Genetic Disruption of Blimp-1 Drastically Augments the Persistence and Anti-Tumour Efficacy of BCMA-Targeting CAR-T Cells

Abstract

Background: Chimeric antigen receptor-T (CAR-T) cells that target B cell maturation antigen (BCMA) have emerged as an effective treatment option for multiple myeloma (MM). However, many patients eventually relapse, signifying a need to improve the function and persistence of this immunotherapy. A major limitation to the therapeutic potential of all CAR-T cells is premature terminal differentiation into effector cells, during both the ex vivo and in vivo expansion stages. Instead, it would be much more favourable for CAR-T cells to develop into long-lived memory cells which are able to proliferate and maintain anti-tumour responses. Different approaches to retain a memory-like phenotype in CAR-T cells, such as selecting particular memory or naïve T-cell subsets as a starting material and altering ex vivo culture conditions, have had varying degrees of success. An alternative strategy to selectively reprogram CAR-T cells is to use gene editing approaches such as CRISPR/Cas9, which have already proven effective in the ablation of inhibitory receptors and generation of allogenic CAR-T cells. Here, we took advantage of this technology to dampen the expression of the Prdm1 gene which encodes Blimp-1, a transcription factor that promotes terminal effector cell differentiation in T cells. Methods: Stimulated T cells transduced with ARI2h, an in-house anti-BCMA CAR that is currently being explored in a clinical trial for relapsed/refractory MM (NCT04309981), were electroporated with either Cas9 alone (WT) or Cas9 + Prdm1-targeting sgRNA (Blimp-1 KO). Blimp-1 gene disruption was assessed by DNA sequencing of the sgRNA target regions. CAR-T cells were analysed by flow cytometry for cell surface marker and effector molecule expression. Anti-tumour responses were measured following in vitro challenges with MM cell lines (ARP-1 or U266) and assessed by injecting CAR-T cells into ARP-1 tumour-bearing mice. Results: CRISPR/Cas9-mediated disruption of the Prdm1 gene to generate Blimp-1 KO CAR-T cells was highly effective (average editing efficiency >80%) and did not influence CAR transduction, cell expansion or the expression of immune checkpoint molecules. Furthermore, Blimp-1 KO CAR-T cells effectively eliminated MM cells following multiple in vitro challenges, although they were slower than WT CAR-T cells. When examining the in vivo function of Blimp-1 KO CAR-T cells in a murine xenograft model of advanced MM, both WT and Blimp-1 KO CAR-T cells initially reduced tumour burden. However, Blimp-1 KO CAR-T cells dramatically slowed disease progression and even completely cleared the tumours in some mice, whereas WT CAR-T cells were unable to prevent rapid relapse (Figure 1A). Correspondingly, mice administered with Blimp-1 KO CAR-T cells had an increased survival time compared to those given WT CAR-T cells (p=0.006) (Figure 1B). As Blimp-1 acts as a master switch to control T-cell memory/effector differentiation, the phenotype of ex vivo Blimp-1 KO CAR-T cells was characterised. Strikingly, Blimp-1 KO CAR-T cells had increased expression of memory cell-associated markers, including CCR7 and CD127 (IL-7 receptor), and diminished levels of the effector molecule granzyme B. Furthermore, following short-term MM cell exposure, Blimp-1 KO CAR-T cells had a limited capacity to upregulate granzyme B, but their ability to produce the cytokines IFNγ, TNFα and IL-2 was uninhibited. To determine whether loss of Blimp-1 influenced CAR-T cell phenotype after long-term exposure to antigen, Blimp-1 KO CAR-T cells were subject to repeated in vitro challenges with MM cells. Interestingly, tumour-induced production of granzyme B was no longer dampened in repeatedly-stimulated Blimp-1 KO CAR-T cells compared to WT CAR-T cells, and in fact, TNFα production was enhanced. Nonetheless, Blimp-1 KO CAR-T cells maintained higher expression of memory cell markers, suggesting that they retained a memory-like phenotype. Conclusions: Genetic disruption of Blimp-1 caused anti-BCMA CAR-T cells to maintain a memory-like phenotype that retained effector function, even following repeated challenges with tumour cells. As a result, Blimp-1 KO CAR-T cells were considerably more effective than WT CAR-T cells at treating advanced MM in a murine model. Overall, we conclude that using gene editing to restrict Blimp-1 expression represents a promising strategy to enhance the sustained efficacy of anti-BCMA CAR-T cells. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal