Genetics of colorectal cancer: hereditary aspects and overview of colorectal tumorigenesis

Genetics of colorectal cancer: hereditary aspects and overview of colorectal tumorigenesis Nicola S Fearnhead *,, Jennifer L Wilding* and Walter F Bodmer* *Cancer Research UK, Weatherall Institute of Molecular Medicine and Department of Colorectal Surgery, John Radcliffe Hospital, Oxford, UK Correspondence to: Sir Walter Bodmer, Cancer and Immunogenetics Laboratory, Cancer Research UK, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Oxford OX3 9DZ, UK Familial adenomatous polyposis and hereditary non-polyposis colorectal cancer are dominantly inherited conditions with 100% and 80% life-time risk of developing colorectal cancer, respectively. The genetic mutations responsible for these two conditions lie in the adenomatous polyposis coli (APC) and mismatch repair genes. These same genes also play a key role in the formation of sporadic colorectal cancers, which arise on a background of a similar spectrum of mutations to the hereditary cancers. This article examines the genetic mechanisms underlying the hereditary colorectal cancers, as well as genetic predisposition to colorectal cancer in the general population in the absence of a clear-cut genetic syndrome. Colorectal cancer arises as the cumulative effect of multiple mutations within the cell, allowing it to escape growth and regulatory control mechanisms. This step-wise progression of mutations facilitates the histological transition from normal mucosa to adenoma to carcinoma. The latter part of this paper focuses on the key genetic events underlying this process and provides an overview of the genetic mechanisms responsible for colorectal tumorigenesis. Colorectal cancer (CRC) is one of the commonest cancers in the Western world. While the majority of cases of CRC are sporadic, a significant minority occur as a result of an inherited genetic mutation. These inherited syndromes of CRC, together with the presence of a readily identifiable precursor lesion in the form of the adenomatous polyp, have greatly facilitated research into the genetic mechanisms responsible for colorectal tumorigenesis. Inherited predisposition to colorectal cancer Familial adenomatous polyposis (FAP), the polyposis syndromes and hereditary non-polyposis colorectal cancer (HNPCC) account for about : 27 43 The British Council 2002

Advances in colorectal cancer Table 1 Clinical syndromes with an increased risk of colorectal cancer and their genetic basis Condition Clinical features Genes responsible Familial adenomatous Hundreds to thousands of adenomatous APC polyposis (FAP) polyps in colon and rectum Hereditary non-polyposis colorectal Family history of CRC and other cancers hmlh1, hmsh2, hmsh6, hmsh3, hpms2 cancer (HNPCC) Peutz Jegher s syndrome Hamartomatous polyps in small bowel LKB1, STK11 and peri-oral pigmentation Juvenile polyposis Hamartomatous polyps in stomach and large bowel SMAD4, PTEN, BMPR1A Hereditary gastric cancer Family history of gastric carcinoma CDH1 and CRC Familial adenomatous polyposis 5% of all colorectal cancers (CRCs). Hereditary factors may then contribute to an estimated further 20% of CRC. Inherited susceptibility should be suspected when CRC is detected in a young patient, when a patient presents with synchronous or metachronous CRCs or adenomatous polyps, or where there is a strong family history of CRC. Even without a family history of CRC, 1 in 21 of the population will develop the disease, usually in later life. Table 1 outlines the inherited conditions which confer increased risk of developing CRC. Familial adenomatous polyposis (FAP) was first clearly described as a dominantly inherited Mendelian trait by Lockhart-Mummery in 1925. The clinical diagnosis of FAP depends upon the detection of hundreds to thousands of adenomatous polyps in the colon and rectum that usually appear during adolescence or the third decade of life. CRC is the inevitable consequence of FAP if the condition is left untreated. The risk of cancer is generally related to polyp number 1. Attenuated FAP is a phenotypic variant of FAP characterised by the presence of less than 100 adenomatous polyps 2. Annual colonoscopy is indicated in individuals at risk of FAP from adolescence, followed by prophylactic colectomy or proctocolectomy to eliminate the risk of subsequent development of CRC only in those proven to have FAP. There is marked variation in the associated phenotypic features of FAP, both between carriers of different mutations and, to a lesser extent, within family members with the same mutation 3,4. Congenital hypertrophy of the retinal pigment epithelium (CHRPE) occurs in about 28

