If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author: Address: Department of Clinical Genetics, Erasmus MC, University Medical Centre Rotterdam, Office Ee2018, PO Box 2040, 3000CA Rotterdam, The Netherlands. Tel.: +31 10 7036915; fax: +31 10 7043072.
Desmoid tumours are rare mesenchymal tumours with unpredictable progression and high recurrence risk. They can occur sporadically or in association with Familial Adenomatous Polyposis (FAP), which is caused by germline APC mutations. The Wnt/β-catenin pathway has a central role in the pathogenesis of desmoid tumours. These tumours can occur due to either a somatic CTNNB1 or APC mutation but can also be the first manifestation of FAP. Because germline APC analysis is not routinely performed in children with desmoid tumours, the diagnosis FAP may escape detection. The aim of this study is to form guidelines for the identification of possible APC germline mutation carriers among children with desmoid tumours, based on CTNNB1 mutation analysis and immunohistochemical analysis (IHC) for β-catenin.
Patients and methods
We performed IHC of β-catenin and mutation analysis of CTNNB1 and APC in 18 paediatric desmoid tumours, diagnosed between 1990 and 2009 in the Erasmus MC, Rotterdam.
Results
In 11 tumours, IHC showed an abnormal nuclear β-catenin accumulation. In this group a CTNNB1 mutation was detected in seven tumours. In two tumours with an abnormal nuclear β-catenin accumulation and no CTNNB1 mutation, an APC mutation was identified, which appeared to be a germline mutation.
Conclusions
Aberrant staining of β-catenin in paediatric desmoids helps to identify children at risk for FAP. We recommend to screen paediatric desmoid tumours for nuclear localisation of β-catenin and consequently for CTNNB1 mutations. For patients with nuclear β-catenin expression and no CTNNB1 mutations, APC mutation analysis should be offered after genetic counselling.
Desmoid tumours (also named aggressive fibromatosis) are rare mesenchymal tumours which can occur sporadically or in association with Familial Adenomatous Polyposis (FAP). These tumours are characterised by unpredictable progression and high recurrence rate and therefore are difficult to treat. The occurrence of desmoid tumours in children is low, with an estimated incidence of 2–4 new diagnoses per million per year.
Currently, no general guidelines are available for genetic analysis of paediatric patients with desmoid tumours.
The Wnt/β-catenin signalling pathway is recognised as having a central role in the pathogenesis of desmoid tumours. The β-catenin protein is a key effector of the pathway, affecting cellular decisions such as stem cell maintenance and cell proliferation through modulating the expression of specific target genes. In normal cells, β-catenin is involved both in cell adhesion, when located at the cell membrane, and in transcriptional regulation, when present in the nucleus.
Several members of the Wnt/β-catenin signalling pathway form an intracellular multiprotein complex, composed of APC, β-catenin, AXIN1, AXIN2 and GSK3β. APC binds to β-catenin at the so-called 15 and 20 amino acid binding sites.
AXIN1/AXIN2 activates casein kinase I alpha, that catalyses a priming phosphorylation on S45 of β-catenin, thereby providing a signal for glycogen synthase kinase 3 to promote the sequential phosphorylation of T41, S37 and S33, which subsequently induces the degradation of β-catenin thereby preventing its signalling activity.
In several tumour types, this pathway is constitutively activated due to ‘loss of function’ mutations of the APC gene leading to inefficient β-catenin degradation and its intracellular stabilisation. In case the APC gene is intact, pathogenic CTNNB1 mutations, encoding β-catenin, are found at the N-terminal phosphorylation sites interfering with its proteolytic degradation.
Either APC or CTNNB1 mutations can lead to an abnormal intranuclear accumulation of β-catenin. The aberrant β-catenin stabilisation is thought to constitutively activate downstream Wnt/β-catenin target genes and trigger a genetic programme resulting in tumour formation.
Several studies identified somatic CTNNB1 mutations in desmoid tumours in a frequency varying between 52% and 87%,
Detection of beta-catenin mutations in paraffin-embedded sporadic desmoid-type fibromatosis by mutation-specific restriction enzyme digestion (MSRED): an ancillary diagnostic tool.
In Familial Adenomatous Polyposis (FAP), caused by germline APC mutations, 10–15% of the patients are affected by desmoid tumours, representing a more than 800-fold increased risk in comparison with the general population.
FAP is a hereditary predisposition to develop hundreds to thousands of colorectal polyps ultimately leading to colorectal cancer. Untreated, the risk of colorectal cancer in FAP patients is 100%. The average age when colorectal polyps are detected is 15 years.
Colonoscopy, started at the age of 10–12 years, and surgery in adolescence prevent colorectal cancer formation in FAP. When FAP is diagnosed in time, health benefits and increased life expectancy can be achieved.
