Prince of Wales Clinical School and Lowy Cancer Research Centre, Faculty of Medicine, University of New South Wales, Australia.
Frequent overexpression of the epidermal growth factor receptor in colorectal cancer was the rationale for the development of anti-epidermal growth factor receptor antibodies. The development of the drug cetuximab, led to considerable expectations in terms of clinical and commercial success. The registration of the anti- epidermal growth factor receptor antibodies, cetuximab and panitumumab, was granted on the basis of improvement in progression free survival. Other drugs targeting the epidermal growth factor receptor, such as the oral tyrosine kinase inhibitors, have minimal efficacy in colorectal cancer when used alone, and are too toxic when combined with chemotherapy. Cetuximab and panitumumab have activity only in patients with metastatic disease who have a reasonable performance status. Retrospective analyses of tumour samples collected from trial enrolees showed the presence of KRAS mutations in exon 2 were a negative predictor of response to the anti-EGFR antibodies. Recent data suggests that patient selection should be based on a more extensive analysis of KRAS, NRAS, BRAF and potentially other genes. The anti-EGFR antibodies have been used alone or in combination with other chemotherapies, however use with oxaliplatin appears to compromise patient outcomes. When used as monotherapy, toxicities include rash and fatigue, however more severe adverse effects are observed when used with chemotherapy. Anti-epidermal growth factor receptor treatments for colorectal cancer, demonstrate the complexity of using targeted treatments. They remain a useful treatment in colorectal cancer but have not fulfilled their initial expectation of being highly effective and non-toxic treatments.
Targeted therapies were expected to deliver a new treatment paradigm for cancer, characterised by improved efficacy and reduced treatment related toxicity. The term ‘targeted therapy’ encompasses a variety of treatments, including biologic agents such as monoclonal antibodies and oral tyrosine kinase inhibitors (TKI). The targets are molecules involved in the initiation or progression of cancer. Some drugs, like the anti-epidermal growth factor receptor (EGFR) antibodies, interact with one receptor. Other drugs, like TKIs, are promiscuous because they have multiple off-target effects. Also, the targets may be found in specific tumour types or in many. In the case of colorectal cancer (CRC), EGFR was identified as a potential therapeutic target, as the receptor is overexpressed as a consequence of gene amplification and other mechanisms.1,2 Over a decade ago, the development of the anti-EGFR antibody cetuximab (also called Erbitux and C225), raised hopes that the era of targeted therapy for CRC had arrived. Since then, a number of clinical trials have clearly demonstrated that anti-angiogenic and anti-EGFR targeted therapies deliver a modest (around three months) delay in progression free survival for patients with metastatic CRC. Unfortunately, these agents have not had any impact on the adjuvant treatment of this disease. The story of targeted therapies for CRC and in particular, the controversies associated with the development and use of anti-EGFR antibodies, highlight many lessons about targeted therapies and help us place their value in perspective.
Cetuximab is a recombinant chimeric human-mouse antibody, which acts by blocking and down regulating EGFR, and promoting the killing of targeted cells by antibody-dependent, cell-mediated cytotoxicity, and complement fixation.3 Cetuximab was initially developed with the support of public funding in 1985.4 In 1994, permission was sought by its patent owner, the New York based biotechnology company ImClone Systems Inc, to conduct human trials. Given the limited treatments for patients with CRC, the successful anecdotes and preliminary data from studies of cetuximab were met with enthusiasm from scientists and clinicians alike.5 The market also responded positively and the share prices for ImClone soared.6 On October 29th 2001, a $2 billion deal between Bristol-Myers Squibb and ImClone came into effect: the CEO of Imclone, Samuel Waksal and his brother Harlan, sold $111 million of their stock in the company. In December 2001, the first application for Food and Drug Administration approval for cetuximab in second line therapy for metastatic CRC was rejected. Among other things, the experts who reviewed the protocol on which the application was based, considered the study was fundamentally flawed and inappropriate for a registration trial. In May 2002, the data on cetuximab was presented at the American Society of Clinical Oncology (ASCO) meeting and therapeutic targeting of EGFR was discussed by Dr John Mendelson at the prestigious Karnofsky lecture. No mention was made of the controversy surrounding the clinical development of cetuximab, yet Sam Waksal resigned as CEO of ImClone a few days following the ASCO meeting in the midst of investigations by the US Department of Justice and Securities and Exchange Commission. Eventually, both Sam Waksal and the business celebrity, Martha Stewart, were convicted of charges related to ImClone insider trading.7,8 In 2004, Food and Drug Administration registration approval of cetuximab was granted on the basis of the phase III BOND trial, which showed an improvement in progression free survival in patients with metastatic CRC treated with cetuximab. Perhaps reassured by the success of using HER2 over-expression to identify breast cancer patients for treatment with Trastuzumab, the FDA approved a companion diagnostic for qualitative immunohistochemistry (IHC) for the detection of EGFR.9 It was proposed that overexpression of EGFR could predict those patients likely to benefit from cetuximab. Unlike Trastuzumab, the relationship between anti-EGFR antibodies and overexpression of EGFR has not stood the test of time.
