Colorectal Cancer: Molecular Features and Clinical Opportunities

Abstract

Colorectal cancer is a heterogeneous disease. There are three main pathways to colorectal cancer: the chromosomal instability pathway, the CpG island methylator phenotype pathway and the pure microsatellite instability pathway. Each of these is characterised by specific pathological precursors, mechanisms of carcinogenesis and natural history. The molecular features of these pathways have been exploited clinically in the diagnosis, screening and management of patients and families with colorectal cancer. This review summarises recent developments in our understanding of colorectal carcinogenesis and examines the interface between scientific discovery and the clinical application of molecular techniques in inherited and sporadic colorectal cancer.

Introduction

Colorectal cancer (CRC) is the second most frequently diagnosed internal malignancy in Australia and the second most common cause of cancer death.¹ In 2005, over 13,000 Australians were diagnosed with, and over 4000 Australians died from, CRC.^1,2 These are sobering statistics for a cancer with well-defined risk factors originating, in the vast majority of cases, from a slowly progressive precursor lesion within reach of, and cured by, colonoscopic polypectomy. Furthermore, surgical resection of stage I CRC is associated with 90% long term survival.³

CRC, however, is a heterogeneous disease. CRCs and their pathological precursors display distinct molecular signatures, distinct pathological features and distinct natural histories. There are at least three major molecular pathways to CRC including the predominant chromosomal instability (CIN) pathway accounting for up to 85% of cases, the CpG island methylator phenotype (CIMP) pathway which is the other major pathway to sporadic CRC⁴ and includes sporadic microsatellite instability (MSI) high cancers and finally the pure MSI pathway resulting from germline mutation in a DNA mismatch repair (MMR) gene. Hereditary nonpolyposis colorectal cancer (HNPCC) develops via the pure MSI pathway.

There is currently a multitiered approach to reducing the burden of CRC through both the earlier diagnosis of cancer and the detection and removal of benign polyp precursors. These clinical strategies have been informed by our understanding of the molecular pathology underpinning colorectal carcinogenesis. For the vast majority of the population that are asymptomatic and without any significant family history of CRC, screening is by an average risk strategy, such as faecal occult blood testing in Australia.² Symptomatic patients, or those with a past history of colorectal neoplasia are subjected to diagnostic and surveillance investigations usually colonoscopy. Patients likely to have familial CRC syndromes are identified and managed by combined endoscopic and molecular testing.

This review explores the molecular basis of colorectal carcinogenesis and touches on the potential clinical opportunities for molecular screening and diagnostic testing in the future.

Colorectal Cancer: the Pathways

CRCs are heterogeneous in terms of molecular momentum, regional distribution, pathology of the invasive and precursor lesions and natural history (Figure).⁵

Figure.

A simplified working-model of sporadic colorectal carcinogenesis: the CIN and CIMP pathways. In ‘Type A’ genes considerable promoter methylation may be observed even in normal tissue and is associated with tissue ageing. ‘Type C’ marker promoter methylation is more specific for neoplastic tissue.

Regardless of the underlying pathway, for a cancer to develop multiple and sequential genetic alterations must occur. Each genetic perturbation establishes a successive clone, with a ‘successful cancer’ requiring up to 10 clonal events, each characterised by relative growth advantage.⁶ These pre-cancerous cells must also develop a cellular environment permissive of future genetic and possibly epigenetic events, that is genomic and epigenomic instability, respectively.⁷ Genomic instability is critical in carcinogenesis to ensure that the subsequent events occur at increasingly greater likelihood. It accelerates the neoplastic evolutionary process by increasing the mutation rate induced by the background mutagenic challenge. Without genomic instability, the acquisition of new mutations would occur far too slowly for a cancer to develop during a person’s lifetime.⁸ As mentioned above, there are at least three main molecular and pathological pathways to CRC - the CIN, the CIMP and the pure MSI pathways.

