Ligand-restricted SynNotch switches enable precision cell therapy

Abstract

Wendell Lim’s group and colleagues demonstrate that synNotch transcriptional circuits engineered into T cells can be used to precisely control location-specific expression of payloads responding to antigen triggers, thus locally inhibiting unwanted immunity or neuroinflammation. With no off-tumor toxicity or systemic immunosuppression upon elimination of mouse brain tumors, this approach can achieve better efficacy than anticipated.

MAIN TEXT

Notch receptors are a known family of transmembrane proteins that undergo proteolytic cleavage upon ligand binding, leading to the release of their intracellular transcription factors (TFs) and downstream gene expression. The intrinsic modularity of Notch receptors has been leveraged to develop the synthetic Notch (synNotch) receptor system, where the extracellular domain is replaced with a customizable antigen recognition domain and the intracellular domain is a user-defined TF, thus enabling customized sensing/response behaviors of engineered cells and delivery of customized immunomodulating payloads during chimeric antigen receptor (CAR) T cell therapy (1–4). CAR T cell therapy has demonstrated remarkable success in B cell malignancies (5), but faces challenges in solid tumors and autoimmune disorders, such as on-target, off-tumor toxicity due to a lack of tumor-specific antigens (6). Moreover, the low peripheral Treg frequency hampers the manufacturing and stability of engineered Tregs for autoimmune treatments (7). To address these challenges, Lim’s research group and colleagues optimized antigen-specific synNotch circuits to convert conventional CD4⁺ T cells into immunosuppressive T cells, reducing unwanted immune responses in targeted areas (8). Using a brain-specific receptor, they showed that synNotch circuits can (i) restrict CAR expression within engineered T cells in the brain of NCG and NSG mice, therefore avoiding off-tumor toxicity when treating primary glioblastoma (GBM) and metastatic breast cancer, respectively, or (ii) inhibit neuroinflammation by locally producing ther immunosuppressive cytokine IL-10 (9). Here, we discuss these two elegant studies that establish novel platforms for location-specific precision delivery of biological payloads.

In the first study, Reddy et al. designed an anti-CD19 synNotch circuit expressing immunoregulatory payloads to convert conventional CD4⁺ T cells into immunosuppressive T cells upon CD19 recognition (8). Screening various payloads, they found that co-expression of TGF-β1 and IL-2Rα (CD25) was the most effective combination in suppressing CD4⁺ and CD8⁺ CAR T cells (8) (Figure 1A, Left). TGF-β1 is a potent immunosuppressive cytokine, while CD25 sequesters IL-2 via competitive binding (10). These TGF-β1/CD25 payload-expressing anti-CD19 synNotch suppressor cells (hereafter suppressor T cells) exhibited comparable inhibitory efficacy in vitro to polyclonal FoxP3⁺ Tregs but outperformed (suppression of anti-tumor CAR activity) anti-CD19 FoxP3⁺ CAR Tregs as well as anti-tumor CAR T cells that were equipped with PD-1-based inhibitory CARs (8). Their suppressive activity was evaluated in vivo in an immunocompromised NSG mouse model. Briefly, dual-antigen (Her2⁺CD19⁺) and single-antigen (Her2⁺) K562 tumors were engrafted in the right and left flanks of mice, respectively. Seven days later, anti-Her2 CAR T cells and suppressor T cells were administered intravenously (i.v.). The authors showed that dual-antigen tumor cells were spared, while single-antigen tumor cells were effectively cleared by CAR T cells (8) (Figure 1A, Right). This suggested that the suppressor T cell inhibitory activity against CAR T cells that was triggered by CD19 on dual-antigen tumor cells, was only effective in the immediate vicinity of these tumor cells. Therefore, these observations highlight the precise, location-restricted control of CAR T cytotoxicity by anti-CD19 synNotch suppressor T cells.

Figure 1. SynNotch transcriptional circuits can be engineered into T cells to enable precision cell therapy.

A) An anti-CD19 synNotch circuit converts conventional CD4⁺ T cells to suppressor T cells upon CD19 recognition, inhibiting CAR T-mediated killing of CD19-expressing target cells. Left: Upon anti-CD19 synNotch receptor binding to CD19, the transmembrane proteolytic site of the receptor is cleaved, releasing the intracellular region that initiates the expression of immunosuppressive cytokine TGF-β1 and the IL-2 receptor α chain, CD25; this effectively convert conventional T cells to suppressor T cells. Right: Suppressor T cells are activated in the presence of CD19-expressing cells, inhibiting CAR T cell cytotoxicity (8). B) Anti-BCAN synNotch-controlled T cell activation enables tissue-specific targeting in the mouse brain. Left: Biological payloads such as CAR and IL-10, are only expressed upon synNotch binding to BCAN in the brain’s extracellular matrix. Middle: The synNotch circuit is activated by BCAN in the brain to express CARs targeting tumor antigens that are specific to neoplastic cells, relative to normal brain tissue. Tumors are effectively cleared in NCG mice without damaging normal neurons. Right: The synNotch circuit is activated by BCAN in the brain to express anti-inflammatory cytokine IL-10, which reduces the proinflammatory capacity of T cells and microglia in neuroinflammation (9). Figure created with BioRender (Biorender.com).

