Tumor MET Expression and Gene Amplification in Chinese Patients with Locally Advanced or Metastatic Gastric or Gastroesophageal Junction Cancer Free

More than 40% percent of the world's gastric cancer cases occur in China. Gastric cancer is the third most common cancer in China, with 405,000 estimated new cases in 2012 (1). Furthermore, gastric cancer causes more than 325,000 deaths annually in China (1). Despite recent advancements in therapy, median survival of patients with advanced gastric cancer remains poor at approximately 9 to 12 months (2–9).

Hepatocyte growth factor (HGF), also known as scatter factor, and its receptor MET appear to be promising therapeutic targets in oncology (10). Activation of the HGF/MET signaling pathway promotes the proliferation, migration, and survival of tumor cells (10). It has also been shown to be associated with cancer pathogenesis, invasion, and metastasis (11).

Potential prognostic and predictive biomarkers in advanced gastric cancer have been identified. For example, overexpression of HER2 has been shown to be predictive of outcomes in patients with advanced gastric cancer treated with trastuzumab (2). Furthermore, MET gene amplification and increased MET protein expression have been associated with advanced disease and poor prognosis in metastatic gastric cancer (12–16).

Although others have explored the relationship between MET expression or MET amplification and prognosis in gastric cancer, most of these studies evaluated early-stage patients or mixed-stage populations (13–20), so the prognostic effect of MET in late-stage disease remains unclear. Cancer stage may be an important factor when studying the prognostic effect of MET, as MET is associated with tumor invasion (11). Moreover, prior studies primarily evaluated samples from resected tissue (13–20). Biopsy samples are more relevant in advanced disease, and limited data about MET status in biopsy samples were available in these studies.

In the current study, we obtained biopsy samples from patients with locally advanced or metastatic gastric or gastroesophageal junction cancer who were treated at a large cancer center in China, and we retrospectively tested the samples for MET expression, phospho-MET (p-MET) expression, and MET gene amplification. Our objective was to evaluate whether MET expression and MET amplification are prognostic for overall survival (OS).

Eligible patients had unresectable locally advanced or metastatic gastric or gastroesophageal junction cancer and were enrolled in multicenter phase II/III clinical trials between 2008 and 2010 at the Peking University Cancer Hospital in Beijing, China. Trials included the following, as identified by their ClinicalTrials.gov identifiers: NCT00548548, NCT01041404, NCT00887822, NCT00678535, NCT01015339, and NCT00842491. Biopsy tumor samples were obtained from all patients by endoscopy from the primary gastric cancer site, prior to chemotherapy. The number of metastatic sites was determined by the number of involved organs at trial enrollment. Tissues were fixed with formalin and paraffin-embedded in blocks, which were prepared at the time of biopsy. Patients with available baseline characteristics and clinical data were included in the analysis. All patients provided written informed consent for the trial participation, and the trials were approved by the Institutional Review Board of the Peking University Cancer Hospital. All patients were followed until December 31, 2013.

MET protein expression, MET gene amplification, phospho-MET (p-MET) expression, and HER2 expression were retrospectively evaluated.

MET expression was determined by an automated MET immunohistochemistry (IHC) Investigational Use Only (IUO) assay using antibody clone MET4 (Dako), as previously described (21). Samples were defined as MET positive if ≥25% of tumor cells had membrane protein staining at any intensity (21). MET amplification was analyzed by FISH using the Research Use Only (RUO) MET/CEN-7 IQFISH Probe Mix assay (Dako). Samples were defined as MET amplified in this analysis if the MET:centromere 7 ratio was >2.0. Several cutoffs have been probed to define MET amplification using the Colorado scoring system as a guide (22). One cutoff that appears to define focal amplification is a MET:centromere 7 ratio >2.0 (21). This cutoff also mirrors the HER2 amplification cutoff, which is widely used in gastric cancer. Prior to assay implementation, technicians and pathologists were trained and certified as proficient by Dako.

Phospho-MET expression was analyzed by IHC using a research grade assay that utilized the monoclonal p-MET antibody (Tyr1234/1235; D26; Cell Signaling Technology, Inc.) Immunopositive cases were defined as those exhibiting membrane protein staining in >10% of tumor cells in the sample.

HER2 expression was analyzed by IHC using the PATHWAY anti-HER-2/neu (4B5) rabbit monoclonal primary antibody (Ventana Medical Systems, Inc.) HER2 positivity was defined as IHC 3+ staining alone or the combination of IHC 2+ staining and FISH positive. In cases of IHC 2+ staining, FISH was performed with PathVysion (Abbott Laboratories). Tumor specimens with a HER2:centromere 17 ratio >2.0 were considered HER2 FISH–positive.

