Editor—Familial Wilms tumour is rare, accounting for only 1-2% of cases, and is not usually associated with other tumour types or congenital malformation.1 The pattern of inheritance has been interpreted as a dominant predisposition to Wilms tumour with incomplete penetrance of approximately 30%. Following the isolation of the WT1 gene in 1990, it soon became clear that WT1 mutation did not account for any of the large Wilms tumour pedigrees and was found in only the minority (∼10%) of sporadic Wilms tumours. Pedigree specific differences in age at onset, incidence of bilateral tumours, and presence of metastases provided further evidence for the existence of more than one gene for familial Wilms tumour.1Recently, two familial Wilms tumour gene loci, designatedFWT1 and FWT2, have been mapped by genetic linkage analysis to 17q12-21 and 19q respectively.2 3

While a series of Wilms tumour families were being analysed for linkage to the putative FWT1 locus on 17q,2 one proband was noted to carry a tumour specific 11p13 deletion (theWT1 locus) and to suffer from genitourinary abnormalities (fig 1). Mutational analysis ofWT1 was therefore undertaken in this pedigree.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Pedigree of WILMS 5. WT = Wilms tumour, C + HS = cryptorchidism and hypospadias. Subjects 201, 301, and 302 are carriers of an intragenic WT1 mutation. 101 is not a carrier. It is unknown whether the mutation arose in 201 de novo or was inherited from her father. However, none of her three sibs nor any of their eight offspring were affected by Wilms tumour.

Family WILMS 5 was identified as a sib pair affected with unilateral Wilms tumour. The sister (No 301) presented at the age of 2 years with a tumour of fetal rhabdomyomatous type and remains well 12 years later. Her brother (No 302) presented at the age of 4 years with a tumour of biphasic histology (stromal and blastemal). He relapsed shortly after finishing standard treatment and subsequently died from progressive tumour. He suffered from severe bilateral cryptorchidism and hypospadias, and surgery showed the presence of persistent Mullerian structures as well as testicular tissue. His constitutional karyotype was 46,XY whereas metaphases from the tumour displayed an interstitial 11p13 deletion and an isochromosome 1q. Both parents were phenotypically normal with no history of Wilms tumour. There was no family history of other cancers or congenital malformation.

Sequencing of exon 7 of WT1, which codes for the first zinc finger, showed a heterozygous 7 bp deletion in constitutional DNA from both affected sibs. This frameshift is predicted to lead to the introduction of 59 novel amino acids following Ser³¹⁸ with a termination codon before the zinc finger region, which is known to be critical for DNA binding by this putative transcription factor. Sequencing of the remaining exons showed no other mutations. Tumour DNA from 302 had only the mutant allele, as expected, since the tumour karyotype showed an interstitial deletion of 11p13. Tumour DNA from 301 showed an intragenic mutation of the secondWT1 allele, a 26 bp insertion also in exon 7, predicted to introduce a termination codon after two novel amino acids. Thus, in both sibs' tumours, WT1 was behaving as a tumour suppressor gene, with loss of function mutations affecting both alleles.

Sequencing of lymphocyte DNA from the mother (No 201), who is phenotypically normal, showed homozygosity for the 7 bp deletion. This was an unexpected finding, since homozygous constitutional mutation ofWT1 in a murine knock out model causes absence of the kidneys and gonads and is incompatible with postnatal survival.4 However, a single case of a Japanese girl with Denys-Drash syndrome (DDS), who is constitutionally homozygous for a missense mutation in zinc finger 2 has been described.5 6Her viability, in contrast to the prediction of the knock out model, may be because of the presence of a mutant form ofWT1, still capable of protein-protein interactions. The truncated WT1 protein produced in the WILMS 5 pedigree could also be capable of such activity through its N-terminal region. An alternative explanation is species specific differences in requirements for WT1function.

Further investigations were therefore carried out to resolve this discrepancy. DNA extracted from buccal cells and from skin fibroblasts exhibited the same homozygous genotype, ruling out the possibility of constitutional mosaicism. However, PCR analysis using a new primer set external to the original pair showed that subject 201 was in fact constitutionally heterozygous for the deletion. The apparent homozygosity was the result of an intronic point mutation, close to the 3′ end of the original reverse primer, preventing amplification of the normal allele (fig 2). This base change was also found in the maternal grandmother (subject 101). No material was available from the maternal grandfather so it was not possible to say whether the inheritedWT1 mutation had arisen de novo in subject 201 or was paternally inherited. Since primer pair 1 had been used in our previous screen for WT1 mutations in leukaemia, a small study was undertaken to see if this intronic base change was common. Constitutional DNA from 16 normal people and two with Wilms tumours was sequenced, but none was found to carry the G to T change, implying that this base change is either a rare polymorphism or is specific to family WILMS 5. This illustrates the importance of verifying homozygous constitutional mutations using a second pair of PCR primers or an independent method.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

