Study Design
This was a phase 1, open label dose escalation trial of orally administered NMP in patients with relapsed/refractory multiple myeloma (RR-MM). Eligible patients were ≥ 18 years old with Eastern Cooperative Oncology Group (ECOG) < 2 and histologically confirmed plasma cell myeloma defined by WHO 2008 criteria with measurable disease (defined by serum M-protein > 5g/L, urine M-protein > 200mg/24hrs, involved serum free light chain > 100mg/L or measurable soft tissue plasmacytoma)29. Patients were required to have relapsed following prior treatment and/or be refractory to or intolerant of both bortezomib and lenalidomide and must have undergone autologous stem cell transplant (unless deemed ineligible at investigator discretion). Exclusion criteria included: central nervous system involvement, active second malignancy within two years, uncontrolled medical conditions contraindicating an investigational agent due to inadequate organ function and concomitant exposure to anti-myeloma therapies.
The primary study endpoint was to determine the MTD of orally administered NMP in patients with RR-MM. Secondary endpoints included safety, PK analyses, IMWG defined response rate and exploratory immunological biomarkers. Adverse events were defined according to Common Terminology Criteria for Adverse Events (CTCAE version 4.03). Haematological DLTs were defined as grade 4 neutropenia unresponsive to G-CSF, grade 4 thrombocytopenia and febrile neutropenia. Non-haematological DLTs were grade 3 or higher and considered probably related to study drug and not attributable to disease progression and those resulting treatment delays greater than two weeks due to drug related toxicity. Self-limited (< 24hrs) gastrointestinal side effects (e.g., nausea, diarrhoea) responding to supportive measures were not classified as DLTs.
Pharmaceutical grade NMP stock solution (Ashland, Columbus, OH, USA) was diluted in ORA SWEET® SF (Medical Flavouring Systems, Vic, Australia) oral syrup vehicle and sterile water to a final concentration of 50 mg/mL and dispensed weekly in amber glass bottles for storage a room temperature. NMP was administered as a single morning dose on an empty stomach at least 30 minutes prior to food. Patients were dosed according to an ‘accelerated’ escalation schedule (one patient per dose level) in the absence of significant toxicity until a DLT was experienced and then in a 3 x 3 ‘standard’ dose escalation phase. Intra-patient dose-escalation at 100mg increments was also permitted whereby the MTD had not been determined and the patient tolerated two consecutive cycles without DLT or an objective disease response. Treatment was administered daily for 28-day cycles until withdrawal from protocol due to toxicity or disease progression.
Pharmacokinetic analysis
Blood samples were taken from patients before NMP and at one, two, four, eight and 24 hours post dose. Plasma samples (20µl) were prepared by adding five volumes of methanol containing deuterated internal standard. The samples were vortexed then centrifuged and the supernatant was transferred to a vial and injected onto the LCMSMS. The LCMSMS system consisted of a Shimadzu UHPLC with a SCIEX 6500QTrap, a KinetexC18 (3 x 50mm, 2.6 µm) column and using a gradient of 0.1% formic acid and acetonitrile. The initial conditions were 0.5 ml/min and 2% acetonitrile, this was held for 1 min then increased to 20% acetonitrile at 2 min then increased to 70% acetonitrile and held for 0.5 min. Equilibration time was 0.5 min. MRMs for NMP were 100→58, 100→69, 100→41 and for the internal standard d9NMP 109→62, 100→61, 100→46. PK analysis was performed using Pkanalix (version 2021R1. Antony, France: Lixolt SAS, 2021)
Blood samples for immunological biomarkers
Peripheral blood was collected at study enrolment, and at monthly intervals after each cycle of treatment. PBMC were isolated using Ficoll-Paque Plus (GE Healthcare, Chicago, IL, USA) and cryopreserved until required for analysis. Ten patients had sufficient PBMC stored for analysis (screening or C1D1 (pre-treatment), and C2D1 or C3D1 (post)). Four patients had C4D1 samples collected, and one patient on long-term NMP treatment had C6D1, and C12D1 samples analysed.
Flow cytometry
PBMC were stained using Zombie Live/dead fixable cell stain (Thermo Fisher Scientific, Waltham, MA, USA) prior to staining with specific antibodies. Antibodies used were CD3 BUV496 (BD Bioscience), CD19 PeCy7 (BD Bioscience), CD56 BB700 (BD Bioscience), CD16 BV650 (BioLegend), CD96 BV421 (BD Bioscience), TIGIT BUV395 (BD Bioscience), DNAM1 (CD226) BUV805 (BD Bioscience). PBMC were stained in Live/Dead cell stain for 15 minutes in PBS at RT, washed in FACS buffer (PBS + 2% FCS), and blocked in Human Fc block (BD Bioscience) for five minutes at RT. Cell staining was performed in FACS buffer for 30 minutes at 4°C, followed by two washes in FACS buffer, and fixation in 2% paraformaldehyde (PFA, Electron Microscopy Sciences). Samples were acquired on a LSR Fortessa (BD Bioscience) flow cytometer and analysed using FlowJo software (BD Bioscience). Flow cytometry gating is shown in Supplementary Fig. 1. NK cells were defined as CD3-CD19-CD56 + viable single lymphocytes. NK cell subsets were defined as mature (CD16 + CD56mid), regulatory (CD16-CD56mid), and immature (CD16-CD56hi). Expression of DNAM1, TIGIT, and CD96 was calculated using the geometric mean of the mature, regulatory and immature NK cell subsets.
Gene Expression Analysis
NanoString gene expression profiling of PBMC was performed as previously described30 using NanoString PanCancer Immune Profiling Panel V1 (NanoString Technologies, Seattle, WA, USA). Data was normalised and analysed using nCounter Advanced Analysis software (version 2.0.115; NanoString Technologies).
Statistical analyses
The study was planned to recruit up to 20 patients, 15 in the dose finding cohort and five as a dose expansion. A pragmatic sample of 15 patients would yield a 95% confidence interval between 4 and 48% for an actual grade 3–4 event rate of 20% which was considered suitable precision for a phase 1 pilot study. Time to progression was calculated from the first day of treatment to progression using the Kaplan-Meier method. Flow cytometry and gene expression data were analysed using GraphPad Prism (version 6.0; GraphPad Software, San Diego, CA). Data was analysed using Wilcoxon’s matched pairs signed rank test compared with baseline samples (*p < 0.05).