A phase 2 study of the insulin-like growth factor-1 receptor inhibitor MK-0646 in patients with metastatic, well-differentiated neuroendocrine tumors


  • Presented as a poster at the 46th Annual Meeting of the American Society of Clinical Oncology; June 4-8, 2010; Chicago, Illinois.



Neuroendocrine tumor (NET) cell lines frequently express both insulin-like growth factor (IGF) ligand and the cognate IGF-1 receptor (IGF-1R) and, as such, potentially depend on the activation of IGF-1R and its downstream effectors for growth and survival. Preclinical studies suggest that somatostatin analogs and mammalian target of rapamycin (mTOR) inhibitors exhibit antitumor activity against NETs through inhibition of IGF-1-dependent signaling, suggesting that IGF-1R inhibition may be a promising therapeutic approach to NETs. Therefore, the authors of this report evaluated the safety and efficacy of MK-0646, a fully human monoclonal antibody (MoAb) that binds to the IGF-1R, as monotherapy in patients with metastatic, well-differentiated NETs.


A phase 2 study was performed in which patients received intravenous MK-0646 10 mg/kg once weekly over 1 hour. Archived pretreatment tumor tissue was obtained and genotyped for v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS), phosphatidylinositol-3-kinase, catalytic, alpha polypeptide (PIK3CA); and v-raf murine sarcoma viral oncogene homolog B1 (BRAF) mutations, and immunohistochemistry was performed to measure the expression IGF-1R.


Twenty-five patients received treatment (40% women; median age, 61 years; age range, 37-83 years), including 15 patients with carcinoid tumors and 10 patients with pancreatic NETs. No partial or complete responses were observed. The median progression-free survival was 4.2 months in the pancreatic NET cohort (range, 0.7-6.7 months) and 2.7 months in the carcinoid cohort (range, 2-3 months). Serious adverse events that were potentially related to MK-0646 included grade 3/4 hyperglycemia in 8 of 25 patients (32%), grade 2 hypersensitivity reaction in 1 of 24 patients (4%), and grade 3 lipase elevation in 1 of 25 patients (4%).


Despite a compelling preclinical rationale, MK-0646 was inactive as a single agent in well-differentiated NETs. Further studies of MK-0646 as a monotherapy in unselected NETs are unwarranted. Cancer 2012. © 2012 American Cancer Society.


The type I insulin-like growth factor receptor (IGF-1R) is a member of a family of transmembrane tyrosine kinases that includes the insulin receptor and the orphan insulin receptor-related receptor.1 Many preclinical and clinical studies have implicated the IGF-1R and its ligands, IGF-1 and IGF-2, in the development and progression of cancer.2-4 The principal pathways for transduction of the IGF signal are the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase/v-akt murine thymoma viral oncogene homolog 1 (PI3K/Akt) pathways.2 MK-0646 is a humanized immunoglobulin G1 (IgG1) kappa antibody that targets IGF-1R. Signaling through IGF-1R mediates cell growth and proliferation as well as resistance to apoptosis in many solid tumors.5 There are 2 possible mechanisms of action of MK-0646: 1) inhibition of IGF-1–mediated cell signaling and 2) antibody-dependent, cell-mediated cytotoxicity.

Preclinical in vitro studies suggest that somatostatin analogs and mammalian target of rapamycin (mTOR) inhibitors exhibit antitumor activity against neuroendocrine tumors (NETs) through the inhibition of IGF-1-dependent signaling.6, 7 The antitumor activity of the latter, however, may be attenuated by the induction of AKT activity upon CREB-regulated transcription coactivator 1 (TORC1) inhibition through the relief of feedback inhibition of upstream receptor tyrosine kinases like IGF-1R. NETs and NET cell lines frequently express both IGFs and the IGF-1R receptor and, as such, potentially depend on autocrine or paracrine stimulation by this pathway for growth and survival.8-11 In a carcinoid tumor cell line (BON), it has been demonstrated that exogenous IGF activates mTOR, increases cellular proliferation, and increases secretion of chromogranin-A, which is a marker protein of neuroendocrine progression.9 Thus, we hypothesized that upstream inhibition at the level of the IGF-1R receptor may have greater antitumor activity than rapalogues in patients with advanced NETs. Therefore, we conducted a phase 2 trial of the anti-IGF-1R monoclonal antibody (MoAb) MK-0646 in patients with advanced NETs.


This was a single-institution, investigator-initiated phase 2 trial in patients with metastatic NETs. The trial was approved by the Institutional Review Board (IRB) and was conducted in accordance with US Department of Health and Human Services guidelines. Written informed consent was obtained from each patient The trial is registered at as National Clinical Trial NCT00610129.

