Colorectal cancer (CRC) is the second leading cause of cancer death in the United States of America, after lung cancer, having the third highest incidence in both women and men. Worldwide, about 1,000,000 cases are diagnosed annually, which represents about 9-10% of all new cancer cases. A peak of the incidence is observed in Europe, USA, Australia and New Zealand, while low rates are observed in India, South America and in the Arab population in Israel. Both the incidence and mortality have steadily decreased in the United States since 1985, a consequence of the increase in screening and diagnostic methods. However, colorectal cancer remains an extremely important health issue, with a high variability in incidence worldwide. Studies on migrant populations have revealed that the incidence of colorectal cancer reflects the country of adoption and not the country of origin, which leads us to the idea that environmental factors play a very important role along with genetic factors in the production of this type of cancer(1).
Most colorectal cancers appear sporadically and are characterized by a sequential carcinogenicity process involving the progressive accumulation of mutations over an average period of 10-15 years. It has also been suggested that the diaphragm between genetic mutations (i.e., BRAF, KRAS and p53 mutations, as well as the instability of microsatellites) and epigenetic changes (i.e., DNA methylation of promoter regions of CPG islands) could play a key role in cancer development(2,3).
This long development interval allows a successful screening, the early detection of cancer and the removal of premalignant lesions (adenomas) leading to a reduction of the incidence and mortality(4).
The standard in the current screening strategy is colonoscopy. Guides recommend that persons aged 45 and above with an average risk of CRC should be subjected to regular screening. However, colonoscopy has a poor patient compliance because the procedure is expensive, invasive and presents risks such as hemorrhage, colon perforation and cardiorespiratory complications. Other reasons for low adhesion are related to bashfulness, discomfort of procedure and bowel preparation(5).
Thus, due to the problems listed above, it becomes clear that the clinical applications of biomarkers in colorectal cancers are not only necessary for the early detection of the disease, but are also essential for forecasting stratification, surveillance and selection of therapy(6).
Recent studies have shown that colorectal neoplasm is not a single disease, therefore several pathological entities could be individualized on the basis of clinical evolution, tumor and genetic markers. This peculiarity of colorectal cancer is a consequence of heterogeneity in its molecular biology that leads to differences in response to different therapies and prognosis. A much more detailed classification of colorectal cancer would therefore be necessary. Currently, the TNM stage, the histopathological characteristics (tumor grade, lymphocyte and perineural invasion, inflammatory infiltration) and the immunohistochemical and genetic characteristics (microsatellite instability, CIN, BRAF gene V600E, APC, TP53, DPC) give us information on the genetic type of cancer, prognosis, and evolution under treatment. In practice, the mutational statuses KRAS and NRAS are used (in metastatic cancers) for predictability to tyrosine kinase inhibitor therapy. It has also been shown that the molecular biology of colorectal cancer implies proinflammatory conditions to promote tumor progression, invasion and metastasis. It is well known that patients with an inflammatory intestinal disease have a higher risk of colorectal cancer. There is a lot of evidence that reinforces the link between inflammation and colorectal cancer. Inflammation involves the interaction between different immune cells, inflammatory cells, chemokines, cytokines and proinflammatory mediators, such as cyclooxygenase (COX) and lipoxygenase (LOX) routes, which may lead to signal the tumor cell proliferation, growth and invasion(7,15).
In order to reduce mortality in colorectal cancer, it is necessary to know very well the mechanisms leading to the emergence and development of this neoplastic site and, last but not least, the biomarkers which help clinicians make the best decisions(8).
The National Health Institute (NIH) defines a biomarker as a biological molecule that is found in blood, other body fluids or tissues, which is the sign of a normal or abnormal process, of a disease or of other health condition. In addition, beside the detection of a disease, a biomarker can be used to assess how well the body responds to a treatment and can also be called a molecular marker or signature molecule. A biomarker may exist in relation to DNA, RNA, micro-RNA, epigenetic changes, proteins and even the expression of antibodies. The discovery of biomarkers is expanding, with the modern growth of medical research in genomics and protein. New biomarkers are developing on the basis of growing information, while research tools are becoming cheaper and more efficient(9).
Tumor markers are macromolecules whose presence and changes in concentration are related to the initiation and development of malignant tumors. Tumor markers can be classified into two types: cellular and humoral. Cell tumor markers are antigens located in cell membranes (cell markers in leukemia), hormonal receptors or growth factors. Tumor markers are substances that can be detected in much higher concentrations than under physiological conditions, in serum, urine or other body fluids. These substances are synthesized and excreted by tumor tissues, released by disintegration of tumor, or formed as a reaction of the organism to the presence of a tumor.
An ideal tumor marker should meet the following criteria:
high specificity (not detectable in benign disease or healthy people);
high sensitivity (detectable even in the presence of a small number of tumor cells);
to be organ-specific;
correlation with tumor stage and mass;
correlation with the disease prognosis;
good predictive value.
