A phase Ib study of everolimus combined with metformin for patients with advanced cancer

Eligible patients were aged ≥18 years with histological or cytological confirmed solid malignancies that were refractory to standard therapies, or for which no standard treatment option was available. Other eligibility criteria included WHO performance status ≤2, estimated life expectancy of ≥3 months, adequate bone marrow (white blood cell count ≥3.0 × 10⁹/l, platelets ≥100 × 10⁹/l), hepatic (serum bilirubin ≤1.5 × upper limit of normal (ULN), ALAT/ASAT ≤2.5 × ULN or in case of liver metastases ≤5 × ULN) and renal function (serum creatinine ≤150 μmol/l). Patients also had to be mentally, physically and geographically able to undergo the study treatment and follow-up. Patients were ineligible if they were known with a serious concomitant medical condition, such as unstable angina pectoris, symptomatic congestive heart failure, myocardial infarction within the last 6 months before screening, serious uncontrolled cardiac arrhythmia, uncontrolled diabetes (fasting serum glucose >2 × ULN), severe infection, cirrhosis, chronic active/persistent hepatitis or severly impaired lung function. Furthermore, patients were ineligible when they had a known hypersensitivity to either study drug, if they had used either study drug in the previous 6 months or when they were concomitantly treated with strong CYP3A4/3A5/2C8 inhibitors/inducers. Women were ineligible when they were pregnant or lactating. All patients gave written informed consent.

The study protocol was approved by the Medical Ethics Committee of the Academic Medical Center (reference number NL_37906.018.12) and was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.

This was a phase I, open-label, single-center, dose escalation study to assess the safety, DLTs, maximum tolerated dose (MTD) and the pharmacokinetic interaction of the combination of everolimus and metformin. The study was conducted at the Academic Medical Center (The Netherlands). At least three patients per dose level were recruited and the cohort at this dose level was to be expanded to six patients if 1/3 patients experienced a DLT. Dose escalation to the next dose level was to be permitted if no DLT occurred in 0/3 or in ≤1/6 patients. In case of DLT(s) in ≥1/3 or in ≥2/6 patients, that dose level was to be declared intolerable and dose-escalation occurred. If dose level −1 was also intolerable, the study was to be terminated without establishment of an MTD. A DLT was defined as any of the following events related to study treatment and occurring during the first treatment cycle, defined by National Cancer Institute Common Terminology Criteria for Adverse Events version 3.0 (CTCAE): neutropenia grade 4 lasting >5 days or febrile neutropenia grade 3 (fever ≥38.5 °C), grade ≥3 anemia, grade ≥4 trombocytopenia or grade 3 trombocytopenia with bleeding, vomiting grade ≥2, diarrhoea grade ≥2 or any other toxicity grade ≥3 (excluding alopecia), despite optimal supportive care. In case of measurable disease, tumor measurements were performed using a CT scan at baseline and every nine weeks and were evaluated in accordance with the Response Evaluation Criteria in Solid Tumors (RECIST 1.1) [25].

During the first seven days of treatment, patients were treated with single-agent everolimus to reach steady-state concentrations. Treatment with metformin bid started on day 8. The everolimus and metformin doses were to be de-escalated or escalated as shown in Table 1. Metformin was administered bid instead of qd at dose level 1 to mimic conventional metformin dosing schedules in T2DM and because of the short elimination half-life (t_1/2) of metformin (5 h) [26]. In dose level −1, the everolimus dose was decreased rather than the metformin dose or the metformin dosing frequency because metformin pills <500 mg were not available and because of the difference in t_1/2 between metformin (5 h) and everolimus (30 h) [26, 27]. In the event of intolerable toxicity at dose level 1, the exposure to the study treatment is probably decreased more effectively by decreasing the everolimus dose than the metformin dose or its dosing frequency.

To determine the pharmacokinetic interaction between everolimus and metformin, patients received (only for pharmacokinetic purposes) one single morning administration of metformin 1 day prior to start of treatment (day −1), at the dose level that the patient would receive at start of treatment. Blood samples were drawn on t = 0 (predose), 1, 2, 4, 6, 8 and 12 h after the morning doses of metformin on day −1 and 15. Blood samples for everolimus were drawn on day 7 (without metformin) and on day 15 (with metformin), both at predose and at 1, 2, 4, 6, 8, 12 and 24 h after administration of everolimus. Using these blood samples, full pharmacokinetic 12/24-h profiles of the metformin and everolimus plasma concentrations were determined, both as single agents and in combination with each other. Everolimus was determined in whole blood and metformin was determined in serum using high-performance liquid chromatography with tandem mass spectrometry (HPLC-MS/MS). According to conventional pharmacokinetic principles [28], the area under the curve (AUC) of the serum concentration curves was calculated as the AUC from t = 0 to infinity for the first administration of a study drug, i.e., single-agent drug administration at day −1 for metformin, and as the AUC from t = 0 to t_last for steady-state drug administration, i.e., at days 7 and 15 for everolimus (as single-agent and in combination with metformin, respectively) and at day 15 for metformin when it was combined with everolimus. T_last was the time point directly preceding the next dose for that agent, i.e., 12 h for metformin bid and 24 h for everolimus qd.

