Introduction

Silver–Russell syndrome (SRS) is a rare (1:30.000-100.000) imprinting disorder characterized by pre- and postnatal growth retardation, relative macrocephaly at birth associated with a triangular face and a prominent forehead, body asymmetry, and feeding difficulties. Clinical diagnosis is based on the occurrence of at least 4 out of 6 clinical signs, in accordance with the Netchine–Harbison Clinical Scoring System (NH-CSS)1. The Silver Russell syndrome is a paradigmatic imprinting disorder, where specific genes undergo to a monoallelic expression of a single parental allele, while the other one is silenced (imprinted). Loss of methylation of the paternal allele at H19/ IGF2:IG-DMR in the 11p15.5 region (IC1_LoM, 30%–60% of cases), leading to IGF2 gene downregulation, and maternal uniparental disomy of chromosome 7 (UPD(7)mat, 5%–10% of cases) involving GRB10:alt-TSS-DMR, PEG10:TSS-DMR, and MEST:alt-TSS-DMR1,2 are the frequent etiopathogenic mechanisms. Further rare abnormalities, such as epimutations in MEG3:TSS-DMR or UPD(14)mat, UPD(16)mat, UPD(20)mat, and pathogenic variants within the IGF2, PLAG1, HMGA2, and CDKN1C genes have been described in approximately 4-5% of SRS patients3,4. However, about 40% of clinical SRS cases remain molecularly undiagnosed1. Cohorts of individuals with IC1_LoM and UPD(7)mat have a very similar phenotype, even if among the UPD(7)mat group, a minor frequency of body asymmetry and a higher incidence of neurodevelopmental delay were observed3,5,6. Most of the reported studies on the UPD(7)mat have associated the onset of the SRS phenotype with the combined altered expression of the three imprinted genes GRB10, MEST, and PEG10. However, SRS patients with segmental UPD(7)mat not involving the three DMRs and SRS, SRS-like patients with a genetic defect impairing singularly GRB107,8,9,10 or MEST11,12,13 have been described. Here, we refer to the first familial case involving the PEG10 gene. PEG10 has aroused interest due to its maternal imprinting and critical role in early development, particularly in placental and fetal growth14. PEG10 is known to be expressed only from the paternal allele, predominantly in the placenta, though its paternal expression is also observed in various other tissues postnatally15. Despite its essential role in growth regulation, the precise contribution of PEG10 to the development of SRS has remained unclear.

In this study, we report a novel paternal deletion involving both the PEG10 and SGCE genes in two siblings with a strong clinical suspicion of SRS. By investigating the molecular effects of the PEG10 deletion, we aim to elucidate the mechanisms underlying the clinical manifestations of SRS and deepen our understanding of its genetic basis.

Materials and methods

Methylation-Specific Multiplex Ligation Probe-dependent Amplification (MS-MLPA)

The MRC-Holland kit (MRC Holland, Amsterdam, Netherlands) ME-030 BWS/RSS, ME032-UPD7-UPD14, and ME031-GNAS were used to investigate the methylation status at SRS imprinted loci. The analyses were performed according to the manufacturer’s protocols. Four control samples were included in each experiment. Raw data were analyzed using Coffalyser.Net software (version 140,701, MRC Holland).

CGH array

Whole-genome array-CGH analysis was performed using the GenetiSure Postnatal Research CGH+SNP Microarray 2x400K platform (Agilent) to detect copy number variants (CNVs) and loss of heterozygosity. Labeling and hybridization were performed according to the manufacturer’s protocol, and CNVs were detected by the Agilent Cytogenomics v5.0.2.5 analysis software. The map positions refer to the Human Genome Building 37 (hg19) assembly. Detected CNVs were compared with the Database of Genomic Variants (http://projects.tcag.ca/variation/, release March 2016) to exclude common copy number polymorphisms (minor allele frequency >1%).

