Multi-locus methylation analyses reveal GNAS methylation defects in three patients with the Beckwith–Wiedemann syndrome phenotype and no molecular defects in the 11p15.5 imprinted region

Abstract

Background

Beckwith–Wiedemann syndrome (BWS) is a congenital imprinting disorder (ID) caused by molecular defects in the 11p15.5 imprinted region, such as hypomethylation of the KCNQ1OT1:TSS-differentially methylated region (KCNQ1OT1-DMR) and hypermethylation of the H19/IGF2:IG-DMR, and maternal CDKN1C pathogenic variants, with various clinical characteristics, including overgrowth and macroglossia. Recently, the concept of Beckwith–Wiedemann spectrum (BWSp) and a clinical scoring system for BWS have been proposed, and cases with four or more points are diagnosed with classic BWS, and 20% of cases with BWS have no molecular defects in the 11p15.5 imprinted region. Pseudohypoparathyroidism type 1B (PHP1B, alias inactivating parathyroid hormone (PTH)/PTH-related protein signaling disorder 3) has characteristics of hormone resistance, particularly PTH, caused by methylation defects in DMRs at the GNAS locus (GNAS-DMRs). Some cases with PHP1B show postnatal overgrowth, which overlaps the BWS-phenotype. However, no studies have conducted a multi-locus methylation analysis for the ID-responsible DMRs other than the DMRs in 11p15.5 in cases with the BWS-phenotype and without molecular defects in the 11p15.5 imprinted region.

Results

We conducted methylation analysis using pyrosequencing in 77 patients showing the BWS-phenotype without molecular defects in the 11p15.5 imprinted region. Consequently, we identified three patients with methylation defects in the GNAS-DMRs. Patients 1, 2, and 3 had 9, 5, and 4 points in a BWSp score, respectively. All three patients had macroglossia and postnatal overgrowth. Further analyses, methylation-specific multiple ligation-dependent probe amplification for multiple DMRs, array-based methylation analysis, exome sequencing, array comparative genome hybridization analysis, and microsatellite marker analysis showed 9p deletion in Patient 1 and paternal uniparental isodisomy of chromosome 20 in Patient 2 together with multiple methylation defects in DMRs other than the GNAS-DMRs. Patient 3 had methylation defects in only the GNAS-DMRs.

Conclusion

Methylation defects in the GNAS-DMRs can cause the BWS-phenotype. For cases with the BWS-phenotype but no molecular defects in the 11p15.5 imprinted region, methylation analysis for the DMRs at the GNAS locus should be considered.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13148-025-01907-y.

Background

Beckwith–Wiedemann syndrome (BWS; OMIM #130650) is a congenital disorder caused by aberrant expression of the imprinted gene(s) on chromosome 11p15.5, namely, an imprinting disorder (ID). Cases with BWS exhibit a variety of clinical characteristics, including overgrowth, macroglossia, exomphalos, ear abnormalities, abdominal organ enlargement, and hyperinsulinemic hypoglycemia [1]. The human chromosome 11p15.5 harbors the H19/IGF2:IG-differentially methylated region (DMR) (H19-DMR) with H19 and IGF2 genes and the KCNQ1OT1:TSS-DMR (KCNQ1OT1-DMR) with KCNQ1OT1, CDKN1C, and KCNQ1 genes [1]. The major molecular etiologies of BWS consist of hypomethylation of the KCNQ1OT1-DMR (frequency: 50%), hypermethylation of the H19-DMR (5%), paternal uniparental disomy (UPD) of chromosome 11 (20%), maternal loss of function mutation of CDKN1C (5%), and pathogenic copy number variants involving 11p15.5 (< 3%) [1, 2]. In addition, up to 10% of cases with UPD of chromosome 11 have mosaic genome-wide paternal UPD [1]. Because of various phenotypes in cases with the molecular etiologies of BWS, BWS was redefined as the Beckwith–Wiedemann spectrum (BWSp), and a clinical scoring system for BWS was proposed by a consensus statement [1]. This scoring system suggests that cases with two or more points in the BWSp score undergo a methylation analysis for the H19-DMR and KCNQ1OT1-DMR. If the cases with BWS have the normally methylated H19-DMR and KCNQ1OT1-DMR, the consensus statement suggests conducting the following analyses: methylation analysis for the H19-DMR and KCNQ1OT1-DMR in the other tissues, such as buccal mucosa or skin fibroblasts, CDKN1C mutation screening, and chromosomal structural analysis. Cases with four or more points in the BWSp score are diagnosed with classic BWS, even if they do not have molecular defects in the 11p15.5 imprinted region; 20% of cases with BWS have no molecular defects in the 11p15.5 imprinted region.