Hereditary aspects and overview of colorectal tumorigenesis 60% of FAP kindreds 5. CHRPE has no impact on sight, nor any malignant potential. It can, however, be detected by ophthalmoscopy from birth, thus helping to identify individuals at risk of FAP from an early age in those kindreds where FAP includes the presence of CHRPE 6. Other manifestations of FAP include upper gastrointestinal tumours, desmoid tumours, papillary carcinoma of the thyroid, medulloblastoma, hepatoblastoma and soft tissue tumours 7. Peri-ampullary carcinoma is the commonest cause of death in FAP patients who have undergone prophylactic colectomy 8,9, making endoscopic screening of the upper gastrointestinal tract mandatory in FAP patients. The incidence of FAP is approximately 1 per 8000 of the population 10, accounting for about 0.5% of all CRCs. The genetic basis for FAP lies in germline (inherited) mutation of the adenomatous polyposis coli (APC) gene (OMIM 175100). Approximately one-quarter of all cases are caused by new mutations that maintain the incidence of FAP 10, despite the strong selective disadvantage of the disease. The first clue to localising the position of the APC gene came from identification of a patient with colorectal polyposis and mental retardation who had a deletion of the chromosomal band 5q21 11. Linkage analysis of families with FAP led to the mapping of the APC gene to 5q21 in 1987 12. The APC gene was then cloned, identified and characterized in 1991 13,14. It consists of 8535 base pairs that encode a 2843 amino acid multidomain protein. Exon 15 comprises more than 75% of the coding sequence of APC and is the most common target for both germline and somatic mutations 15. Germline mutations in the APC gene have been demonstrated in most FAP patients 16. The vast majority (95%) of APC mutations are nonsense or frameshift mutations that result in a truncated protein product with abnormal function. As expected from Knudsen s two-hit hypothesis, colorectal tumours from FAP patients nearly all harbour either additional somatic APC mutations or loss of heterozygosity at the APC locus in addition to the original germline mutation 17. The type of germline APC mutation in FAP appears to determine the nature of the second somatic hit to APC. If the germline mutation occurs between codons 1194 and 1392, then there is strong selection for allelic loss of APC as the second hit in the development of a colorectal adenoma. If the germline mutation lies outside this region, the second hit in tumorigenesis is most likely to produce a truncating mutation in the somatic mutation cluster region (MCR) between codons 1286 and 1513 17. Up to one-third of all FAP-causing mutations occur at APC codons 1061 and especially 1309 15, although this figure is probably artificially high due to reporting bias. Apart from these two peaks (due to a relatively high mutation rate at these two positions), germline mutations in APC are spread fairly evenly between codons 200 and 1600, with 29

Advances in colorectal cancer Other polyposis syndromes mutations occurring only rarely beyond codon 1600. The risk of developing specific manifestations of FAP is often correlated to the position of the inherited APC mutation. A particularly severe phenotype is seen in patients with mutations between codons 1250 and 1464, and especially at codon 1309 18. Attenuated polyposis, in contrast, is usually attributed to mutations at the extreme 5 - or 3 -ends of the APC gene 2, or in the alternatively spliced region of exon 9 19. The presence of CHRPE correlates with APC mutations between codons 457 and 1444, while desmoid tumours are mostly limited to patients with mutations between codons 1403 and 1578 7. It is highly likely that this correlation between genotype and phenotype will become increasingly important in the future as a means of targeting genetic testing in FAP to the most likely regions, rather than screening the whole gene. Genetic tests for FAP require DNA, usually extracted from leucocytes found in blood although any source of DNA can be used. The method of mutation detection varies, with direct sequencing, mutation screening with either single strand conformational polymorphism (SSCP) or denaturing gradient gel electrophoresis (DGGE), and the protein truncation test (PTT) all in current use. The sensitivity of the PTT in FAP relies on the fact that the majority of disease-causing mutations result in the formation of a truncated APC protein product. With rigorous testing and the use of multiple screening methods, the mutation detection rate in classical FAP may be as high as 87% 20. Identification of a disease-causing mutation in the APC gene in affected and at-risk family members allows the implementation of appropriate screening programmes. Annual colonoscopy is indicated from adolescence, along with periodic upper gastrointestinal endoscopy in view of the significant risk of gastroduodenal tumours. The timing and nature of prophylactic surgery will also need discussion. Recent evidence suggests that celecoxib, a cyclo-oxygenase-2 (COX-2) inhibitor may reduce the number of colorectal polyps in FAP 21, suggesting that it may be a useful adjunct prior to definitive surgery and help in suppressing rectal polyps in patients with ileorectal anastomosis. Peutz Jegher s syndrome (OMIM 175200) is an autosomal dominant condition characterised by the presence of hamartomatous polyps in the small bowel, together with pigmentation of the peri-oral region, hands and feet. Small bowel intussusception and gastrointestinal haemorrhage are the commonest form of presentation. There is a significant risk of malignant transformation from the small bowel polyps. The syndrome is caused by germline mutations in the STK11, LKB1 and possibly other genes. 30

Hereditary aspects and overview of colorectal tumorigenesis Familial juvenile polyposis (OMIM 174900) is also an autosomal dominant disease in which affected individuals are predisposed to hamartomatous polyps, with the stomach and colorectum most commonly affected. Gastrointestinal cancer affects variable numbers (from 9% to 68%) of juvenile polyposis patients. Germline mutations responsible for juvenile polyposis have so far been identified in the PTEN, SMAD4 (DPC4) or BMPR1A genes. Mutations in PTEN also account for Cowden s syndrome, which carries an increased risk of early-onset breast cancer, thyroid cancer, ovarian cancer, and hamartomatous lesions of the skin. The hereditary mixed polyposis syndrome (OMIM 601228) is characterised by a spectrum of tumours confined to the colorectum. Juvenile polyps, serrated adenomas (hyperplastic polyp with dysplasia), adenomatous polyps and colorectal cancers have all been described. Hereditary non-polyposis colorectal cancer (HNPCC) HNPCC is caused by mutations in one of the mismatch repair genes, most commonly hmlh1, hmsh2 and hmsh6 (OMIM 120436, 120435 and 600678). The condition is inherited as an autosomal dominant with an 80% life-time risk of CRC, and 50 60% life-time risk of endometrial cancer in women 22. Carriers are at particular risk of synchronous and metachronous CRC, and are at significant risk of endometrial, ovarian, upper gastrointestinal, urological, sebaceous and central nervous system cancers 23. There is a preponderance of proximal colonic cancers in HNPCC, making colonoscopy mandatory as the clinical screening tool of choice. The identification of HNPCC kindreds has been facilitated by the Amsterdam criteria defined by the International Collaborative Group on HNPCC in 1991 and refined in 1999 24. The Amsterdam criteria state that CRC should be present in at least three relatives where one is a first-degree relative of the others, that CRC should occur in at least two generations of a family, and that at least one CRC should have been diagnosed before the age of 50 years. FAP must also be formally excluded within the kindred. The revised Amsterdam criteria acknowledge the importance of extracolonic HNPCC-associated tumours 24. These criteria were intended principally as a research, rather than clinical, tool. Mismatch repair gene mutations have been identified in 50 60% of families that fulfil the Amsterdam criteria. Mutations have, however, also been found in families who do not fulfil the criteria, particularly those with a preponderance of HNPCC-related extracolonic tumours. As with FAP, there appears to be a correlation between the particular genetic mutation and the phenotypic appearance of the disease. For example, 31

