The genetic basis of colonic adenomatous polyposis syndromes

Classic FAP (OMIM #175100) refers to patients who are diagnosed with FAP due to the development of more than 100 adenomatous colorectal polyps from early childhood (typically at age 16) who harbour an APC germline mutation. On average, cancer develops a decade after the appearance of adenomas and if the colon is left untreated most patients develop CRC by 40 years of age [8]. Other gastrointestinal manifestations include fundic gland polyps (which occurs in approximately 90% of FAP patients and are mostly benign [24]), adenomatous polyps in the duodenum and periampullary region (lifetime risk has been reported to reach 100% [25, 26]), and small bowel adenomas [8]. Extra-colonic manifestations are common but rarely malignant and include [8]; desmoid tumours (benign soft-tissue tumours that can be fatal due to progressive invasion into surrounding tissues [5, 27]), cutaneous lesions such as fibromas, lipomas, sebaceous and epidermoid cysts (present in Gardner syndrome [28], a phenotypic variant of FAP), congenital hypertrophy of the retinal pigment epithelium (CHRPE, which is a lesion causing discoloration in the ocular fundus – low-grade adenocarcinoma has been described in these lesions [29]), brain tumours (mainly medulloblastoma, described in Turcot’s syndrome, another phenotypic variant of FAP), hepatoblastoma, dental abnormalities, cancer of the pancreas, thyroid, gallbladder, bile duct and adrenal glands [8, 30‐32].

AFAP is a phenotypic variant of FAP; patients develop less than 100 polyps, delayed polyp growth and later age of cancer onset. Germline APC mutations are also present in these patients, which are mainly observed in three sections of the gene (first 5 exons, exon 9 and in the distal 3′end of APC) [33]. Mean age of polyp diagnosis in AFAP patients is variable but on average in fourth to fifth decade of life, with cancer developing 10–15 years later [8, 34]. Screening is suggested to start late teens to mid-20s [34]. As with FAP the most common extra-colonic manifestations are upper gastrointestinal polyps, duodenal and gastric adenomas and fundic gland polyps [33]. Extra-colonic manifestations in AFAP are rare but hepatoblastoma, gastric and breast adenocarcinoma have been documented [8, 33].

In families with FAP, considerable variability in disease expression is observed within and between families harbouring identical APC mutations [13, 35‐37] and it has been shown that the greater the number of colorectal adenomas, the greater the risk of CRC [38]. It has been demonstrated that there is significant variation with respect to age of onset of intestinal symptoms and the development of CRC, even in patients with the same mutations [13]. Haplotype reconstruction from pedigrees have revealed there is no evidence for a specific APC haplotype associated with disease severity [39]. Genotype-phenotype correlations have been associated with the location of germline mutations within APC that are related to disease severity and the expression of extra-colonic disease [40‐42], see figures in Half et al. [8] and Macrae [10]. Patients with mutations in the mutation cluster region (MCR), located between codons 1286 and 1513 [43], have generally a worse prognosis with earlier disease onset than those with mutations outside this region [44]. Germline mutations at codon 1309 is associated with most severe disease [45], while milder forms with less than 100 adenomas and later ages of onset (AFAP) are associated with codons <157, 312–412 and >1595 [33, 41]. CHRPE has been associated with mutations between codons 457 and 1444 and susceptibility to desmoid tumours is correlated with mutations between codons 1395 and 2000, with slight variability in codon ranges between reports [8, 10, 46]. There is evidence of large phenotypic variation among patients with identical germline mutations [13], strongly suggesting the existence of FAP modifier alleles.

The phenotypic variation of APC ^Min mouse model of FAP reveal among different inbred strains the importance of modifier alleles [47]. There is evidence to suggest that these phenotypic differences are caused by segregating modifier alleles that impact adenoma number [47]. Several have been found in the min mouse model. The best known modifier is possibly Mom1 (modifier of Min 1), which is semi-dominant - each copy affects tumour multiplicity by a factor of approximately 2 [47]. Pla2g2a (found at the same region as Mom1) has also been shown to affect the net growth rate of adjacent tumours [48]. The exact mechanism by which it influences tumorigenesis remains unresolved [47] and the effort of linking PLA2G2A to FAP in humans has failed [49, 50], illustrating the difficult task of searching for modifier alleles in FAP. Many different genetic modifiers of the Apc knockout mouse models have been found, affecting karyotypic stability, DNA mutation rate, recombination rates, differentiation, DNA methylation, stromal regulation, cell growth and proliferation (reviewed in [47, 51]). There are claims that mouse models are essential in identifying modifiers of human disease and by using an Apc (Min/+) model have identified seven genes that are the most likely candidates for the Mom5 modifier [52]. Recently it has been reported that a new Xenopus tumour model might be especially useful for identifying or characterising modifier genes associated with APC mediated tumour formation [53].