In children, a desmoid tumour can occur due to either a somatic CTNNB1 or APC mutation, but importantly can also be the first manifestation of FAP. This is of importance for the management of the child, but might also have implications for other (asymptomatic) family members. Currently, no established procedure is available to identify FAP carriers among children with this tumour type. Here, we present an immunohistochemical and mutational analysis on 18 paediatric desmoid tumours, based on which we formulate guidelines to identify possible APC germline mutation carriers.
2. Materials and methods
2.1 Patients
Between January 1990 and June 2009, 20 desmoid tumours were operated in 19 patients under the age of 21 years in the Erasmus MC, University Medical Centre, Rotterdam. Ten tumours were localised in the head and neck region, five in the extremities, three in the abdominal wall and two in the back (Table 1). All tumours were classified as deep fibromatoses. The medical records and family history of these patients were analysed. Tissues were obtained from the initial operative procedure and embedded in paraffin after formalin fixation. All tissues were revised by a pathologist and confirmed as being desmoid tumours. Unfortunately, two samples (D14 and D18) were not suitable for molecular analysis.
Table 1Clinical characteristics of the 19 children with desmoid tumours.
Patient no.
Tumour
Gender
Age at Dx (years)
Tumour localisation
Family history
1
D1
M
2.4
Mandible
2
D2
M
8.2
Spina iliaca posterior superior left
3
D3
F
1.4
Upper extremity
4
D4
F
4
Neck
5
D5
F
5
Mandible
6
D6
F
12.3
Lower extremity
7
D7
M
1.4
Trunk (back, subcutaneous),
Sister with a fibrous hamartoma of infancy
D8
upper extremity
8
D9
M
5.7
Mandible
9
D10
F
14.8
Abdominal wall (intra-abdominal)
10
D11
F
15.3
Abdominal wall
This patient and her father diagnosed with FAP
11
D12
M
0.8
Neck
12
D13
F
0.6
Neck (parotid)
13
D14
M
0, Congenital
Upper extremity
14
D15
F
2
Mandible
15
D16
M
0.8
Trunk (groin)
16
D17
M
1
Trunk (back, subcutaneous)
17
D18
F
6
Neck
18
D19
F
1.3
Earlobe
19
D20
M
1.5
Parotid gland
Father diagnosed with FAP after diagnosis of the desmoid tumour in the patient
All sections were evaluated under a light microscope after Mayer haematoxylin counterstaining.
2.3 DNA isolation
Normal and tumour DNA was extracted from Formalin-Fixed, Paraffin-Embedded (FFPE) tissue fragments using proteinase K and 5% Chelex® 100 resin, as previously described.
For the identification of germline APC mutations, DNA isolated from peripheral blood cells was used. Isolation of this DNA was performed according to standard procedures.
2.4 CTNNB1 mutation analysis
Sequence analysis of CTNNB1 exon 3, encoding the mutational cluster region (MCR) of β-catenin was performed on tumour DNA as previously described.
The APC MCR was amplified in 4 overlapping fragments according to standard procedures.
For the identification of germline APC mutations, DNA isolated from peripheral blood cells was used. All coding exons and intron–exon boundaries of APC were sequenced according to standard procedures. In addition, Multiplex ligation-dependent probe amplification (MLPA) analysis was performed using the SALSA MLPA kit P043 (MRC Holland, Amsterdam, The Netherlands) as previously described.
Loss of heterozygosity (LOH) analysis of the APC locus was performed using DNA isolated from normal and tumour tissues. Three microsatellite markers (D5S433, D5S656, D5S421), mapping to chromosome 5q21.2-5q22.2 were selected. PCR amplification was performed as described above for CTNNB1 mutation analysis with FAM-labelled primers. The FAM-labelled PCR fragments were run on an ABI 3130xl genetic analyser (Applied Biosystems) and data were analysed with Genemarker version 1.8 software (SoftGenetics, State College, PA). Peak heights of the alleles were compared between normal and tumour DNA samples. If one of the alleles showed relative loss in the tumour sample this was considered to be due to LOH (scored manually).
Sequence of all primers is available on request.
3. Results
Eighteen desmoid tumours from 17 patients could be included in the study (Table 1, Table 2). Patient 7 showed two localisations of desmoid tumours (D7 and D8). His sister had a fibrous hamartoma of infancy. In patient 10 there was a prior family history of FAP. The father of patient 19 was known with polyposis. After diagnosis of a desmoid tumour in patient 19 at 1.5 years of age, we could identify a germline APC mutation in the father as well as in the child during subsequent counselling. The youngest age at diagnosis was 0 years (a congenital desmoid tumour) and the oldest patient was 15 years. The latter was the patient from the previously known FAP family. The mean age was 4.4 years. The male: female ratio was almost equal (1:1.1).
Table 2Results of immunohistochemical analysis (IHC) and mutation analyses of CTNNB1 and APC.