All samples were analyzed and evaluated in the Department of Pathology at Peking University Cancer Hospital. Pathologists were blinded to the clinical and molecular characteristics of the patients.

The Fisher exact test was used to evaluate the following: (i) association between MET expression and MET amplification, (ii) association of MET expression and MET amplification with clinical characteristics, and (iii) association between MET expression or MET amplification and p-MET expression. Survival curves for OS were plotted using the Kaplan–Meier method and were compared using the log-rank test. The HR for OS between MET-positive and MET-negative patients was evaluated using the Cox proportional hazards model. Statistical significance was defined as P < 0.05. For the 95% confidence interval (CI) of a proportion, a binomial exact CI was provided.

Overall, 168 patients were eligible for the study. Of these, 155 patients had available tumor samples for testing: 137 patients had samples evaluable for MET IHC, 113 patients had samples evaluable for MET FISH, and 134 patients had samples evaluable for p-MET IHC. A patient flow diagram is shown in Fig. 1. Patient demographics, disease characteristics, and treatment are shown in Table 1. All patients with available tumor samples received first-line systemic chemotherapy for metastatic disease. Besides one patient receiving paclitaxel alone, all other patients received doublet or triplet chemotherapy. Among the 137 patients with evaluable MET IHC, 60% received first-line platinum-based therapy (fluoropyrimidine/platinum or fluoropyrimidine/platinum/taxane), and 53% received first-line taxane-based therapy (fluoropyrimidine/taxane, fluoropyrimidine/taxane/platinum, or taxane alone; Table 1).

Of the 137 patients evaluable for MET IHC, 53 patients (39%; 95% CI, 30%–47%) had MET-positive tumors. Forty-one patients (30%) had tumors with no MET membrane staining of any intensity and 43 patients (31%) had tumors with >0% and <25% of cells with MET membrane staining of any intensity. Of the 113 patients evaluable for MET FISH, 8 patients (7%; 95% CI, 3%–13%) had tumors with MET amplification.

Among the 137 patients with evaluable IHC samples, no significant association was observed between MET expression and clinical characteristics (Table 1). Among the 113 patients with evaluable FISH samples, MET amplification was not associated with sex, primary tumor site, distant metastasis (liver vs. lung), or number of metastases (Table 1). However, among the 111 patients for whom Lauren classification data were available, Lauren classification was associated with MET amplification (P = 0.04; Table 1); seven of 46 diffuse samples (15%) were MET amplified, whereas one of 49 intestinal samples (2%) were MET amplified.

Ninety-five patients had samples evaluable for both MET IHC and MET FISH, and of these, eight patients had MET-amplified tumors. MET expression was significantly associated with MET amplification (P < 0.001). All eight tumor samples with MET amplification had ≥90% MET expression (Fig. 2).

Of the 141 patients with available survival data, the median follow-up time was 10.9 months. Of 15 censored patients with available survival data, the median follow-up time was 35.0 months. Follow-up time was measured from the start date of chemotherapy for advanced gastric cancer.

Of the 137 evaluable patients for MET IHC, 133 had available survival data. In this population, no statistically significant difference in OS was observed between the MET-positive and MET-negative patients (P = 0.80; Fig. 3A). The HR (95% CI) was 1.049 (0.725–1.518), and the median OS times were 12.3 and 10.9 months in the MET-positive and MET-negative patients, respectively (Supplementary Table S1).

We also evaluated the prognostic value of MET expression according to treatment disposition. In patients receiving first-line platinum/fluoropyrimidine therapy (n = 62), there was a nonsignificant trend toward shorter OS in the MET-positive patients versus MET-negative patients (P = 0.12; Fig. 3B); the HR (95% CI) was 1.530 (0.892–2.625), and the median OS times were 10.6 and 11.9 months, respectively (Supplementary Table S1). In patients receiving first-line taxane-based therapy (n = 71), no statistically significant difference in OS was observed between the MET-positive patients and MET-negative patients (P = 0.26; Fig. 3C); the HR (95% CI) was 0.736 (0.431–1.256), and the median OS times were 12.6 and 10.8 months, respectively (Supplementary Table S1). In patients receiving first-line platinum-based therapy (n = 80), no statistically significant difference in OS and no apparent trends were observed between the MET-positive and MET-negative patients (P = 0.37; Fig. 3D); the HR (95% CI) was 1.243 (0.774–1.996), and the median OS times were 11.3 and 11.8 months, respectively (Supplementary Table S1).

Of the 134 patients evaluable for p-MET IHC, eight (6%; 95% CI, 3%–11%) had p-MET–positive tumors. Representative p-MET IHC images are shown in Fig. 4A.

