A Phase I Study of Cixutumumab (IMC-A12) in Combination with Temsirolimus (CCI-779) in Children with Recurrent Solid Tumors: A Children's Oncology Group Phase I Consortium Report

Deregulated insulin-like growth factor signaling through the type 1 receptor is common to many childhood solid tumors (1), including Ewing sarcoma (2), rhabdomyosarcoma (3), osteosarcoma (4, 5), neuroblastoma (6–8), brain tumors (9–13), Wilms tumor (14), and hepatoblastoma (15). Similarly, the cell signaling pathways that activate the mTOR are altered in many pediatric cancers, making them susceptible to mTOR inhibition by sirolimus and its analogues (16–18). Although we have previously shown preclinical activity of mTOR inhibitors in pediatric cancers (17), monotherapy with these agents has had limited clinical activity (16, 19). This may, in part, be explained by a feedback loop in which inhibition of mTOR signaling leads to upregulation of IGFIR signaling through loss of S6K1 phosphorylation, leading to activation of PI3K and Akt (20–22). Indeed, Akt activation has been demonstrated in cancer cell lines and in patients treated with mTOR inhibitors (23). IGFIR inhibition should prevent rapamycin-induced Akt activation, hence abrogating at least one survival pathway and synergizing with mTOR inhibitors (20, 21, 24).

Cixutumumab (IMC-A12) is a fully recombinant IgG1 monoclonal antibody that specifically targets the human IGFIR1 with high affinity. It acts as an antagonist of IGFI and IGFII ligand binding and signaling. It also blocks ligand binding to IGFIR and inhibits downstream signaling of the two major insulin-like growth factor pathways: MAPK and PI3K/AKT. Cixutumumab has potent dose-dependent in vitro and in vivo antitumor activity in a variety of cell lines and xenografts.

Temsirolimus is a small-molecule inhibitor of mTOR. Like sirolimus and everolimus, temsirolimus forms a gain-of-function complex with FK506-binding protein 12 (FKBP12) that binds and inhibits mTOR, leading to antiproliferative effects, including G₁-phase cell-cycle arrest (25), and apoptosis. The primary downstream targets of mTOR include eIF4E-binding protein (4E-BP1; refs. 26, 27) and p70S6 kinase, important in the translation regulation of mRNA encoding proteins involved in G₁ phase progression. mTOR inhibitors have potent in vitro activity against many human cancer cell lines and xenograft models. The Pediatric Preclinical Testing Program (17, 28) and others have reported preclinical single agent and synergistic combination activity of these agents in many solid tumors (24, 29, 30).

We report the results of a phase I trial of cixutumumab in combination with temsirolimus in children with recurrent or refractory solid tumors. The primary objectives were to estimate the MTD, determine dose-limiting toxicities (DLT), and characterize the pharmacokinetics of IMC-A12 and temsirolimus administered once weekly in combination, to children with refractory solid tumors. The secondary objectives were to assess the biologic activity of temsirolimus by measuring levels of phospho-S6^Ser235/236, phospho-AKT^Ser473, and phospho-4EBP1^Ser65 in peripheral blood mononuclear cells (PBMC).

Patients >12 months and <22 years with measurable or evaluable solid tumors refractory to therapy were eligible. Histologic verification of malignancy was required except for patients with intrinsic brainstem glioma. Other eligibility criteria included Lansky or Karnofsky score ≥50; recovery from the acute toxic effects of prior therapy; ≥3 months since total body irradiation, craniospinal or hemi-pelvic radiation, and ≥2 months since a stem cell transplant; adequate bone marrow function [peripheral absolute neutrophil count (ANC) ≥ 1,000/μL, platelets ≥ 100,000/μL (transfusion independent), hemoglobin ≥ 8.0 g/dL]; adequate renal function (age-adjusted normal serum creatinine or a glomerular filtration rate ≥ 70 mL/min/1.73 m²); adequate liver function [total bilirubin ≤ 1.5× institutional upper limit of normal for age, SGPT (alanine aminotransferase, ALT) ≤ 5× institutional upper limit of normal for age and albumin ≥ 2 g/dL]; prothrombin time and international normalized ratio < 1.2× upper limit of normal. Patients receiving corticosteroids had to be on a stable or decreasing dose for ≥7 days before study enrollment.

