Understanding BRAF-Mutant Colorectal Cancer

Understanding BRAF-Mutant Colorectal Cancer

By Jeremy C. Jones, MD, and Axel Grothey, MD

Article Highlights

  • Patients with BRAF-mutant (BRAFMut) colorectal cancer (CRC) have a poor prognosis and an unmet need for improved therapeutic options.
  • Patients with BRAFMut CRC are generally older and female, and their tumors are more likely to have microsatellite instability, be higher grade, have more lymph node involvement, and have more advanced primaries.
  • BRAF mutational status is a strong predictor for overall survival not only in the metastatic setting but also in earlier-stage diagnoses.
  • Rational clinical trial designs are likely the key to exploiting BRAF mutations as targets for treatment rather than as markers of poor prognosis.

Patients with BRAF-mutant (BRAFMut) colorectal cancer (CRC) have a poor prognosis and, as a group, clearly have an unmet need for improved therapeutic options. Appropriately identifying and treating patients with BRAFMut CRC remains a particularly challenging clinical undertaking. The need for further understanding of this disease’s pathogenesis, pathways of metastasis, and novel approaches to treatment is only amplified by the aggressive clinical phenotype with which the disease manifests. During the past decade, we have seen far-reaching advances in the field of targeted therapeutics; yet, the success that has been realized by targeting BRAF in other malignancies, such as malignant melanoma, remains elusive in CRC. This article will serve to review the pertinent epidemiologic, pathologic, and clinical characteristics of BRAFMut CRC, with a focus on therapeutic challenges and the clinical trials designed to address these difficulties.                

Overall, BRAF mutations are found in approximately 7% of all human cancers, but the relative frequency with which they occur varies widely among a number of different malignancies.1 For instance, malignant melanoma has a BRAF mutation frequency of approximately 50%, whereas CRCs harbor such a mutation in only approximately 10% of cases.2 

Biology of the MAPK Signaling Pathway

The oncogenic BRAF mutation was initially described in 2002 by Davies et al.1 from the Cancer Genome Project after finding that approximately 66% of malignant melanoma cell lines harbored a specific mutation within the kinase domain of BRAF that allowed cells to grow in the absence of external signaling. Since this time, more than 100 unique BRAF mutations within exon 11 and 15 on the long arm of chromosome 7 have been described.3 The most common of these, the V600E mutation, occurs in exon 15 and results in a substitution from valine to glutamic acid at position 600 within the BRAF kinase domain. This mutation disrupts the kinase function, which leads to constitutive activation and downstream signaling along the MAPK signaling pathway.

Under normal physiologic conditions, the MAPK signaling pathway is activated when any of a number of extracellular growth signals is transduced via membrane-bound, receptor-linked tyrosine kinases. These kinases lead to activation of RAS (rat sarcoma) proteins (KRAS, NRAS, and HRAS) that, in turn, activate RAF (rapid accelerated fibrosarcoma) serine/threonine kinases (BRAF, CRAF, and ARAF), which ultimately leads to downstream activation of MEK (mitogen-activated protein) and, finally, ERK (extracellular signal-regulated kinases). In whole, this pathway plays an intricate role in translating extracellular growth signals into nuclear changes, including transcription-factor activation, DNA replication, and histone modifications critical for cellular proliferation and survival (Fig. 1).4

Detection and Screening

Although there are a number of known mutations that occur within the BRAF gene, the vast majority of our clinical and laboratory understanding in CRC is derived from patients with the V600E mutation. As next-generation sequencing becomes more commonplace, we have found a number of non-V600 BRAF mutations that have unclear clinical consequences.5 In a sample of patients from the Mayo Clinic with metastatic CRC who had next-generation sequencing performed on their tumor specimens, we found that non-V600 mutations made up approximately 20% of all BRAF-mutant tumors. Many of these mutations have been examined in laboratory models and have been found to have various effects on kinase activity, including activation of, no effect on, or in some cases decreased kinase activity.6 In one of the only reports of clinical outcomes in patients with non-V600 BRAFMut CRC, Cremolini et al.7 describe 10 patients with codon 594 and 596 BRAF mutations. They found that patients with these mutations had markedly different molecular features, pathologic characteristics, and clinical outcomes compared with patients who had a V600 mutation. Further investigation is needed in patients with these rare mutations to better elucidate which mutations confer the same prognostic significance as V600E.