As with tumor markers for other sites, the ones for colorectal cancer can be obtained from multiple sources, in this case from tumor, stool or blood.
Tumor genetic biomarkers
a) Microsatellite instability
Colorectal tumors can be classified according to molecular markers. The current conventional molecular tests used to evaluate patients with colorectal cancer include the analysis of microsatellite instability (MSI) and the analysis of BRAF and KRAS mutations. The presence of MSI and/or certain mutations or epimutations in colorectal tumors provides clinicians with the information they need to choose the appropriate treatment. Due to the high variability of clinical responses to colorectal cancer treatment, it is necessary to identify new predictive and prognostic molecular classifications in order to make the best treatment decisions for the patient, taking into account their prognosis, as well as the predicted response to chemotherapy.
Mutations and/or gene epimutations involved in the DNA Repair System (MMR), MLH1, MSH2, MSH6 and PMS2 lead to very repeated changes in DNA sequences (microsatellites). MSI is a hallmark of Lynch syndrome, an inherited colorectal cancer syndrome, and is used as a diagnostic marker for this disease. In sporadic colorectal cancer, 10-15% of tumors have MSI, although somatic mutations are rarely found in MMR genes. MLH1 methylation is responsible for the majority of sporadic colorectal cancer with MSI.
A panel of five mononucleic markers (Bat-25, Bat-26, NR-21, NR-24 and MONO-27) is currently used by most clinical laboratories to detect MSI. The definition of MSI (also known as MSI-high) is based on the fact that it has ≥30% unstable loci using mononucleotide and dinucleotide markers; tumors with 10-29% unstable loci in the panel are considered MSI-low. The absence of MMR protein expression in tumor tissue is also used as an indicative surrogate test of MSI. Sporadic colorectal tumors with MSI are located mostly in the proximal colon, present with mucinous or ring histology, are poorly differentiated, have an abundance of tumor infiltrating lymphocytes, and have BRAF mutations. There is still controversy as to whether tumors with low MSI are a colorectal cancer subtype.
The accumulated evidence suggests that MSI status may be a sporadic prognostic and predictively useful colorectal cancer marker. MSI is associated with increased patient survival and a favorable prognosis. Data from retrospective studies, population studies and meta-analyses claim that patients with colorectal MSI tumors perform better than patients with stable tumors in microsatellites. In a population-based study of 607 patients, those with MSI had a more favorable prognosis and a low probability of lymph node and systemic metastasis.
Further studies argue that MSI may be an independent predictor of less aggressive disease and better results; however, the use of MSI as a prognostic biomarker has not yet been implemented in the clinic. Although controversial, several studies have reported an association between MSI and the development of additional colorectal carcinomas in sporadic colorectal tumor patients, suggesting that MSI status may also be useful as a predictor of the risk of developing metachronous colorectal cancer. Further studies are needed to provide strong evidence that the sporadic management of colorectal cancer patients should take into account MSI status.
In confirmed cases of hereditary nonpolyposis colorectal cancer (also known as Lynch syndrome), mutations in the MMR gene are used to make treatment decisions. Tumors that are MSI-high typically mean better outcomes and higher rates of survival compared with tumors that are MSI-low or stable; however, this particular mutation doesn’t respond as well to 5-flurouracil (5-FU)-based chemotherapy compared with the same as previous (MSI-low or stable).
With familial adenomatous polyposis, tumors that feature the CIN phenotype and APC gene mutations are markers of a poor prognosis, but they do help to identify high-risk patients with stage II disease who could benefit from adjuvant chemotherapy(10).
The accumulated evidence claims that MSI status can predict the ability to respond to adjuvant chemotherapy. Reports from clinical trials, retrospective cases and meta-analysis reveal that patients with MSI tumors do not benefit from 5-fluorouracil adjuvant chemotherapy compared to patients with stable microsatellite tumors (MSS). In a study-based meta-analysis that layered the overall survival of the patients with 5-FU therapy, the survival data in patients with colorectal MSI tumors were heterogeneous compared to patients with MSS tumors. A significantly improved overall survival was found only in patients with MSS tumors after 5-FU therapy. In addition, some studies have also proposed MSI status as a predictive marker for irinotecan response. The value of using MSI as a predictive marker for chemosensitivity remains controversial and is still being evaluated.
b) KRAS mutations
Gene mutations associated with chemoresistance to certain compounds are currently used as predictive markers in colorectal cancer to identify the best treatment regimen for patients. The detection of KRAS mutations is currently the most widely used predictive marker for response to anti-EGFR (epidermal growth factor receptor), cetuximab and panitumumab-based therapies. However, recent studies have reported compelling evidence that, in addition to KRAS, mutations in NRAS predict nonresponse to anti-EGFR therapy. These studies support the use of extended RAS mutational analyses (KRAS and NRAS) as negative predictive markers for anti-EGFR therapy in metastatic colorectal cancer (mCRC).