Patients were excluded if they had known bone marrow involvement; had received prior temsirolimus or monoclonal antibody therapy targeting IGFIR; were pregnant or lactating; had an uncontrolled infection; were receiving enzyme inducing anticonvulsants (EIACD), insulin, growth hormone therapy, or any of the following CYP3A4 inducers or inhibitors: erythromycin, clarithromycin, ketoconazole, azithromycin, itraconazole, grapefruit juice, or St. John's wort or other noncytotoxic anticancer agents. Also excluded were patients with a history of allergic reactions attributed to compounds of similar chemical or biologic composition to cixutumumab or temsirolimus, or patients who had undergone major surgery within 6 weeks before study enrollment. The Institutional Review Boards of participating institutions approved the protocol. Informed consent and assent, as appropriate, were obtained according to local institutional guidelines.

Cixutumumab [5 mg/mL (250 mg) or 10 mg/mL (500 mg) vials] and temsirolimus (25 mg/mL), supplied by the Cancer Therapy Evaluation Program (NCI, Bethesda, MD), were administered weekly as a 1 hour and 30 minute infusion, respectively, in 28-day cycles. In the absence of disease progression, patients could receive a total of 25 cycles of therapy.

The starting dose of cixutumumab was 6 mg/kg (one dose level below the recommended phase II dose of 9 mg/kg in the pediatric phase I study; ref. 31) and of temsirolimus was 15 mg/m², with planned dose escalations to a maximum of 9 mg/kg for cixutumumab and 35 mg/m² for temsirolimus. The starting dose of temsirolimus was based on published studies in adult combination trials (32). Because of toxicity, the study was amended to incorporate several dose de-escalations as shown in Table 1.

Toxicities were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) v3.0. (http://ctep.cancer.gov/protocolDevelopment/electronic_applications/docs/ctcaev3.pdf). Hematologic DLT was defined as grade 4 neutropenia for >7 days, or before drug administration on days 8, 15, or 21; platelet count <20,000/mm³ on 2 separate days or requiring platelet transfusions on 2 separate days within a 7-day period, or myelosuppression that caused a delay of ≥14 days between treatment cycles. Nonhematologic DLT was defined as grade 3 or 4 nonhematologic toxicity with the specific exclusion of grade 3 nausea and vomiting of <3 days duration, grade 3 transaminase elevations that returned to grade ≤ 1 or baseline before the time for the next treatment cycle, grade 3 fever or infection, nonhematologic toxicity that caused a delay of ≥14 days between treatment cycles.

Any patient who experienced DLT at any time during protocol therapy was considered fully evaluable for toxicity. Patients without DLT who received at least 85% of the prescribed dose during the first 28-day cycle were considered evaluable for toxicity as well as response. Dose escalation proceeded using a modified 3 + 3 cohort design in which 3 patients were initially studied at each dose level. If none of these 3 patients experience DLT, the dose was escalated to the next higher level. If one of three patients experienced DLT, then up to three more patients were accrued at the same dose level. If none of these three additional patients experienced DLT, then the dose was escalated. If one or more of these three additional patients experience DLT, the MTD was exceeded, unless one of the DLTs did not appear to be related to dose or the DLTs were of different classes and the toxicities were readily reversible. In that circumstance, the cohort could be expanded to 12. The MTD was the maximum dose at which fewer than one third of patients experienced DLT. Once the MTD for the combination was defined, up to 6 patients with recurrent or refractory solid tumors were to be enrolled to acquire additional pharmacokinetic data in patients <12 years.