Genome wide association studies (GWASs) have identified approximately 40 CRC susceptibility loci, where each loci gives a small increased risk of CRC [54]. The risk associated with each variant is too small on their own for translation to testing in clinical practice but the development of algorithms estimating cumulative risk are expected to lead to clinical application [54]. Two of these SNPs have been associated with Lynch syndrome [55, 56] and recently the same SNPs (rs16892766 and rs3802842) have been associated with adenoma number in APC mutation carriers causing a more sever FAP phenotype [57].

Several studies suggest that low-penetrant susceptibility genes may play an important role in the development of sporadic CRC [58‐63]. There is evidence to show that the variation in FAP severity (which have been shown to be independent of APC mutations and most likely the action of modifier alleles), is expected to result in different rates of adenoma number rather than differences in tumour progression [64]. Modifier genes can influence individual susceptibility to cancer by enhancing or suppressing disease initiation, growth and/or progression. The pattern of intra-familial variation in colonic FAP severity is consistent with the action of modifier genes [39, 64‐66]. As described above there is plenty of evidence from animal models for the existence of FAP modifiers and knowledge of modifier genes will contribute to better prophylactic measures for FAP patients [67]. It is important that the search for modifier genes/alleles continues.

MAP (OMIM #608456) is an autosomal recessive disease caused by biallelic mutations in the base excision repair gene MUTYH. MUTYH is involved in base excision repair and is necessary in the amelioration of reactive oxygen species DNA damage prior to cell division [68]. In a recent study, 23% of APC mutation negative samples (FAP samples screened for APC mutations) were found to harbour pathogenic mutations in MUTYH [69]. Patients usually present with <100 colorectal polyps at an average age of disease diagnosis at around 50 years of age (which is similar to AFAP) and a high risk of CRC [69, 70]. The age of onset of polyposis has been shown to be significantly delayed for biallelic MUTYH carriers compared to APC mutation carriers [69]. MAP has been associated with malignancies of the duodenum, ovaries, urinary bladder and skin—occasionally resembling the phenotype of LS [71]. In a recent report describing extra-colonic disease, biallelic carriers are at high risk of urinary bladder and ovarian cancer, while there is some evidence that monoallelic carriers are at risk of gastric, hepatobiliary, endometrial and breast cancer [72]. No increased risk of other extra-colonic cancers associated with FAP was observed in this study [72].

A recently described autosomal recessive polyposis condition has been named NAP (OMIM #616415). Patients have germline homozygous or compound heterozygous mutations in the base excision repair gene NTHL1 [73]. Due to its recent discovery the clinical manifestation is not set, but it points towards an extended spectrum of cancer diagnosis in these patients; endometrial, duodenal, skin (basal cell carcinoma) and others [74]. Given such disease heterogeneity, Dutch researchers suggest it is a novel cancer syndrome. This is supported by Canadian researchers who also identified biallelic NTHL1 mutations in a woman with multiple primary tumours [75].

PPAP is associated with mono- and biallelic mutations in the genes POLE and POLD1 [76], both genes a part of the mismatch repair (MMR) pathway. PPAP is in an autosomal dominantly inherited CRC predisposition [77]. Variants in POLE and POLD1 are known to increase the somatic mutation rate in tumours [78], thereby increasing the risk of tumour development. The somatic mutation landscape can display great diversity [79], which could be a reason for the differences observed in the location of primary tumours between patients. Both POLE and POLD1 have been associated with an increased risk of endometrial cancer [76, 80]. POLD1 has been associated with breast and brain tumours in addition to CRC and endometrial cancer [81]. Multi-tumour phenotypes such as colon/pancreas/ovaries/small intestine [82] and colon/ovarian/endometrial/brain [80] have been seen in POLE mutation carriers. In addition, POLE has been linked to an early onset cancer case raising the question whether this specific POLE mutation may confer a more severe phenotype than previously reported POLE/POLD1 mutations [83].