Two-Stage Priming of Allogeneic Natural Killer Cells for the Treatment of Patients with Acute Myeloid Leukemia: A Phase I Trial

Abstract

Human Natural Killer (NK) cells require at least two signals to trigger tumor cell lysis. Absence of ligands providing either signal 1 or 2 provides NK resistance. We manufactured a lysate of a tumour cell line which provides signal 1 to resting NK cells without signal 2. The tumor-primed NK cells (TpNK) lyse NK resistant Acute Myeloid Leukemia (AML) blasts expressing signal 2 ligands. We conducted a clinical trial to determine the toxicity of TpNK cell infusions from haploidentical donors. 15 patients with high risk AML were screened, 13 enrolled and 7 patients treated. The remaining 6 either failed to respond to re-induction chemotherapy or the donor refused to undergo peripheral blood apheresis. The conditioning consisted of fludarabine and total body irradiation. This was the first UK trial of a cell therapy regulated as a medicine. The complexity of Good Clinical Practice compliance was underestimated and led to failures requiring retrospective independent data review. The lessons learned are an important aspect of this report. There was no evidence of infusional toxicity. Profound myelosuppression was seen in the majority (median neutrophil recovery day 55). At six months follow-up, three patients treated in Complete Remission (CR) remained in remission, one patient infused in Partial Remission had achieved CR1, two had relapsed and one had died. One year post-treatment one patient remained in CR. Four patients remained in CR after treatment for longer than their most recent previous CR. During the 2 year follow-up six of seven patients died; median overall survival was 400 days post infusion (range 141–910). This is the first clinical trial of an NK therapy in the absence of IL-2 or other cytokine support. The HLA-mismatched NK cells survived and expanded in vivo without on-going host immunosuppression and appeared to exert an anti-leukemia effect in 4/7 patients treated.

Trial registration: ISRCTN trial registry ISRCTN11950134.

Conflict of interest statement

Competing Interests: MWL is a consultant to Coronado Biosciences which has licensed this technology from UCL. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Trial enrollment summary.

Fig 2. Patient conditioning regimen.

Fig 3. Donor NK chimerism.

Donor NK chimerism was measured by flow cytometry using monoclonal antibodies specific for an MHC Class I allele which was restricted to the donor cells. Concomitant labeling with anti-CD3 and anti-CD56 allowed detection of donor T, NKT and NK cells in patients’ peripheral blood and bone marrow. Preliminary experiments confirmed sensitivity (A) and specificity; indeed the method consistently underestimated the dosed proportion of donor cells (B). Three evaluable patients showed consistent presence of low levels of donor NK cells (filled circles and solid line) in peripheral blood (C, D, E). Patient 01 (C) and Patient 03 (D) showed absence of donor T cells (open triangles and dashed line). Patient 04 who suffered profound marrow aplasia and required allogeneic stem cell rescue was the only one to have detectable donor T cells in repeated blood samples (E).

Fig 4. Clinical outcome.

Duration of most recent previous CR (black blocks) or PR (white block) prior to ATIMP administration (white arrow) is shown relative to the duration of CR post IMP infusion (gray blocks). Patients 01, 03, 04 and 11 all experienced a longer duration of remission post IMP infusion than their previous CR which was not seen in the other patients.

Fig 5. Detection of donor NK cells in bone marrow.

A single bone marrow aspirate taken from Patient 03 48 days after receipt of a second infusion of primed donor NK cells showed absence of HLA-A24+ donor T cells (A and B). Donor NK cells (C) expressing HLA-A24 represented almost 1/3 of the total NK cells in the bone marrow (D).

Fig 6. Presence of activated NK cells in peripheral blood.

All patients were monitored for the proportions of resting and activated NK cells in their peripheral blood prior to and post primed allogeneic NK cell infusion. NK cell activation in peripheral blood increased within the first two weeks of infusion and did not return to pre-infusion levels until week 12 (A—data shown are mean +/- SD n = 7). A single patient (01) was consistently monitored for the presence of activated NK cells in her peripheral blood which lysed NK-resistant Raji cells in vitro at a fixed effector:target ratio of 5:1 (B—closed circles). The presence of functional primed NK cells closely mirrored the presence of CD69+ NK cells (B—open squares).

Fig 7. Haploidentical tumor-primed NK cells do not suppress hematopoietic progenitor cells (HPC) in vitro.

A—Tumor primed NK cells co-incubated with haploidentical CD34 selected HPC did not suppress BFU-E (white bars), CFU-GM (gray bars) or GFU-GEMM (black bars). B—T cells from the same haploidentical donor substantially inhibited growth of donor CFU-GM (gray bars) C—In vivo donor T cell (Open triangles) and NK (Open Squares) cell chimerism.