Disease evaluations were obtained at baseline, at the end of cycle 1, and after every other cycle. Tumor response was reported using the RECIST (33).

Blood samples (5 mL) for pharmacokinetic analysis of cixutumumab were obtained before administration on days 1, 8, 15, 22, and 28 of cycle 1, and days 15 and 28 of cycle 2. In patients who agreed to have additional pharmacokinetic sampling, 3 mL of blood were collected at the end of the day 1 infusion, and 1, 3, 6, and 24 (±2) hours, and 4 days (±1 day) after the first dose. All blood samples were collected at a site distant from the infusion, in tubes without anticoagulant, allowed to clot at room temperature, centrifuged at 1,500 rpm for 15 minutes, and stored at less than −20°C. A validated ELISA was used to quantify levels of cixutumumab as previously described (31).

Blood samples (1 mL) for pharmacokinetic analysis of temsirolimus were obtained before temsirolimus administration on days 1 and 28 of cycle 1, and days 15 and 28 of cycle 2. In patients who agreed to have additional pharmacokinetic sampling, blood was collected prior at the end and at 15 minutes, 30 minutes, 1, 3, 6, and 24 (±2) hours after the infusion on day 1. All blood samples were collected at a site distant from the infusion.

Temsirolimus and sirolimus concentrations were determined using validated isotope-dilution high-performance liquid chromatography–tandem mass spectrometry assays based on previously described principles (34).

The maximum plasma concentration (C_max) and time to maximum plasma concentration (T_max) were determined by visual inspection of the plasma concentration-time profiles for subjects participating in the extended pharmacokinetic sampling. Temsirolimus clearance (CL), steady-state volume of distribution (Vd_ss), half-life (t_1/2), and area under the plasma concentration versus time curve (AUC_{0–24 and} AUC_0–∞) were analyzed using standard noncompartmental methods (NCA; Phoenix WinNonlin, Pharsight Corp). Sirolimus pharmacokinetic parameter estimates for t_1/2 and AUC_0–∞ were generated using model-based Bayesian analysis (MW/Pharm; Version 3.60, Mediware; ref. 35).

Whole blood (4 mL) was collected in BD Vacutainer CPT sodium citrate tubes before administration of protocol therapy on days 1, 8, and 28 of cycle 1 in consenting patients. Within 1 hour of acquisition, PBMC lysate was prepared as described (36), snap-frozen in liquid nitrogen, and stored at −80ºC until phosphoprotein analysis. Phosphorylated and total S6K1, Akt, and 4E-BP1 detection and quantification were performed via Western blot analysis followed by densitometric analysis with ImageQuant 5.0 (Molecular Dynamics). The ratio of phosphorylated protein to total (phosphorylated and unphosphorylated) protein in each patient sample was then calculated.

Of 39 patients enrolled, all were eligible, six were not fully assessable for toxicity: two never received protocol therapy because of clinical deterioration, two experienced progressive disease during cycle 1, and two withdrew consent for non-DLTs [grade 2 ALT/AST (aspartate aminotransferase; n = 1) and grade 2 mucositis (n = 1)]. Table 2 summarizes the characteristics of the eligible patients. Patients received a median of 2 cycles (range, 1–20) of therapy.