(A) Automated sequencing in the region of the deletion found in family WILMS 5. Top panel, normal sequence, showing the 7 bp that are deleted (boxed). Centre panel, sequence from subject 201 using primer pair 1, showing apparent homozygosity for the 7 bp deletion. Lower panel, sequence from subject 201 using primer pair 2, showing heterozygosity for the 7 bp deletion. (B) Sequence of the intron downstream of exon 7 from subject 201 around the annealing site for primer 1R. The mismatch at the 3′ end of the primer is shown in bold. Mutational analysis for exons 1-10 of WT1 was performed on DNA extracted from blood, buccal cells (201 only), and tumour material (301 and 302) as previously described.11 Primer sequences for exon 7 were: pair 1 1F 5′- GACCTACGTGAATGATCACATG -3′ and 1R 5′-ACAACACCTGGATCAAGACCT -3′; pair 2 (flanking primer pair 1) 2F 5′-CACCCCTTCTTTGGATATAC -3′and 2R 5′- CTGACCTCTGTAATAAAGGAT -3′. PCR conditions for primer pair 2 (449 bp product) were 1.5 mmol/l MgCl₂, 55°C annealing. PCR products were sequenced using the ABI Prism Dye Terminator Cycle Sequencing kit with AmpliTaq FS using a modified method and capillary electrophoresed on a 310 Genetic Analyser (Perkin Elmer ABI).15 The results were analysed using Factura software and by eye to identify heterozygotes. Samples containing mutations were resequenced using a second independent PCR product as a template. Subcloning and sequencing of PCR products was carried out as previously described, where necessary.11

This is only the fourth Wilms tumour pedigree where an inheritedWT1 mutation has been identified. In the previously described pedigrees, one combined familial Wilms tumour with aniridia owing to transmission of a t(2;11)(q32;p13) with presumed complete deletion of the adjacent WT1 andPAX6 genes at 11p13.7 A father and son both had Wilms tumour associated with a constitutional 1 bp deletion in exon 6 of WT1, causing truncation of the protein before the zinc finger region.8Only the son had urogenital abnormalities. In the third family, a nonsense mutation in zinc finger 2 was transmitted by an unaffected father to three daughters all affected by Wilms tumour.9 The same mutation has also been reported in a person with Denys-Drash syndrome but there were no features of this syndrome in this pedigree.10

The mutation described in the pedigree in this report is novel, although mutations resulting in truncation ofWT1 in exons 6 or 7, just before the zinc finger region, are described in sporadic Wilms tumours and in acute myeloid leukaemia.11 In this pedigree, germlineWT1 mutation was associated with severe genitourinary malformation in the male whereas the two females had normal development; the mother is clearly fertile and the daughter has gone through normal puberty. A similar sex difference in severity of genital malformation is seen in patients with Denys-Drash syndrome, now that the syndrome has been expanded to include genotypic females with the characteristic nephropathy and Wilms tumour. This pedigree also shows the variable nature of tumour behaviour resulting from a common initiating event.

It has been suggested that there is a genotype-phenotype correlation in the histological appearance of Wilms tumours withWT1 mutations, with a predominance of differentiation into heterologous mesodermal derivatives, commonly muscle, cartilage, bone, or fat.12 Subject 301's Wilms tumour did indeed show an extreme form of heterologous differentiation towards muscle, with the fetal rhabdomyomatous subtype. However, her brother's tumour had biphasic histology (stromal and blastemal) with no heterologous differentiation. This difference may be because of the nature or timing of the second WT1 mutation during renal or tumour development. However, it is equally likely that different morphologies result from an identical initiating mutation owing to the inherent plasticity of renal development. This is clearly seen in Wilms tumours resulting from mutation in theFWT1 gene, where histologies range from classical triphasic Wilms to the fetal rhabdomyomatous subtype (Rahman and Pritchard-Jones, unpublished observations).

Germline WT1 mutations are also found in cases of Wilms tumour without a family history. The greatest number has been described in patients with DDS, but a few have been described in some patients with bilateral tumours or genitourinary abnormalities, particularly cryptorchidism, and in a small number of unilateral cases without congenital abnormalities.12-14 In a few of these cases, the mutation is inherited from an unaffected parent. Germline mutations causing truncations in exon 7 ofWT1 have been described in four patients with sporadic Wilms tumour, all of whom had normal genitourinary development.12 This is in contrast with the severe abnormalities found in subject 302 and further indicates the extreme variability in phenotypic expression of WT1mutations.

Given the excellent survival following treatment for Wilms tumour without the need for gonadotoxic chemotherapy, the rarity of familialWT1 mutation could be because of pleiotropic effects of germline WT1 on genitourinary development, causing reduced fertility in affected subjects. Constitutional WT1 mutations are more commonly found in certain congenital malformation syndromes predisposing to childhood genitourinary cancers (WAGR, Denys-Drash syndrome, and Frasier syndrome). Nevertheless, sporadic Wilms tumours occurring in conjunction with genitourinary abnormalities or showing heterologous differentiation can be the result of germlineWT1 mutations. Screening of such patients and their relatives is advisable, particularly as unaffected subjects can carry mutations that might have severe consequences for their offspring. Ten of 21 patients with Wilms tumours showing heterologous differentiation had germline WT1mutations.12 A recent report by the National Wilms Tumour Study Group found germline WT1 mutations in 7/28 Wilms tumour patients with cryptorchidism and 4/12 with hypospadias, but only 1/110 patients without genitourinary malformation.13 WT1 mutational screening can therefore probably be safely reserved for cases selected on the basis of associated abnormalities or tumour histology.