Impact of Insulin-Like Growth Factor Type 1 Receptor, Epidermal Growth Factor Receptor, and HER2 Expressions on Outcomes of Patients with Gastric Cancer

Globally, gastric cancer (GC) is the third most prevalent malignancy. Although its incidence is declining, ∼930,000 cases are newly diagnosed each year (1). Despite the identification and development of several new classes of anticancer agents, GC remains an aggressive malignancy, with a median survival of 7 to 10 months in patients with metastatic or unresectable disease (2, 3). Once metastatic, GC is incurable, and chemotherapy is palliative. Increasing emphasis on the need for improved techniques for the prediction of treatment response and survival may facilitate the tailoring of chemotherapy and risk-related therapy, resulting in significantly better survival.

Insulin-like growth factor type 1 receptor (IGF-IR) is a cell membrane receptor that is activated by its ligands, IGF-I and IGF-II. IGF-IR participates in cell proliferation, differentiation, and prevention of apoptosis (4, 5). Because IGF-IR is also involved in malignant transformation (5), development of IGF-IR–directed cancer therapy has been initiated. IGF-IR is frequently overexpressed in human cancers, and the association between IGF-IR expression and outcomes has been assessed for breast cancer and other solid tumors (6, 7). However, IGF-IR expression in GC remains poorly understood.

Previous studies have indicated that IGF-IR can interact with epidermal growth factor receptor (EGFR) to augment the malignant behavior of tumors (8, 9). EGFR and its homologues HER2 (also known as erbB-2) are members of the erbB gene family. These receptors encode for transmembrane receptor–type tyrosine kinases, for which therapeutic approaches do exist. They play a crucial role in tumor cell proliferation, survival, adhesion, migration, and differentiation and also participate in tumor angiogenesis (10). On ligand stimulation, these receptors form either homodimers or heterodimers, which activate their cytoplasmic domain. Through EGFR, epidermal growth factor and transforming growth factor-α stimulate DNA synthesis and cell growth in various systems, including the gastrointestinal tract (11). HER2 is the preferred co-receptor for the formation of dimers with EGFR, HER3, or HER4; the heterodimers consisting of HER2 and these other receptors have a greater capacity for translating mitogenic signals than the homodimers and act synergistically to promote cellular transformation (12). In many cancers, the expression of these receptors may be related to patient survival (13). However, whether the expressions of EGFR and HER2 are significant predictors of survival in patients with GC remains controversial.

This study was designed to determine the expression frequencies of IGF-IR, EGFR, and HER2 in GC by immunohistochemical assays and to assess the prognostic relevance of these receptors. The relations of the expressions of these receptors to clinicopathologic characteristics were also examined.

Patients. Patients with a diagnosis of histologically proved advanced GC were eligible for the study. Inclusion criteria were as follows: unresectable, recurrent, or metastatic disease; no prior chemotherapy and no prior adjuvant/neoadjuvant chemotherapy; specimens of primary gastric adenocarcinomas were obtained by surgical resection before the start of chemotherapy at National Cancer Center Hospital; first-line chemotherapy was received at National Cancer Center Hospital; radiographically measurable disease; and written informed consent. The tissue samples were collected retrospectively from patients who met these criteria. Measurable disease was assessed by a standardized computed tomographic examination every 2 mo in all patients. Response was evaluated according to the standard Unio Internationale Contra Cancrum guidelines as complete response, partial response, no change, or progressive disease (14). Tumor response and survival times as of December 2006 were confirmed in all patients. The following clinical characteristics were included in analyses: age, performance status, liver metastases and peritoneal metastases at the start of first-line chemotherapy, and histologic type and grade of the surgically removed primary tumor. Peritoneal metastases included only measurable nodules arising in the peritoneum, without bowel obstruction. Written informed consent was obtained before treatment and the evaluation of tumor samples. This study was approved by the institutional review board of National Cancer Center Hospital.