Patient Selection

Eligible patients were required to have histologic or cytologic confirmation of well-differentiated NET with measurable, metastatic disease according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria. There was no limit on the number of prior treatment regimens. Patients were required to have an Eastern Cooperative Oncology Group performance status of 0 or 1, they had to be aged ≥18 years, and they had to have a life expectancy >3 months. Adequate bone marrow and kidney function were required, as was a bilirubin level <1.5 times the upper limit of normal. Patients were excluded if they had received any prior IGF-R–directed agents or if they had inadequately controlled diabetes mellitus, defined as a fasting blood glucose level >120 mg/dL or a hemoglobin A1c level >8%. Patients were allowed to continue on somatostatin analog therapy. The decision to stop the somatostatin analog before initiating MK-0646 was at the discretion of the treating physician. All patients who had hormone symptoms (ie, functional) continued on octreotide therapy.


Patients received MK-0646 as an intravenous, 1-hour infusion at a dose of 10 mg/kg. Treatment was continued weekly until patients developed disease progression or unacceptable toxicity, as defined below, or until the patient withdrew consent. The National Cancer Institute Common Adverse Event Criteria (NCI-CTCAE) version 3.0 was used to grade adverse effects. For all grade 3 or 4 toxicities that were possibly or probably related to MK-0646, treatment was withheld until symptoms returned to baseline or to grade 2 with treatment; then, the patient was restarted with the dose reduced by 7.5 mg/kg and 6 mg/kg for grade 3 or 4 toxicity, respectively.

For the assessment of hyperglycemia (diabetes), NCI-CTCAE version 3.0 endocrine terminology was used. For grade 3 hyperglycemia (“symptoms interfering with activities of daily living” or “insulin indicated”), MK-0646 was withheld if symptoms were present or if the serum glucose level was ≥300 mg/dL; MK-0646 was resumed when the patient was asymptomatic, glucose was consistently <220 mg/dL, and the patient was on a stable insulin regimen. In such cases, the MK-0646 dose was reduced to 7.5 mg/kg. For grade 4 hyperglycemia, MK-0646 therapy was withheld until the patient was asymptomatic, glucose was consistently <220 mg/dL, and the patient was on a stable insulin regimen. When treatment resumed, the MK-0646 dose was reduced to 6 mg/kg.

Evaluation Criteria

Contrast-enhanced computed tomography scans of the chest, abdomen, and pelvis with liver triphasic imaging were obtained at baseline and every 12 weeks. Responses were categorized according to standard RECIST (version 1.0) criteria. Response and progression were determined by a single reference radiologist. The primary objective of the study was to evaluate the effect of MK-0646 on the objective response rate. The secondary objective was evaluation of progression-free and overall survival.

Correlative Laboratory Analysis of Tissue Samples

Immunohistochemical analysis

Pretreatment archival tissue was collected and assayed for IGF-1R expression by immunohistochemistry (IHC). IHC for IGF-1R was performed using rabbit polyclonal antibodies to IGF-1R (Ventana Medical Systems, Inc., Tucson, Ariz). The staining was scored based on the intensity (from 0 to 3+) and the percentage of stained tumor cells.


For genotyping studies, hematoxylin and eosin-stained slides were reviewed to ensure adequate tumor content by a reference pathologist before DNA extraction. Sections (5 to 6 μm thick) cut from paraffin-embedded tumor samples were available for 10 of 25 patients and were evaluated for v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS), v-raf murine sarcoma viral oncogene homolog B1 (BRAF), and phosphatidylinositol-3-kinase, catalytic, alpha polypeptide (PIK3CA) mutations using a chip-based, matrix-assisted, laser desorption-time-of-flight (MALDI-TOF) mass spectrometry-based assay (Sequenom, San Diego, Calif) and confirmed with Sanger sequencing, the details of which have been published previously.12-14 An additional 15 well-differentiated, low-grade NET samples from our NET registry (Protocol 08-105) also were included for genetic analysis after IRB waiver of approval was provided.

Statistical Considerations

The initial planned sample size for the first stage of a Simon 2-stage design was a total of 36 patients (18 carcinoid and 18 pancreatic NETs). Because of poor accrual, the study statistics were modified so that, if 2 or more (of the first 15 patients) in each stratum had an objective response in an arm, then enrollment would be extended to 18 patients. If 3 or more confirmed major responses were observed of 18 patients in a stratum, then enrollment would be extended to a total of 44 patients in each stratum. On the basis of a Simon 2-stage design with a null response rate of 13% and an alternative of 30% and with type I and type II error rates of 10%, at the end of the stage 2, a regimen was to be considered promising if ≥9 of 44 treated patients in that arm experienced an objective response. Overall and progression-free survival curves were estimated using the Kaplan-Meier method. Because of poor accrual in the pancreatic NET cohort, that arm was closed after enrolling 10 patients. This was done because the probability of observing no responses in 10 patients was 0.03 assuming an objective response rate of 30%, which made it very unlikely that promising activity in the first stage would have been observed had the study accrued to the prespecified enrollment of 15 patients in the first stage.