The GNAS locus on chromosome 20q13.32 harbors four DMRs (GNAS-DMRs) which comprise the GNAS-A/B:TSS-DMR (A/B-DMR), GNAS-XL: Ex1-DMR (XL-DMR), GNAS-AS1:TSS-DMR (AS1-DMR), and GNAS-NESP:TSS-DMR (NESP-DMR). Pseudohypoparathyroidism type 1B (PHP1B, alias inactivating parathyroid hormone (PTH)/PTH-related protein signaling disorder 3) is a rare endocrine disorder caused by methylation defects in the GNAS-DMRs, and cases with PHP1B commonly show hypomethylation of the A/B-DMR. PHP1B has PTH resistance, characterized by increased serum PTH level, hypocalcemia, and hyperphosphatemia, and PTH resistance gradually manifests after birth [3]. In addition, some cases with PHP1B show resistance to thyroid-stimulating hormone (TSH) and have Albright’s hereditary osteodystrophy, defined by short stature, ectopic ossification, brachydactyly, and a round face. Furthermore, postnatal overgrowth, common in BWS, is frequently observed in cases with PHP1B [3, 4].

Some clinical features overlap among different IDs. For example, cases with BWS and PHP1B present overgrowth during the infantile period. Moreover, recent advances in analytical techniques have revealed multi-locus imprinting disturbance (MLID) with aberrant methylation of multiple DMRs in the different imprinted regions [5]. Among cases with IDs having MLID, the BWS-phenotype is the most frequently observed. About 12% of cases with BWS caused by epimutation show MLID [6–9]. Two cases with MLID showing both BWS- and PHP1B-phenotypes, including macroglossia and PTH resistance, have hitherto been reported, and these cases had aberrant methylation of the DMRs in the 11p15.5 imprinted region and the GNAS region on chromosome 20q13.32 [10, 11]. A recent study using whole-exome sequencing (WES) in BWSp cases without methylation defects in 11p15.5 or pathogenetic variants of CDKN1C detected the etiologies of other genetic diseases in 14.5% [12]. However, no studies have conducted multi-locus methylation analysis for the ID-responsible DMRs other than 11p15.5 in cases with BWSp having no molecular defects in the 11p15.5 imprinted region.

Here, we conducted multi-locus methylation analysis for the ID-related DMRs other than the H19-DMR and KCNQ1OT1-DMR in patients with the BWS-phenotype and no molecular defects in the 11p15.5 imprinted region, identified three patients with methylation defects in the DMRs on the GNAS locus on chromosome 20q13.32, and expanded the diversity of the etiologies leading to the BWS-phenotype.

Results

Identification of methylation defects in the GNAS-A/B:TSS-DMR

The study flowchart is shown in Fig. 1. Briefly, we conducted methylation analysis for the PLAGL1:alt-TSS-DMR on chromosome 6, PEG10:TSS-DMR on chromosome 7, MEG3/DLK1:IG-DMR and MEG3:TSS-DMR on chromosome 14, SNURF:TSS-DMR on chromosome 15, and A/B-DMR on chromosome 20 by pyrosequencing in 77 patients showing the BWS-phenotype without methylation defects in the H19-DMR and KCNQ1OT1-DMR or CDKN1C pathogenic rare variants. Consequently, we identified three patients with methylation defects in the A/B-DMR (Fig. 1).

Fig. 1.

Study flowchart. BWS, Beckwith–Wiedemann syndrome; DMR, differentially methylated region

Case presentation

The clinical findings of the three patients are shown in Table 1, Fig. 2, and Additional file 1: Supplemental Fig. 1. Patient 1 was a Japanese girl conceived naturally by healthy and unrelated parents. Fetal abnormality was not observed during the pregnancy. She was born at 37 weeks of gestation by vaginal delivery. Her birth weight (BW) was within the normal range of gestational age and sex-matched neonates, and her newborn screening result was negative. She was hospitalized due to a fever during the neonatal period, and BWS-like clinical features, such as persistent hyperinsulinemic hypoglycemia, macroglossia, and lateralized overgrowth, were noticed by attending physicians, and her BWSp score was nine points. After hospitalization, she started receiving treatment with diazoxide for hypoglycemia. As shown in Fig. 2 and Additional file 1: Supplemental Fig. 1, she showed rapid weight gain in infancy, and her body mass index (BMI) was over 90%ile at 3 months of age, and a round face became apparent. She began taking antiseizure medication due to status epilepticus not related to hypoglycemia or hypocalcemia at 2 months of age. She had an elevated serum TSH level and started receiving treatment with levothyroxine at 6 months of age. At 1 year of age, she had temporary weight loss, probably due to infection and excessive levothyroxine doses. Hitherto, her serum intact PTH (iPTH), calcium (Ca), or inorganic phosphate (IP) levels were within the normal range (Fig. 2). She still requires treatment with diazoxide.

