In 2008, there will be approximately 148,810 new cases of colorectal cancer in the United States. In addition, 50,000 individuals die annually from this disease.1 The survival of colorectal cancer clearly correlates with the stage of the disease at diagnosis. Patients with stage I disease have more than a 90% 5-year survival, whereas the survival for metastatic disease is only about 5%.2-4 Approximately 20% of patients have stage IV disease at their initial presentation. Over the past decade, treatment for metastatic colorectal cancer has significantly improved with the approval of several new agents, including 3 targeted agents: bevacizumab, cetuximab, and panitumumab. The median survival of stage IV disease has significantly increased to more than 20 months (more recently approaching 24 months) with the addition of the newer cytotoxic and targeted biologic agents.5 Over the past several years, more research interests have been focused on identifying patients who will benefit from these targeted therapies. This review focuses on the impact of the K-ras mutation in the development, clinical outcomes, and treatment selection of colorectal cancer.
Some of the most significant therapeutic advances in the treatment of cancer have occurred in the management of colorectal metastases. The introduction of new cytotoxic chemotherapeutic and biologic agents has changed the approach to these patients from both an oncologic and a surgical perspective. In addition, an understanding of the molecular mechanisms by which these agents affect tumors is developing. This molecular information will be critical in the future in designing therapeutic regimens based on an individual tumor's genetic profile rather than treatment for a specific tumor type. The rapidly evolving treatment of colon cancer has provided several interesting genetic biomarkers/pathways/genes-/kinases that have been targeted or seem to play an important role. Of particular interest is the blockade of epidermal growth factor receptor (EGFR) with monoclonal antibodies. This treatment is efficacious when used alone or combined with chemotherapy. However, recent data revealed that patients with tumors positive for the K-ras mutation do not benefit from EGFR blockade. Compelling evidence has indicated that mutated K-ras is an important oncogene involved at the early stage of the development of colorectal cancer. Furthermore, mutations in the K-ras gene have been associated with aggressive tumor biology. K-ras mutational analysis is an important step in the overarching goal of developing personalized medicine. New treatment strategies are needed to more effectively treat patients with the K-ras mutation. Cancer 2009. © 2009 American Cancer Society.
K-ras Gene Mutation and the Development of Colorectal Cancer
Ras proteins are GTP-coupled proteins that are important in many receptor tyrosine kinase signaling. Activation of Ras results in initiation of its complex signaling network. One of the earliest defined receptor tyrosine kinase pathways is the Ras/Raf/Erk/Map kinase pathway. Activation of the Erk/Map kinase pathway via growth factors and their receptors ultimately leads to gene transcription and cell proliferation. It is well established that the Erk/Map kinase pathway plays an important role in Ras-mediated oncogenesis.6, 7 Aside from the classic Erk/Map kinase pathway, Ras is a key activator for many other signaling pathways leading to distinct biological effects. For instance, both phosphainostide 3-kinase (PI3 kinase) and Ral proteins are downstream effectors of Ras.8-10 Yet, the biological consequences are very different. The PI3 kinase/AKT pathway mainly regulates cell survival and apoptosis, while the Ral pathways may be related to cell transformation.11, 12 So far, at least 9 pathways have been identified as the result of Ras activation.13 The complexity of the Ras network and cross-talk between its downstream pathways pose challenges in the development of therapeutic strategies.