KRAS protooncogene encodes a small G protein (guanosine triphosphate binding protein/guanosine diphosphate) downstream of EGFR in the PI3K/PTEN/AKT and RAF/MEK/ERK signaling pathways. Most active mutations, about 90%, are found in codons 12 and 13 of exon 1. Almost 5% of mutations are found in codon 61 of exon 2. NRAS (homologous viral oncogene of RAS neuroblastoma), a gene closely related to KRAS, is mutated to 3-5% of CRC cases, especially in codon 61. It has been associated with weak responses to cetuximab and panitumumab-FOLFOX4.
The activation of KRAS mutations has been significantly associated with a lack of response to anti-EGFR therapies in patients with mCRC. Tumors with KRAS mutations have response rates to cetuximab and panitumumab ranging from 26% to 41%, respectively from 11-17%. Approximately 60-70% of wild-type mCRC patients with KRAS mutations have a limited response to EGFR antibody therapy, suggesting that additional mutations may contribute to the response to anti-EGFR treatment. Clinical guidelines established by the NCCN and ASCO recommend KRAS mutational analysis for patients with mCRC prior to the use of cetuximab and panitumumab. Furthermore, there are reports that codon mutations 12 and 13 are associated with weaker prognosis and poor survival in mCRC patients. However, it has not been clearly established whether the worse prognosis associated with KRAS mutations is independent of the treatment regimen used. The use of KRAS as a predictive biomarker for response to anti-EGFR therapies is the standard of care in patients with mCRC and the first instance of personalized medicine for these patients.
c) BRAF mutations
BRAF, an RAF serum/threonine kinase gene, is the immediate downstream effector of KRAS in the Ras/Raf/MAPK signaling path. BRAF gene mutations have been associated with CRC development and are present in 40-50% of sporadic CRC with high MSI. They are absent in patients with Lynch syndrome, which makes the condition of the BRAF mutation a very useful diagnostic tool to distinguish between familial CRC and sporadic CRC. A wrong mutation that results in a substitution of vanillic acid to glutamic (V600E) is the most common mutation observed. KRAS and BRAF mutations are generally mutually excluded in colorectal tumors. Recent studies suggest that BRAF mutations can be also used as predictive markers for EGFR-oriented therapy. Mutations in BRAF are associated with poor prognosis; however, it is not yet clear whether this combination is independent of treatment. Due to their increasing importance, NCCN recommends testing the BRAF and RAS mutations (exon 2 KRAS and non-exon 2 NRAS) for the diagnosis of CRC in stage IV. Based on this evidence, BRAF mutations can be used as a biomarker for patients with CRC metastasis who may benefit from anti-EGFR antibody therapy(11).
A meta-analysis performed by Rowland et al. to evaluate the effect of BRAF mutation on anti-EGFR antibody therapy revealed that there was insufficient evidence that BRAF could be used as a predictive marker for anti-EGFR antibody therapy success. Although the hazard ratio for overall survival (OS) and progression-free survival (PFS) with anti-EGFR antibody therapy was higher in RAS wild-type/BRAF mutated tumors than in RAS wild-type/BRAF wild-type tumors (OS: 0.97; 95% CI, 0.67-1.41 versus 0.81; 95% CI, 0.70-0.95, and PFS: 0.89; 95% CI, 0.61-1.21 versus 0.62; 95% CI, 0.50-0.77, respectively), the results were not statistically significant (P=0.43 and P=0.07, respectively)(12).
The rationale behind stool-based molecular diagnostic tests for CRC is based on the mechanisms that lead to the presence of these markers in the stool. The presence of tumor markers in the stool can be attributed to leakage, exfoliation or secretion. Disruption of blood vessels by tumor growth leads to the leakage of markers into the lumen of the colon. This process may not be continuous and also occurs in neoplastic lesions. As a result, short markers have limited sensitivity and specificity. The exfoliated and secreted markers come from colonocytes, both vital and apoptotic, thrown into the colorectal lumen. It is assumed that these types of markers from the stool are very specific, because they are derived directly from the tumor. Although considerable efforts are being made to identify DNA, RNA and protein markers present in the stool and to develop new detection methods, only three stool-based clinical tests are currently available for the diagnosis of CRC: gFOBT, fecal immunochemical test (FIT), and the detection of vimentin methylation.
gFOBTs are based on the detection of the activity of hem pseudoperoxidase in stool samples resulting from bleeding in adenomatous or neoplastic lesions, thus detecting the presence of occult blood. The most commonly used gFOBT are Hemocult II and Hemocult II SENSA with guaiac infusion (which has an improved sensitivity; Beckman Coulter, Brea, CA). One of the disadvantages of this method is that, since colorectal bleeding may be intermittent, this test should be performed several times to be sensitive. In addition, gFOBTs are not specific to human pseudoperoxidase and can detect bleeding from any place in the gastrointestinal tract. Before testing, patients should join a three-day diet that eliminates meat and NSAID, as ingestion of certain foods and drugs can cause false-positive results. However, randomized controlled clinical trials in the United States, England and Denmark have shown that gFOBT testing once or twice a year significantly reduces CRC mortality.
