Insulin-like growth factor-I receptor blockade by a specific tyrosine kinase inhibitor for human gastrointestinal carcinomas

Signals from a variety of growth factors are required for tumorigenesis and cancer development in human malignancies (1), including gastrointestinal (GI) carcinomas (2). Tyrosine kinase receptors are one new group of targets and can be targeted by small-molecule tyrosine kinase inhibitors (TKI) or monoclonal antibodies. Insulin-like growth factor-I receptor (IGF-IR) signaling is able to potently stimulate tumor progression and cellular differentiation (3) and is a promising new molecular target in human malignancies (2, 4).

IGF-IR consists of two α and two β subunits in a heterotetramer (5). Binding of the ligands IGF-I and IGF-II to IGF-IR causes receptor autophosphorylation and activates multiple signaling pathways, including the mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) and phosphatidylinositide 3-kinase/Akt-1 pathways (6, 7). IGF-IR pathway activation is also regulated by IGF-binding proteins and IGF-II receptor (8–10).

Dysregulation of the IGF-I system has been implicated in the proliferation of numerous neoplasms (2, 11). Elevation of serum IGF-I increases the risk of developing several cancers, including colon (9), and IGF-IR is also important for tumor maintenance (12, 13). Blocking IGF-IR has been shown to induce apoptosis in tumors but produce only growth arrest in untransformed cells (1), implying that receptor blockade may have a greater therapeutic index than other strategies targeting fundamental cell mechanisms such as DNA synthesis or the cell cycle. The fact that IGF-IR knockout mice are viable indicates that relatively normal tissue development and differentiation occurs in the absence IGF-IR signaling (14).

IGF stimulate the proliferation of GI cancer cells, including colon, esophagus, hepatocellular, and pancreas, and blocking of IGF-IR signaling inhibits tumor growth (8, 15–22). Soluble IGF-II receptor rescues Apc(Min/+) intestinal adenoma progression induced by loss of IGF-II imprinting (10). Intestinal fibroblast-derived IGF-II has been shown to stimulate proliferation of intestinal epithelial cells (23). Expression of both IGF-II and IGF-IR are increased in GI cancers (16, 18, 24, 25). IGF-I also has been shown to antagonize the antiproliferative effects of cyclooxygenase-2 inhibitors (26). IGF-IR signaling is also important in tumor dissemination through the control of adhesion, migration, and metastasis. These findings suggest a potential basis for tumor selectivity in therapeutic applications in GI cancers.

The insulin receptor (IR) is also a key component of the IGF signaling pathway. IR activation leads cell proliferation in addition to glucose metabolism. In addition to insulin, IR can also bind IGF-II and initiate mitogenic signaling (27). Two isoforms of IR are generated by alternative splicing of exon 11, A and B isoforms (28). IGF-IR and IR can form hybrid receptors that bind IGF at physiologic concentrations. IR and IGF-IR/IR hybrid receptors may also be involved in cancer biology as both insulin and IGF-I contribute to the development and progression of adenomatous polyps (29).

Several possible approaches to blocking IGF-IR signaling have been reported. Humanized monoclonal antibodies for IGF-IR are available (30–32) and some are in phase I trials. TKI for IGF-IR have been developed, including NVP-ADW742 (33), BMS-536924 (34), and BMS-554417 (dual targeting for IR and IGF-IR; ref. 35). We have constructed two dominant-negative inhibitors for IGF-IR (IGF-IR/dn), which are active as plasmids and recombinant adenovirus vectors in GI malignancies (18, 36–38).

The orally available compound NVP-AEW541, a pyrrolo[2,3-d]pyrimidine derivative, is a novel TKI for IGF-IR (39). It is highly selective against IGF-IR compared with other tyrosine kinases. There are only two reports about the effects of AEW541 on GI malignancies, hepatoma and neuroendocrine tumors (40, 41). As GI cancers are heterogeneous diseases, we analyzed the efficacy of this compound in esophageal squamous cell carcinoma, colorectal and pancreatic adenocarcinoma, and hepatoma. In this study, we assessed the effect of this TKI on signaling blockade, growth, apoptosis induction, migration, and in vivo therapeutic efficacy in s.c. xenografts. We also analyzed effects of this drug on insulin signaling and hybrid receptor formation. These observations strengthen the rationale for using NVP-AEW541 alone or in combination with chemotherapy in the molecular targeted therapy of GI cancers.

