CCR 20^th Anniversary Commentary: RAS as a Biomarker for EGFR-Targeted Therapy for Colorectal Cancer—From Concept to Practice

Summary

Clinical data support the use of EGFR mAbs in patients with metastatic colorectal cancer (mCRC) with wild-typeRAS status. This notion, hypothesized in the review article by Camp, Ellis, and colleagues in the January 1, 2005, issue of Clinical Cancer Research, serves as an example of the successful application of basic science principles to clinical practice. The exclusion of patients with mCRC with Ras mutated tumors from therapy with EGFR mAbs has led to improved outcomes while sparing patients unnecessary and potentially harmful therapy.

Introduction

The therapeutic potential of mAbs targeting the EGFR was established by promising pre-clinical studies and subsequent early phase clinical trials in patients with metastatic colorectal cancer (mCRC) (1). Ultimately, in 2004, the FDA approved the human-mouse chimeric EGFR mAb, cetuximab, for use in chemotherapy-refractory patients with mCRC. Similarly, panitumumab, a fully human EGFR mAb, was FDA approved for similar indications in 2006, and ultimately both drugs were subsequently approved in the first line setting in combination with chemotherapy. However, the earliest clinical trials investigating EGFR mAbs in mCRC, demonstrated that these agents only provided modest benefit when used in combination with chemotherapy. Around this time, a concerted effort to identify predictive markers for EGFR targeted therapies was undertaken simultaneous to advances in genomic sequencing.

As principal investigator (PI) of a laboratory, the senior author of this manuscript (L.M.E.) learned early on in his career that the best way to educate a trainee on a topic was to have them write a review article. This strategy exposes trainees to the entire field of the topic of interest, requiring a critical evaluation of the literature, and identifying knowledge gaps. In 2004, the first author of this manuscript (E.R.C.) entered the laboratory and was tasked with writing a review on resistance to EGFR targeted therapies. At this time, EGFR mutations in lung cancer were being identified as markers of sensitivity to EGFR tyrosine kinase receptor inhibitors (2). Our laboratory had been interested in growth factor receptors and downstream intracellular signaling, with a focus on Src in collaboration with G. Gallick at MDACC. In reviewing the literature and applying our own knowledge of cell signaling, the general topics of this review inClinical Cancer Research (CCR) focused on the following as resistance factors to EGFR targeted therapies: 1) the presence of redundant tyrosine kinase receptors,2) increased angiogenic signaling,3) existence of specific EGFR mutations, and 4) constitutive activation of downstream mediators (3). This last proposed mechanism of resistance was not based on actual data, but a simple hypothesis: if an important intracellular pathway is constitutively activated by a mutation, then blocking signaling at the cell surface would be ineffective.

In our original schematic drawing describing this hypothesis, we listed several activated downstream intermediates as potential resistance pathways. One, in particular, has demonstrated clinical utility in selecting patients with mCRC for EGFR mAb therapy. The therapeutic relevance of downstreamRAS mutations on EGFR mAb efficacy in patients with mCRC highlights the importance of understanding the molecular basis of malignant disease in order to improve personalized medicine. We recognize that this concept is the basis of many preclinical and clinical studies: but at the time of writing this review, there were no publications that had discussed or hypothesized that mutatedRAS or other constitutively activated signaling intermediates could serve as markers of resistance (or even detriment) in patients with mCRC. However, at that time there were several studies examining the role of mutatedRAS as a resistance marker for EGFR tyrosine kinase inhibitors in lung cancer (4,5); other investigators had hypothesized thatRAS could be a resistance marker for EGFR targeted therapies, and this occurred simultaneous to, or shortly after, our publication. Soon after the publication of our review in January 2005, there were several studies examining the role ofKRAS in EGFR mAb resistance, suggesting that others had this idea and had already begun studies. However, we have not identified any publication prior to ours in CCR addressing this topic. Of course, since then we have learned that additionalRAS mutations may render patients resistant to EGFR mAb therapy (6).