An alternative to gFOBT is FIT, which has a higher sensitivity and an improved detection rate for advanced neoplasia. This test uses monoclonal or polyclonal antibodies to detect globin in human hemoglobin using an enzyme-linked immunosorbent assay (ELISA). FIT is specific to bleeding from the distal gastrointestinal tract due to the fact that globin gradually degrades as it passes through the intestine. Two types of FIT have been developed: a qualitative test that requires visual interpretation, and a quantitative test which is automatically analyzed and determines the amount of hemoglobin present in the sample.
Extensive research has been done to identify CRC-specific antigens in the blood. However, there are only two blood-based biomarkers available to monitor patients with colorectal cancer: CEA and carbohydrate antigen 19-9 (CA19-9). CEA, a high molecular weight glycoprotein, is found in embryonic tissue and colorectal malignancies. It was discovered in 1965 and is the only acceptable tumor marker that has monitored CCR recurrence to date. Elevated CEA levels are considered a poor prognostic factor for resectable CRC and correlate with cancer progression. Using this marker, the sensitivity increases depending on the stage of the tumor; CEA levels decrease after tumor resection. However, high blood levels are not specific for CRC and can be caused by other diseases, such as irritable bowel syndrome, liver disease, pancreatitis and other malignancies. Moreover, the detection of this serum antigen marker is not an effective method for CRC screening, as elevated CEA levels are detected only in advanced stages of a fraction of all CRC patients. CA19-9 antigen, compared to CEA, is less sensitive and specific for CRC. However, it is the best marker available for detecting pancreatic-biliary malignancies. Therefore, CEA is still the antigen chosen to be used as a prognostic marker after diagnosis and to monitor disease progression(13).
In one study, Yangfeng Gao et al. examined the sensitivity and reliability of single/multiple serum markers for diagnosis and also assessed their relationship to pathological parameters in a total of 279 colorectal cancer patients. The study showed that carcinoembryonic glycoprotein antigen has the highest sensitivity among the single marker in the order CEA>CA 72-4>CA19-9>ferritin>CA125, while the most sensitive markers combined for two to five are the following: CEA + CA72-4; CEA + CA72-4 + CA125; CEA + CA19-9 + CA72-4 + CA125; and respectively CEA + CA19-9 + CA72-4 + CA125 + ferritin. They also showed that patients who had preoperative serum positive for CEA, CA19-9 or CA72-4 were more prone to lymph node invasion, CA125-positive were more prone to vascular invasion, and CEA-positive or CA125-positive were correlated with perineural invasion. In addition, positive CA19-9, CA72-4 or CA125 were associated with a poorly differentiated tumor, while CEA, CA19-9, CA72-4 and CA125 levels were positively correlated with tumor – nodule – metastasis pathological stages. The study concluded that combined serum markers can be used not only to diagnose colorectal cancer, but also to assess tumor status in order to guide treatment, assess curative effect and predict prognosis in patients(14).
Colorectal cancer is one of the leading causes of death by cancer worldwide. This disease is very heterogenous, with a lot of alterations in multiple molecular pathways. For clinicians, the most important is the applicability of biomarkers in CRC for the early detection of the disease, being also essential for prognostic stratification, surveillance and therapy selection.
The most well-known and used in practice is represented by mutation in RAS, along with the mismatch repair gene deficiency. This mutation is routinely tested in clinics. This kind of biomarkers provide information for patient’s prognostic and for the choice of the adequate treatment. Unfortunately, reliable and powerful prognostic markers for recurrence of CRC, to establish who might benefit from adjuvant chemotherapy, are missing at that time. For that reason, genomic information has increasingly gained interest as a potential method for determining the risk of recurrence. Because there are several limitations of gene-based signatures, this methodology has not yet been clinically implemented.
Biomarkers can greatly improve the selection of treatment strategies for CRC patients. At this time, most of these markers are used to inform clinicians of the overall prognosis of the disease, but they fail to guide physicians for therapeutic decisions(15). Most frequently used biomarkers are: KRAS and BRAF genes and MSI status. The conclusion is that additional studies are needed to identify and validate novel biomarkers in order to improve the outcomes for CRC patients. Taking into account that immunotherapy is increasingly used in the treatment of cancer, it is necessary to determine new biomarkers that predict the efficacy of immune-based therapies for patients with colorectal cancer.