Anti-Akt-1 (c-20), anti-ERK1 (K-23), anti-phosphorylated ERK1 (E-4), ant-IGF-I (G-17), anti-IGF-IRα (2C8), and anti-IGF-IRβ (C20) were purchased from Santa Cruz Biotechnology and anti-phosphorylated Akt (Ser⁴⁷³), anti-phosphorylated p44/42 mitogen-activated protein kinase (Thy²⁰²/Tyr²⁰⁴), and PathScan Multiplex Western Cocktail-I were from Cell Signaling Technology. Cisplatin and 5-fluorouracil (5-FU) were purchased from Sigma. Recombinant human IGF-I and IGF-II were purchased from R&D Systems. All human GI cancer cell lines were obtained from Japanese Cancer Collection of Research Bioresources Cell Bank. Cells were passaged in RPMI 1640 and DMEM with 10% fetal bovine serum. Specific pathogen-free female BALB/cAnNCrj-nu mice, 6 weeks old, were purchased from Charles River. The care and use of mice were according to our university' guidelines.

NVP-AEW541 was kindly provided by Novartis Pharma. Stock solution of this drug was prepared in DMSO and stored at −20°C.

Cells were treated with 20 ng/mL IGF-I, 40 ng/mL IGF-II, or 10 nmol/L insulin as indicated in the text. Cell lysates were prepared as described previously (36). Equal aliquots of lysates (100 μg) were separated by 4% to 20% SDS-PAGE and immunoblotted onto polyvinylidene Hybond-P membrane (Amersham). Analysis was done using indicated antibodies, and bands were visualized by enhanced chemiluminescence (Amersham).

To analyze IGF-IR or IR signaling, 1 mg cell lysates were immunoprecipitated overnight at 4°C with anti-IGF-IR or anti-IR antibodies followed by Western blotting.

Four thousand cells were seeded into a 96-well plate and each was treated with several concentrations of AEW541. Cell growth was measured using WST-1 reagent (Roche) as described previously (18).

Cells (3 × 10³ per plate) were seeded onto 60-mm culture plates and incubated for 24 h. The cells were then treated with NVP-AEW541 and were incubated for 14 days. After air drying, cells were fixed with methanol and stained with Giemsa solution. Colonies containing ≥50 cells were counted.

Caspase-3 colorimetric protease assay was done following the manufacturer's protocol (MBL). In brief, the caspase-3 activity of lysates (100 μg) was measured by colorimetric reaction at 400 nm. Early apoptosis were quantified by staining with Annexin V-FITC according to the manufacturer's protocol (BD Biosciences) and measured by flow cytometry. Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling assays were done with in situ apoptosis detection kit (Takara) following the manufacturer's protocol.

HT29 cells (1 × 10⁶) were s.c. injected into nude mice. After tumors were palpable (on day 24), animals were treated p.o. twice daily for 2 weeks either with NVP-AEW541 (40 mg/kg) or control. Three days after treatment, tumors were collected for analysis. Tumor diameters were serially measured with calipers and tumor volume was calculated using the formula: tumor volume (mm³) = (width² × length) / 2.

After HT29 tumors were palpable, animals were treated p.o. twice daily for 2 weeks with either NVP-AEW541 (40 mg/kg) or control. Both groups were then divided into pair-matched 5-FU-treated and control groups (administered i.p. once weekly for four doses). Mice were euthanized when tumors reached 2 cm in size or they developed clinically evident symptoms.

Sections (5 μm) from formalin-fixed, paraffin-embedded tumor xenografts were prepared. After removed paraffin, the sections were pretreated with DakoCytomation Target Retrieval Solution (Dako) in a microwave (10 min). Then, endogenous peroxidase activity was blocked. Antibodies were applied after blocking with normal goat serum. Sections were incubated with the anti-rabbit secondary antibody (Santa Cruz Biotechnology) and a streptavidin-horseradish peroxidase (Dako) followed by exposure to the diaminobenzidine tetrahydrochloride substrate (Dako).

Measurement of immunostained area in pictures was done using the public domain NIH Image program.⁴

Tumor samples were homogenized in Western blot lysis buffer and protein content was measured. Concentrations of IGF-IR, IR, and IGF were measured using the following ELISA kits according to the manufacturer's protocols: human IGF-I DuoSet, human total insulin R Duoset IC, and human total IGF-IR DuoSet IC (R&D Systems) and Active IGF-II ELISA Kit (Diagnostic Systems Laboratories).

