Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed cancer in men and the second in women worldwide. In Brazil, it is the third most frequent type of cancer with an estimated 30,000 new cases of CRCs in 2012 (1). In addition, CRC is the second leading cause of mortality worldwide and the fifth in Brazil (2,3). Therefore, it is highly important to improve strategies for CRC prevention and early detection aiming to decrease its incidence and mortality (4).
The majority of CRC cases develop through a stepwise evolution of normal mucosa to precursor lesions and ultimately to a malignant tumor. Adenoma is the principal precursor lesion of CRC (5,6) but, recently, serrated polyp was recognized as an alternative precursor lesion of CRC and follows an alternative pathway in which serrated polyp replaces the traditional adenoma as the precursor lesion to serrated CRC, accounting for ~10% of all CRCs (7,8). Serrated polyps form a heterogeneous group of colorectal lesions that include hyperplastic polyps (HPs), sessile serrated adenoma (SSA), traditional serrated adenoma (TSA) and a combination of two or more characteristics, formerly classified as mixed polyps (MP) (9). HPs are the most common serrated polyp and they have been increasingly suggested to be precursor lesions, since they may develop into other serrated polyps as SSA, TSA or MP to CRC (10).
Colonoscopy is considered the main method for detection and removal of precursor lesions during screening and surveillance of CRC (11). However, it can still miss up to 26% of adenomas and 2% of advanced adenomas (11). Therefore, novel and complementary approaches to detect these potential malignant lesions are required and the application of molecular biomarkers has been considered in the context of CRC screening (12). One of the most challenging issues in biomarker screening is the knowledge of the different molecular pathways implicated in colorectal carcinogenesis and hence the identification of relevant and reliable biomarkers for colorectal screening and surveillance.
The molecular mechanism underlying the adenoma-to-carcinoma sequence has been extensively studied and involves a cumulative acquisition of mutations in tumor suppressor genes, such as APC, and oncogenes such as KRAS leading to a phenotype of genomic instability (5). On the other hand, the mechanisms related to serrated carcinoma development are less understood. Mutations of BRAF and, less frequently, KRAS, are likely to be the initiating events and serrated carcinomas are characterized by microsatellite instability (MSI) and/or CpG island methylator phenotype (CIMP) (7). MSI is a hallmark of CRC arising in the context of hereditary nonpolyposis colorectal cancer (HNPCC) or Lynch syndrome (13). However, ~15–20% of sporadic CRCs are MSI (13,14). The MSI is caused by the loss of mismatch-repair genes, which leads to an increased susceptibility to accumulate mutations in genes with microsatellite regions (13,14). Both KRAS and BRAF encode kinases that belong to the mitogen-activated protein kinase (MAPK) cascade that mediates the cellular signaling involving cell proliferation, apoptosis and differentiation (15). In adenomas, mutations in KRAS occur during the early to advanced adenomas in the adenoma-to-carcinoma sequence. However, in the other precursor lesions, there is considerable variability in the frequency of KRAS and BRAF reported (7).
Considering the wide divergence in the frequency of KRAS and BRAF mutations in the precursor lesions of CRC and the absence of data in the Brazilian population, the aim of this study was to research the frequency of KRAS and BRAF mutations and MSI phenotype in precursor lesions of a Brazilian population referred for colonoscopy and to associate molecular alterations with histological and morphological characteristics. Moreover, we compared these findings with molecular alterations found in a series of Brazilian CRC.
Materials and methods
Patients
A total of 155 patients (>50 years old) referred to the Department of Endoscopy of Barretos Cancer Hospital for colonoscopy, from January to October 2011, were prospectively included in this study. A total of 342 lesions were endoscopically removed from 82 (52.9%) men and 73 (47.1%) women with a mean age of 66 years (range 50–89). The main indication for colonoscopy was surveillance after colectomy for CRC (36.4%), followed by surveillance after polypectomy (14.6%), CRC (12.6%) and abdominal pain (9.8%). Ninety-two (59.4%) patients had more than one lesion of the same or different histological type (mean 2.2; range 1–9). Patients with a known family history, hereditary CRC or bowel inflammatory disease were excluded.
For the comparative analysis of molecular alterations, 47 patients with sporadic colorectal adenocarcinoma were retrospectively retrieved from the Department of Pathology of the same hospital and randomly included in the study. The study was approved by the Ethics Committee of Barretos Cancer Hospital.
