Introduction

FAP is caused by the dominant inheritance of germline mutations of the adenomatous polyposis coli (APC, MIM 175100) tumour suppressor gene. Truncating germline mutations in the APC gene have been found to be responsible for 70–90% of FAP cases and families.(Lipton and Tomlinson 2006) Mutational inactivation of APC leads to the accumulation and nuclear translocation of β-catenin, resulting in aberrant activation of the canonical Wnt signaling pathway implicated in colon cancer development.

Cloned in 1991, the APC gene is located on chromosome 5q21 and consists of 21 exons.(Groden et al. 1991) The major APC transcript has an 8,538 bp open reading frame that represents 15 exons and encodes a 312 kDa protein. Additionally, APC can express multiple transcripts of various lengths in a tissue-specific manner and a recent study identified 9 different alternatively or aberrantly spliced transcripts, three of which occurred only in FAP patients (Horii et al. 1993). Two distinct promoters have been identified for the APC gene. The major transcript is initiated by its main promoter, denoted as promoter 1A (Esteller et al. 2000; Horii et al. 1993). Hypermethylation of promoter 1A has been shown to impair APC expression and to contribute to the development of approximately 18% of primary colorectal carcinomas and adenomas. Some APC transcripts are thought to arise from a second promoter, termed 1B (Lambertz and Ballhausen 1993). However, this promoter does not appear to be methylated (Esteller et al. 2000; Hiltunen et al. 1997; Wong et al. 2008) and details of this minor promoter and on how it regulates transcription of the APC gene remain poorly understood.

More than 700 pathogenic APC mutations have been reported to date, but the most common germline mutations involve the introduction of a new stop codon, leading to truncation of the protein product (Galiatsatos and Foulkes 2006). Large deletions comprise approximately 2% of all reported mutations, including deletions that extend from the promoter into the coding region (Aretz et al. 2005). However, to date, promoter-specific deletions have not been reported and/or characterized. Here we describe a large Canadian Mennonite family with FAP in which a novel germline deletion in the APC promoter region has been identified.

The Mennonites, named for the Dutch priest Menno Simons, are a religious group that arose from the sixteenth century European Anabaptist movement. They are divided into two ancestral groups: Dutch–German and Swiss–German. With some exceptions, Swiss–German groups are concentrated in the Kitchener-Waterloo area of Ontario, Canada and in Pennsylvania State, USA whereas the Dutch-German groups are most often located in Western Canada, the American Midwest and throughout South America. Separated by geography and religious differences these two ancestral groups rarely intermarry. Although a few genetic disorders, including infantile hypophosphatasia (Greenberg et al. 1993), incomplete X-linked congenital stationary night blindness (Boycott et al. 2000), and several immunodeficiency syndromes including severe combined immunodeficiency syndrome and Fanconi Anemia (unpublished data), have been previously anecdotally reported as somewhat over-represented in the Manitoba Dutch-German Mennonite population, the incidence of cancer has not been formally reported to be increased.

Patients and methods

Patients

The proband was a 28-year-old Mennonite gentleman (Fig. 1a, II-4) who was referred for genetic evaluation because of a confirmed diagnosis of familial adenomatous polyposis (FAP). His father had died from metastatic disease and two of his five siblings were known to be affected. The patient had undergone prophylactic total colectomy and proctectomy with ileo-anal reconstruction in 2002. Pathological examination of the bowel revealed multiple (approximately 90) tubular and flat adenomata consistent with FAP. The proband was referred to the Molecular Diagnostic Laboratory at the Department of Pathology and Laboratory Medicine, Mount Sinai Hospital, University of Toronto, for molecular genetic testing to confirm a diagnosis of FAP. Patient accrual, sample collection, and genetic screening were performed according to clinical practice at the Winnipeg Health Sciences Centre and to guidelines of the Ethics Committee of the University of Toronto.

Fig. 1
figure 1

a Pedigree of the polyposis family screened by linkage analysis and MLPA (round parentheses). The APC haplotypes for I:1 have been extrapolated (square parentheses). The proband is indicated by an arrow. Shaded symbols represent affected with polyposis. Open triangles deleted; NT not tested. b Detection of large genomic APC deletions by multiplex ligation-dependent probe amplification (MLPA). The representative positive sample from the proband is shown (center panel) in comparison to negative and positive samples from two kindred members (upper and lower panels, respectively). The reduced genomic copy number of Promoter 1A and 5′ untranslated regions of APC are indicated (arrows). Unlabeled peaks are reference control probes, including the peak [C] between the arrows. Exon 15 is captured by three probes (15A, 15B, and 15C). c An illustration of APC promoters 1A and 1B, exons and the locations of the MLPA probes targeting the promoter region (Probes 1–3) and Exon 1 (Probe 4). Linkage analysis markers situated in the promoter region (rs2019720) and Exon 11 (rs2229992) are also shown. The distances between these probes and markers are indicated. Relative sizes of intronic regions are not to scale. Promoter regions 1A and 1B are depicted as black boxes. The 5′ UTRs are shaded grey

