Folate receptor beta targeting of tumor-associated macrophages improves lymphoma cell responses to CD19 CAR T cells

Abstract

Abstract Chimeric antigen receptor (CAR) T cell therapies have revolutionized the treatment of patients with relapsed and refractory (R/R) diffuse large B-cell lymphoma (DLBCL). However, approximately half of the patients experience lymphoma progression after CD19 CAR T cell therapy. Immunosuppressive tumor microenvironment (TME), and M2-like tumor-associated macrophages (TAMs) in particular, can facilitate resistance to CAR T cell therapy. To determine whether CAR T cell responses in DLBCL can be improved by reprogramming the immunosuppressive TME, we set out to identify M2-TAM-specific biomarkers within DLBCL tumor tissue. First, we performed CIBERSORTx deconvolution analysis on eight cohorts of primary (n=3234) and a cohort of R/R (n=170) DLBCL patients treated with immunochemotherapy and CAR T cells, respectively. Each cohort showed heterogeneous TAM proportions among patients, and a high M2 TAM proportion was associated with poor OS in six of the eight primary and in the R/R DLBCL cohorts. We then identified folate receptor beta (FRβ), encoded by FOLR2, as a specific M2-like TAM cell surface marker and a predictor of poor survival across five cohorts. The poor survival association was even more evident when assessing the high co-expression of FOLR2 and the common M2 macrophage marker CD163, and it was further validated at the protein level on a tissue microarray consisting of 117 DLBCL samples (5-year OS, 68% vs 85%, p = 0,0097). We next investigated the impact of macrophages on DLBCL and CD19 CAR T cells using in vitro differentiated primary M2 and M1-like macrophages, as well as a lymphoma macrophage co-culture spheroid model. Three DLBCL cell lines (OCI-ly7, suDHL10, and suDHL5) were transduced to express a luciferase reporter, enabling us to measure lymphoma cell viability under co-culture conditions. This revealed reciprocal interactions between lymphoma cells and M2-like macrophages. Lymphoma cells promoted the differentiation of monocytes into M2-like macrophages, as shown by confocal imaging of the co-culture spheroids displaying M2-like cell morphology and CD163 positivity. M2-like macrophages, in turn, fostered lymphoma cell growth and impaired CD19 CAR T cell-mediated killing of lymphoma cells. Lastly, we generated FRβ-targeting CAR T cells, which were first tested for their ability to specifically target M2-like macrophages. When co-administered with CD19 CAR T cells, the FRβ CAR T cells improved lymphoma cell killing of the M2-like co-cultured spheroids, actively counteracting the immunosuppressive effect of the M2-like macrophages. Taken together, the results indicate active crosstalk between lymphoma cells and M2-like macrophages, as well as TAM-mediated resistance mechanisms against CD19 CAR T cells, which can be circumvented by FRβ-CAR T cells. Additionally, the findings suggest that cotargeting of FRβ+ macrophages and CD19+ lymphoma cells could be a potential way to improve the efficacy of CAR T cell therapy in DLBCL.