Mutations in the Ras protein usually cause constitutive activation of Ras GTPase, which leads to overactivation of downstream signaling pathways, resulting in cell transformation and tumorigenesis.14-17 Mutations in the K-ras gene have been found in multiple human cancers including colorectal, lung, and pancreatic cancers.15-18 However, the relationship of the K-ras mutation and colorectal cancer was not delineated until Fearon and Vogelstein established a stepwise hypothesis in 1990.19
The development of colorectal cancer is a multistep process involving cumulative genetic mutations. Oncogenes, tumor-suppressor genes, and chromosomal deletions are required to ultimately transform colonic epithelium. Among those genes, K-ras is considered to be critical components in the development of colorectal cancer.20-22 In clinical specimens, it has been reported that approximately 30% to 50% of colon cancers harbor K-ras mutations.16, 17, 23 Evidence also suggests that the K-ras mutation is an early event in a subset of colorectal cancers. For example, the K-ras mutation has been detected in precancerous aberrant crypt foci of the colonic epithelium.24, 25 Furthermore, transgenic mice model expressing the mutated k-Ras protein using the colonic epithelial-specific promoter, villin 1 promoter, developed aberrant crypt foci-like morphology throughout the colon.26 Clinically, the K-Ras mutation has been found in early adenomatus polyps.27 It has been shown that larger polyps (> 1cm) with severe dysplastic histology usually have a higher frequency of K-Ras mutations.27
In addition to the K-Ras mutation and silencing of p53 and APC tumor suppressors, multiple alleic losses (17q-, 5q- and 18q-) and epigenetic changes, such as DNA hypermethylation, play roles in the development of colorectal cancer.19, 27-29 Fearon and Vogelstein postulated that the K-Ras mutation occurs in a preexisting small adenoma, and through clonal expansion, it generates larger and more dysplastic tumors. Alleic deletions likely occur at a later stage of cancer development.19
K-ras Gene Mutation and Clinical Outcomes
The consequence of a mutated K-ras gene affects multiple aspects of colorectal cancer. Constitutively activated K-Ras not only promotes tumor initiation but also tumor growth, survival, progression, local invasion, metastasis formation, angiogenesis, and even immune response.30 The effect of mutant K-Ras on epithelial cell proliferation is well documented.31 This is supported by the identification of K-Ras mutations in early cancerous lesions in clinical specimens.24, 25 However, it is not clear whether a K-Ras mutation is absolutely required for colorectal cancer proliferation at later stage or in metastatic foci. In addition, mutated K-Ras promotes angiogenesis and, in turn, aids tumor progression. The mutant K-Ras protein also has been shown to up-regulate VEGF through the PI3 kinase pathway in human colon cancer cells.32 Clinically, antiangiogenic therapy with bevacizumab has proven to be effective even in K-ras mutant colorectal cancer.33, 34
Aside from tumor growth and angiogenesis, oncogenic K-Ras also plays a critical role in promoting tumor metastasis. To invade into the local stroma or disseminate from the primary tumor, colorectal cancer cells need to pass through the epithelial basement membrane. Invading colorectal cancer cells produce matrix metalloproteases (MMPs), cysteine proteases, serine proteases, and urokinase plasminogen activator (uPA) that break down and facilitate migration through the basement membrane. Yamamoto et al showed that mutated K-Ras stimulates MMP-7 production, a determinant of metastatic potential.35 In addition, Jankun and colleagues demonstrated that expression of activated K-Ras significantly increased uPA release in fibroblasts.36 Mutant K-Ras is required for high-level expression of the uPA receptor in human colorectal cancer cells.37 Buo et al revealed that the expression of uPA and uPA receptor is often associated with tumor local invasion in human colorectal cancers.38 Conversely, deletion of the endogenous K-ras from HCT-116 colorectal cancer cells markedly reduced cell motility.39 On the basis of these compelling findings, it is reasonable to conclude that constitutively activated mutant K-ras facilitates tumor invasion and metastasis.
The relations among K-ras mutational status, tumor stage, and survival in early studies are not very consistent. Two large studies have been conducted to address this question. In 1998, RASCAL I study enrolled 2721 patients and was reported by Andreyev and colleagues.23 The study showed no association of K-ras mutation with tumor stages. However, the presence of the mutation significantly increases the risk of recurrence and death. Interestingly, the study also found that poorly differentiated cancers less frequently had a mutated in K-ras gene (P < .01). Subsequently, Andreyev and colleagues published the largest study thus far, RASCAL II, in 2001.40 The RASCAL II study collected data from 4268 patients. From this larger cohort, 3439 patients could be evaluated by multivariate analysis. The study demonstrated that a codon 12 mutation, glycine to valine, had a significant impact on disease-free survival (P = .004, hazard ratio [HR], 1.3) and overall survival (P = .008, HR, 1.29). Patients with the mutation had a much poorer prognosis. Further evaluation revealed that this mutation had a greater survival impact on Duke C (disease-free survival, P = .008; HR, 1.5; overall survival, P = .02; HR, 1.45) than Duke B cancers (disease-free survival, P = .46, HR, 1.12; overall survival P = .36; HR, 1.15). On the basis of these 2 studies, the authors hypothesized that the presence of the codon 12 mutation, glycine to valine, in K-ras gene may predict a more aggressive biological behavior of the cancer.
