Patient selection
For dose escalation phase patients were included over 18 years of age with either relapsed or refractory AML after standard therapy in which conventional salvage therapy was inappropriate or who were previously untreated for the reason of age, poor prognosis, or concurrent medical conditions and therefore were considered unfit for standard induction therapy, or patients with T-ALL or Philadelphia chromosome negative BCP-ALL suffering from relapse (after at least induction and consolidation chemotherapy) or showing refractory disease and having no option to be treated with standard therapy, or patients with relapsed or refractory Philadelphia chromosome- or BCR-ABL1-positive BCP-ALL or CML-BP after first- and second-line treatment (including at least two tyrosine kinase inhibitors (TKIs)) or showing molecular relapse or molecular progression in the minimal residual disease (MRD) measurement including the detection of the T315I mutation or other mutations conferring resistance to available TKI treatment. MRD analysis was based on the detection of BCR-ABL1 transcripts by RT-PCR. These patients could not be eligible for allogeneic stem cell transplant (SCT) at the time of enrolment. In the presence of a T315I mutation, prior treatment with a second TKI was not required.
Additional eligibility criteria included serum bilirubin ≤ 1.5 x the institutional upper limit of normal (ULN) except with known Gilbert’s Syndrome, alanine transaminase (ALT) and aspartate transaminase (AST) ≤ 3 x the ULN (or ≤ 5.0 x ULN in case of hepatic infiltration by leukemia), INR ≤ 1.5, serum creatinine ≤ 2 mg/dl or creatinine clearance ≥50 ml/min, fasting plasma glucose (FPG) ≤ 160 mg/dL, HbA1c ≤ 8 % and WBC ≤ 30 x 109/L.
Cytoreductive treatment was allowed up to one week prior to first dosage for typical prephase or maintenance medication for ALL like vincristine, mercaptopurine, low-dose (≤15 mg/m²) MTX and low-dose cyclophosphamide (cumulative dose ≤ 1g/m²) and treatment with steroids and hydroxyurea were allowed up to 1 day before first dosage.
Patients were excluded showing acute promyelocytic leukemia (APL), diabetes mellitus (insulin dependent or history of gestational diabetes), reduced cardiac function (left ventricular ejection fraction < 45% and/or QTcF > 480 msec on screening ECG) and patients with active ≥ grade II graft versus host disease (GVHD). Patients requiring treatment with medication having the potential to interact with BEZ235 in the respect of QT prolongation or metabolization via p450 cytochrome enzymes were also excluded.
The BID schedule investigated in this trial was selected based on preliminary data obtained from the solid tumor phase I study CBEZ235ZUS07T.
Study design
This was a phase I, single center, open-label study designed to assess the safety, tolerability, preliminary efficacy and PK of BEZ235. During the initial dose escalation phase, successive patient cohorts were scheduled to receive BEZ235 orally twice a day during 28-day cycles. Dose escalation was based on a “rolling-six” design, a modification of the 3 + 3 design. The first cohort of subjects received a starting dose of 400 mg/BID. Dose escalation for subsequent patient cohort of subjects was guided by the incidence of BEZ235-related adverse events as graded by NCI Common Terminology Criteria for Adverse Events (CTCAE) v4.03 (http://evs.nci.nih.gov/ftp1/CTCAE) in the first four weeks of dosing (dose limiting toxicities (DLT) evaluation period). Up to 6 subjects could be enrolled to a dose level, although only 3 subjects were required to complete the DLT evaluation period prior to enrolling subjects in the next higher dose cohort. The patient population for the determination of MTD consisted of patients who fulfilled the minimum safety evaluation requirements of the study, which were met if in cycle 1, the patient has been treated with BEZ235 for ≥ 21 days, observed for ≥ 28 days following the first dose, and completed all relevant safety evaluations, or experienced a DLT during cycle 1. Patients who did not meet these minimum safety evaluation requirements were regarded as ineligible for the MTD determining population and were replaced.
