記住我
Research ArticleEndocrinologyGenetics
Open Access | 
10.1172/jci.insight.185874
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by
Li, D.
in:
JCI
|
PubMed
|
Google Scholar
|
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by Jan de Beur, S. in: JCI | PubMed | Google Scholar
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by Hou, C. in: JCI | PubMed | Google Scholar
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by Ruzhnikov, M. in: JCI | PubMed | Google Scholar
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by Seeley, H. in: JCI | PubMed | Google Scholar
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by
Cutting, G.
in:
JCI
|
PubMed
|
Google Scholar
|
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by Sheridan, M. in: JCI | PubMed | Google Scholar
1Center for Applied Genomics, and
2Division of Human Genetics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
3Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
4Division of Endocrinology and Metabolism, University of Virginia School of Medicine, Charlottesville, Virginia, USA.
5Neurology and Neurological Sciences, Pediatrics, Division of Medical Genetics, and
6Division of Pediatric Endocrinology, Stanford University and Lucile Packard Children’s Hospital, Palo Alto, California, USA.
7Department of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.
8Division of Endocrinology and Diabetes and The Center for Bone Health, The Children’s Hospital of Philadelphia, and Department of Pediatrics University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Address correspondence to: Michael A. Levine, Division of Endocrinology and Diabetes, The Children’s Hospital of Philadelphia, ARC 510A, 3615 Civic Center Boulevard, Philadelphia, Pennsylvania 19104, USA. Phone: 215.590.3618; Email: levinem@chop.edu.
Find articles by
Levine, M.
in:
JCI
|
PubMed
|
Google Scholar
|
Published November 14, 2024 - More info
Published in Volume 9, Issue 24 on December 20, 2024
AbstractPseudohypoparathyroidism type 1B (PHP1B) is associated with epigenetic changes in the maternal allele of the imprinted GNAS gene that inhibit expression of the α subunit of Gs (Gsα), thereby leading to parathyroid hormone resistance in renal proximal tubule cells where expression of Gsα from the paternal GNAS allele is normally silent. Although all patients with PHP1B show loss of methylation for the exon A/B differentially methylated region (DMR), some patients with autosomal dominant PHP1B (AD-PHP1B) and most patients with sporadic PHP1B have additional methylation defects that affect the DMRs corresponding to exons XL, AS1, and NESP. Because the genetic defect is unknown in most of these patients, we sought to identify the underlying genetic basis for AD-PHP1B in 2 multigenerational families with broad GNAS methylation defects and negative clinical exomes. Genome sequencing identified small GNAS variants in each family that were also present in unrelated individuals with PHP1B in a replication cohort. Maternal transmission of one GNAS microdeletion showed reduced penetrance in some unaffected patients. Expression of AS transcripts was increased, and NESP was decreased, in cells from affected patients. These results suggest that the small deletion activated AS transcription, leading to methylation of the NESP DMR with consequent inhibition of NESP transcription, and thereby provide a potential mechanism for PHP1B.
IntroductionPseudohypoparathyroidism type 1B (PHP1B; also termed inactivating PTH/PTHrP signaling disorder, iPPSD; ref. 1) is a disorder of hormone resistance in which the inability of parathyroid hormone (PTH) to activate endocrine signaling processes in target cells leads to a state of functional hypoparathyroidism (2–4). The pathophysiology of PHP1B is the result of genetic or epigenetic defects in the GNAS gene that reduce expression or function of the α subunit of the heterotrimeric stimulatory G protein (Gsα), a signaling protein that couples heptahelical receptors to activation of adenylyl cyclase (4). GNAS is a complex imprinted gene on chromosome 20q13.3 (5) that utilizes multiple alternative promoters and first exons to generate various transcripts based on the parent of origin of each allele (6–8) (Figure 1A). Gsα is the principal product of GNAS and is encoded by exons 1–13; Gsα is expressed from both parental alleles in most tissues, but in some cells such as proximal renal tubular epithelium, pituitary somatotrophs, gonads, thyroid epithelial cells, and regions of the central nervous system, transcription of Gsα occurs predominately from the maternal GNAS allele (9–17). The mechanism that accounts for allelic silencing of paternal Gsα in some tissues is uncertain, but restricted expression of additional GNAS transcripts is achieved by use of promoters that are in differentially methylated regions (DMRs) (Figure 1A). Transcription of NESP from the paternal GNAS allele is inhibited by methylation of CpG dinucleotides within the NESP DMR (GNAS-NESP:TSS-DMR), whereas generation of transcripts from exons XL, AS1, and A/B (termed exon 1A in mice) is inhibited from the maternal GNAS allele by methylation of the corresponding DMRs (GNAS-AS1:TSS-DMR, GNAS-XL:Ex1-DMR, and GNAS A/B:TSS-DMR) on that allele (7). In addition, a fifth DMR, termed GNAS-AS2:TSS-DMR (18) and which appears to consist of 2 subdomains (19), has recently been identified and is located telomeric of GNAS-AS1:TSS-DMR.