Clinical Presentation

When compared with patients who have non-mutated disease, patients with BRAFMutCRC are generally older and female. Their tumors are more likely to have microsatellite instability, be higher grade, have more lymph node involvement, and have more advanced primaries. BRAFMuttumors also have a striking predilection for proximal (i.e., right-sided) tumor locations, as well as early peritoneal and distant lymph node metastasis.8,9

Prognosis and Management

Fig. 1.
BRAF mutational status is a strong predictor for overall survival not only in the metastatic setting but also in earlier-stage diagnoses. In the adjuvant setting, patients with BRAFMut CRCs tend to have shorter disease-free survival and worse overall survival after recurrence. Roth et al.10 examined specimens from 1,404 patients with stage II-III colon cancer who were enrolled on the PETACC-3 trial for BRAF and KRAS mutations. They found that patients with BRAFMuttumors had significantly worse overall survival rates than those with wild-type tumors. These differences were even greater when patients were stratified by microsatellite stable versus instable because of the relatively good prognosis associated with the latter. Interestingly, there was no significant difference in disease-free survival between patients with BRAFMut and patients with wild-type disease, which suggests that the poor overall survival was driven largely by mortality after relapse and/or metastasis. Similar findings were reported by Sinicrope et al.11 in their report of 2,720 patients with stage III disease enrolled on the North Central Cancer Treatment Group (NCCTG) N0147 adjuvant study. In this trial, BRAF mutation was prognostic of overall and disease-free survival.  Again, these differences in survival were particularly clear when microsatellite instability was accounted for.

In the oligometastatic setting, patients with isolated liver metastases of BRAFMut CRC who undergo curative resection achieve, on average, fewer than 6 months of relapse-free survival and experience significantly worse overall survival than patients with BRAF wild-type cancers.12,13 Furthermore, in the noncurative, metastatic setting, outcomes are even more dismal. As a whole, median survival for patients with BRAFMut CRC treated with standard cytotoxic chemotherapy is approximately one-half to one-third that of patients with BRAF wild-type tumors.8,14-16BRAF mutational status is also likely predictive of lack of response to EGFR inhibitors when given without concomitant BRAF inhibition, as discussed below.15,17,18

A recent trial by Loupakis et al.19 has shown promising outcomes for patients with BRAF-mutated tumors treated with a highly aggressive chemotherapy regimen (i.e., FOLFOXIRI plus bevacizumab). In this trial, patients with metastatic CRC who received FOLFOXIRI plus bevacizumab achieved progression-free survival that was approximately 2.5 months longer than in those who were treated with FOLFIRI. Unfortunately, even with this regimen, survival for patients with BRAFMut CRC was still less than half that of patients with wild-type tumors (median overall survival, 19.0 months vs. 41.7 months).

Targeted Therapies

The relative lack of durable response to standard cytotoxic agents, along with the successes found in treating BRAFMut melanoma, has led to a multitude of studies designed to exploit BRAF as an important driver mutation. One of the first of these trials, which used single-agent vemurafenib (a selective BRAFV600E inhibitor) in 21 patients with BRAFMut CRCs, yielded surprisingly poor results.20 Only one patient achieved a partial response, which equated to an overall response rate of approximately 5% compared with single-agent response rates for the same drug used in melanoma, for which overall response rates are generally closer to 60% or 75%.21,22

A number of insights were gained from this trial. First, it became abundantly clear that mutational status alone was not sufficient to predict response to targeted therapies. Second, it provided a focal point for contrast of the cellular mechanisms at play in two disparate malignancies that are thought to be driven by a shared mutation. In fact, when researchers examined levels of phosphorylated ERK (pERK, the downstream effector of BRAF activation) in melanoma and CRC cell lines treated with BRAF inhibitors, they found significantly higher levels of pERK in CRC cell lines, suggesting inadequate suppression of the MAPK pathway via BRAF inhibition alone.23,24

Table
A second trial paired the BRAF inhibitor dabrafenib with trametinib, a MEK inhibitor, in 43 patients with metastatic BRAFMut CRC.25 Four patients (9%) developed a partial response, and one patient (2%) had a durable complete response. Again, despite the disappointing response rates, a number of key points can be taken from this study. First, it is interesting to note that the patient who developed a complete and durable response was the only one to have not received prior systemic therapy. This certainly could be a coincidence, but it could also suggest that cytotoxic therapy may disrupt the cellular milieu at play in BRAFMut CRC feedback activation. Second, unlike the previous study, pretreatment and day-15 biopsies were obtained in nine patients. All nine patients showed decreased pERK signaling on day-15 biopsies compared with baseline. However, the decrease in pERK signaling levels was significantly less for patients with CRC who received dual blockade (median decrease, 37%) than for patients with melanoma who received BRAF inhibitors alone (median decrease, 84%). On one hand, this trial proved that improved clinical response rates can be achieved with more effective blockade of the MAPK signaling cascade; conversely, it further demonstrated the difficulties in fully and effectively blocking MAPK signaling in CRC.