Endoscopic analysis and tissue specimens
All colonoscopies were performed with high-resolution magnification endoscopes (Fujinon 4400 and Olympus CV GIF 180; Tokyo, Japan) and with targeted dye spraying of the colon using 0.4% indigo carmine solution. The cecum was reached in all cases and all lesions detected were removed. The lesions were characterized according to Paris classification [type 0–I, polypoid (0–Is, sessile; 0–Isp, semi-pedunculated; 0–Ip, pedunculated); type 0–II, non-polypoid (0–IIa, slightly elevated; 0–IIb, flat; 0–IIc, slightly depressed; type 0–III, excavated); LST, laterally spreading type] (16). The site and size of each lesion was annotated and for the purpose of analysis, lesions located in the cecum, ascending colon and transverse colon were regarded as right colon and those from descending colon, sigmoid colon and rectum were regarded as left colon. All lesions removed during colonoscopy were submitted to histological analysis and re-evaluated in a blind manner from the initial pathology classification. The lesions were classified based on WHO criteria (17). The combination of more than one histological type in the same lesion was regarded as MPs. Advanced adenomas were classified if at least 10 mm size or with villous architecture or high-grade dysplasia. For molecular analysis, 103 lesions (one from each patient) were randomly selected, to have a balanced distribution of the different histological subtypes.
DNA isolation
Serial 5-μm unstained sections of formalin-fixed paraffin-embedded blocks were cut, and one adjacent hematoxylin and eosin-stained (H&E) section was taken for pathologist identification and selection of the precursor lesion and tumor tissue. DNA was isolated from 1 unstained section from each specimen as previously described (18). Briefly, tissues were deparaffinized at 80°C and serial washed with xylene and ethanol (100, 70 and 50%). Selected areas of tumor or precursor lesions were macrodissected using a sterile needle (18G × 1 ½) (Becton Dickinson Ind Cirúrgicas Curitiba-PR, Brazil) and carefully collected into a microtube. DNA was extracted using QIAamp DNA Micro Kit (Qiagen, Hilden, Germany), following the manufacturer’s instructions. DNA quantity and quality was evaluated by Nanodrop 2000 (Thermo Scientific, Wilmington, DE, USA). DNA samples were diluted to a final concentration of 50 ng/μl and stored at −20°C for further molecular analysis.
Mutational analysis of KRAS and BRAF
The hotspots regions of the oncogenes KRAS (codons 12 and 13) and BRAF (codon 600) were analyzed by polymerase chain reaction (PCR), followed by direct sequencing, as previously described by our group (18,19).
For KRAS, PCR reaction was performed in a final volume of 15 μl, under the following conditions: 1.5 μl buffer (Qiagen), 2 mM MgCl2 (Qiagen), 100 mM dNTPs (Invitrogen, Carlsbad, CA, USA), 0.2 mM of both sense and anti-sense primers (Sigma Aldrich, St. Louis, MO, USA), 1 unit of HotStarTaq DNA polymerase (Qiagen) and 1 μl of DNA. The KRAS primers used were: GTGTGACATGTTCTAATATAGTCA (sense) and GAATGGTCCTGCACCAGTAA (antisense) (19). For BRAF the PCR reaction was realized in a final volume of 15 μl, under the following conditions: 1.5 μl buffer (Qiagen), 2 mM MgCl2 (Qiagen), 100 mM dNTPs (Invitrogen), 0.3 mM of both sense and antisense primers (Sigma Aldrich, St. Louis, MO, USA), 1 unit of HotStarTaq DNA polymerase (Qiagen) and 1 μl of DNA. The BRAF primers used were: TCATAATGCTTGCTCTGATAGGA (sense) and GGCCAAAAATTTAATCAGTGGA (antisense) (18,19). The PCR was performed in Veriti Termociclador (Applied Biosystems, Austin, TX, USA) using Taq polymerase (Qiagen). The PCR products were evaluated by electrophoresis in agarose gel.
The PCR products of each analyzed exon were firstly purified with EXO-SAP (GE Technology, Cleveland, OH, USA), then, PCR products were submitted to a sequencing reaction using 1 μl of BigDye (Applied Biosystems), 1.5 μl of sequencing buffer (Applied Biosystems) and 1 μl of primer. The sequencing reaction was followed by post-sequencing purification with EDTA, alcohol and sodium citrate. The products of PCR were eluted in HiDye (formamide) and incubated at 95°C for 5 min and at −4°C for at least 5 min. Direct sequencing was realized in 3500 series Genetic Analyzer (Applied Biosystems).
All lesions with mutations were confirmed twice with a new PCR and direct sequencing. Additionally, for quality control, in 10% of cases, a new DNA isolation and further mutation analyses were performed.