Toxicity and safety assessment
Adverse events data and serious adverse events were analyzed taking into account the NCI-CTCAE v4.03 and were presented in frequency tables by grade. Haematological and clinical biochemistry toxicities were assessed from laboratory test parameters. For patients with multiple occurrences of the same event the maximum grade (worst) was used. The safety analyses were performed in the safety population.
Response evaluation
Assessment of response was first performed at the end of cycle 1 (day 29) and after every subsequent cycle until CR or CRi were achieved. After CR or CRi, ongoing response was evaluated on day 29 of every even cycle and at end of treatment.
In case of ALL CR was defined as: absence of circulating blast cells, Polymorphonuclear > 1000 /mm3, platelets > 100 000 /mm3, no extra medullary involvement and a normal bone marrow smear. In case of AML CR was defined as: < 5% leukemic blasts in bone marrow aspirate and absence of Auer rods, no leukemic blasts in peripheral blood, neutrophils ≥ 1.0 x 109/L, platelets ≥ 100 x 109/L and no evidence of extramedullary disease. CRi was defined as: < 5% leukemic blasts in bone marrow aspirate and absence of Auer rods, no leukemic blasts in Peripheral blood, neutrophils < 1.0 x 109/L, platelets < 100 x 109/L and no evidence of extramedullary disease.
Pharmacokinetic analysis
Blood for Pharmacokinetic (PK) analysis was collected on day 1 of cycle 1 prior to and 2h, 4h, and 8h after the first administration of BEZ235, after 24 hours immediately preceding the second administration of BEZ235 and on day 15 of cycle 1 immediately prior to administration of BEZ235. All blood samples were taken by either direct vein puncture or an indwelling cannula in accordance with the assessment schedule, (table 1). If a patient had a central line, blood sampling was also obtained from this source. The sample tubes were to be labelled with pre-printed labels which contain the following information: protocol number, centre number, patient number, patient initials, sample number, date sample taken and actual time of sample. All samples were given a unique sample number and the exact clock time of dosing, as well as actual sample collection date and time will be captured on the CRF page. Sampling problems were noted in the comments field of the CRF page. All pharmacokinetic specimens were stored frozen at least at -20°C until shipment. Samples were packed carefully with suitable packing material and dry ice to keep them frozen during shipment. Laboratory analyses of specimens collected in this study for determination of drug or metabolite concentrations were conducted by the central laboratory at the University of Dresden.
Pharmacokinetic analyses were performed in 21 patients and in all over 235 plasma samples. BEZ235 plasma levels were determined by HPLC and fluorescence detection. For the HPLC assay zinc sulfate (5g/100ml) and acetone were added to plasma samples to precipitate protein. 300μl plasma were combined with 200μl ZnSo4 and 300μl acetone, were vortexed for 5 min and subsequently centrifuged for 10min followed by HPLC analysis. HPLC analysis was performed using a ZirChrom PBD column, which was equilibrated with 95% water, 1ml TEMED, 5% methanol and adjusted to a pH of 4 with phosphoric acid (800μl; pH3.6 / l Eluent). For elution an MN 125-4, 5μm, Rp SelectB column was used. The eluent was composed of 45% water, 1 ml TEMED and 55% methanol, adjusted to pH 2 with phosphoric acid (2250μl pH 2/l Eluent). Detection was performed with a fluorescence detector at 270nm extinction and 480nm emission. This methodology was validated to a detection limit of 1ng/ml BEZ235. The intra- und inter-assay variation for 10 samples each per measurement was below 10%. PK parameter analysis was performed for every patient individually based on a 2-compartment model using the TOPFIT PK computer program. For weighting the concentration was set to 1/y. All correlation coefficients were ≥ 0,85.
Pharmacodynamic parameters
Pharmacodynamic (PD) analysis included assessment of phosphorylation of AKT, S6 and 4EBP1 by WB and flow cytometry. Primary patient samples adequate for Western blot PD and flow cytometry analysis were obtained from bone marrow and or peripheral blood prior to therapy and on days 1, 2, 15, 28 and/or end of treatment (EOT). Jurkat T-ALL cells were used as positive control in each case.