ConstitutiveRAS Pathway Activation Impact on EGFR mAb Therapy

In 2006, a retrospective study of 30 patients with mCRC treated with cetuximab (mostly combined with chemotherapy) was the first clinical study to correlate KRAS mutational status with resistance to EGFR mAb therapy (7). In this small study, patients with tumors harboring aKRAS mutation demonstrated a 0% response rate. Overall survival (OS) in cetuximab treated patients was significantly shorter in those patients withKRAS mutations when compared with patients with wild-type (WT)KRAS (6.9 versus 16.3 months, p=0.016). A subsequent investigation confirmed the clinical significance ofKRAS mutations as predictive markers for cetuximab or panitumumab therapy for patients with mCRC (8). Furthermore, this study was the first to demonstrate a causal relationship between activatingKRAS mutation (Gly¹²Val) and decreasedin vitro response to cetuximab therapy in human colon cancer cell lines.

A landmark retrospective analysis of a prospective randomized trial in 392 chemotherapy refractory patients with mCRC comparing cetuximab treatment with best supportive care (BSC) solidified the significance ofKRAS mutational status in clinical practice (9). OS in patients with WTKRAS tumors was significantly longer in patients treated with cetuximab compared to patients who received BSC (9.5 versus 4.8 months, p<0.001). In contrast, patients with mutatedKRAS tumors experienced an equivalent OS between cetuximab therapy and BSC. Even more concerning, results from the randomized Phase 2 OPUS trial comparing first-line FOLFOX4 with or without cetuximab in patients with mCRC suggested that the addition of an EGFR mAb may actually have a detrimental effect on outcomes in mCRC patients with mutatedKRAS tumors (10) and other studies have confirmed these findings (6). These initial translational investigations laid the foundation for personalized molecular medicine in mCRC by demonstrating the predictive capability ofKRAS mutational status for improving patient selection and outcomes for EGFR mAbs therapy. In response to the growing evidence supporting the predictive implications ofKRAS status for EGFR mAb therapy, the FDA restricted use of cetuximab and panitumumab to patients with WTKRAS mCRC in 2009.

The Clinical Implication of ExpandedRAS Status on Cetuximab Therapy

The initial studies investigating the role ofKRAS mutation status on the efficacy of cetuximab in patients with mCRC focused on the most prevalentKRAS mutations involving codons 12 and 13 in exon 2 which are present in ~40% of patients with mCRC. Even though multiple studies confirmed that exon 2 mutatedKRAS rendered patients with mCRC resistant to EGFR mAb therapy, EGFR mAbs were found to be effective in some, but not all, patients withKRAS WT tumors; this suggested that alternative molecular abnormalities may be influencing therapeutic response. Extending this line of investigation further, an expanded or “allRAS” mutation analysis including allKRAS andNRAS mutations has been proposed for identifying patients with mCRC who would benefit from EGFR mAb therapy. The first study to identify the negative predictive value of driver mutations, includingNRAS, reported on a large retrospective cohort of chemotherapy refractory patients with mCRC who were subsequently treated with irinotecan plus cetuximab (11). In this study, patients with tumors harboring driver mutations beyond KRAS identified even more patients unlikely to respond to EGFR mAbs, thus improving response rates in those patients with tumors without these mutations. Subsequently, a retrospective analysis of 1,060 patient tumor samples from the randomized PRIME trial in patients with mCRC treated with first-line FOLFOX4 with or without panitumumab highlighted the significance of an expandedRAS assessment to predict response for EGFR mAb containing regimens (6). In this trial, 52% of the analyzed tumors were identified asRAS mutated including anyKRAS orNRAS mutations in exon 2, 3, or 4. Of the original 641 patients categorized as exon 2 WTKRAS, an additional 108 (17%) had an alternativeRAS mutation. Patients with tumors expressing anyRAS mutation who received panitumumab with FOLFOX4 experienced a significantly decreased progression-free survival (PFS) as well as OS compared to patients with WTRAS tumors. The survival outcomes were similar between exon 2KRAS mutations and tumors with alternativeRAS mutations suggesting that anyRAS mutation equally negates the therapeutic benefit achieved with EGFR mAbs. However, the numbers of patients with these less frequentRAS mutations limits the ability to firmly state that they are indeed markers of resistance. Similar to the findings from the OPUS trial, patients with anyRAS mutation actually experienced a significantly decreased PFS and OS with the addition of panitumumab to FOLFOX4 reinforcing the concern that EGFR mAbs may actually be detrimental in this setting. Similarly, a recent meta-analysis of nine randomized, controlled trials with EGFR mAb for mCRC patients evaluatingRAS mutational status further solidified the predictive impact of an expandedRAS mutational profile on PFS as well as OS (12). In general, expandedRAS analysis may increase the percent of patients ineligible for EGFR mAb therapy by ~15 % compared toKRAS exon 2 mutations alone. Based on the potential detrimental impact of anyRAS mutation in patients receiving EGFR mAb therapy, the FDA endorsed an extendedRAS assessment for EGFR mAb therapy in 2013 and, subsequently, the National Comprehensive Cancer Network (NCCN) published clinical guidelines recommending prospective expandedRAS mutational assessment for all mCRC patients considered candidates for EGFR mAb therapy.

