Targeting Deacetylases As a Novel Strategy for Prevention of Acute Gvhd

Introduction: Histone deacetylase inhibitors (HDACi) are an important class of anti-cancer agent and have been shown by us and others to attenuate experimental GVHD in multiple murine models. We and others have shown that HDACi reduce proinflammatory cytokines, modulate APC functions through induction of IDO in a STAT-3-dependent manner, and increase Treg numbers (PNAS 2004; JCI 2008; Nat Med 2007). We have now translated our experimental observations from murine models into a first-in-human clinical trial to test the hypothesis that HDAC inhibition is safe and reduces GVHD severity in patients undergoing matched related donor (MRD) reduced intensity conditioning (RIC) HSCT.

Methods: 45 subjects were enrolled from two centers in a phase I/II trial that added the HDACi, vorinostat 100 mg BID, to a standard GVHD prophylaxis regimen (tacrolimus and MMF). The primary endpoint was the incidence of day 100 grade 2-4 acute GVHD with a risk of 25% compared with 42% in historical controls. The RIC regimen consisted of fludarabine 40 mg/m² (4 days) and busulfan 3.2 mg/kg (2 days). Based on our experimental observations, we obtained 4 blood samples from study subjects: (1) prior to study drug administration (pre-therapy), (2) day 1, (3) day 30, and (4) day 100 to perform correlative studies.

Results: Vorinostat was safe, tolerable, and feasible to administer after MRD RIC HSCT. All study subjects have engrafted and there have been no DLTs. Donor chimerism was 95%, 97.5%, and 100% by day 30, 100 and 365, respectively. We performed pharmacokinetic studies of vorinostat in blood samples on day 1 at: pre-dose (immediately before the first dose of the day), 0.5, 1, 2, 3, 4, 6, and 12h (immediately before the second dose of the day). The mean half-life of study subjects was 1.8h (1.3-2.3h) resulting in a mean AUC, Cmax and Tmax 2.5±1.4μM*h, 0.49±1.0μM, and 2.6h (1-4h), respectively. We then investigated the pharmaco-dynamic effects of HDAC inhibition on day 30 after BMT. Study patients demonstrated enhanced H4 acetylation in PBMCs compared with historical controls (P=0.009), suggesting that HDAC enzymes were inhibited. We also determined the functional impact of HDACi on PBMCs from study patients and found that HDACi acetylated STAT-3 (P=0.005), enhanced IDO expression (P<0.0001), and reduced in vivo TNF-α (P=0.039) and IL-6 (P=0.048) secretion and also ex vivo following LPS stimulation (P<0.05). Study patients showed increased numbers of CD4⁺CD25⁺CD127^- T cell population (Treg) (P=0.04) and Foxp3⁺ expression by qPCR (P=0.006). Consistent with experimental and laboratory data, the incidence of GVHD at day 100 (primary endpoint) was reduced to 22% for grade 2-4 acute GVHD (only 4%: grade 3-4) with no increase in relapse (17% at 1-year).

Conclusions: These data demonstrate translation of our experimental observations into a novel proof-of-concept phase I/II clinical trial of vorinostat in reducing the incidence of GVHD.