Background: Chimeric antigen receptor T (CAR-T) cell therapy is more effective in relapsed or refractory diffuse large B cell lymphoma (DLBCL) than other therapies, but a high proportion of patients relapse after CAR-T cell therapy owing to antigen escape, limited persistence of CAR-T cells, and immunosuppression in the tumor microenvironment. CAR-T cell exhaustion is a major cause of relapse. Epigenetic modifications can regulate T cell activation, maturation and depletion; they can be applied to reduce T cell depletion, improve infiltration, and promote memory phenotype formation to reduce relapse after CAR-T cell therapy. Purpose: We propose to develop and validate Methods: We designed an shRNA sequence targeting LSD1 mRNA, and created a vector with the following elements: the U6 promoter driving expression of the LSD1 shRNA sequence, the EF1a promoter driving a second-generation anti-CD19 CAR sequence encoding an anti-CD19 single-chain variable fragment (FMC63), the CD8 hinge and transmembrane structural domains, the CD28 co-stimulatory structural domain, and the CD3ζ-activating structural domain. The MFG-LSD1 shRNA anti-CD19 CAR plasmid was first constructed, then packaged in retroviral vectors and transduced into human primary peripheral blood mononuclear cell-derived T cells to generate the corresponding CAR-T cells. We examined by flow cytometry the efficiency of two CAR-T cells in killing U-2932 cells (a human DLBCL line) upon co-culture with RNAU6 anti-CD19 CAR-T cells or LSD1 shRNA anti-CD19 CAR-T cells. We analyzed Ki-67 staining of the CAR-T cells by flow cytometry on days 0, 5, and 10, and counted the cells to assess expansion. We also used flow cytometry to detect the central memory T cell (TCM) proportion. Results: We detected the expression of the CAR in the CAR-T cells by flow cytometry, and observed transduction rates of 31.5% for RNAU6 anti-CD19 CAR-T cells and 60.7% for LSD1 shRNA anti-CD19 CAR-T cells. The killing efficiency of LSD1 shRNA anti-CD19 CAR-T cells was significantly higher than that of RNAU6 anti-CD19 CAR-T cells at the low effector target ratio. We further found that LSD1 shRNA anti-CD19 CAR-T cells secreted more IFN-γ and granzyme B than RNAU6 anti-CD19 CAR-T cells. CAR-T cells proliferated after U-2932 cell stimulation and were able to sustain proliferation. After stimulation via U-2932 cell co-culture, both RNAU6 anti-CD19 CAR-T and LSD1 shRNA anti-CD19 CAR-T populations had increased proportions of cells with the TCM phenotype, with a higher percentage among LSD1 shRNA anti-CD19 CAR-T cells. Conclusion: We developed a novel, feasible CD19-LSD1 shRNA CAR-T cell strategy for the treatment of DLBCL. Our