A 2012 Nature article shed light on one potential mechanism of resistance to BRAF inhibition in CRC.23 In it, Prahallad et al. described a dramatic increase in EGFR activation encountered when CRC cells were exposed to BRAF inhibition. This activation led to increased signaling through not only the MAPK pathway but also through other established CRC signaling pathways, namely PI3K and AKT. Indeed, when an EGFR inhibitor was added to BRAF inhibition, there was a marked decrease in MEK-, ERF-, and AKT-mediated signaling. This was presumably accomplished by neutralizing the BRAF inhibitor–mediated activation of EGFR. 

These findings have spurred numerous clinical trials that combine BRAF and EGFR inhibitors (Table 1). Thus far, four such trials including a total of 83 patients are currently ongoing or have recently been completed.28,30-32 Most of these trials remain in early phases, but, as of the most recent analyses, 12 patients (14%) have achieved a partial response by Response Evaluation Criteria in Solid Tumors (RECIST). Reported progression-free survival times have ranged from approximately 3 months to 4 months among all patients. Although response rates and progression-free survival data are still far from optimal, it does appear that, given the continued incremental improvement, dual BRAF and EGFR inhibition represents at least one component of a successful treatment strategy.

A few new trials combining BRAF and EGFR inhibition with a third agent for patients with metastatic BRAFMut CRC are worth noting. The first combines the MEK inhibitor trametinib with dabrafenib and panitumumab. As of the latest update, nine (26%) of 35 patients who received the triplet regimen have achieved at least a partial response.24 The second of these triple-agent trials combines the BRAF and EGFR inhibitors encorafenib and cetuximab with the a-specific PI3K inhibitor alpelisib. The latest update from this trial, given at the 2014 European Society for Medical Oncology Annual Meeting, showed that nine (32%) of 28 patients who received triple therapy achieved at least a partial response.23 The last of these triple-agent trials combined the BRAF and EGFR inhibitors vemurafenib and cetuximab with irinotecan.32 At last update, 17 patients with metastatic BRAFMut CRC were evaluable for response, of which six (35%) had achieved at least a partial response.

Future Directions

During the past decade, we have gained vast amounts of clinical knowledge about BRAFMut CRC; however, at this point, this knowledge has not yet translated into improved patient outcomes. BRAF mutations are clearly pathogenic drivers of aggressive disease; yet, unlike in melanoma, BRAF inhibition alone is likely only a required but not sufficient part of effectively targeting this pathway in CRC. Rational clinical trial designs, such as those discussed earlier, are likely the key to unlocking the ultimate goal of exploiting BRAF mutations as targets for treatment rather than as markers of poor prognosis. Effectively targeting BRAF will likely also require the blockade of various parallel feedback pathways, including, but likely not limited to, EGFR, MEK, and PI3K. One interesting approach that has already shown intriguing results is the combination of BRAF and EGFR inhibition with cytotoxic chemotherapies.32 It is possible that the addition of BRAF and EGFR inhibitors to combination chemotherapy regimens could result in significant improvements in clinical outcomes by neutralizing the driver of the aggressive pathology. 

Given the relatively high rates of microsatellite instability found in BRAFMut CRC, it is certainly plausible that these patients would respond favorably to immunotherapy PD-1 or PD-L1 inhibitors. Could these agents be combined with BRAF and/or EGFR inhibition for a synergistic effect?

Last, an important and often overlooked aspect to better understanding not only BRAFMutCRCs but all solid tumors is the development of longitudinal trials designed to assess the mechanisms of secondary resistance. It is likely that, in evaluating these secondary resistance patterns and mechanisms of resistance, we will not only be able to understand the myriad escape pathways for cellular survival but also potentially overcome them and restore effective cytotoxicity.