The story of insulin-like growth factor 1 receptor (IGF-1R) inhibitors exemplifies the excitement and disappointment in the development of new targeted therapies in oncology. IGF-1R is a part of a complicated system of the insulin-like growth factor (IGF)/insulin family of growth factors comprising of 3 receptors, 3 ligands, and 6 circulating IGF-binding proteins; the IGF-1R part of the system appears to be responsible for the control of apoptosis, cell growth, and differentiation. It is a cell membrane-bound receptor, and it gets activated upon binding of the ligands to the extracellular part; this process leads to the autophosphorylation of tyrosine residues located within the intracellular kinase domain of the receptor and further activation of 2 main signaling cascades: the phosphoinositide 3-kinase/Akt/mammalian target of rapamycin (PI3K/Akt/mTOR) pathway and the mitogen-activated protein (MAP) kinase pathway. Activation of both pathways is heavily regulated; however, when activation is dysregulated, as in many cancers, it can lead to uncontrollable cell proliferation, decreased apoptosis, and increased cell survival.
More than 25 years ago, it was reported that the tumor cells from surgically removed primary breast and colon cancers contained specific IGF-1 binding sites and that monoclonal antibodies that blocked the binding domain of IGF-1 inhibited the growth of breast cancer cell lines both in vitro and in vivo, and that was the first suggestion of the importance of this pathway in cancer biology.[3, 4] Despite the aforementioned findings, there was limited interest in translating these basic research results into clinical practice until the early 2000s. Since then, myriad different molecules have been tested in multiple clinical trials and in multiple types of malignancies. This development was supported by encouraging findings of the preclinical research. The drugs generally can be grouped into 3 categories: monoclonal antibodies blocking the extracellular ligand binding domain of IGF-1R, small molecules serving as tyrosine kinase inhibitors (TKIs) of the intracellular domain, and monoclonal antibodies targeting the IGF ligands.[1, 6] The first group has been studied in the clinic most extensively. The list of tested anti-IGF-1R antibodies includes: AMG-479 (ganitumab; Amgen, Thousand Oaks, Calif), CP-751871 (figitumumab; Pfizer, New York, NY); IMC-A12 (cixutumumab; ImClone, Bridgewater, NJ), MK-0648 (dalotuzumab; Merck, Whitehouse Station, NJ), R1507 (Roche, Basel, Switzerland), and Sch-717454 (Schering-Plough, Kenilworth, NJ). BMS-754807 (Bristol-Myers Squibb, New York, NY) and OSI-906 (linsitinib; Astellas Pharma, Tokyo, Japan) are small-molecule TKIs. Finally, MEDI-573 (MedImmune, Gaithersburg, Md [a subsidiary of AstraZeneca, London, United Kingdom]) and BI836845 (Boehringer Ingelheim, Ingelheim, Germany) block the ligands. Thus, as pharmaceutical companies embarked on a journey through the “dangerous seas” of IGF-1R pathway inhibition testing, the bad news on the clinical activity has been coming: ganitumab produced disappointing activity in patients with breast and pancreatic cancer, and likewise with figitumumab in lung cancer and Ewing sarcoma, dalotuzumab in colorectal cancer, cixutumumab in sarcomas and colorectal cancer, and R1507 in Ewing sarcoma. Based on these results Pfizer, Roche, Amgen, and Bristol-Myers Squibb halted further development of their products.[6, 7] We are still awaiting the results from trials of OSI-906, MEDI-573, and BI836845.
In this issue of Cancer, Pappo et al present results from a phase 2 trial of the IGF-1R blocking antibody R1507 in patients with rhabdomyosarcoma, osteosarcoma, synovial sarcoma, and other soft tissue sarcomas. In the category of sarcoma clinical trials, it was a large study that enrolled 163 patients. The response rate by World Health Organization criteria was the primary endpoint of the study; and, unfortunately, the treatment with R1507 resulted in a disappointing response rate of 2.5%; that is, only 4 patients (2 with osteosarcoma, 1 with embryonal rhabdomyosarcoma, and 1 with alveolar soft part sarcoma) responded to the therapy. In addition, the benefit was short lived, with a median response duration of 12 weeks. The treatment was not associated with significant side effects. These results supported Roche's decision not to continue the development of R1507.
