Introduction
Fluoropyrimidine-based chemotherapy plus antibody therapy is currently the standard first-line treatment for metastatic colorectal cancer (mCRC). Infusional 5-fluorouracil (5-FU) and leucovorin (LV) or capecitabine with either oxaliplatin (FOLFOX/XELOX) or irinotecan (FOLFIRI), both of which were shown to have manageable toxicity profiles and to be able to improve treatment efficacy (1–3), are administered as the chemotherapy backbones in such treatment. Due to their proven anticancer activity, the monoclonal antibodies (mAbs) cetuximab, panitumumab and bevacizumab have also been approved for use as first-line chemotherapy in mCRC patients in combination with FOLFOX/XELOX and/or FOLFIRI.
Progress has been achieved with drugs targeting the vascular endothelial growth factor (3) or the epidermal growth factor receptor (EGFR) (4). The EGFR, a receptor tyrosine kinase, triggers a downstream signaling cascade through mechanisms such as the RAS/RAF/MAPK and PI3K/AKT pathways, which are involved in cell proliferation, survival and motility. Based on our knowledge of this cascade, the administration of cetuximab and panitumumab, two mAbs targeting EGFR, was established as a novel treatment option for mCRC patients.
Among the predictive biomarkers used to identify the mCRC patients most likely to benefit from cetuximab and panitumumab treatment, the best established is the KRAS gene. Mutations in KRAS produce a constitutively active RAS protein, leading in turn to EGFR-independent activation of the RAS/RAF/MAPK pathway (5). The identification of this phenomenon has led to the compelling hypothesis that the activation of KRAS mutations may preclude response to anti-EGFR mAb therapy, a hypothesis supported by earlier clinical observations (6,7). Between 2007 and 2008, 6 randomized clinical trials were conducted, in which the KRAS status was retrospectively assessed in tumor samples from mCRC patients who had been randomly assigned to receive panitumumab (8,9) or cetuximab treatment (10–13).
The activation of mutations of BRAF, another component of the EGFR/MAPK signal transduction pathway, is also prevalent among mCRC patients. As with the KRAS mutation, it is plausible that BRAF mutations may confer resistance to anti-EGFR therapy, although their lower prevalence makes this hypothesis more difficult to test clinically. However, two previous studies reported that BRAF mutation at codon 600 (V600E), resulting in strong activation of the BRAF protein downstream of KRAS, is associated with a shorter progression-free survival (PFS) and overall survival in mCRC chemorefractory patients treated with anti-EGFR mAb therapy (14,15).
Tumor-derived mutant PI3K was shown to stimulate the AKT pathway and promote cell growth in several types of cancer, including CRC. Tumors with PIK3CA mutations have been associated with poor prognosis, with mutations in the PIK3CA gene found to significantly impair the response to anti-EGFR mAb treatment in mCRC patients. In support of these findings, a large-scale European study reported that acquiring knowledge regarding the combined KRAS, BRAF, PIK3CA and NRAS mutation status may improve the sensitivity of prediction of the response to anti-EGFR mAb therapy (16).
In order to counsel mCRC patients harbouring KRAS mutations (and possibly other gene mutations), we need to establish whether they may still benefit from standard chemotherapeutic options. Therefore, this study pursued three objectives: to evaluate the efficacy of the currently available chemotherapeutic protocols for the treatment of mCRC patients, to investigate the value of predictive biomarkers in the personalization of 5-FU-based chemotherapy and to test the hypothesis that mutations in several of the targeted oncogenes are correlated with treatment outcomes in patients receiving different first-line regimens. To test this hypothesis, we first evaluated the predictive significance of KRAS, BRAF, PIK3CA, NRAS and AKT1 mutations in a cohort of mCRC patients who had undergone 5-FU-based chemotherapy; subsequently, using a uniform catalog of retrospective but detailed clinical data, we determined the predictive value of these mutations regarding patient outcomes following completion of the most common therapeutic regimens. This analysis allowed for assessment of the predictive significance of KRAS, BRAF, PIK3CA, NRAS and AKT1 mutations independent of anti-EGFR therapy and their predictive value regarding benefit from oxaliplatin- or irinotecan-based therapy.