Advances in colorectal cancer extracolonic tumours are more commonly found in patients with mutations in hmsh2 as compared to hmlh1. Similarly, women with hmsh6 mutations appear to be more likely to develop endometrial cancer 25. Individuals at risk in HNPCC kindreds are heterozygous for mutations in the mismatch repair genes and so their normal cells do not have an elevated mutation rate 26. Loss of DNA mismatch repair function in HNPCC, therefore, requires both the germline mutation and a somatic hit, so that the cell loses its ability to correct errors made during DNA replication. The most vulnerable areas to loss of mismatch repair mechanisms are poly-oligo tracts and base pair repeats known as microsatellites. Disruption of these sequences is seen in over 90% of CRCs arising in HNPCC patients 27, a phenomenon known as replication error (RER + ) or microsatellite instability (MSI). MSI is also seen in about 15% of sporadic CRCs 28. Genetic testing in HNPCC is complicated by the number of genes potentially involved. Nonsense and missense mutations form a large proportion of the disease-causing mutations 29, making PTT unreliable. In addition to this, a significant proportion of hmsh2 disease-causing mutations consist of very large deletions that are not detected by standard methods of mutation detection. Current methods of germline mutation detection in HNPCC utilise direct sequencing, SSCP, DGGE, PTT and dhplc (high performance liquid chromatography). Sensitivity for all methods is low at about 50%. Commercial tests are now available for hmlh1 and hmsh2. An alternative approach to testing for HNPCC by searching for germline mutations is to test CRCs for MSI if there is some evidence for familial involvement or an early age of onset. As over 90% of HNPCC tumours have a marked degree of MSI (known as MSI-H) when screened with a panel of microsatellite markers, detection of a sporadic MSI-H tumour suggests that the patient should be tested for a germline defect in the mismatch repair genes if there is any indication of cancer in the family. The Bethesda criteria (see Table 2) provide guidelines for MSI Table 2 The Bethesda guidelines which provide recommendations for microsatellite instability testing in colorectal cancer * Individuals with a family history fulfilling the Amsterdam criteria * In patients with two or more HNPCC-related cancers, including synchronous and metachronous CRC * Where a patient with CRC has a first-degree relative with CRC or an HNPCC-related cancer diagnosed under the age of 45 years, or with an adenoma diagnosed under 40 years Individuals with CRC or endometrial cancer diagnosed under the age of 45 years Individuals with undifferentiated or right-sided CRC diagnosed under the age of 45 years Individuals with signet-ring type CRC diagnosed under the age of 45 years Individuals with colorectal adenomas diagnosed under the age of 40 years 32

Hereditary aspects and overview of colorectal tumorigenesis testing in CRC in order to identify HNPCC patients. It should be remembered that up to 15% of sporadic CRC are also MSI positive. Immunohistochemical staining for the hmlh1 and hmsh2 proteins may also provide an estimate of the presence or absence of mismatch repair function in a tumour. Current recommendations suggest that screening for HNPCC should begin from the age of 25 years, or at 5 years younger than the youngest affected relative. Colonoscopy is then performed at one to two yearly intervals. The International Collaborative Group on HNPCC also recommends gastroduodenoscopy, renal tract ultrasound, urine cytology, gynaecological examination, endometrial curettage, transvaginal ultrasound and serum CA-125 measurements on an annual or biennial basis 24. A controlled trial has established that 3-yearly colonoscopy and polypectomy are of benefit in HNPCC both in reducing CRC rates and mortality 30. The efficacy of aspirin in prevention of CRC in HNPCC is currently under assessment by the Concerted Action Polyp Prevention (CAPP) trial 31. Prophylactic colectomy with ileorectal anastomosis and subsequent rectal surveillance is offered to HNPCC carriers in some centres. Moderate inherited risk of colorectal cancer Apart from the classical autosomal dominant diseases, the genetic mechanisms responsible for inherited predisposition to CRC are likely to include less deleterious mutations in the same genes responsible for FAP and HNPCC. The APC variant I1307K in Ashkenazi Jews and the rarer E1317Q in Caucasians 32 have been described in patients with multiple adenomas or a carcinoma developing at a young age. These mutations result in APC proteins with amino acid substitutions in functionally critical areas, thus apparently conferring an advantage with respect to tumour function. Incomplete penetrance of such mutations result in increased risk, of perhaps 20 50%, of developing CRC, as distinct to the 70 100% risk associated with the clearly inherited syndrome. Analogous variants in other candidate genes involved in the progression of the adenoma-carcinoma sequence may also carry an increased risk of developing colorectal tumours. The Harvey ras-1 variable number tandem repeat (HRAS1-VNTR) polymorphism and methylenetetrahydrofolate reductase (MTHFR) valine/valine polymorphism have both been found to be associated with an increased risk of colorectal tumorigenesis on meta-analysis 33. Missense mutations in the CDH1 (E-cadherin) gene have been suggested to be associated with an increased risk of colorectal and gastric cancer in the Korean population 34. Genetic variability probably contributes substantially to multifactorial disease inheritance in this way. 33