Whole Exome Sequencing (WES)

WES was performed at the BIODIVERSA srl service (Milan, Italy). WES libraries were prepared with the SureSelect XT PreCap Human All Exon V8 kit (Agilent Technologies Inc., Santa Clara, CA, USA) and sequenced on a NovaSeq 6000 system (Illumina Inc, San Diego, CA) to generate 150 bp paired-end reads with a mean target coverage of ≥100×. FASTQ files were quality-checked with FastQC v0.11.9 and processed using the DRAGEN Bio-IT Platform v3.8.4 in WES mode, which performed read alignment to GRCh38, duplicate marking, realignment, base recalibration, and germline variant calling. Variant annotation was performed using WANNOVAR16, incorporating functional predictions and population frequency data. A virtual panel of 2508 growth-related genes was designed to disclose causative variants by reviewing the literature and using PanelApp17. All variants identified were filtered by minor allele frequency (< 1%) in the 1000 Genomes, Genome Aggregation Databases, and Exome Aggregation Consortium databases. The interpretation of the variants was based on the classification by the InterVar, VarSome, and Franklin by Genoox databases18,19 in accordance with the American College of Medical Genetics and Genomics/Association for Molecular Pathology guidelines20,21. All the variants herein reported were confirmed by Sanger sequencing.

Reverse Transcription Quantitative Real-Time PCR (RT-qPCR) and statistical analysis

Total RNA of patients, their relatives, and healthy controls was collected using Tempus Blood RNA tubes (Thermo Fisher Scientific) and isolated using the Tempus Spin RNA Isolation kit (Thermo Fisher Scientific). RNA was reverse-transcribed into cDNA using the SuperScript VILO cDNA Synthesis Kit (Thermo Fisher, Waltham, MA, USA). RT-qPCR reactions for the PEG10 (F_GACCCCATCCTTCCTGTCTTC, R_CCCCTCTTCCACTCCTTCTTT), IGF2 (F_CCGGCTTCCAGACACCAAT, R_GGTAAGCAGCAATGCAGCAC), and CDKN1C (F_CGGCGATCAAGAAGCTGTC, R_GCTGATCTCTTGCGCTTGG) genes were carried out using the SYBR Green Universal Master Mix (Thermo Fisher Scientific). They were performed using the QuantStudio 12K Flex Real-Time PCR System (Thermo Fisher Scientific). Data were analyzed using the QuantStudio 12K Flex Software v1.2.3 (Thermo Fisher Scientific). The amounts of mRNA were calculated using the 2^-ΔΔCt method, normalized to the housekeeping genes HMBS and ACTB, and replicated three times. We established the proper range of gene expression in 12 healthy adult controls (6 males and 6 females, 20-40 years). Statistical analysis was performed using GraphPad Prism 9.0 software. Data obtained from expression studies were analyzed using the Kruskal-Wallis test and Dunn’s multiple comparisons test.

Results

Clinical presentation

The two affected brothers were born to healthy unrelated parents with a normal height (mother 160 cm; father 175 cm; target height 170.6 cm, -0.6 SDS). Both pregnancies occurred spontaneously and were complicated by maternal cholestasis. Data on gestational age as well as birth weight, birth length, and head circumference were obtained from birth charts and converted to SD according to Bertino et al22. Auxological data were evaluated according to Tanner growth charts23. Growth charts of the two boys are displayed in Supplementary Figure 1. Parents did not allow the publication of the photos. The main clinical features of the sibs are summarized in Figure 1.

Fig. 1
Fig. 1
Full size image

Main clinical features of our patients and segregation of the 7q21.3 deletion. The family pedigree shows that the deletion is present in both affected siblings, as well as in the unaffected father and paternal grandmother, while it is absent in all other family members. The clinical features of the sibs are summarized in the boxes. AGA = appropriate for gestational age; GER = gastroesophageal reflux.

