The insulin-like growth factor (IGF) signaling system is comprised of the IGF ligands (IGF-1 and IGF-2), the cell surface receptors that mediate the biological effects of the IGFs (IGF-1 receptor (IGF-1R), IGF-2 receptor and the insulin receptor), as well as a family of circulating IGF-binding proteins. Insulin-like growth factors 1 and 2 have a critical physiological role in the growth and development of many tissues, and are also implicated in various pathological conditions, particularly in multiple myeloma (MM). Primary cells from patients or MM cell lines are particularly sensitive to IGF-1 and IGF-1R inhibition.1, 2 IGF-1R (CD221) is aberrantly expressed on MM cells in about 75% of the cases and is associated with disease severity.1, 2, 3 It was also shown that IGF-1R (CD221) signaling attenuates responsiveness of MM cells to proapoptotic agents, including dexamethasone, cytotoxics, or proteasome inhibitors.2 IGF-1R is therefore an attractive therapeutic target for MM treatment based on the hypothesis that inhibition of IGF-1R function would result in growth inhibition and apoptosis of tumor cells, particularly in combination with other chemotherapeutic drugs and especially proteasome inhibitors.4
AVE1642 is a humanized version of the murine monoclonal antibody, EM164, raised against the human IGF-1R.5 It has been shown to bind specifically, with a high affinity, to human and monkey IGF-1R, to inhibit IGF-1 binding and receptor activation, and to downregulate the receptor by internalization and degradation. AVE1642 has been able to delay growth and survival of some cancer cells in vitro, human tumor xenografts in nude mice, and to inhibit proliferation and survival of most MM cells (primary cells from patients or MM cell lines).4
The present phase 1, multicenter, 2-part study, open-label, was to explore the dose to be selected for further development of AVE1642 (dose escalation design, part 1), to evaluate pharmacokinetics (clearance of the antibody) and pharmacodynamics (IGF-1 serum levels) of AVE1642 alone and in combination with bortezomib, and to assess the safety of the combination of AVE1642 (at the selected dose), with the recommended dose of bortezomib in patients with relapsed MM (Part 2).
A total of 26 patients with relapsed/refractory MM, median age 66 years, were included, from four participating centres (three in France and one in Italy). At the time of study entry, the median number of lines of previous therapy was four. For part 2 of the study, previous exposure to bortezomib was allowed, provided no Grade3 toxicity related to the drug had been observed during previous therapy. Overall, in the first part of the trial dedicated to the definition of the maximum tolerated dose according to the occurrence of dose-limiting toxicities (DLTs) during cycle 1, 15 patients received 51 cycles of AVE1642 at doses ranging from 3 (four patients), 6 (six patients), 12 (four patients) to 18 mg/kg (one patient). AVE1642 was infused IV, on day 1 of 21-day cycle. The median exposure was two cycles for all dose levels, with the highest number of 12 cycles administered at dose level 6 mg/kg in one patient. One of the first three patients treated in cohort no. 2 (6 mg/kg of AVE1642) experienced a DLT (Grade 3 hyperglycemia) at cycle 1. Three additional patients received 6 mg/kg, and no DLT occurred. No DLT was reported in cohort no. 3 (12 mg/kg) and in the last patient treated at the highest dose (18 mg/kg). Overall, after the treatment of these 15 patients, the maximum tolerated dose was not reached, the toxicity profile of AVE1642 was favorable, and pharmacokinetics plus pharmacodynamics analyses on appropriate parameters showed evidence of the saturation of the IGF-1Rs such as AVE1642 clearance and circulating level of IGF-1: concomitant plateaus of both parameters at doses of 3, 6 and 12 mg/kg were deemed to reflect a maximal saturation of accessible receptors to AVE1642. Therefore, the investigators considered that AVE1642 could be tested in combination with bortezomib in the second part of the trial. Moreover, assessment was negative for Human antihumanized antibody in all 15 patients. Overall, in the second part of the trial, 11 patients received 45 cycles of AVE1642 at doses ranging from 0.5 (three patients), 6 (four patients) to 12 mg/kg (four patients) in combination with IV bortezomib at the fixed dose of 1.3 mg/m2 (days 1, 4, 8 and 11). The median exposure was four cycles for all dose levels, with the highest number of eight cycles administered at initial dose level 0.5 mg/kg of AVE1642 in one patient. No dose-limiting toxicity was observed, and the majority of adverse events described were related to bortezomib infusion. Four (36.4%) patients experienced five serious adverse events among which three (disease progression, hypercalcemia and renal vein thrombosis) were Grade 3–4. Assessment was negative for Human antihumanized antibody in all 11 treated patients.