Molecular screening strategy

Genetic screening for germline APC mutations was initiated with combined methods of the protein truncation test (PTT) and exon-by exon sequencing of lymphocyte genomic DNA. APC codons # 1-774 spanning the first 14 exons were examined by sequence analysis, while the major coding exon (exon 15, codons #686-2843) was screened for truncating mutations by PTT. Since 2006, patient samples that are found negative by these two methods are further screened by multiplex ligation-dependent probe amplification (MLPA) analysis, which detects large genomic deletions and/or duplications. Prior to 2006, when a mutation was not identified by combined methods of PTT and genomic DNA sequencing, whenever possible, linkage analysis was used to predict the likelihood that an APC mutation-bearing haplotype segregated with the clinical phenotype. The major limitation of this strategy was the reliance on obtaining samples from multiple relevant family members and identifying informative markers for each family.

PTT and sequence analysis

Blood samples were obtained from patients and lymphocytes were isolated using NH4Cl–Tris. DNA extraction was performed using a saturated salt solution. RNA extraction was performed using TRIzol (Invitrogen Life Technologies, Burlington, Canada) according to the manufacturer’s protocol for RT-PCR analysis. The in vitro synthesized PTT assay was performed by use of a commercial kit (TNT T7 Quick Coupled Transcription/Translation System; Promega, Madison, WI, USA) as previously described (Charames et al. 2002).

Sequencing reactions were carried out in forward and reverse orientations by using the ABI BigDye Terminator Ready Reaction Mix (Applied Biosystems, Foster City, CA, USA) and analyzed on an ABI 3130XL Genetic Analyzer according to the manufacturer’s protocol. Primer sequences are included in Table 1.

Table 1 Primer Sequences for the detection of germline mutations by sequence analysis of exons 1–14 of the APC gene

Multiplex ligation-dependent probe amplification (MLPA)

MLPA is a PCR-based assay that identifies large genomic deletions and duplications in a patient sample as compared to a normal or wild-type control, without the requirement of additional samples from the patient’s family. The MLPA-APC kit was obtained from MRC-Holland (PO43, Amsterdam, The Netherlands). MLPA reactions were performed according to the manufacturer’s instructions. Products were analyzed using the ABI 3130XL Genetic Analyzer and the ABI GeneMapper 4.0 software (Applied Biosystems, Foster City, CA, USA).

Linkage analysis

For all living family members, genotypes were determined at seven different polymorphic marker loci. The following markers were intragenic to APC: (1) an A/G polymorphism (National Center for Biotechnology Information single-nucleotide polymorphism cluster ID: rs2019720) located within the promoter region detected by RsaI digests, (2) a T/C polymorphism located at nucleotide 1458 (exon 11) detected by RsaI digests (rs2229992), (3) a G/A polymorphism located at nucleotide 4479 (exon 15) detected by BsaJ1 digests (rs41115), (4) a G/A polymorphism located at nucleotide 5880 (exon 15) detected by MspI analysis (rs465899), and (5) a T/C polymorphism located within the 3′ untranslated region detected by SspI digests (rs41116). Primer sequences and protocols are as previously described (Sieber et al. 2002). In addition, family members were genotyped for two microsatellite markers (D5S318 and D5S346) located 3′ of APC, also previously described (Bapat et al. 1993; Wijnen et al. 1991).

Results and discussion

This is the first report of a genomic deletion in the APC promoter region which results in gene silencing and is responsible for familial polyposis. The proband of this Mennonite FAP family (II-4), was clinically affected with polyposis, as were three of his first-degree relatives (father I-1, and two sisters II-2 and II-6) (Fig. 1a). Thus germline testing for APC mutations was warranted. A combination of PTT and sequencing analysis of the entire APC coding region was initially performed; but a germline mutation was not identified. As this case was originally referred for predictive testing for the unaffected first degree relatives, linkage analysis was pursued to confirm linkage of the polyposis phenotype in affected family members to the APC gene (Fig. 1a). All three clinically affected family members (II-2, II-4, II-6) shared the same mutation-bearing APC haplotypes that was presumed to be inherited from their deceased clinically affected father. This supported the role of the APC gene in the polyposis phenotype. II-2 has inherited a maternal APC haplotype different from that inherited by II-4 and II-6. Comparative assessment of APC haplotypes among these three individuals suggested the presence of a deletion is more likely than a recombination event at the APC-promoter-RsaI (rs2019720) locus.