Minimal residual disease
MRD assessment was performed after each cycle of BEZ235. Disease specific MRD markers were used for each patient individually. All transplanted patients were monitored additionally via regular chimerism analysis.
Mutation analysis
The presence of PI3KCA, AKT or PTEN mutations was evaluated by direct sequencing of exons with known mutation hotspots.
Whole genome and RNA sequencing
Samples from representative time points during treatment with BEZ235 were used for whole genome and RNA sequencing in the best responding patient described below. Herein a skin punch biopsy was used as non-malignant tissue control and a bone marrow sample from relapse after allogeneic SCT before initiation of therapy with BEZ235 as tumor sample.
DNA from tumor (mononuclear bone marrow cells) and normal tissue (skin punch biopsy) were extracted using the DNeasy Blood & Tissue kit (Qiagen). Sequencing libraries were prepared using the TruSeq DNA Nano kit (Illumina) and sequenced on an Illumina HiSeq X sequencer in paired and mode with a read length of 150 bp.
For analysis of WGS data the OneTouch Pipeline (OTP) developed at the German Cancer Research Center (DKFZ) has been employed. Details of the pipeline as described in Reisinger E et al.38 In brief, raw reads were aligned using bwa mem and SNVs were called using samtools (mpileup, bcftools) and platypus. Indels were called with platypus. The used alignment, SNV- and InDel-calling are based on the established workflows in the International Cancer Genome Consortium (ICGC) and are basically described in Alioto et al.39 Structural variants were called with SOPHIA (https://bitbucket.org/utoprak/sophia/src/master/) and CNVs were determined using ACESeq (https://www.biorxiv.org/content/10.1101/210807v1.full). The WGS pipeline of the OTP has been extensively validated and used for WGS data analysis in multiple international (ICGC) as well as national (MASTER, INFORM) cancer sequencing programs.
Total RNA was extracted from mononuclear bone marrow cells using the RNeasy Mini kit (Qiagen). Sequencing libraries were prepared using the TruSeq Stranded Total RNA kit (Illumina) and sequenced on an Illumina HiSeq 4000 in paired end mode with a read length of 100 bp. RNA sequencing data was analyzed using STAR38,40 and fusion transcripts were detected using STAR-Fusion (https://www.biorxiv.org/content/early/2017/03/24/120295).
Statistical considerations
This was a phase I multiple ascending dose study in which a following variant of the 3+3 rule (“Rolling Six”), a minimum of 6 and a maximum of 12 patients was recruited for the dose escalation and another 5 were planned for the expansion phase in ALL/AML. Demographics and baseline Characteristics: standard descriptive statistics, such as the mean, median, range and proportion, were used to summarize the patient sample and to estimate parameters of interest. Ninety-five per cent confidence intervals were provided for estimates of interest where possible. Statistical tests for comparison of different patient cohorts were not feasible due to low patient number.
Safety population
All patients who received at least one dose of study medication and had at least one post-baseline safety assessment (as evaluated by the existence of at least one adverse event CRF, including the case where no adverse event is reported) were included.
MTD determining population
The MTD determining population enclosed all patients from the safety population who either received enough treatment and had sufficient safety evaluations or discontinued due to unacceptable toxicity. The minimum safety evaluation requirements were met if, in cycle 1, the patient has been treated with BEZ235 for ≥ 21 doses. In addition, the patient must have been observed for ≥ 28 days following the first dose, and must have completed sufficient safety evaluations or the patient experiences DLT during cycle 1. Patients who did not meet these minimum safety evaluation requirements were regarded as ineligible for the MTD determining population and were replaced.
Trial registration
ClinicalTrials.gov, identifier NCT01756118. registered 19th December 2012, https://clinicaltrials.gov/ct2/show/NCT01756118