The Chromosomal Instability (CIN) Pathway

Approximately 70%–85% of CRCs develop via the CIN pathway.⁹ In the CIN pathway molecular aberrations occur in significant part through the accumulation of numerical or structural chromosomal abnormalities (aneuploidy).⁶ The earliest identifiable lesion in this pathway is the dysplastic aberrant crypt focus (ACF),¹⁰ a microscopic mucosal lesion that precedes the development of a polyp.¹¹ The CIN pathway is associated with mutation in APC and/or loss of chromosome 5q that includes the APC gene, mutation of the KRAS oncogene, loss of chromosome 18q and deletion of chromosome 17p, which contains the important tumour suppressor gene TP53.⁹ Only a very small minority of CRCs characterised by CIN, however, possess a full complement of these molecular abnormalities.¹² It is possible that several of these clonal events can be bypassed by other genetic or epigenetic aberrations in order to deliver the necessary biological consequences.¹³ Sequencing the human genome has provided the opportunity for more detailed analysis of tumour-specific mutations with many potentially important genetic events now identified outside of the genes referred to above.¹⁴ Many of these will ultimately be shown to be additional, alternative, and perhaps complementary steps to these traditional genes within the adenoma-carcinoma sequence.

APC is an extremely important tumour-suppressor gene in the CIN pathway to CRC. Pathogenic mutations in APC frequently truncate the APC protein and interrupt the binding of APC to β-catenin. The binding of APC to β-catenin helps to suppress the Wnt-signaling pathway.¹⁵ Wnt signalling regulates growth, apoptosis and differentiation and is particularly relevant in maintaining tissue specific stem cell compartments.¹⁶ Loss of functional APC might also interfere with the careful regulation of mitosis contributing to CIN.¹⁷ The frequency of APC or β-catenin mutation in early adenomas has been reported to be as high as as 80%,¹¹ although the rate of APC mutation is significantly lower in some series.¹⁸ Mutation of APC is found in approximately 60% of colonic and 82% of rectal cancers.¹⁸

KRAS (12p12) is another important gene within the CIN pathway. KRAS encodes a GTP-binding protein which, when mutated, can cause a loss of inherent GTPase activity and thus constitutive signalling through the downstream, RASRAF- MEK-ERK pathway.¹⁹ BRAF is another important factor in this signalling cascade particularly relevant in the CIMP pathway, which is discussed below. Activating KRAS mutations are found in 35–42% of CRCs and in a similar number of advanced adenomas.^19,20 The role of KRAS is not unique to the CIN pathway, however, and KRAS has an important role in the CIMP pathway as well.²¹

DCC, SMAD2 and SMAD4 are all located at 18q21.1 and allelic loss at this site is found in up to 60% of CRCs.²² SMAD2 and SMAD4 are involved in the TGF-β signalling pathway, which is important in regulating growth as well as apoptosis. Germline mutation of SMAD4 can cause generalised juvenile polyposis syndrome, which is associated with CRC.²³

Finally, impairment of TP53 (17p13) usually through allelic loss of 17p is often a late event in the traditional pathway accompanying the transition from adenoma to adenocarcinoma. TP53 abnormalities, either mutation or loss of heterozygosity, increase relative to the advancing histological stage of the lesion being studied, with 4–26% of adenomas, 50% of adenomas with invasive foci, and 50–75% of CRCs having impaired function of TP53.¹⁹ The p53 protein normally acts to increase the expression of cell-cycle genes, to slow the cell cycle and provide sufficient time for DNA repair. Furthermore, when the genetic damage sustained is too great for the cell to repair, p53 induces pro-apoptotic genes, thus containing the genetic insult through programmed cell death.²⁴

The CIN pathway and its accompanying pathological adenoma-carcinoma sequence has provided a foundation for the molecular classification of colorectal carcinogenesis and has established a reference against which one may contrast other CRC molecular profiles, but it is now clear that CRC may develop by other means.

