tumor development in the beckwithâ€“wiedemann syndrome is

© 2001 Oxford University Press Human Molecular Genetics, 2001, Vol. 10, No. 26 2989–3000

Tumor development in the Beckwith–Wiedemann syndrome is associated with a variety of constitutional molecular 11p15 alterations including imprinting defects of KCNQ1OT1Rosanna Weksberg1,2,3,4,*, Joy Nishikawa3, Oana Caluseriu2, Yan-Ling Fei2, Cheryl Shuman1,2,3, Cuihong Wei5, Leslie Steele3,5, Jessie Cameron2, Adam Smith2,4, Ingrid Ambus1, Madeline Li1, Peter N. Ray2,3,5, Paul Sadowski3 and Jeremy Squire6,7,8

1Division of Clinical and Metabolic Genetics and the Department of Paediatrics and 2Research Institute, Hospital for Sick Children, Toronto, Ontario, Canada, 3Department of Molecular and Medical Genetics and 4Institute of Medical Sciences, University of Toronto, Toronto, Ontario, Canada, 5Department of Paediatric Laboratory Medicine, Hospital for Sick Children, Toronto, Ontario, Canada, 6Ontario Cancer Institute, Toronto, Ontario, Canada, 7Department of Laboratory Medicine and Pathobiology and 8Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada

Received August 10, 2001; Revised and Accepted October 29, 2001

Dysregulation of imprinted genes on humanchromosome 11p15 has been implicated inBeckwith–Wiedemann syndrome (BWS), an over-growth syndrome associated with congenitalmalformations and tumor predisposition. The molecularbasis of BWS is complex and heterogeneous. Thesyndrome is associated with alterations in twodistinct imprinting domains on 11p15: a telomericdomain containing the H19 and IGF2 genes and acentromeric domain including the KCNQ1OT1 andCDKNIC genes. It has been postulated that disordersof imprinting in the telomeric domain are associatedwith overgrowth and cancer predisposition, whereasthose in the centromeric domain involve malforma-tions but not tumor development. In this study of125 BWS cases, we confirm the association oftumors with constitutional defects in the 11p15 telo-meric domain; six of 21 BWS cases with uniparentaldisomy (UPD) of 11p15 developed tumors and one ofthree of the rare BWS subtype with hypermethylationof the H19 gene developed tumors. Most importantly,we find that five of 32 individuals with BWS andimprinting defects in the centromeric domain devel-oped embryonal tumors. Furthermore, the type oftumors observed in BWS cases with telomericdefects are different from those seen in BWS caseswith defects limited to the centromeric domain.Whereas Wilms’ tumor was the most frequent tumorseen in BWS cases with UPD for 11p15 or H19 hyper-methylation, none of the embryonal tumors with

imprinting defects at KCNQ1OT1 was a Wilms’ tumor.This suggests that distinct tumor predispositionprofiles result from dysregulation of the telomericdomain versus the centromeric domain and thatthese imprinting defects activate distinct geneticpathways for embryonal tumorigenesis.

INTRODUCTION

Beckwith–Wiedemann syndrome (BWS) (OMIM 130650;http://www3.ncbi.nlm.nih.gov/htbin-post/Omim/dispmim?130650) is a phenotypically and genotypicallyheterogeneous overgrowth syndrome (1–4). The most commonphenotypes associated with BWS are pre- and post-natal over-growth, organomegaly, hemihyperplasia, omphalocele, earlobe and renal abnormalities. Children with BWS have a7–21% risk for the development of embryonal malignancies,most notably Wilms’ tumor of the kidney (5–8). However, awide variety of benign and malignant tumors have beenreported (1,5). These malignancies include hepatoblastoma,adrenocortical carcinoma, rhabdomyosarcoma and neuro-blastoma (5).

BWS has been associated with a diversity of genetic andepigenetic alterations on chromosome 11p15 (3,4). Theimprinted gene cluster on human chromosome 11p15.5 hasbeen proposed to consist of two domains (3,4,9) (Fig. 1). Thetelomeric domain includes the paternally expressed gene IGF2(insulin-like growth factor 2) and the maternally expressedgene H19, whereas the centromeric domain includes thematernally expressed genes CDKNIC (p57KIP2), KCNQ1(KvLQT1) and the paternally expressed gene KCNQ1OT1(KvLQT1-AS, LIT1). Two non-imprinted genes TSSC4 and

*To whom correspondence should be addressed at: Division of Clinical and Metabolic Genetics, Hospital for Sick Children, 555 University Avenue,Toronto, Ontario M5G 1X8, Canada. Tel: +1 416 813 6386; Fax: +1 416 813 5345; Email: [email protected]

2990 Human Molecular Genetics, 2001, Vol. 10, No. 26

PHEMX (10) map to the 11p15 region between the twoimprinted gene clusters.

In the telomeric domain, H19 encodes an apparently untrans-lated RNA (11). The function of this gene is presentlyunresolved but it is involved in regulating the expression of theIGF2 gene, which encodes a fetal growth factor. A putativeimprinting control center, named H19DMR, is locatedupstream of H19, and regulates H19 and IGF2 expression.

On the expressed maternal H19 allele, H19DMR is non-methylated and is able to bind the CTCF protein. This insulatesthe maternal IGF2 promoter from two enhancers located 3′ ofthe H19 gene and prevents expression of IGF2. On the paternalallele, the H19 DMR is methylated. This not only preventsexpression of the imprinted paternal H19 alleles, but it alsoblocks the binding of the CTCF protein (12). The IGF2promoter now has access to the 3′ enhancers and the paternalallele is expressed via enhancers and boundary elements in theregion (13,14).

In the centromeric domain, CDKNIC encodes an inhibitor ofcyclin-dependent kinases and is involved in cell cycle regula-tion. KCNQ1 encodes a voltage-gated potassium channel. TheKCNQ1OT1 gene encodes a paternally expressed, untranslatedtranscript which is read in antisense orientation to and over-lapping with the KCNQ1 gene (9,15). KvDMR1 is a maternallymethylated CpG island at the 5′ end of the KCNQ1OT1 tran-script (9,15,16). Recently, Horike et al. (17) generated modi-fied human chromosomes carrying targeted deletions ofKvDMR1. In these cells the KCNQ1OT1 gene was silenced andthe KCNQ1 and CDKNIC genes were overexpressed from thepaternal chromosome bearing the deletion. These resultssuggest that KvDMR1 may function as a regional imprintingcenter (BWSIC2) that regulates the centromeric 11p15imprinted domain.

The genetics of BWS is complex. Approximately 15% ofcases are familial and 1–2% have translocations and duplica-tions of 11p15 (1,2). The majority (∼85%) of BWS cases aresporadic with normal karyotypes. Of the sporadic cases, ∼20%exhibit somatic mosaicism for paternal uniparental disomy(UPD) of both the centromeric and telomeric domains (18,19).

Sporadic cases of BWS without 11p15 UPD exhibit changes ina number of imprinted genes, including mutations in theCDKNIC gene (∼5% of sporadic cases) (20,21) and loss ofimprinting of IGF2 resulting in biallelic expression (22). Thisloss of imprinting of IGF2 can occur in two situations: rarely inconjunction with hypermethylation of H19 in 1–2% of BWScases (23), or more commonly, accompanied by normalmethylation of the H19 gene (22,24). Recently, it has beenshown that ∼50% of sporadic BWS cases show loss of methy-lation at KvDMR1 (9,15,25,26). Lee et al. (9) found that loss ofmethylation at KvDMR1 was associated with biallelic expressionof KCNQ1OT1 in eight of 16 BWS cases.