Figure 1Schematic depiction of the region from the STX16 gene to the GNAS complex locus and variants identified in patients with AD-PHP1B. (A) Schematically depicted locations of STX16, GNAS, AS, and putative GNAS ICRs. Boxes represent exons of STX16, GNAS, and AS, and filled black dots represent putative GNAS ICRs. (B) Distribution of microdeletions in patients with AD-PHP1B overlapping the NESP/AS region and 2 variants identified in this study. The deletion reported in Danzig et al. (28), presented as a dashed line, was confirmed via genome sequencing but was not specified in the original publication.
Imprinting of GNAS leads to variable phenotypes based on the parental origin of the defective allele. Mutations that affect exons 1–13 of GNAS and directly reduce expression or function of Gsα are the basis for tissue resistance to multiple hormones plus features of Albright hereditary osteodystrophy (AHO) in patients with PHP1A (OMIM 103580, also termed iPPSD type 2; ref. 2) when they are on the maternal allele and for pseudopseudohypoparathyroidism (PPHP; OMIM 612463), characterized by AHO only, when they lie on the paternal allele (20, 21). By contrast, PHP1B (OMIM 603233; also termed iPPSD3) is associated with characteristic epigenetic changes in the maternal GNAS allele that lead to reduced expression of Gsα. Clinically, this manifests most notably in the renal proximal tubule and thyroid follicular cells, and results in renal resistance to PTH and sometimes resistance to thyroid-stimulating hormone (TSH) (2–4, 19). To date, all patients with PHP1B have loss of methylation (LOM) of the maternal DMR for exon A/B. LOM can be restricted to exon A/B, as occurs in most familial forms of PHP1B (autosomal dominant PHP1B [AD-PHP1B]). Alternatively, A/B DMR LOM can be associated with broader methylation defects that include LOM at all 3 maternal DMRs (XL, AS1, and A/B) and gain of methylation (GOM) at the paternal DMR associated with NESP (22–25), as found in rare families (see below) or most frequently in patients with sporadic PHP1B (80%–85% of PHP1B patients). Although in some cases partial (23–25) or complete (4, 22, 23, 26, 27) paternal uniparental disomy (patUPD) for chromosome 20 has been found, the genetic basis for most cases of sporadic PHP1B remains unresolved (19, 28) and has been proposed to result from either a failure in imprint establishment during oogenesis (29) or a lack of imprint maintenance after fertilization (28).