Table 1.

Clinical characteristics in patients 1–3

	Patient 1	Patient 2	Patient 3
Sex	Female	Male	Male
Etiologies	9p deletion	UPD(20)pat	Epimutation on the GNAS locus
Multi-locus methylation defects of the DMRs	+	+	–
Birth history
Gravidity (G) and parity (P)	G1P1	G1P1
Birth length, cm (SDS)	50.5 (1.3)	40.0 (–0.8)	44.0 (–1.1)
Birth weight, g (SDS)	2964 (0.8)	1415 (–1.4)	2366 (–0.5)
Gestational age (weeks:days)	37:6	32:0	36:3
Paternal age at birth (years)	19	29	36
Maternal age at birth (years)	19	26	39
Age at (epi)genetic testing (months)	6	5	5
Age at present (years: months)	2:9	2:3	3:9
BWSp characteristics
Macroglossia	+	+	+
Exomphalos	–	–	–
Lateralized overgrowth	+	–	–
Hypoglycemia	+ + (lasting > 1 week)	+ + (lasting > 1 week)	–
Tumor	–	–	–
Ear creases and/or pits	+	–	+
Organomegaly	–	–	–
Naevus flammeus	+	–	–
Umbilical hernia	+	+	+
Polyhydramnios and/or placentomegaly	–	–	–
BWSp score	9	5	4
PHP characteristics
Brachydactyly	–	–	–
Ectopic ossification	–	–	–
Short stature	–	+	+
Round face	+	+	+
Obesity	+	+	+
PTH resistance	–	–	+
TSH resistance	+	+	+
Hormonal and biochemical findings at the time of genetic examination
Intact PTH (pg/mL)	21	39	21
Serum calcium (mmol/L)	2.45	2.63	2.63
Serum inorganic phosphorus (mmol/L)	1.60	2.27	1.76
TSH (μIU/mL)	13.4	14.5	2.4
Free thyroxine (ng/dL)	1.09	0.98	1.30
Other clinical features	Intellectual disability Thalamic stroke	Intellectual disability Cleft palate	None

Reference ranges: intact PTH, 10–65 pg/mL; Ca, 2.20–2.60 mmol/L; IP, 1.22–2.08 mmol/L, TSH, 0.4–4.0 μIU/mL; and free thyroxine, 0.99–1.90 ng/dL

DMR, differentially methylated region; SDS, standard deviation score; BWSp, Beckwith–Wiedemann spectrum; PHP, pseudohypoparathyroidism; PTH, parathyroid hormone; TSH, thyroid-stimulating hormone; and UPD(20)pat, paternal uniparental disomy of chromosome 20

Fig. 2.

Clinical information of Patients 1–3. A Growth chart of three patients. Their height and weight are plotted on the reference growth curves for Japanese children (http://jspe.umin.jp/medical/chart_dl.html). SD, standard deviation. B Trends of electrolytes and endocrine hormones. Gray background indicates a normal range. The black arrows indicate starting treatment with levothyroxine in Patients 1 and 2. iPTH, serum intact parathyroid hormone; Ca, serum calcium; IP, serum inorganic phosphorus; and TSH, serum thyroid-stimulating hormone

Patient 2 was a Japanese boy conceived naturally by healthy and unrelated parents. He was born by emergency cesarean section due to his mother’s gestational hypertension at 32 weeks of gestation. His BW was within the normal range of gestational age of sex-matched neonates. He had hyperinsulinemic hypoglycemia from just after birth to 3 months of age, but was not treated with diazoxide. His newborn screening test revealed an elevated TSH level, and treatment with levothyroxine was initiated. As shown in Fig. 2 and Additional file 1: Supplemental Fig. 1, he showed rapid weight gain in infancy, and his BMI was over 97%ile at 6 months of age, and a round face became prominent in infancy. He had a neurodevelopmental disorder with a developmental quotient of 66 at 18 months of age. His BWSp score was five points, including macroglossia, transient hypoglycemia, and umbilical hernia. So far, his serum iPTH, Ca, or IP levels were within the normal range (Fig. 2).