Analysis of MSI
The MSI evaluation was performed using a multiplex PCR comprising five quasimonomorphic mononucleotide repeat markers (NR27, NR21, NR24, BAT 25 and BAT26), as described by our group (20). The MSI status of the lesion was analyzed using GeneMapper 4.1 software (Applied Biosystems). Cases exhibiting instability at two or more markers were considered to have high MSI (MSI-H), those with instability at one marker were defined as having low MSI (MSI-L) and finally those that showed no instability were defined as microsatellite stable (MSS). In cases with MSI-H, DNA was isolated from adjacent normal tissue and instability of markers was assessed. DNA from cell lines HCT15 (MSI-H) and DNA of healthy people (MSS) were used as controls. Analyses of samples with an abnormal profile were repeated twice.
Statistical analyses
Statistical analyses were performed in SPSS Software® for Windows, version 19.0. The casuistic was characterized by means of descriptive statistics. Categorical variables were compared using the chi-square or Fisher’s exact tests, depending on the expected values in the contingency tables. The significance level was set at 5%.
Results
Endoscopic and histopathological features of the colorectal precursor lesions removed by colonoscopy
The endoscopic and histopathological characteristics of the colorectal precursor lesions removed by colonoscopy are summarized in Table I and are illustrated in Fig. 1. For association analysis, serrated polyps were stratified in two groups: SAs (SSA and TSA) (Fig. 1E and F) and HPs (Fig. 1C and D) based on malignant potential differences between them. Due to the presence of more than one histopathological type, MP cases were not considered in the association analysis. HPs were located predominantly in the left colon when compared with adenomas (Fig. 1A and B) and SAs (P<0.001, Table II). Non-polypoid type was more likely to be more frequent among SAs compared to adenomas or HPs (P=0.06). A significant association between lesion size and histological type was observed (Table II). Lesions >10 mm were more common among SAs than HPs and adenomas (P=0.009, Table II).
Molecular alterations in colorectal precursor lesions removed by colonoscopy
After the morphological characterization of all lesions, we selected 103 lesions (one from each patient) consisting of 50 adenomas and 53 serrated polyps (13 SSAs, TSAs and 38 HPs) for KRAS and BRAF mutation analysis and MSI status analysis. The frequency and mutation description are summarized in Table III and Table IV.
Mutations in KRAS and BRAF were respectively detected in 14 (13.6%) and 9 (8.7%) out of 103 lesions and they were mutually exclusive events (Fig. 2). None of the precursor lesions exhibited MSI-H phenotype.
KRAS mutations were observed in 7 (14.0%) out of 50 adenomas and in 7 (13.2%) out of 53 serrated polyps. None of the SAs were KRAS mutated (P=0.223; Tables IV and V). The majority of KRAS mutations were found in codon 12 (86.7%), and the most frequent mutation type was Gly12Asp, observed in 8 cases (61.5%) (Table IV). A tubular adenoma had two KRAS mutations (Gly12Ala and Gly13Asp).
BRAF mutations were found in 9 (17.0%) out of 53 serrated polyps and in no adenoma. All BRAF mutations were V600E (Val600Glu) (Table IV). BRAF mutations were significantly associated with SAs when compared with adenomas and HPs (P<0.001; Table V).
We further analyzed the association between KRAS and BRAF status and endoscopic characteristics (Table VI). Twelve (85.7%) lesions with KRAS mutations were located in the left colon and only 2 (14.3%) in the right colon, while 52 (58.4%) wild-type KRAS were located in the left colon and 37 (41.6%) in the right colon (P=0.05). Regarding morphology and size of the lesions, no association was found with KRAS status. On the other hand, KRAS mutations were significantly more common in advanced adenomas (33.3%) than in non-advanced adenomas (5.7%) (P=0.020).
No association was found between BRAF status and localization, morphology or size of the lesions.
Comparison of precursor CRC lesions and colorectal adenocarcinomas
We further compared the frequency of molecular alterations found in precursor lesions of CRC with molecular findings in 47 CRCs from the same institution. All CRCs were adenocarcinomas, and their clinical and clinical-pathological characteristics are detailed in Table VII. The description of clinical and demographic characteristics of adenocarcinomas harboring KRAS or BRAF mutations is shown in Table VIII. KRAS mutations were detected in 22 (46.8%) cases and BRAF mutations were found in 3 (6.5%) colorectal adenocarcinomas. KRAS mutations were significantly more frequent in colorectal adenocarcinomas than in precursor lesions (P<0.001). As found in precursor lesions, the most frequent mutations in KRAS were at codon 12 (81.8%), and the Gly12Asp was the most frequent mutation (44.5%). All BRAF mutations were V600E (Val600Glu), as well as in precursor lesions. MSI-H was detected in 10.6% of all cancers and in no precursor lesions.