Advances in colorectal cancer Genetic testing for inherited syndromes of colorectal cancer The process of genetic testing in individuals considered to be at risk for a clearly inherited susceptibility to CRC includes identification of affected and at-risk individuals, genetic counselling, laboratory testing and accurate interpretation of results. Genetic testing may be carried out in an individual suspected of having a hereditary form of colorectal cancer, or in relatives of a known genetic carrier. If at all possible, the proband (original patient presenting with the disease) should be genotyped first. This has the advantage of identifying the responsible mutation and increasing the sensitivity of its subsequent detection in at-risk relatives. This may not always be feasible where a proband has died or refused genetic testing. Corroborating genetic test results with affected individuals from the same family should then confirm the pathogenicity of any mutation found. Testing for any genetic disorder is inappropriate in the absence of trained genetic counselling and good clinical follow-up. Counselling sessions may focus either on the individual or on a family group. The counsellor s multiple roles include ascertaining precise details of an individual s personal and family history, making an assessment of genetic risk, description of genetic testing, and written informed consent. It is also important to: (i) provide appropriate information about the genetic basis of inherited colorectal cancer; (ii) explain the consequences of genetic testing in the event of positive, negative or ambiguous results; and (iii) explain the actual results of any genetic tests performed on family members. Support should also be provided in the follow-up period with emphasis placed on compliance with recommended screening programmes. If a mutation is detected in the proband, it is then possible to test family members for the same mutation. Family members with negative results may be re-assured that they do not need to undergo conventional screening, but it should be stressed that they still have the background population risk for CRC. Family members with the mutation should start or continue with screening programmes, and the option of prophylactic surgery should be discussed. There will always be cases where results of genetic testing in the proband are ambiguous or negative. In these cases, it is inappropriate to perform genetic analysis on other family members. Linkage analysis may provide some useful information in large families. If no mutation is identified, all at-risk family members should continue to be clinically screened. The same situation arises if the proband declines genetic testing. 34

Hereditary aspects and overview of colorectal tumorigenesis Genetic screening for colorectal cancer or multiple adenomas Genetic screening generally refers to the testing of members of a defined subset of the population (e.g. above a certain age or from a particular ethnic group) for specific genetic changes that give rise to particular disease susceptibility in the absence of any clinical symptoms. This should only be done when there is clear benefit that can be provided for those individuals shown to carry the genetic susceptibility factor. For example, essentially all births in the UK are tested for the presence of mutations causing phenylketonuria (PKU) since the disease onset (severe mental retardation) can then be prevented by a diet deficient in the amino acid phenylalanine. Similarly, individuals carrying, for example, the APC I1307K or E1317Q mutations (which predispose to multiple adenomas) could undergo colonoscopic screening regularly at appropriate intervals after a certain age, and any adenomas removed prophylactically. This strategy should completely eliminate any risk of these individuals developing CRC. As the technology for large-scale screening for mutations develops, there may eventually be a strong case for some form of genetic screening for the variety of mutations which predispose to multiple adenomas, and so to CRC. This may already be justified with respect to screening for the APC I1307K mutation in individuals of Ashkenazi Jewish descent, amongst whom the mutation occurs with a frequency of up to 10%. Overview of colorectal tumorigenesis All cells exist under strict regulation of signals for growth, apoptosis, differentiation, cell cell interactions and cell extracellular matrix interactions. Tumorigenesis is a process involving multiple sequential genetic events resulting in activation of oncogenes and inactivation of tumour suppressor genes, which allow escape from the tight constraints that control normal cells. The nature of these events includes mutation, loss of heterozygosity (LOH), epigenetic silencing of gene transcription by promoter hypermethylation, gene amplification, gain of function mutations, transcriptional up-regulation and translocations that generate chimeric proteins with oncogenic activity. Somatic mutations occur randomly. A mutation that provides a cell with a relative survival advantage at the time of occurrence will be selected for and result in clonal expansion of the cell in which it arises. The sequence of acquired mutations, or genetic pathway, in any developing tumour is a reflection of the constraints that become rate limiting at different stages in the evolution of the tumour. For example, whilst early mutations conferring escape from growth control or adhesion may be initiating factors, they may also promote apoptosis. 35