Patient 3 was a Japanese boy conceived naturally by healthy and unrelated parents. He was born by emergency cesarean section due to his mother’s gestational hypertension at 36 weeks of gestation. His BW was within the normal range of gestational age and sex-matched neonates. His newborn screening test showed low free thyroxine (T4) and normal TSH levels, but the retest of free T4 was within the normal range. As shown in Fig. 2 and Additional file 1: Supplemental Fig. 1, he showed rapid weight gain in infancy, and his BMI reached 90%ile at 3 months of age, and a round face became prominent. His BWSp score was four points, including macroglossia, ear creases and pits, and umbilical hernia. He had an elevated TSH level since the age of 1 year and an elevated iPTH level since the age of 2 years, but he did not receive medication (Fig. 2).

Molecular studies

The results of methylation-specific multiple ligation-dependent probe amplification (MS-MLPA) analysis targeting multiple ID-related DMRs using genomic DNA from patient leukocytes are shown in Fig. 3. All three patients showed aberrant methylation of the GNAS-DMRs: hypermethylation of the NESP-DMR and hypomethylation of the A/B-DMR, AS1-DMR, and XL-DMR, without methylation defects in the H19-DMR and KCNQ1OT1-DMR. To evaluate the methylation levels of the H19-DMR and KCNQ1OT1-DMR in tissues other than leukocytes, we conducted methylation analysis using pyrosequencing on samples of buccal mucosa. Methylation analysis showed very slight hypomethylation of the KCNQ1OT1-DMR in Patient 2 and high normal methylation levels of the H19-DMR in Patient 3 together with extremely low methylation levels of the A/B-DMR in Patients 1–3 (Fig. 3). Array-based methylation analysis for all known DMRs using the Infinium MethylationEPIC Kit (EPIC) (Illumina) revealed that Patients 1 and 2 showed methylation defects in several DMRs other than the GNAS-DMRs. Patient 3 had the lower limit of normal methylation in the XL-DMR and no methylation defects in the DMRs other than the NESP-DMR, AS1-DMR, and A/B-DMR (Fig. 4, Additional File 2: Supplementary Table 1). Patients 1 and 2 had aberrant methylation of the DMRs commonly in the DIRAS3:EX2-DMR, DIRAS3:TSS-DMR, JAKMIP1:Int2-DMR, SVOPL:alt-TSS-DMR, MCTS2P:TSS-DMR, NNAT:TSS-DMR, L3MBTL1:alt-DMR, WRB:alt-TSS-DMR, and SNU13:alt-TSS-DMR, but aberrant methylation patterns in these DMRs were not identical. Furthermore, we analyzed methylation array data using a more stringent analysis method (analysis method 2, see Methods section), and the DMRs related to PYHIN1 and OBSCN, SVOPL:alt-TSS-DMR and WRB:alt-TSS-DMR in Patient 1 and the DIRAS3:Ex2-DMR, JAKMIP1:Int2-DMR, WDR27:Int13-DMR, SVOPL:alt-TSS-DMR, ZNF597:3′-DMR, WRB:alt-TSS-DMR, and SNU13:alt-TSS-DMR in Patient 2, classified as normally methylated DMRs (Fig. 4, Additional File 2: Supplementary Table 1).

Fig. 3.

Results of methylation analysis using MS-MLPA and pyrosequencing. MS-MLPA shows copy number changes (upper panel) and methylation levels (lower panel). Probe ratios below 0.7 (red line) and above 1.3 (blue line) show a heterozygous deletion and a heterozygous duplication, respectively (upper panel). Probe ratios below the mean of methylation values among reference samples –0.3 (red line) and above the mean of methylation values among reference samples + 0.3 (blue line) show loss and gain of methylation (lower panel), respectively. A Patient 1, B Patient 2, C Patient 3. D Results of methylation analysis on samples of the buccal mucosa of Patients 1–3 using pyrosequencing for nine DMRs related to major imprinting disorders. Gray vertical bars indicate the normal ranges of MIs (minimum–maximum in 50 control subjects) in blood. MS-MLPA, methylation-specific multiple ligation-dependent probe amplification; H19, H19/IGF2:IG-differential methylated region (DMR); MEG3, MEG3:TSS-DMR; NESP, GNAS-NESP:TSS-DMR; PLAGL1, PLAGL1:alt-TSS-DMR; GRB10, GRB10:alt-TSS-DMR; MEST, MEST:alt-TSS-DMR; KCNQ1OT1, KCNQ1OT1:TSS-DMR; MEG8, MEG8:TSS-DMR; SNURF, SNURF:TSS-DMR; PEG3, PEG3:TSS-DMR; AS, GNAS-AS1:TSS-DMR; XL, GNAS-XL:Ex1-DMR; A/B, GNAS-A/B-TSS-DMR; PEG10, PEG10:TSS-DMR; and IG, MEG3/DLK1:IG-DMR

Fig. 4.

