Introduction: CD28 is a costimulatory protein incorporated into two FDA-approved second-generation CD19-directed CAR T cell products and utilized extensively preclinically. Within CAR T cells, CD28 induces complex signaling including activation of the PI3K/AKT/mTOR pathway, with ingrained signaling redundancies. Although PI3K signaling enhances proliferation, excessive growth signals may lead to T cell exhaustion and terminal differentiation, correlating with suboptimal clinical responses. Importantly, PI3K and AKT inhibitors have been used to modulate CAR T cell signaling, enriching memory-like and reducing highly differentiated T cells, leading to a more potent cellular product. We aimed to attenuate CD28 signaling redundancies through the development of a “synthetic rheostat” that tempers PI3K/AKT/mTOR pathway signaling.

Methods:We mutated CD28 to reduce PI3K signaling and assessed the function of this mutant (referred to as SAVVY) compared to a native CD28 in second-generation CD19-targeting CAR T cells against NALM6, a B cell acute lymphoblastic leukemia (B-ALL) cell line. CD19-targeting CAR T cells were produced by gammaretroviral transduction of peripheral blood-derived primary T cells. In vitro, we evaluated CAR T cell proliferation, target lysis, CAR recycling, T cell mitochondrial mass/function, T cell immunophenotype (memory/exhaustion), and T cell cytokine secretion. In vivo, we evaluated tumor control and overall survival in cell-line derived xenograft murine models. In vitro and in vivo, models of normal antigen density, low antigen expression, and PD-L1 overexpression were evaluated given the clinical relevance and prevalence of antigen-positive relapses, as well as CD19 antigen-low/negative relapses post-treatment with FDA-approved CAR T cell therapies.

Results:Although CD19-targetingSAVVY-CAR T cells demonstrated no significant difference in long-term anti-tumor efficacy compared to native CD28-CAR T cells against NALM6 (effector:target (E:T) ranging from 1:5 to 1:40), they exhibited superior proliferation in vitro. When cocultured long-term in vitro with low antigen density tumor (NALM6-CD19low) and PD-L1-overexpressing tumor (NALM6-PD-L1+), SAVVY-CAR T cells demonstrated comparable cytotoxicity and proliferation at an E:T of 1:5 but superior tumor control at a lower E:T of 1:40. The SAVVY modification did not perturb blastogenesis or expression of early activation markers (CD69), but it protracted late activation marker expression (CD25, CD71, HLA-DR). SAVVY-CAR T cell activation protraction was not dependent on CAR recycling, mitochondrial mass, or mitochondrial transmembrane potential. SAVVY-CAR T cells demonstrated lower expression of co-inhibitory markers on excessive antigen encounter, and there was no significant difference in cytokine secretion (interferon-gamma, IL-2, TNF-a, granzyme B) when cocultured with NALM6 or NALM6-CD19low. CD8+ SAVVY-CAR T cells retained a memory immunophenotype (CD62L+) after antigen encounter relative to native CD8+ CD28-CAR T cells. SAVVY-CAR T cells showed superiority in vivo compared to native CD28-CD19 CAR T cells in xenograft models utilizing disseminated NALM6, NALM6-CD19low, and NALM6-PD-L1+, with mice exhibiting significantly prolonged survival in all three models.

Conclusions:The SAVVY modification imparts enhanced functionality to second-generation CAR T cells evident in vitro and in vivo, leading to significantly improved tumor control in an aggressive disseminated B-ALL model under normal, low antigen-density, and PD-L1-high conditions. Recent findings have shown that CD28 is more active in low antigen density disease compared to 4-1BB, raising concerns about attenuating the former's signaling. However, SAVVY-CAR T cells demonstrate enhanced anti-tumor effects compared to native CD28-CAR T cells in a low antigen density model, which could potentially lead to longer-lasting remissions and reduced antigen-low relapses. Given these robust findings, SAVVY has been incorporated into a CD371 (CLL-1, CLEC12A)-directed CAR T cell product with constitutive secretion of interleukin-18, which is currently being investigated in a phase 1 trial for relapsed/refractory AML patients (CLEAR-AML; NCT06017258). Furthermore, the impact of SAVVY is currently being evaluated in other CAR constructs targeting various hematologic and solid tumor malignancies.

Disclosures

Brentjens:Triumvira, Cargo Tx, CoImmune: Membership on an entity's Board of Directors or advisory committees; BMS: Research Funding; BMS, Caribou, Sanofi: Other: licensed intellectual property to and collects royalties from; BMS, Atara Biotherapeutics Inc. and was a consultant for Triumvira, Cargo Tx, CoImmune: Consultancy. Daniyan:Tigen Pharma SA: Patents & Royalties: Intellectual Property Rights; Promicell Therapeutics, Inc.: Consultancy, Current holder of stock options in a privately-held company; Shoreline Biosciences, Inc.: Consultancy; Caribou Biosciences, Inc: Patents & Royalties: Intellectual Property Rights.

This content is only available as a PDF.
Sign in via your Institution