Several studies have shown that dysregulation of imprintingin the telomeric domain is associated with predisposition totumors in BWS (27). Lee et al. (9) suggested that the two11p15 imprinting domains have overlapping but non-identicalfunctions in BWS, with the telomeric domain primarilyinvolved in cancer predisposition. Moreover, Tycko (27)proposed that H19 hypermethylation was strongly associatedwith tumorigenesis either in the context of 11p15 UPD or as anisolated epigenetic event. These findings were supported byreports of cases of BWS and tumors associated with 11p15UPD (25,26) and with the rare epigenotype of hypermethyla-tion of H19 (25,27). However, mutation of CDKNIC, a mater-nally expressed candidate tumor suppressor gene located in thecentromeric imprinting domain, has been associated withtumor development in only two BWS cases to date (28,29). Noinformation regarding tumor status was reported for 30 casesof BWS with imprinting defects of KvDMR1 (9,15,16). Inaddition, there have been only two studies of BWS patientswith loss of methylation at KvDMR1 which reported no embry-onal tumors (25,26). Recently, only one malignant tumor wasreported for a non-UPD BWS patient with demethylation atKvDMR1 (29). Thus, to date, only two BWS molecularsubgroups have been clearly implicated in the tumor predis-position associated with BWS, i.e. UPD of 11p15 and hyper-methylation of H19 (30). Heretofore, imprinting defects at theKCNQ1OT1/KvDMR1 locus have not been considered topredispose to embryonal tumor development in BWS (25,26).

Figure 1. Genes and their imprinting status on human chromosome 11p15 (not all genes are shown). Expressed genes are indicated by open boxes and silencedgenes by closed boxes. DMRs are indicated as a hatched box for the H19 DMR and a gray box for the KvDMR. For differentially methylated maternal and paternalalleles, CH3 indicates the methylated parental chromosome. Alternate names for the following approved gene names are in parentheses: CDKNIC (p57KIP2),KCNQ1 (KvLQT1), KCNQ1OT1 (LIT1 and KvLQT1-AS).

Human Molecular Genetics, 2001, Vol. 10, No. 26 2991

In this work, we examine the molecular lesions that occur incases of BWS with tumors. We confirm that 11p15 UPD andH19 hypermethylation are indeed associated with tumor devel-opment. We demonstrate that BWS patients with imprintingdefects at KCNQ1OT1/KvDMR1 are also at risk for tumor

development. Furthermore, the data suggest that the types oftumors found in BWS cases with constitutional imprintingdefects involving the telomeric domain are distinct from thoseassociated with disruption of imprinting of the centromericdomain.

Figure 2. Quantitative multiplex-PCR analysis of STR polymorphisms used to detect paternal chromosome 11 UPD. (A) Child with BWS and non-UPD 11 asindicated by the normal dosage of alleles transmitted from father and mother at DNA markers within the 11p15.5 region [D11S1997 (blue), D11S2362 (green) andTH (black)] and at two DNA markers on the q arm of chromosome 11 [D11S1974 (green) and D11S1998 (black)]. (B) Child with BWS and paternal UPD 11 asindicated by the increase in dosage of the paternally derived allele compared to the maternally derived allele at DNA markers within the 11p15.5 region(D11S1997, D11S2362 and TH) and the normal dosage of two DNA markers on the q arm of chromosome 11. (C) Child with BWS and paternal UPD 11 detectedin skin fibroblast cells derived from side of the body with hyperplasia (D11S2362 and TH) and non-UPD 11 in skin fibroblast cells from the side of the bodywithout hyperplasia.


RESULTS

Chromosome 11 uniparental disomy

Analysis for UPD of chromosome 11 was carried out usingquantitative multiplex-PCR with highly polymorphic STRmarkers. At least two informative markers in the 11p15 regionwere required to complete the analysis. Normal bi-parentalinheritance of these 11p15 markers was found in all 12 controlsamples (Fig. 2A). Of 125 BWS cases tested, we detected 21cases of UPD for 11p15 (Fig. 2B). In 17 of these cases,analysis of parental samples confirmed paternal origin of theUPD. In the other four cases, parental samples were not avail-able. For all 21 cases of 11p15 UPD, somatic mosaicism wasevident, the percentage varying from 25 to 95%. In fact, fortwo cases of BWS, 11p15 UPD was detected in only some ofthe available tissue samples. Figure 2C shows data for onesuch individual with hemihyperplasia affecting one whole sideof the body. The skin fibroblast cells derived from the arm withhemihyperplasia showed UPD for 11p15 whereas the fibro-blast strain from the other arm and from lymphocytes showednormal biparental inheritance of 11p15 markers. Six (29%) ofthe 21 cases of UPD developed malignant tumors. Theseincluded five Wilms’ tumors and one hepatoblastoma.

CDKNIC mutations

DNA from BWS patients was screened for coding mutations inthe CDKNIC gene. Mutations were identified in five patientsof the 125 patients screened. All mutations were inherited fromthe mother except in one case where paternal transmission wasnoted. Four of the five mutations have been previouslypublished (31). The new mutation was identified in a sporadicBWS case and is a C→T transition at nucleotide 432. Thischange results in a single amino acid substitution at residue 58converting a proline (CCG) to a leucine (CTG) in the CdKinhibitory domain (P58L mutation).

Methylation status of H19

In order to assess the methylation status of the H19 gene, weanalyzed the differentially methylated SmaI site near thepromoter of H19 by Southern blot analysis. A methylationindex (MI) was calculated as formulated by Reik et al. (32).For 10 controls (Fig. 3A), MIs gave a mean of 0.50 with a SDof ±0.05. We defined hypermethylation at the H19 locus as aMI of ≥0.60 (mean + 2SD). We found hypermethylation of theH19 gene in 24 of 125 BWS cases. Of these, 21 cases had UPD(Fig. 3B), leaving three of 125 cases with hypermethylation ofH19 as an independent finding (Fig. 3C). All 12 familial casesshowed a normal methylation pattern for H19 (data notshown). An anaplastic Wilms’ tumor was found in one (33%)of the three BWS cases with H19 hypermethylation.

Methylation status of KvDMR1

The methylation status of KvDMR1 was analyzed using thedifferentially methylated NotI site in the CpG island upstreamof KCNQ1OT1, KvDMR1. The MI was calculated as for H19.For 10 controls (Fig. 4A), a mean MI of 0.50 was obtained with

a SD of ±0.04. A MI ≤0.42 (mean – 2SD) was defined ashypomethylation at KvDMR1.

We analyzed 72 BWS cases for methylation at KvDMR1 meth-ylation. As expected, all 13 patients (18%) with UPD exhibitedhypomethylation at KvDMR1 (Fig. 4B). In addition, 35/59 (59%)of the remaining 59 non-UPD BWS cases exhibited hypomethyl-ation at KvDMR1 (Fig. 4C and Table 1). In 11 cases, completedemethylation was observed; in 25 cases the demethylation wasincomplete. We found malignant tumors in five (14%) of the 36non-UPD BWS cases with hypomethylation at KvDMR1.

Expression analysis of KCNQ1OT1

In order to evaluate whether hypomethylation at KvDMR1 wasconsistently associated with altered expression of theimprinted transcript KCNQ1OT1 we assayed its expressionusing PCR-based assays of three polymorphic SNPs (9). All 12controls, informative for at least one SNP, showed monoallelicexpression of KCNQ1OT1 (Fig. 5A and Table 1). This findingis consistent with the normal MIs obtained for KvDMR1 of0.48–0.53. We found at least one informative SNP inKCNQ1OT1 in 67 of 72 BWS cases. Of these 67 cases, 35(52%) showed biallelic expression of KCNQ1OT1 (Fig. 5Band Table 1). These 35 cases all demonstrated hypomethy-lation of KvDMR1 with MIs ranging from 0 to 0.33. Theremaining 32 BWS cases showed monoallelic expression of

Figure 3. DNA methylation analysis of H19. Southern blots of DNA samplesdigested with PstI and SmaI hybridized with an H19 cDNA probe. The upperband (1.8 kb) is of paternal origin, the lower band (1.0 kb) of maternal origin.Blots obtained for (A) a control with a MI of 0.55, (B) a BWS patient with11p15 UPD and a MI of 0.74 and (C) a non-UPD BWS patient with a MI of0.84. (D–H) Normal MIs of 0.47, 0.51, 0.47, 0.47 and 0.49, respectively, infive cases of BWS.

Figure 4. DNA methylation analysis of KvDMR1. Southern blots of DNAsamples digested with NotI and EcoRI hybridized with a [32P]dCTP-labeledDMR probe. The upper band (4.2 kb) is of maternal origin, the lower band(2.7 kb) of paternal origin. Blots showing (A) a control with a MI of 0.50,(B) a BWS patient with paternal UPD of 11p15 and a MI of 0.24 and (C) anon-UPD BWS patient with an MI of 0.09.


