Methods and results

CLINICAL DETAILS

The index case presented at 29 years of age with a firm, slightly tender swelling within the right rectus abdominus muscle. A 6 x 5 cm tumour was locally excised and conventional histological examination showed infiltrative fibromatosis. The tumour recurred after six years and was resected again along with 30 cm of adherent small bowel. A year later, a further abdominal wall recurrence was resected and on this occasion fresh tissue was submitted for cytogenetic analysis. Full colonoscopy before referral to the genetics service showed no evidence of colonic adenomas throughout the colon. Repeat colonoscopy after the gene mutation was identified still failed to show any colonic pathology, although contrast dye spray was not undertaken on either occasion. The only relevant family history was that her father had a previous history of a sigmoid colectomy carried out at 56 years of age for a carcinoma of the colon. He had been discharged from follow up after seven years during which endoscopy had shown no further pathology and he has remained well and symptom free since. He had never had any palpable lumps. Review of the histopathology from the resection specimen showed a Dukes B adenocarcinoma of the colon with six adenomas in the surrounding colonic mucosa.

CYTOGENETIC ANALYSIS

Fresh desmoid tumour was subjected to cytogenetic analysis which showed normal fibroblasts mixed with abnormal cells showing an interstitial deletion involving chromosome 5q22.

MOLECULAR ANALYSIS

DNA from the index case was examined for mutations in the APC gene using denaturing high performance liquid chromatography (DHPLC). A single base substitution G>A was identified at nucleotide position 7511, codon 2504, which changes tryptophan (TGG) to a stop codon (TAG). The mutation was present in both the index case and her father.

IMMUNOHISTOCHEMISTRY

The colonic tumour and the desmoid tumour were examined for β-catenin expression using immunohistochemistry. This showed intense nuclear and cytoplasmic β-catenin expression in the colonic tumour. A similar, although less intense, pattern of nuclear expression with some cytoplasmic expression was apparent in the desmoid tumour from the index case.

Discussion

Classical familial adenomatous polyposis coli (FAP) is recognised by the presence of many hundreds of colonic adenomas throughout the colon. Since the identification of the gene responsible for FAP in 1991, it has been possible to recognise certain genotype/phenotype correlations which may not only allow the clinician to direct the molecular geneticist towards the most likely region where a mutation may be located, but conversely may predict the likely natural course of the disease associated with a given mutation and direct management strategies accordingly.7 Where mutations fall into the 5′ and 3′ extremes of the gene, an attenuated colonic phenotype is usually seen with fewer and later onset polyps.8 Whereas mutations at the 5′ end of the gene (exons 1-4) are not associated with desmoid tumour development, mutations arising in the 3′ half of the gene are more often associated with desmoid tumours, and in rare families this is the dominant phenotype with little else to suggest the diagnosis of polyposis on examination of the colon.3 5 Although cytogenetic analysis of desmoid tumours is technically difficult and requires fresh tumour tissue, it may be a useful additional investigation where the phenotype for FAP is absent. Cytogenetic deletion of 5q in desmoid tumours is more frequently associated with FAP than apparently sporadic tumours,9 10 so, even in the absence of typical FAP features, a cytogenetic deletion of 5q in the tumour may direct a careful search for possible underlyingAPC gene mutations.

APC is involved in many critical cellular functions including cell adhesion (via E-cadherin) and in negatively regulating β-catenin levels. β-catenin is involved in the transcriptional activation of several genes involved in cell cycle entry and progression. APC mutation leads to reduced degradation of β-catenin (reflected in overexpression of β-catenin shown using immunohistochemistry in the tumours described in this case).

The natural history of desmoid tumours is variable but they can follow a highly aggressive course. However, even in families where the desmoid phenotype is 100% penetrant,3 5 the clinical course is unpredictable. The explanation for the particular predilection of 3′ mutations to be associated with desmoid tumour formation is still a matter of speculation. Low levels of truncated protein product have been reported in similarly distal mutations, possibly owing to an unstable mRNA; the stability of the mRNA may vary according to tissue type.3 5 There are likely to be both environmental and genetic modifiers exerting an effect on APCgene mutations in view of the variation in phenotype between subjects with the same mutation. It is possible that the effect of these putative genetic modifiers will differ according to tissue type. Detection of such modifiers and determination of their mode of action could be extremely helpful in selecting potential medical treatment for patients with desmoid tumours, since surgery is best avoided if the tumours are not life threatening. A recent paper looking at second hits in adenomas and desmoid tumours in FAP included data only for patients with germline mutations 5′ of codon 1462. In desmoid tumours from these patients, there was either loss of the wild type allele or a somaticAPC mutation. The somatic mutations were consistently 3′ of codon 1426. In our patient, there was loss of heterozygosity indicated by a cytogenetic deletion but a 3′ germline mutation. This suggests that the siting of first and second hits may be interchangeable with respect to the end result (a desmoid tumour), but if somatic mutations in colonic epithelium are in some way reliant on the germline mutation this may explain the very different clinical phenotype presented by many patients with unusually 3′ mutations.11

For families where the desmoid phenotype is the dominant clinical feature and the risk of colorectal cancer is likely to be lower than where adenomatous polyps are frequent, colonic resection should be reserved for situations where there is significant colonic pathology. There may be an argument for favouring subtotal colectomy and ileorectal anastomosis if colectomy becomes necessary. Giving clear information to at risk members of the family reported here is difficult, as the position of the mutation leads us to suspect a high risk for desmoid tumours although the clinical picture is not of 100% desmoid tumour penetrance. Although the colon cancer risk must also be increased, the position of the mutation in the family described, the association of desmoid tumours with trauma (and surgery in particular), the late onset, and the presence of only a few adenomas in the only identified gene carrier with colon cancer and the normal appearance of the colonic mucosa in the index case in her late thirties all suggest that a conservative approach to surgery is appropriate. We have offered genetic testing to at risk family members and suggested that a screening colonoscopy every two years starting at 20 years, with dye spray or random mucosal biopsies to determine whether any microadenomas are present, would be a reasonable course of action for a gene carrier or for a family member at 50% risk who chose not to have a predictive genetic test.

We know this disease is caused by a mutation in theAPC gene and could therefore be called familial adenomatous polyposis (FAP) or perhaps better attenuated adenomatous polyposis coli (AAPC). This acknowledges that there is an increased colon cancer risk; however, this potentially underplays the significance of the desmoid tumour phenotype which may be a very important factor to take into account in making decisions about surgery.