In samples D14 and D18, IHC of β-catenin and DNA-isolation did not succeed due to low cellular density and the poor quality of FFPE stored tumour samples (no sequences obtained).
Immunohistochemistry succeeded in all 18 tumours. Eleven tumours (61%) showed strong nuclear staining for β-catenin (Table 2 and Fig. 1a and c). No nuclear expression of β-catenin was present in normal cells in these tissues. In seven other tumours at most a light nuclear, membrane or cytoplasm staining was present (Fig. 1b and d).
Fig. 1Microscopy and immunohistochemistry. (A) and (B), hematoxylin and eosin (HE) stained section of two desmoid tumours. The same tumours were used for β-catenin immunohistochemistry: (C) note the strong nuclear accumulation of β-catenin in the tumour cells whereas the accompanying normal cells are negative; (D) note the absence of nuclear staining and the presence of membranous staining of β-catenin in the tumour cells. The blood vessels show a normal positive membranous and cytoplasmic β-catenin staining.
Sequence analysis of exon 3 of CTNNB1 succeeded in 16/18 tumours. No data were obtained from the remaining two samples due to poor quality DNA within these samples. In 7/16 tumours (44%) point mutations were found (Table 2 and Fig. 2a). Six mutations were situated in the expected positions (residues S45 and T41). One mutation was detected at position 35 (p.I35S), which has been described previously.
Frequent deletions and mutations of the beta-catenin gene are associated with overexpression of cyclin D1 and fibronectin and poorly differentiated histology in childhood hepatoblastoma.
Analysis for exon 3 deletions revealed no obvious deletions, but small deletions could have been missed, due to the relative large size of the residual product and poor quality of DNA extracted from FFPE material. All oncogenic CTNNB1 mutations were identified in tumours showing a strong nuclear accumulation of β-catenin (7/11).
Fig. 2(A) Sequence analysis of CTNNB1 (D12) showing heterozygosity for the p.S45F mutation; (B) sequence analysis of APC (D11) showing heterozygosity for the p.Q1406X mutation.
Mutation analysis of the APC MCR succeeded in 14/18 tumours. An APC mutation was found in two samples (14%), both causing a premature stopcodon (Table 2 and Fig. 2b). Both mutations were also found in the germline DNA of the patients. Ten samples harboured the heterozygous c.4479G>A change, known as polymorphism (data not shown). No APC mutation was detected in the CTNNB1 mutation positive samples. Both desmoid tumours in which we identified the APC mutation also showed a strong accumulation of nuclear β-catenin (2/11).
3.4 LOH analysis
To further study the possible involvement of APC in the development of the desmoid tumours, LOH analysis of the APC locus was performed in all tumours. Only one tumour, i.e. D20 in which we had previously identified the p.R1450X germline mutation, showed LOH of the APC locus (Fig. 3).
Fig. 3(A) Sequence analysis of APC in the tumour (D20), showing hemizygosity for the c.4348C>T p.R1450X mutation; (B) Identification of the germline APC mutation in patient no 19, showing heterozygosity for the c.4348C>T p.R1450X mutation; (C) Loss of heterozygosity (LOH) analysis in tumour D20, showing loss of heterozygosity of the APC locus (marker D5S656). The arrow indicates the loss of one allele in the tumour DNA.
We analysed 18 paediatric desmoid tumours for aberrations in the Wnt/β-catenin signalling pathway. Immunohistochemical analysis for β-catenin indicated that deregulation of this pathway was involved in the development of 11 tumours, of which two harboured point mutations in the APC gene (D11, D20, Table 2). These were confirmed as being a germline APC mutation. Patient 10 was a member of a known FAP family. The father of patient 19 was known with polyposis. During the following genetic counselling, we identified both father and son as being a carrier of the p.R1450X germline APC mutation. We identified loss of the remaining wild type allele (LOH of APC) in the desmoid tumour of patient 19. This is in line with Knudson’s two-hit hypothesis,
further supporting that the germline APC mutation underlies formation of this tumour.
The results of the molecular and immunohistochemical analyses of the desmoid tumours of patient 10 and 19 indicate that these analyses can lead to the identification of FAP carriers.
Mutations in CTNNB1 and APC have been shown to occur in a mutually exclusive manner in all tumour types studied so far. As germline oncogenic CTNNB1 mutations are not compatible with adult life, the identification of such a mutation provides direct evidence of the sporadic nature of the tumour. CTNNB1 mutations are more common than APC mutations in adult desmoid tumours, suggesting that most desmoid tumours are of sporadic origin.
Detection of beta-catenin mutations in paraffin-embedded sporadic desmoid-type fibromatosis by mutation-specific restriction enzyme digestion (MSRED): an ancillary diagnostic tool.