In addition to these 2 large studies, there are several smaller studies that investigated the impact of the K-ras mutational status and colorectal cancer.41-44 Some studies showed positive relationship between K-ras mutation and advanced stages while others did not. A common criticism of these studies focused on the nonstandardized methods for detection of K-ras mutations. The methods used in these studies varied widely, which included restriction fragment length polymorphism (RFLP), single-stranded conformational polymorphism (SSCP), allele-specific hybridization, and direct sequencing of polymerase chain reaction (PCR)-amplified genomic DNA. Recently, Conlin et al took tumor samples from 107 patients with colorectal cancer treated with surgery from November 1997 to December 1999 in Tayside, Scotland and subjected them to K-ras mutation analysis. Clearly, tumors with K-ras mutation had a significantly poorer overall survival (P = .0098).45 Multivariate analysis after adjusting for stage, age, and sex confirmed that K-ras mutation carriers had a much higher risk of dying from their disease (HR, 2.9, confidence interval [CI], 1.4-6.2, P = .004). Interestingly, the presence of APC and p53 mutations did not affect survival.
In 2000, Samowitz et al reported the first large US study assessing K-ras mutation in 1413 individuals with colon cancer.46 Approximately 32% tumor had either codon 12 or 13 mutations. Codon 12 mutations were found to be much more common in proximal tumors. The study also demonstrated that the codon 12 mutation was associated with an advanced stage of colon cancer, but it did not correlate with cancer-related deaths. The authors also determined that a codon 13 mutation had a 40% increase in short-term mortality from colon cancer. This piece of data is partially consistent with the study published by Bazan et al.47 Bazan and colleague showed that codon 12 mutation was associated with mucinus histology, whereas codon 13 is associated with advanced Duckes stage. Collectively, the effect of K-ras mutational status on clinical outcomes in patients with colorectal cancer has not been well defined. Although, there is some evidence suggesting that tumors with the K-ras mutation may have a more aggressive biologic behavior. In addition, the relationship between the K-ras mutation and tumor stage is also not conclusive.
K-ras Gene Mutation and Epidermal Growth Factor Receptor Targeting
Epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase that controls cell growth, differentiation, and transformation. It is ubiquitously expressed in human tissues. Upon activation of the EGFR, the receptor undergoes dimerization and activation of its kinase activity. Subsequently, the downstream signaling cascade is initiated, which includes activation of Ras-GTPase and Erk/Map kinase resulting in cell proliferation. EGFR expression is detected in up to 80% of colorectal cancers.48, 49 Tumors with a high level of EGFR expression usually have poor prognosis.50 These data have led to the clinical use of EGFR blockade in the treatment of colorectal cancer with cetuximab. Cetuximab is a chimeric monoclonal antibody directly against the extra cellular domain of the EGFR and blocks the signal cascade through the EGFR. In 2004, Cunningham et al showed that addition of cetuximab to irinotecan achieved a 20% response rate (RR) in irinotecan-refractory patients. This suggests that cetuximab can overcome irinotecan resistance.51 Similarly the CRYSTAL trial published by Van Cutsem and colleagues, demonstrated a significant increase in response rate when cetuximab was combined with FOFIRI in the first-line setting (38.7% for FOLFIRI; 46.9% for cetuximab + FOLFIRI; P = .0038).52 This higher response rate also translated into a superior progression-free survival (progression-free survival, HR = 0.637, 95% CI, 0.432-0.941, P = .023).
As K-Ras is the downstream effector of EGFR and it is frequently mutated in colorectal cancer, the effectiveness of cetuximab in K-ras mutated tumors has become a topic of many recent studies. One of the initial reports came from Personeni and colleagues at the 2007 Gastrointestinal Cancers Symposium.53 In their study, the authors demonstrated that an increasing copy number of EGFR in colorectal cancer by FISH analysis predicted the response to cetuximab. However, tumors with the K-ras mutation did not respond to cetuximab therapy. This observation was further confirmed by several studies presented at 2007 American Society of Clinical Oncology (ASCO) annual meeting.54, 55 In a recently published report, De Roock et al found that K-ras mutations precluded tumor response to cetuximab.55, 56 However, all of these studies suffered from small patient cohorts. Recently, 2 large studies were published correlating K-ras mutations with the efficacy of cetuximab and panitumumab. Amado and colleagues examined K-ras mutations in 427 patients with metastatic colorectal cancer treated with panitumumab or best supportive care.57 Patients with wild-type K-ras clearly benefited from panitumumab with a significantly longer progression-free survival and overall survival. However, this benefit diminished in57 patients with mutant K-ras. Hence, the authors recommend that K-ras mutational analysis should be perfomed before initiating panitumumab therapy. Similar results were demonstrated by Karapetis and colleagues.58 A total of 394 tumor samples from patients treated with cetuximab or best supportive care were tested for the K-ras mutation. Their results supported the finding that patients with K-ras mutations did not benefit from cetuximab treatment.
