Introduction

Human chromosome 14q32.2 harbors an imprinted region with several paternally expressed genes such as DLK1 and RTL1 and maternally expressed genes such as MEG3 and RTL1as, together with the germline-derived primary DLK1-MEG3 intergenic differentially methylated region (IG-DMR) and the post fertilization-derived secondary MEG3-DMR.1, 2 Consistent with this, maternal uniparental disomy 14 (UPD(14)mat) results in clinically discernible features such as pre- and post-natal growth failure, characteristic face with prominent forehead and micrognathia, small hands, muscular hypotonia, and precocious puberty.3 These UPD(14)mat clinical features are also caused by microdeletions involving paternally derived RTL1 and/or DLK1 and by epimutation (hypomethylation) affecting the normally methylated IG-DMR and MEG3-DMR of paternal origin.2, 4, 5, 6, 7 Recently, such a clinically recognizable condition has been referred to as ‘Temple Syndrome’ (TS).8

Clinical features of TS partially overlap with those of other imprinting disorders. Indeed, pre- and post-natal growth failure, small hands, and hypotonia during early infancy are also observed in Prader–Willi Syndrome (OMIM 176270),9 and UPD(14)mat and epimutations involving the IG-DMR and the MEG3-DMR have been identified in several patients diagnosed as having Prader–Willi Syndrome.5, 7, 10 Furthermore, pre- and post-natal growth failure, prominent forehead, micrognathia, and muscular hypotonia during early infancy are often found in Silver–Russell Syndrome (SRS) (OMIM 180860).11 To our knowledge; however, UPD(14)mat has been identified only in a single patient diagnosed as having SRS with no description of detailed phenotype.12

Here, we report on epimutations of the IG-DMR and the MEG3-DMR in two patients with SRS-compatible phenotype, and discuss on phenotypic overlap between SRS and TS.

Patients and methods

Patients

We studied 85 Japanese SRS patients in whom underlying genetic factors remained unknown from our previous study for 138 SRS patients13 who satisfied the mandatory criteria and at least three of the five scoring system criteria proposed by Netchine et al14 (for details of the criteria, see footnote of Table 1). In the previous study,13 we identified H19-DMR hypomethylation (epimutation) in 43 patients (31.2%) and UPD(7)mat in nine patients (6.5%), and revealed a microdeletion at chromosome 17q24 in a single patient by analyzing copy number alterations for chromosome 11p15.5, 7p12.2, 12q14, and 17q24 that have been identified in rare SRS patients.15, 16, 17, 18

Table 1 Assessment of Silver–Russell Syndrome (SRS) clinical findings
Full size table

The 85 patients had a less-typical SRS phenotype (for details, see Fuke et al13). Indeed, of the 85 patients, none showed all of the five Netchine scoring system features, and 19 and 66 patients manifested four and three scoring system features, respectively. By contrast, of the 43 patients with H19-DMR epimutations, 10 patients were positive for all the five Netchine scoring system features, and 16 and 17 patients exhibited four and three scoring system features, respectively. This phenotypic difference was primarily due to the difference in the frequencies of relative macrocephaly at birth (35.6% vs 100%) and body asymmetry (32.2% vs 81.1%) between the two groups; the frequencies of the remaining three scoring system features were similar between the two groups. As our previous study included a large number of such patients with less-typical SRS phenotype, this would explain why the prevalence of H19-DMR epimutations was lower in our previous study than in Western European studies reported in the literature.11, 14, 19 The phenotypes of the nine UPD(7)mat patients fell between those of the 85 idiopathic SRS patients and those of the 43 epimutation-positive patients, with the frequencies of relative macrocephaly at birth and body asymmetry being 77.8% and 33.3%, respectively. This appeared to be consistent with the prevalence of UPD(7)mat being similar between our previous study and Western European studies.11, 15, 19, 20, 21

Ethical approval and samples

This study was approved by the Institute Review Board Committees of National Center for Child Health and Development and Hamamatsu University School of medicine, and performed using peripheral leukocyte samples after obtaining written informed consent.

Molecular studies

We first performed pyrosequencing analysis for four CpG dinucleotides (CG1–CG4) within the IG-DMR and five CpG dinucleotides (CG5–CG9) within the MEG3-DMR, using bisulfite-treated leukocyte genomic DNA samples (Figure 1). The procedure was as described in the manufacturer’s instructions (Qiagen, Valencia, CA, USA). Subsequently, methylation indices (MIs, the ratio of methylated clones) were obtained using PyroMark Q24 (Qiagen). We also studied six UPD(14)mat patients for comparison and 50 control subjects to define the reference ranges of MIs.