Patient 1 (Pt1) was born at 36 + 6 weeks of gestation with a birth weight (BW) of 2830 g (−0.07 SDS), a birth length (BL) of 47 cm (−0.56 SDS), and an occipital-frontal circumference (OFC) of 35.5 cm (1.66 SDS). Facial dysmorphisms included a triangular face with a prominent forehead and frontal bossing, micrognathia, downturned corners of the mouth, and thin lips. He experienced feeding difficulties with gastroesophageal reflux and episodes of hypoglycemia. Muscular hypotrophy, fifth finger clinodactyly, left valgus foot and knee, toenail dystrophy, extra hair whorls, and fetal pads were also observed. At 24 months, he weighed 8.6 kg (−3.79 SDS), measured 79 cm in length (−2.45 SDS), and had an OFC of 49 cm (0.2 SDS). He fulfilled 4 out of 6 NH-CSS criteria. Due to a severe growth delay, two growth hormone (GH) tests were performed at the age of four years, revealing an insufficient GH peak, which is suggestive of GH deficiency. GH treatment was therefore started, resulting in an improvement in growth velocity.

A mild motor delay was reported with autonomous walking at 18 months and difficulties in fine motor skills. Slight distal tremors were first noticed at the age of seven, along with easy fatigue. At the last clinical assessment at the age of 8 years and 9 months, his weight was 22 kg (−1.87 SDS), his height was 128.2 cm (-0.4 SDS), and his OFC was 54 cm (1.3 SDS). Body Mass Index (BMI) was 13.8 Kg/m2 (-1.65 SDS according to World Health Organization, WHO).

Patient 2 (Pt2) was born at 35 + 2 weeks of gestation with a BW of 2580 g (0.12 SDS), a BL of 47 cm (0.12 SDS), and an OFC of 35 cm (1.76 SDS). At birth, he required resuscitation with CPAP. Feeding difficulties, requiring a nasogastric tube for 20 days after birth, muscular hypotonia, right-sided aortic arch, atrial septal defect, phimosis, broad halluces with toenail dystrophy, and varus feet were reported. At the last evaluation, when the child was 4 years and 8 months old, his weight was 13.7 kg (−2.33 SDS), his height was 94.4 cm (−2.5 SDS), and his OFC was 51.4 cm (0.3 SDS). BMI was 15.6 Kg/m2 (0.13 SDS).

He displayed a triangular face, protruding forehead, frontal bossing, saddle nose, downturned corners of the mouth, straight eyebrows, and short palpebral fissures. He reached an NH-CSS of 4/6. Furthermore, a mild motor delay was observed, with independent walking achieved at 16 months and a wide-based gait.

Brain Magnetic Resonance Imaging (MRI) was normal. Because of short stature, 2 GH stimulation tests were performed at the age of 2 and at the age of 4 years, which revealed a normal GH peak (9.64 ng/ml).

These patients undergo auxo-endocrinological assessments every six months and neurological evaluations every year as part of their follow-up.

Molecular analysis

Following the Pt2’s SRS clinical diagnosis, methylation was analysed by MS-MLPA in the imprinted regions on chromosomes 11, 7, 14, and 20, without disclosing (epi)genetic or genetic alterations. In line with our diagnostic workflow3, sequencing of the SRS-related genes, HGMA2, IGF2, PLAG1, CDKN1C, and IGF1R was carried out. After the birth of the second child, displaying light cardiac anomalies, a high-resolution CGH Array revealed a paternal deletion of approximately 109 Kb at 7q21.3, arr[GRCh37]7q21.3(94141411x2,94191291_94300309x1,94306364x2) encompassing the entire SGCE gene, associated to the autosomal dominant Myoclonic Dystonia 11 (DYT11, MIM#159900), as well as the maternally imprinted PEG10 gene (Figure 2). As depicted in Figure 1, the deletion, also present in the eldest brother, was segregated from the healthy father and paternal grandmother and was absent in their paternal uncle. The familial segregation was consistent with the hypothesis of an imprinting disorder. To rule out the occurrence of an alternative etiology, WES was analysed, but no pathogenic or likely pathogenic variant correlated to the growth disorder, GH deficit, and myoclonic dystrophy was identified. Table 1 summarizes the clinical features of patients previously reported in the literature who carry deletions with a maximum size of 2.5 Mb at 7q21.3; Figure 3 depicts the exact deletion sizes and gene content, demonstrating that our case is the smallest reported to date and the only one including only PEG10 and SGCE genes.