The Microsatellite Instability (MSI) Pathway

Another important type of genomic instability is MSI. Microsatellites are nucleotide repeat sequences scattered throughout the genome and MSI refers to a discrepancy, and thus instability, in the number of nucleotide repeats found within these microsatellite regions in tumour versus germline DNA. DNA polymerase is particularly susceptible to making errors when copying these short repeat sequences and thus mismatch repair (MMR) dysfunction results in MSI. The MMR system is composed of at least seven proteins, mlh1, MLH3, msh2, msh3, msh6, pms1 and pms2, which associate with specific partners to form functional heterodimers.⁷ MLH1 and MSH2 are essential in the mismatch repair machinery and form five functional heterodimeric proteins (MSH2- MSH3; MSH2-MSH6; MLH1-PMS1; MLH1-PMS2; MLH1- MLH3). Mutations in MLH1, MSH2, MSH6 and PMS2, have all been implicated in HNPCC. Many CRCs with an intact MMR system, however, will have frameshift mutations at a small number of microsatellites and thus a standardised panel of microsatellites was devised to provide uniformity of definition for research and practice.²⁵ The currently endorsed panel includes two mononucleotide (BAT25 and BAT26) and three dinucleotide microsatellites (D5S346, D2S123, and D17S250). Considerable MSI or MSI-high (MSI-H) is defined as MSI at ≥2 (40%) of the five specified sites, MSI-low (MSI-l) as MSI at one site, and microsatellite stable (MSS) when no instability is demonstrated at these markers. MSI leads to a dramatic increase in genetic errors and several microsatellites are present in genes implicated in colorectal carcinogenesis, such as MSH3, TGFBR2, BAX, CASP5, MSH6, CTNNB1, APC, IGF2, and E2F4.²⁶ MSI-H cancers are usually diploid, as MMR dysfunction provides the mechanism of genomic instability without the biological imperative for concomitant CIN.

Whilst HNPCC causes the pure form of MSI, the majority of MSI-H CRCs occur sporadically in the context of DNA methylation of the MLH1 promoter and the consequent transcriptional silencing of MLH1 expression. Such cancers exhibit both CIMP and MSI, and are considered in this article as part of the CIMP pathway. MSI-H tumours, whether sporadic or inherited, however, share similar biology.

The CpG Island Methylator Phenotype (CIMP) Pathway

The CIMP pathway is the second most common pathway to sporadic CRCs. The CIMP pathway accounts for approximately 15% of sporadic cases.^4,27–30 The CIMP pathway provides the epigenetic instability necessary for sporadic cancers to methylate the promoter regions of, and thus epigenetically inactivate the expression of, key tumour suppressor genes such as MLH1. CIMP-positive CRCs are currently defined by a panel of CpG island methylation markers, that are classified as having or not having DNA methylation on the basis of certain thresholds. The CIMP panel of genes and or markers is analogous to the panel of microsatellites used to determine microsatellite status.²⁷ There is not yet, however, one universally endorsed panel of CIMP markers or even a gold-standard technique for characterising methylation for the diagnosis of CIMP. There has been a great deal of research into the best markers and in our laboratory we have adopted a five marker panel CACNA1G, IGF2, NEUROG1, RUNX3, and SOCS1, examined by a methylation specific, probe-based real-time PCR technique called MethyLight.²⁷ In this panel, methylation is defined quantitatively and cancers are characterised according to their percent of methylated reference (PMR). Cancers with a PMR of ≥10 at three or more of these gene promoter sites are classified as CIMP positive.²⁷ Using this panel, there is a strong association between CIMP-positive cancers and BRAF mutation.²⁷ Some investigators, however, rightly advocate for the further refinement of the CIMP-positive group into CIMP-low (or CIMP2) and CIMP-high (or CIMP1) categories.²¹ A three-tiered system probably provides a more biologically consistent means of classification, with CIMP2 CRCs showing a closer association with KRAS rather than BRAF mutation, and there is the suggestion that CIMP2 CRCs may even develop from unique pathological precursors, possibly tubulovillous adenomas.^21,30 Dissecting out these detailed pathways is an area of intensive ongoing research.