Advances in colorectal cancer Expansion of the developing clone relies on the rate of cell proliferation outmatching the rate of cell death. Thus, in this situation, apoptosis becomes rate limiting and the greatest selective pressure for subsequent mutations would be for those conferring escape from apoptosis. Similarly, larger increases in tumour bulk without adequate vascularisation would place evolutionary pressure for mutations stimulating angiogenesis. Cells from a particular tissue have to escape the same micro-environmental restrictions in order to form a cancer, and would be expected to accumulate mutations along a similar genetic pathway. The presence of a readily evident precursor lesion in colorectal cancer, the adenomatous polyp, has facilitated the characterization of the histopathological and genetic pathways during the transition from adenoma to carcinoma. The genetic basis for the adenoma-carcinoma sequence The histological progression of colorectal cancer from adenoma to carcinoma was first described by Morson and colleagues 35 and the genetic pathway which parallels this transition has been suggested by Fearon, Vogelstein, Bodmer and others (Fig. 1) 36,37. Histological evidence of dysplasia is usually taken as the first recognizable step in the adenoma-carcinoma sequence. Small areas of Fig. 1 Basic outline of the adenoma to carcinoma sequence. The temporal order in which key genes may be affected is shown above the histological stages of disease during which they are thought to occur. Broken lines are used where the order of accumulation of genetic events is uncertain. Functional pathways affected are indicated at the bottom of the diagram. 36

Hereditary aspects and overview of colorectal tumorigenesis epithelium with irregular glandular architecture, termed aberrant crypt foci (ACF), have been reported to harbour mutations in APC, K-ras and TP53 38. However, the relationship between aberrant crypt foci and adenomas is still not clear. Progression of an adenoma is more likely with increased size, severe features of dysplasia villous rather than tubular architecture. Finally, overt features of carcinoma are characterized by local invasion and eventual metastasis. The earliest genetic change associated with adenomatous polyps is most frequently mutation and/or loss of the APC gene. The exact sequence of commonly acquired genetic changes accumulated subsequent to inactivation of APC is variable. K-ras mutations are found in ~50% of colorectal cancers and are thought to be relatively early events that correlate histologically with early to late adenomas. There is good evidence to suggest that p53 mutations occur more frequently in high-grade dysplastic polyps and are thought to mark the transition from adenoma to carcinoma 36,39. Disruption of the TGF-βIIR/SMAD4 pathway and mutations in mismatch repair genes (e.g. hmlh1, hmsh2) and cyclin-dependent kinase inhibitors (e.g. CDKN2A) have all been identified as key factors in the development and progression of CRC. However, the temporal order of disruption of these genetic pathways as they relate to histological progression remains uncertain, and is most probably often different in different tumours. Each tumour is, after all, the result of an independent somatic evolutionary process involving a series of genetic or epigenetic changes each of which gives the tumour a further growth advantage. The role of APC in colorectal carcinogenesis The APC protein has many well-characterized functional domains and interacts with numerous other proteins. It is involved in a wide variety of cellular processes including migration, adhesion, proliferation, and even perhaps aspects of chromosome stability and cytoskeletal organization. Regulation of cytoplasmic β-catenin levels requires appropriate complexing of APC, β-catenin, glycogen synthase kinase 3β (GSK3β) and axin 40. GSK3β is then able to phosphorylate β-catenin on specific serine and threonine residues, thus targeting it for ubiquitin-mediated degradation. In the absence of this regulation, β-catenin escapes degradation and translocates to the nucleus where it complexes with one of the TCF/LEF transcription factors and initiates transcription of a wide variety of genes 41. The downstream transcriptional activation targets of β-catenin include a number of genes involved in the development and progression of colorectal carcinoma, including cyclin D1 42 and the oncogene c-myc 43. Consistent with its definition as a tumour suppressor 37

Advances in colorectal cancer Mutation rate and genomic instability gene, bi-allelic disruption of the APC gene occurs in both FAP and sporadic CRCs. Approximately 95% of the mutations are nonsense or frameshift mutations in the form of small deletions/insertions or point mutations that result in truncation of the APC protein. The highest frequency of point mutations in both familial and sporadic disease arises from transitions at CpG dinucleotides. Spontaneous de-amination of methylated cytosine to thymidine at CGA arginine sequences generates a TGA or stop codon, and so the most frequent nonsense mutation is from an arginine to a stop codon. In the case of somatic mutations, approximately 60% of these truncating mutations fall within the mutation cluster region (MCR) between codons 1286 and 1513 44. There appears to be an intriguing interdependence between the first and second hits to the APC gene in both FAP and sporadic cancers. Where the first mutation (either germline in the case of FAP or somatic in sporadic tumours) is a truncating mutation falling within the MCR, the second somatic hit is almost invariably a deletion or loss of heterozygosity (LOH). Conversely, when the first hit falls outside the MCR, the second hit usually falls within the MCR 17,45. Homozygous deletions of APC are almost never found. Thus it appears there is a strong selection for one truncating mutation within the MCR of APC. Together with the tight clustering of the MCR, this strong selection for truncating mutations implies that mutations truncated specifically in this region have a strong dominant-negative effect and so may confer oncogenic activity to the truncated protein. The region immediately 3 to the MCR contains three of seven 20 amino acid repeat domains essential for β-catenin binding as well as three SAMP repeat motifs at axin-binding sites. The truncated APC product present in the vast majority of colorectal cancers lack SAMP domains necessary for complexing with axin. This may explain why the truncated APC protein acts in a dominant negative fashion namely by disrupting the complex that ultimately targets β-catenin for destruction. The evidence suggests that one of the main selective advantages of truncated APC protein is loss of regulation of β-catenin turnover. This hypothesis is further supported by the observation that gain-of-function mutations in exon 3 of the β-catenin gene CTNNB1 appear to be found predominantly in those CRCs that do not have APC mutations 46,47. Approximately 90% of all sporadic CRCs have constitutive activation of the β-catenin mediated transcriptional pathway, consequent on either APC or CTNNB1 (β-catenin) mutations. There has been much debate regarding the requirement for an increased mutation rate or genomic instability in the generation of cancers to 38