The results of array-based methylation analysis using the Infinium MethylationEPIC Kit (Illumina). The heatmap indicates 78 examined DMRs. Each row represents a DMR, each column represents a patient. Methylation defects in DMRs are classified into nine categories based on the degree. Warm colors show gain of methylations, and cool colors show loss of methylations using analysis method 1. The degree of color intensity indicates the degree of abnormal methylation level. Black and gray boxes represent hypomethylation and hypermethylation, respectively, as determined by analysis method 2. DMR, differentially methylated region; Chr, chromosome; Pt, patient; SD, standard deviation; MML, median methylation level; 3 SD, array-based methylation analysis using analysis method 1; and CH-t, array-based methylation analysis using analysis method 2

We conducted WES on Patient 1 and her mother, Patient 2 and his parents, and Patient 3, respectively, and examined the pathogenic variants in the known or candidate MLID-causative genes, genes related to BWS-like phenotypes, and causative genes related to known genetic disorders (see Methods). There were no known pathogenic variants in genes leading to BWS-like phenotypes, methylation defects in the DMRs, and known genetic disorders [1, 7, 8, 12]. Because WES in Patient 1 showed loss of heterozygosity on chromosome 9, we conducted array comparative genome hybridization analysis and identified a heterozygous deletion on 9pter-p24.2, which ranged approximately 4.4 Mb including WASHC1, FOXD4, CBWD1, DOCK8, KANK1, DMRT1, DMRT2, DMRT3, SMARCA2, VLDLR, KCNV2, PUM3, RFX3, and GLIS3 (Fig. 5). Next, we conducted microsatellite analysis using genomic DNA from Patients 2 and 3 and their parents to determine etiologies of the methylation defects in the GNAS-DMRs, as previously reported [13, 14]. Patient 2 had only paternally inherited chromosome 20, suggesting paternal uniparental disomy of chromosome 20 (UPD(20)pat). Patient 3 showed biparental origin of chromosome 20, suggesting epimutation (Additional File 3: Supplementary Table 2).

Fig. 5.

Results of array-based comparative genomic hybridization analysis in Patient 1. Black dots show normal (log₂ ratio from + 0.5 to − 1.0) copy numbers and green dots show low (log₂ ratio less than − 1.0) copy numbers

Discussion

In this study, three patients with the BWS-phenotype having methylation defects in the GNAS-DMRs but no molecular abnormalities in the 11p15.5 imprinted region were identified for the first time. The previously reported 31 cases with methylation defects in the A/B-DMR showing the BWS-phenotype also had hypomethylation of the KCNQ1OT1-DMR [6, 7, 9–11, 15–17]. Our three patients had BWSp scores of four or more meeting the clinical diagnostic criteria of classic BWS with macroglossia and postnatal rapid weight gain, and showed a round face, obesity, elevated TSH levels, and low normal free T4 levels as PHP-suggestive clinical features. Because macroglossia has been reported in five PHP1B cases with methylation defects only on the GNAS locus [18–20], macroglossia may be rarely observed in cases with PHP1B. In addition, postnatal rapid weight gain with high normal birth weight is frequently observed in PHP1B [4, 21]. The high prevalence of BWS (1/10,340) [1] compared to the prevalence of PHP (0.34–1.1/100,000) [3], and the presence of macroglossia frequently observed in patients with BWS may result in the diagnostic bias by attending physicians. As shown in Fig. 2, we identified an elevated iPTH level only in Patient 3. It is difficult to recognize an elevated iPTH level in the infantile period due to the gradual manifestation of PTH resistance after birth in patients with PHP [22]. For the early diagnosis of PHP1B in cases presenting with the BWS-phenotype, methylation analysis for the GNAS region should be considered in cases with the BWS-phenotype and no molecular defects in the 11p15.5 imprinted region.