CIMP-positive CRCs are characterised by a well defined cluster of clinicopathological features, including proximal location and a gender and age bias for the development of CIMP in older women.^28–30 The ultimate phenotype is influenced by the presence or absence of concomitant microsatellite instability, which may arise from gene promoter methylation-induced transcriptional silencing of MLH1. Classically, the CIMP-positive CRCs that are MSI-H share MSI-H characteristics, specifically the relative good prognosis, but in the absence of MSI-H, the CIMP-positive phenotype is characterised by more advanced pathology, poorer clinical outcome and an absence of tumour-infiltrating lymphocytes.³⁰ Very importantly, CIMP-positive CRCs differ from the other pathways with respect to their precursor lesion.^28–30 CRCs developing via the CIN pathway, and also in HNPCC, originate from adenomatous polyps.^31,32 In the CIMP pathway, however, sessile serrated adenomas are the chief pathological precursor.

Colorectal Polyps: Pathway Precursors

Adenomatous polyps with their well-recognised pathological progression from simple to advanced adenoma and ultimately invasive adenocarcinoma are the established precursors to the CIN pathway as well as HNPCC cancers.^18,33–35 As noted above, the serrated polyp precursors to the CIMP pathway are now also recognised as important polyps, although knowledge about the exact pathological sequence of serrated neoplasia is still evolving.^{18,30,34,36,37} These CIMP-precursor polyps are referred to as ‘serrated’ in order to describe the sawtooth appearance of their crypt lumina.³⁷ Their serrated crypt columns are generated due to an underlying defect in apoptosis shared by these polyps, in turn, the consequence of overactive RAS-RAF-MEK-ERK signalling. This leads to a build-up of colonocytes and thus the sawtooth appearance.³⁷

The earliest histopathology within the CIMP pathway may be the serrated, non-dysplastic ACF. There is some evidence, however, that there may be disturbed DNA methylation even within the apparently “normal” colorectal mucosa.^{38, 39} Therefore, these lesions may develop in the context of a primed or predisposed epithelial field. These early nondysplastic ACFs probably evolve into simple hyperplastic polyps, further characterised as mucin-poor, goblet-cell-rich or microvesicular. The goblet-cell-rich hyperplastic polyp is the typically diminutive rectosigmoid hyperplastic polyp, a common finding at colonoscopy and believed to have negligible malignant potential. In contrast, it is the microvesicular hyperplastic polyp that shows a high rate of BRAF mutation and thus is the most likely candidate to be the polypoid precursor to the CIMP pathway. BRAF mutation and CIMP marker methylation are seldom found in tubular adenomas.^40–44 Higher CIMP levels (methylation at four or more CIMP markers), in particular, are specific for the serrated pathway precursors, with approximately 30% of advanced serrated polyps being positive compared to 0% of advanced adenomas.⁴⁰ Sessile serrated adenomas, which are a more advanced polyp, are characterised by increased serration of the surface and luminal crypt epithelium, dilation of the crypt base, horizontal crypt branching, increased intracellular and/or luminal mucin, cytological atypia in the upper crypt, enlarged vesicular nuclei with prominent nucleoli and an increased epithelial to stromal cell ratio.³⁶ The pathological progression of serrated neoplasia may then include mixed polyps and traditional serrated adenomas. In the traditional serrated adenomas the dysplasia is an integrated component of the serrated polyp, whilst in the mixed polyp, the dysplasia exists as a discrete element, separate to the serrated architecture. All of the serrated polyps are likely to be related, but whether each represents a different stage of a uniform and sequential pathological process is still uncertain. Many of these polyps, however, share the common BRAF mutation and other characteristics of the serrated neoplasia pathway.⁴⁴ The serrated morphology of the precursor polyps is often lost following malignant transformation into the CIMP-positive cancer. Nevertheless, whilst all CIMP-positive cancers are not serrated adenocarcinomas, those that are share many molecular features with their serrated precursors.⁴⁵

Translating Discoveries into Practice

There are many examples of CRC pathology informing practice, particularly in familial CRC syndromes. The bench-to-bedside approach in such families has established colorectal medicine at the forefront of translational research and these families have greatly informed our understanding of CRC overall. Below we review some of the clinical circumstances in which molecular colorectal carcinogenesis research has informed our current practice.