Approximately 10% of patients with PHP1B show autosomal dominant transmission of the disorder that is associated with epigenetic defects on the maternally inherited GNAS allele. In most cases, this is associated with a methylation defect that is limited to LOM of the exon A/B DMR and is caused by overlapping heterozygous microdeletions within the maternal STX16 allele approximately 170 kb centromeric of the GNAS locus that include exon 4 and the adjacent portion of intron 4 (28, 30, 31). Exon 4 of STX16 presumably represents an imprint control region (STX-ICR) that exerts long-range cis-regulatory control of an ICR located in the NESP55 exon (NESP-ICR) that is required for methylation and transcriptional silencing of the GNAS exon A/B (32). In addition, a large inversion located 7,225 bp downstream of GNAS exon XL (33) and retrotransposon insertions located approximately 1,200 bp telomeric of GNAS exon XL have also been identified as causes of altered methylation that is limited to GNAS exon A/B (34, 35). By contrast, a small number of patients with AD-PHP1B have broad methylation defects with GOM at DMRs for NESP55, and LOM at the XL, AS1, and exon A/B. In most cases, this is due to heterozygous deletions affecting the NESP55 and/or AS4/AS3 exons on the maternal GNAS allele (28, 36–39) within a region that represents a second ICR (NESP-ICR) (Figure 1A) (32). Finally, broad methylation defects in some patients with AD-PHP1B have been associated with rare duplications or complex rearrangements that involve DMRs in the alternative first exon(s) (40, 41). Nevertheless, apart from these rare defects, the majority of patients with PHP1B, both familial and sporadic, with broad methylation defects at the GNAS cluster do not have an obvious genetic alteration. Here, we describe the identification of 2 previously unidentified and recurrent small variants within the NESP-ICR of GNAS in individuals with familial and sporadic PHP1B who have global methylation defects. Our findings suggest that these deletions activate expression of the maternal GNAS antisense transcript and thereby convert the maternal GNAS allele to a paternal epigenotype.
ResultsFamily 1. We performed genome sequencing (GS) on DNA samples from affected individuals III-1 and III-4 and confirmed normal sequences for GNAS exons 1–13 and the intervening introns. Due to our prior demonstration of linkage between the GNAS locus and AD-PHP1B in this family (42, 43), we systematically analyzed the 20q13.3 locus, where GNAS is located, for structural variations (SVs). Despite using multiple in silico algorithms (i.e., BreakDancer, Manta, Wham, CNVnator, and Lumpy), we were unable to identify an SV in either of the 2 affected patients. Next, we scrutinized the GNAS locus for relatively smaller insertions/deletions (indels) and duplications (i.e., <50 bp), which disclosed a heterozygous 6-bp deletion (20[GRCh37]:g.57419071–57419076delTTCATT) associated with a nearby single-nucleotide transversion (20[GRCh37]:g.57419082C>A) in cis in these 2 patients (Figure 2, A and B). The 2 variants are on the same allele, which was also confirmed by subsequent Sanger sequencing in the entire family (hereafter we refer to this defect as the 6-bp deletion). The 6-bp deletion was intronic to both NESP55 and AS1 — in the first intronic region of the NESP55 transcript and the second intronic region of the antisense AS1 transcript (Figure 1). The variant was absent in 1000 Genomes Project, gnomAD v2 with 15,708 genomes, gnomAD v4 with 76,215 genomes, and additional genome-sequencing data from over 10,000 samples in the Human Longevity database. Furthermore, we analyzed nucleotide sequences of the deletion and surrounding 20 bp. It is noteworthy that there are 6-bp inverted repeat sequences flanking the deletion that could form a small stem-loop structure with one mismatch (Figure 2C, green marked). The 6-bp deleted sequence (20[GRCh37]:g.57419071–57419076delTTCATT) resides at the loop of the putative structure (Figure 2C, red highlighted) and the base change (20[GRCh37]:g.57419082C>A) at the stem of the structure could have occurred during DNA replication in the context of slipped-strand mispairing (44).
Figure 2Pedigree of family 1, identified 6-bp NESP55/NESPAS intronic deletion, and proposed mechanism of deletion formation. (A) Pedigree of family 1 with AD-PHP1B. *Tested negative for the NESP55/NESPAS intronic variant (20[GRCh37]:g.57419071–57419076delTTCATT) associated with a nearby single-nucleotide transversion (20[GRCh37]:g.57419082C>A) in cis. **Tested positive for the NESP55/NESPAS intronic variant. (B) Sanger sequencing showed the presence of the heterozygous NESP55/NESPAS intronic variant. (C) Proposed stem-loop structure with 1 mismatch to explain the variant formation. N.I., normal imprinting (methylation); L.O.I., loss of imprinting (abnormal methylation); PTH resist, resistance to parathyroid hormone.