The relation between the K-ras mutational status and the efficacy of anti-EGFR therapy was further assessed by several other groups who reported similar results at the 2008 Gastrointestinal Cancer Symposium and the ASCO annual meeting.57, 59-61 Van Cutsem and colleagues performed a retrospective study to assess K-ras mutational status in patients who participated in the CRYSTAL trial.62 In this study, 348 patients (64.4%) had wild-type K-ras, while 192 patients (35.6%) had mutant K-ras. In patients with wild-type K-ras, cetuximab significantly improved the response rate (43% for FOLFIRI, 59% for cetuximab + FOLFIRI, P = .0025) and progression-free survival (8.7 months for FOLFIRI, 9.9 months for cetuximab + FOLFIRI, P = .017). However, cetuximab had no effect on tumors bearing the K-ras mutation (RR: 40% without cetuximab vs 36% with cetuximab, P = .46; progression-free survival: 8.1 months without cetuximab vs 7.6 months with cetuximab, P = .75). This observation was confirmed by another large retrospective study presented by Bokeneyer et al in the OPUS study.61 In the OPUS study, patients with EFGR-expressing metastatic colorectal cancer were randomized to receive FOLOFX4 or cetuximab plus FOLFOX4. The K-ras mutation was correlated with RR and progression-free survival. Bokemeyer and colleagues found that the addition of cetuximab to FOLFOX4 significantly increases RR in K-ras wild-type tumors (37% without cetuximab vs 61% with cetuximab, P = .011). This enhanced RR was not observed in K-ras mutant tumors (49% without cetuximab vs 33% with cetuximab, P = .106). Similarly, patients with wild-type K-ras benefited from cetuximab with a better progression-free survival (7.7 months with cetuximab vs 7.2 months without cetuximab, P = .016). Surprisingly, cetuximab had a detrimental effect on progression-free survival in patients with K-ras mutation (5.5 months with cetuximab vs 8.6 months without cetuximab, P = .0192). Similar results were reported by Tejpar and colleagues in the EVEREST.63 In that study, the authors tried to address the relationship between development of skin rash, K-ras mutation and the efficacy of cetuximab. Patients were initially treated with standard dose (400 mg/m2 loading dose followed by weekly 250mg/m2) cetuximab with irinotecan. Those with grade 0 of 1 skin rash were further randomized to receive standard dose of cetuximab (250 mg/m2 weekly) or dose escalation of cetuximab (increase by 50 mg/m2 every 2 weeks). The authors were able to demonstrate that patients with wild-type K-ras benefited from cetuximab dose escalation. Overall, progression-free survival was found to be better in K-ras wild-type population (173 days for wild-type vs 83 days for mutant, P < .0001). The development of cetuximab related rash was an independent predictor for cetuximab. The results from these studies will undoubtedly change our current practice pattern. The results of these trials are summarized in Table 1.
|Studies||Study Size||Wild-Type K-ras||Mutant K-ras||P|
|Cetuximab alone or with irinotecan||RR: 22.2%||RR: 0%||.045|
|PFS: ND||PFS: ND|
|OS: ND||OS: ND|
|Cetuximab alone||RR: 26.5%||RR: 6.3%||.02|
|TTP: 6.3 mo||TTP: 3.7 mo||.07|
|OS: 10.8 mo||OS: 8.3 mo||.2|
|Cetuximab alone or with irinotecan||37||RR: 21.6%||RR: 0%||<.01|
|TTP: ND||TTP: ND|
|OS: ND||OS: ND|
|Cetuximab and irinotecan||RR: 21.8%||RR: 6.25%||NS|
|TTP: ND||TTP: ND|
|OS: ND||OS: ND|
|Cetuximab with FOLFOX, FOLFIRI, or irinotecan||RR: 57.5%||RR: 0%||.003|
|TTP: ND||TTP: ND|
|OS: ND||OS: ND|
|Panitumumab alone||RR: 17%||RR; 0%||ND, likely significant|
|PFS: 12.3 wk||PFS: 7.4 wk||.0001|
|OS: 8.1 mo||OS: 4.9 mo||Significant HR, 0.67 (95% CI, 0.55-0.82)|
|Cetuximab with FOLFIRI CRYSTAL Trial||RR: 59%||RR: 36%||ND, likely significant|
|PFS: 9.9 mo||PFS: 7.6 mo||ND, likely significant|
|OS: ND||OS: ND||ND|
|Cetuximab with FOLFOX OPUS Trial||RR: 61%||RR: 33%||ND, likely significant|
|PFS: 7.7 mo||PFS: 5.5 mo||.0009|
|OS: ND||OS: ND|
|Dose escalation of cetuximab with irinotecan EVEREST Trial||RR: 46.4%||RR: 0%||ND|
|PFS: 173d||PFS: 83 d||ND|
|OS: ND||OS: ND|
K-ras mutational analysis should be performed in all metastatic colorectal cancer patients at the beginning of treatment to more accurately guide their therapy. Figure 1 is a proposed algorithm to stratify patients based on their K-ras mutational status. Cetuximab should be offered only to those with wild-type K-ras. Similarly, Amado and colleagues have demonstrated that panitumumab, a fully humanized monoclonal antibody to EGFR, exhibits efficacy only in wild-type K-ras tumors.57 Hence, the use of panitumumab should be restricted to patients with K-ras wild- type tumors.
