A Phase 1 Study of NKX101, a Chimeric Antigen Receptor Natural Killer (CAR-NK) Cell Therapy, with Fludarabine and Cytarabine in Patients with Acute Myeloid Leukemia

Introduction

Relapsed or refractory (r/r) acute myeloid leukemia (AML) has a dismal prognosis. Treatment options are limited and include fludarabine (Flu) and cytarabine (Ara-C), +/- anthracyclines (e.g., FLAG-Ida). Recent progress has been limited to targeted therapies, which benefit only a subset of the population; efforts to develop cellular therapies have not been met with success. NKX101 is an NK cell therapy derived from healthy donors and engineered to express an NKG2D ligand-directed chimeric antigen receptor (CAR) to enhance killing of malignant cells, as well as a membrane bound form of interleukin (IL)-15 to promote persistence and activity. Because Ara-C is known to upregulate NKG2D ligand expression, use of Flu/Ara-C as an alternative to standard Flu/cyclophosphamide (Flu/Cy) for lymphodepletion (LD) was tested for NKX101.

Methods

Six patients with r/r AML were enrolled in this cohort of NKX101-101, a Phase 1 safety study. All patients had received at least one prior line of therapy, as well as approved targeted therapies for those with underlying genetic mutations. Patients received NKX101 at a dose of 1.5 billion (B) viable CAR+ cells/dose on Days 0, 7, and 14 after LD comprising Flu (30 mg/m2) and Ara-C (2 g/m2) each once daily for five days and 2 days of rest. Efficacy assessments were performed on Day 27 and included bone marrow aspiration and biopsy. Additional treatment cycles of LD and NKX101 were given to some patients to deepen or consolidate responses. Pharmacokinetic (PK) and cytokine sampling were performed throughout the treatment cycle(s).

Results

At baseline, five out of six patients had poor-risk genetic features, including TP53 mutation; the median bone marrow blast count was 35% (range: 20 - 86%). Patients had received a median of two prior lines of therapy, with all patients having previous venetoclax exposure (Table 1). All patients received at least three doses of NKX101 at 1.5 B cells/dose. There were no cases of cytokine release syndrome (CRS), immune cell associated neurotoxicity syndrome (ICANS), or graft-versus-host disease of any grade in any patients (Table 2). Myelosuppression and infection were the most common higher-grade toxicities. Four out of six patients had CR/CRi, with three out of six achieving CR. Three of these patients had no detectable MRD by flow, and one had 0.18% RUNX1-RUNX1T1 translocation via polymerase chain reaction testing on bone marrow. One subject was taken to consolidative hematopoietic cell transplant (HCT) and remains in CR. Another subject received a consolidative cycle of therapy and remains in CR. A third subject had three cycles of treatment with successive decrease in disease burden and remains in CRi. The fourth subject had CR after one cycle of treatment.

PK profiling showed that NKX101 was consistently detected and correlated with infusion days (Figure 1).

Cytokine data showed minimal post-infusion elevations above baseline; no association was observed between clinical response and elevation of serum cytokines (including IL-15, IL-6, IFNγ, IL-10, and IL-8).

Data on the expression of NKG2D ligands (MICA, MICB, ULBP1, and ULBP3) in subject bone marrow samples were obtained via multiplex immunohistochemistry. This was evaluated as a composite H-score. Responders were split between those with relatively low ligand expression and those with higher ligand expression, suggesting that further factors may contribute to NK cell resistance.

Conclusions

Alternative LD with Flu/Ara-C followed by NKX101 shows promising early responses including MRD negativity in r/r AML. The toxicity profile of this regimen is manageable and was consistent with underlying AML and exposure to LD, without any events of CRS or ICANS. Expansion of enrollment and longer follow-up are needed.

Despite being an allogeneic, non-HLA-matched cell product, NKX101 persisted for up to three weeks in PK testing.

Directly comparing the data from patients treated with Flu/Cy with NKX101 versus Flu/Ara-C with NKX101 suggests that Flu/Ara-C can replace the Flu/Cy regimen common to CAR T cell therapy as LD without compromising overall NKX101 exposure.