Correlation of Serum Inflammatory State and Macrophage Phenotype with CAR T Cell Expansion and Clinical Outcomes

Introduction: Chimeric antigen receptor (CAR) T-cell therapy has shown excellent responses in patients with relapsed refractory hematologic malignancies but in a subset of patients, CAR-T cells fail to prevent tumor progression. Tumor microenvironment and higher serum inflammatory proteins have been implicated in the poor response to CAR T cell therapy. High levels of monocytic myeloid-derived suppressor cells (M-MDSc) are associated with decreased expansion of CAR T-cells, however the mechanisms of their phenotypic transformation are not fully understood. Tumors are known to impact myelopoiesis via chemokines, as well as evade anti-tumor immune responses via tumor associated macrophages (TAMs), also known as M2 macrophages. It has been established that certain interleukins can direct the differentiation of macrophages towards immunosuppressive phenotypes. Our study aims to explore the interplay of inflammation induced macrophage phenotype with its impact on CAR T cell expansion and clinical outcomes.

Methods: We obtained pre-lymphodepletion serum samples of 30 patients treated with CAR-T cell therapy at Roswell Park Comprehensive Cancer Center. Monocytes and T-cells were isolated from single healthy commercial donor PBMCs. Transduction and spinoculation of T cells was done and high CAR transduction efficiency was achieved. Isolated monocytes were stimulated to M0 macrophages with M-CSF alone and M2 macrophages with M-CSF along with IL-4, IL-10 and IL-13. The CAR-T cells, Raji lymphoma cells and macrophages (1:1:0.5) were co-cultured with patient’s serum. Three groups, one without macrophages (No mac), one with M0 (undifferentiated phenotype) and one with M2 (suppressive phenotype) macrophages served as control. The expansion of CAR T-cells was determined by an IncuCyte S3 live cell analysis system (Essen Bioscience).

Results: Out of serum samples of 30 patients, 3 samples were excluded due to collection more than 10 days prior to receiving CAR-T cell infusion. Out of 27 samples, 18 were from patients with non-Hodgkin B-cell lymphomas (NHL) and 9 were from patients with multiple myeloma (MM). Due to the different disease biology of NHL and MM, only 18 samples from NHL patients were utilized for this analysis. Majority of the patients were male (89%, n=16) with median age of 66 years (range 37-79) at the time of CAR-T cell infusion.

CAR-T expansion index of 32-fold was determined in control samples with no macrophages (No mac) whereas the expansion index was 17 and 8-fold in control samples with M0 (undifferentiated phenotype) and M2 (suppressive phenotype) macrophages, respectively. Serum from 7 patients produced CAR-T expansion like M2 macrophage co-cultured control group. Contrary to that, serum from 11 patients yielded expansion of CAR-T cells similar or better than M0 macrophage co-cultured control group. Closer analysis revealed that serum from non-responder NHL patients produced CAR-T expansion (≤ 8 folds) similar to M2 co-cultured CAR-T cells whereas responder patients had CAR-T expansion index > than M2 co-cultured CAR-T cells.

Conclusion: Pre-lymphodepletion serum samples of patients without response to CAR-T cell therapy cause lower CAR-T expansion as seen in the presence of M2 macrophages. These results indicate the role of serum inflammatory proteins in modulating macrophages to M2 differentiation and suppression of CAR-T expansion, which is associated with suboptimal clinical responses. More studies are needed to understand the role of macrophages in efficacy of CAR-T cell function, and we are currently evaluating the cytokine profiles of the patients’ serum samples that induced M2-like macrophage phenotypes.