Hereditary aspects and overview of colorectal tumorigenesis account for the large number of genetic alterations present in most tumours. Two forms of genetic instability occur in colorectal cancer. Approximately 85% of tumours are characterised by global chromosome instability resulting in aneuploidy and frequent LOH 48,49. The remaining 15% of CRC harbour bi-allelic inactivation of one of the mismatch repair genes 50,51. As in HNPCC, these replication error positive (RER + ) tumours typically retain a near diploid karyotype but exhibit features of deficient DNA mismatch repair in the form of microsatellite instability (MSI) and an increased mutation rate. It was initially assumed that such an increased tolerance for DNA damage would allow accelerated tumour development and that this would be the main reason for selection of a mismatch repair mutation in RER + tumours or mutations promoting chromosomal instability in replication error negative (RER ) tumours. However, such mutations would not confer any relative survival advantage to the cell in which they first occur. Intriguingly, recent evidence suggests that some genes implicated in genetic instability also play key roles in the regulation of apoptosis. For example, mutations in both the mismatch repair genes hmlh1 and hmsh2 52 and in TP53 53,54 have been shown to confer resistance to DNA damage-induced apoptotic signals. It is, therefore, likely that the initial selection for such mutations derives from the need to escape apoptosis. The attendant increase in mutation rate may then be a by-product of the primary selection against apoptosis and only confer a long-term selective advantage 55,56. Common and divergent pathways in colorectal tumorigenesis Although RER and RER + tumours share early initiating events (involving, in particular, APC or β-catenin mutations), tumours with loss of mismatch repair probably develop along a different genetic pathway to sporadic RER tumours. The same environmental constraints exist for both pathways and it is likely that although different sets of genes may be targeted by mutation, there will be overlap of the signalling pathways affected. For example, inactivation of transcription growth factor-β (TGF-β) signalling may occur via mutation within the poly-a tract in the TGF-β receptor II (TGF-βRII) in RER + tumours whilst mutation of the downstream signalling molecule SMAD4 may be the most efficient means to inactivate this pathway in RER tumours 57. Similarly, TP53 mutations are apparently less common in RER + tumours. However, a p53-induced effector of apoptosis, Bax, may be a common target in RER + tumours and so may provide an alternative to TP53 mutations in RER tumours. 39

Advances in colorectal cancer Ulcerative colitis associated colorectal carcinoma Colorectal cancer may arise through a third genetically and clinically distinct pathway against a background of ulcerative colitis. Such tumours are termed ulcerative colitis associated colorectal carcinomas (UCACRCs). The precursor lesion for UCACRCs is thought to be the flat dysplasia-associated lesion or mass (DALM). Genetically, UCACRCs appear to be characterised by early TP53 mutations, together with much lower frequencies of APC and K-ras mutations than in sporadic CRCs 58. Epigenetic events in colorectal carcinogenesis In addition to deletions and inactivating mutations, epigenetic events have now been recognised as an important mechanism of gene silencing. Specifically, hypermethylation of CpG islands in the promoter region of a gene often impairs the ability of transactivating factors to bind and initiate gene transcription. The importance of this mechanism in colorectal cancer has been highlighted by reports of homozygous promoter hypermethylation of the hmlh1 gene in about 80% of sporadic RER + tumours 59,60. In addition, promotor hypermethylation has been shown to play a part in the silencing of CDKN2A (p16) in some colorectal cancer cell lines and patients 61. Key points for clinical practice Up to a quarter of all colorectal cancers may occur due to some form of inherited susceptibility to colorectal cancer, either in the form of a recognised clinical syndrome or due to germline variants which carry an increased risk of colorectal cancer. Familial adenomatous polyposis arises as a result of germline mutation in the APC gene, and is characterised clinically by the presence of hundreds to thousands of adenomatous polyps in the large bowel. Hereditary non-polyposis colorectal cancer should be suspected when colorectal cancer is detected in a young patient, when a patient presents with synchronous or metachronous colorectal cancers or where there is a strong family history of colorectal cancer. Hereditary non-polyposis colorectal cancer arises as a result of germline mutation in one of the mismatch repair genes, and results in microsatellite instability in the tumour. 40