Immunohistochemical assays. Primary antibodies used for immunohistochemical assays were commercially available mouse monoclonal antibodies against human IGF-IRα subunit, Ab-1 (clone 24-31; Lab Vision Corporation), an EGFR PharmDx kit, and a HercepTest kit (Dako). All tumor specimens used in our study were derived from routine formalin-fixed, paraffin-embedded tissue samples obtained from the resected primary GC specimens. One block that included the site of deepest invasion was selected from each specimen, after reviewing slides of the surgical specimens stained with H&E. Sections (4-μm thick) were cut from the paraffin blocks and mounted on silanized slides. Immunohistochemical staining of EGFR and HER2 were done by using a EGFR pharmDx kit and a HercepTest kit according to the manufacturer's instructions, using the reagents supplied with the kits. For IGF-IR staining, the sections were deparaffinized in xylene and dehydrated with graded ethanol. After washing with distilled water, the sections were placed in the supplied buffer. For antigen retrieval, the slides were heated at 95°C for 40 min and then cooled for at least 20 min at room temperature. After washing with distilled water and with Tris-buffered 0.9% NaCl solution containing Tween 20 (pH 7.6), tissue sections were covered for 5 min with peroxidase blocking reagent (Dako) to block endogenous peroxidase, followed by an additional washing with the supplied buffer. Individual slides were then incubated for 30 min at room temperature with anti–IGF-IR antibody, Ab-1 (dilution 1:50), in the antibody diluent buffer. The slides were washed thrice with the buffer and then incubated with a peroxidase-labeled polymer conjugated to goat anti-mouse IgG (Dako ENVISION+, mouse/horseradish peroxidase, code K4000; Dako) for 30 min at room temperature. After extensive washing with PBS, the color reaction was developed in DAB Liquid System (code K3465; Dako) according to the manufacturer's instructions. The sections were then counterstained with Meyer's hematoxylin, dehydrated, and mounted. In the negative controls, the primary antibody solution was substituted with a buffer containing mouse IgG1 (negative control).

Scoring system of immunostaining. To assess membranous staining for IGF-IR, EGFR, and HER2, slides were independently evaluated by two investigators (J.M. and Y.H.), and the results were then reviewed by an experienced gastrointestinal pathologist (T.S.). All of them were blinded to the clinical follow-up data. Although cytoplasmic staining of the tumor cells was occasionally noted, most likely resulting from internalized or nascent receptor molecules, only staining of the tumor cell membranes was considered to represent specific expression. Immunostaining on each slide was scored according to a semiquantitative four-grade scale: negative (no positively stained cells or 1% to 10% of positively stained cells in tumor), low (>10-40% of positively stained cells in tumor), moderate (>40-70% of positively stained cells in tumor), high (>70% of positively stained cells in tumor). Samples assigned scores of low, moderate, or high were considered to be positive for expression of the respective protein. The cutoff values for positive expression were the same as those used in previous studies (15, 16).

Statistical analysis. Associations of IGF-IR, EGFR, or HER2 expression with clinicopathologic characteristics were assessed with Spearman's rank-correlation test or Mann-Whitney U test, as appropriate. Overall survival time was defined as the period from the date of starting first-line chemotherapy until the date of death from any cause or until the date of the last follow-up, at which point, data were censored. We used the Kaplan-Meier method to plot overall survival curves according to immunohistochemical results, and the statistical significance of differences was assessed with the log-rank test. Univariate and multivariate Cox proportional hazard models were used to estimate the relations of protein expressions and clinical characteristics to overall survival. All reported P values are two-sided, and the level of significance was set at P < 0.05. Variables for multivariate analysis were selected by means of a forward stepwise approach, using a significance level of P < 0.10 for entering into or remaining in the model. All analyses were done with the use of the statistical software package StatView, version 5.0 (SAS Institute, Inc.).

A total of 87 patients were eligible for the study. Chemotherapy began in July 1997 in the first patient and May 2004 in the last patient. The demographic characteristics of the patients at the start of first-line chemotherapy are shown in Table 1. There were 70 men (80%) and 17 women (20%), with a median age of 64 years. The associations of IGF-IR, EGFR, and HER2 expression on immunohistochemical assay to clinical outcome could be assessed in all patients. At the time of analysis, 79 patients (91%) had died and eight patients (9%) were alive.

The chemotherapy regimens received by the patients and the response rates are also listed in Table 1. The response rates to first-line chemotherapy in our study are comparable with those reported previously (2, 17–19).

Expression frequencies of IGF-IR, EGFR, and HER2 in primary tumors and associations with clinicopathologic features. All 87 of the samples showed positive immunohistochemical staining compared with the negative controls, i.e., without the primary antibodies. Semiquantitative data are summarized in Table 2, and typical examples of positive staining are shown in Fig. 1. Membranous expression was evaluated to be positive for IGF-IR in 67 tumors (77%), positive for EGFR in 55 (63%), and positive for HER2 in 16 (18%).

IGF-IR expression was significantly more common in intestinal type tumors than in diffuse type (P = 0.002, Mann-Whitney U test; Table 2). HER2-positive tumors were uncommon among diffuse type cancers.