Materials and methods
Patients and treatment methods
This study was approved by the Ethics Committee of Tohoku University School of Medicine and included a total of 194 mCRC patients who had received various forms of first-line chemotherapy at the study site between February, 2005 and October, 2010.
The mFOLFOX6 regimen consisted of a 2-h infusion of 85 mg/m2 of oxaliplatin on day 1, a 2-h infusion of 200 mg/m2 of LV on day 1, a bolus of 400 mg/m2 of 5-FU on day 1 and a 46-h infusion of 2,400 mg/m2 of 5-FU/day on days 1–2. The FOLFIRI regimen consisted of a 1.5-h infusion of 150 mg/m2 of irinotecan on day 1, a 2-h infusion of 200 mg/m2 of LV on day 1, a bolus of 400 mg/m2 of 5-FU on day 1 and a 46-h infusion of 2,400 mg/m2 of 5-FU/day on days 1–2. The treatments had been administered on day 1 and repeated on day 2 of a 14-day treatment cycle. The IRIS regimen consisted of continuous administration of 150 mg/m2 of irinotecan for 90 min on day 1, followed by twice-daily administration of S-1 for a 2-week period on days 3–16. The administered dose of S-1 had been determined as follows: for a body surface area (BSA) of <1.25 m2, 80 mg/day; for a BSA of 1.25–1.5 m2, 100 mg/day; and for a BSA of >1.5 m2, 120 mg/day as a 3-week course.
Tumor collection and processing
Formalin-fixed, paraffin-embedded (FFPE) samples of tumor tissue from archival specimens that had been collected at the time of diagnosis and stored at Tohoku University Hospital were investigated. The assays of the tissue samples for KRAS, BRAF, PIK3CA, NRAS and AKT1 mutations were performed at the Department of Clinical Oncology, Institute of Development, Aging and Cancer, Tohoku University. All samples were screened for KRAS mutations in codons 12, 13 and 61; BRAF V600E; PIK3CA mutations in exons 9 and 20; NRAS mutations in codons 12, 13 and 61; and AKT1 E17K. All samples were also classified as mutant or wild-type.
Nucleotide sequence analysis
Mutation analyses of KRAS, BRAF, PIK3CA, NRAS and AKT1 were performed by extraction of genomic DNA from FFPE tissue slides or sections. DNA was extracted using the QIAamp DNA FFPE tissue kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Analyses of the DNA sequences were performed using the automated CEQ2000XL DNA analysis system (Beckman Coulter, Fullerton, CA, USA) under specific cycle and temperature conditions. The PCR products were analyzed by 1.0% agarose gel electrophoresis. Appropriate positive and negative controls were included for the KRAS, BRAF, PIK3CA, NRAS and AKT1 analyses. To minimize bias, all researchers who performed the mutation analyses were blinded to the clinical outcomes.
Statistical analysis
All patients for whom data regarding KRAS, BRAF, PIK3CA, NRAS and AKT1 mutation status were available, were included in the analysis. The response rate (RR) was determined according to the Response Evaluation Criteria in Solid Tumors (RECIST), version 1.0. According to RECIST, the patients were categorized as responders if they achieved complete response (CR) or partial response (PR) and as non-responders if they exhibited stable disease (SD) or progressive disease (PD). The associations between treatment response or patient characteristics and mutational status were assessed using the χ2 test. PFS was defined as the time interval between the initiation of chemotherapy and the first objective evidence of disease progression or death from any cause. The PFS was determined using the Kaplan-Meier method and compared using the log-rank test. Statistical significance was set at a level of P<0.05 for a bilateral test. Through such means, the KRAS mutational status was investigated and the hypothesis that PFS varies according to the type of first-line regimen (oxaliplatin- or irinotecan-based) was tested.
Results
Study objective
This retrospective study investigated the efficacy of first-line chemotherapeutic protocols in 194 mCRC patients according to gene status and its association with several patient clinical characteristics. As tumor samples and complete end-point data were available for all patients, RR and PFS were determined for the entire patient sample (100%).