Patient 1 with 9p deletion and Patient 2 with UPD(20)pat had multiple methylation defects in DMRs other than the GNAS-DMRs. Patient 1 showed hypomethylation of the DIRAS3:EX2-DMR, DIRAS3:TSS-DMR, the DMRs related to PYHIN1 and OBSCN, JAKMIP1:lnt-DMR, SVOPL:alt-TSS-DMR, WRB:alt-TSS-DMR, SNU13:alt-TSS-DMR, and hypermethylation of the MCTS2P:TSS-DMR, NNAT:TSS-DMR, and L3MBTL1:alt-TSS-DMR, these are DMRs in which methylation defects are observed in many cases with MLID [6, 7]. The SMARCA2 gene involved in the deleted region in Patient 1 encodes the central catalytic subunit of the chromatin remodeling complex [23]. The pathogenic missense variants and intragenic deletion in SMARCA2 lead to Nicolaides–Baraitser syndrome, which has a disease-specific episignature including hypomethylation of the DMR related to OBSCN observed in Patient 1 [24]. Methylation defects in the DMRs in cases with 9p deletion have not been reported other than in our patient. Further accumulation of the methylome analysis data in cases with 9p deletion is expected. Patient 2 showed methylation defects in the DMRs on chromosome 20 and hypomethylation of the WDR27:lnt13-DMR, and hypermethylation of the DIRAS3:Ex2-DMR, DIRAS3:TSS-DMR, JAKMIP1:lnt-DMR, SVOPL:alt-TSS-DMR, ZNF597:3′-DMR, WRB:alt-TSS-DMR, and SNU13:alt-TSS-DMR. Abnormal methylation of the DIRAS3:Ex2-DMR, DIRAS3:TSS-DMR, JAKMIP1:lnt-DMR, SVOPL:alt-TSS-DMR, MCTS2P:TSS-DMR, NNAT:TSS-DMR, L3MBTL1:alt-TSS-DMR, WRB:alt-TSS-DMR, and SNU13:alt-TSS-DMR were observed in Patients 1 and 2; however, these DMRs were abnormally hypomethylated in Patient 1 and hypermethylated in Patient 2. Patients 1 and 2 showed opposite abnormal methylation patterns in the DIRAS3:TSS-DMR, MCTS2P:TSS-DMR, NNAT:TSS-DMR, and L3MBTL1:alt-DMR by both analysis methods (Fig. 4, Additional File 2: Supplementary Table 1). Because methylation defects in the DIRAS3:TSS-DMR are frequently observed in cases with various IDs with both abnormal hypermethylation and hypomethylation [7, 9], these issues may cause opposite methylation patterns between Patients 1 and 2. Etiologies of Patients 1 and 2 were epimutation and UPD(20)pat, respectively. Hypomethylation of the MCTS2P:TSS-DMR, NNAT:TSS-DMR, and L3MBTL1:alt-TSS-DMR in Patient 2 is reasonable in UPD(20)pat. On the other hand, we speculated that hypermethylation of the MCTS2P:TSS-DMR, NNAT:TSS-DMR, and L3MBTL1:alt-TSS-DMR in Patient 1 is accompanying MLID. In MLID, loss of methylation of the DMRs is more common than gain of methylation of the DMRs, as previously reported [8]. The 9p deletion detected in Patient 1 may have some effect on the methylation status of the DMRs on chromosome 20, including the GNAS-DMRs. Patient 2 may have undetected genetic abnormalities causing the gain of methylation in the DMRs.

In Patient 3, the XL-DMR showed mild hypomethylation in MS-MLPA, and this DMR was classified as normal in array-based methylation analysis. The median methylation level (MML) of the XL-DMR in Patient 3 was slightly over the lower limit of the normal range of MML in the XL-DMR and was defined as normal. These findings suggest no significant difference in methylation levels between MS-MLPA and array-based methylation analysis using EPIC, although differences in methylation analysis method and the number of controls may result in this slight difference.

Methylation analysis using pyrosequencing on samples of buccal mucosa showed slightly low levels compared to the normal range of the methylation index of the KCNQ1OT1-DMR in leukocytes from Patient 2. In a previous study, methylation analysis in the skin or peritumoral tissue of cases with BWS having no 11p15.5 methylation defects in leukocytes showed 11p15.5 methylation defects in 35.7% (5/14) of them [25]. Thus, it is difficult to judge whether slight hypomethylation of the KCNQ1OT1-DMR in the buccal cells of Patient 2 affects the development of the BWS-phenotype.