We evaluated the association between protein expression levels and the anatomic extent of disease at the time of gastrectomy using the Japanese classification (20) to define pathologic stage. Pathologic stage did not correlate with the expression of IGF-IR, EGFR, or HER2 in primary tumors (Table 2).

Expressions of IGF-IR, EGFR, and HER2 according to response to first-line chemotherapy. In patients given S-1 monotherapy as first-line treatment, no significant associations were found between tumor response and the protein expressions of the primary tumors assessed according to a four-grade scale (IGF-IR, P = 0.87; EGFR, P = 0.23; HER2, P = 0.50; Mann-Whitney U test). In patients who received cisplatin + irinotecan as first-line chemotherapy, there were also no associations between tumor response and protein expressions (IGF-IR, P = 0.91; EGFR, P = 0.39; HER2, P = 0.48; Mann-Whitney U test). Other first-line regimens were not examined because the number of patients who responded to treatment was too small.

Expressions of IGF-IR, EGFR, and HER2, clinical characteristics, and overall survival since the start of first-line chemotherapy in all patients. The overall median survival time in our study was 14.1 months. Patients with advanced GC who had IGF-IR–positive tumors had slightly poorer survival (Fig. 2A). EGFR expression was unrelated to overall survival (Fig. 2B). HER2 expression was also unrelated to overall survival (Fig. 2C).

On univariate Cox regression analyses, no clinical characteristic significantly correlated with overall survival. A multivariate Cox regression analysis showed that IGF-IR–positive expression (hazard ratio 2.14, 95% confidence interval 1.20-3.82; P = 0.01), performance status 1 or 2 (hazard ratio 1.83, 95% confidence interval 1.15-2.91; P = 0.01), and diffuse type tumors (hazard ratio 1.71, 95% confidence interval 1.08-2.70; P = 0.02) were significant predictors of poor survival (Table 3).

Coexpressions of IGF-IR, EGFR, and HER2 and overall survival. Positive coexpression of IGF-IR and EGFR was found in 48 tumors (55%), that of IGF-IR and HER2 in 16 (18%), and that of EGFR and HER2 in 13 (15%; Table 4). Although a definite correlation was not found among the four-grade scores of these protein expressions, IGF-IR expression weakly correlated with EGFR and HER2 (Table 4).

No combination of protein expressions was significantly related to overall survival. Coexpression of IGF-IR and EGFR (n = 48) was associated with a median overall survival of 13.2 months, and patients with negative expression of either or both of these proteins (n = 39) had a median overall survival of 14.7 months (P = 0.50, log-rank test). Coexpression of IGF-IR and HER2 (median overall survival, 16.0 months n = 16 versus 13.6 months in other patients n = 71; log-rank P = 0.31), and that of EGFR and HER2 (median, 13.2 months n = 13 versus 14.1 months in other patients n = 74; log-rank P = 0.65) were also not significantly related to overall survival.

In our study, 67 of the 87 cases (77%) of advanced GC were positive for IGF-IR expression on immunohistochemical assay, and such expression was a significant independent predictor of poor outcomes, as were well-recognized prognostic factors, such as performance status and the histologic type of tumor. In contrast, the expression of EGFR or HER2 was unrelated to overall survival. The frequency of EGFR expression (63%) in the present study was higher than those in previous studies (range, 31-47%) using immunohistochemical techniques to evaluate GC (21–23), whereas the frequency of HER2 expression (18%) was in agreement with previous findings (range, 10-23%; refs. 24–27). Although EGFR expression in GC remains poorly understood, a possible reason for our high frequency of EGFR expression was that our study was underpowered. Differences in EGFR expression among studies may also be ascribed to the lack of an established immunohistochemical scoring system commonly used to evaluate GC.

Because reports that the IGF system is involved in cancer progression, angiogenesis, metastasis, and resistance to apoptosis, IGF-IR has received considerable attention as a potential target for cancer therapy (28–31). During the past few years, intensive efforts have been directed toward the development of anti–IGF-IR drugs, such as receptor-specific blocking monoclonal antibodies and small molecule tyrosine kinase inhibitors. Two phase I/phase II clinical trials of new monoclonal antibodies are now under way (32, 33), as are many phase I and preclinical trials of other monoclonal antibodies and tyrosine kinase inhibitors. In GC, evidence supporting an association of IGF-IR expression with clinicopathologic characteristics and survival remains scant thus far. Our study showed a high rate of IGF-IR–positive expression in GC and provided evidence that such expression is related to poor outcomes. Our findings suggest that membranous IGF-IR expression in GC might provide a good target for therapy with specific monoclonal antibodies or tyrosine kinase inhibitors, potentially contributing to improved outcomes of patients.

