Introduction
The interest of the Tenovus Centre for Cancer Research in targeting the epidermal growth factor receptor (EGFR) has its origins in the early to mid-1980s when it was first recognised that the EGFR, together with its family member HER2 (c-erbB-2), were found at high levels in approximately 25% of breast cancers where their presence often heralded a more aggressive cancer phenotype and consequently poor patient outlook [1,2]. At that time, it was thought that if the over-expression of these receptor tyrosine kinases was directing unfavourable tumour characteristics, such as increased cell growth and invasiveness, then their subsequent targeting and inactivation might restore a more normal phenotype which would be to the benefit of the breast cancer sufferer. Twenty years later, we are now amassing the pharmacological tools to selectively attack these receptor proteins which include antibodies such as traztuzumab (Herceptin™) and small molecule tyrosine kinase inhibitors such as gefitinib (Iressa™) and it is encouraging that tumour remissions have been observed in breast cancer patients treated with these drugs [3,4].
Disappointingly, however, the recent clinical experience across a range of cancer types including breast cancer is revealing that despite anti-EGFR and anti-HER2 agents demonstrating some anti-tumour activity, there is a high level of de novo resistance to such treatments [5–10], where over-expression of the EGFR and HER2 is clearly not the sole determinant of response to anti-erbB therapies [3,10–12], and moreover, even in responders, the acquisition of resistance and hence disease relapse, occurs fairly rapidly [6–10]. Such adverse phenomena, which serve to limit the overall therapeutic impact of these new agents, implies the existence of a greater complexity involved in EGFR/HER2 signalling network than was previously assumed. Indeed, evidence is accumulating which demonstrates that signalling interplay occurs between the EGFR/HER2 and other growth factor receptors. In this light, the article will focus on the emerging concept of growth factor pathway switching as a means of influencing the effectiveness of anti-growth factor regimes, paying particular attention to the interplay between EGFR/HER2 and the insulin-like growth factor -1 receptor (IGF-1R) where our laboratory observations clearly demonstrate that altered IGF-1R signalling can influence the effectiveness of gefitinib and trastuzumab.
Growth factor pathway switching
Cancer cells are generally regarded as having a high level of plasticity with respect to their use of growth regulatory mechanisms. Stemming from this, it is increasingly evident that despite tumour cells having adopted a dominant growth pathway as a result of the aberrant expression of a particular class of growth factor receptor, they are not necessarily incapable of responding to other paracrine, endocrine or even autocrine growth factors that may be present. Conceptually, therefore, an ability to switch to using these alternative growth signalling pathways during drug treatment represents an obvious means of limiting the actions of highly targeted therapies. As proof of principal of this concept, the substantial anti-tumour response to gefitinib demonstrated by our EGFR positive tamoxifen resistant breast cancer cells, is easily reversed by several non-EGF like growth factors that reactivate cell proliferation and survival signalling, including heregulins and IGFs [13,14]. Thus, although gefitinib is able to efficiently reduce EGFR signalling onto MAPK and AKT in these cells to inhibit their growth [15,16], IGFs acting through the IGF-1R readily reactivate these signalling molecules and thereby re-establish tumour cell growth [14]. Clearly, an in vivo cellular environment rich in non-EGF like growth factors which have the capacity to impinge on growth-related pathways, would be expected to protect cancer cells against agents which specifically target the EGFR (and HER2) and compromise their anti-tumour properties.
Significantly, the phenomenon of growth factor pathway switching is not just a passive process dependent upon the availability of alternative growth factors arising in a paracrine or autocrine manner. We have recently demonstrated that it contains a strong adaptive component that may be directly induced by the selective inhibition of a previously aberrant growth factor pathway. In our experience this often involves increased signalling through the IGF-1R which appears a common mechanism of promoting resistance to either EGFR or HER2 blockade [17]. For example, our in vitro models of acquired gefitinib resistance in breast cancer have shown that despite showing good initial growth inhibitory responses to gefitinib, chronic exposure of both our tamoxifen and fulvestrant (Faslodex™) resistant cell lines to the inhibitor resulted in the eventual emergence of gefitinib resistant variants which showed up-regulation of multiple IGF-1R signalling components and increased sensitivity to IGF-1R inhibition with respect to growth [17,18]. Indeed, the involvement of the IGF-1R signalling in mediating the acquired gefitinib resistant phenotype has also been shown in other model systems generated from prostate and lung cancer cells [18,19] which highlights its importance across multiple cancer types. Interestingly, although IGF-1R signalling plays a role in the maintenance of growth in the established acquired resistant phenotype, its ability to prevent apoptosis and promote survival [20] may also be critical in limiting the effectiveness of anti-EGFR/HER2 strategies in the early phase of use of the inhibitors. This appears to stem from the ability of EGFR/HER2 to suppress signalling from other growth factor receptors such as the IGF-1R, where the subsequent blockade of EGFR/HER2 would relieve the inhibition on the suppressed pathway to efficiently promote its signalling and instigate tumour cell growth. As such, we have shown that although total IGF-1R expression remains unaltered, exposure to gefitinib can induce a marked increase in the activity of the IGF-1R in A549 lung cancer cells which demonstrate partial sensitivity (approximately 40% growth inhibition) to the inhibitor [19]. Mechanistically, this may originate from the ability of EGFR to sequester IRS-1, a classically described adaptor protein for IGF-1R as we have shown that although challenge of our tamoxifen resistant breast cells with gefitinib immediately results in a decrease in activated IRS-1 at the 896 tyrosine residue [21] which is known to bind Grb-2, leading to MAPK signalling [22], it never the less simultaneously promotes IRS-1 activity at the 612 tyrosine residue [21] which is known to result in the formation of a pYMxM consensus motif for PI3-kinase binding and subsequent activation of AKT [22], a well-established potent survival signal [23]. Indeed, evidence is accumulating which indicates that an association exists between elevated IGF-1R expression/signalling and its downstream components such as AKT and resistance to drugs which inhibit erb family signal transduction in various cancers. For example, IGF-1R via PI3-kinase/AKT activation has been shown to mediate resistance to the selective EGFR tyrosine kinase inhibitor AG1478 in glioblastoma cells [24] and also the anti-EGFR monoclonal antibody 225 in DiFi human colorectal cancer cell line [25].