Some clinical features of cases with 9p deletion overlap with those of cases with BWS, such as ear abnormality, macroglossia, congenital hyperinsulinemic hypoglycemia, and hypothyroidism, which were identified in Patient 1 [26–28]. Of genes within the deleted region in Patient 1, SMARCA2 and RFX3 have been reported as causative genes for hyperinsulinemic hypoglycemia and GLIS3 for congenital hypothyroidism [28, 29]. Consistent with this, a previously reported case with 9p deletion syndrome had 10 points in BWSp score together with a neurodevelopmental disorder [12]. Neurodevelopmental disorders are inevitable in 9p deletion syndromes [26] but are not common in BWS. The BWS-phenotype observed in Patient 1 may be caused by the synergic effects of methylation defects in the GNAS-DMRs and 9p deletion. For cases meeting the BWS clinical diagnostic criteria complicated with neurodevelopmental disorder, attending physicians need to consider differential diagnoses such as 9p deletion. Intellectual disability was also observed in Patient 2, but not in Patient 3. The previous studies showed that cases with MLID had more neurodevelopmental disability than unaffected children [6, 7]. Therefore, the intellectual disability in Patient 2 may have been due to some effect of aberrant methylation in the non-ID-related DMRs, although the aberrant methylation of these DMRs may have had a small effect on phenotypes of Patient 2. Furthermore, in Patient 2, his prematurity may also be associated with intellectual disabilities and transient hypoglycemia, and transient hypoglycemia itself may also affect intellectual development.

This study has a limitation. Because of the mosaicism of the methylation defects in the DMRs in the 11p15.5 imprinted region, the unexamined tissues, such as tongue and skin, may affect the BWS-phenotype. Despite this limitation, our study is valuable in demonstrating the diversity of (epi)genetic causes of the BWS-phenotype.

Conclusion

Because rapid weight gain in early childhood and macroglossia are observed both in cases with PHP1B and BWS, we recommend conducting methylation analysis for the DMRs at the GNAS locus in cases with the BWS-phenotype but no molecular defects in the 11p15.5 imprinted region.

Methods

This study was approved by the Institutional Review Board Committee at the National Center for Child Health and Development.

Patients

For 199 patients with suspected BWS by their attending physicians, we conducted methylation analysis for the H19-DMR and KCNQ1OT1-DMRs in the 11p15.5 imprinted region using pyrosequencing, MS-MLPA (SALSA MS-MLPA Probe-mix ME030) (MRC-Holland, Amsterdam, Netherlands), and/or combined bisulfite restriction analysis as previously reported [30] and identified methylation defects in these DMRs in 101 patients. Of the 98 patients without methylation defects in 11p15.5, three patients and a single patient had the diagnosis of Angelman syndrome and Tatton-Brown syndrome, respectively, after additional analysis by attending physicians. We conducted mutation analysis for CDKN1C in 89 patients, excluding five patients without sufficient DNA samples, and detected a known pathological variant (p.Gln230Ter) in a single patient. Subsequently, we conducted methylation analysis in 77 patients using pyrosequencing for the PLAGL1:alt-TSS-DMR on chromosome 6, PEG10:TSS-DMR on chromosome 7, MEG3/DLK1:IG-DMR and MEG3:TSS-DMR on chromosome 14, SNURF:TSS-DMR on chromosome 15, and A/B-DMR on chromosome 20 [31]. For patients with methylation defects in the GNAS-DMRs, we conducted MS-MLPA (SALSA MS-MLPA Probe-mix ME034) (MRC-Holland) for targeting multiple ID-associated DMRs to confirm the methylation status by a different method. Furthermore, we conducted methylation analysis on the H19-DMR and KCNQ1OT1-DMR in samples from the buccal mucosa in the three patients as previously reported [32].