KCNQ1OT1 transcript (Fig. 5C and Table 1), including twocases with hypermethylation of the H19 gene (Fig. 5D andTable 1). Informative parental samples available in 24 casesdemonstrated that this monoallelic expression was of paternalorigin. For 19 of the 32 BWS cases with monoallelic expression ofthe KCNQ1OT1 transcript, we observed normal methylation atKvDMR1 (MIs of 0.46–0.53). The remaining 13 cases all hadUPD for 11p15 and showed hypomethylation at KvDMR1(MIs of 0.35–0.10) (Fig. 5E and Table 1). That is, apparentdiscordance between methylation status at KvDMR1 andtranscription of KCNQ1OT1 is a hallmark of BWS cases withUPD. This profile reflects the differential parental contributions inthe 11p15 region in the absence of a primary imprinting error.

Combined molecular analysis of 11p15

We have completed molecular analysis for UPD of 11p15,methylation of the H19 gene and CDKNIC mutations in 125cases of BWS. Methylation at KvDMR1 and allelic transcrip-tion of KCNQ1OT1 were assessed in 72 BWS cases. Fromthese data, we have grouped the patients into five molecularclasses: group I, UPD of 11p15; group II, methylation errors atH19; group III, methylation/transcription errors at KvDMR1/KCNQ1OT1; group IV, CDKNIC mutations; group V, normalmethylation patterns of KvDMR and H19. A summary of theresults is listed in Table 2. There were no statistical differencesin tumor frequencies among groups I, II, III and IV.

Delineation of BWS patient molecular subgroups

Group I: UPD. This group is defined by (i) MI for H19 of≥0.60, (ii) MI for KvDMR1 of ≤0.42, (iii) monoallelic expres-sion of KCNQ1OT1 and (iv) greater paternal than maternalcontribution of 11p15 alleles.

We assigned 21 of 125 BWS cases to the UPD group basedon a greater paternal than maternal 11p15 allelic contributionas well as a MI for H19 of 0.60. Thirteen of the UPD caseswere available for testing for KvDMR1. All of these showed aMI for KvDMR1 of ≤0.42 and normal monoallelic expressionof KCNQ1OT1. Six of the 21 UPD cases developed tumors,including five Wilms’ tumors and one hepatoblastoma(Table 2).

Group II: methylation error of H19. This group is defined by(i) MI at H19 of ≥0.60, (ii) MI at KvDMR1 of 0.50 ± 2SD, (iii)monoallelic expression of KCNQ1OT1 and (iv) normalbiparental 11p15 allelic contributions. Three of 125 (2%) had a

MI at H19 of ≥0.60–0.72, 0.78 and 0.84 with correspondingMIs at KvDMR1 of 0.52, 0.54 and 0.49, respectively. All threeshowed monoallelic expression of KCNQ1OT1 and normalbiparental inheritance of 11p15 markers. One (33%) of thesethree patients developed a Wilms’ tumor.

Group III: methylation error of KvDMR1 and biallelic expres-sion of KCNQ1OT1. This group is defined by (i) MI at H19 of0.50 ± 2SD, (ii) MI at KvDMR1 of ≤0.42, (iii) biallelic expres-sion of KCNQ1OT1 and (iv) normal biparental 11p15 alleliccontributions.

Thirty-five (59%) of 59 BWS cases (Table 2) showednormal biparental inheritance of 11p15 markers andhypomethylation at KvDMR1 with concomitant biallelicexpression of KCNQ1OT1. All 35 showed normal methylationat H19 (MIs of 0.46–0.55), supporting the hypothesis that H19and KvDMR1 methylation errors occur independently. Of the35 BWS cases in this group, five (14%) developed tumors.Surprisingly, none of these was a Wilms’ tumor. The tumorsincluded two hepatoblastomas, two rhabdomyosarcomas andone gonadoblastoma. These five BWS patients have normalbiparental inheritance of 11p15 markers (data not shown) andnormal H19 MIs (Fig. 3D–H), thereby excluding mosaicismfor 11p15 UPD and defects of H19 methylation. However, allfive cases show loss of maternal methylation at KvDMR1(Fig. 6A–E). Furthermore, allelic expression studies ofKCNQ1OT1 are available for three of these five cases and allshow biallelic expression of KCNQ1OT1 (Fig. 6F–H). There-fore, the constitutional molecular defects in these five cases ofBWS with non-Wilms’ embryonal tumors must involveprimary epigenetic alterations at KvDMR1/KCNQ1OT1.

Group IV: CDKNIC mutations. This group is defined by (i) MIat H19 of 0.50 ± 2SD, (ii) MI at KvDMR1 of 0.50 ± 2SD, (iii)monoallelic expression of KCNQ1OT1, (iv) normal biparental11p15 allelic contributions and (v) mutations in CDKNIC.

Five of 125 BWS cases had mutations in CDKNIC. All fivehad normal molecular analyses for H19 and KvDMR1 methy-lation, monoallelic expression of KCNQ1OT1 and normalbiparental 11p15 allelic contributions. None of these BWScases had embryonal tumors.

Group V: normal methylation. This group is defined by (i) MIat H19 of 0.50 ± 2SD, (ii) MI at KvDMR1 of 0.50 ± 2SD, (iii)monoallelic expression of KCNQ1OT1 and (iv) normal bipar-ental 11p15 allelic contributions. Seventeen (25%) of 67 BWS

Table 1. KCNQ1OT1 transcription and KvDMR1 methylation in BWS and controls

aIncludes two cases with hypermethylation at H19.bAll 13 in this category are UPD cases.

KvDMR1 and KCNQ1OT1 status Number of controls Number of BWS patients

KvDMR1 MI = 0.50 ± 2SD and monoallelic KCNQ1OT1 expression 12 19a

KvDMR1 MI = 0.50 ± 2SD and biallelic KCNQ1OT1 expression 0 0

KvDMR1 MI ≤ 0.42 and monoallelic KCNQ1OT1 expression 0 13b

KvDMR1 MI ≤ 0.42 and biallelic KCNQ1OT1 expression 0 35

Totals 12 67


cases (Table 2) showed normal biparental 11p15 contributionswith normal methylation at H19 and KvDMR1 accompanied bymonoallelic expression of KCNQ1OT1. Four (24%) of 17 hadWilms’ tumors. These four cases all had classical features ofBWS including macrosomia, macroglossia and hemihyper-plasia and none of these cases was familial.

Comparison of our data to published data on tumor risk and BWS molecular subgroups

Table 3 compares our data to the published data for tumor riskassociated with specific BWS molecular subgroups in the telo-meric and centromeric domains. Our data found six (29%)tumors in 21 cases of BWS with 11p15 UPD, consistent with

that previously reported by several other groups (Table 3). Allof these studies attribute tumor development in BWS caseseither to UPD of 11p15 or methylation defects at H19. Blieket al. (25) found four (18%) tumors among 22 cases of BWSwith UPD of 11p15. Both Engel et al. (26) and Gaston et al.(29) compiled previous and new data reporting two (9%)tumors in 22 BWS cases with 11p15 UPD and four (36%) in 11cases of BWS with 11p15 UPD, respectively. For the BWSpatients with H19 methylation errors, although only smallnumbers are available, our findings (one of three) areconsistent with those previously reported (five of 14) in thatpatients in this molecular subgroup are at increased risk todevelop tumors (Table 3). Overall, the data highlight thesignificant tumor risk associated with molecular defects in the

Figure 5. Allelic expression analysis of KCNQ1OT1. Allelic expression was determined by DNA sequencing of gel-purified RT–PCR products from lymphoblastsor fibroblasts of patients informative for SNP1, SNP2 and SNP3 polymorphisms as described by Lee et al. (9). The genomic sequences for these SNPs are: SNP1,AGCTCTGACC(G/A)TCAGACCCCC; SNP2, GAAATGTGTA(C/T)GGCATGTTGT; SNP3, CTAGACAGTG(C/T)GGCCCTCTCC. (A) A control hetero-zygous for SNP3 in genomic DNA with monoallelic KCNQ1OT1 expression (MI of 0.50 at KvDMR1). (B) A BWS case heterozygous for SNP2 in genomic DNA,with biallelic expression of KCNQ1OT1. The MIs were 0.24 at KvDMR1 and 0.54 at H19. (C) A BWS case heterozygous for SNP2 in genomic DNA with mono-allelic expression of KCNQ1OT1. MIs were 0.48 at KvDMR and 0.51 at H19. (D) A BWS case with H19 hypermethylation (MI = 0.72) and heterozygous forSNP3 in genomic DNA with monoallelic expression at KCNQ1OT1 and a MI of 0.47 at KvDMR1. (E) A BWS case with 11p15 UPD heterozygous for SNP1 ingenomic DNA showing monoallelic expression of KCNQ1OT1 with MIs of 0.25 at KvDMR1 and 0.74 at H19. gDNA, genomic DNA; cDNA, reverse-transcribedRNA. The arrows on the Figure indicate informative SNPs.