Figure 1
figure 1

Representative molecular findings. (a) Methylation analysis by pyrosequencing analysis. Top panel: schematic representation indicating of four CpG dinucleotides (CG1–CG4) within the IG-DMR and five CpG dinucleotides (CG5–CG9) within the MEG3-DMR. The cytosine residues at the CpG dinucleotides are usually methylated after paternal transmission (filled circles) and unmethylated after maternal transmission (open circles). A 164 bp segment encompassing CG1–CG4 and a 167 bp segment harboring CG5–CG9 were PCR amplified with primer sets (PyF1-PyR1 and PyF2-PyR2) hybridizing to both methylated and unmethylated clones, and sequence primers (SP1 and SP2) were hybridized to single-stranded PCR products. Middle panel: pyrosequencing data in cases 1 and 2, a UPD(14)mat patient, and a control subject. Bottom panel: summary of MIs. (b) Microsatellite analysis. The data are consistent with biparental origin of the chromosome 14 pairs. Unequal amplification of the heterozygous peaks in each individual is consistent with short products being more easily amplified than long products, and the patterns of heterozygous peak heights for D14S292 are comparable between case 1 and the father and between case 2 and the mother, with no disproportionally increased heights of maternally derived peaks.

Full size image

When hypomethylation was identified, we performed microsatellite analysis for nine loci on chromosome 14, FISH analysis for the IG-DMR and the MEG3-DMR, and array comparative genomic hybridization for the 14q32.2 imprinted region using a custom-build oligo-microarray containing 12 600 probes (Agilent Technologies, Palo Alto, CA, USA).22 We also performed pyrosequencing for the H19-DMR (ICR1) and the PEG1/MEST-DMR to re-confirm the absence of the known causes for SRS, and for the KvDMR (ICR2), the SNRPN-DMR, the PLAGL1-DMR, and the GNAS exon A/B-DMR to examine the occurrence of multiple methylation defects.23 Primers utilized in this study are shown in Supplementary Table 1.

Results

Molecular studies

Pyrosequencing identified hypomethylation of the IG-DMR and the MEG3-DMR in two of the 85 SRS patients (cases 1 and 2) (Figure 1). The MIs in case 1 were around the lower limit of the MIs in the six UPD(14)mat patients and much lower than the reference range in the 50 control subjects, whereas the MIs in case 2 were above the maximum MIs in the six UPD(14)mat patients, except for the MI of CG4, and below the reference range in the 50 controls, except for the MI of CG3. The MIs were obviously lower at the MEG3-DMR than at the IG-DMR in case 1 and the six UPD(14)mat patients, whereas the MIs were not so different between the IG-DMR and the MEG3-DMR in case 2 and the 50 control subjects.

In cases 1 and 2, microsatellite analysis showed biparental transmission of the homologs of chromosome 14, with similar patterns of peak heights for heterozygous alleles between cases and the parents (eg, comparable patterns of peak heights for the 108 bp and the 112 bp alleles of D14S292 between case 1 and the father and between case 2 and the mother) (Figure 1 and Supplementary Table 2). FISH analysis delineated two copies of the IG-DMR and the MEG3-DMR, and array comparative genomic hybridization revealed no discernible copy number alteration in the 14q32.2 imprinted region (Supplementary Figure 1). Furthermore, the MIs for the six DMRs other than the IG-DMR and the MEG3-DMR were invariably within the normal range in cases 1 and 2 (Supplementary Table 3).

Clinical findings of cases 1 and 2

Both cases 1 and 2 showed severe prenatal growth failure, the mandatory criteria (ie, birth length and/or birth weight ≤−2 SD), and four of the five scoring system criteria (ie, relative macrocephaly at birth, postnatal short stature (≤−2 SD) at ≥2 year of age, prominent forehead during early childhood, and body asymmetry) for the diagnosis of SRS, whereas both of them lacked feeding difficulties (Table 1 and Figure 2). In addition, both cases 1 and 2 exhibited triangular face and clinodactyly, and case 1 manifested irregular teeth, brachydactyly, Single palmar crease, muscular hypotonia, and speech delay. Notably, relative macrocephaly with prominent forehead was no longer recognizable with age in both cases. Consistent with this, although the facial appearance was fairly characteristic of SRS in both cases in infancy to early childhood, it became less characteristic in both cases with age (Figure 2).

Figure 2
figure 2

Photographs of case 1. (a) At 3 5/12 years of age. He exhibits triangular face with prominent forehead and micrognathia, and clinodactyly of the 5th fingers. (b) At 9 6/12 years of age. He exhibits slight central obesity, with the body mass index above the mean. Although this photo suggests mild scoliosis, this is primarily due to body asymmetry with asymmetric leg length. No scoliosis has been identified at the sitting position. He also manifests irregular teeth, joint hypermobility, and clinodactyly of the 5th fingers.

Full size image

Both cases 1 and 2 also exhibited TS (UPD(14)mat) clinical features (Supplementary Table 4). In particular, several features characteristic of TS rather than SRS were observed, such as the body mass index above the mean at 9 years of age (though not assessed as obese), joint hypermobility, and small hands in case 1, and small hands and early onset of puberty in case 2.