Hereditary aspects and overview of colorectal tumorigenesis References Colonoscopic screening programmes should be tailored to the individual s genetic risk of developing colorectal cancer. Population-wide genetic screening for individuals at risk of colorectal cancer may be technically possible in the future. Colorectal tumorigenesis consists of a step-wise accumulation of advantageous genetic events which allow the cell to escape constraints on its growth. The genetic events in colorectal tumorigenesis mirror the pathological sequence of transition from adenoma (precursor lesion) to colorectal cancer. Selection occurs in favour of mutations which confer a survival advantage on the cell at the time the mutation happens. Somatic mutation in the APC gene occurs in the majority of sporadic colorectal cancers. 1 Debinski HS, Love S, Spigelman AD, Phillips K. Colorectal polyp counts and cancer risk in familial adenomatous polyposis. Gastroenterology 1996; 110: 1028 30 2 Spirio L, Olschwang S, Groden J et al. Alleles of the APC gene: an attenuated form of familial polyposis. Cell 1993; 75: 951 7 3 Giardiello F M, Krush AJ, Petersen GM et al. Phenotypic variability of familial adenomatous polyposis in 11 unrelated families with identical APC gene mutation. Gastroenterology 1994; 106: 1542 7 4 Nugent KP, Phillips RK, Hodgson SV et al. Phenotypic expression in familial adenomatous polyposis: partial prediction by mutation analysis. Gut 1994; 35: 1622 3 5 Blair NP, Trempe CL. Hypertrophy of the retinal pigment epithelium associated with Gardner s syndrome. Am J Ophthalmol 1980; 90: 661 7 6 Diaz-Llopis M, Menezo JL. Congenital hypertrophy of the retinal pigment epithelium in familial adenomatous polyposis. Arch Ophthalmol 1988; 106: 412 3 7 Fearnhead NS, Britton MP, Bodmer WF. The ABC of APC. Hum Mol Genet 2001; 10: 721 33 8 Jagelman DG, DeCosse JJ, Bussey HJ. Upper gastrointestinal cancer in familial adenomatous polyposis. Lancet 1988; 1: 1149 51 9 Offerhaus GJ, Giardiello FM, Krush AJ et al. The risk of upper gastrointestinal cancer in familial adenomatous polyposis [see comments]. Gastroenterology 1992; 102: 1980 2 10 Bisgaard ML, Fenger K, Bulow S, Niebuhr E, Mohr J. Familial adenomatous polyposis (FAP): frequency, penetrance, and mutation rate. Hum Mutat 1994; 3: 121 5 11 Herrera L, Kakati S, Gibas L, Pietrzak E, Sandberg AA. Gardner syndrome in a man with an interstitial deletion of 5q. Am J Med Genet 1986; 25: 473 6 12 Bodmer WF, Bailey CJ, Bodmer J et al. Localization of the gene for familial adenomatous polyposis on chromosome 5. Nature 1987; 328: 614 6 13 Groden J, Thliveris A, Samowitz W et al. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 1991; 66: 589 600 14 Kinzler KW, Nilbert MC, Su LK et al. Identification of FAP locus genes from chromosome 5q21. Science 1991; 253: 661 5 15 Beroud C, Soussi T. APC gene: database of germline and somatic mutations in human tumors and cell lines. Nucleic Acids Res 1996; 24: 121 4 41

Advances in colorectal cancer 16 Cottrell S, Bicknell D, Kaklamanis L, Bodmer WF. Molecular analysis of APC mutations in familial adenomatous polyposis and sporadic colon carcinomas [see comments]. Lancet 1992; 340: 626 30 17 Lamlum H, Ilyas M, Rowan A et al. The type of somatic mutation at APC in familial adenomatous polyposis is determined by the site of the germline mutation: a new facet to Knudson s two-hit hypothesis. Nat Med 1999; 5: 1071 5 18 Nagase H, Miyoshi Y, Horii A et al. Correlation between the location of germ-line mutations in the APC gene and the number of colorectal polyps in familial adenomatous polyposis patients. Cancer Res 1992; 52: 4055 7 19 Young J, Simms LA, Tarish J et al. A family with attenuated familial adenomatous polyposis due to a mutation in the alternatively spliced region of APC exon 9. Hum Mutat 1998; 11: 450 5 20 O Sullivan MJ, McCarthy TV et al. Familial adenomatous polyposis: from benchside to bedside. Am J Clin Pathol 1998; 109:521 6 21 Steinbach G, Lynch PM, Phillips RK et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000; 342: 1946 52 22 Aarnio M, Sankila R, Pukkala E et al. Cancer risk in mutation carriers of DNA mismatch repair genes. Int J Cancer 1999; 81: 214 8 23 Watson P, Lynch HT. Extracolonic cancer in hereditary non-polyposis colorectal cancer. Cancer 1993; 71: 677 85 24 Vasen HF, Watson P, Mecklin JP, Lynch HT et al. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative Group on HNPCC. Gastroenterology 1999; 116: 1453 6 25 Charames GS, Millar AL, Pal T, Narod S, Bapat B. Do MSH6 mutations contribute to double primary cancers of the colorectum and endometrium? Hum Genet 2000; 107: 623 9 26 Tomlinson IP, Hampson R, Karran P, Bodmer WF. DNA mismatch repair in lymphoblastoid cells from hereditary non-polyposis colorectal cancer (HNPCC) patients is normal under conditions of rapid cell division and increased mutational load. Mutat Res 1997; 383: 177 82 27 Borresen AL, Lothe RA, Meling GI et al. Somatic mutations in the hmsh2 gene in microsatellite unstable colorectal carcinomas. Hum Mol Genet 1995; 4: 2065 72 28 Liu B, Nicolaides NC, Markowitz A et al. Mismatch repair gene defects in sporadic colorectal cancers with microsatellite instability. Nat Genet 1995; 9: 48 55 29 Peltomaki P, Vasen HF. Mutations predisposing to hereditary nonpolyposis colorectal cancer: database and results of a collaborative study. The International Collaborative Group on Hereditary Nonpolyposis Colorectal Cancer. Gastroenterology 1997; 113: 1146 58 30 Jarvinen HJ, Aarnio M, Mustonen H et al. Controlled 15-year trial on screening for colorectal cancer in families with hereditary non-polyposis colorectal cancer. Gastroenterology 2000; 118: 829 34 31 Burn J, Chapman P, Bishop DT, Mathers J. Diet and cancer preventions: the Concerted Action Polyp Prevention (CAPP) studies. Proc Nutr Soc 1998; 57: 183 6 32 Frayling IM, Beck NE, Ilyas M et al. The APC variants I1307K and E1317Q are associated with colorectal tumors, but not always with a family history. Proc Natl Acad Sci USA 1998; 95: 10722 7 33 Houlston RS, Tomlinson IPM. Polymorphisms and colorectal tumor risk. Gastroenterology 2001; 121: 282 301 34 Kim HC, Wheeler JMD, Kim JC et al. The E-cadherin gene (CDH1) variants T340A and L599V in gastric and colorectal cancer patients in Korea. Gut 2000; 47: 262 7 35 Muto T, Bussey HJ, Morson BC et al. The evolution of cancer of the colon and rectum. Cancer 1975; 36: 2251 70 36 Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61: 759 67 37 Bodmer WF. The somatic evolution of cancer. The Harveian Oration of 1996. J R Coll Physicians Lond 1996; 31: 82 9 38 Jen J, Powell SM, Papadopoulos N et al. Molecular determinants of dysplasia in colorectal lesions. Cancer Res 1994; 54: 5523 6 39 Rodrigues NR, Rowan A, Smith ME et al. p53 mutations in colorectal cancer. Proc Natl Acad Sci USA 1990; 87: 7555 9 42

