Background: Relapsed/refractory diffuse large B-cell lymphoma (r/r DLBCL) has a poor prognosis. Double-expressor (DEL) or TP53 abnormal r/r DLBCL patients(pts) are associated with even worse outcomes. Class I and IIb histone deacetylases (HDACs) are overexpressed in DLBCL and have been identified as a therapy target. Purinostat Mesylate (PM) is a high selective HDAC I/IIb inhibitor. Phase I dose-escalation trial of PM (1.2, 2.4, 4.0, 6.0, 8.4, 11.2, 15 mg/m²) by i.v was conducted in 29 hematologic malignancies at day 1, 4, 8, 11 of a 21-day cycle. PM was generally well tolerated with no DLTs. 61.1% (11/18) ORR was observed in r/r lymphoma pts. Based on these data, we conducted a phase 2a to further explore efficacy and safety of PM and mechanism of actions. Methods: This randomized, multicenter, open-label, phase 2a study was conducted from Nov.2022 to the present(NCT05563844). Key eligibility include r/r DLBCL pts with prior therapy including anti-CD20 antibody and anthracycline-based chemotherapy; ECOG≤2. Thirty pts were randomized 1:1 received PM at 8.4 and 11.2 mg/m² on Day 1, 4, 8, 11 of a 21-day cycle. Pts continued to receive PM until disease progression or unacceptable toxicity. Primary outcome was ORR and safety. Multiple cell lines and PDX mouse models were used to evaluate the PM activity and mechanism of action in vitro and in vivo. ATAC-seq, bulk RNA-seq, and scRNA-seq from both PDX models and pts were investigated for the activated immune response of PM. Results: Thirty patients were enrolled and 28 patients were evaluable. The ORR (20/28) was 71.4% (95%CI:51.3-86.8) with 5 CR and 15 PR. Fifteen pts at 8.4 mg/m² achieved an ORR of 66.7%(95%CI:38.4-88.2) with 1 CR and 9 PR. Thirteen pts at 11.2 mg/m² achieved ORR of 76.9%(95%CI:46.2-95.0) with 4 CR and 6 PR. As of the data cut off in Feb. 2024, 7 pts remained on treatment and the longest treatment has lasted 17 cycles. Median PFS was 4.3m (95%CI:2.8-8.5), OS were immature. In subgroup analyses, 7 DE DLBCL pts obtained 42.9%(3/7) ORR and 11 pts with TP53 abnormal by FISH or NGS test achieved 45.5%(5/11) ORR. Fifteen pts with non-DE or without TP53 abnormal achieved ORR of 86.7%(13/15). The most frequently reported Grade≥3 TRAE were thrombocytopenia, neutropenia, lymphocytopenia. No PM-related death was reported. PM monotherapy showed stronger and superior antitumor effects in DE DLBCL and TP53 mutations PDX models than selinexor and R-CHOP. PM significantly down-regulates the proteins c-MYC, EZH2, and mutated P53. ATAC-seq, bulk RNA-seq and scRNA-seq revealed that PM can stimulate the proliferation and activation of cytotoxic T cells and NKT cells, up-regulate the expression of B-cell tumor MHC I and II and inhibit tumor cell immune escape. Conclusions: This study further supports the recommended dose 11.2 mg/m² PM as the phase 2b for r/r DLBCL. Currently, the phase 2b, open-label, multicenter study has enrolled in 37 sites in China. Clinical trial information: NCT05563844.