Treatment regimens
Combination chemotherapy was administered as first-line treatment to 174 patients (89.7%), either with or without mAb supplementation (Table I). 5-FU was administered as the only cytotoxic agent to 17 patients (8.8%). A total of 109 patients (56.2%) were treated with oxaliplatin in the first-line setting. The oxaliplatin-containing regimen consisted of only the FOLFOX regimen (infusion and bolus 5-FU plus oxaliplatin). A total of 65 patients (33.5%) were treated with irinotecan in the first-line setting. The irinotecan-containing regimen consisted of the FOLFIRI regimen (infusion and bolus 5-FU with irinotecan) for 43 patients and S-1 plus irinotecan for 22 patients. As first-line treatment, bevacizumab was administered as part of an oxaliplatin-containing regimen to 27 patients (13.9%) and with irinotecan to 26 patients (13.4%). As bevacizumab was not approved until 2007 in Japan, the percentage of mCRC patients who received bevacizumab in this study was low (27.3%).
Mutation analyses of KRAS, BRAF, PIK3CA, NRAS and AKT1
Table II lists the mutations detected by direct sequencing. A relatively rare mutation in codon 61 was analyzed in addition to the common mutations in codons 12 and 13 in order to increase the sensitivity of mutation detection. KRAS mutations at codons 12, 13 and 61 were observed in 78 (40.2%) of the tumor samples. Of the 78 detected mutations in codons 12 and 13, the most frequent was G12D (12.9%), followed by G13D (11.3%), G12V (10.3%), G12C (1.5%), G12A (1.0%), G12R (0.5%) and G13C (0.5%). In codon 61, Q61H and Q61R were detected in 4 samples (2.0%). Three common KRAS mutations, G12D, G13D and G12V, were also frequently detected. V600E was detected in 10 samples (5.2%), all of which harboured wild-type KRAS. PIK3CA mutations in exon 9 (E542K, E545K, E545G and Q546K) were detected in 17 samples (8.8%) and PIK3CA mutations in exon 20 (H1047R, H1047L and H1047Y) in 6 (3.1%). Mutations in KRAS and PIK3CA were detected in 9 samples (4.7%). NRAS mutations at codons 12, 13 and 61 were detected in 3 samples (1.5%); and an AKT1 mutation at codon 17 (E17K) was detected in two samples (1.0%).
Patient characteristics
The characteristics of the 194 mCRC patients (median age, 63 years; range, 16–82 years) from whom primary tumor tissue samples had been collected were retrospectively analyzed (Table III). The most frequent type of tumor was tumor of the rectum (75 patients; 38.7%), followed by tumor of the ascending and sigmoid colon (42 patients; 21.6%), transverse colon (15 patients; 7.7%) and descending colon and cecum (10 patients; 5.2%). The most frequent site of metastasis was the liver (114 patients; 58.8%), followed by the lungs (100 patients; 51.5%), intra-abdominal lymph nodes (72 patients; 37.1%) and the peritoneum (36 patients; 18.6%).
Effect of mutation status on the outcome of first-line chemotherapy
Tables IV and V show the results of the analysis of the association between clinical response in terms of RR and median PFS (mPFS) and the presence or absence of gene mutations. There were no significant differences in RR or mPFS between the wild-type and mutant KRAS subgroups who had received oxaliplatin- or irinotecan-based treatment as first-line therapy. Furthermore, there was no significant difference in RR or mPFS between the wild-type KRAS, BRAF, PIK3CA, NRAS and AKT1 subgroups and the respective mutant subgroups in any of the 5 genes.
The mPFS of the wild-type and mutant KRAS subgroups who had received oxaliplatin-based treatment was 8.6 (n=66) and 6.8 months (n=43), respectively (P=0.41; Fig. 1A). Of the 109 assessed patients, 16 of the 66 (24.2%) patients in the wild-type subgroup and 11 of the 43 (25.6%) patients in the mutant KRAS subgroup had received bevacizumab in combination with FOLFOX (P=0.87). The mPFS of the wild-type KRAS, BRAF, PIK3CA, NRAS and AKT1 subgroups and that of their respective mutant subgroups was 9.7 (n=51) and 7.2 months (n=58), respectively (P=0.10; Fig. 1B).