Discussion
Colorectal cancer (CRC) is a major health problem in Brazil with an increase in its incidence in the last decade. Approximately 25,000 new cases of CRCs were expected in 2006 and 30,000 in 2012 (1). The only way to change this is through prevention strategies with early detection and resection of their precursor lesions. The morphological and molecular characterization of these lesions has helped us in the understanding of the sequence of events by which normal cells develop into cancer. In line with that, this study sought to contribute to the morphologic and molecular characterization of the different types of colorectal precursor lesions removed during colonoscopy in a Brazilian population with an increased risk for CRC. To the best of our knowledge, this is the first study to describe molecular alterations in colorectal precursor lesions in a Brazilian population.
In our series, adenomas were the most frequent (70.2%) colorectal precursor lesion removed during colonoscopy, as described by other authors in different populations (51–67%) (21–24). Tubular adenomas were more prevalent than tubulovillous and villous adenomas. HPs accounted for ~24% of all serrated lesions followed by SSAs (3.8%), TSAs (<0.6%) and MPs (21,24–26). Most HPs were left-sided and <10 mm. There was a significantly higher number of SAs and conventional adenomas in the right colon than in HPs. In addition, SAs had a tendency to be non-polypoid lesions compared with adenomas and HPs. These findings are in agreement with previous published studies, which showed that HPs are the most common serrated polyp of the colon accounting for 10–15% of all polyps of the colon and SSAs account for approximately 3–9% of all the colorectal polyps (21–23). Previous studies have also shown that most of the HPs are small (<5 mm) and located in the distal colon (75–80% in the rectosigmoid) and SSAs are generally located in the right colon (27,28). HPs located in the distal colon have been considered indolent lesions, without the need of removal or further endoscopic vigilance. On the other hand, HPs >0.5 cm and located in the right side colon have been associated with increased cancer risk and their removal has been recommended (29). In contrast, SAs should be submitted to the same vigilance as patients with conventional adenomas (30).
In the classic adenoma-carcinoma sequence model of colorectal tumorigenesis proposed by Fearon and Vogelstein, HPs were described as harmless non neoplastic lesions with no malignant potential (5). This concept was challenged since the description of cancer occurrence in patients with hyperplastic polyposis syndrome (31) and in sporadically occurring serrated polyps (32). Approximately 10% of sporadic CRCs, known as serrated adenocarcinoma, will arise via serrated polyp-carcinoma sequence (13). In this context, HPs were recently recognized as neoplastic lesions included in the serrated group and may predispose to cancer. Therefore, efforts have been made to better differentiate serrated lesions without malignant potential from those with high risk. Herein, we performed an analysis of KRAS and BRAF mutations, and MSI status, to better understand the malignant potential of such lesions.
In the literature, there is considerable variability in KRAS and BRAF mutation frequencies among colorectal precursor lesions, mainly among serrated polyps. In our series, KRAS mutations were detected only in adenomas and HPs. Notably, all mutation types found in precursor lesions were those usually detected in CRC. KRAS mutations were significantly associated with advanced adenomas, which have a greater risk of developing into malignant tumors than non-advanced adenomas. Yadamsuren et al demonstrated that 57.5% of advanced adenomas harbored KRAS mutations compared with 31.0% of non-advanced adenomas in a series of 164 sporadic adenomas (33). These findings are in agreement with the Fearon and Vogelstein model where KRAS mutation is responsible for the intermediate stage of adenoma progression (33). At variance, we did not observe KRAS mutations in SAs, contrasting with some studies that report the presence of KRAS mutations, yet at lower frequencies (8–16,5%) (24,34). This discrepancy in KRAS mutations status can be justified by the small number of cases studied, methodology issues or it can be related to differences in patient ethnic population pertaining to the distinct genetic background of patients. Nonetheless, the absence of KRAS mutations in SAs may be an indication that, in this Brazilian population, KRAS is not responsible for the serrated pathway.
Our results indicated that BRAF is a prevalent marker in the serrated pathway. We observed BRAF mutation in ~40% of SAs and this is in line with previous studies that demonstrated a frequency of BRAF mutations (V600E) ranging from 32 to 82.9% (33). Collectively, our findings, as well as those of others, are in agreement with a recent study using BRAF V600E knock-in murine models that demonstrate the pivotal role of BRAF mutations in the initiation of the serrated pathway (35).
In the present study, MSI-H was not detected among colorectal precursor lesions, in accordance with international literature, suggesting that MSI is a late event in the serrated adenocarcinoma progression (13).
In conclusion, the present clinical and molecular characterization of colorectal lesions may contribute to the identification of molecular diagnostic biomarkers as a tool for strategies of screening and early detection of CRC in the Brazilian population. Nevertheless, further studies are required to validate the present findings in a large number of patients, and to extend it not only to high risk CCR populations, as in our study, but also to average risk populations.