MLPA was available to capture large genomic deletions in the APC gene and therefore the proband’s sample was tested using this strategy. A distinct advantage of MLPA over more cumbersome methods such as Southern blotting, is that it is a PCR-based technique. Screening of proband’s DNA by MLPA identified a novel deletion mutation corresponding to promoter 1A and 5′ untranslated regions of APC (Fig. 1b, c). This large deletion was also detected among the proband’s other clinically affected siblings (II-2, II-6), and was subsequently identified in other at-risk members (II-7) including both of his children (III-1, III-2). The proband’s unaffected sisters (II-1 and II-3) do not carry the same APC germline promoter deletion.

The existence of two distinct APC promoters has been well documented, as is the occurrence of multiple APC transcripts generated under a variety of conditions and/or in a tissue-specific manner.(Lambertz and Ballhausen 1993) Promoter 1B (Genbank accession: D13981) is located approximately 30 kb upstream of Promoter 1A (Genbank accession: U02509).(Tsuchiya et al. 2000) By MLPA, we could not establish the precise limits of this large promoter deletion, and were unable to determine whether the deletion encompasses neighboring alternate promoter sequences such as promoter 1B. However, several studies investigating promoter 1A methylation have proved that this is the major transcript and therefore leads to gene silencing and the polyposis phenotype (Brabender et al. 2001; Esteller et al. 2000; Tsuchiya et al. 2000; Wong et al. 2008). Therefore, we sought to characterize the functional consequence of this novel deletion spanning promoter 1A, by examining its effect on the transcription of APC. We investigated whether family members who were heterozygous and informative for downstream intragenic APC polymorphic markers at the genomic level, had lost their heterozygosity at the transcript level. RNA was available from two of the proband’s siblings: II-1, who was found to be APC promoter deletion negative and II-7 who was found to be positive for this promoter deletion by MLPA. We selected intragenic exon 11 Rsa1 polymorphic linkage marker for this analysis for two reasons: (1) II-1 and II-7 were heterozygous and informative at this polymorphic site (T/C). (2) Linkage analysis had shown that the wild-type T allele of exon 11 polymorphism was linked to the novel APC promoter deletion in this family (Fig. 1a). We compared APC exon 11 sequences generated from genomic DNA and cDNA of II-1 and II-7 (Fig. 2). This analysis showed that both T and C alleles were retained at the genomic and cDNA level for II-1, while the individual with the promoter deletion (II-7) lost the T allele at the transcript level. Taken together, this data suggests that the genomic deletion mutation of the APC promoter and 5′ UTR likely results in a lack of transcription of the deletion-bearing mutant allele, leading to FAP in this family.

Fig. 2
figure 2

APC Exon 11 sequencing analysis of genomic and complementary DNA from MLPA negative and positive family members (II:1 and II:7, respectively). The site of the c.1458T>C polymorphism is indicated (arrow)

A 62-year-old proband from another Mennonite polyposis family, was referred in 2002 for genetic counselling as she had been recently diagnosed with metastatic adenocarcinoma of the colon secondary to FAP. She died shortly thereafter. She had 6 children from two marriages. One son had died at 22 years of age of colon cancer and he apparently had multiple colonic polyps, while one daughter was known to have multiple colonic polyps. This proband was from a very large extended Manitoba Mennonite family with several individuals in three successive generations (including her father and two sisters) known to either have multiple colonic polyps or adenocarcinoma of the bowel. APC gene sequencing of all coding exons had been pursued elsewhere and no disease-causing mutation was identified. Given the similar ethnic background, MLPA analysis was then pursued and a similar promoter deletion was identified in her and her affected niece (daughter of one of her deceased sisters). Prior to this molecular result, this 62-year-old woman was not known to be related to our first proband. However, it is now confirmed that one of her paternal uncles was the paternal grandfather of proband II-4 (Fig. 1a); thus, molecular analysis helped to confirm the genealogical relationships in this expanded polyposis family.

The Old Colony Mennonites have previously been recognized as a genetic and religious isolate with a distinct pattern of inherited disease (Jaworski et al. 1989). The Mennonite population in Manitoba has been affected by other diseases with a seemingly rare origin. A multiple congenital anomalies syndrome with autosomal recessive inheritance, termed CODAS, has been reported in three children worldwide (Innes et al. 2001). Two of these children are of Canadian Mennonite ancestry. Additionally, homozygosity for the p.Gly317Asp missense mutation in an alkaline phosphatase gene (ALPL) has been implicated in the infantile form of hypophosphatasia, a disorder characterized by defective bone mineralization.(Greenberg et al. 1993) This mutation has been specifically linked to the Canadian Mennonite population, since it was not described in controls nor in other non-Mennonite probands with both forms of hypophosphatasia. A single nonsense MLH1 mutation, p.Trp714X, is responsible for many but not all Manitoba Mennonites diagnosed with HNPCC (unpublished data). These studies strengthen the likelihood that this novel APC promoter mutation is segregating in this unique population as a founder mutation. However, further characterization and mapping is still required.