Background
Many efforts have been focused in better understanding the mechanisms of malignant transformation, resulting in the identification of molecules playing a crucial role in tumor growth. The race to discover compounds that specifically inhibit these targets is giving promising results, and many of these drugs successfully entered clinical trials, opening the era of the "targeted therapies" [
1].
Cancer is a multigenic disease arising from the accumulation of different alterations of genes controlling cell proliferation and/or apoptosis [
2]. However, recent studies in preclinical models demonstrated that tumor cells may be dependent on a single oncogene for their proliferation and survival. In fact, the specific inactivation of that oncogene leads to apoptosis of cancer cells and to tumor regression. This phenomenon, known as "oncogene addiction" [
3], provides a further rationale for the use of targeted therapies. However, only a fraction of patients respond to these therapies, even if the molecular target of the drug is present in the cell. Moreover, almost invariably, responsive patients develop pharmacological resistance and undergo relapse, often due to the activation of alternative signaling pathways [
4]. One of the major challenges of targeted therapies is, therefore, to know in advance which pathways could mediate resistance to the treatment and to find ways to circumvent these hurdles.
Gastric cancer is the second leading cause of mortality in the world and the first one in Asia. Despite the improvement of surgical techniques and the recent availability of new chemotherapic regimens, the outcome of patients with clinical advanced disease is usually poor. The identification of molecules altered in gastric cancers has led to the possibility of hitting them by use of specific targeted drugs. Among them is the receptor for Hepatocyte Growth Factor (HGF), encoded by the
MET gene, that promotes a complex biological program called "invasive growth", inducing cells to break intercellular junctions, acquire a motile/invasive phenotype and escape apoptosis [
5]. The improper activation of this program, due to MET deregulated activation, confers proliferative and invasive/metastatic ability to cancer cells [
6]. Recent studies demonstrated that MET plays a role in a high percentage of human tumors [
7]. In gastric cancers this receptor is frequently constitutively activated; activation is usually associated with receptor overexpression, that can be due to gene amplification. Moreover, MET activation can also result from infection of gastric cells by Helicobacter Pylori, a known predisposing factor for development of gastric cancer.
We and others have shown that gastric cancer cells bearing amplification of the
MET gene and overexpression of the receptor, are "addicted" to this oncogene, since its inhibition results in impairment of tumor growth [
8‐
10]. On these bases, MET is considered a good target in gastric cancer.
Recently, molecules targeting MET have gained access to clinical trials and results are expected soon [
11]. Experience acquired from other RTKs has shown that only a percentage of patients respond to targeted therapies, even in the presence of the altered molecular target, and that almost invariably also responding patients develop resistance during treatment. Therefore, we were interested in identifying pathways whose activation could vicariate the signaling driven by MET. Several studies have shown the presence of a biochemical and functional interplay between MET and the HER (Human Epidermal Receptor) family of RTK (reviewed in [
12,
13]). This family of receptors is frequently altered in gastric cancers where they are constitutively activated, mainly as consequence of gene amplification. Moreover, in patients with advanced gastric cancer, co-expression of c-Met and HER2 has been associated with poorer survival compared to overexpression of either one [
14].
In our work we show that in gastric cancer cell lines "addicted" to MET, activation of HER family members, through ligand stimulation or mutational activation, contributes to overcome MET inhibition. This is due to the partial overlap of downstream signaling pathways common to MET and HER family. Moreover, we provide evidence that resistance to MET inhibition generated in cell lines by treatment with high doses of PHA-665752 is largely due to HER members overexpression.
Discussion
The clinical experience derived from use of drugs targeting molecules that play critical roles in human tumors has shown that their efficacy critically depends on the presence of the altered target in the neoplasm [
26]. However, even in these conditions, a response to the inhibitor is seen only in a fraction of patients (primary resistance) [
27]. Moreover, even in responding patients in which the drugs are initially successful in impairing tumor growth, their efficacy decreases or is abrogated in a short time period, due to appearance of "secondary resistance" [
4]. Most commonly, primary resistance is due either to constitutive activation of pathways downstream to the targeted molecule or to the engagement of alternative or redundant parallel signaling pathways that vicariate the lack of signal due to target inhibition. Secondary resistance can be due either to the same mechanisms, or to genetic alterations of the target, such as gene amplifications (rendering the amount of available drug not sufficient to block the target) or the appearance of point mutations (that prevent the interaction between the target and the drug). The recent availability of drugs that simultaneously inhibit multiple targets or the possibility to perform association therapies able to block synergistic signal transduction pathways has underlined the importance of identifying these functional and biochemical interactions, potentially involved in the appearance of resistance to targeted drugs.