Wounding assays were done using a modification of the procedure described previously (18). Briefly, six-well chambers were prepared by scratching registration marks onto the slide surface. Cells were plated, grown normally for 48 h, and starved overnight. Cells were cut with a cell scraper and five images were captured 24 h later on an Olympus IX 71S1F-2 microscope using a ×20 objective. For each experiment, two independent observers counted the number of migrating cells.

The results are presented as mean ± SE for each sample. The statistical significance of differences was determined by Student's two-tailed t test in two groups and done by one-way ANOVA in multiple groups and two-factor factorial ANOVA. Survival curves were constructed according to the method of Kaplan-Meier. P values < 0.05 were considered to indicate statistical significance.

To investigate the effect of NVP-AEW541 on the IGF/receptor axis, we examined four GI cancer cell lines: HT29, TE1, BxPC3, and PLC/PRF/5. First, we evaluated NVP-AEW541 activity in HT29 by immunoprecipitation with anti-IGF-IRα and immunoblotting for phosphorylated tyrosine and IGF-IRβ. NVP-AEW541 reduced IGF-I-induced phosphorylation of IGF-IR in a dose-dependent manner and 1.0 μmol/L NVP-AEW541 blocked it completely (Fig. 1A). Then, we assessed the downstream signaling pathways of IGF-IR. This TKI blocked phosphorylation of Akt-1 and ERK both basal and that induced by both ligands in a dose-dependent manner (Fig. 1B).

Similarly, in TE1 cells, NVP-AEW541 reduced ligand-induced IGF-IR autophosphorylation and phosphorylation of Akt-1 and ERK in a dose-dependent manner (Fig. 1A and B). In BxPC3 and PLC/PRF/5 cells, 1 μmol/L NVP-AEW541 also reduced the ligand-induced phosphorylation of IGF-IR and its downstream signal activity (Fig. 1C and D). Thus, AEW541 effectively interrupted phosphorylation of IGF-IR and its downstream signals in GI cancers.

NVP-AEW541 reduced the growth on plastic of all GI cancer cells in a dose-dependent manner as analyzed by the WST-1 assay (Fig. 2A). Moreover, AEW541 dramatically reduced the in vitro tumorigenicity of all cells dose-dependently as assessed by colony formation assay (Fig. 2B). These results indicate that NVP-AEW541 effectively blocks the carcinogenicity and proliferation of multiple GI cancers.

To assess the effects of NVP-AEW541 on migration of GI cancer cells, we did wounding assays. AEW541 suppressed migration of HT29 (Fig. 2C). The number of migrated cells treated is significantly reduced compared with control. The compound significantly inhibited motility of other cells. The results indicate that NVP-AEW541 may have potential prevention for metastasis and tumor progression.

NVP-AEW541 also dramatically increased terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling–positive cells in GI cancers (Fig. 3A). These results were confirmed by Annexin V assay (Fig. 3B), NVP-AEW541 increased the percentage of cells in early apoptosis. In BxPC3 cells, apoptotic induction was observed in 0.5 μmol/L NVP-AEW541.

We assessed the effect of NVP-AEW541 on stressor-induced apoptosis. Caspase-3 assay revealed that serum starvation did not induce apoptosis in HT29 but did induce apoptosis in the presence of NVP-AEW541 (Fig. 3C). Heat-induced caspase-3 activity and NVP-AEW541 up-regulated heat shock–induced apoptosis in both HT29 and BxPC3 (Fig. 3C). This result was confirmed by Annexin V assay in HT29, BxPC3, and PLC/PRF/5 (Fig. 3D). In TE1, NVP-AEW541 increased the percentage of cells undergoing apoptosis in response to ethanol (Fig. 3D).

The effects of NVP-AEW541 in combination with chemotherapy was assessed. Colony formation assays revealed that cisplatin reduced the number of colonies in a dose-dependent manner in all cells and NVP-AEW541 significantly enhanced the effect of cisplatin (Fig. 4A). Both cisplatin and 5-FU alone induced caspase-3 activity in all cells (Fig. 4B and C), and NVP-AEW541 significantly enhanced chemotherapy-induced apoptosis with all combinations, except for cisplatin in PLC/PRF/5 and 5-FU in TE1. These results suggest that IGF-IR may play critical roles in tumor progression and that NVP-AEW541 enhances the effects of chemotherapies.