Together with the increased dependency on IGF-1R signalling, we have observed that the acquistion of resistance to gefitinib also promotes cross-resistance to traztuzumab. Thus although we have shown that our HER2 positive tamoxifen resistant breast cancer cells are substantially growth inhibited by this anti-HER2 agent, sensitivity to trastuzumab is lost in our gefitinib resistant variant [17]. Paradoxically, this occurs despite the fact that they have elevated levels of HER2 expression and activity [17]. Importantly, IGF-1R signalling has also been implicated in modulating responses to trastuzumab [26] and it has also been observed that IGF-1R can uni-directionally activate HER2, which involves a physical association of the two receptors [27]. Similarly, our gefitinib resistant breast cancer cells also show evidence of the existence of a physical interaction between the IGF-1R and HER2, which furthermore, co-localise at the tumour cell membranes [17]. Interestingly, we have also observed that vice versa, the acquisition of resistance to traztuzumab is accompanied by a cross-resistance to gefitinib and moreover, the traztuzumab resistant breast cancer cells also show an up-regulation of components of IGF-1R signalling [28]. Such data further adds to the observation that the IGF-1R is increasingly becoming a key therapeutic target to potentially improve the efficacy of erbB inhibitors.
Growth factor pathway switching can occur not only as a result of changes such as those seen during the development of acquired resistance, but can also occur very rapidly and may modulate initial sensitivity to EGFR inhibitory agents. Thus, together with the facilitation of IGF-1R activity, following blockade of EGFR signalling [19], we have recently identified an important reverse element to this pathway switching that involves a capacity of the IGF-1R to transphosphorylate the EGFR on specific tyrosine residues to further contribute to gefitinib resistant growth, with this reverse signalling mechanism again being evident in the partially responsive A549 cells and which contain high IGF-1R signalling [19]. Pre-existence of IGF-1R signalling, or its rapid induction by anti-EGFR strategies, may thus act to efficiently limit the anti-tumour activity of gefitinib and most likely, trastuzumab.
Future therapeutic implications stemming from growth factor pathway switching
An appreciation of the phenomenon of growth factor pathway switching may in the future, allow us to devise strategies to improve the anti-tumour properties of drugs such as gefitinib and trastuzumab. As such, we envisage that the targeting of these additional growth factor pathways may extend the efficacy of anti-EGFR and anti-HER2 strategies and possibly even create responses in previously unresponsive cells. Certainly, we have already shown that the acquisition of gefitinib resistance in our tamoxifen resistant breast cancer cells can be delayed or even prevented by the combination of gefitinib plus an IGF-1R inhibitor, confirming that IGF-1R is a key compensatory cell survival mechanism acquired during gefitinib resistance [29]. Indeed, it has also been shown that growth inhibition and the apoptosis-inducing effects of gefitinib can be enhanced by the co-inhibition of IGF-1R signalling in a range of breast cancer cell lines [30]. Co-targeting the EGFR and the IGF-1R has also been effective in synergistically sensitising glioma cells to apoptosis [31] and recently, the blockade of IGF-1R signalling with the humanised anti-IGF-1R antibody h7C10 was shown to enhance the anti-tumour effects of vinorelbine and the anti-EGFR antibody 225 in human breast and lung cancer xenografts [32]. It is interesting to note that a bispecific antibody, known as Di-diabody, that targets both the IGF-1R and EGFR has been developed and has been shown to possess in vivo anti-tumour effects in colorectal and pancreatic carcinoma xenografts [33].
In conclusion, a body of preclinical evidence is accumulating which implicates the signalling interplay between the EGFR and the IGF-1R as playing an important role in mediating responses to anti-EGFR/HER2 strategies. It is our hope that the identification of key players involved in growth factor pathway switching will ultimately lead to the more productive use of anti-growth factor agents in the treatment of breast cancer and the lessons learnt in this important disease will find general applicability in other prevalent cancer types.
Acknowledgements
The authors wish to acknowledge the Tenovus Centre for Cancer Research and AstraZeneca for financial support. Thanks to the Tenovus Tissue Culture Unit and Immunocytochemistry Unit for technical assistance. Helen E. Jones, Julia M.W. Gee and Robert I. Nicholson are in receipt of funding from AstraZeneca, the latter additionally being a member of an advisory board for AstraZeneca.