The mPFS of the wild-type and mutant KRAS subgroups who had received irinotecan-based treatments was 7.7 (n=39) and 9.7 months (n=26), respectively (P=0.43; Fig. 2A). Of the 65 assessed patients, 17 of the 39 (43.6%) patients in the wild-type subgroup and 9 of the 26 (34.6%) patients in the mutant KRAS subgroup had received bevacizumab in combination with FOLFIRI or IRIS treatment (P= 0.47). The mPFS of the wild-type KRAS, BRAF, PIK3CA, NRAS and AKT1 subgroups and that of their respective mutant subgroups were 7.1 (n=32) and 10.0 months (n=33), respectively (P=0.76; Fig. 2B).
Discussion
This analysis of various mutations of the KRAS, BRAF, PIK3CA, NRAS and AKT1 genes in 194 Japanese mCRC patients resulted in the detection of KRAS mutations at a frequency (78/194; 40.2%) similar to that described in a previous study of Japanese patients (17). As the pattern of KRAS mutations was also found to be similar to that reported in a previous study on Caucasian patients (18,19), the results indicated that KRAS mutations do not differ significantly between Japanese and Caucasian populations in terms of frequency and mutation spectrum. The prevalence of PIK3CA mutation (23/194; 11.9%) was also found to be similar to that reported by previous studies (10–20%) (16). By contrast, the prevalence of BRAF mutations (10/194; 5.2%) was found to be lower compared to that reported in studies on Caucasian patients (20), possibly reflecting the genetic differences between the populations. Of the mutations detected, E542K, E545K and H1047R were identified as hotspot mutations, whereas the E545G mutation was rarely detected (16,21). A large-scale analysis is required to elucidate whether this discrepancy in the mutation spectrum is the result of genetic differences among different populations. Previous studies identified mutations in the NRAS and AKT1 genes in 2.6% (16) and 5.9% (22) of mCRC patients, respectively. In this study, NRAS and AKT1 mutations were detected in 1.5% (3 patients) and 1.0% (2 patients) of the sample, respectively.
Similar to previous investigations, the present study analyzed the association between gene mutations and patient characteristics in order to determine whether such associations may predict the efficacy of a first-line regimen. Sartore-Bianchi et al (23) reported that KRAS mutations were significantly more prevalent among females compared with males, whereas PIK3CA mutations were not found to be significantly associated with gender. In accordance with Watanabe et al (17), who reported a higher prevalence of KRAS mutations among Japanese female (40.9%) compared to male mCRC patients (35.5%; P=0.001), a higher prevalence of KRAS mutations was detected among the samples obtained from female (48.2%) compared with those obtained from male patients (34.2%, P=0.050) in the present study.
The individualization of drug therapy for mCRC patients is becoming increasingly feasible. Studies on patients receiving first-line and subsequent lines of treatment demonstrated that those with KRAS mutations do not respond to or experience any survival benefit from treatment with anti-EGFR mAb therapy. Based on this finding, all mCRC patients are currently offered KRAS testing to determine whether their tumor is wild-type KRAS and, if so, counseled that they would likely benefit from anti-EGFR mAb therapy. As such, it is crucial to establish whether the mutation may affect the ability to benefit from anti-EGFR mAb therapy (or any other form of therapy), or whether prognosis is independent of treatment. Retrospective analyses of KRAS mutations in mCRC patients treated with bevacizumab plus chemotherapy revealed that the clinical benefit of bevacizumab is independent of the KRAS status (24,25). Other studies investigated the association between KRAS status (as well as other gene statuses) and clinical benefit from oxaliplatin- or irinotecan-based treatment in the first-line setting (26,27). Those studies reported that the clinical benefit of oxaliplatin- or irinotecan-based treatment is independent of the KRAS mutational status.
In this study, the patients who had received oxaliplatin treatment exhibited longer mPFS in the wild-type KRAS alleles. By contrast, the patients who had received irinotecan treatment exhibited longer mPFS in the mutant KRAS alleles. Although there were no statistically significant differences in the distinct KRAS status between the oxaliplatin- and irinotecan-based treatment groups, the KRAS status is likely to affect the outcome of these treatments in some of the patients. The results of the present study indicated that mCRC patients with activation of KRAS mutations, even those treated with oxaliplatin- and irinotecan-based regimens as first-line treatments, may benefit from cytotoxic drug therapy. We also provided evidence that both the wild-type and mutant KRAS subgroups of mCRC patients may benefit from oxaliplatin- and irinotecan-based therapy.