Table 2. Summary of combined data for BWS molecular groups

a35/59 (59%) for biallelic expression at KCNQ1OT1; 36/59 (61%) for loss of methylation at KvDMR1.

Group I, 11p15 UPD Group II, H19 hypermethylation

Group III, KCNQ1OT1 imprinting defect

Group IV, CDKNIC mutations

Group V, normal imprinting

11p15 allelic contributions Paternal > maternal Paternal = maternal Paternal = maternal Paternal = maternal Paternal = maternal

H19 methylation Hypermethylation Hypermethylation Normal Normal Normal

KvDMR1 methylation Hypomethylation Normal Hypomethylation Normal Normal

KCNQ1OT1 Monoallelic Monoallelic Biallelic Monoallelic Monoallelic

Frequency 21/125 (17%) 3/125 (2%) 35/59 (59%)a 5/125 (5%) 17/67 (25%)

Number of tumors 6/21 (29%) 1/3 (33%) 5/35 (14%) 0/5 4/17 (24%)

Tumor types Wilms’ tumor (5), hepatoblastoma (1)

Wilms’ tumor (1) Hepatoblastoma (2), rhabdomyosarcoma (2), gonadoblastoma (1)

– Wilms’ tumor (4)

Frequency of hemihyperplasia in group

19/21 (90%) 0/3 0/59 0/5 8/17 (47%)


telomeric domain. Combining our data with those in the litera-ture, 25 tumors have been observed in the 107 BWS cases withtelomeric molecular defects and a tumor incidence of 24%.

In the centromeric domain, although we did not find anytumors in five cases of BWS with CDKNIC mutations, thereare two neuroblastomas in approximately 33 BWS individualswith such mutations. Overall, the tumor incidence forCDKNIC mutation is 2/38 or 5.2% (Table 3).

Of particular interest are the data on tumor incidence in BWScases with methylation and transcription errors at KvDMR1and KCNQ1OT1. Our results showing five tumors in 35 BWScases with imprinting defects at KvDMR1/KCNQ1OT1 arestatistically different from those previously reported for BWScases with such constitutional alterations (Fisher test P < 0.01).Engel et al. (26) and Bliek et al. (25) each reported no tumorsin 29 cases of BWS with methylation errors at KvDMR1. Onerecent study (29) identified a single tumor in 30 cases of BWSwith KvDMR1 methylation defects which further supports ournew findings. Thus, to date, of 123 BWS cases reported withimprinting defects at KCNQ1OT1, six (5%) have developed amalignant tumor.

A review of the tumor types associated with specific molecularsubgroups of BWS demonstrates clear differences (Table 4).

Our data show that whereas Wilms’ tumor is the commontumor seen in BWS cases with molecular defects in thetelomeric domain, it is not represented in the five tumorsobserved in BWS cases with imprinting errors at KCNQ1OT1(Table 4). Instead, we see a variety of non-Wilms’ tumorsincluding rhabdomyosarcoma and gonadoblastoma. Hepato-blastoma is seen with molecular defects in both domains. Areview of such data from the literature further supports ourfinding. Combining all the available data for BWS cases withknown molecular defects and tumors (Table 4) it appears thatfor 11p15 UPD and methylation errors at H19, 22 (88%) of25 tumors reported are Wilms’ tumors. The other three tumorsare hepatoblastoma, neuroblastoma and pheochromocytoma.However, in the centromeric domain of the eight tumorsreported to date none is a Wilms’ tumor (Table 4). The presentstudy demonstrates two BWS cases with rhabdomyosarcomas,two BWS cases with hepatoblastomas, and one BWS case withgonadoblastoma and an imprinting defect at KCNQ1OT1. Inaddition, one other group has reported a thyroid carcinoma in aBWS case with an imprinting defect at KCNQ1OT1 (29).However, since thyroid carcinoma has not previously beenreported in BWS, it is not clear whether the thyroid carcinomaoccurred by chance or is truly associated with the imprintingdefect at KCNQ1OT1. For CDKNIC mutations, also in thecentromeric domain, two cases of neuroblastoma have beenreported (28,29).

For the group of BWS cases in which the tests includingUPD analysis, H19 methylation KCNQ1OT1, and CDKNICmutation analysis are negative (group V), the rate of tumors issix (22%) of 27 cases. In addition, all six of these group Vpatients, four in our study and two from Bliek et al. (25) hadhemihyperplasia. The tumor type is Wilms’ tumor in all sixcases, a profile suggestive of a telomeric domain moleculardefect.

DISCUSSION

In this paper, we have shown that several constitutional BWSmolecular defects are represented in BWS cases with tumors.We confirm that these molecular defects include UPD of11p15 and biallelic methylation of the H19 gene. Most import-antly, this is the first report of embryonal tumors in BWS caseswith constitutional imprinting defects at KCNQ1OT1.Dysregulation of imprinting at KCNQ1OT1 and/or KvDMR1,was found to be independent of methylation of the maternalH19 gene or of UPD at 11p15. These findings expand oncurrent molecular models of tumor predisposition on 11p15(3,4,30), and suggest that both the telomeric and centromericimprinted domains are implicated in embryonal tumorigenesis.Furthermore, our data suggest a molecular basis for the diver-sity of tumors observed in BWS. Our data imply that tumorrisk associated with UPD of 11p15 and H19 methylation errorsis different from that observed in BWS individuals withimprinting defects at KCNQ1OT1 and/or KvDMR1. Specific-ally, Wilms’ tumor is associated with BWS molecular defectsin the telomeric domain, whereas rhabdomyosarcoma andgonadoblastoma are associated with molecular lesions in thecentromeric domain. Hepatoblastoma occurs in BWS caseswith molecular lesions in either domain.

Figure 6. DNA methylation analysis of KvDMR1 and allelic expressionanalysis of KCNQ1OT1 in BWS cases with tumors. (A–E) DNA methylationanalyses for KvDMR1. These are Southern blots of DNA samples digestedwith NotI and EcoRI hybridized with a [32P]dCTP-labeled DMR probe. Theupper band (4.2 kb) is of maternal origin, the lower band (2.7 kb) of paternalorigin. (F–H) Allelic expression analyses of KCNQ1OT1 determined by DNAsequencing of gel-purified RT–PCR products from lymphoblasts of patientsinformative for SNP1 or SNP2 as described by Lee et al. (9). (A) A BWS casewith rhabdomyosarcoma. The MI for KvDMR1 is 0.30. This patient, hetero-zygous for SNP2 in genomic DNA, shows biallelic expression of KCNQ1OT1(F). (B) A second BWS case with rhabdomyosarcoma; the MI for KvDMR1 is0.09. This patient, heterozygous for SNP1 in genomic DNA, shows biallelicexpression of KCNQ1OT1 (G). (C) A BWS case with hepatoblastoma; the MIfor KvDMR1 is 0.20. This patient, heterozygous for SNP1 in genomic DNA,shows biallelic expression of KCNQ1OT1 (H). (D) A second BWS case withhepatoblastoma; the MI for KvDMR1 is 0.01. No cell line was available forallelic expression analysis of KCNQ1OT1. (E) A BWS case with gonado-blastoma; the MI for KvDMR1 is 0. Allelic expression for KCNQ1OT1 wasnot done because this individual is uninformative for all three transcribedSNPs of KCNQ1OT1. gDNA, genomic DNA; cDNA, reverse-transcribedRNA. The arrows on the Figure indicate informative SNPs.