Clinical survey also revealed that case 2 was born after in vitro fertilization-embryo transfer, whereas case 1 was born after natural conception. Furthermore, case 1 was treated with growth hormone for short stature from 6 to 8 years of age, and case 2 received growth hormone therapy for short stature since 5 years of age and gonadotropin-releasing hormone analog therapy for precocious puberty since 7 years of age.

Discussion

The present study showed that the IG-DMR and the MEG3-DMR were severely hypomethylated in case 1 with the MIs comparable to those of UPD(14)mat and moderately hypomethylated in case 2 with the MIs between those of UPD(14)mat patients and those of control subjects, in the absence of UPD(14)mat and microdeletion or copy number alteration involving the DMRs. Furthermore, although such hypomethylation patterns, especially the moderate hypomethylation in case 2, could be caused by post zygotic mosaicism with maternal full or distal 14q segmental isodisomy involving the imprinted region,24 microsatellite analysis indicated no disproportionally increased height of the maternally inherited alleles, thereby arguing against the possible mosaicism. Taken together, the results imply the occurrence of epimutation (hypomethylation) of the IG-DMR and the MEG3-DMR in cases 1 and 2.

Cases 1 and 2 satisfied SRS diagnostic criteria proposed by Netchine et al.14 In addition, UPD(14)mat has been identified in a single patient diagnosed as having SRS, although detailed clinical findings are unknown (No. 445 in Table 1).12 Furthermore, phenotypic assessment of TS patients with UPD(14)mat reported in the literature reveals that such patients frequently exhibit clinical features utilized as the mandatory and the scoring system criteria for SRS (Table 1). Indeed, pre- and post-natal growth failure, prominent forehead, and feeding difficulties are shared in common by SRS and TS (Table 1 and Supplementary Table 4). In this regard, although the presence or the absence of body asymmetry is not described in most TS patients, it is unlikely that body asymmetry was not reported despite its presence (body asymmetry has been described in a single patient with UPD(14)mat and Prader–Willi Syndrome-like phenotype).25 Thus, it is inferred that a considerable degree of phenotypic overlap exists between SRS and TS, except for body asymmetry that is apparently characteristic of SRS, and that epimutations of the IG-DMR and the MEG3-DMR were identified in cases 1 and 2 who exceptionally manifested body asymmetry.

Several matters should be pointed out in this study. First, the MIs were obviously lower at the MEG3-DMR than at the IG-DMR in case 1 and the six UPD(14)mat patients, whereas the MIs were not so different between the IG-DMR and the MEG3-DMR in case 2 and the 50 control subjects. As the IG-DMR and the MEG3-DMR function as the imprinting centers in the placenta and the body, respectively,26 hypomethylation may be more strictly established in the MEG3-DMR of leukocytes in patients with UPD(14)mat and definitive epimutation. Second, multiple methylation defects was not detected in cases 1 and 2. Although the examined DMRs were rather limited, this may argue that isolated epimutation of the IG-DMR and the MEG3-DMR can lead to SRS phenotype. Third, relative macrocephaly with prominent forehead became clinically non-recognizable with age in cases 1 and 2. Thus, although clinical features of the two cases were compatible with SRS with no specific finding that serves to distinguish the two cases from other SRS patients in infancy to early childhood, they became less characteristic for SRS with age. Indeed, except for body asymmetry, their recent clinical features were more similar to those of patients with TS4, 8 or those of patients with short stature born small-for-date with no catch-up growth.27 Such phenotypic change with age, in addition to TS-like clinical features such as recent body mass index gain in case 1 and early onset of puberty in case 2, might be characteristic of SRS patients with an aberrant chromosome 14 imprinted region. Fourth, case 2 was born after i n vitro fertilization. As i n vitro fertilization could be a risk factor for the occurrence of epimutation (hypomethylation),28 i n vitro fertilization may be related to the moderate degree of epimutation in case 2. Lastly, epimutation was identified only in two of the 85 SRS patients who were free from epimutation of the H19-DMR and UPD(7)mat. Poole et al12 also have identified UPD(14)mat in one of 127 SRS patients, although clinical assessment remained fragmentary in 127 patients. Thus, UPD(14)mat and related genetic aberrations account for only a small fraction of SRS patients, and underlying factor(s) still remain to be clarified in many SRS patients. Nevertheless, analysis of the chromosome 14 imprinted region is worth attempting in SRS patients, especially in those with neither hypomethylation of the H19-DMR nor UPD(7)mat.

In summary, we identified epimutations affecting the IG-DMR and the MEG3-DMR in two patients with SRS-compatible phenotype. Further studies will permit to define the phenotypic spectrum of TS with aberrations of the chromosome 14 imprinted region.