In addition to III-1 and III-4, the other 2 affected members of this family (proband II-4 and III-3) also carried the variant (Figure 2A). All affected members of generation III had inherited the variant from their mother and had broad defects in methylation. DNA was not available from individuals I-1 and I-2, however, so it was not possible to identify the parent of origin of this defect.
Two of the proband’s unaffected sisters (II-2 and II-6), an unaffected great niece (IV-2), and an unaffected great nephew (IV-1) had inherited the mutation on a maternal GNAS allele, but had completely normal methylation patterns and lacked evidence of PHP1B, consistent with our original proposal that there is incomplete penetrance of this GNAS defect (42, 45). These observations, plus the absence of any other genetic variations in the GNAS locus, provide additional evidence that the 6-bp defect is associated with AD-PHP1B in affected members of this family.
Family 2. The molecular basis for PHP1B in kindred 2 (Figure 3A) was not disclosed by previous exome sequencing. We performed GS on DNA samples from all 6 available family members and identified a point variant (20[GRCh37]:g.57,416,931G>A; Figure 3B) in the first intronic region of the NESP55 transcript and near the donor splice site of AS1 exon 4 (NR_002785.2: n.819+61C>T; Figure 1) in all 3 affected individuals with broad methylation defects (III-1, III-2, and III-3) and their unaffected mother (II-2) with normal methylation of DMRs, consistent with maternal transmission of PHP1B. DNA samples were not available from the parents of II-2, who presumably carries the variant on a paternal GNAS allele. Both the unaffected father (II-1) and unaffected sister (III-4) had wild-type GNAS alleles and normal methylation patterns (see below).
Figure 3Pedigree of family 2 and identified intronic point variant. (A) Pedigree of family 2 with AD-PHP1B. *Tested negative for the point intronic variant. **Tested positive for the point intronic variant. N.I., normal imprinting (methylation); L.O.I., loss of imprinting (abnormal methylation); PTH resist, resistance to parathyroid hormone. (B) Sanger sequencing in available family members of family 2. (C) A UCSC screenshot showing the intronic point variant overlapping with a cis-regulatory element (ccRE) annotated by ENCODE (accession number EH37E1205658).
This variant was not present in the 1000 Genomes Project, gnomAD datasets (v2 and v4), and the Human Longevity database. To further annotate the variant, we queried ENCODE data and found the variant is in a cis-regulatory element (ccRE; ENCODE accession number EH37E1205658) with the GNAS gene nearby. This ccRE has maximum DNase, H3K4me3, H3K27ac, and CTCF z scores of 2.25, 4.99, 2.75, and 1.48, respectively, highly suggested it has a promotor-like property regulating GNAS expression (Figure 3C). However, since ENCODE data were primarily developed to annotate the relationship between ccREs and protein-coding gene expression, this ccRE’s effect on noncoding gene expression is unknown.
Replication cohort. To further investigate the prevalence of the 2 genetic defects that we identified, we performed a replication study using an independent cohort of individuals with PHP1B. We performed Sanger sequencing on 64 additional unrelated individuals with PHP1B who were not previously reported in the literature and who have broad methylation defects (see above). We identified 2 additional patients, 3472 (II-1 in family 3, Figure 4A) and 6891-35 with apparent sporadic PHP1B, who carried the same complex 6-bp deletion (associated with the transversion in cis) as members of family 1 (Figure 4B and Figure 5), indicating that this small deletion is an uncommon but recurrent variant in PHP1B patients with global methylation defects. DNA was available from additional unaffected relatives of individual II-1 (family 3, Figure 4A) and Sanger sequencing demonstrated the same 6-bp deletion in his unaffected mother, brother, and nephew (Figure 4, A and B). DNA methylation analysis (Figure 5) showed that the affected individual (II-1) had a global defect in methylation that is consistent with inheritance of a defective GNAS allele from his unaffected mother (I-2), who had normal methylation status. By contrast, the proband’s brother, II-2, who had also inherited the defective GNAS allele from their mother, was unaffected and had normal methylation on GNAS DMRs. Finally, the proband’s unaffected nephew, individual III-1, also had normal methylation of the defective GNAS allele that he had inherited from his father, II-2. Therefore, the genetic and epigenetic findings in family 3 confirm and extend the findings in family 1 and further illustrate that maternal inheritance of this 6-bp deletion does not always lead to the methylation defects that cause PHP1B.