To assess the effect of this drug on in vivo GI cancer, HT29 cells were inoculated in nude mice and allowed to form evident tumors. Oral administration of 40 mg/kg NVP-AEW541 (twice daily for 2 weeks) significantly inhibited HT29 tumor growth on mice (Fig. 5A).

As this TKI up-regulated 5-FU-induced apoptosis in vitro, the effect of combined IGF-blockade with 5-FU on established tumors was then assessed. HT29 tumor-bearing mice were treated with NVP-AEW541 along with 5-FU or PBS. Although either 5-FU or NVP-AEW541 suppressed tumor growth, the combination therapy was much more effective than either treatment alone (Fig. 5B). These treatments did not have an adverse effect on the body weight of the treated mice, suggesting tolerable toxicity. This indicates that NVP-AEW541 has significant potential to enhance the effect of chemotherapy in GI cancers.

To evaluate the effect of AEW541 on IGF-IR downstream signals in tumors, immunohistochemical analysis (Fig. 5C) and ELISA (Fig. 5D) were done. Immunohistochemically positive areas were quantitatively measured using densitometry and NIH image software, and this revealed that AEW541 inhibited both activation of Akt and ERK (Fig. 5C) without influence on IGF-IR expression. This compound blocked phosphorylation of Akt-1 more than that of the ERK. HT29 did not express IGF-I by ELISA at all (data not shown), and AEW541 did not affect on the expression of IGF-II, IGF-IR, and IR (Fig. 5D) in HT29 tumors on mice. These data suggest that this TKI effectively blocks IGF-IR signals without affecting the expression of IGF-IR/IR pathway proteins.

As IGF-IR has homology with the IR, we assessed the effects of this inhibitor on insulin signaling. First, we assessed the expression of IR isoforms in GI cancer cells. All cells expressed both isoforms (Fig. 6A). Although both HT29 and BxPC3 express A isoform mainly, both PLC/PRF/5 and TE1 express both isoforms equally.

To define the effects of NVP-AEW541 on downstream signaling of IR in HT29, Western blotting was done. Although insulin induces both phosphorylation of Akt and the ERK, the TKI did not block this phosphorylation (Fig. 6B). Immunoprecipitation with anti-IR followed by immunoblotting for phosphotyrosine or IR revealed that this inhibitor did not block insulin-induced phosphorylation of IR (Fig. 6B). The result is confirmed in TE1 (Fig. 6B). These data indicate that NVP-AEW541 can block IGF signaling without detectable effects on insulin signaling under these conditions.

We assessed heterodimerization of IGF-IR and IR in both HT29 and TE1. IGF-IR was detected in all samples immunoprecipitated with anti-IR (Fig. 6B), showing that NVP-AEW541 does not influence this heterodimerization. This was confirmed by the observation that IR was also identified in all immunoprecipitations with anti-IGF-IR, although AEW541 blocked IGF-induced phosphorylation of IGF-IR (Fig. 6C). In BxPC3 and PLC/PRF/5 (Fig. 6D), IGF-IR was detected equally in the presence or absence of this drug after immunoprecipitation with anti-IR. TKI also did not affect IR pull-down in cells immunoprecipitated with anti-IGF-IR. These results indicate that low dose of 1 μmol/L NVP-AEW541 selectively blocks IGF-IR signaling.

GI cancers are frequently diagnosed in advanced stages with lymph node and distant metastases and peritoneal dissemination, and there is a paucity of meaningful treatment strategies for these tumors. IGF-mediated growth responsiveness is found in most GI cancer cells (15, 18, 19, 22). Aberrant activation of IGF-IR by paracrine and autocrine loops has been reported (22, 42), and the expression of IGF-IR/IGF-II may be useful for the prediction of recurrence and poor prognosis in esophageal squamous cell carcinoma (18). Here, we used a new IGF-IR-specific TKI, NPV-AEW541, for the accurate dissection of the responsible signaling pathways. Even in nanomolar concentrations, this agent suppressed colony formation efficiency and enhanced chemotherapy-induced apoptosis in vitro and chemotherapeutic efficacy in vivo. These results indicate that this strategy is promising for the treatment of GI cancers not only as a monotherapy but also to enhance the efficacy of chemotherapy. We show that AEW541 is effective against a wide range of GI carcinomas, with low toxicity in animal models. Blockade of phosphatidylinositide 3-kinase abrogates IGF-induced transmigration (43). As migration is one of critical steps for cancer progression, especially invasion and metastasis, it is worthwhile that NVP-AEW541 can inhibit the mobility of all four GI cancers studied.

