Prevention of Acute Gvhd By Ex Vivo Expanded Umbilical Cord Blood Derived Regulatory T Cells (Treg)

We previously showed that adoptive transfer of 3 million Treg per kilogram recipient body weight could reduce the risk of acute GVHD (aGVHD) (39% vs. 61%, p-0.05; Blood 117:1061) in recipients of double UCBT. We hypothesized that control of GVHD was suboptimal, likely related to the low ratio of Treg to T cells in the UCB graft (ie, 0.3:1.0). Our group previously demonstrated optimal murine GVHD control with Treg:Teffector ratio ≥ 1:1. A genetically modified K562 cell line, termed KT64/86, that expressed CD64, a high affinity Fc receptor loaded with anti-CD3 antibody, and CD86, the natural ligand of CD28/CTLA-4, was used to manufacture higher doses of Treg from a single UCB unit, that retained the CD4+CD25+FoxP3+ CD127- phenotype and had potent in vitro and in vivo (murine xenogeneic) suppression of T cell proliferation. Using this methodology, we performed a phase I dose escalation trial starting with 3 million donor Treg/kg. Patients who were candidates for nonmyeloablative double UCBT were eligible. Ten patients received Cy50/FLU200/TBI200 nonmyeloablative conditioning with sirolimus and mycophenolate mofetil immunosuppression. Thawed Tregs underwent a single CD25 positive selection and were cultured for 18 ±1 days in the presence of KT64/86 (1:1) supplemented with 300 units/mL rhIL-2; cultures were restimulated with KT64/86 on day 12. In the dose escalation, 3 to 100 million Tregs/kg were infused on day +1 and patients monitored for infusional toxicity and GVHD. Fifteen concurrent patients received same treatment regimen but no Treg served as controls for clinical endpoints. The median age of Treg treated patients was 62 (r:45-69), 5 were male, 6 had acute leukemia, 4 had CLL/lymphoma. The median infused combined nucleated and CD3+ cell doses of the UCB graft were 4 x10e7 (r:4-9) and 16.5 x 10e6 (r:14-18), respectively. None had a severe adverse event with the infusion. One Treg treated patient developed a skin rash on day +16; biopsy was equivocal for engraftment syndrome to be conservative we considered as aGVHD. Thus, the incidence of grade 2-4 aGVHD by day +100 was 10% (95%CI, 0-28%) as compared to 53% (95%CI, 27-79%) in the controls (p=.04, FIGURE). One patient developed grade 3 late gastrointestinal aGVHD in the Treg group (day 119).  Engraftment was similar between Treg (90%, 95%CI, 74-99%) and control (87%, 95%CI, 65- 98%) groups (p=.92). The median follow-up is too short for assessment of chronic GVHD and relapse. In summary, we 1) demonstrated the ability to consistently manufacture Treg doses up to 100 x10e6/kg which allows for Treg:Teffector ratio of up to 7:1, and 2) observed a remarkably low risk of aGVHD. This promising result is the first clear demonstration that ex vivo expanded UCB Treg are potent suppressors of aGVHD in humans and sets the stage for a future pivotal clinical trial that will assess the clinical impact of Treg GVHD prophylaxis on immune recovery and risk of leukemia relapse.