Hereditary aspects and overview of colorectal tumorigenesis 40 Behrens J, Jerchow BA, Wurtele M et al. Functional interaction of an axin homolog, conductin, with beta-catenin, APC, and GSK3beta. Science 1998; 280: 596 9 41 Polakis P. The oncogenic activation of beta-catenin. Curr Opin Genet Dev 1999; 9: 15 21 42 Shtutman M, Zhurinsky J, Simcha I et al. The cyclin D1 gene is a target of the betacatenin/lef-1 pathway. Proc Natl Acad Sci USA 1999; 96: 5522 7 43 He TC, Sparks AB, Rago C et al. Identification of c-myc as a target of the APC pathway. Science 1998; 281: 1509 12 44 Miyoshi Y, Nagase H, Ando H et al. Somatic mutations of the APC gene in colorectal tumors: mutation cluster region in the APC gene. Hum Mol Genet 1992; 1: 229 33 45 Rowan AJ, Lamlum H, Ilyas M et al. APC mutations in sporadic colorectal tumors: a mutational hotspot and interdependence of the two hits. Proc Natl Acad Sci USA 2000; 97: 3352 7 46 Ilyas M, Tomlinson IP, Rowan A, Pignatelli M, Bodmer WF. Beta-catenin mutations in cell lines established from human colorectal cancers. Proc Natl Acad Sci USA 1997; 94: 10330 4 47 Sparks AB, Morin PJ, Vogelstein B, Kinzler. Mutational analysis of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer Res 1998; 58: 1130 4 48 Lengauer C, Kinzler KW, Vogelstein B et al. Genetic instabilities in human cancers. Nature 1998; 396: 643 9 49 Cahill DP, Kinzler KW, Vogelstein B, Lengauer C. Genetic instability and Darwinian selection in tumours. Trends Cell Biol 1999; 9: M57 60 50 Perucho M. Cancer of the microsatellite mutator phenotype. Biol Chem 1996; 377: 675 84 51 Markowitz S. DNA repair defects inactivate tumor suppressor genes and induce hereditary and sporadic colon cancers. J Clin Oncol 2000; 18 (Suppl 21): 75S 80S 52 Aebi S, Fink D, Gordon R et al. Resistance to cytotoxic drugs in DNA mismatch repair-deficient cells. Clin Cancer Res 1997; 10: 1763 7 53 el-deiry WS, Harper JW, O Connor et al. WAF1/CIP1 is induced in p53-mediated G1 arrest and apoptosis. Cancer Res 1994; 54: 1169 74 54 Weller M. Predicting response to cancer chemotherapy: the role of p53. Cell Tissue Res 1998; 292: 435 45 55 Tomlinson IP, Bodmer WF. Failure of programmed cell death and differentiation as causes of tumors: some simple mathematical models. Proc Natl Acad Sci USA 1995; 92: 11130 4 56 Tomlinson IP, Novelli MR, Bodmer WF. The mutation rate and cancer. Proc Natl Acad Sci USA 1996; 93: 14800 3 57 Woodford-Richens KL, Rowan AJ, Gorman P et al. SMAD4 mutations in colorectal cancer probably occur before chromosomal instability, but after divergence of the microsatellite instability pathway. Proc Natl Acad Sci USA 2001; 98:9719 23 58 Tarmin L, Yin J, Harpaz N et al. Adenomatous polyposis coli gene mutations in ulcerative colitis-associated dysplasias and cancers versus sporadic colon neoplasms. Cancer Res 1995; 55: 2035 8 59 Herman JG, Umar A, Polyak K et al. Incidence and functional consequences of hmlh1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci USA 1998; 95: 6870 5 60 Wheeler JMD, Beck NE, Kim HC et al. Mechanisms of inactivation of mismatch repair genes in human colorectal cancer cell lines: the predominant role of hmlh1. Proc Natl Acad Sci USA 1999; 96: 10296 301 61 Burri N, Shaw P, Bouzourene H et al. Methylation silencing and mutations of the p14 ARF and p16 INK4a genes in colon cancer. Lab Invest 2001; 81: 217 29 43