KRAS mutations: variable incidences in a Brazilian cohort of 8,234 metastatic colorectal cancer patients

Personalised medicine is an evolving field that seeks to target cancer therapies based on unique genetic characteristics of the tumour and/or the patient [1]. One of the most significant advances towards personalised care in the field of oncology was the establishment of KRAS gene mutation as a validated biomarker predicting efficacy in epidermal growth factor receptor (EGFR) targeted therapies – such as cetuximab and panitumumab - in the treatment of metastatic colorectal cancer (mCRC) [2, 3]. According to Siegel et al. [4], CRC is the third cause of new cancer cases and of death by cancer in the United States with an estimate of 73,420 new cases for males and 70,040 new cases for females in 2012. In Brazil, data published by the Brazilian National Cancer Institute predicted the number of new cases for 2012 of 14,180 for males and 15,960 for females with a frequency variation depending on the country’s region (data available online at http://​portal.​saude.​sp.​gov.​br/​resources/​ses/​perfil/​gestor/​homepage/​estimativas-de-incidencia-de-cancer-2012/​estimativas_​incidencia_​cancer_​2012.​pdf). The Southeast region shows the highest incidence of CRC, being the second most frequent cause of cancer for both men (22/100,000) and women (23/100,000). The differences observed in the incidence of CRC according to different Brazilian regions could relate to the idea that differences in patients origins might contribute to the incidence of somatic mutations in candidate cancer genes such as KRAS[5].

In order to understand KRAS function it is important to address how Ras proteins are activated. The Ras protein family belongs to a group of small GTPases, which are able to cycle between an inactive (GDP-bound form) and an active state (GTP-bound form) leading to the activation of several effector kinases. These proteins are involved in cell proliferation, differentiation and survival, hence its importance in cancer research [6]. Mutations in RAS proto-oncogenes (comprising H-, N- and K-RAS) are among the most common in malignant tumours and although RAS isoforms are very similar, KRAS is more frequently found mutated in cancers occurring in 22% of all tumours analysed compared to 8% for NRAS and 3% for HRAS[7]. In mCRC, mutation in KRAS gene result in continuous activation of intracellular EGFR pathway regardless the pharmacological blocking of the receptor [8]. Thus, proliferation, invasion, survival and metastasis of the tumour are maintained. Clinically, patients with KRAS wild-type tumours are more likely to respond to anti-EGFR therapy whereas those with mutant KRAS show lack of benefit [9‐14]. Therefore, anti-EGFR monoclonal antibodies are only indicated in patients with KRAS wild-type tumours [15].

Since June 2008, a Merck Serono Oncology sponsored program began to reimburse for KRAS mutation analyses in mCRC patients in Brazil. Almost all KRAS mutations tests were performed in a single molecular biology facility. A total of 8,234 patients had their primary tumour and/or metastasis analysed through the program. Here, some epidemiological characteristics of the KRAS mutations are described.

In the colorectal carcinogenesis model initially described by Vogelstein [20], specific genetic events would be related to morphological tissue changes. Among these genetic changes, different studies have shown that mutations in the KRAS gene were found in over 30% of CRC and advanced adenomas [21, 22]. The present study is consistent with this showing that among the 8,234 cases analysed, a total of 2,623 (31.9%) corresponded to mutant KRAS (Figure 1). From the 2,623 mutated cases, 83% were in codon 12 versus 17% in codon 13, being the aminoacid change Gly to Asp the most common for both codons and genders (Table 3). The prevalence of Gly12Asp mutation over other mutations corroborates with data from other populations around the world [23, 24]. Despite variations in the frequency of types of mutations per region (Table 2), our results showed that more than 30% of the patients in each region harboured KRAS mutations (Table 1). Regarding prognosis, previously published data showed that Gly12Val substitutions are more aggressive having a poorer prognosis than those with Gly12Asp mutation, thus revealing a connection between survival and KRAS mutation type [23, 25‐27]. In relation to codon 13, some studies have shown that mutations in this codon could be less aggressive than in codon 12 and that patients with KRAS Gly13Asp mutant tumours could benefit from anti-EGFR therapies [28, 29]. In addition, recent in vitro studies have confirmed that Gly13Asp mutations are associated with sensitivity to anti-EGFR antibody treatments [30, 31]. Unfortunately, we were unable to provide prognosis comparisons since we could not retrieve the follow up from many patients in our cohort.

Gender was another variable we analysed in our cohort. The percentages of KRAS mutations pointed out a female over male prevalence (Table 4, p < 0.05). Actually, the female predominance of KRAS mutations in CRC patients has been reported in few recent studies performed in smaller cohorts from Asia. In China, a cohort of 167 CRC patients were tested for mutations on KRAS codons 12 and 13 and their results detected a higher rate of KRAS mutations in female compared to male patients [32] and this finding did not significantly correlate with the patient age, tumour site, differentiation grades and histological types. Differently to our data, their results referred to Gly13Asp as the most frequently mutation identified [32]. In another Chinese study, Shen et al. [33] observed that in a cohort of 118 CRC patients, there was also a higher incidence of KRAS mutations in female patients compared to male patients (44.7% vs 28.2%, p = 0.037). In Japan, researchers suggested that gender and age were independent risk factors for KRAS mutations [34]. Another publication did not show correlation between gender and KRAS mutational status [35], and one possible explanation for this could be the ethnological differences in populations studied. Nevertheless, it is important to address that the greatest prevalence of KRAS mutation in women has been previously observed in other cancer types such as in patients with lung adenocarcinoma, leading to the hypothesis of possible hormonal influence [36].