Table 1 summarizes the DLTs across all dose levels. At the starting dose [cixutumumab (6 mg/kg); temsirolimus (15 mg/m²)], one patient in the initial cohort of three had dose-limiting hypercholesterolemia (grade 3). The cohort was expanded to include three more evaluable patients, one of whom had dose-limiting mucositis (grade 3). Because the observed DLTs observed were readily reversible, of different classes and unrelated to dose, the statistical design allowed enrollment of up to six additional patients at this dose level. Three different DLTs were observed among this expanded cohort of 6 (grade 3 ALT, fatigue, and prolonged thrombocytopenia >14 days), resulting in a dose de-escalation. At dose level 0 [cixutumumab (6 mg/kg); temsirolimus (10 mg/m²)], 1 of 3 patients experienced grade 3 dose-limiting mucositis. Three additional patients were enrolled, one of whom experienced grade 3 dose-limiting mucositis. As a result, the protocol was amended to incorporate dose level −1 in which the doses of both agents were further decreased [cixutumumab (4 mg/kg) and temsirolimus (8 mg/m²)]. None of the patients at dose level −1 experienced a DLT. An intermediate dose level [cixutumumab (6 mg/kg) and temsirolimus (8 mg/m²)] was subsequently added and one of six patients experienced grade 3 mucositis; none of 6 patients experienced a DLT among an expanded cohort for children <12 years.

Table 3 summarizes all non-DLTs at least possibly attributable to cixutumumab or temsirolimus in the 33 patients fully assessable for toxicity.

One partial response, confirmed on central review, was observed following cycle 1 in a patient with medulloblastoma enrolled at dose level 1. Unfortunately, tumor progression was noted on surveillance imaging studies performed after cycle 4. Stable disease for ≥3 cycles was observed in 3 patients (median, 12 cycles; range, 11–20) with epithelioid sarcoma, neuroblastoma, and alveolar rhabdomyosarcoma, respectively.

Cixutumumab single-dose pharmacokinetics were studied in 19 patients, and steady-state pharmacokinetics were studied in 36 patients who received the combination of cixutumumab with temsirolimus. There were no changes in cixutumumab clearance as the temsirolimus dose was increased (Table 4). The median weight-adjusted clearance and half-life values for all patients were 0.17 mL/hr/kg (range, 0.07–0.43 mL/hr/kg) and 122 hours (range, 64–446 hours), respectively. Cixutumumab serum trough concentrations (C_trough) exhibited marked interpatient variability, increased throughout the first cycle of treatment (Supplementary Fig. S1), and a median C_trough of approximately 200 μg/mL was achieved on day 28 following a 6 mg/kg dose (Supplementary Table S1). The median cixutumumab C_trough at the recommended phase II dose was 197 μg/mL (range, 60–544).

Results of the pharmacokinetic data (n = 19) following the first dose of temsirolimus at 8, 10, and 15 mg/m²/dose are summarized in Table 5. There was marked interpatient and interdose group variability. Sirolimus, the metabolite of temsirolimus, exposure did not appreciably increase with dose. The steady-state volume of distribution (V_ss) was large and increased with dose, ranging from 46.3 L (after 8 mg/m² dose) to 74.9 L (after 15 mg/m² dose). Similarly, temsirolimus clearance from whole blood increased with increasing dose, from 2.3 L/h after an 8 mg/m² dose to 6.7 L/h after a 15 mg/m² dose.

At the recommended phase II dose, a total of 23 PBMC samples were obtained from 9 consenting patients. Four patients had suitable samples from the days 1, 8, and 28 timepoints, three had samples from days 1 and 8, and one had samples from days 1 and 28. Baseline variation in total S6K1 protein was observed among patients, but intrapatient levels were consistent at each time point (days 1, 8, and 28). Phospho-S6K1 (T389) was reduced at day 8 in the seven evaluable patients when compared with baseline (day 1), and in the majority of cases was further reduced at day 28; in only one patient was S6K1 phosphorylation higher at day 28 when compared with baseline. Unlike S6K1, total and phospho-4EPB1 (S65) followed expression levels at each time point, with no observed change in phosphorylation as a function of treatment on days 8 or 28. Akt phosphorylation (S473) was more variable than S6K1, showing a reduction at day 8 in five patients, and slight or no increase in two others. By day 28, Akt phosphorylation in the five evaluable patients remained variable, with evidence of an increase in three patients, and decrease in the other two, when compared with baseline. Figure 1 depicts a Western blot of representative patient PBMC samples showing inhibition of target proteins p-S6K1 T389 and p-AKT S473.