Fig. 2
Fig. 2
Full size image

Genomic profile of the 7q21.3 deletion. Array-CGH profile of chromosome 7 (hg19 genome assembly) showed an interstitial deletion of 109 Kb in the long arm of the chromosome (7q21.3), involving the maternally imprinted PEG10 and SGCE genes.

Table 1 Clinical and molecular features of the reported patients with a contiguous 7q21.3 gene deletion smaller than 2.5 Mb.
Full size table
Fig. 3
Fig. 3
Full size image

Representation of contiguous gene deletions in the 7q21.3 chromosomal region involving PEG10 and SGCE genes. In the figure are displayed only the deletions smaller than 2.5 Mb. The deletion reported in this study is red highlighted.

Interestingly, a paternal missense variant in the COL7A1 gene (NM_000094):c.7048C>T, resulting in the substitution of proline 2350 with a serine, was disclosed in both probands. Alterations in this gene are associated with different phenotypes, including the autosomal dominant Nail-Disorder Nonsyndromic Congenital 8 (NDNC8, MIM#607523), characterized by toenail dystrophy. Since the nail phenotype of the father was not assessed, the variant was classified as of Unknown Significance (PM2, PM1, PP3).

According to the paternal expression of the gene in this tissue, the paternally inherited deletion of PEG10 was expected to affect its expression. RNA expression level was then studied in blood samples of the patients, their parents, and paternal uncle, while the paternal grandmother was not available for the collection of a new sample for the RNA study. The RT-qPCR analyses showed a downregulation of the PEG10 expression in both probands compared to healthy controls, but not in their relatives (Figure 4). The next step was aimed at correlating the loss of function of PEG10 with the SRS phenotype. Given that PEG10 regulates the expression of cyclin-dependent kinase inhibitor genes, including CDKN1C, and IGF2 is the main SRS causative gene, we investigated the expression of these genes by RT-qPCR analyses. Both CDKN1C and IGF2 turned out to be upregulated in the Pt1, taking GH, while only CDKN1C was overexpressed in Pt2 (Figure 4). A normal expression of both genes was observed in their relatives.

Fig. 4
Fig. 4
Full size image

Expression studies of PEG10, IGF2, and CDKN1C genes. Results from qPCR studies in blood from the probands and their relatives for different genes. The expression of the target genes was normalized against the housekeeping gene HMBS. Each experiment was replicated three times. The control group consisted of blood expression derived from 12 healthy individuals (* < 0.05, ** < 0.01, *** < 0.001). Error bars represent standard deviation.

The deletion of SGCE leads to SGCE myoclonus-dystonia (SGCE-M-D), a maternal imprinting disorder, characterized by a brief muscle contractions, repetitive movements, occurring mainly in the upper part of the body. The disease has a variable penentrance and an onset before age of 18 years. The deletion of SGCE may explain the motor impairment observed in patient 1.

