Abstract 505: Oral azacitidine modulates the immune microenvironment in acute myeloid leukemia (AML) patients in remission: A subanalysis from the QUAZAR AML-001 Trial

• Published in: Cancer research (Chicago, Ill.) Ill.), 2021-07, Vol.81 (13_Supplement), p.505-505
• Main Authors: Menezes, Daniel L., See, Wendy L., Risueno, Alberto, Ma, Jianglin, La Torre, Ignazia, Skikne, Barry, Beach, CL, Kumar, Keshava, Thakurta, Anjan
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2021-505

BACKGROUND The immunologic effects of maintenance therapy in patients (pts) with AML in remission are not well-characterized but of high clinical interest, as rapid recovery of bone marrow (BM) after intensive chemotherapy (IC) may help delay relapse. Post IC, immunological interactions in the BM microenvironment present several immunosuppressive mechanisms. PD-L1 is commonly overexpressed on AML blasts, which is associated with worse prognosis. Oral azacitidine (Oral-AZA [CC-486]) is a hypomethylating agent recently approved in the US for pts with AML in complete remission (CR) or CR with incomplete hematologic recovery (CRi). To better understand the effects of Oral-AZA on immune cells, checkpoint expression of PD-L1/2 on AML blasts and normal myeloid progenitors (MPs), and the kinetics of T cell recovery and activation/exhaustion (eg, PD1, TIM3) were assessed. METHODS Biomarker-evaluable pts aged ≥ 55 years with AML were randomized 1:1 to Oral-AZA 300 mg (n=56) or placebo (PBO, n=52) post IC within 4 months of achieving CR. Flow cytometry evaluations of BM aspirates were performed at screening (ie, baseline [BL]), every 3 cycles until cycle 24 and every 6 cycles thereafter to cycle 36, or as clinically indicated. Correlative analyses of baseline immune parameters with median (med) relapse-free survival (RFS) were computed using Kaplan-Meier methods. RESULTS In the biomarker-evaluable pts, PD-L1 and PD-L2 expression at BL were higher on AML blasts (med intensity 7.1 and 2.9) than normal MPs (0.7 and 1.6). Most AML blasts were PD-L2+ (79%), whereas only 1.9% were PD-L1+. When stratified by the med, higher BL CD3 T cell numbers (as a % of total BM ) were associated with favorable RFS in both Tx arms (Oral-AZA: ≥ med, 562 days[d] and < med, 235d [P = .0308]; PBO: ≥ med, 325d and < med, 155d [P = .0391]). At cycle 3, pts in the Oral-AZA arm had a 1.7-fold increase in CD3 T cells from BL (PBO, 1.1; P = .0450), suggesting Oral-AZA can promote immunologic recovery during early Tx cycles. There was an inverse correlation between T cell exhaustion marker phenotypes (PD1/TIM3+) with CD4 (r = -.5967; P < .0001) and CD8 (r = -.2484; P = .0095) T cell numbers. An increase in RFS was seen in the PBO arm with lower PD1/TIM3+ CD4 numbers (< med, 429d; ≥ med, 155d; P = .0037), with a nominal increase observed in the Oral-AZA arm (< med, 428d; ≥ med, 303d; P = .6764). In a subset of pts, Oral-AZA appeared to suppress CD4 T cell exhaustion (PD1/TIM3+) compared with PBO. CO