11p15 UPD in BWS patients

BWS patients with UPD of 11p15 (group I) have an increasedpaternal genomic contribution. Because the majority of thesecases exhibit somatic mosaicism, abnormal methylationanalyses of genes in the 11p15 region will reflect the degree towhich the paternal methylation contribution is over-repre-sented. ‘Hypermethylation’ of H19 is observed, whereas‘hypomethylation’ of KvDMR1 is seen. As previously noted byBliek et al. (25), the MIs generated at each locus are skewed tothe same extent, but in opposite directions by the increasedcontribution of paternal alleles. Thus, for cases of UPD of11p15, the combined MIs for H19 and KvDMR1 approachunity. Furthermore, the degree of methylation distortion isproportional to the level of mosaicism in the tissue beinganalyzed. Whereas methylation assays for cases of UPD for11p15 will be skewed, transcription assays should demonstratethe integrity of the cis-acting imprinting regulatory mech-anisms of 11p15. As anticipated, we have shown in this work,despite hypomethylation at KvDMR1, monoallelic paternaltranscription of KCNQ1OT1 is maintained in cases of 11p15UPD. However, since we assayed the allelic contributions byPCR, we have not assessed quantitatively the steady-statelevels of KCNQ1OT1 RNA. These may be elevated by thepresence of UPD for 11p15 (Fig. 7).

The frequency of 11p15 UPD in our series was 17%. Thisis in keeping with data obtained by other groups(18,19,25,26,29,33–36). However, our demonstration of twocases where UPD was found in skin fibroblasts from one sideof the body but not the other suggests that 11p15 UPD mayexhibit tissue specificity or mosaicism and be more frequentthan the literature reports, since analysis is usually limited toblood.

The tumor risk (29%) associated with 11p15 UPD in ourpatient series is comparable to that reported by other investiga-tors (Table 3). However, because the tumor types observed inour study group were more diverse than in most previousstudies, we were able to demonstrate a specific associationbetween 11p15 UPD and the development of Wilms’ tumor.This association between Wilms’ tumorigenesis anddysregulation of the 11p15 telomeric imprinted domain paral-lels a body of molecular data generated for sporadic Wilms’

tumor (30). That is, maternal-specific loss of heterozygosity in11p15 has been observed often in Wilms’ and other embryonaltumors, involving a region of at least 800 kb that is proposed toharbor the WT2 tumor susceptibility gene (37). Specificpatterns of acquired gene dysregulation in tumors associatedwith the development of sporadic Wilms’ tumor indicate a rolefor somatic dysregulation of the 11p15 imprinted telomericdomain during tumorigenesis (38).

H19 methylation errors in BWS patients

This group of BWS cases demonstrates methylation of thenormally unmethylated maternal H19 gene as a result of anepigenetic alteration in BWSIC1, the putative imprintingcenter upstream of H19. Concomitant changes in allelic tran-scription have been reported, with loss of maternal H19expression and biallelic IGF2 expression (30). In our study,one of three patients with this epigenotype developed a tumor,a finding supported by a series of studies (23,25,26,33). As formost of the 11p15 UPD cases, the tumors reported by othershave been exclusively Wilms’ tumors. In fact, H19 hypermethy-lation is a well-established epigenotype associated withsporadic Wilms’ tumor. Usually, hypermethylation anddecreased expression of the H19 gene with concomitant loss ofimprinting of the maternal allele of the IGF2 gene are seen(30,39,40). Thus parallel changes in epigenotype occur in thisclass of BWS patients and in sporadic Wilms’ tumors.

Imprinting defects of KvDMR1/KCNQ1OT1 in BWS patients

The differentially methylated region (DMR), KvDMR1, hasbeen postulated to represent BWSIC2, the imprinting centerfor domain 2. An imprinting defect at this locus results in lossof methylation of the maternal KvDMR1 region. To date, suchcases with loss of maternal methylation at KvDMR1 have beenshown to exhibit concomitant expression of the normallysilenced maternal KCNQ1OT1 in only a small number (n = 8)of patients. In keeping with a model in which regulation of thiscentromeric imprinted domain is an independent process,aberrant imprinting at both KvDMR1 and KCNQ1OT1 was notaccompanied by abnormal methylation of the H19 gene.

Table 3. Tumor frequency in BWS molecular subgroups

aCases screened for 11p15 UPD, H19 hypermethylation, CDKNIC mutations and KCNQ1OT1 imprinting defects.bThe data from the literature were obtained from Henry et al. (19), Hatada et al. (20,41), Lam et al. (21), Reik et al. (23), Bliek et al. (25), Engel et al. (26), Leeet al. (28), Gaston et al. (29), Catchpoole et al. (33), Slatter et al. (34), Schneid et al. (35), Dutly et al. (36), O’Keefe et al. (42) and Okamoto et al. (43). SomeBWS cases were represented in more than one publication. They were counted only once in this Table.

Telomeric domain Centromeric domain No. of tumors in BWS cases with no identified molecular defecta

No. of tumors in BWS cases with UPD of 11p15

BWS cases with H19 hypermethylation

No. of tumors in BWS cases with CDKNIC mutations

No. of tumors in BWS cases with KCNQ1OT1 imprinting defects

This work 6/21 1/3 0/5 5/35 4/17

Data from literatureb 13/69 5/14 2/33 1/88 2/10

Subgroup totals 19/90 6/17 2/38 6/118 6/27

Domain totals 25/107 8/156


The frequency of KvDMR1 defects in our series of BWSpatients is 61%, which is comparable to that in the literature.We found 11/36 cases with complete demethylation and 25/36cases with partial demethylation. These data are consistentwith those of Bliek et al. (25). Moreover, we analyzed allelictranscription for KCNQ1OT1 in these patients and found that31 (46%) of 67 cases had biallelic expression. Those withbiallelic expression all had partial or complete demethylation(MI ≤ 0.42), whereas patients with normal monoallelic expres-sion had normal methylation unless UPD for 11p15 waspresent. These data substantially increase the previous numberof cases studied for both KvDRM1 methylation status andKCNQ1OT1 transcription (9). Our data support the contentionthat there is excellent correlation between KvDMR1 methyla-tion status and allelic transcription of KCNQ1OT1.

We have shown that BWS individuals with constitutionalimprinting defects at KvDMR1/KCNQ1OT1 do indeed developtumors. The tumors observed include hepatoblastoma, rhab-domyosarcoma and gonadoblastoma. What is of particularinterest, in our series, is that the rarer types of tumors associ-ated with BWS (rhabdomyosarcoma and gonadoblastoma) areseen in BWS cases with constitutional imprinting defects atKvDMR1/KCNQ1OT1 but usually not in BWS cases with UPDof 11p15 or methylation errors of H19. Moreover, the mostcommon tumor observed in BWS cases, Wilms’ tumor was notseen in patients with KvDMR1/KCNQ1OT1 imprinting errors.This profile of tumor specificity is consistent with availabledata from molecular analyses of sporadic tumors. For example,imprinting defects of the KCNQ1OT1 transcript or KvDMR1have not been observed in sporadic Wilms’ tumor (16,30). Todate, no data have been reported on the imprinting status ofKCNQ1OT1 or KvDMR1 in non-Wilms’ embryonal tumors.

Normal imprinting status of 11p15 in BWS patients

Patients with normal imprinting status of 11p15 (group V)have normal biparental inheritance of 11p15 markers, normal

methylation at H19 and KvDMR1 and normal monoallelicexpression of KCNQ1OT1. The frequency of tumors for thisgroup in our study is 24%, comparable to 20% in a previousstudy (25). In addition, all tumors reported to date in this groupare Wilms’ tumors. These data, both in terms of frequency andtumor type, are very similar to those seen for BWS moleculardefects in the telomeric domain. Given that hemihyperplasia isa frequent finding in group V BWS cases with Wilms’ tumorand that hemihyperplasia is a common finding in BWS caseswith 11p15 UPD, we propose that such individuals in group Vmay represent cryptic cases of 11p15 UPD. Since UPD for11p15 usually occurs as a somatic mosaic event, one wouldpredict that sampling of multiple tissues from group V BWSpatients could potentially demonstrate whether UPD for 11p15 isin fact the underlying molecular defect in at least some group Vcases.