The study was run by the Sarcoma Alliance for Research through Collaboration (SARC), which is a nonprofit organization established by a group of sarcoma physicians who recognized that clinical trials in patients with rare diseases like sarcoma would be successful only when multiple institutions joined their efforts. SARC has served for the last 10 years as a forum in which the new ideas of clinical trials can be introduced, and SARC has also had stronger negotiating power when it comes to discussing drug development with pharmaceutical companies. SARC should be congratulated on conducting this clinical trial, which, after including the Ewing sarcoma part, enrolled 272 patients in less than 2 years, and also on reporting the negative results that hopefully will stimulate further basic research on the IGF-1R pathway and will initiate a discussion on the conduct of trials in sarcoma patients in the era of targeted therapy.
The currently discussed trial is not the only negative trial in patients with sarcoma. A study of figitumumab in 107 patients with Ewing sarcoma demonstrated a response rate of 14.2%; a study of ganitumab in 38 patients with Ewing sarcoma or desmoplastic small round cell tumor demonstrated a response rate of 6%; a study of cixutumumab in 113 patients with a variety of sarcomas demonstrated a response rate of 0.2% and, when cixutumumab was used in combination with temsirolimus, (174 patients) of 5%.
It is unavoidable that such several negative clinical trials with IGF-1R–blocking antibodies must raise several questions. Did we not have enough preclinical and early clinical evidence of the activity of these drugs before launching large clinical trials in sarcoma? To the contrary, we have strong laboratory evidence of the importance of the pathway. Although mutated IGF-1R has not been identified in any cancer, IGF-1R does appear to be up-regulated in alveolar rhabdomyosarcoma, its signaling is required for survival of Ewing sarcoma cell lines, and the growth of osteosarcomas and desmoplastic small round cell tumors can be inhibited by IGF-1R blockade.
Why were all trials with IGF-1R antibodies negative despite strong basic research evidence of the importance of this pathway in cancer biology? It is well known that the trials of targeted therapy were successful when a reliable biomarker existed. Had we not been able to identify amplification of her-2/neu in patients with breast cancer as a biomarker, most probably trastuzumab, a breakthrough development in oncology, would have never been approved and available for these patients. The level of expression of IFG-1R intuitively would appear to be the most feasible biomarker, but the level of IGF-1R was not measured in the trial of R1507. At the same time, other clinical trials attempted to stratify patients based on the presence of IGF-1R in the pretreatment samples, and it did not appear that the level of expression detected by immunohistochemistry correlated with the degree of response. The trial of dalotuzumab included only patients whose tumor samples expressed IGF-1R, and such preselection did not lead to increased efficacy. The trial of R1507, as well as other studies, demonstrated that serum levels of both IGF-1 and insulin increase gradually upon blocking IGF-1R,[1, 6] and this theoretically can lead to increased signaling through the alternate receptors, such as heterodimers of IGF-1R and insulin receptor or currently unknown receptors of negligent biologic activity in the presence of unblocked IGF-1R, but plays an important rescue role when the traditional receptor is blocked.