In 2008, a fully humanised recombinant anti-EGFR monoclonal antibody, panitumumab, was granted Food and Drug Adminstration approval on the basis of an improvement in progression free survival in a non-blinded phase III study that compared panitumumab with best supportive care in patients with treatment refractory CRC. The development of panitumumab has in many respects mirrored cetuximab. However, unlike cetuximab, a benefit in overall survival has not been demonstrated in any trials of panitumumab, although this has been explained by the fact that patients in the control arm were allowed to cross-over to panitumumab on progression.
The small molecule TKIs, such as erlotinib and gefitinib, interact with the intracellular domain of the EGFR. These orally available drugs have not proved to be effective in CRC. Erlotinib and gefitinib as single agents have shown minimal activity in metastatic CRC.10,11 Clinical trials of combinations of erlotinib and cytotoxic chemotherapy, with or without the vascular endothelial growth factor inhibitor bevacizumab, were closed prematurely because of significant toxicities.12 Erlotinib has also been investigated in combination with other treatments in different schedules, including maintenance and intermittent dosing schedules, however no benefits in response rate, progression free survival or overall survival have been observed.13,14
A number of factors are influential in selecting patients for treatment with anti-EGFR antibodies. These include the molecular profile of the tumour, disease stage and the performance status of the patient.
Although it might be intuitive to propose that over-expression of EGFR is a likely biomarker for response to the anti-EGFR antibodies, this assumption has not proven to be correct. Specifically, response to antibody therapy does not correlate with EGFR over-expression as determined by immunostaining, somatic EGFR mutations or EGFR gene amplification. Certainly patients lacking EGFR expression have responded to cetuximab and panitumumab.5,15 The signal transduction downstream of the EGFR receptor includes multiple cell signalling pathways connected through complex cross-talk and feedback loops.3 Most cancers show mutations in one or more of the downstream signalling pathways such as RAS-RAF-MAPK and PI3K-AKT. Preclinical studies show that tumours with mutations in KRAS have downregulation and suppression of upstream EGFR signalling and are thus refractory to drugs which block EGFR. This observation has been validated in the clinic by a retrospective analysis of KRAS status in tumours from patients enrolled in clinical trials of cetuximab or panitumumab. This analysis showed that patients with KRAS mutant tumours did not derive a benefit from either antibody.16-18 Since KRAS mutations are present in about 40% of CRCs, analysis of KRAS status has proven to be a useful, if imperfect, negative predictor of response to the anti-EGFR antibodies.16
The results of a pooled meta-analysis of 11 trials of anti-EGFR antibodies is summarised in table 1.16 This analysis shows that 50-65% of patients with KRAS wild type CRC remain resistant to the anti-EGFR antibodies. This finding could be explained by tumour heterogeneity, the presence of undetected less common KRAS mutations, or other RAS mutations such as in NRAS. Until very recently, all analyses have confined mutation testing to codons 12 and 13 of exon 2 of KRAS. There is now increasing evidence that gene mutations in NRAS, BRAF (V600E), or PI3K (PIK3CA), or loss of PTEN expression are associated with lack of response to anti-EGFR antibodies.19,20
The benefits of targeted therapies in CRC remain confined to individuals with metastatic disease. Targeted therapies have failed to improve outcomes when used in the adjuvant setting. There are several possible explanations for this. Firstly, the increased toxicity of combination treatment may lead to decreases in the dose of cytotoxics or premature discontinuation of adjuvant therapy. Further, some investigators have postulated that there is a negative interaction between some chemotherapy agents and anti-EGFR antibodies in patients with micro-metastatic disease. Finally it has been suggested that in the adjuvant setting, tumour cells are undergoing epithelial-mesenchymal transition and are not dependent on EGFR signalling.21
Age was not used as part of the selection criteria for most clinical trials of the anti-EGFR antibodies, however enrolment was restricted to patients with a functional status of ECOG 2 or less. The side-effects of the anti-EGFR antibodies include skin reactions, including acneiform rash (in up to 90% of patients), dry skin, pruritus and nail changes, diarrhoea, infusion related reactions (including hypersensitivity reactions), cardiac events and hypomagnesaemia. A high degree of fatigue and asthenia has also been reported. When used in combination with chemotherapy, severe side-effects are more frequent. In the adjuvant N0147 trial, elderly patients had more toxicity.22 Thus the goal of minimal side-effects from a targeted treatment has not been realised in the case of the anti-EGFR antibodies, and performance status is very influential in selecting patients for treatment.
