ^−/− mice (n=5–8) treated with or without WT CAR T-vax therapy. All mice bear EGFRvIII⁺CT-2A tumors. Error bars are mean ± 95% CI, ***, p<0.0001; **p<0.01; *, p<0.05, ns, not significant by Student’s t-test for A-B, by one-way ANOVA with Tukey’s post-test for C-F, I, by two-way ANOVA with Tukey’s post-test for J-K, and by Log-rank (Mantel-Cox) test for H and L.

Figure. 6.. DCs regulate CAR T-cell-induced antigen spreading through enhanced tumor antigen acquisition and IFN-γ-IL-12 crosstalk.

(A) EGFRvIII⁺ CT-2A cell killing by WT, IFN-γ^−/−, or IFNGR1^−/− CAR T-cells in vitro (n=3). (B) Enumeration of tumor-infiltrating immune cells in mice (n = 4–5) receiving WT CAR T ± vax. See supplemental methods for phenotyping details. (C) IFN-γ ELISPOT. Tumor-bearing WT or Batf3^−/− mice (n=5) treated with WT CAR T ± vax. (D) Tumor-bearing mice were left untreated (n=4) or treated with WT or IFN-γ^−/− CAR T-vax (n=5). Shown are chemokine expression in tumors 7 days post treatment. (E-F) Ki67 (E) and CCR7(F) expression in intratumoral CD103⁺ DCs and CD11b⁺ DCs from mice (n=5) treated with WT CAR T ± vax. (G-H) Mice bearing ZsGreen⁺EGFRvIII⁺CT-2A tumors were treated with WT CAR T, WT CAR T-vax or IFN-γ^−/− CAR T-vax (n=5), shown are tumor antigen (ZsGreen) uptake by intratumoral CD103⁺ DCs (G) and CD11b⁺ DCs (H). (I-J) IFN-γ ELISPOT (I) and tumor growth (J) in mice (n = 5) treated with WT or IFNGR1^−/− CAR T ± vax. (K-N) IFN-γ ELISPOT. (K) WT vs. CD11c-specific IFNGR1 KO tumor-bearing mice (n=5) following WT CAR T ± vax. (L) Tumor-bearing mice (n=5) following WT CAR T or WT CAR T-vax + anti-IFN-γ or anti-IL12(p70). (M) Tumor-bearing WT mice (n=5) following WT or IL12rb2^−/− CAR T-vax therapy or in IL12p40^−/− mice following WT CAR T-vax. (N) Tumor-bearing WT mice (n=7) following WT CAR T-vax or IFNGR1^−/− CAR T ± vax. (O) Tumor growth in mice (n=5–7) left untreated or treated with WT or IFNGR1^−/− CAR T ± vax. All mice except those in G-H bear EGFRvIII⁺CT-2A tumors. Error bars are mean ± 95% CI, ***, p<0.001; **, p<0.01; *, p<0.05; ns, not significant by Student’s t-test for B and E-F, by one-way ANOVA with Tukey’s post-test for A, C-D, G-I and K-N, by two-way ANOVA with Tukey’s post-test for J and O.

Figure. 7.. Engineering CAR T-cells for increased IFN-γ expression synergizes with vaccine boosting to enhance antigen spreading and rejection of solid tumors with pre-existing antigen heterogeneity.

(A-D). Heterogenous CT-2A tumors were established in C57BL/6 mice. (A)Tumor growth and (B)survival of mice (n=10) after treatment with WT CAR T-vax therapy ± anti-IFN-γ. (C)Tumor growth and (D)survival of mice (n=8) left untreated, receiving WT CAR T, or WT CAR T-vax ± anti-IL12 (p75). (E) IFN-γ secretion from WT or NFAT-IFN-γ CAR T-cells ± anti-CD3/CD28 beads (n=3). (F) IFN-γ ELISPOT. EGFRvIII⁺ CT-2A tumor-bearing mice (n=6) treated with WT or NFAT-IFN-γ CAR T ± vax. (G-L) Mice bearing heterogenous CT-2A tumors (n=5) treated with WT or NFAT-IFN-γ CAR T ± vax therapy. Enumeration of CAR T (G) and endogenous CD8⁺ T cells (J) infiltrated into tumors as well as the expression of IFN-γ (H for CAR T, K for host CD8 T) and granzyme B (I for CAR T, L for host CD8 T). (M-N) Tumor growth (M) and overall survival (N) of mice bearing heterogenous CT-2A tumors (n=10) treated with WT or NFAT-IFN-γ CAR T ± vax. (O) Schematic overview of CAR T-vax therapy triggered antigen spreading. Created with BioRender.com. Heterogenous CT-2A tumors are EGFRvIII⁺:EGFRvIII⁻ cells mixed at 80:20 ratio. Error bars are mean ± 95% CI. ***, p<0.001; **, p<0.01; *, p<0.05; ns not significant by one-way ANOVA with Tukey’s post-test for E-L, by two-way ANOVA with Tukey’s post-test for C, and Log-rank (Mantel-Cox) test for B, D and N.