A major hurdle to the targeting of IGF-IR is the close homology of the IGF-IR and the IR kinase domains. The ATP-binding site of these two receptors displays 100% sequence identity, whereas the entire kinase domains share 84% sequence identity, both with each other and across species (44). Therefore, it is very important that any strategy designed to block IGF-IR signaling has specificity for IGF-IR and without significant influence on IR signaling. We show here that AEW541 does not suppress insulin-induced IR or Akt phosphorylation, indicating a high degree of receptor selectivity at tumor-active doses with this molecular targeted therapy. Both receptors can form not only homodimers but also heterodimers. However, AEW541 does not influence this heterodimerization. Although high doses of AEW541 can block several tyrosine kinase receptors including IR, we show that concentrations of this compound active against GI cancers (≤1.0 μmol/L) inhibit IGF-IR signals selectively without affecting insulin signals. NVP-AEW541 shows a 27-fold inhibitory selectivity for the IGF-IR versus IR (39) and is slightly more active (IGF-IR kinase IC₅₀, 0.086 mol/L) and is more selective than another TKI, NVP-ADW742 (33, 45). Although plasma glucose and insulin were not monitored in mice treated with active doses of NVP-AEW541 to assess the potential risk of hypoglycemia, body weight was monitored and not significantly affected by TKI therapy. These results confirm our biochemical findings. A recent article reported an initial weight loss, with a significant decrease of serum glucose due to increased glucose uptake at cellular level, in mice bearing Ewing's sarcoma, similar to the weight loss observed in mice treated with vincristine (46). The weight changes in our study are similar to their cohort of mice, which started to grow again after the first days of treatment. This lack of effect on the insulin pathway is very important for avoiding adverse effects in clinical applications. In addition, NVP-AEW541 has the advantage of being orally bioavailable.

However, some articles have recently reported several effects of IGF-IR blockade on IR. Both a chimeric single-chain antibody against IGF-IR (scFv-Fc) and a murine antibody for IGF-IR (EM164) resulted in decreased levels of IR in vitro and in vivo despite their lack of reactivity against IR (29). The down-regulation of IR by those anti-IGF-IR monoclonal antibodies requires the coexpression of IGF-IR and might be due to the endocytosis of hybrid IR/IGF-IR. IGF-IR down-regulation by small interfering RNA did not affect IR levels but, interestingly, did sensitize cells to insulin activation of downstream signaling pathways in breast cancer cells (47). IGF-IR small interfering RNA treatment diminished hybrid receptor formation, suggesting that specific down-regulation of IGF-IR resulted in enhanced holo-IR formation and insulin sensitivity. The expression IGF/IGF-IR and IR were not changed by this TKI, indicating that NVP-AEW541 might induce neither positive nor negative feedback in IGF-IR/IR axis in mice.

There are thus two opposing strategies for blockade IGF-IR signaling. One is to avoid adverse effects by shutting down only IGF-IR signaling without influencing IR signaling. Such selective IGF-IR blockade can be achieved by low-dose NVP-AEW541 or IGF-IR/dn for prevention of recurrence or maintenance of remission. Low dose of AEW541 shows clear antitumor effects not only by itself but also in combination with commonly used chemotherapy drugs. The other approach is to achieve maximum antitumor effects by blocking both IGF-IR and IR simultaneously using scFv-Fc, BMS-554417, or high-dose NVP-AEW541. This latter approach is likely to be significantly more toxic.

Recently, a new role for the IGF-IR has been identified with the demonstration that there is significant crosstalk between the IGF-R and other tyrosine kinase receptors. For example, many breast cancer patients who achieve an initial response to trastuzumab acquire resistance within 1 year from treatment initiation. One mechanism of resistance has been shown to be overexpression of IGF-IR (48) and another is the formation of IGF-IR/Her2 heterodimers in the resistant tumors (49). These data suggest that selective IGF-IR blockade may be specifically effective for breast cancer patients with trastuzumab resistant tumors.

Thus, in this study, we show that NVP-AEW541 suppresses tumorigenicity, survival, and migration by blocking Akt and ERK activation both in vitro and in animal models. It also potentiated chemotherapy efficacy without significantly influencing insulin signals at therapeutically effective doses. These studies thus validate IGF-IR as a therapeutic target in GI malignancies and suggest that NVP-AEW541 may be a promising anticancer therapeutic.

No potential conflicts of interest were disclosed.