Tumor rates in different BWS molecular groups

Children with BWS have a greatly elevated risk of developinga wide variety of pediatric embryonal tumors (1,5,8). Our datashow that all four molecular groups are represented in theBWS cases with tumors. We have also reviewed the tumorrates defined for each molecular group from data in the litera-ture. Previous studies that have assessed tumor rates in specificBWS subgroups have concluded that tumors are associatedwith only two molecular subgroups UPD for 11p15 and H19hypermethylation. However, such studies have had limitedrepresentation of tumors other than Wilms’ tumor and occa-sionally hepatoblastoma. Two of the largest studies are thoseof Engel et al. (26) with three Wilms’ tumors and Bliek et al.(25) with seven Wilms’ tumors. Only a recent report by Gastonet al. (29) included a broader tumor spectrum and in this study,one BWS case with an imprinting defect at KvDMR1 wasnoted to have thyroid carcinoma. Therefore, given thatimprinting defects at KCNQ1OT1/KvDMR1 are associatedwith tumors other than Wilms’ tumor, it is not surprising that

Table 4. Tumor types in BWS molecular subgroups

aCases screened for 11p15 UPD, H19 hypermethylation CDKNIC mutations and KCNQ1OT1 imprinting defects.bThe data from the literature were obtained from Henry et al. (19), Hatada et al. (20,41), Lam et al. (21), Reik et al. (23), Bliek et al. (25), Engel et al. (26), Leeet al. (28), Gaston et al. (29), Catchpoole et al. (33), Slatter et al. (34), Schneid et al. (35), Dutly et al. (36), O’Keefe et al. (42) and Okamoto et al. (43). SomeBWS cases were represented in more than one publication. They were counted only once in this Table.

Telomeric domain Centromeric domain Tumor types in BWS cases with no identified molecular defecta

Tumor types in BWScases with UPD of 11p15

Tumor types in BWScases with hypermethylation ofH19

Tumor types in BWS cases with CDKNIC mutation

Tumor types in BWS cases with KCNQ1OT1imprinting defects

This work Five Wilms’ tumors,one hepatoblastoma

One Wilms’ tumor – Two rhabdomyosarcomas,two hepatoblastomas,one gonadoblastoma

Four Wilms’ tumors

Data from literatureb

Eleven Wilms’ tumors, one neuroblastoma, one pheochromocytoma

Five Wilms’ tumors Two neuroblastomas One thyroid carcinoma Two Wilms’ tumors

Totals Sixteen Wilms’ tumors, one hepatoblastoma, one neuroblastoma, one pheochromocytoma

Six Wilms’ tumors Two neuroblastomas Two rhabdomyosarcomas, two hepatoblastomas,one gonadoblastoma,one thyroid carcinoma

Six Wilms’ tumors


investigators have not previously noted the tumor risk forBWS cases with imprinting defects at KCNQ1OT1.

Model for tumor development in BWS

From our data, we propose that specific molecular defects inthe telomeric H19 and IGF2 domain have a role in Wilms’tumorigenesis, whereas the KCNQ1OT1/KvDMR1 centro-meric domain is associated with predisposition to other typesof embryonal tumors in BWS. Since UPD for 11p15 generallyinvolves both imprinted domains, it is important to define thefeatures which distinguish 11p15 UPD from primaryimprinting alterations in the centromeric domain (Fig. 7). Inthe telomeric domain, it is likely that the critical lesion both forBWS and Wilms’ tumorigenesis is extinction of the H19 geneproduct via either UPD for 11p15 or methylation of thematernal H19 region (38,39).

We propose that the critical lesion associated with tumordevelopment in the centromeric domain is loss of imprintedgene regulation in cis on the maternally derived chromosome.We expect that this would lead to dysregulation of one or moreas yet unidentified downstream targets important for thedevelopment of rhabdomyosarcoma, gonadoblastoma andhepatoblastoma. Although CDKNIC is a plausible candidatedownstream target, only two tumors, both neuroblastoma, havebeen reported to date in approximately 38 BWS cases withmutations in CDKNIC (20,21,28,31,41–43). Thus, the keyelements in the centromeric domain relevant to non-Wilms’tumorigenesis remain to be identified. In this regard, other

potential downstream targets of KvDMR1 would be excellentcandidate genes.

Whereas it is evident that there are at least two genetic path-ways leading to tumors in BWS, the finding of biallelic IGF2expression in some patients with imprinting defects atKCNQ1OT1/KvDMR1 (9) suggests that these pathways maybe able to cooperate in tumorigenesis. Loss of imprinting forthe IGF2 gene is reported to occur in a large variety of tumors(30). In Wilms’ tumor it may be associated with either H19hypermethylation or it may occur without any concomitantidentified alteration on 11p15. Constitutional loss of imprint ofthe maternal IGF2 gene has also been reported for sporadiccases of BWS, rarely in association with hypermethylation ofthe H19 gene, and more commonly without another identified11p15 defect. Since BWS cases with imprinting defects atKCNQ1OT1/KvDMR1 can show either loss of imprinting ornormal imprinting at IGF2 (9), imprinting defects at IGF2 maywell be part of the tumorigenic pathways controlled by both thecentromeric and telomeric imprinted domains of 11p15.

In summary, we have broadened the scope of epigeneticmolecular changes on 11p15 documented for BWS patientswho have developed tumors and have defined new associationsbetween specific molecular lesions and the tumor types seen inthese patients. These data further elucidate some of thecomplex genetic pathways leading to BWS tumorigenesis,demonstrating that BWS provides a unique model system tostudy the networks of interacting genetic pathways subject to

Figure 7. Role of methylation at KvDMR1 in KCNQ1OT1 expression. (A) The KCNQ1OT1 gene (gray box) is normally methylated (CH3) on the maternal alleleand expression of KCNQ1OT1 occurs only from the paternal allele. Paternal and maternal contributions are denoted by P or M, respectively. (B) Patients with UPDhave two copies of the paternal allele but are mosaic and have a varying proportion of cells that are normal. Expression (increased) of KCNQ1OT1 and methylation(decreased) at KvDMR1 is predicted to be altered in proportion to the percentage of paternal UPD cells in the tissue studied. (C) In patients with KvDMR1 imprint-ing defects partial or complete loss of methylation of the maternal allele results in expression from both the paternal and normally silent maternal allele. TheKCNQ1OT1 transcript is represented as a wavy line and is either paternal or maternal in origin as indicated by P or M, respectively. Transcripts can be distin-guished by RFLPs represented in this Figure by A (paternal) or B (maternal).


constitutional epigenetic alteration that influence tumor risk inchildren.

MATERIALS AND METHODS

Patient material

One hundred and sixty-two patients carrying a diagnosis ofBWS were referred for molecular testing. Following review ofthe clinical features, we confirmed a clinical diagnosis of BWSto 125 patients based on the presence of at least three of thefollowing features: macrosomia, macroglossia, hemihyper-plasia, ear creases/pits, abdominal wall defect (omphalocele orumbilical hernia), embryonal tumor; or, at least two of thepreceding features plus one of neonatal hypoglycemia, abdom-inal organomegaly or renal malformation. Eight of the 125patients had positive family histories. The number of child-hood tumors reported in this clinically defined group of BWScases was 16 (12.8%) and included a broad range of tumortypes, specifically Wilms’ tumor, hepatoblastoma, rhabdomyo-sarcoma and gonadoblastoma.

Control samples were obtained from normal controlindividuals as well as individuals with balanced chromosomerearrangement not involving chromosome 11.

DNA or RNA was obtained from patient samples foranalyses of chromosome 11 UPD, methylation status at H19and KvDMR, and for KCNQ1OT1 allelic expression. Thesestudies were approved by the Research Ethics Board of theHospital for Sick Children, Toronto, Canada.

Cell cultures

Lymphoblast lines were maintained in RPM1 1640 mediasupplemented with 15% fetal calf serum. Fibroblast strainswere maintained for fewer than 10 passages in α-MEM supple-mented with 10% fetal calf serum.