Figure 4Pedigree of family 3 and the identified 6-bp deletion. (A) Pedigree of family 3 with AD-PHP1B. **Tested positive for the 6-bp deletion (20[GRCh37]:g.57419071–57419076delTTCATT) associated with a nearby single-nucleotide transversion (20[GRCh37]:g.57419082C>A) in cis. (B) Sanger sequencing in available family members of family 2. N.I., normal imprinting (methylation); L.O.I., loss of imprinting (abnormal methylation); PTH resist, resistance to parathyroid hormone.
Figure 5Methylation analyses. Colors were used to indicate methylation levels: red for marked reduction, blue for marked increase, and yellow for normal levels. Shaded red or blue represent moderate reductions or increases in methylation.
We also identified 3 additional unrelated patients, 6891-1, 6891-23, and 6891-6, with sporadic PHP1B, who carried the same point variant (20[GRCh37]:g.57,416,931G>A) in the first intronic region of the NESP transcript and near the donor splice site of AS1 exon 4 as members of family 2 (Figure 5).
Methylation analysis of DMRs. We performed a comprehensive analysis of the methylation status of CpG sites within 6 regions that correspond to the 5 GNAS DMRs (Figure 1) for a control group of normal individuals, members of families 1 and 2, a replication cohort of patients with PHP1B who had global methylation defects, and members of family 3 who were related to the affected proband included in the replication cohort (Figure 5). DNA from healthy individuals showed 41%–52% methylation at DMRs corresponding to NESP, AS1, XL, and A/B. Remarkably, normal individuals showed moderate reductions in methylation at GNAS-AS2-1:TSS-DMR and GNAS-AS2-2:TSS-DMR, while individuals with sporadic PHP1B and global methylation defects showed marked reductions in methylation at these 2 sites, similar to findings reported by Hanna et al. (19). All patients in our study with sporadic PHP1B or AD-PHP1B had global methylation defects and showed increased methylation at GNAS-NESP:TSS-DMR and marked reductions in methylation at GNAS-AS1:TSS-DMR and both regions of the GNAS-AS2-2:TSS-DMR. By contrast, Hanna et al. (19) reported apparently increased methylation at GNAS-AS2:TSS-DMR in 3 affected members of an AD-PHP1B family who did not have an STX16 deletion, but these individuals differed from those whom we studied as they did not have evidence for global methylation defects and showed normal methylation at GNAS-NESP:TSS-DMR, GNAS-XL:EX1-DMR, and GNAS-AS1:TSS-DMR.
Quantification of GNAS transcript expression. We assessed the effect of the 6-bp deletion on transcription of AS and NESP by quantitative real-time reverse transcription PCR (RT-qPCR). These quantitative analyses showed that PHP1B individuals with the 6-bp deletion had significantly greater levels of the AS transcripts (Padj < 0.0001; Figure 6A) and lower levels of the NESP transcripts (Padj < 0.05; Figure 6B) in lymphoblastoid cell lines (LCLs) than did control normal individuals, which was consistent with the loss of imprinting at the AS promoter. By contrast, unaffected relatives in family 1 with the 6-bp deletion (carrier) or with wild-type GNAS alleles had normal levels of AS and NESP transcripts (Figure 6), consistent with normal methylation at the XL/AS DMR. Finally, LCLs from PHP1B patients with STX16 deletions, who have normal methylation at the AS promoter, showed normal levels of AS and NESP transcripts.
