EGFR mutations and HER2/3 protein expression and clinical outcome in Chinese advanced non-small cell lung cancer patients treated with gefitinib

Introduction

Non-small cell lung cancer (NSCLC) represents a major cause of cancer death worldwide (Jemal et al. 2006), with platinum-based chemotherapy being the standard treatment for patients with advanced disease (Schiller et al. 2002). Although new chemotherapy strategies have been developed, the prognosis of patients with inoperable disease remains extremely poor (Schiller et al. 2002). Several new molecules interfering with specific cancer-related pathways have been studied in the last few years, and promising results have been observed with agents targeting the epidermal growth factor receptor (EGFR) (Mendelsohn and Baselga 2000). The EGFR family includes four distinct transmembrane receptors, EGFR/erbB-1, HER2/erbB-2, HER3/erbB-3, and HER4/erbB-4 (Mendelsohn and Baselga 2000; Salomon et al. 1995). Ligand-induced EGFR activation triggers phosphorylation of the intracellular EGFR tyrosine kinase domain and ultimately results in tumor cell survival, proliferation and invasion (Asahina et al. 2006; Cappuzzo et al. 2003). Gefitinib (ZD 1839, Iressa^®, AstraZeneca, UK) and erlotinib (OSI 774, Tarceva^®, Genentech, US) are orally active EGFR tyrosine kinase inhibitors (TKIs) that have been extensively studied in NSCLC (Arteaga 2002), with erlotinib gaining Food and Drug Administration approval for the monotherapy treatment of NSCLC patients who have failed at least one chemotherapy regimen.

EGFR-TKIs have demonstrated significant antitumor activity in approximately 10% of unselected pretreated NSCLC patients (Thatcher et al. 2005; Shepherd et al. 2005). Retrospective analyses of large phase II and III trials with either gefitinib or erlotinib have shown a significant association between the response to EGFR-TKIs and female gender, adenocarcinoma histology, Asian ethnicity and, most of all, never-smoking history (Thatcher et al. 2005; Shepherd et al. 2005).

Biological mechanisms underlying TKI sensitivity and resistance have been extensively investigated. Several reports indicate that the presence of EGFR gene mutations, particularly exon 19 deletions, and EGFR gene gain, both of which are associated with a never-smoking history, predict response to TKIs (Lynch et al. 2004; Paez et al. 2004; Tsao et al. 2005). Whereas assessment of the relationship between EGFR protein expression and response has generated conflicting reports (Hirsch et al. 2006; Cappuzzo et al. 2003; Pao et al. 2005). Conversely, the presence of mutations in the K-ras gene and in EGFR exon 20 have been linked with resistance to these agents (Pao et al. 2005; Kobayashi et al. 2005).

HER2 is the major partner for EGFR, and preclinical studies have shown that HER2-overexpressing cancer cell lines or xenografts are sensitive to the inhibitory effects of gefitinib (Lenferink et al. 1998; Moasser et al. 2001). However, in a small retrospective study of gefitinib in NSCLC patients, no association was observed between HER2 expression, assessed by immunohistochemistry (IHC), and efficacy of the drug (Cappuzzo et al. 2003). Nevertheless, a recent Italian retrospective study in a cohort of NSCLC patients treated with gefitinib has shown that increased HER2 gene copy number, evaluated by fluorescence in situ hybridization (FISH), is associated with a higher response rate and prolonged time to progression (TTP), with a trend toward longer survival (Cappuzzo et al. 2005a, b, c).

HER3 is commonly overexpressed together with HER2 in a variety of cancers (Naidu et al. 1998; Krahn et al. 2001), and has been found to modulate drug resistance in HER2-overexpressing tumors (Chen et al. 2000). In addition, Anido et al. (2003) have demonstrated that gefitinib inhibits the growth of HER2-overexpressing cancer cells, possibly by sequestration of HER2 and HER3 receptors in an inactive heterodimer configuration with the EGFR. Nevertheless, Engelman et al. (2007) recently highlighted an additional mechanism of acquired or pre-existing resistance to gefitinib treatment in NSCLC. This involves proto-oncogene c-MET amplification loading to a HER3-dependent PI3K/Akt persistent activation. In this model, beyond the number of c-MET copies, the levels of phosphorylated HER3 and Akt following gefitinib treatment would be negatively related to the clinical outcome. Therefore it appears that no convincing data are available on HER3 protein expression as a possible predictor for TKI activity in the clinical setting.