Gastric cancer is an aggressive cancer, constituting a major cause of cancer-related deaths worldwide. Even if traditional therapies such as surgery, chemotherapy and radiotherapy have improved in recent years, patients with advanced disease have a poor prognosis, with a 5-year survival of less than 30%. For this reason, there is an absolute need for the integration in the treatment of this cancer of new drugs, targeting the genetic lesions present in the tumor. Molecular analyses performed in gastric cancer samples have shown that among the genes frequently altered in this tumor are tyrosine kinase receptors of the MET and HER families. The MET gene has been shown to be amplified in human gastric cancers and gastric cancer cell lines; amplification is known to be responsible for receptor overexpression and ligand-independent constitutive activation. Activating mutations have also been identified in some tumors of this histotype [
6]. The role of the
MET gene in human tumors has been firmly established [
11] and it has also been demonstrated that genetic alterations of
MET can be selected for the long-term persistence of the transformed phenotype as gene amplification is more frequent in metastatic lesions rather than in primary tumors [
28]. Moreover, "in vitro" and preclinical models have shown that tumor gastric cells displaying
MET gene amplification are "addicted" to the constitutive activity of this receptor for their growth and maintenance [
8‐
10], thus suggesting that patients affected by this cancer could be ideal candidates for anti-MET targeted therapies. Indeed, clinical trials evaluating the effect of MET inhibition in these patients are ongoing [
11]. It is also very puzzling to note that Helicobacter Pylori, a well known risk factor for this neoplasm, requires MET activation to exert its pro-tumorigenic effects [
29]. Several reports have also identified in gastric cancers quantitative and qualitative alterations of members of the HER family, the most frequent being gene overexpression and amplification, even if also activating mutations have been detected [
30,
31]. Clinical trials targeting HER family members are thus ongoing in patients affected by gastric cancers [
32]. It is important to note that in patients with advanced gastric cancer, co-expression of c-MET and HER2 has been associated with poorer survival compared to overexpression of either one[
14]. These data thus suggest that, in some cases, co-targeting of both these molecules could be of clinical importance.
Several experimental evidences suggest the existence of biochemical and functional interplays between the members of the HER family and MET. Moreover, recent studies have shown that resistance to Gefitinib can be due to
MET amplification [
33]. In this case, MET overexpression and constitutive activation leads to HER3 trans-phosphorylation and activation of HER3-dependent survival pathways. In these cells, co-inhibition of MET and EGFR reverted resistance to Gefitinib. Since MET plays a role in mediating resistance to EGFR inhibition, we wondered if also the reversal was true. Some works have shown that,
in vitro, activation of HER family members can lead to MET phosphorylation, but the role of this interplay has never been evaluated
in vivo and in the contest of cells resistant to MET inhibitors [
34,
35].
As works conducted on other RTKs highlighted the ability of laboratory models to identify clinically relevant mechanisms of drug resistance, the aim of our work was to try to evaluate, in vitro and in preclinical models, the possible role of HER family receptors in mediating primary resistance to MET inhibition.
We took advantage of gastric MET-addicted tumor cell lines that stop proliferating upon treatment with specific MET inhibitors. We found that activation of HER family members in MET addicted cells, after MET inactivation, is able to increase cell viability
in vitro, and to recover tumorigenicity
in vivo. This observation is important if translated into a clinical context. In fact, gastric tumors that display
MET gene amplification (10% of cases) are potentially addicted to MET expression and can be considered ideal targets for anti-MET therapies; however, aberrant activation of HER family members has also been shown to be concomitant in these tumors [
36,
37]. This means that the effect of MET inhibition could potentially be neutralized or attenuated by the parallel activation of receptors of the HER family. This implies that combinatorial inhibition of both MET and HER could likely improve the therapeutic effect. It is critical to underline that not all the growth factor-activated pathways can compensate for the lack of signal due to MET inhibition, as shown by data reported in this paper. Differently from previous observations in HER-addicted cells, the biological effects due to HER members activation was not due to their ability to trans-phosphorylate MET. In fact, the resistance was present not only in cells in which MET was inhibited by the specific small molecule, but also in cells in which the receptor was no longer present - and thus not available for trans-phosphorylation - due to shRNA-mediated silencing. These results suggest that the resistance induced by HER members activation may be rather due to their ability to activate signaling pathways that are critically overlapping with those generated by MET, such as activation of the AKT/MAPK pathways [
20].
Finally, we have generated gastric cells resistant to a MET specific inhibitor and, upon ruling out the presence of MET gene amplification or mutations in either MET itself or other downstream signalling molecules such as RAS, Raf or PI3K (all of them being implicated in the acquisition of resistance to treatment with small molecules inhibitors), we found that the levels of HER2 and HER3 were significantly increased in these resistant cells. Moreover, HER3 silencing led to reversion of the resistance to MET inhibitors and to decreased cell viability. These data suggest that a molecular mechanism exploited by addicted cells to overcome the pro-apoptotic effect of MET inhibition may be the increased expression of HER family members, enhancing the sensibility to their cognate growth factors, which are usually available in the tumour microenvironment.
Competing interests
P.M. Comoglio received consultation fees from Bayer-Shering, Boehringer- Ingelheim, Johnson & Johnson and Servier. The other authors have no competing interests.
Authors' contributions
SC and EG: study concept and design; execution of experiments; acquisition of data; analysis and interpretation of data; drafting of the manuscript. VC and JRS: design and execution of experiments (generation of resistant cell lines). CM: analysis and interpretation of data and statistical analysis. AB, LT and PMC: critical revision of the manuscript for important intellectual content. SG: study supervision and drafting of the manuscript. All authors read and approved the final manuscript.