Patient Characteristics

Twenty-five patients were treated on this trial. Patient characteristics are outlined in Table 1. Overall, 18 of 25 patients (72%) had received prior somatostatin analog therapy. All patients had documented disease progression within 6 months of starting therapy with MK-0646. The median number of systemic treatments before initiating therapy was 1, and approximately 75% of the patients had received 1 or 2 systemic regimens before enrollment.

Table 1. Patient Characteristics
CharacteristicNo. of Patients (%)
CarcinoidPancreatic NET
  1. Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; NET, neuroectodermal tumor; VIPoma, vasoactive intestinal peptide tumor.

Total no. of patients15 (100)10 (100)
 Men8 (53.3)7 (70)
 Women7 (46.7)3 (30)
Age: Median [range], y63.9 [37-83]59.1 [48-70]
Time since initial diagnosis [range], mo131.75 [137.83-6.08]49.24 [50.23-.99]
Previous octreotide9 (60)9 (90)
Primary diagnosis
 Foregut: Lungs, stomach6 
 Midgut: Small bowel, appendix5 
 Hindgut: Colon, rectum3 
 Functioning (hormone-secreting)3 
 Nonfunctioning (nonhormone-secreting)2 
 Gastrinoma 1
 Insulinoma 1
 VIPoma 0
 Glucagonoma 1
 Other 0
 Nonfunctioning 7


All 25 patients were assessable for toxicity. Table 2 lists the toxicities of MK-0646 that were possibly, probably, or definitely attributable to the drug. The most common adverse event that potentially was related to MK-0646 was hyperglycemia (8 of 25 patients [32%]; grade 3-4). Twenty-four of 25 patients who had data available had laboratory glucose elevations at least 1 NCI-CTCAE grade above baseline at some point during the study. None of the hyperglycemia events had associated clinical symptoms. Other severe adverse events included 1 of 25 patients (4%) patient with grade 3 lipase elevation and 1 of 25 patients (4%) with grade 2 hypersensitivity reaction; the latter patient was given diphenhydramine and received subsequent treatments without complication.

Table 2. Reported Toxicities Possibly, Probably, or Definitely Related to MK-0646
Toxicity Attributed to MK-0646No. of Patients (%)
Grade 1 fatigue3 (12)
Grade 2 fatigue2 (8)
Grade 2 hypersensitivity reaction1 (4)
Grade 1 skin rash1 (4)
Grade 1 polyuria1 (4)
Grade 1 metallic taste1 (4)
Grade 3 lipase1 (4)
Grade 1 muscle ache1 (4)
Grade 1 diarrhea1 (4)
Grade 1 hyperglycemia24 (96)
Grade 2 hyperglycemia17 (68)
Grade 3 or 4 hyperglycemia8 (32)


Twenty-eight patients were registered, although 3 were deemed ineligible and were never treated. Of the 25 patients who were eligible for the study, all received at least 1 dose of therapy and were assessable for response and toxicity. The median number of treatment cycles was 2.76 (range 1-7 cycles; median, 4 cycles). No patient had an objective response. The median progression-free survival was 4.2 months in the pancreatic NET cohort (range, 0.7-6.7 months) and 2.7 months (range, 2-3 months) in the carcinoid cohort. Kaplan-Meier curves illustrating progression-free and overall survival curves are provided in Figures 1 and 2, respectively. At a median follow-up of 18 months, the median overall survival in the carcinoid study arm was 10.5 months, and the median overall survival was not yet reached in the pancreatic NET cohort.

Figure 1.

These Kaplan-Meier estimates illustrate progression-free survival (PFS) for the intention-to-treat population.

Figure 2.

These Kaplan-Meier estimates illustrate overall survival for the intention-to-treat population.

Correlative Assessments

Immunohistochemistry testing

Fourteen pretreatment formalin-fixed, paraffin-embedded tissue samples were assayed for IGF-1R expression by IHC. Three of 9 carcinoid tumors demonstrated 1+ IGF-1R expression, whereas the other 6 carcinoid tumors were negative for IGF-1R by IHC. One of 4 pancreatic NETs was 1+ positive for IGF-1R expression.