Tumor responses were calculated as the relative difference between the volume of the target lesions at the radiological assessment during study inclusion and the radiological assessment that showed the best overall response. The cut-off marks of the total sum of the volume of the target lesions for progressive disease, stable disease and partial response were +20% and −30%, respectively. Complete response was defined as a disappearance of all target lesions.

Descriptive statistics were used for evaluation of the adverse events, safety and efficacy of everolimus and metformin. The pharmacokinetic parameters were calculated using PKSolver [29]. All P values were calculated using the R statistical programming language (paired Student’s t-test).

In total, 12 patients were screened for eligibility between January 2013 and February 2014. Three patients did not meet the study inclusion criteria; one because of inadequate renal function, one because of inadequate bone marrow function, and one because of dysphagia. Table 2 summarizes the baseline characteristics of the nine eligible patients that were enrolled in the study. All patients had received one or more lines of treatment prior to study enrolment. Six patients received at least four weeks of study treatment and were thus evaluable for toxicity assessments (Table 3). Patients stayed on the study for a median duration of 48 days (range: 4–78). Two patients temporarily interrupted treatment: in one case due to trombocytopenia and in one case due to skin rash. Following treatment interruption, these patients received a 50% dose reduction of everolimus and metformin, respectively.

Nine patients started the study regimen consisting of one week of everolimus followed by the addition of metformin. Of the eight evaluable patients that received everolimus and metformin, five patients discontinued their study participation due to toxicity reasons and three patients because of progressive disease. Of the three patients who entered the study at the starting dose level of 10 mg everolimus qd and 500 mg metformin bid, one experienced a DLT (thrombocytopenia). The cohort receiving dose level 1 was then expanded, where after two more patients experienced a DLT (one case of thrombocytopenia and one case of skin rash). According to the protocol, the dose level was de-escalated to 5 mg everolimus qd and 500 mg metformin bid, at which no more DLTs were observed, but the toxicity of study treatment persisted (Tables 3 and 4). Subsequently, patient enrollment in the study was terminated. Regarding SAEs, one patient had to be hospitalized for a sepsis, one patient had to undergo surgery for a bile duct stenosis and obstructed esophagus (PTC drain and stent biliary duct, neo-gastric tube and stent) and one patient had to undergo an intervention for cholangitis and a liver abscess (PTC drain revised and abscess drainage).

The treatment-related CTC-graded adverse events per dose level are listed in Table 4. Regarding severe (grade 3/4) adverse events, one patient with a grade 4 sepsis, two patients with grade 3 thrombocytopenia, one patient with a grade 3 rash, one patient with a grade 3 melena and one patient with a grade 3 hyperbilirubinemia were observed. The most frequently reported clinical toxicities of any grade included epistaxis, fatigue (both 33%), anorexia, coughing, dry skin, mucositis, skin rash and stomatitis (all 22%). The combination of everolimus and metformin did not induce significant changes in hematological or clinical chemistry parameters. Interestingly, glucose and insulin levels increased instead of decreased during study treatment (Supplementary Table 1 ).

Concentration-time curves of everolimus and metformin, both as single agents and in combination, were obtained from seven patients and are shown in Fig. 1 . The observed pharmacokinetic parameters for patients with complete concentration time-curves are shown in Table 5 and Supplementary Table 2 . An increased everolimus clearance was observed when it was administered in combination with metformin than as single agent (clearance: 0.018 l/h vs. 0.0097 l/h; P = 0.0018), but this did not affect everolimus exposure (area under the curve, AUC). No difference in clearance was detected between metformin administration as single agent or in combination with everolimus (61.1 l/h vs. 59.9 l/h, P = 0.86). Nevertheless, the t_1/2 and elimination rate constant (λ_z) of metformin were increased when co-administered with everolimus as compared with single-agent administration (t_1/2: 7.95 h vs. 4.16 h; P = 0.017 and λ_z: 0.17 h⁻¹ vs. 0.09 h⁻¹; P = 0.005). The metformin AUC was comparable between single-agent administration and co-administration with everolimus. There was no association between the extent of pharmacokinetic interaction between everolimus and metformin and the occurrence of toxicity in the individual patients. Three patients (#5, #8 and #9) had a higher exposure to metformin when administered in combination with everolimus than as single agent due to slower drug elimination (Supplementary Figs. 1 and 2), but patients #5 and #8 experienced only minor toxicity and discontinued study treatment due to progressive disease. Although patients #8 and #9 had incomplete concentration-time curves for metformin, they had an evidently increased metformin exposure when administered in combination with everolimus. However, due to the incompleteness of their data, they could not be included in the quantitative pharmacokinetic analyses shown in Table 5.

After nine weeks of treatment, three patients were still on study and all had stable disease, with tumor responses of −14%, +3% and −5% of the volume of all target lesions. The three patients that received nine weeks of treatment had a longer survival than the five patients who had to discontinue study treatment prematurely due to toxicity (median overall survival: 184 days vs. 82 days; P = 0.0482, Table 3). Of the three patients that received nine weeks of treatment, two eventually discontinued study treatment due to progressive disease and one because of toxicity reasons (Table 3). The nine-week time point for this survival analysis was prespecified because patients were to receive a CT scan to evaluate tumor responses at this time point.