Clinical evidence suggests potential sex-related differences in the development and prognosis of CRC, which could be associated with sex hormones. Estrogens and androgens regulate growth, differentiation and functioning of different tissues, including the gastrointestinal tract. Estrogen is an important mitogen capable of sending its signal to the nucleus via interaction with estrogen receptors (ER) on target cells. The two distinct estrogen receptors (ERα and ERβ) tend to respond differently to estrogen. The high proliferative activity triggered by estrogen is related to its connection to the ERα, which can help tumoral development by increasing the probability of genetic mutations [37]. In contrast, the ERβ forms heterodimers with the ERα blocking their proliferative activity by suppression of oncogenic transcription factors (e.g. c-myc, cyclin D1 e cyclin A) and by stimulating the expression of tumour suppressing genes (e.g. p21 e p27) [38]. The expression of the ERβ is significantly reduced in adenomatous tissues and in colon tumours, when compared to normal mucosa in both genders, however with a slight reduction in females (p < 0.002) [39‐41]. Nevertheless, there are no observed changes in the expression of ERα between the different tissues. Also in pre-cancerous lesions with a high risk of CRC development, a decrease in ERβ expression could indicate a promoting factor for the development of cancer [42]. In hereditary nonpolyposis colon cancer syndrome (HNPCC) characterised by a dominant susceptibility acquired in the early appearance of symptoms of CRC, the average age for CRC diagnosis is earlier in males than females (38.8 vs 47.2; p < 0.05) [19], indicating that female sexual hormones could be acting as protective factors. This estrogen protectiveness was also the conclusion after guinea pigs that were treated with PhiP (2-Amino-1-methyl-6-phenylimidazo [4,5-b] pyridine), an inducing agent for colon cancer, resulted in the average number of aberrant crypt foci higher in males than in females (p < 0.001) [43]. In regard to hormone replacement therapy, meta-analysis studies have confirmed the reverse association with the risk of developing CRC [44‐47].

The role that male sexual hormones play in the risk for CRC is still unclear. There is evidence of reverse association between the serum level of the dehydroepiandrosterone sulfate and the risk of CRC [48]. The increase of CAG trinucleotide repetitions in the coding sequence of the androgen receptor seems to be related to the lowest trancriptional activation of this receptor resulting in a lower androgenic action on the tissues, increasing the risk of developing colon cancer [49]. Gillessen et al. [50] confirmed in a large retrospective evaluation of 107,859 patients with prostate cancer that patients treated with Gonadotrophin-releasing hormone (GnRH) antagonist or orchiectomy presented an increase of 30-40% risk of developing CRC when compared to the subgroup that were not submitted to androgenic deprival.

Kato et al. [51, 52] showed that the estrogen receptor is involved in the KRAS mediated transcription and its implication in the senescence escape. Furthermore, an association between KRAS and estrogen receptor was also observed when mutant KRAS (Gly12Val) was overexpressed in NIH-3 T3 cells, which resulted in increased levels of the endogenous estrogen receptor. In addition, the RAS signal intensified the estrogen receptor activity as a transcription factor leading to cell transformation [53]. In a pre-clinical study it was shown that RAS oncogenes might remain latent in the mammary gland of guinea pigs until estrogen exposure. This suggests that normal proliferative processes, such as the ones induced by estrogen in the development of the mammary gland, could be necessary to lead cells with the latent RAS oncogenes to neoplastic development [54]. Collectively, these findings could indicate an association between mutant KRAS and sexual hormones.

Our observations of differences in KRAS mutational status according to sex led us to analyse also age. KRAS was more frequently found mutated in females than males for the ages ≤40, 40–50, and 50–60 (Figure 2). The age range 40–60 coincides with the menopausal period, which is associated with a drop in estrogen levels. In contrast, ages 60–70 and ≥70 showed a different result, where males had more mutations than females (p > 0.05). Although the p value was higher for the ages 60–70 and ≥70, this result could indicate an age-linked difference associated with several causes including men andropause. Actually, a recent study suggested that men with lower androgenicity (resulting either from reduced androgen receptor activity or lower circulating dehydroepiandrosterone sulfate) have a higher risk for colorectal cancer, however they did not correlate the hormonal influence with levels of gene mutations [55]. One theory to explain our results could be that male and female hormones may act as protective factors not exerting pressure on the expansion of KRAS mutant cells. However, upon menopause and andropause (events that occur at different ages for males and females) the decrease in hormonal levels could generate a pressure to stimulate a molecular switch in favour of clonal selection of cells containing KRAS mutations. Nevertheless, future studies on a large cohort corresponding to late age ranges are necessary to further confirm these results.

We found that variables such as region, age and sex-linked factors can correlate with KRAS mutational status. This is the largest study to point out a statistically difference in the prevalence of KRAS mutations between genders in a cohort of brazilian mCRC patients. The exact interaction between KRAS mutations, sexual hormones and the development of CRC is still not well defined. In the adenoma-carcinoma sequence, the proliferative hormonal exposure in the presence of ERα and the decrease of ERβ could be a factor to select KRAS mutant clones in the adenocarcinoma histologic subtype in females. Although estrogen could serve as a protective hormone in CRC, it might not prevent mutations in KRAS. Our group is currently evaluating experimentally the reasons behind the differences observed in this study. Hopefully, future research will be able to elucitate the molecular links between hormones, KRAS mutations, age and CRC development.