Discussion

The investigation of the etiology of Silver-Russell syndrome uncovered many (epi)genetic and genetic defects. Although UPD(7)mat was the first molecular alteration identified in clinically suspected SRS24, the main deregulated genes on chromosome 7 and their role in pre- and postnatal development remain incompletely understood. To date, three maternally imprinted domains with a putative clinical relevance have been identified on chromosome 7: GRB10:alt-TSS DMR at 7p12.1; PEG10:TSS DMR at 7q21.3; and MEST:alt-TSS DMR at 7q3225. GRB10 is a growth suppressor gene biallelically expressed in most fetal and postnatal tissues but only maternally expressed in the placenta26, while MEST is a paternally expressed gene whose involvement in growth deficit is not well documented10. Maternal duplications of GRB10 and paternal deletions of MEST are linked to SRS-like clinical features7,8,9,11,12,13. Recently, an intragenic paternal deletion of GRB10 has been reported in a couple of monozygotic twins with postnatal SRS clinical features10. Furthermore, segmental UPD(7q)mat involving the MEST gene was found associated with classic or partial SRS phenotype27,28. While molecular and clinical evidence support the involvement of GRB10 and MEST genes in the etiology of SRS, the role and function of the maternally imprinted PEG10 gene have remained unclear10.

For the first time, we describe a paternally inherited deletion encompassing PEG10 and SGCE genes in two siblings with a strong clinical suspicion of SRS, providing new evidence for PEG10 involvement in SRS pathogenesis. As shown in Figure 3, larger deletions in the region, typically spanning the COL1A2 gene, have been documented previously and are associated with Osteogenesis Imperfecta (OI)29,30,31,32as well as SGCE-related myoclonus dystonia. Patients with these larger deletions commonly exhibited bone fractures, dentinogenesis imperfecta, hearing loss, short stature, and joint laxity29, except for the patients described by Dale et al., in which a 170 Kb deletion affecting only SGCE and CASD1 genes was reported33. In contrast, our family presents the smallest deletion reported so far at 7q21.3, affecting only PEG10 and SGCE genes. Unlike those with larger deletions, our patients exhibit no OI-related features, aside from short stature, most likely because COL1A2 and other genes are not impacted.

The PEG10 gene is expressed only from the paternal allele, predominantly in the placenta of both humans and mice, but also shows high postnatal expression in various tissues such as adipose, kidney, brain, muscle, and lung14,15. In mouse models, embryos lacking PEG10 appear morphologically normal at embryonic day 9.5 (E9.5), but by E10.5 they exhibit severe growth retardation of both the embryo and placenta34. In humans, PEG10 expression is low during early gestation but increases significantly around weeks 11–12, suggesting that PEG10 plays an essential role in later stages of placental development35. Clinically, hypermethylation and aberrant PEG10 expression have been associated with several pregnancy complications, including spontaneous miscarriage, intrauterine growth restriction (IUGR), and preeclampsia36,37. Doria et al. found a reduced expression of PEG10 in fetal tissues at the third trimester, comparing fetuses from miscarriages versus control fetuses. Despite the critical role of PEG10 in both mouse and human placental development, our patients did not present any structural placental abnormalities or IUGR, and were born appropriate for gestational age. However, both patients were born preterm, and patient 2 required resuscitation. These observations may reflect distinct features of early human placental development30,38. Indeed, most previously reported cases involving paternal deletions at 7q21.3, which affect PEG10, SGCE, and additional genes, have not been associated with placental abnormalities or IUGR, supporting the idea that our patients’ isolated deletion of only PEG10 and SGCE may underlie the observed lack of placental phenotype aside from short stature.

