Recombinant Human IL-15 Promotes in Vivo Expansion of Adoptively Transferred NK Cells in a First-in-Human Phase I Dose Escalation Study in Patients with AML

Abstract 894

Adoptive transfer of haploidentical NK cells can induce remissions in patients with refractory AML. However, many do not expand NK cells and fail to respond. While IL-2 can promote NK cell proliferation, it also promotes the expansion of regulatory T cells, which impede NK cell expansion. Because IL-15 has different effects on regulatory T cells than IL-2, we initiated a phase I dose escalation trial of recombinant human IL-15 to enhance adoptive transfer of NK cells. Patients with refractory AML receive a lymphodepleting preparative regimen of fludarabine 25 mg/m² × 5 days and cyclophosphamide 60 mg/kg × 2 days. Haploidentical NK cells (CD3- and CD19-depleted and overnight activated with IL-15 10 ng/ml) were infused Day 0, followed by 12 daily doses of intravenous IL-15 (Biopharmaceutical Development Program, NCI Frederick) in planned dosing cohorts of 0.25, 0.5, 1, 2 and 3 mcg/kg. To date 9 patients have been treated (Table). The first 6 patients (0.25 and 0.5 mcg/ml cohorts) received all 12 planned doses of IL-15, with no dose-limiting toxicities (DLTs). Apart from transient fevers, the IL-15 was well tolerated. While donor-derived NK cells were detected at Day 7 in all patients, none achieved the primary endpoint of >100 donor-derived NK cells/μl circulating in blood at Day 14. At the completion of IL-15 dosing, all patients in cohorts 1–2 had a lymphocytosis comprised of host T cells, which were mainly CD8⁺ (mean 81±6%). Five of 6 recovered neutrophils by Day 28 (range 15–26 days), and one with residual leukemia remained neutropenic leading us to conclude that IL-15 does not impede neutrophil recovery. Three of the 6 cleared their leukemic blasts, fully recovered and proceeded to allogeneic hematopoietic cell transplant (HCT). In contrast to the first 2 cohorts, the first patient at the 1 mcg/kg dose experienced a DLT (grade 4 dyspnea due to diffuse alveolar hemorrhage requiring high dose steroids) and received only 8 does of IL-15 making him not evaluable for in vivo expansion. Due to the DLT, the 1 mcg/kg dose cohort will expand to 6 evaluable patients, 2 of whom have completed IL-15 dosing. Both are evaluable, each having received the required minimum of 9 doses, but additional planned doses 10–12 were held in both patients due to high fevers and transient hypoxia possibly related to infection that did not constitute DLT. Both cleared refractory leukemia at Day 14 and successfully expanded donor NK cells (2094 and 448 cells/ml) at the end of IL-15 dosing. The in vivo expanded NK cells exhibited potent function, with 81.5% and 82.3% cytotoxicity against K562 targets at a 20:1 E:T ratio. Thus 1 mcg/ml dosing of IL-15 is significantly more likely to induce successful donor NK cell expansion at day 14 than the 0.25 or 0.5 mcg/ml doses (p = 0.04). Since endogenous IL-15 may heterodimerize with its receptor, IL-15Ra, to provide more stability and potent signalling to NK cells, batched serum samples are in process to measure free and IL-15Ra complexed IL-15. In summary, this platform of adoptive transfer of haploidentical NK cells with IL-15 has, in this early experience shown to be an effective treatment for refractory AML allowing patients to achieve remission and subsequent allogenenic HCT. The 1 mcg/kg dose is associated with more toxicity and limited ability to deliver all 12 doses, but toxicity was transient. 9 daily doses were sufficient to promote robust in vivo expansion of highly functional donor-derived NK cells. Further dose modifications (perhaps with continuous infusion or using fewer doses given subcutaneously) may be required to enhance safety. Based on this preliminary experience, IL-15 should emerge as the optimal cytokine to promote expansion and activation of adoptively transferred NK cells without Treg stimulation, which should be effective therapy for AML.

In vivo expansion of donor-derived NK cells is dependent on IL-15 dosing.