We showed that 55% of tumors coexpressed IGF-IR and EGFR and that all tumors that were HER2 positive (18%) also expressed IGF-IR (Table 4). Although there was no significant association between coexpression of these proteins and overall survival, previous studies have suggested that interactions among families of growth factor receptors augment the malignant behavior of tumor cells (8, 34). Furthermore, cross-talk between IGF-IR and EGFR or HER2 has been implicated in the development of tumor cell resistance to therapy with EGFR inhibitors and anti-HER2 monoclonal antibodies (35, 36).

Recently, combined use of IGF-IR–targeted therapy with anti-EGFR or anti-HER2 therapeutic strategies has been shown to synergistically enhance antitumor activity in vitro (37, 38). Approaches for targeting EGFR or HER2 have successfully been used to treat colorectal and breast cancers (39–41). In GC, the results of phase II clinical trials of cetuximab (anti-EGFR monoclonal antibody), trastuzumab (anti-HER2 monoclonal antibody), or lapatinib (a dual tyrosine kinase inhibitor of EGFR and HER2) were reported in 2007 (42–44). Because antagonists of IGF-IR, EGFR, and HER2 will be approved for the clinical treatment of GC in the near future, there is an urgent need for a better understanding of the clinical significance of IGF-IR expression and for further clarifying interactions between IGF-IR, EGFR, and HER2 at the molecular level. Simultaneous targeting of IGF-IR and EGFR or HER2 (or both of these receptors) may be a useful new strategy for antineoplastic treatment, potentially helping to solve the pressing problem of how to overcome the resistance of tumor cells to these new drugs.

Patients who had IGF-IR–positive tumors or diffuse type tumors showed slight trends toward poorer survival (Fig. 2A; Table 3). However, IGF-IR expression was more common among intestinal type tumors than diffuse type tumors (Table 2; 88% versus 68%). This paradoxical finding may have made the result of univariate analysis insignificant in our study. The trend toward poorer survival in patients with diffuse type GC is consistent with the findings of previous studies (45). Other possible reasons why the relation between IGF-IR expression and survival was insignificant on univariate analysis, but significant on multivariate analysis, were an underpowered study or the fact that only a single methodology of immunohistochemistry was used.

On the other hand, patients with HER2-positive tumors showed a slight but insignificant trend toward better survival (Fig. 2C). This finding might be related to the fact that HER2-positive expression was uncommon among diffuse type tumors (a well-known fact) which are significantly associated with poorer survival (Tables 2 and 3). Some studies have indicated that HER2 expression is an independent predictor of poor survival in GC (26, 46), whereas others have not (24, 25). Expression of EGFR in GC has also been linked to shorter overall survival, more advanced tumor stage, and lymph node metastases in some studies, but not in others (23, 47–49). EGFR expression was not related to any of these factors in our study. The results of immunohistochemical assays can be affected by many variables, including tissue fixation, choice of primary antibodies, and scoring systems, potentially leading to conflicting relations between the expressions of growth factor receptors and clinical outcomes. Immunohistochemical scoring systems of IGF-IR and EGFR differ among studies (16, 23, 49). Even for HER2, it remains unclear whether the scoring system used for breast cancer is valid for GC. The utilization of fluorescent in situ hybridization as an adjunct in GC or the automated quantification of immunohistochemical results may circumvent the subjective nature of immunohistochemical analyses. Nonetheless, a common scoring system for GC should be urgently established to accurately predict the clinical response to antagonists of these receptors, as well as to predict patient survival.

In conclusion, our study provides evidence that IGF-IR expression in GC specimens, poor performance status, and diffuse type cancer are significant predictors of poor survival in patients with advanced GC. We also showed that coexpression of IGF-IR and EGFR or IGF-IR and HER2 is relatively common in GC. Because the expression of IGF-IR has been associated with resistance to anti-EGFR and anti-HER2 therapies (35, 36), the potential therapeutic benefits of simultaneously targeting such receptors in patients with GC should be critically evaluated. Taken together, our findings suggest that anti–IGF-IR strategies may prove valuable in patients with GC.

No potential conflicts of interest were disclosed.

We thank Hideko Morita, Hiromi Orita, and Mari Araake for help in collecting and organizing the clinical samples and Hideko Morita for providing excellent technical assistance.