Ninety-six patients were evaluable for both p-MET IHC and MET FISH, and of these patients, seven had p-MET–positive tumors. Of the seven patients with p-MET–positive tumors, all had MET amplification; only one patient with a MET-amplified tumor was not p-MET–positive (Table 2, Fig. 4B). There was a significant association between p-MET expression and MET amplification (P < 0.001).

Of 118 patients evaluable for both p-MET IHC and MET IHC, 8 had p-MET–positive tumors (Fig. 4C). Of the 8 patients with p-MET–positive tumors, all had MET expression. There was a significant association between p-MET expression and MET expression (P < 0.001).

Seventy patients were evaluable for HER2 status and MET expression. Of these, 12 (17%) were HER2-positive and 58 (83%) were HER2-negative (Table 3). MET expression was not associated with HER2 status (P = 0.75). Of the 12 HER2-positive tumors, 6 (50%) showed MET-positive expression.

Forty-four patients were evaluable for HER2 status and MET amplification. Of these, 11 (25%) were HER2-positive and 33 (75%) were HER2-negative; no HER2-positive tumors were MET amplified (Table 3). Likewise, no association was observed between MET amplification and HER2 status (P = 0.56).

To our knowledge, this study is the first to evaluate MET protein expression and MET gene amplification in biopsy samples from Asian patients with advanced gastric or gastroesophageal junction cancer by the MET4 IHC IOU and MET/CEN-7 IQFISH Probe Mix RUO assays. To date, most studies in Chinese patients have employed small sample sizes with patients of variable baseline characteristics (23–25). Our study included 155 patients with available tumor samples, of whom 137 patients were evaluable for MET IHC, and 113 patients were evaluable for MET FISH, all with advanced cancer. Currently, there is no globally recognized standard testing platform for MET. Interpretation of IHC results across studies is hindered by lack of uniform scoring criteria for the different IHC testing methods or MET FISH (14). Recent studies have evaluated MET expression by the Ventana assay (12, 14, 16, 26–29), with MET-positive expression defined as IHC 3+ or 2+ staining in variable percentages of tumor cells. Our study showed a MET-positive expression prevalence rate of 39% by IHC in biopsy samples from Chinese advanced gastric or gastroesophageal junction patients when MET-positive expression was defined as ≥25% tumor cells with membrane protein staining at any intensity using the MET4 antibody.

For the HGF/MET pathway, gene amplification may be a crucial biomarker for MET-targeted agents, especially for small-molecule MET inhibitors. The prevalence of MET amplification in gastric and gastroesophageal junction cancer has been reported in several studies by different methodologies, including Southern blot analysis, quantitative PCR (qPCR), and in situ hybridization (ISH) technologies [FISH or silver ISH (SISH); refs. 12–14, 16, 26–32]. Earlier studies using Southern blot analysis or qPCR reported relatively high rates of MET amplification of approximately 10% (13, 30, 31), but recent studies using FISH or SISH found that MET amplification rates were lower, varying from 2% to 8% (12, 26, 27, 29, 31, 32). In our study, the MET amplification rate was 7%, which is consistent with the 8% rate reported by An and colleagues who evaluated Chinese patients with locally advanced or metastatic gastric or gastroesophageal junction cancer using a similar FISH method (12). The prevalence of MET amplification may be related to disease stage; our MET amplification rate (7%) and that reported by An and colleagues (8%) are relatively high compared with other studies in which the majority of patients had stage I–III disease. For example, a large cohort study by Lee and colleagues, with only 12% of patients having stage IV disease, found MET amplification by SISH in 13 of 381 patients (3%) (26). However, in the same study, MET amplification was observed in four of 41 patients with stage IV disease (10%). Recent studies of AMG 337, INC280, and crizotinib have shown promising tumor response in MET-amplified gastric or non–small cell lung cancers (33–35).

A notable finding of our study was the strong association between MET-positive expression and MET amplification in the biopsy tumor samples (P < 0.001), consistent with other studies (12, 26). In the small number of samples with MET amplification (n = 8), all had ≥90% MET expression. Furthermore, all samples with ≥90% MET expression and valid FISH results showed MET amplification, except one. However, the number of MET-amplified samples is too small to draw strong conclusions. As MET amplification may be an important biomarker for MET-targeted agents, it is crucial to find a convenient and cost-effective approach to detect MET amplification. Further research is needed to investigate whether IHC may act as a potential prescreening method for MET amplification.