Based on the above data, we performed a retrospective study aimed at identifying the role of different EGFR gene mutations and HER2 and HER3 expression as predictive factors for gefitinib sensitivity in a cohort of Chinese pretreated NSCLC patients.

Patients and methods

Results

Discussion

EGFR TKIs have shown remarkable activity in subsets of NSCLC patients, with a clear benefit being observed in women, adenocarcinomas, Asian ethnicities and, most of all, never-smokers (Thatcher et al. 2005; Shepherd et al. 2005; Lynch et al. 2004; Paez et al. 2004). The molecular mechanisms underlying sensitivity to these agents have been extensively investigated. In particular, the presence of EGFR activating mutations has been associated with response to either gefitinib or erlotinib (Lynch et al. 2004; Paez et al. 2004; Tsao et al. 2005). Not surprisingly, the incidence of these mutations was significantly higher in clinical subgroups who were more likely to respond to TKIs (Lynch et al. 2004; Paez et al. 2004; Tsao et al. 2005).

In this study, EGFR mutations were observed in about one-third of our cohort of unselected NSCLC patients, this is within the range previously reported in Asian populations treated with TKIs, and exon 19 deletions were the most common subtype (Lynch et al. 2004; Paez et al. 2004; Hirsch et al. 2006). We found a significant association between EGFR mutations and adenocarcinoma histology, while never-smoking status did not predict the presence of EGFR genetic alterations, an observation supported by other investigators (Huang et al. 2004). These data might indicate that selected Asian populations present a high incidence of EGFR activating mutations regardless of smoking status.

As previously reported by several other authors (Lynch et al. 2004; Paez et al. 2004; Tsao et al. 2005), we observed a significant association between EGFR mutations and higher ORR, longer TTP and OS. However, although available data suggest that patients with EGFR mutations are sensitive to TKIs, some observations indicate that such genetic alterations may have a prognostic rather than a predictive value. Retrospective analyses of the largest trials with TKIs showed no survival benefit for patients harboring EGFR mutations receiving a TKI (Shepherd et al. 2005; Eberhard et al. 2005). In addition, retrospective analyses of the INTACT trials, which compared chemotherapy plus gefitinib versus chemotherapy alone in advanced NSCLC patients, showed a significant survival advantage for patients with EGFR mutations treated with chemotherapy alone compared to individuals with EGFR wild-type, suggesting a possible better natural history for tumors harboring EGFR mutations. Indeed, Shepherd and Tsao remarked a growing evidence that EGFR mutations are favorable prognostic markers of survival and that they are predictive markers of response, in terms of tumor shrinkage. Therefore, the subgroup of patients with mutations, clearly appears biologically distinct (Shepherd and Tsao 2006). Indeed Haneda et al. showed that patients affected by adenocarcinoma with bronchioloalveolar carcinoma components with EGFR gene mutation have a significantly better survival than those with adenocarcinoma with wild type (Haneda et al. 2006).

In our study TTP was longer in patients with EGFR/HER1 mutant as compared to wild type gene also in non-responders to gefitinib. These results appears to cast doubts whether EGFR/HER mutations have a prognostic value rather than predictive on the response to gefitinib.

Both patient’s clinical characteristics and molecular tumor markers may be used to predict the likelihood of response to EGFR TKIs therapy. In fact, in almost all studies, mutations were seen more frequently in females, lifetime non-smokers, and patients with adenocarcinoma (Shepherd 2008). Nevertheless, at this time, molecular selection of patients for therapy with EGFR TKIs is not quite ready. The results of prospectively designed randomized trials are awaited.

Moreover, retrospective studies suggested that different mutations might play different roles in patients exposed to TKIs in terms of response and survival. Particularly, patients with deletions have been reported to have a better response to TKIs when compared with other mutations (Jackman et al. 2006; Riely et al. 2006). In addition, longer TTP and OS have been observed for patients with exon 19 deletions over those with exon 21 substitution (Jackman et al. 2006; Riely et al. 2006). In our study, no difference in terms of response and survival was observed across the different mutational classes, although patients with exon 19 deletions had significantly longer TTP than those with other types of mutations, with a trend for prolonged TTP favoring exon 19 deletions over L858R substitution. Prospective trials should be performed to further define the prognostic and predictive roles of different EGFR mutations.