Chromosome 11 UPD analysis using quantitative PCR

Genomic DNA was extracted from either peripheral blood,cultured lymphoblasts or skin fibroblast cells from the probandand parents using a QIAamp spin column (QIAGEN) method(as per manufacturer’s instructions). Quantitative multiplex-PCR using highly polymorphic STR markers was performedusing three markers within (TH, D11S2362 and D11S1997)and two markers distal (D11S1998 and D11S1974) to theBWS critical region at 11p15.5 in order to detect somatic cellrearrangements giving rise to paternal UPD of the chromo-some 11p15 region. Amplification products were separated ona 6% denaturing polyacrylamide gel on an ABI 377 PRISMDNA Sequencer (PE Biosystems, Boston, MA). The allelesizes and corresponding peak areas were determined usingGenescan software. The percentage of paternal UPD of allelesat 11p15.5 in the proband was determined from informativealleles at a minimum of two DNA markers within the BWScritical region showing an increase in dosage of >20% based onthe following calculation: (Peak area of paternal allele – Peak areaof maternal allele) / (Peak area of paternal allele + Peak area ofmaternal allele).

Analysis of allele-specific KCNQ1OT1 expression

Total RNA was isolated from lymphoblasts or fibroblastsusing the TRIZOL® reagent (Gibco BRL, Burlington, Ontario,Canada). mRNA was extracted from total RNA using a Quick-Prep™ Micro mRNA purification kit (Amersham PharmaciaBiotech, Little Chalfont, UK). M-MuLV reverse-transcriptase(MBI Fermentas, Burlington, Ontario, Canada) was used forreverse transcription reactions. Screening for contamination bygenomic DNA was carried out in an identical tube withoutreverse transcriptase. RT–PCR products were gel purified andsequenced to determine allele-specific expression.

Southern blot analysis of KvDMR1 and H19 methylation

For analysis of methylation of KvDMR1, genomic DNA wasdigested with EcoRI and NotI; for analysis of methylation ofthe H19 gene, genomic DNA was digested with PstI and SmaI.For both methylation assays, digestion products were electro-phoresed through 0.8% agarose gels and were then transferredto a GeneScreen Plus membrane (NEN, Boston, MA). Foranalysis of methylation of KvDMR1, blots were hybridizedwith the 400 bp [α32P]dCTP-labeled DMR probe, a kind giftfrom M.Higgins (15). For analysis of H19 methylation, blotswere hybridized with the H19 promoter probe described byReik et al. (32). The probes were labeled using the randomprimed DNA labeling kit (Roche, Mannheim, Germany). Theblots were analyzed using a Molecular Dynamics Storm Phos-phorImager. The H19 MI was determined by dividing theoptical density of the 1.8 kb band by the combined densities ofthe 1.0 and 1.8 kb bands. For KvDMR1 the MI was determinedby dividing the optical density of the 4.2 kb band by thecombined densities of the 4.2 and 2.7 kb bands.

CDKNIC mutation screening

The entire coding region of CDKNIC was analyzed by PCR/heteroduplex screening as previously described by Li et al.(31).

ACKNOWLEDGEMENTS

We thank all the families who participated in this study. Wewould like to acknowledge the expert statistical input ofDr Gideon Koren and the excellent secretarial assistance ofNancy Taylor and Sarah Petty. The KvDMR (DMR) probe wasgenerously provided by Dr M.Higgins. This research wassupported by the National Cancer Institute of Canada withfunds from the Canadian Cancer Society. Ingrid Ambus wasthe recipient of a Starbucks Clinical Genetics ResearchStudentship Award.

REFERENCES

1. Pettenati, M.J., Haines, J.L., Higgins, R.R., Wappner, R.S., Palmer, C.G. and Weaver, D.D. (1986) Wiedemann–Beckwith syndrome: presentation of clinical and cytogenetic data on 22 new cases and review of the literature. Hum. Genet., 74, 143–154.

2. Li, M., Squire, J. and Weksberg, R. (1998) Molecular genetics of Wiedemann–Beckwith syndrome. Am. J. Med. Genet., 79, 253–259.

3. Maher, E.R. and Reik, W. (2000) Beckwith–Wiedemann syndrome: imprinting in clusters revisited. J. Clin. Invest., 105, 247–252.

4. Reik, W. and Murrell, A. (2000) Genomic imprinting. Silence across the border. Nature, 405, 408–409.


5. Wiedemann, H. (1983) Tumors and hemihypertrophy associated with the Wiedemann–Beckwith syndrome. Eur. J. Pediatr., 141, 129.

6. DeBaun, M.R. and Tucker, M.A. (1998) Risk of cancer during the first four years of life in children from The Beckwith–Wiedemann Syndrome Registry. J. Pediatr., 132, 398–400.

7. Sotelo-Avila, C., Gonzalez-Crussi, F. and Fowler, J.W. (1980) Complete and incomplete forms of Beckwith–Wiedemann syndrome: their oncogenic potential. J. Pediatr., 96, 47–50.

8. Schneid, H., Vazquez, M.P., Vacher, C., Gourmelen, M., Cabrol, S. and Le Bouc, Y. (1997) The Beckwith–Wiedemann syndrome phenotype and the risk of cancer. Med. Pediatr. Oncol., 28, 411–415.

9. Lee, M.P., DeBaun, M.R., Mitsuya, K., Galonek, H.L., Brandenburg, S., Oshimura, M. and Feinberg, A.P. (1999) Loss of imprinting of a paternally expressed transcript, with antisense orientation to KVLQT1, occurs frequently in Beckwith–Wiedemann syndrome and is independent of insulin-like growth factor II imprinting. Proc. Natl Acad. Sci. USA, 96, 5203–5208.

10. Lee, M., Brandenburg, S., Landes, G., Adams, M., Miller, G. and Feinberg, A. (1999) Two novel genes in the center of the 11p15 imprinted domain escape genomic imprinting. Hum. Mol. Genet., 8, 683–690.

11. Brannan, C.I., Dees, E.C., Ingram, R.S. and Tilghman, S.M. (1990) The product of the H19 gene may function as an RNA. Mol. Cell. Biol., 10, 28–36.

12. Hark, A.T., Schoenherr, C.J., Katz, D.J., Ingram, R.S., Levorse, J.M. and Tilghman, S.M. (2000) CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus. Nature, 405, 486–489.

13. Caspary, T., Cleary, M.A., Baker, C.C., Guan, X.J. and Tilghman, S.M. (1998) Multiple mechanisms regulate imprinting of the mouse distal chromosome 7 gene cluster. Mol. Cell. Biol., 18, 3466–3474.

14. Kanduri, C., Holmgren, C., Pilartz, M., Franklin, G., Kanduri, M., Liu, L., Ginjala, V., Ulleras, E., Mattsson, R. and Ohlsson, R. (2000) The 5′ flank of mouse H19 in an unusual chromatin conformation unidirectionally blocks enhancer–promoter communication. Curr. Biol., 10, 449–457.

15. Smilinich, N.J., Day, C.D., Fitzpatrick, G.V., Caldwell, G.M., Lossie, A.C., Cooper, P.R., Smallwood, A.C., Joyce, J.A., Schofield, P.N., Reik, W. et al. (1999) A maternally methylated CpG island in KvLQT1 is associated with an antisense paternal transcript and loss of imprinting in Beckwith–Wiedemann syndrome. Proc. Natl Acad. Sci. USA, 96, 8064–8069.

16. Mitsuya, K., Meguro, M., Lee, M.P., Katoh, M., Schulz, T.C., Kugoh, H., Yoshida, M.A., Niikawa, N., Feinberg, A.P. and Oshimura, M. (1999) LIT1, an imprinted antisense RNA in the human KvLQT1 locus identified by screening for differentially expressed transcripts using monochromosomal hybrids. Hum. Mol. Genet., 8, 1209–1217.

17. Horike, S., Mitsuya, K., Meguro, M., Kotobuki, N., Kashiwagi, A., Notsu, T., Schulz, T.C., Shirayoshi, Y. and Oshimura, M. (2000) Targeted disruption of the human LIT1 locus defines a putative imprinting control element playing an essential role in Beckwith–Wiedemann syndrome. Hum. Mol. Genet., 9, 2075–2083.