Familial adenomatous polyposis (FAP) is an autosomal dominant syndrome characterised by >100 colorectal adenomas. Without surgical intervention, CRC occurs in essentially all cases of FAP by the age of 50 years.^2,46 FAP is due to a pathogenic mutation in APC, a very important gene in the CIN pathway, as described above. Interestingly, the site of mutation within APC influences the resultant phenotype both in terms of severity as well as the associated extracolonic features.⁴⁶ Mutations within the central region of the APC gene are generally associated with a higher colorectal burden of adenomas, whilst mutations at the 5′ or 3′ ends of the gene usually result in a milder phenotype, so-called attenuated FAP (AFAP). AFAP is characterised by <100 colorectal adenomas, and the development of CRC is delayed by approximately 15 years compared to classical FAP.⁴⁶ Another inherited polyposis syndrome, MUTYH-associated polyposis (MAP) is an autosomal recessive cause of multiple colorectal adenomas and cancer. MAP is often similar in phenotype to AFAP.^47,48 In contrast to FAP, MAP is primarily a DNA repair disorder. MUTYH is a DNA glycosylase which helps to repair mispaired bases that develop following oxidative DNA damage. MUTYH protects against mutations in genes such as APC and KRAS, critical in the CIN pathway to CRC.^46,48 Bi-allelic mutations in MUTYH confer a 93-fold increased risk of CRC.^49,50 Although local recommendations vary, clinical genetics referral, counselling and testing for germline APC and MUTYH mutations should be considered in patients with 20 or more colorectal adenomas (including metachronous lesions) or in patients younger than 60 years with >/= 5 adenomas and either a personal history of, or a first- or second-degree relative with, CRC or adenoma with high-grade dysplasia also before 60 years. The sequence of testing is often directed by the family history.

Several clinicopathological criteria have been developed to identify patients and families that are at higher risk of HNPCC. These criteria help to identify at-risk tumours that should undergo further molecular testing often by MSI analysis and immunohistochemistry for the four MMR proteins implicated in HNPCC i.e. MLH1, MSH2, MSH6 and PMS2 (see Table).^2,51–54 If the index CRC is MSI-H or has selective loss of protein expression in the tumour, then germline testing of the proband is organised. The first gene to analyse is directed by the immunohistochemistry results. It is important to appreciate, however, that loss of one member of a mismatch repair heterodimer pair, as described above, can lead to loss of the complementary member of the heterodimer. For instance MLH1 mutation can lead to the loss of both MLH1 and PMS2 staining on tumour immunohistochemistry.

Table.

The Revised Bethesda Guidelines for further testing of colorectal tumours, such as immunohistochemistry for MLH1, MSH2, MSH6 and PMS2 and microsatellite instability.⁵³

Tumours should undergo further molecular testing in the following situations:
Colorectal cancer diagnosed in a patient who is less than 50 years of age.
Presence of synchronous, metachronous colorectal, or other HNPCC-associated tumours,* regardless of age.
Colorectal cancer with the MSI-H** histology*** diagnosed in a patient who is less than 60 years of age.
Colorectal cancer diagnosed in one or more first-degree relatives with an HNPCC-related tumour, with one of the cancers being diagnosed under age 50 years.
Colorectal cancer diagnosed in two or more first- or second-degree relatives with HNPCC-related tumours, regardless of age.

^*HNPCC-related tumours include colorectal, endometrial, stomach, ovarian, pancreas, ureter and renal pelvis, biliary tract, and brain tumours, sebaceous gland adenomas and keratoacanthomas in Muir Torre syndrome and carcinoma of the small bowel.