Array-based methylation analysis using EPIC

We conducted genome-wide methylation analysis using Infinium MethylationEPIC Kit (EPIC) (Illumina) with genomic DNA from leukocytes of three patients with methylation defects in the A/B-DMR on chromosome 20 and obtained β values indicating the methylation levels for 825 CpGs in 78 imprinted DMRs as previously reported [33]. We defined aberrantly methylated DMR based on previous reports [6, 7]. In brief, the median β value for each CpG within a DMR was determined as the MML of the DMR. An aberrantly hypermethylated DMR was defined as MML > + 3 SD obtained from the mean of MML in 16 healthy child controls. An aberrantly hypomethylated DMR was defined as MML < –3 SD. The aberrantly methylated DMRs were further categorized as mild (ΔMML < 0.1), moderate (0.2 ≥ ΔMML ≥ 0.1), and extreme (ΔMML > 0.2) by the difference between the MML of each patient and the mean MML of the controls (analysis method 1). In addition, we also analyzed the methylation array data by applying more stringent bioinformatic parameters using the Crawford–Howell t-test, as previously reported [7]. Briefly, a probe with an absolute ∆β value (|∆β|) over 0.1 and a false discovery rate less than 0.05 was defined as differentially methylated. When we detected two or more consecutive differentially methylated probes within a DMR (including at least four probes), we defined the DMR as aberrantly methylated (analysis method 2).

Whole-exome sequencing

We conducted WES in Patient 1 and her mother, Patient 2 and his parents, and Patient 3. We used SureSelect Human All Exon V6 (Agilent Technologies) for WES as previously reported [7]. We searched for a variant(s) of reported MLID-related genes (NLRP2, NLRP5, NLRP7, PADI6, KHDC3L, ZFP57, and ZNF445), candidate MLID-causative genes (OOEP, ZAR1, TLE6, ARID4A, UHRF1, NLRP14, DPPA3, DNMT3A, DNMT3B, DNMT3L, DNMT1, SETDB2, TRIM28, and WHSC1), genes leading to BWS-like phenotypes (GPC3, DIS3L2, HRAS, NSD1, EZH2, NFIX, PTEN, PIK3CA, SUZ12, and PIGW), and genes leading to congenital hyperinsulinism (ABCC8, KCNJ11, CACNA1D, SLC16A1, GLUD1, GCK, HADH, UCP2, HK1, PGM1, PMM2, HNF4A, HNF1A, and FOXA2) [1, 7, 8, 12, 34]. We extracted rare variants with minor allele frequencies of ≤ 0.01 in public databases and in-house database as previously reported [33]. We also searched for other causative genes for genetic diseases. We evaluated pathogenicity of identified rare variants using the following in silico analyses: (1) CADD (http://cadd.gs.washington.edu/), (2) PP2_HVAR (http://genetics.bwh.harvard.edu/pph2/), (3) SIFT (http://sift.jcvi.org/), and (4) MutationTaster (http://www.mutationtaster.org/).

Abbreviations

A/B

GNAS-A/B:TSS

AS1

GNAS-AS1:TSS-DMR

BMI

Body mass index

BW

Birth weight

BWS

Beckwith–Wiedemann syndrome

BWSp

BWS spectrum

Ca

Calcium

DMR

Differentially methylated region

EPIC

Array-based methylation analysis using Infinium MethylationEPIC Kit (Illumina)

GNAS-DMRs

DMRs on the GNAS locus

H19

H19/IGF2:IG

ID

Imprinting disorder

IP

Inorganic phosphate

iPTH

Intact PTH

KCNQ1OT1

KCNQ1OT1:TSS

MLID

Multi-locus imprinting disturbance

MML

Median methylation level

MS-MLPA

Methylation-specific multiple ligation-dependent probe amplification

NESP

GNAS-NESP:TSS-DMR

PHP1B

Pseudohypoparathyroidism type 1B

PTH

Parathyroid hormone

T4

Thyroxine

TSH

Thyroid-stimulating hormone

UPD

Uniparental disomy

UPD(20)pat

Paternal uniparental disomy of chromosome 20

WES

Whole-exome sequencing

XL

GNAS-XL:Ex1

Author contributions

TU performed the molecular and data analyses and wrote the paper. YK performed the molecular and data analysis. NA and AU obtained patients’ clinical information. MF reviewed the paper and supervised the project. TU and MK designed the project. MK wrote the paper and gave the final approval for the publication of the submitted version. All authors read and approved the final manuscript.

Funding

This work was supported by grants from the National Center for Child Health and Development (2022B-5), the Japan Society for the Promotion of Science (JSPS) (22K07858 [C]), the Japan Agency for Medical Research and Development (AMED) (23ek0109587h0002), the Naito Science & Engineering Foundation, the Japanese Society for Pediatric Endocrinology Future Development Grant supported by Novo Nordisk Pharma Ltd., and the Takeda Science Foundation.

Availability of data and materials

Data are provided within the manuscript or supplementary information files.

Declarations

Consent for publication

We obtained written informed consent from the patients or the patients’ parents to publish the patients’ clinical and molecular information.

Competing Interests

The authors declare no competing interests.