The recommended pediatric phase II doses of the combination of cixutumumab and temsirolimus in children with refractory solid tumor are 6 mg/kg and 8 mg/m², respectively. Similar to adults, observed DLTs included mucositis, fatigue, hypercholesterolemia, and transaminitis. However, the MTD of temsirolimus for this combination in children was significantly lower than observed in adults with advanced cancers (37), relapsed Ewing sarcoma [ref. 38; cixutumumab 6 mg/kg i.v. and temsirolimus 25-mg flat dose (14 mg/m²)], or metastatic prostate cancer [cixutumumab 6 mg/kg i.v. and temsirolimus 20-mg flat dose (11 mg/m²); ref. 39]. It was noted in the prostate cancer trial that the weekly schedule was difficult to maintain long term due to chronic toxicities. Pharmacokinetic data were not available for any of the adult trials. Interestingly, in the adult trials, grade 3 mucositis was observed in 3% to 27% of patients (37–39). In our study, 4 of 27 patients or 15% experienced grade 3 mucositis, leading to repeated dose de-escalations of the temsirolimus and cixutumumab dose.

Although the pharmacokinetics of cixutumumab (31) and temsirolimus (40) have been previously reported in single-agent pediatric studies, this is the first description of the disposition of these drugs when used in combination in children. Cixutumumab clearance, C_max, and half-life at 6 mg/kg when combined with temsirolimus at the MTD of 8 mg/m² were similar to those reported earlier in the pediatric phase I single-agent study of cixutumumab (31). Furthermore, cixutumumab half-life and steady-state trough concentrations of 122 hours and 197 mg/mL, respectively, at 6 mg/kg were similar to the values of 209 hours and 145 mg/mL reported for adults at that same dose (41).

Although there was large interpatient and interdose variability in pharmacokinetic parameter estimates for temsirolimus and its active metabolite sirolimus, the temsirolimus and sirolimus AUCs did not increase significantly with dose, which has previously been reported (42, 43). The steady-state volume of distribution (V_ss) of temsirolimus was generally large, increased with dose, and ranged from 230 L (after a 25-mg dose) to as high as 900 L (after a 250-mg dose). The distribution of temsirolimus into red blood cells appeared to be saturable at higher doses. In addition, the clearance from whole blood increased with increasing dose, as observed with V_ss, with mean values ranging from 20 L/h after a 25-mg dose to 100 L/h after a 250-mg dose (coefficient of variation, approximately 16%–27%; ref. 43). We compared the pharmacokinetic parameters in the pediatric single-agent temsirolimus phase I trial with the current study. Spunt and colleagues (40) reported no MTD for temsirolimus at dose levels ranging from 10 to 150 mg/m². The median temsirolimus C_max and AUC values at the MTD of 8 mg/m² in our study were comparable with median values reported at the 25 and 75 mg/m² dose level reported by Spunt and colleagues By contrast, the sirolimus AUC and C_max were substantially lower in our study at similar doses, with lower sirolimus to temsirolimus ratios at all dose levels tested. The AUCs and C_max of temsirolimus and AUCs and C_min of sirolimus and cixutumumab observed concentrations that did not show an association with toxicity or inhibition of target. The lower MTD observed in this study is possibly pharmacodynamic in nature and related to the combination of IGFIR and m-TOR inhibition, but requires further evaluation in a larger patient cohort. As these inhibitors affect signaling pathways that regulate glucose, lipid, and energy metabolism, it is possible that children may be biologically more sensitive to these drugs, particularly in combination; however, this determination would require further evaluation in a larger patient cohort.

