Inhibitors of class I/II histone deacetylases (HDACi) possess anti-leukemic activity and have been reported to modulate the function of immune effector cells. Thus, they could provide specific clinical benefit in the allogeneic hematopoietic stem cell transplantation (HSCT) setting. Panobinostat (PAN) is a potent, orally available pan-HDACi reported to either suppress or stimulate regulatory T cells (T reg), depending on the administered dose (Shen L & Pili R, OncoImmunology 1;7:948, 2012). The feasibility and efficacy of PAN treatment following HSCT for patients (pts) with high risk acute myeloid leukemia (AML) has not been established. We report clinical and translational results of the dose-escalation phase of the PANOBEST study with PAN as post-HSCT maintenance.
Primary objectives were, based on dose-limiting toxicity (DLT), determining the maximum tolerated dose (MTD) and/or the recommended phase 2 dose (RP2D) of PAN in patients with high risk AML or myelodysplastic syndromes (MDS) in complete remission (CR) after reduced-intensity conditioning HSCT. Secondary objectives were the evaluation of safety, tolerability and immunoregulatory properties of PAN, and overall (OS) and disease-free survival (DFS) of treated patients.
PAN was started at 10 mg p.o. three times a week (TIW) and escalated to 20 and 30 mg TIW using a 3+3 design. Treatment was initiated 60-150 days after HSCT and continued for up to one year. Eligibility criteria included: recovery of peripheral blood counts, adequate organ function and no severe graft versus host disease (GvHD). All pts gave written informed consent. DLT was defined as prolonged grade 4 hematologic or grade 3/4 non-hematologic toxicity within 28 days of the first PAN dose. Immunophenoytyping of lymphocyte subsets was performed pre-treatment and on days 3, 8, 30, 90, 180 and 360.
12 pts (11 AML, 1 MDS), median age of 52 years (21-62) were enrolled. PAN was started within a median of 73 days (60-126) after HSCT, which was performed with active disease (n=11) or in CR2 (n=1). The RP2D was determined to be 20 mg TIW based on one DLT (fatigue grade 3) at 20 mg and two DLTs (nausea/emesis and colitis grade 3 each) at 30 mg. Grade 2-4 adverse events (AEs) were reported in 10 out of the 12 pts (83%). Grade 3/4 AEs included hematologic toxicity (50% of pts), laboratory alterations (33%), gastrointestinal symptoms (25%), fatigue, pulmonary infection (17% each), sepsis, herpes stomatitis, diabetes, syncope, deep vein thrombosis and pulmonary embolism (8% each). Toxicity was reversible and required at least one PAN dose reduction in 3 pts. Acute GvHD grade 2 (1 pt) and 3 (2 pts) was responsive to steroids in 2 pts or salvage therapy in 1 pt. Four pts developed mild (n=3) or moderate (n=1) chronic GvHD. To date, 5 pts have completed one year of PAN and 2 pts remain on treatment (days 238, 290). Five pts discontinued treatment prematurely after 10-217 days due to grade 3 toxicities (n=4) or AML relapse (n=1). With a median follow up of 579 days (129-911), 11/12 pts are alive and 10/12 in continuous CR after HSCT. Seven pts received prophylactic donor lymphocyte infusions (DLIs, 1-5 doses of 0.1-20x10
6 CD3
+ cells/kg). Immunophenotyping revealed no impact of PAN on absolute T reg numbers (9 pts), but a significantly reduced proportion of CD4
+CD25
++CD127
dim/- T reg to CD3
+CD4
+ T helper (Th) cells by day 8 after 3 doses of PAN (mean±SEM: 14.6±2.6 vs. 9.6±1.2%, p value of t test =0.03). While Treg/Th proportion continuously decreased in pts with ongoing CR, it again increased after PAN discontinuation or remained stable under PAN treatment in both relapsing patients. Outcome of the study population was compared with a historical cohort of 29 consecutive pts with active AML transplanted in Frankfurt in 2000-2009. Both cohorts were similar in age, gender, disease stage or BM blasts at time of HSCT, donor type and use of DLIs. In a landmark analysis including all pts who were in CR and without severe GvHD on day 73 after HSCT, probabilities for DFS and OS at 18 months after HSCT were 83±11% vs. 55±9% (p=0.145, log-rank test) and 92±8% vs.66±9% (p=0.085) in the PANOBEST vs. historical cohort (
Fig 1).
PAN is well tolerated after HSCT at a RP2D of 20 mg TIW. Comparison with a historical control cohort of pts transplanted with active AML shows a low relapse rate, which appears to be associated with a PAN-induced modulation of the T reg/Th proportion.
Disclosures:
Bug:Novartis: Honoraria, Travel grants, support for the clinical study Other. Ottmann:Novartis: Consultancy, Membership on an entity’s Board of Directors or advisory committees, Research Funding, Speakers Bureau.