Figure 6GNAS transcript expression. RT-qPCR showing higher levels of AS transcripts (A) and lower levels of the NESP transcripts (B) in patients with 6-bp deletion (n = 5) compared with patients with pathogenic STX16 deletion (n = 7), unaffected 6-bp deletion carriers (n = 2), or unrelated controls (n = 10). Values were normalized to 10 unaffected controls. Normalized values are illustrated by the box-and-whisker plot, where the center line represents the median, the box limits represent the interquartile range, and the whiskers represent the minimum to maximum data range. Comparisons between multiple groups were done by 1-way ANOVA followed by Dunnett’s test using GraphPad Prism. *Padj < 0.05; ****Padj < 0.0001.
DiscussionPHP1B is now recognized as an epigenetic disorder in which expression of Gsα from the maternal GNAS allele is disrupted. Several mechanisms have been identified as the cause of epigenetic diseases, including stable changes in DNA methylation, posttranslational histone modification, and/or production of noncoding RNA. These defects are termed epimutations when they are directly involved as the molecular basis of the disease and can be separated into 2 types, primary and secondary, the latter occurring as the result of a DNA variant that affects a cis- or trans-acting factor. The genetic basis for the aberrant methylation of DMRs in PHP1B appears to be heterogeneous. Recurrent deletions within the maternal STX16 allele that include exon 4 and which disrupt a putative ICR termed STX16-ICR represent the predominant cause of AD-PHP1B and are associated with an imprinting defect that is limited to LOM at exon A/B (32). By contrast, the genetic basis for most cases of sporadic PHP1B, as well as some cases of AD-PHP1B, that are associated with more extensive imprinting defects affecting methylation at all DMRs has been more elusive. Rare maternally inherited heterozygous deletions involving the NESP55 exon and/or AS exons and introns have been identified in some PHP1B patients with broad methylation defects at the GNAS cluster (summarized in Figure 1), as these deletions presumably affect a primary ICR, termed NESP-ICR, that controls DMRs throughout the GNAS cluster (32). These observations have led to the notion that primary somatic epimutations rather than genetic defects account for the imprinting abnormalities in most PHP1B with broad methylation defects (28). Here we describe the application of GS to identify 2 small, intronic GNAS variants in patients with PHP1B who have extensive methylation defects. One variant is a complex 6-bp deletion (20[GRCh37]:g.57419071–57419076delTTCATT) associated with a nearby single-nucleotide transversion (20[GRCh37]:g.57419082C>A) that is intronic to both NESP55 and AS and is within the NESP-ICR. We found this variant on the maternal GNAS alleles of all affected individuals of 2 AD-PHP1B kindreds, families 1 and 3, as well as 1 individual with apparently sporadic PHP1B. Notably, this variant showed reduced penetrance in both familial cases, as not all individuals who carried the variant on a maternal GNAS allele manifested the imprinting defect that causes PHP1B. The identification of this variant in family 1 provides molecular confirmation of our previous linkage analysis that first showed a lack of concordance between aberrant methylation and a putative genetic defect within the maternal GNAS locus (42). A similar lack of co-segregation between aberrant methylation and a small deletion within the NESP-ICR of the maternal GNAS allele was subsequently reported in another PHP1B family with broad methylation defects (37). The incomplete penetrance of these small deletions may indicate that they are located near a boundary for the cis-acting element on the maternal GNAS allele that is required for methylation of the primary ICR for GNAS.
We also identified a point variant in affected members of the second multigenerational family with AD-PHP1B as well as 3 additional individuals with apparently sporadic PHP1B that is intronic to AS2 and AS3 exons. Therefore, this appears to be a second small variant that can cause broad methylation defects. In both cases, the identification of these variants in other, nonrelated patients with PHP1B suggests that the 2 changes are either recurrent pathogenic variants or very rare deleterious polymorphisms.