Existing data indicate that a significant fraction of patients with EGFR mutations (12–84%) does not respond to TKIs (Hirsch et al. 2006; Cappuzzo et al. 2005a). Moreover, dramatic responses can be achieved in patients with wild-type receptor (Shepherd et al. 2005). In our study, ten patients with wild-type EGFR achieved PR. These observations suggest that mechanisms other than EGFR mutations can be involved in sensitivity to TKIs. Several studies reported an improved outcome for patients with EGFR gene gain, regardless of the methods adopted for its assessment (Tsao et al. 2005; Hirsch et al. 2006; Cappuzzo et al. 2005a; Cappuzzo 2008). Furthermore, preclinical data indicate that other members of the EGFR family play a critical role in driving sensitivity to TKIs (Naidu et al. 1998; Krahn et al. 2001; Hirata et al. 2005).

In our study we observed an higher evidence of EGFR/HER1 mutation in non-smokers (20 vs. 35.2%) and female patients (37.7 vs. 23.6%) even these difference did not reach statistical significance, probably as a consequence of the limited number of patients differently from the findings of a recent paper with a more consistent number of patients (Wu et al. 2007).

Overexpression of HER2 has been associated with increased gefitinib activity in preclinical models (Moasser et al. 2001). Moreover, Hirata et al. (2005) have recently demonstrated that effects of gefitinib are influenced by HER2 levels rather than EGFR expression in one NSCLC cell line transfected with HER2 gene. Increased HER2 gene copy number assessed by FISH has been associated with significantly better ORR and prolonged TTP, and with a trend toward longer survival, in a cohort of NSCLC patients exposed to gefitinib (Cappuzzo et al. 2005a). To the best of our knowledge, only one small Italian study investigated HER2 protein expression assessed by IHC in NSCLC patients treated with gefitinib, observing no influence of HER2 levels on clinical outcome (Cappuzzo et al. 2003). In our study, HER2-positive patients had significantly better ORR than those negative for HER2 expression, regardless of EGFR mutational status, suggesting a possible key role of HER2 in driving TKI sensitivity.

HER3, which is commonly overexpressed together with HER2 in a variety of malignancies, has been implicated with TKI sensitivity due to its ability to couple EGFR to the PI3K-Akt pathway (Naidu et al. 1998; Krahn et al. 2001; Engelman et al. 2005). However, increased HER3 gene copy number assessed by FISH did not predict a better response or a longer survival in NSCLC patients who received gefitinib (Cappuzzo et al. 2005a). Fujimoto et al. (2005) evaluated HER3 protein expression by IHC in 42 NSCLC patients treated with gefitinib. The authors reported higher levels of HER3 in patients who responded or achieved disease stabilization to gefitinib as compared with those who had disease progression. Similarly, in the present study we observed a higher ORR for HER3-positive patients when compared with HER3-negative, although the difference was not statistically significant. Finally, when combining EGFR mutational status with HER2 and/or HER3 expression, we observed that EGFR mutated patients had a better outcome compared with patients with wild-type EGFR regardless of HER2 and/or HER3 protein levels. Taken together, these data indicate that presence of EGFR mutations represents the most powerful biological predictor for sensitivity to TKIs.

A significant fraction of patients who achieve initial tumor response to TKIs will ultimately develop progressive disease. Molecular mechanisms involved in resistance to TKIs are currently under investigation. The threonine-to-methionine substitution at amino acid position 790 (T790M) in exon 20 of the EGFR gene has been described in progressing lesions under gefitinib or erlotinib treatment (Kobayashi et al. 2005). The T790M substitution seems to modify the structure of the EGFR tyrosine kinase domain, thus conferring resistance in EGFR mutant cell lines usually sensitive to TKIs (Kobayashi et al. 2005; Greulich et al. 2005). We examined three patients whose primary tumors had an EGFR mutation and had developed resistance to gefitinib after exhibiting an initial response to the drug, we did not find any secondary mutation (data not shown). This finding suggests that mechanisms of acquired resistance to gefitinib in patients with EGFR-mutant NSCLC requires additional investigation.

Overall, our data confirm that the presence of EGFR mutations is a useful marker for selecting advanced NSCLC patients who are likely to benefit from gefitinib regardless of their subtype. These findings are emphasized by recent results of perspective trials (Asahina et al. 2006; Inoue et al. 2006; Paz-Ares et al. 2006; Sutani et al. 2006; Yoshida et al. 2007; Cappuzzo et al. 2007; Sunaga et al. 2007; Sequist et al. 2008). Evaluation of either HER2 or HER3 expression does not add significantly predictive power to EGFR mutation analysis.