Abstract
Background: BRAF gene encodes a serinethreonine kinase that inhibits the RAS/MAPK intracellular pathway. BRAF mutations occur at an early stage of colorectal cancer and their presence, 10-20% of colorectal cancer (CRC), is usually associated with inferior prognosis. Materials and Methods: From 41 consecutive CRC confirmed referrals from 1,446 suspected cancer cases (mean age=67.99+13.451, male=21, female=20), we retrospecti vely analyzed collected data from haemoglobin (Hb) and symptoms at presentation, location of tumor and stage of the disease, including lymphovascular invasion (LVI). Gene profile analysis data (KRAS, BRAF) were retrospectively collected and associated with the presentation profile above. Results: There was no significant correlation in presentation Hb levels and eventual disease staging (p>0.05 for all associations). Patients with right-sided tumours were found to have a lower Hb level than patients with either left-sided colonic or rectal tumours. Hb levels were also significantly lower in patients with the BRAF V600E mutation. KRAS status or LVI status did not have a specific correlation with Hb levels. Conclusion: BRAF V600E mutation might be associated with right-sided tumors and subsequently related unexplained iron-deficiency anaemia (IDA) at presentation. This finding may affect the choice of clinical strategy for investigation of unexplained IDA. Further research should be conducted in order to identify and support the potential biological explanation of the findings above.
Colorectal cancer is the third most common cancer worldwide and there has been a significant progress in understanding its pathogenesis (1). Currently, there are three distinct molecular pathways to explain its complex pathogenesis. The first one is chromosomal instability (CIN) where the aetiology is attributed to numerical (aneuploidy) or structural loss of chromosomes, whereas loss of heterozygosity (LOH) is a classic component of this theory. CIN is linked with several point mutations in genes, which are thought to be involved in the CRC pathway (APC, KRAS, BRAF, etc.) (1, 4). The second pathway consists of a fault of the mismatch repair (MMR) system of the DNA and this is defined as microsatellite instability pathway (MSI) (1, 5, 6). MSI is a unique feature of Lynch syndrome (LS), although it was also found in sporadic cancer. Lastly, a newer theory is that of C-pho sphate-G (CpG) island methylation (CIMP), which can result in CIMP-high and CIMP-low phenotypes. BRAF tends to be associated with CIMP-high tumour phenotype (1, 7, 8).
V-raf murine sarcoma viral oncogene homolog B (BRAF) gene is involved as an inhibitor in the RAS/RAF/MAPK pathway (9, 10). Soma tic mutations in BRAF gene are found in 10-20% of colo rectal cancer cases (11) and result in activation of BRAF serine-threonine kinase and subsequent up regulation of the RAS/RAF/MAPK signalling pathway (9, 12). This can result in abnormal cellular growth, invasion and metastasis (2).
BRAF V600 E mutation is the most well -studied mutation and occurs in 90% of BRAF mutations (12-14). It results in a substitution of glutamic acid for a valine at amino acid position 600 (c.f. 1799T>A) (15, 16).
In terms of BRAF mutations and their relationship to other genetic events in CRC pathogenesis, there has been documented an association between BRAF-mutant status and sporadic MSI-high tumours (11, 17-19). Lochhead et al. state that combined BRAF is related with MSI profile through its association with high level CIMP and MLH1 promoter methylation. The same study concluded that BRAF/MSI status could be used as a prognostic tool (11).
There are existing data which associate BRAF status with certain macroscopic features. Roth et al. 3 concluded that BRAF mutations are signi fi cantly associated with female sex, right-sided location, older age, high grade and MSI-high tumours (20). A recent metanalysis 3 concluded that BRAF V600E is linked with advanced TNM stage, poor differentiation, mucinous histology, MSI, CIMP, as well as female gender, older age, proximal location and mutL homolog 1 (MLH1) methylation (21).
Cancer stage in the study patients.
There has been a long discussion about the prognostic and predictive value of BRAF mutational profile (22-24). Recent studies associate BRAF mutations with poorer pro gnosis (11, 24) and support their role in the response to chemotherapy (15, 16, 25).
Materials and Methods
All our data derive from the urgent referral (“2 week-wait” or “2WW”) Cancer Database. 2WW is a pathway where general practitioners (GP) refer suspected colorectal cancer cases to a specialized unit in the UK. We used as a source of information the original GP referral letter, as well as the Colorectal Clinic letters in order to identify the presenting symptoms of each patient. Biochemistry profile information (haemoglobin (Hb)), histology reports and molecular diagnostics were kept online on Electronic Patients' Records (EPR) database.
We retrospectively identified 41 confirmed CRC cases from 1,446 consecutive referrals for suspected cancer. From those patients, 21 were male (51.2%), 20 female (48.8%). Mean age was 67.99 years and standard deviation (SD) 13.451 years (Figure 1).