What can be the basis for the extensive imprinting defect within all GNAS DMRs, with GOM at the DMR for NESP55 and LOM at the DMRs for XL, AS1, and exon A/B, in these patients with such small mutations? Nearly all previously described PHP1B patients with broad methylation defects had larger deletions that included the NESP55 exon and/or AS exons on the maternal GNAS allele (28, 36–39) (Figure 1). By removing the unmethylated maternal NESP55 DMR, these deletions lead to an artifactual GOM at the DMR for NESP55 and disrupt normal transcription of NESP55, which is maternally expressed for approximately 80 kb across the entire GNAS locus. More importantly perhaps, it is likely that these deletions disrupt normal transcription of NESP55. Experimental studies have shown that transcription of NESP55 is required for proper methylation of downstream DMRs, as truncation (46, 47) or deletion (48) of Nesp in mice prevents acquisition of normal methylation at the downstream Nespas/Gnasxl and exon 1a DMRs in the oocyte. Therefore, deletions that include NESP55 and/or portions of AS exons lead to a broad methylation defect that has been termed H-L-L (or H-L-L-L), as shown in Figure 1. By contrast, discrete deletions that are limited to NESP55 (28, 49) result in a loss of imprinting that is limited to exon A/B (H-N-L; Figure 1), suggesting that transcription from the NESP55 promoter occurs but is prematurely terminated upstream of exon A/B.
The broad methylation defects in the PHP1B patients that we have studied are consistent with disruption of NESP55 transcription, but how can such small variants account for this profound defect in imprinting? Control of NESP55 transcription has been intensively studied over the past decade and important observations have demonstrated that transcription of AS regulates methylation of the NESP-ICR. The GNAS cluster contains 2 CpG islands with the characteristics of a germline ICR; these regions are differentially methylated in the germline and the differential methylation is maintained in the somatic tissues of the offspring. The principal ICR for the GNAS cluster contains the AS promoter (Figure 1A), and lies within the AS and XL DMR (50, 51); a second germline ICR is within the exon A/B DMR (52), and specifically controls maternal expression of A/B transcripts and the imprinted expression of transcripts encoding Gsα. Importantly, the third DMR in the GNAS locus is the NESP-ICR, which is a somatic DMR that becomes methylated on the paternal allele after fertilization as a consequence of transcription of AS (50, 52, 53). By contrast, acquisition of CpG methylation within the maternal AS ICR during oogenesis prevents transcription of AS in somatic cells after fertilization and thereby assures that the NESP DMR is not methylated, with consequent transcription of NESP from the maternal allele. As we had LCLs from only patients with the 6-bp deletion, we were unable to investigate the effect of both mutations on AS and NESP55 transcription. Nevertheless, our RT-qPCR data showed that transcription of NESP55 was markedly reduced only in affected patients who had increased methylation of the NESP-ICR. Certainly, lack of NESP transcription provides a cogent explanation for the loss of methylation at the downstream DMRs. Surprisingly, our data also showed that transcription of AS was increased in these same patients, which could provide an explanation for methylation of the maternal NESP-ICR. Thus, we propose that the methylation status of the maternal GNAS DMRs in the patients we describe here is consistent with acquisition of NESP-ICR methylation as a result of abnormal activation of AS transcription. Taken together, these observations provide a putative mechanistic framework to explain the epigenetic changes observed in PHP1B patients who have genetic defects that disrupt transcription of NESP55 upstream of exon A/B (33–35, 54) and align with the concept that DNA methylation is established in the oocyte in response to active transcription rather than by specific DNA sequence motifs or properties (55, 56). Thus, transcription initiating at the NESP55 promoter and proceeding through the GNAS locus is required for acquisition of maternally methylated germline DMRs at XL/AS and exon A/B.
While these studies were in progress, Iwasaki et al. reported the application of human embryonal stem cells (hESCs) as a human cellular model to investigate the mechanistic basis for epigenetic defects in PHP1B in general and loss of methylation at the A/B DMR in particular (32). Specifically, using hESC clones with maternal NESP-ICR ablation, they showed that the NESP-ICR, and NESP55 transcription, is required for methylation and transcriptional silencing of maternal A/B. However, it remains unexplained why this model failed to show hypomethylation at DMRs containing the AS and XL promoters, as occurs in PHP1B patients with defects in
留言 (0)