Pharmacokinetic and pharmacodynamic parameters were assessed at cycle 1 and cycle 2 in part 1 and at cycle 1 in part 2 (Figures 1 and 2). AVE1642 serum concentrations showed biphasic profiles, with a rapid distribution phase and a slow elimination phase, as typically observed for IgG1 monoclonal antibodies. The shape of the 0.5 mg/kg curve suggested more rapid elimination, than for the higher 3–18 mg/kg groups, which showed parallel elimination profiles. No significant accumulation were observed between cycle 1 and cycle 2 and visual inspection of the trough profiles suggested that the steady-state was attained from the second to third administration. Exposure was approximately dose-proportional. Similarly, exposure was comparable between study part 1 (AVE1642 as single agent) and study Part 2 (AVE1624+bortezomib), suggesting that bortezomib had no major effect on AVE1642 serum concentrations. AVE1642 serum half-life was around 10 days, as reported for similar anti-IGFR monoclonal antibodies. Clearance was similar across the 0.5–18 mg/kg dose range. These data suggest that the plateau in clearance was attained from the low dose of 0.5 mg/kg. Pharmacodynamic analysis showed in general large interindividual variability and lack of dose/pharmacological response relationship. The relationship between the observed pharmacological effect (IGF-1 serum concentration) and AVE1642 serum concentration resulted in counter-clockwise hysteresis loops,6 indicating a delayed pharmacological response. Following the first administration of AVE1642, IGF-1 serum concentrations increased from physiological baseline level to maximal concentration at around 14 days. In the 0.5 mg/kg cohort, IGF-1 concentrations declined after peak and returned to baseline level at the end of cycle 1. At the higher 3–18 mg/kg doses, at the subsequent AVE1642 administration, IGF-1 concentrations remained elevated and reached a plateau. No apparent dose-concentration relationship were observed above the dose of 0.5 mg/kg, suggesting that the plateau of IGF-1 levels was attained at or below 3 mg/kg.
Disease responses were evaluated according to the EBMT criteria. In the first part of the study (AVE1642 single-agent, dose-escalation phase), out of 15 patients treated, the best overall response was a minor response in one patient, with stable disease observed in seven cases. Disease progression was described at cycle 1 in four patients, at cycle 2 in three patients, at cycle 3 in two patients and after at least four cycles of treatment in five patients. In the second part of the trial, out of 11 patients treated with the combination therapy, five responses were observed: one complete response, one partial response (PR=2/11, 18%) and three minor response (among these patients, only one patient who experienced minor response had been previously exposed to bortezomib and achieved PR during his first exposure to the drug). Three patients had a stable disease. Disease progression was described at cycle 1 in two patients, at cycle 3 in two patients and after at least 6 cycles of treatment in two patients.
Overall, despite the good toxicity profile of the antibody, the response rates for patients treated with AVE1642 in this study, as a single agent or in combination with bortezomib, were considered insufficient to merit further development of AVE1642 in MM.
These unsatisfactory results might be partially explained by two factors. First, inclusion criteria did not allow the selection of the ideal population of patients for AVE1642 therapy. The phenotype of tumor cells, especially CD221 (IGF1-R) and CD45 expression, was not systematically evaluated at study entry. We have previously shown using flow cytometry that CD221 was aberrantly expressed at a significant level in 73% of patients at the time of diagnosis, in relation with disease severity, but the expression of this molecule, albeit very frequent, is not constant in MM.3, 7 Potentially, to increase the likelihood of response to AVE1642 therapy, the antibody should be tested only in patients with MM that exhibits significant expression of CD221. Moreover, CD221 expression needs to be evaluated together with CD45 phenotype.7, 8 CD45 is an important tyrosine phosphatase that can modulate signal transduction thresholds and impair the activation of the AKT pathway by IGF-1.8 The lack of CD45 expression then allows a response to IGF-1, that is, growth through the AKT pathway. We have previously shown that AVE1642 was able to inhibit the proliferation of CD221 positive-CD45 negative, but not that of CD221 positive-CD45 positive human myeloma cell lines, and that the overexpression of CD221 in CD45-positive human myeloma cell lines was not sufficient to induce AVE1642 sensitivity.4, 8 On the contrary, extinction of CD45 by shRNA in CD45 positive human myeloma cell lines was able to induce AVE1642 sensitivity. In the present study, CD45 phenotype was evaluated in 23 out of 26 patients, and was negative in 16 cases (70%). The second potential explanation is that AVE1642 was not able to neutralize insulin/IGF1 hybrid receptors. Indeed it has been shown that IGF1-R and insulin receptor can heterodimerize leading to the formation of insulin/IGF-1 hybrid receptors, which comprise one α- and one β-subunit of each receptor. A recent paper from Sprynski et al.9 clearly demonstrate that these hybrid receptors can be activated by both IGF-1 and insulin supporting myeloma cell survival and proliferation. Therapeutic strategies targeting the IGF/IGF-1R pathway have now to take into account the existence of these hybrid receptors that need to be neutralized to provide an optimal activity. Unfortunately, we did not generate in vitro and in vivo data supporting that AVE1642 was able to target and fully neutralized these hybrid receptors. This study also illustrates the difficulties of the development of targeted therapies based on phenotype analysis in MM. The ideal target should be expressed on 100% of tumor cells, 0% on normal cells, easily saturated to induce systematic apoptosis. Such therapy remains a huge challenge in MM, one of the more heterogeneous disease.
About this article
Frontiers in Immunology (2018)