Editor—Silver-Russell syndrome (SRS) is a condition characterised by pre- and postnatal growth restriction, triangular facies, and limb and truncal asymmetry.1 2 The aetiology of the syndrome is heterogeneous and there is no clearly established Mendelian basis. A number of chromosomal abnormalities are associated with the SRS phenotype in a minority of cases. To date, 37 cases of maternal uniparental disomy for chromosome 7 (mUPD(7)) have been reported, representing approximately 10% of all cases.3-18

Five mUPD(7) probands were found to show heterodisomy for the complete length of chromosome 7, ruling out the exposure of a mutant recessive gene as the basis of SRS.19 These findings indicate that one or more genes on chromosome 7 are imprinted and involved in the pathogenesis of the syndrome. Lack of a paternal gene(s) expression could result in the syndrome, as could the overexpression of a maternal gene involved in growth inhibition. Recently, two unrelated SRS probands with maternal duplications of 7p have been reported defining a candidate gene region that may contain the gene(s) responsible for the phenotype associated with mUPD(7).20 21 In addition, a single case of partial mUPD(7) has been reported with biparental inheritance for the majority of the chromosome, with mUPD(7) 7q32-qter when analysed using fluorescent microsatellite PCR.22 This indicates that there are at least two imprinted regions on chromosome 7 that are involved in the pathogenesis of SRS.

It is possible that certain features of the mUPD(7) SRS phenotype, particularly the asymmetry, could result from an undetected mosaicism for chromosome 7. This may be analogous to mosaicism for pUPD(11), which is thought to be the cause of hemihypertrophy seen in Beckwith-Wiedemann syndrome (BWS).23 24

It has clearly been established that the UPD in SRS and BWS arise by different mechanisms. All reported BWS cases of pUPD(11) are isodisomic, resulting in a UPD/normal cell mosaic, most likely occurring from a postzygotic mitotic non-disjunction. In contrast, most reported cases of mUPD(7) SRS probands exhibit mixed hetero/isodisomy, and are most likely to arise from meiosis I non-disjunction followed by trisomic rescue owing to heterodisomy at the centromeres. Chromosomal mosaicism has previously been ruled out in the two cases of SRS presenting with cystic fibrosis, the latter resulting from unmasking of the maternal recessive CFTR mutation by isodisomy.3 4

Any mosaicism in the cases of heterodisomy would involve both trisomic and UPD cell lines. Four cases of mUPD(7) heterodisomy have been shown to have arisen from trisomic rescue, with mosaicism for trisomy 7 confined to the placenta.7 25 Evidence for an association between SRS and potential mosaicism has been found in a case of maternal long arm and paternal short arm isodisomy for chromosome 7.26 There has also been one reported case of an asymmetrical mUPD(7) SRS proband with a mosaic partial paternal ring chromosome 7.16 However the function of this ring has not been established.27

To test the hypothesis that the SRS phenotype may result from low level mosaicism in the cases of mUPD(7), we used a similar Southern blot strategy as originally reported by Henry et al 23 to show mosaicism in pUPD(11) BWS to investigate three SRS probands with mUPD(7). High molecular weight DNA was extracted from lymphocytes and fibroblasts for two of the mUPD(7) SRS probands and one non-mUPD(7) SRS proband with lateral asymmetry. In addition, DNA was extracted from lymphocytes for the third mUPD(7) SRS and all parents (probands previously described by Preeceet al,11 Preeceet al,19 and Russoet al 18). The two heterodisomic mUPD(7) SRS probands reported by Preece et al 11 19 were growth restricted but were not asymmetrical. The heterodisomic mUPD(7) SRS proband reported by Russoet al 18 presented with both growth restriction and lateral asymmetry. Six μg of DNA was digested with 30 units of Hinf1 (Promega) for six hours at 37°C, electrophoresed on 0.8% agarose gel overnight, then blotted onto a Hybond N+ filter (Amersham). Filters were hybridised with the D7S22 VNTR probe (PIC=0.97) to detect the inheritance of parental alleles. DNA probes were radiolabelled with [α³² P] dCTP using a nick translation kit (Amersham). Hybridisation was carried out overnight and the filters were washed down with increasing stringency to 0.5 × SSC/0.1% SDS and exposed to autoradiographic film (Kodak) for three weeks at −80°C so that the presence of any paternal bands could be detected.

Fluorescence in situ hybridisation (FISH) was performed on cultured lymphoblasts and fibroblasts for one mUPD(7) SRS proband and one non-mUPD(7) SRS proband (probands 2 and 4) using standard methods. The cells were probed with chromosome 7p specific PAC probes containingGRB10 (RP1-108E23) andEGFR (RP1-069I12). Standard miniprepped DNA was nick translated with the direct incorporation of either Spectrum Green or Spectrum Red (Vysis) following the manufacturer's protocol.

Using the 7q36 D7S22 VNTR DNA marker, three mUPD(7) probands and one non-mUPD(7) SRS control proband and both parents were genotyped by Southern blot analysis. In the three mUPD(7) probands, direct examination with D7S22 from overexposed Southern blots failed to disclose the presence of paternal alleles. Since the Southern blot technique can only detect mosaicism as low as 10-15% (unpublished observation), FISH was performed on single nuclei to detect lower levels of mosaicism, in preference to fluorescence PCR reported by Slater et al.24 FISH investigations of 200 nuclei also failed to show any potential mosaicism.

To overcome the problem that studies solely on lymphocyte DNA would not rule out tissue specific mosaicism, fibroblasts were also studied where possible. No evidence for mosaicism was found in either cell line in any of the mUPD(7) SRS probands studied. Studies using skeletal related tissues would have been preferred, but were unavailable.

As it is likely that the mUPD arose from trisomic rescue, it is probable that their placentas were mosaic. The possibility that CPM may influence the phenotypic prenatal growth failure has previously been addressed. Kalousek et al 28reported a study of 14 pregnancies with CPM for trisomy 7. DNA analysis in nine cases showed that the fetal UPD was present in only one case whereas the other eight were biparental. One trisomy resulted from meiosis non-disjunction and was associated with the fetal UPD. The infant with the UPD presented with severe IUGR. The other eight cases with biparental inheritance had normal birth weights. This provides evidence that intrauterine growth is not affected by the presence of a trisomic cell line for chromosome 7 in the placenta, therefore making it unlikely that CPM is responsible for the significant growth restriction seen in probands with mUPD(7).

The findings we report here encourage continued research into the molecular basis of SRS as a consequence of disrupted imprinting. We have presented evidence against somatic mosaicism in three mUPD(7) probands in two different tissues using both Southern blot and FISH analyses. Comparisons were also made of a case of non-mUPD(7) SRS with lateral asymmetry and no mosaicism was observed.