Besides the placental function, PEG10 has been reported as an oncogene implicated in tumor cell proliferation, metastasis, and apoptosis15. Several studies showed that PEG10 is upregulated in different cancer types, with its overexpression enhancing cancer cell proliferation15. Conversely, reduced expression of PEG10 is associated with marked growth inhibition of tumor cells15,39. The oncogenic properties of PEG10 are linked to its ability to bind DNA via a CCHC-type zinc finger domain40. Other studies have shown that PEG10 functions as a transcriptional repressor by binding to the promoters of key cell cycle-related genes active during the G0/G1 phase. These include cyclin-dependent kinase inhibitors (such as p16, p18, p21, p27, and p57) as well as specific cyclins (CCNE1, CCNE2, CCND1)15,39,40,41. Downregulation of PEG10 in tumor cells leads to G0/G1 phase arrest, mediated by the upregulation and accumulation of the above inhibitory factors. The cited studies highlight the role of PEG10 in promoting cell cycle progression from G0/G1 to S phase41. In an attempt to characterize the molecular impact of the paternal PEG10 deletion in our patients, we evaluated the expression of PEG10 and SRS genes IGF2 and CDKN1C by RT-qPCR studies. Consistent with the maternal imprinting of PEG10 in blood42, its downregulated expression in UPD(7)mat SRS10, and the segregation pattern of the deletion within the family, we observed a complete absence of PEG10 expression in the two affected brothers, whereas expression levels were normal in both parents and paternal uncle. Although the literature indicates that PEG10 does not bind the promoters of cell-growth genes such as IGF2, PLAG1, and HMGA240, we assessed IGF2 expression due to its central role in SRS. Interestingly, IGF2 expression was normal in Pt2 but also increased compared to the healthy relatives in Pt1. We can hypothesize that IGF2 upregulation in Pt1 may be attributed to the ongoing GH therapy, according with the positive physical response observed. However further experiments will be required to sustain this observation. Human IGF2 expression is regulated by GH through promoter binding, and GH-treated patients often exhibit increased serum IGF2 levels after treatment initiation43,44,45. The normal IGF2 expression in Pt2 contrasts with the reduced expression typically reported in SRS patients with UPD(7)mat and UPD(7q)mat10. These discrepancies may be explained by hypothesizing that a complementary effect of the three chromosome 7 imprinted genes is necessary to obtain a complete phenotype and classic gene downregulation. These findings support the idea that the deregulation of PEG10 in SRS does not involve the IGF2 gene, aligning with previous reports on the known targets of PEG10 binding40. To enforce this consideration, there is also the occurrence of GH deficit only in one of the two affected children. Regarding the CDKN1C gene, and in line with the role of PEG10 in regulating cell cycle progression40, RT-qPCR analysis revealed upregulation of CDKN1C expression in both patients, with higher levels in Pt2, likely because he has not yet initiated growth hormone therapy. The finding is consistent with the established role of CDKN1C as a growth inhibitor and the documented pathogenic gain-of-function variants in CDKN1C in other SRS cases46,47,48,49. In contrast, unexpected downregulation of CDKN1C expression was recently reported in SRS patients with UPD(7)mat and UPD(7q)mat 10. Additionally, rare paternal microdeletions affecting the KCNQ1OT1:TSS-DMR in the IC2 region of 11p15.5 not involving the CDKN1C gene, and maternal microduplications of IC2 have been linked to upregulation of gene expression and have been reported in cases of SRS or growth retardation phenotype50,51,52,53. Therefore, an increased functionality of CDKN1C, as observed in our patients, may contribute to cell cycle arrest and ultimately may result in growth restriction.

In conclusion, our findings suggest considering the PEG10 gene as a gene involved in the pathogenesis of Silver-Russell Syndrome and provide new insights into the molecular mechanisms underlying the disorder. The paternal deletion of PEG10 and SGCE identified in our patients, along with the observed upregulation of the key growth-related gene CDKN1C, suggests that this genetic alteration contributes to the characteristic SRS clinical phenotype, and that locus-specific structural alterations involving PEG10 and SGCE may represent a previously underappreciated mechanism contributing to SRS in a subset of patients. This also allows us to hypothesize the existence of an intricate interplay between imprinted gene dosage and growth regulatory pathways mediated by CDKN1C in the syndrome’s molecular etiology. However, a limitation of this study is the use of blood samples for gene expression analysis, as PEG10 shows higher expression in placental and brain tissues. Altogether, these observations invite further investigation of paternal deletions of PEG10 and SGCE, utilizing relevant tissue samples to provide more definitive evidence.