Furthermore, Reddy et al. investigated local immunosuppression using anti-CD19 synNotch suppressor T cells to prevent host immune rejection in allogeneic organ transplantation models (8). Enriched β cell (eBC) organoids were differentiated from HLA-A2⁺ human pluripotent stem cells (hPSCs) and engineered to express CD19. Co-culturing anti-HLA-A2 CAR T cells with eBC organoids led to CAR T cell-mediated killing of the organoids. However, addition of suppressor T cells prevented this killing. In vivo, HLA-A2⁺CD19⁺ eBC organoids transplanted under the kidney capsule of immunocompromised NSG mice were rapidly rejected upon infusion of anti-HLA-A2 CAR T cells, mimicking graft rejection. Notably, suppressor T cells effectively protected the graft, reduced CAR T cell expansion, and caused no detectable systematic immunosuppression (Figure 1A, Right). In contrast, HLA-A2⁺CD19⁻ eBC organoids were not protected under similar conditions, indicating that this protection was directly mediated by the CD19-recognizing SynNotch circuit (8). These findings confirmed the aforementioned observations with the dual-antigen tumor cell system that showed that synNotch circuits can reprogram conventional CD4⁺ T cells into suppressor T cells upon encountering synNotch ligand-expressing cells (8).

In the second study, Simic et al. developed synthetic T cells whose synNotch circuits could be selectively activated in the brain (9). To achieve this, the researchers engineered T cells to express a synNotch receptor that specifically recognized brevican (BCAN), a protein uniquely and uniformly presented in the extracellular matrix of the brain. This design enabled the targeted expression of therapeutic payloads such as CAR or IL-10 within the brain tissue of NCG and NSG mice (Figure 1B, Left). To target primary GBM tumors, CAR T cells were engineered to express anti-EphA2/anti-IL13Rα2 CARs under the control of anti-BCAN synNotch switch. EphA2 (ephrin type A receptor 2) and IL13Rα2 (interleukin-13 receptor α2) are attractive therapeutic targets because of their common overexpression in neoplastic GBM cells and their absence in healthy brain tissues. However, EphA2 and IL13Rα2 are also expressed in some non-brain normal tissues, posing a risk of on-target, off-tumor toxicity. By using the anti-BCAN synNotch switch and anti-EphA2/anti-IL13Rα2 CARs as payloads, i.v.-infused engineered T cells cleared GBM patient-derived xenograft tumors (GBM6) inoculated in the brain of NCG mice while sparing BCAN knockout GBM6 cells that were engrafted in the flank (9). This precise local expression of BCAN-induced CARs minimized off-tumor and systemic effects, as evidenced by flow cytometric analysis (9). The authors also extended their strategy to target secondary brain tumors, using anti-BCAN synNotch to activate anti-HER2 CAR for clearing HER2⁺ breast cancer and anti-TROP2 CAR for eliminating triple negative breast cancer metastases in NSG mice (9) (Figure 1B, Middle).

The scope was then expanded to treat neuroinflammatory diseases. Experimental autoimmune encephalomyelitis (EAE) is a widely used mouse model for multiple sclerosis that highlights the role of T cells and microglia in neuroinflammation. IL-10 is a potent anti-inflammatory cytokine, but it shows no efficacy against neuroinflammation when administered i.v., likely due to its short half-life in serum and inability to cross the blood-brain barrier (BBB) (11). Thus, Simic and coworkers engineered CD4⁺ T cells to express IL-10 under the control of the anti-BCAN synNotch receptor. Upon BCAN-mediated activation in vitro, these engineered CD4⁺ T cells reduced TNF-α and IL-6 secretion by microglia and IFN-γ secretion by myelin oligodendrocyte glycoprotein (MOG)-specific T cells, therefore reducing the proinflammatory capacities of the cells; in vivo, the engineered T cells ameliorated EAE symptoms compared to controls (9) (Figure 1B, Right). Consequently, the authors successfully used anti-BCAN synNotch circuits to precisely deliver therapeutic payloads to the brain, targeting residing tumors or modulating local immune responses (9).

The ability to control the release of payloads in response to specific molecular cues makes synNotch circuits a powerful tool in precision medicine (12). By coupling antigen-specific synNotch receptors with various therapeutic payloads, Reddy et al. enabled localized inhibition of unwanted immune responses, avoiding systemic immunosuppression (8). In turn, Simic et al. demonstrated that engineered T cells can serve as cargo, crossing the highly selective BBB, and precisely delivering therapeutic cells or proteins into the brain (9). Furthermore, this anti-BCAN synNotch circuit platform could, in principle, be extended to achieve precision therapeutic delivery to other tissues. These two groundbreaking studies hold promise for treating complex diseases, including cancer, autoimmune disorders, graft-versus-host disease, and neuroinflammation, with ideally greater specificity and fewer side effects, certainly meriting further investigation.