We further evaluated p-MET protein expression and the association between p-MET IHC and MET FISH. Phospho-MET expression indicates activation of the MET pathway. Therefore, inhibiting the MET pathway in patients with positive p-MET expression might be a viable therapeutic approach. Moreover, p-MET staining is more convenient compared with MET FISH and may act as a substitute method for choosing patients for anti-MET treatment. To date, several studies have evaluated p-MET expression in gastric cancer (36–39), and the results were inconsistent. The p-MET–positive rate of 6% (8/134) observed in our study is similar to the rate reported by Janjigian and colleagues, who also used the same antibody that recognizes p-MET at Y1234/1235 (37). Of interest, we found that there was a strong association between p-MET expression (IHC) and MET amplification (FISH). All p-MET–positive samples with valid FISH results showed MET amplification, and all samples with MET amplification had valid p-MET results and were p-MET positive, except one. To our knowledge, we are the first to show the potentially strong association between p-MET IHC and FISH in patients with advanced gastric or gastroesophageal junction cancer. More studies are needed to confirm whether p-MET IHC may be an alternative and cost-effective method to FISH to select MET-amplified patients as a prescreening tool.

In the current study, no significant association was found between MET expression and survival, a result that is inconsistent with observations reported by others that patients with MET-positive gastric cancer have poorer outcomes than those with MET-negative disease (14, 21). When we evaluated OS by different treatment subgroups, there was a nonsignificant trend toward shorter OS for MET-positive patients versus MET-negative patients receiving first-line platinum/fluoropyrimidine therapy (n = 62; 10.6 months vs. 11.9 months, P = 0.12).

We did not observe a significant association between HER2 status and MET amplification in our study, which is consistent with other studies (40). Of note, among the 44 patients with samples evaluable for both HER2 status and MET amplification, all four MET-amplified samples were HER2 negative. Our interpretation of this finding is limited by the small sample size, but alteration of either one of these genes may be sufficient to support tumor growth.

In our study, MET amplification but not MET expression was associated with Lauren classification. A majority of MET-amplified samples (seven of eight, 88%) were found in patients with diffuse gastric cancer according to the Lauren classification. These results do not align with the HER2-positive status, which is found predominantly in tumors of intestinal classification (2, 41). The difference in Lauren classification between HER2-positive tumors and MET-amplified tumors provides early evidence that these two tumor types may represent different molecular subtypes.

Several limitations of this study should be noted. This was a retrospective study, and the sample size was small, especially in the subgroups. Tumor biopsy samples were not available for all eligible patients, and all samples were not evaluable for IHC and/or FISH. Moreover, the study was limited by the variable first-line treatments among patients. Conclusions regarding outcomes based on MET expression may be confounded by differences in subsequent therapies and selection bias for patients with available samples.

In our study, no statistically significant survival difference was observed between the MET-positive and MET-negative populations, based on MET protein expression, in Chinese patients with advanced gastric or gastroesophageal junction cancer. A significant association was found between MET gene amplification and MET protein expression as well as between MET gene amplification and p-MET protein expression. Prospective and larger studies where patients receive uniform treatment are needed to further evaluate potential prognostic markers in Chinese patients with advanced gastric or gastroesophageal junction cancer.

K.S. Oliner has ownership interest (including patents) in Amgen, Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Z. Peng, J. Gao, K.S. Oliner, Y.-J. Hei, L. Shen

Development of methodology: Z. Peng, K.S. Oliner

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z. Peng, Z. Li, J. Gao, M. Lu, J. Gong, K.S. Oliner, L. Shen

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z. Peng, J. Gao, M. Liu, E.-T. Tang, J. Gong, K.S. Oliner, Y.-J. Hei

Writing, review, and/or revision of the manuscript: Z. Peng, Z. Li, J. Gao, M. Lu, J. Gong, E.-T. Tang, K.S. Oliner, Y.-J. Hei, H. Zhou, L. Shen

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Peng, Z. Li, J. Gao, M. Lu, J. Gong, E.-T. Tang, L. Shen

Study supervision: Z. Peng

The authors acknowledge Lixin Zhou of Peking University Cancer Hospital for assistance with the sample analysis; Ning Go of Amgen Inc. and Karsten Bork Nielsen of Dako for assistance with the MET IQFISH RUO assay; and Scott Webster of Dako for assistance with the MET IHC IUO assay. Medical writing support was provided by Laura Evans on behalf of Amgen Inc., and Jenilyn Virrey, and Micah Robinson of Amgen Inc.

This work was supported by grants from the National Natural Science Foundation of China (81172110 and 81472789; to L. Shen), the National High Technology Research and Development Program (2012AA 02A 504; to L. Shen), the Beijing Natural Science Foundation (7142034; to Z. Peng), and the National Basic Research Program of China (2014CBA02002; to L. Shen). This work was funded by Amgen Inc.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.