Numerous additional genetic mutations are presently being investigated to tailor therapy to these specific gene alterations. One such gene that has similar implications is PTEN. Loupakis and colleagues examined the expression of PTEN and K-ras mutation in metastatic colorectal cancers to correlate with the response to cetuximab and irinotecan treatment.64 The authors demonstrated that patients with tumors expressing PTEN and wild-type K-ras had a better prognosis and higher response rate. One possible mechanism is that the loss of PTEN leads to activation of PI3 kinase/AKT pathway which inhibits cell apoptosis triggered by cytotoxic agents or cetuximab. An additional piece of evidence came from a preclinical study in prostate cancer in which C4-2 Cap, a prostate cancer cell line deficient of PTEN, was investigated. This cell line is essentially resistant to EGFR inhibitors such as gefitinib and lapatinib. However, restoration of PTEN in C4-2 Cap cells resensitized cells to EGFR inhibitors.65
Most recently, mutations in other signaling molecules along Ras/Raf/Map kinase pathway have been explored. In 2002, Davies et al identified gain-of-function BRaf mutations in colorectal cancer and melanoma.66 All point mutations occurred within the kinase domain of BRaf protein leading to constitutive activation of BRaf kinase. Ras is not required for the activated BRaf-induced transformation in NIH 3T3 cells. After Davies' discovery, several other groups also evaluated BRaf mutations specifically in colorectal cancer. A French study published by Barault et al identified a 13.3% BRaf mutation rate in a cohort of 586 patients. Unlike the K-ras mutation, mutations in BRaf frequently occurred in females, patients older than 75 years of age, proximal lesions, and tumors with higher frequency of microsatellite instability.67 These important discoveries opened up more questions. Are K-ras and BRaf mutations mutually exclusive in colorectal cancer? What is the efficacy of EGFR-blocking therapy in colorectal cancers with BRaf mutations? And finally, are currently available BRaf inhibitors such as sorafenib effective in colorectal cancers with the BRaf mutation?
These breakthrough findings on the molecular mechanisms of cancers are changing not only our current practice paradigm but also the traditional method by which clinical trials are designed in this new era of molecular medicine. Clinical trials involving certuximab have all undergone modifications in the randomization strategies based on the K-ras mutation status. CALGB 80,405, a phase 3 study assessing cetuximab and bevacizumab in combination with either FOFOX or FOLFIRI as a first-line therapy, will randomize patients based on their K-ras mutational status. The SWOG/NCCTG study evaluating the continuation of bevacizumab versus cetuximab in the second-line setting will also stratify patients using K-ras analysis. Finally, ECOG-N0147/NCCTG-N0147, an adjuvant trial for stage III colon cancer, will adopt the same strategy. We propose an algorithm in the management of metastatic colorectal cancer as illustrated in Figure 2.
K-ras genetic testing is an important development in the treatment of colorectal cancer. Numerous other genes are currently being studied to further define the role of specific targeted therapies. Molecular profiling of individual tumors, an in-depth understanding of cancer genomics, and epigenomic alterations will lead to more effective therapies in the near future. The individualization of therapy for a particular patient will be a revolution in the treatment of cancer patients when its potential is fully explored and recognized in the upcoming era of personalized medicine.
Conflict of Interest Disclosures
Supported in part by grants from the Dallas, Park, Cantu, and Smith families, and the River Creek Foundation.