Recently, Bo et al. showed that CTNNB1 is also frequently mutated in paediatric desmoid tumours (25/32), but FAP patients were excluded in their study.
We observed a somewhat lower overall mutation frequency than previous reports, which may be explained by the limited number of cases in our cohort. However, the less frequent involvement of Wnt signalling pathway in paediatric desmoid tumours has been previously described as well.
Also in our set, no simultaneous APC mutations were detected in the tumours with a somatic CTNNB1 mutation, and as such these tumours can be considered as sporadic tumours.
Interestingly, we observed two samples with intranuclear staining for β-catenin in which neither CTNNB1 nor APC mutations were detected. However, we still do suspect an involvement of the Wnt/β-catenin signalling pathway in these samples. An APC-mutation localised outside the analysed region cannot be excluded. Alternatively, these samples could contain an alteration of other Wnt-related genes, such as AXIN1 or AXIN2, although mutations in these genes have not been reported in desmoid tumours previously. Also epigenetic alterations may play an important role in these tumours. Recently, Okpanyi et al. showed activation of the Wnt/β-catenin pathway in paediatric germ cell tumours due to APC promoter methylation and LOH of APC.
In colorectal cancer an epigenetic inactivation of secreted frizzled related proteins (SFRPs), which normally suppresses Wnt signalling, has been suggested to contribute to colon cancer formation.
As the epigenetic mechanisms in desmoids tumours are not well understood, it should be investigated in the future.
In four desmoid tumours, neither nuclear β-catenin staining nor CTNNB1 or APC mutations were detected. At this moment we have no evidence of Wnt/β-catenin pathway involvement in these tumours, and the aetiology of these tumours remains unclear. Previous studies described a lack of CTNNB1 and APC gene mutations in superficial fibromatoses.
However, since all tumours in our cohort were classified as deep fibromatoses, this feature represents an unlikely explanation for the lack of mutation in these mutation-negative tumours. Interestingly, patient 7 developed two desmoid tumours with both presence and absence of intranuclear staining of β-catenin in different tumours. His sister had a fibrous hamartoma of infancy. Previously it has been reported that Wnt/β-catenin pathway plays an important role in desmoid formation, but does not appear to play a role in the pathogenesis of other myofibroblastic lesions in children.
Pathological diagnosis of fibroblastic lesions can be challenging. As such, it cannot be entirely excluded that some of the immuno-negative lesions might not be desmoid tumours, although central review by an experienced pathologist has been performed.
Despite the limited size of our cohort, our data support current ideas about the molecular background in desmoid tumours and illustrate that a combination of IHC of β-catenin and the mutation analysis of CTNNB1 and APC can help to identify APC carriers among the children with desmoid tumours. This may be especially valuable in de novo APC mutation carriers, with a negative family history for FAP. It is known that approximately 20–25% of individuals with FAP have a de novo APC mutation.
Unfortunately, not in all our FFPE samples sufficient DNA quality could be obtained, most likely due to the well-known problems associated with over-fixation and long-term storage in paraffin. To this aim, we recommend that fresh frozen tissue of the desmoid sample is stored specifically for DNA isolation, in addition to the FFPE sample. The latter is still required for a proper histological evaluation and for β-catenin IHC, as it is not possible to detect nuclear β-catenin in frozen sections.
In conclusion, children with desmoid tumours can be carriers of a germline APC mutation. Therefore, we recommend to analyse the lesions for nuclear staining of β-catenin by IHC to identify the desmoid tumours with an underlying defect in β-catenin signalling. If consequently pathogenic CTNNB1 mutations are detected, this makes an increased risk of FAP in the patient unlikely. If, however, CTNNB1 screening turns out negative, genetic counselling of such children and their parents is warranted and germline APC analysis should be offered. These recommendations should be evaluated by future studies in larger cohorts.
Conflict of interest statement
None declared.
Acknowledgement
E.M.H. Mathus – Vliegen, MD, PhD, Department of Gastroenterology & Hepatology, Academic Medical Centre, Amsterdam, The Netherlands
Michael A. den Bakker, MD, PhD, Department of Pathology, Josephine Nefkens Institute, Erasmus MC, University Medical Centre, Rotterdam, The Netherlands
References
Reitamo J.J.
Scheinin T.M.
Hayry P.
The desmoid syndrome. New aspects in the cause, pathogenesis and treatment of the desmoid tumor.
Detection of beta-catenin mutations in paraffin-embedded sporadic desmoid-type fibromatosis by mutation-specific restriction enzyme digestion (MSRED): an ancillary diagnostic tool.
Frequent deletions and mutations of the beta-catenin gene are associated with overexpression of cyclin D1 and fibronectin and poorly differentiated histology in childhood hepatoblastoma.
00009-3&id=gr1.jpg){kind=link}
00009-3&id=gr2.jpg){kind=link}
00009-3&id=gr3.jpg){kind=link}