^**MSI-H refers to changes in two or more of the five National Cancer Institute-recommended microsatellite markers.

^***Presence of tumour infiltrating lymphocytes, Crohn’s-like lymphocytic reaction, mucinous/signet-ring differentiation, or medullary growth pattern.

BRAF is not mutated in HNPCC cases and thus there is an emerging role for BRAF testing in MSI-H cancers, particularly when wishing to confirm that the cancer developed via the sporadic CIMP pathway and thus exclude HNPCC.^28,55 BRAF testing of the tumour may be helpful for clarifying the relevance of MSI-H cancers diagnosed in older patients, in whom MSI-H cancers are more likely to have resulted from CIMP related methylation of MLH1 rather than HNPCC.^28,54 It is important to note that not all HNPCC adenoma precursors will demonstrate abnormal immunohistochemistry or MSI as these are both relatively late events in the HNPCC adenoma-carcinoma sequence.⁵⁶ Using a combination of immunohistochemistry and MSI testing may miss up to a quarter of HNPCC adenomas, albeit that it is of course clinically very helpful when an abnormality is found.⁵⁶ Finally, it is important to note that specific MMR gene mutations are associated with differing clinical phenotypes. For instance, patients and families with germline mutations of PMS2 and MSH6 have a lower penetrance of CRC than those with mutations of either MLH1 or MSH2.^57,58

Beyond these familial examples, molecular analysis of CRC is also now mainstream in planning the management of advanced CRC. The monoclonal antibodies cetuximab and panitumumab, which target the epidermal growth factor receptor (EGFR) and thus the RAS-RAF-MEK-ERK pathway, have been an important development in the management of metastatic CRC. It became clear, however, that some patients were resistant to the clinical benefits of these agents. Patients with KRAS mutant cancers or those with wild-type KRAS but mutations in BRAF or PIK3CA are less likely to respond to these EGFR-directed therapies.⁵⁹ Cancers are now routinely tested for KRAS mutations prior to starting these therapies.

There is burgeoning interest in the application of molecular techniques for the screening and surveillance of patients at average risk of CRC. Serum technologies have been explored, including the testing for circulating tumour DNA encoding detectable mutations in APC, TP53 and KRAS.^60,61 Currently it is the molecular analysis of stool-based mutations, however, that may hold the greatest promise.^62–65 The multiple pathways to cancer, however, mandate either a very broad, inclusive marker panel or alternatively several specific panels each designed for use in distinct clinical circumstances. Molecular CRC screening in average risk people promises to develop into an important strategy in the detection of advanced colorectal neoplasia, which will ultimately be necessary for the prevention of CRC.

Summary

The discussion above outlined the three main pathways to CRC: the CIN, the CIMP and the pure MSI pathways. The CIN pathway is characterised by aneuploidy and is associated with the stepwise mutation or loss of APC, KRAS, SMAD2, SMAD4 and DCC, and TP53. The Wnt-signalling pathway is important in many of these cancers. The RAS-RAF-MEKERK pathway is another important pathway in CRC, in both the CIN and CIMP pathways. The CIN pathway passes through simple and then advanced adenomatous stages and, in the minority that develop into invasive disease, an MSS cancer. A minority of cancers, less than 5%, develop via the pure MSI pathway resulting from a germline mutation in a mismatch repair gene and develop from traditional adenomas. The majority of MSI-H cancers, however, occur sporadically following the epigenetic silencing of MLH1 and are thus part of the CIMP pathway, which also includes CIMP positive, MSS cancers. The CIMP pathway is usually associated with BRAF mutation but in BRAF wild-type cancers may be associated with KRAS mutation. These molecular features are exploited in determining which cancers are likely to have been caused by HNPCC. In the future the well-established partnership between the clinic and the laboratory will continue to identify new approaches to reduce the burden of CRC on our community.