Data were collected on clinical presentation profile, i.e. changes in bowel habit (CIBH), rectal bleeding (RB), unexplained irondeficiency anaemia (IDA), weight loss, abdominal mass and family history (FH). Moreover, we have focused on the Hb levels at presentation, as well as tumour location and staging profile including lymphovascular invasion (LVI).
KRAS and BRAF mutations ware defined using the pyrosequencing technique from extracted tumour DNA. The whole series of those were performed in our Advanced Diagnostics Lab.
Data were analysed on IBM SPSS Statistics for Macintosh (Version 22.0. Armonk, NY: IBM Corp.) using t-test and ANOVA test. Finally, cross-referencing of our results was performed using existing data in the literature.
Results
The mean overall initial Hb level was 115.8 (SD=20.4) g/dl, with 10 patients being referred with a diagnosis of IDA (24.4%). Amongst those referred with IDA, the mean Hb was 95.3 g/dl, significantly lower than those not identified as having IDA who had a mean Hb of 124.3 g/dl (t-test, p<0.0001). Twenty patients were referred with rectal bleeding (48.8%), 26 (63.4%) with a change in CIBH, 6 (14.6%) with weight loss and 1 (2.4%) with a significant family history.
Relationships of Hb level to clinical features (ANOVA associations).
Hb levels were significantly lower in patients referred with rectal bleeding (107.1 g/dl with and 126.3 g/dl without, p=0.002). There was no significant difference in Hb level between patients according to CIBH (119.6 g/dl with CIBH and 111.5 g/dl without, p=0.245) or abdominal mass (109.3 g/dl with a mass, 118.3 g/dl without, p=0.323).
All patients had a confirmed diagnosis of CRC: 10 right-sided (24.4%), 31 left-sided (75.6%), comprised of 16 patients with left-sided colonic (39.0%) and 15 with rectal tumours (36.6%). Overall, there were 26 patients with a colonic primary (63.4%) and 15 (36.6%) with a rectal primary.
There were 8 patients (19.5%) with stage I disease, 12 stage II (29.3%), 15 Stage III (36.6%) and 6 stage IV (14.6%). There was an uneven distribution of patients by stage, with relatively more patients with stage II and III disease than I and IV disease (see Table I and Figure 2, p=0.003), which was not altered significantly between those with colonic or rectal primaries (Table II, p=0.212). The presence of LVI was available in 36 patients of whom 25 had no LVI (69.4%) and 11 with LVI (30.6%).
Of the 41 patients, KRAS status was available in 40, of whom 28 were wild-type (wt, 70.0%), 7 mutation 12 (17.5%) and 5 mutation 13 (12.5%). BRAF was available in 24 patients of whom 14 were wt (82.5%) and 3 with V600E mutation (17.6%).
Age distribution.
There was no significant correlation in presentation Hb levels and eventual disease staging (p>0.05 for all associations). Patients with right-sided tumours were found to have a lower Hb level than patients with either left-sided (p=0.017) or rectal tumours (p=0.009). Hb levels were also significantly lower in patients with the BRAF V600E mutation (p=0.009). KRAS status was not found to be associated with Hb levels (p>0.05). LVI status was also not found to be linked with Hb levels (p>0.05).
Discussion
There has been a long discussion about clinical significance of BRAF V600E mutation in colorectal cancer (20, 23, 24). Recent studies conclude that BRAF V600E mutation is associated with poorer prognosis (11, 24) and support its role in the response to chemotherapy (10, 16, 26). More specifically, Mao et al. state that BRAF V600E is associated with resistance to anti-epidermal growth factor receptor (anti-EGFR) monoclonal antibodies in patients with metastatic CRC and wt KRAS (15). Phipps et al. (12, 27) conclude that BRAF mutational prognostic value can vary, depending on patient and tumour characteristics. Ciombor et al. (28) highlight the significance of BRAF in terms of prognostic value, as well as a new CRC therapy target.
However, there is also emerging data, which liaise BRAF V600E with MSI-high tumours (2). In those cases, there is still a negative effect of the V600E mutation on overall survival (29), although some other researcher still doubt this and support that BRAF V600E has a negative prognostic effect only in microsatellite stable (MSS) tumours (30, 31).
Cancer stage.
BRAF V600E is generally linked in the literature with certain clinicopathological features, i.e. female sex and mucinous differentiation (1, 2, 32, 33). There is a well-defined relationship between V600E mutation and generally poorer prognosis; however, no convincing explanation has been given yet.