We used a chip-based MALDI-TOF mass spectrometry assay (Sequenom) to genotype 25 samples for hotspot mutations in KRAS, BRAF, neuroblastoma v-ras oncogene homolog (NRAS), and PIK3CA. A single pancreatic NET harbored a glutamic acid to lysine mutation of codon 542 (E542K) in the helical domain of PIK3CA. Because PIK3CA mutations had not been reported previously in patients with pancreatic NET at the time of our analysis, we further assayed an additional 15 NETs from our tissue bank for PIK3CA mutations. A second tumor (a presumed small bowel carcinoid) also was identified that harbored an identical E542K PIK3CA mutation for a total frequency of 8%. No kinase domain mutations in PIK3CA were noted. All tumors were also wild-type for KRAS, BRAF, and NRAS.


Given the promising preclinical data on IGF-1R–directed therapies and its unequivocal activity with mTOR inhibition, why was there no evidence of antitumor activity? In other solid tumors, mutations downstream of the directed therapy confer resistance to the upstream targeted therapy; antiepidermal growth factor receptor-directed therapies, for example, are restricted to patients with KRAS wild-type tumors.15, 16 Nevertheless, in all of our patients, the tumors were wild-type KRAS and BRAF tumors; therefore, we cannot explain this lack of activity. It is noteworthy that 2 of the 25 tumors tumors (8%) were identified with PIK3CA mutations (E542K in both tumors). This mutation rate is higher than the 1.4% rate recently reported last month by Jiao et al.17 In vitro studies have indicated that tumor cells may be even more sensitive to mTOR inhibitors in the presence of PIK3CA mutations. The presence of these mutations, however, failed to produce a response to MK-0646 in our patients.18

A second possible explanation for the lack of efficacy of MK-0646 observed in this trial is that IGF-1R may not have been highly activated in the patients enrolled. To address this possibility, we determined IGF-1R expression levels using IHC. In that analysis, we observed modest staining (1+) for IGF-1R in only 4 of 17 tumors. It should be highlighted that total IGF-1R IHC expression may be a poor indicator of the degree of IGF-1R activation. A review of these patients suggested that they did not fare better than those who did not express IGF-1R.

A third explanation may be that we did not hit the target. Phase 1 data, however, demonstrate evidence of IGF-1R signal blockade in tumors. In that study, tumor biopsies before and after treatment were obtained for pharmacodynamic analyses. Thirteen patients had a decrease in IGF-1R expression and inhibition of IGF-1R signaling (decreases in phosphorylated AKT, phosphorylated MAPK, and phosphorylated S6 protein), and a decrease in Ki-67 was observed at doses >5 mg/kg.19 In addition, many of our patients developed hyperglycemia, which we believe indirectly indicates that the target is being hit. A fourth explanation is that these were heavily treated patients who were exposed to other targeted agents, leading to the treatment of more resistant tumors with MK-0646. However, as noted above, the median number of treatments received before enrollment was 1, and no patients were exposed to mTOR or vascular endothelial growth factor inhibitors. In addition, 80% of our patients went on to receive additional therapy after disease progression, and 50% of patients underwent elective hepatic embolization.

In summary, despite uncommonly compelling preclinical data, our study failed to demonstrate that targeting a pathway upstream of mTOR by binding to the IGF-1R with MK-0646 was clinically helpful. Further studies evaluating combinations of, perhaps, dual mTOR and anti-IGF-1R inhibition as well as other combinations of targeted therapies, including antiangiogenic agents, may be warranted to further explore for possible uses of anti IGF-1R strategies in advanced NETs.

A lesson reiterated by this clinical trial is the large heterogeneity of the disease. To compensate for this, we only included patients who had progressive disease, and we believe the short progression-free survival reflects the natural history of this later phase of the disease. All patients had an Eastern Cooperative Oncology Group performance status of 0 or 1, however, and 80% went on to receive additional therapy after progressing on this trial: All of this suggests that their disease was not so far advanced that we could have missed a signal. Consistent eligibility and clearly defined patient populations like the population in the current trial are critical for future studies. We also believe that pretreatment and post-treatment biopsies should be obtained in future trials whenever possible. To date, the paucity of correlative studies in NET clinical trials has limited our ability to develop a basic scientific understanding of the relevant tumor signaling mechanisms. This negative trial also highlights the need for multisite studies to continue efforts that explore new targets in the treatment of this disease.


This study was supported in part by Merck, Incorporated. Correlative studies were supported by the 2008 American Society for Clinical Oncology Young Investigator Award (Principal Investigator, Diane Reidy-Lagunes).


The authors made no disclosures.