18. Henry, I., Bonaiti-Pellie, C., Chehensse, V., Beldjord, C., Schwartz, C., Utermann, G. and Junien, C. (1991) Uniparental paternal disomy in a genetic cancer-predisposing syndrome. Nature, 351, 665–667.

19. Henry, I., Puech, A., Riesewijk, A., Ahnine, L., Mannens, M., Beldjord, C., Bitoun, P., Tournade, M.F., Landrieu, P. and Junien, C. (1993) Somatic mosaicism for partial paternal isodisomy in Wiedemann–Beckwith syndrome: a post-fertilization event. Eur. J. Hum. Genet., 1, 19–29.

20. Hatada, I., Ohashi, H., Fukushima, Y., Kaneko, Y., Inoue, M., Komoto, Y., Okada, A., Ohishi, S., Nabetani, A., Morisaki, H. et al. (1996) An imprinted gene p57KIP2 is mutated in Beckwith–Wiedemann syndrome. Nat. Genet., 14, 171–173.

21. Lam, W.W., Hatada, I., Ohishi, S., Mukai, T., Joyce, J.A., Cole, T.R., Donnai, D., Reik, W., Schofield, P.N. and Maher, E.R. (1999) Analysis of germline CDKN1C (p57KIP2) mutations in familial and sporadic Beckwith–Wiedemann syndrome (BWS) provides a novel genotype–phenotype correlation. J. Med. Genet., 36, 518–523.

22. Weksberg, R., Shen, D.R., Fei, Y.L., Song, Q.L. and Squire, J. (1993) Disruption of insulin-like growth factor 2 imprinting in Beckwith–Wiedemann syndrome. Nat. Genet., 5, 143–150.

23. Reik, W., Brown, K.W., Schneid, H., Le Bouc, Y., Bickmore, W. and Maher, E.R. (1995) Imprinting mutations in the Beckwith–Wiedemann syndrome suggested by altered imprinting pattern in the IGF2-H19 domain. Hum. Mol. Genet., 4, 2379–2385.

24. Joyce, J.A., Lam, W.K., Catchpoole, D.J., Jenks, P., Reik, W., Maher, E.R. and Schofield, P.N. (1997) Imprinting of IGF2 and H19: lack

of reciprocity in sporadic Beckwith–Wiedemann syndrome. Hum. Mol. Genet., 6, 1543–1548.

25. Bliek, J., Maas, S.M., Ruijter, J.M., Hennekam, R.C., Alders, M., Westerveld, A. and Mannens, M.M. (2001) Increased tumour risk for BWS patients correlates with aberrant H19 and not KCNQ1OT1 methylation: occurrence of KCNQ1OT1 hypomethylation in familial cases of BWS. Hum. Mol. Genet., 10, 467–476.

26. Engel, J.R., Smallwood, A., Harper, A., Higgins, M.J., Oshimura, M., Reik, W., Schofield, P.N. and Maher, E.R. (2000) Epigenotype–phenotype correlations in Beckwith–Wiedemann syndrome. J. Med. Genet., 37, 921–926.

27. Tycko, B. (2000) Epigenetic gene silencing in cancer. J. Clin. Invest., 105, 401–407.

28. Lee, M.P., DeBaun, M., Randhawa, G., Reichard, B.A., Elledge, S.J. and Feinberg, A.P. (1997) Low frequency of p57KIP2 mutation in Beckwith–Wiedemann syndrome. Am. J. Hum. Genet., 61, 304–309.

29. Gaston, V., Le Bouc, Y., Soupre, V., Burglen, L., Donadieu, J., Oro, H., Audry, G., Vazquez, M.P. and Gicquel, C. (2001) Analysis of the methylation status of the KCNQ1OT and H19 genes in leukocyte DNA for the diagnosis and prognosis of Beckwith–Wiedemann syndrome. Eur. J. Hum. Genet., 9, 409–418.

30. Tycko, B. (2000) Genomic imprinting and human neoplasia. In Ehrlich, M. (ed.), DNA Alterations in Cancer. Eaton Publishing, Natwick, MA, pp. 333–349.

31. Li, M., Squire, J., Shuman, C., Atkin, J., Pauli, R., Smith, A., Chitayat, D. and Weksberg, R. (2001) Imprinting status of 11p15 genes in Beckwith–Wiedemann syndrome patients with CDKN1C mutations. Genomics., 74, 370–376.

32. Reik, W., Brown, K.W., Slatter, R.E., Sartori, P., Elliott, M. and Maher, E.R. (1994) Allelic methylation of H19 and IGF2 in the Beckwith–Wiedemann syndrome. Hum. Mol. Genet., 3, 1297–1301.

33. Catchpoole, D., Lam, W.W., Valler, D., Temple, I.K., Joyce, J.A., Reik, W., Schofield, P.N. and Maher, E.R. (1997) Epigenetic modification and uniparental inheritance of H19 in Beckwith–Wiedemann syndrome. J. Med. Genet., 34, 353–359.

34. Slatter, R.E., Elliott, M., Welham, K., Carrera, M., Schofield, P.N., Barton, D.E. and Maher, E.R. (1994) Mosaic uniparental disomy in Beckwith–Wiedemann syndrome. J. Med. Genet., 31, 749–753.

35. Schneid, H., Seurin, D., Vazquez, M.P., Gourmelen, M., Cabrol, S. and Le Bouc, Y. (1993) Parental allele specific methylation of the human insulin-like growth factor II gene and Beckwith–Wiedemann syndrome. J. Med. Genet., 30, 353–362.

36. Dutly, F., Baumer, A., Kayserili, H., Yuksel-Apak, M., Zerova, T., Hebisch, G. and Schinzel, A. (1998) Seven cases of Wiedemann–Beckwith syndrome, including the first reported case of mosaic paternal isodisomy along the whole chromosome 11. Am. J. Med. Genet., 79, 347–353.

37. Karnik, P., Paris, M., Williams, B.R., Casey, G., Crowe, J. and Chen, P. (1998) Two distinct tumor suppressor loci within chromosome 11p15 implicated in breast cancer progression and metastasis. Hum. Mol. Genet., 7, 895–903.

38. Schofield, P.N., Joyce, J.A., Lam, W.K., Grandjean, V., Ferguson-Smith, A., Reik, W. and Maher, E.R. (2001) Genomic imprinting and cancer; new paradigms in the genetics of neoplasia. Toxicol. Lett., 120, 151–160.

39. Dao, D., Walsh, C.P., Yuan, L., Gorelov, D., Feng, L., Hensle, T., Nisen, P., Yamashiro, D.J., Bestor, T.H. and Tycko, B. (1999) Multipoint analysis of human chromosome 11p15/mouse distal chromosome 7: inclusion of H19/IGF2 in the minimal WT2 region, gene specificity of H19 silencing in Wilms’ tumorigenesis and methylation hyper-dependence of H19 imprinting. Hum. Mol. Genet., 8, 1337–1352.

40. Steenman, M., Westerveld, A. and Mannens, M. (2000) Genetics of Beckwith–Wiedemann syndrome-associated tumors: common genetic pathways. Genes Chromosomes Cancer, 28, 1–13.

41. Hatada, I., Nabetani, A., Morisaki, H., Xin, Z., Ohishi, S., Tonoki, H., Niikawa, N., Inoue, M., Komoto, Y., Okada, A. et al. (1997) New p57KIP2 mutations in Beckwith–Wiedemann syndrome. Hum. Genet., 100, 681–683.

42. O’Keefe, D., Dao, D., Zhao, L., Sanderson, R., Warburton, D., Weiss, L., Anyane-Yeboa, K. and Tycko, B. (1997) Coding mutations in p57KIP2 are present in some cases of Beckwith–Wiedemann syndrome but are rare or absent in Wilms tumors. Am. J. Hum. Genet., 61, 295–303.

43. Okamoto, K., Morison, I.M., Reeve, A.E., Tommerup, N., Wiedemann, H.R. and Friedrich, U. (1998) Is p57KIP2 mutation a common mechanism for Beckwith–Wiedemann syndrome or somatic overgrowth? J. Med. Genet., 35, 86.

tumor development in the beckwithâ€“wiedemann syndrome is

Documents