After years of separate development, a direct comparison of the efficacy of cetuximab and panitumumab in the chemo-refractory setting was reported this year.23 Overall survival, progression free survival and response rates were similar in this non inferiority study. Much has been made of the chimeric nature of cetuximab when compared with panitumumab, with claims that increased immunogenicity would be both positive in promoting a stronger immune and tumour response, and negative, with increased side-effects. While panitumumab has less infusion related side-effects than cetuximab and a different administration schedule, otherwise the drugs are currently interchangeable. This year, the first phase III trial to prospectively test for the KRAS mutation status to determine patients’ randomisation and treatment in chemo-refractory CRC was reported.24 The PICCOLO study revealed no overall survival benefit when panitumumab was added to irinotecan, when compared to irinotecan alone in patients known to be wild type for KRAS (without the common mutations in codons 12, 13 and 61). The secondary endpoints of progression free survival and response rates were improved in the combination arm. A grouped analysis for mutations such as BRAF, NRAS, KRAS (codon146) and PIK3CA was performed and showed that treatment outcomes for this group were compromised compared with patients without mutations.
A number of studies have noted a negative interaction between cetuximab and oxaliplatin. In all studies with oxaliplatin as the chemotherapy partner (OPUS,26 COIN,27 NORDIC VII,28 N014729 and PETACC 830), none were able to reach a significant result for the primary study end points. This observation was confirmed in a meta-analysis for both cetuximab and panitumumab, when used in combination with oxaliplatin based chemotherapy, showing no benefit in survival or response rates.31 Possible explanations for this include negative interactions for oxaliplatin on the action of cetuximab and vice versa. Src has been observed to be activated by oxaliplatin in CRC cell lines suggesting a possible mechanism of induced resistance through activation of a downstream signalling pathway to the EGFR receptor.32 A suggested mechanism for the latter is reduced effectiveness of oxaliplatin in KRAS wild type CRC cells when treated with cetuximab. Oxaliplatin efficacy relies on an intracellular redox reaction which is inhibited by the cetuximab EGFR interaction through the Nox1 pathway.33 Supporting these observations, the PACCE trial showed that the addition of panitumumab to bevacizumab and FOLFOX (oxaliplatin, fluorouracil and leucovorin) chemotherapy increased toxicity and decreased progression free survival.34
Having shown some activity as a single agent or as a partner with irinotecan in the treatment refractory metastatic CRC setting, the anti-EGFR antibodies were tested in first line and adjuvant settings. Cetuximab in combination with first line FOLFIRI (irinotecan, fluorouracil and leucovorin) chemotherapy has been shown to have a small statistically significant improvement in progression free survival in unselected patients (8.9 vs 8.1 months). The progression free survival outcome is enhanced by selecting KRAS wild type patients for the combination treatment (9.5 vs 8.1 months).35 The OPUS study used overall response rates as the primary endpoint and found in a retrospective subset analysis that there was a 2.54 times increase in response rates in patients with KRAS wild type tumours receiving cetuximab and FOLFOX (oxaliplatin, fluorouracil and leucovorin). When given as neoadjuvant treatment prior to liver resection, the rate of R0 resection achieved was doubled, however this was based on small patient numbers (6/61 vs 3/73 patients).26,25
The duration of antibody treatment has been studied as a continuous therapy until progression or unacceptable toxicity in all the randomised trials. A recent study in patients whose disease progressed after initial response, found acquired new KRAS mutations in six out of 10 cases, suggesting that this is a common resistance pathway to anti-EGFR treatment.36 Another study suggested that treatment with anti-EGFR antibodies select resistant clones present in heterogenous tumours at the outset, and that the time to recurrence is simply the interval required for the subclone to repopulate the lesion.37 This raises the question as to whether continuous treatment with anti-EGFR antibodies is advisable.
The long history of development of anti-EGFR therapies for CRC illustrates the conceptual fallacy of the single target-single treatment model. It was tantalising to think that EGFR inhibitors targeting the commonly overexpressed EGFR would translate to a paradigm shift in CRC therapy. In fact, the EGFR has proven to be an inexact target, with downstream signals compromising the efficacy of anti-EGFR antibodies. For CRC, anti-EGFR antibodies help a minority of patients and every day more contraindications, such as the presence of other RAS and BRAF mutations, are being uncovered. Optimists may see an expected role for the anti-EGFR antibodies within select patient populations, yet others may question the declining marginal benefit of such treatments in an era of limited public resources and competing cancer care needs.
Anti-EGFR antibodies are effective in the treatment of metastatic but not early CRC.
Patients are selected for treatment on the basis of the molecular profile of their cancer.
The effectiveness of standard chemotherapy maybe reduced when used with anti-EGFR antibodies.