Background: Chimeric Antigen Receptor (CAR) T-cell therapy represents a paradigm shift in therapy of advanced haematological malignancies in particular Acute Lymphoblastic Leukaemia. We have engineered multi-antigen specific CAR T-cells targeting three acute myeloid leukaemia (AML) associated antigens through incorporation of nanobody binding domains. Target antigens were selected for their broad expression on both on AML blasts as well as a high expression on Leukaemic Stem Cells (LSC). LSCs are a primitive self-renewing population that can contribute to relapse and may be phenotypically distinct from bulk AML blasts within the same patient.
Aims: Our multi-antigenic approach thus aims to overcome antigen negative escape, overcome intra-tumour heterogeneity, and clear both LSCs and bulk disease by designing a tandem CAR capable of sensing CD33, CD123 and CLL1 on target cells.
Methods: We initially screened libraries of nanobody binders for each antigen in a second generation CAR format. Once lead binders for each antigen were identified, Tandem CAR (TanCAR) structures were designed taking into consideration data on antigen structure from structural prediction servers and published crystal structures, nanobody domain mapping and antigen expression profiles. A TanCAR was then rationally designed comprising of three nanobodies joined by G4S linkers, fused to a hinge spacer, CD8 transmembrane domains, and 41BBz signalling domains. TanCAR T-cell function was compared to that of CAR T-cells targeting single antigens through in vitro co-culture assays with engineered target cells expressing single antigens (n=3). Antigen-specific cytotoxicity and proliferation were measured by flow cytometry assays; IL-2 and IFNy production was assayed by ELISA.
Results: The ability of nanobody binding domains to bind individual antigens within the TanCAR format were confirmed using soluble antigen formats with a murine Fc tags via flow cytometry (Fig 1). In co-culture assays against engineered cell lines expressing single antigens (E:T=1:8, 24hrs) the TanCAR demonstrated antigen specific cytolytic activity (median cytotoxicity = 63%, 59%, and 47% against SupT1-CD33, SupT1-CD123, and SupT1-CLL1 respectively) approximately equal to or greater than single antigen targeting nanobody CAR T-cells (Fig 2a). In further co-culture experiments TanCAR T-cells produced high levels of IL-2 and IFNy when co cultured with cell lines expressing CD33 or CD123 only (n=3, median IL-2 = 9237pg/ml and 8190pg/ml, median IFNy 3924pg/ml and 4842pg/ml respectively) compared to SupT1 NT target cells, whereas low level cytokine production was observed with cell lines that express CLL-1 only (Fig 2b). Proliferation data matched this trend with greater levels of antigen specific proliferation observed when TanCAR T-cells were co-cultured with cells expressing CD33 and CD123 (7day co-culture, E:T=1:10, median fold expansion = 22, 8, and 5 fold for TanCAR co-cultured with cell lines expressing CD33, CD123, and CLL1 respectively), compared to antigen negative SupT1 target cells (Fig 2c).
Summary/Conclusion: Collectively, our data demonstrate that TanCAR T-cells operate as an OR gate mediating potent reactivity against target cell lines expressing single AML antigens in vitro. We now aim to test for optimal TanCAR formats and potential additive or synergistic responses where TanCAR T cells are co-cultured with AML cell lines in-vivo (Molm14, PDX).
Ghorashian:Novartis: Honoraria; UCLB: Patents & Royalties; Amgen: Honoraria. Pule:Mana Therapeutics: Other: entitled to share of revenue from patents filed by UCL; UCLB: Patents & Royalties; Autolus: Current Employment, Other: owns stock in and receives royalties, Patents & Royalties.