The failure of efficacy with EGFR mAbs in a significant number of patients with WTKRAS tumors fueled further investigations evaluating alternative downstream mediators of EGFR. The first report to extend the molecular analysis beyondRAS in patients with mCRC investigated the role of mutatedBRAF^v600e as a marker of cetuximab resistance in WTKRAS tumors (13).BRAF is a downstream mediator in theRAS pathway, but of course can be constitutively activated as it is mutated in ~5–10% of CRCs. In this study, patients with tumors expressing theBRAF^v600e mutation experienced a shorter PFS and OS in response to cetuximab compared with WTBRAF tumors. Although initial reports suggested thatBRAF might be a predictive marker for EGFR mAb, subsequent studies have not always confirmed this association. For example, in the retrospective analysis of the PRIME trial, patients with tumors expressing theBRAF^v600e mutation demonstrated an equally poor PFS and OS in response to FOLFOX4 alone compared with the addition of panitumumab (6). In contrast, the randomized PICCOLO trial investigating the role of irinotecan plus panitumumab in fluorouracil-refractory patients with mCRC demonstrated a significantly worse OS if patient’s tumors harbored aBRAF mutation. Despite the inconsistent results regarding BRAF as an EGFR mAb predictive marker,BRAF mutational status appears to have strong negative prognostic significance as highlighted several trials. Compared to patients with WTRAS/BRAF tumors,BRAF^v600e mutations conferred a ~50% reduced PFS and OS in both arms of the trial. Given the aggressive biology associated withBRAF^v600e mutation, the actual impact regarding the therapeutic efficacy of EGFR mAbs may be difficult to determine. Similar toBRAF, PIK3CA mutations and loss ofPTEN have demonstrated an inconsistent impact on the therapeutic efficacy of EGFR mAbs. Currently, testing forBRAF andPIK3CA/PTEN is not recommended for EGFR mAb mCRC patient selection, although these mutations/alterations may be of benefit in determining prognosis and eligibility for clinical trials.

The improved objective response with the addition of EGFR mAb observed in “allRAS” wild type tumors highlights the potential of molecular-driven personalized cancer therapy. The role for prospectiveRAS mutational analysis for cetuximab selection has been explored in recent cooperative group clinical trials, which analyzed a comprehensive (“all-RAS”) molecular profile to assess the benefit of using cetuximab in combination with FOLFIRI or mFOLFOX6 (14). The CALGB/SWOG 80405 trial evaluated response to cetuximab containing regimens in 526 “allRAS” WT CRC patients. In the “all-RAS” WT cohort, patients receiving cetuximab plus chemotherapy achieved an improved median PFS and OS of 11.4 and 32 months, respectively compared with historical controls.

Taken together, the evidence is growing that mutations inRAS can predict for resistance to EGFR mAb in patients with mCRC. NCCN clinical guidelines, the FDA, and European Medicines Agency (EMA) mandate restricted use of EGFR mAb for patients withRAS WT mCRC. The foundation of this clinical practice change can be traced back to the fundamentals learned from preclinical studies of signaling pathways. This dramatic change in clinical practice highlights the continued need for adequate funding for translational research in order to improve oncologic outcomes for our patients.