Vaccine-mediated co-engagement of CAR T and endogenous T-cells eradicates tumors with pre-existing antigenic heterogeneity

Introduction::
Chimeric antigen receptor (CAR) T-cell therapy in solid tumors has been largely disappointing. A key obstacle is the pre-existing antigenic heterogeneity in solid tumors, where not all tumor cells express the antigen targeted by the CAR. Strategies to prevent tumor escape caused by antigenic heterogeneity are of paramount importance to achieve the full promise of CAR T therapy. Antigen spreading (AS) is the induction and amplification of immune responses to secondary antigens distinct from the original therapeutic target. AS can play an important role in the expansion and broadening of the peripheral T cell repertoire that mediates the rejection of antigenically heterogeneous tumors. However, to date there is limited clinical evidence of CAR T therapy itself inducing therapeutically meaningful AS, and mechanisms by which AS could be promoted during CAR T therapy remain poorly understood. We recently developed a synthetic vaccine (amph-vax) to enhance CAR T cell activity against solid tumors (Ma et al, Science 2019, Ma et al, BioRxiv, 2023). Amph-vax boosting of CAR T cells (CAR T-vax) substantially enhanced tumor rejection by CAR T cell therapy in multiple solid tumor models (Figure 1A) and, unexpectedly, was accompanied by potent AS that couldn’t be achieved by CAR T therapy alone. These results highlight the underappreciated impact of vaccine boosting in broadening anti-tumor immunity, and exploiting AS as a target for therapeutic intervention has the potential for enhancing CAR T therapy against tumors with pre-existing antigenic heterogeneity.




Materials and Methods::
We used a novel amphiphilic vaccine (amph-vax) to specifically stimulate and enhance CAR T therapy directly through the CAR in vivo. The amph-vax molecules were generated by linking a CAR-ligand (i.e., pepvIII in this proposal as the ligand for the EGFRvIII CAR) to albumin-binding poly (ethylene glycol)-phospholipid (PEG-DSPE) conjugates. These lipid-polymer conjugates preferably accumulate in the lymph nodes (LNs) and insert in the antigen-presenting cell (APC) membranes on arrival in LN allowing APCs to stimulate CAR T cells in tandem with natural costimulation and cytokine support, enhancing CAR T expansion and functionality. CAR T cells were produced by retrovirally transducing an EGFRvIII-targeting CAR or into primary mouse T cells ex vivo. CAR T cells were infused into recipient mice intravenously. The vaccine was formulated with 10μg peptide equivalent per mouse and 25μg cyclic-di-GMP (STING agonist) as the adjuvant and injected subcutaneously at the tail base of animals. All the animal experiments in this study were carried out in a CT-2A murine glioblastoma model expressing a tumor-specific surface antigen EGFRvIII.




Results, Conclusions, and Discussions::
In the CT-2A murine glioblastoma models, amph-vax boosted CAR T therapy (termed “CAR T-vax therapy”), but not CAR T therapy alone, elicited pronounced antigen spreading (AS) in both CD4+ and CD8+ T-cell compartments (Figure 1B). Single-cell RNA-seq of 21,835 tumor infiltrating-endogenous T cells confirmed a significant increase of cytotoxic CD8 T cells and induction of Th1 CD4 T cells. Transcriptomic analysis demonstrated that vaccine boosting significantly enhanced CAR T cell metabolism, including oxidative phosphorylation (OXPHOS). AS was reduced by ~50% using PGC-1α-deficient CAR T-vax therapy. IFNγ blockade using an anti-IFNγ antibody (Figure 1C-D) or IFNγ-knockout CAR T abolished AS. Blockade of IL12 pathway using anti-IL12 (p75) antibody or IL12rb2 knockout mice phenocopied IFNγ blockade, resulting in negligible AS (Figure 1E). We further observed a significant reduction of AS in BatF3-deficient mice or mice with deficient IFNγ signaling in CD11c+ DCs (Figure 1F). Finally, we showed that IFNGR1- or IL12rb2-deficient CAR T-vax therapy failed to induce AS. Using a heterogeneous tumor model with EGFRvIII+ and EGFRvIII- CT-2A cells mixed at pre-defined ratios, we found that CAR-T vax therapy cured ~50% animals bearing tumors with up to 20% EGFRvIII^- cells, and further elevating CAR T-intrinsic IFNγ expression increased the cure rate to ~80% (Figure 1G). And the therapeutic response was completely lost in Rag1-/- mice which have endogenous T cells and therefore no antigen spreading (Figure 1G).









In conclusion, we demonstrated vaccine boosting of CAR T therapy as a clinically-translatable strategy to promote control of antigenically heterogeneous tumors via antigen spreading. Mechanistically, we showed that IFN-γproduction by CAR T-cells and sensing of IFN-γ by both host cells and the CAR T-cells are critical to enabling AS (Figure 1H). CAR T-derived IFN-γ were propagated via a positive feedback regulation through crosstalk with DC-derived IL-12 (Figure 1H). Engineering CAR T cells to forcibly express IFN-γ in response to antigen receptor stimulation enhances AS, thereby enhancing control of antigenically heterogeneous tumors. As few solid tumors express target antigens on >90% of tumor cells, the findings described here provide guidance for engineering CAR T therapies to more effectively treat solid tumors with pre-existing antigenic heterogeneity.




Acknowledgements (Optional): : We thank the Koch Institute Swanson Biotechnology Center for technical support, specifically the flow cytometry core facility.D.J.I was supported by the NIH (award EB022433), the Marble Center for Nanomedicine, and the Mark Foundation, and L.M was supported by an American Cancer Society postdoctoral fellowship, the Cell and Gene Therapy Collaborative at CHOP, and the W.W. Smith Charitable Trust.

References (Optional): :