Our study concludes that there is a well-defined relationship between significantly lower Hb levels and V600E mutant neoplasms comparing to the wt samples. There is no doubt that this could be attributed to the fact that all those patients presented with right-sided tumours and also by the fact that all BRAF V600E specimens were advanced-stage tumours (T3b and above). However, this observation could alter the significance of the presence of IDA in the clinical investigations pathway by demonstrating that IDA may signify a higher risk cancer. This may well affect the choice of treatment and would we should look for BRAF mutation in unexplained IDA, where colorectal cancer is suspected. Of note is the fact that we could not identify any statistical significance between Hb levels and disease stage (I-IV).
The main question that this study raises is at what stage of the carcinogenesis pathway does BRAF interfere. According to the literature, most of the BRAF V600E specimens appear to be in advanced stage (20, 21, 32). It is, thus, quite important to establish in which time point BRAF interferes as it appears that this happens on a mature stage in the pathway. Taking into consideration the latter, it is quite important to interpret the potential link between IDA at presentation and BRAF V600E.
A recent promising study (33) suggests a multitarget stool DNA test with faecal immuno chemical test (FIT) as a non-invasive diagnostic technique for CRC screening. The sensitivity for detecting CRC was 92.3% with DNA testing and 73.8% with FIT detecting versus 42.4% with DNA testing and 23.8% with FIT in advanced precancerous lesions. This test does not include BRAF. The main question is whether there would be any clinical importance to incorporating BRAF screening in these tests, especially in those cases where IDA is noted at presentation. This would help in increasing the sensitivity of these methods and IDA could be used as an alarm point where advanced cancer with mutant BRAF is suspected. In terms of precancerous lesions, again, our question regarding the stage of carcinogenesis that BRAF interferes with, could be a crucial point in order to improve the sensitivity of this method, as well as stratifying the risk of potential progression to cancer. In other words, if we knew the exact point where V600E mutation appears, then we would be able to tell if a precancerous lesion has a potentially higher risk of progression to cancer.
Moreover, we tried to liaise potential clinical presentation features with the molecular status of KRAS and BRAF, which are thought to play a significant role in prognosis of CRC, as well as response to neoadjuvant chemotherapy. The main question was whether there would be any unique clinical presentation or histopathology features, which could potentially be liaised with a known mutation of KRAS or BRAF.
With regards to KRAS mutational status, there was no significant relationship between disease stage and KRAS status (p>0.05). This is in line with literature, as KRAS is not an established prognostic marker (20). It is well known that in many centres around the world, KRAS can be used as a predictive tool to identify whether a patient would benefit from anti-EGFR chemotherapy agents (16, 33, 35, 36). Never theless, despite its well known predictive value, there is no conclusion as to whether KRAS status could be used as a prognostic tool, with the vast majority of researchers doubting its significance in terms of defining disease-free survival (33, 35). An interesting point from or results is that there was no direct relationship between Hb levels or clinical presentation and KRAS status (p>0.05). This could be easily explained by the fact that CRC pathogenesis is quite complex and a single gene is difficult to be linked to specific macroscopic presentation features unless there is an established strong relationship between certain histopathology characteristics and molecular status. Similarly, there was no defined statistical relationship between LVI and KRAS mutational status (p>0.05), which could be attributed again to its controversial prognostic value.
KRAS mutations tend to appear in earlier stages of the carcinogenesis pathway (1). Contrary to BRAF V600 E neoplasms, where there is an established relationship with proximal location, KRAS-mutant tumours tend to have a bimodal distribution in the proximal-distal axis and tend to appear more proximally and to the caecum (37).
On the other hand, BRAF V600E mutation is well associated with proximally located tumours (32). On our sample, even though it is quite small, there was a well-defined relationship between right-sided neoplasms and V600E mutation.
Nevertheless, there was no defined statistical relationship between LVI and BRAF status (p>0.05) and this is difficult to comment based on the limitations of our sample. Positive LVI is associated with a slightly lower Hb level comparing to LVI negative (109.7 g/dl vs. 115.1 g/dl).
In conclusion, given the complexity of the molecular pathways described, as well as the fact that many molecular agents interfere one each other, it is quite difficult to establish a clear picture for the unique impact of every gene contributing to CRC pathogenesis. Our study concludes by raising a question regarding the stage of carcinogenesis where BRAF appears to be mutant. It is quite important to consider whether IDA could also be used as a potential marker of advanced cancer and, given its relationship with proximal tumours and subsequently with BRAF V600E mutation, this could be used as an starting point to question whether V600E mutation could be incorporated into DNA screening methods that were recently released. This would be indisputably useful in cases where IDA could potentially be a sign of more advanced disease.
- Received December 30, 2014.
- Revision received January 22, 2015.
- Accepted January 26, 2015.
- Copyright© 2015 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