We demonstrated that this combination significantly inhibited the phosphorylation of S6K1, but not 4E-BP1 in PBMCs at all dose levels. Phospho-S6K1 was reduced at day 8 in the 7 patients with evaluable samples, and was further reduced in 4 of 5 patients with samples at day 28. Although S6K1 and 4E-BP1 are both downstream effectors of mTORC1, these data are consistent with reports that in some cells, rapamycin and the rapalogs are incomplete inhibitors of mTORC1, affecting S6K1 phosphorylation more strongly than 4E-BP1 (44). Similar to previous pediatric and adult reports (40, 45), the phosphorylation status of S6K1, 4E-BP1, and Akt did not correlate with objective responses or prolonged stabilization of disease, limiting their use as biomarkers. There was considerable variation in Akt phosphorylation on days 8 and 28, with either an increase or decrease in Akt phosphorylation when compared with baseline. The pAkt variability is consistent with an earlier pediatric temsirolimus study that demonstrated a lack of correlation between temsirolimus dose and phosphorylation of Akt, S6, and 4E-BP1 (40). However, in the current study, phosphorylation trends of the mTORC1 targets S6K1 and 4E-BP1 were consistently decreased or unchanged, respectively. In 5 patients, Akt S473 phosphorylation was generally decreased in response to treatment, paralleling S6K1 phosphorylation and consistent with the inhibition of mTOR complex 2 (mTORC2) observed in cells exposed to long-term treatment with rapamycin (46). This decrease may also reflect the efficacy of cixutumumab, but the contribution of IGFIR signaling to PI3K and Akt may be minor in PBMCs. In two patients, there was an inverse correlation between pS6K1 (decreased) and pAkt (increased), which may be explained by temsirolimus-mediated inhibition of the mTORC1-S6K1–negative feedback loop (47–49), leading to increased PI3K signaling to Akt. The variability of pAkt among patients does not appear to correlate with the magnitude of S6K1 inhibition, but instead may be due to a multitude of factors that are patient-specific, including variability in the makeup of the lymphocytes and monocytes in the PBMC population, and unknown changes in the status of the immune system following prolonged treatment with temsirolimus and cixutumumab.

This study demonstrates that the combination of cixutumumab and temsirolimus can be safely given to children with recurrent solid tumors, albeit at lower doses than in adult combination studies. One patient with a medulloblastoma had a confirmed partial response and 3 patients experienced prolonged disease stabilization for ≥3 cycles. Target inhibition was noted in PBMCs at all dose levels as evidenced by a decrease in phosphor-S6K1 and pAKT. A phase II study of this combination has recently been completed in children with recurrent tumors in the Children's Oncology Group.

D.A. Krueger reports receiving a commercial research grant from, is a consultant/advisory board member for, and reports receiving speakers bureau honoraria from Novartis. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M. Fouladi, J.P. Perentesis, P.J. Houghton, L.A. Doyle, H. Chen, B. Weigel, S.M. Blaney

Development of methodology: M. Fouladi, A.A. Vinks, G. Thomas, P.J. Houghton

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Fouladi, L.M. Wagner, C.A. Mercer, D.A. Krueger, P.J. Houghton, B. Weigel

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Fouladi, A.A. Vinks, J.M. Reid, C. Ahern, C.A. Mercer, D.A. Krueger, S.M. Blaney

Writing, review, and/or revision of the manuscript: M. Fouladi, J.P. Perentesis, L.M. Wagner, J.M. Reid, C. Ahern, C.A. Mercer, D.A. Krueger, P.J. Houghton, L.A. Doyle, B. Weigel, S.M. Blaney

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Fouladi, H. Chen

Study supervision: M. Fouladi, J.P. Perentesis, L.A. Doyle, B. Weigel

Others (population pharmacokinetic modeling): A.A. Vinks

The authors thank Melissa Asman for excellent technical assistance as well as Biljana Georgievska, Thalia Beeles, and Mark Ingle of the COG Phase 1/Pilot Consortium Coordinating Center for outstanding administrative and data management support throughout the development and conduct of this trial.

This work was supported by NCI grant nos. U01 CA97452 and UM1 CA097452-11.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.