What can we learn from nonrandomized clinical trials? For the last few decades, the development of new anticancer medications has followed the same pathway, starting from dose-finding phase 1 trials, through signal-finding phase 2 trials, and ending with the efficacy-proving randomized phase 3 trials. Traditionally, an improvement in overall survival has been the most meaningful endpoint of registrational clinical trials; a lot of phase 2 trials have concentrated on measuring a response rate. The new targeted therapies frequently do not result in a significant decrease in the size of measureable tumors; rather, they stop or slow down tumor growth. Therefore, progression-free survival (PFS) has been used more frequently; and new meaningful endpoints like the clinical benefit rate have emerged. In sarcoma, a PFS rate of 30% at 12 weeks has been considered as a benchmark, and the current study of R1507 produced a PFS rate of only 17%. PFS and the clinical benefit rate are heavily influenced by the natural history of the disease and patient selection, and they can be meaningful only when an appropriate control group exists. It is doubtful that the use of historic controls in such a heterogeneous malignancy as sarcoma is appropriate, and a control arm, even in phase 2 trials, should be mandated. In addition, the control arm would allow us to identify a changed toxicity profile. No significant toxicities have been associated with the use of IGF-1R inhibitors, but the results from at least 2 clinical trials demonstrating their detrimental effect on overall survival must not be disregarded.[16, 17]
Does nobody benefit from the therapy with IGF-1R inhibitors? A small number of patients experience significant tumor shrinkage and remain on the therapy for a prolonged time. The study of R1507 by Pappo et al revealed that 22 patients were alive more than 2 years after enrollment, and 1 patient was receiving the treatment for at least 175 weeks. Similar results were observed with cixutumumab, figitumumab, and ganitumab.[10, 11, 18] Tumor histology, expression of IGF-1R, and serum levels of IGF-1 did not predict for the benefit. These findings underscore the importance not only of pharmacokinetic analyses, which were performed very diligently, but also of pharmacodynamic testing and the essential need for tumor tissue analysis upon the drug exposure. The limitations of the currently available tissue testing originate from concentrating on the analysis of isolated signaling pathways and not on system biology. No future studies of IGF-1R inhibitors should be planned unless the currently available tissue biopsies are carefully examined in the basic research laboratories and possible predictors for response have been identified.
The experience with the clinical research of IGF-1R inhibitors is also a lesson about how new drugs should not be developed. At least 4 different anti-IGF-1R antibodies (figitumumab, cixutumumab, ganitumab, and R1507) were tested in larger clinical trials devoted only to sarcoma patients.[9, 11, 12] Those trials enrolled a total of more than 700 patients, ie, more than 700 patients were exposed to inactive (or sporadically active) therapy, the trials did not result in the emergence of new therapies for sarcoma patients, and, moreover, we cannot even predict how to identify rare responders. Several factors influence such random plans for clinical drug development. Multiple pharmaceutical companies develop drugs with the same mechanism of action simultaneously, and certainly there would be a significant incentive for them if their product were approved first. Clinical investigators who enroll patients in trials, provide excellent care, and follow all Good Clinical Practice requirements but are only mentioned in the acknowledgment section do not get any academic credit, which is so important for their advancement. They may be tempted to open smaller clinical trials, which may have a lower chance for success; however, even if the results are negative, they will have a stronger impact on their career. Basic research scientists try to justify development of another compound with the same mechanism of action by pointing out minimal differences and predicting a significant clinical impact. Finally, patients who learn quickly of anecdotal responses of other individuals with the same type of cancer to a new drug apply pressure on clinicians to get access to new, unproven therapies. Can we change it? I believe we can. I know that clinicians, scientists, pharmaceutical companies, and the US Food and Drug Administration work devotedly to improve patients' well being, but they have different tools and different points of view. Therefore, all groups must communicate with each other transparently, exchange ideas selflessly, allow access to their data to other scientists, and encourage patients to contribute their cancer tissue to a repository that has broad access for researchers. Thus, as we develop more targeted therapies based on a deeper understanding of complex oncogenic events and harnessing the system biology that describes crosstalk among multiple biologic pathways, we will have to change the design of our clinical research: the trials will enroll fewer patients, the population of patients will be more homogeneous, and the endpoints for clinical benefit will be different. We will witness more and more drugs approved without large, randomized, phase 3 clinical trials. This approach can only be successful if we spend enough time at the bench before we take it to the bedside; and, if we see that we are not successful at the bedside, then we must return to the bench immediately. We also will have to simplify the conduct of clinical research in such a way that we do not jeopardize patients' health and do not spend all of our energy on regulatory requirements.