Intragenic Deletions of GNAS in Pseudohypoparathyroidism Type 1A Identify a New Region Affecting Methylation of Exon A/B

Abstract

Context

Pseudohypoparathyroidism type 1A (PHP1A) and pseudopseudohypoparathyroidism (PPHP) are caused by inactivating mutations in the exons of GNAS that encode the alpha-subunit of the stimulatory G protein (G_sα). In some cases abnormal methylation of exon A/B of GNAS, a hallmark of PHP1B, has been reported.

Objective

To identify the underlying genetic basis for PHP1A/PPHP in patients in whom molecular defects were not detected by GNAS sequencing and microarray-based analysis of copy number variations.

Methods

Whole genome sequencing (WGS) and pyrosequencing of differentially methylated regions (DMRs) of GNAS using genomic deoxyribonucleic acid from affected patients.

Results

We identified 2 novel heterozygous GNAS deletions: a 6.4 kb deletion that includes exon 2 of GNAS in the first proband that was associated with normal methylation (57%) of exon A/B DMR, and a 1438 bp deletion in a second PHP1A patient that encompasses the promoter region and 5′ untranslated region of G_sα transcripts, which was inherited from his mother with PPHP. This deletion was associated with reduced methylation (32%) of exon A/B DMR.

Conclusions

WGS can detect exonic and intronic mutations, including deletions that are too small to be identified by microarray analysis, and therefore is more sensitive than other techniques for molecular analysis of PHP1A/PPHP. One of the deletions we identified led to reduced methylation of exon A/B DMR, further refining a region needed for normal imprinting of this DMR. We propose that deletion of this region can explain why some PHP1A patients have reduced of methylation of the exon A/B DMR.

Pseudohypoparathyroidism type 1A (PHP1A) is a developmental disorder characterized by resistance to multiple hormones and a constellation of somatic defects termed Albright hereditary osteodystrophy (AHO; OMIM 103580), which includes short stature, brachydactyly, and heterotopic ossification. PHP1A results from heterozygous mutations within exons 1-13 of the maternal allele of GNAS, including both point mutations and rare structural rearrangements, that reduce the expression or function of G_sα, the alpha-subunit of the heterotrimeric stimulatory G protein that couples heptahelical receptors to activation of adenylyl cyclase (1, 2). By contrast, subjects with identical mutations on paternally inherited GNAS alleles manifest AHO but have normal hormone responsiveness, a condition termed pseudopseudohypoparathyroidism (PPHP) (3, 4). The variable expression of hormone resistance in different tissues and patients with GNAS mutations has been attributed to genomic imprinting of GNAS, a highly complex genetic locus on chromosome 20q13.3 (5-9) that contains multiple alternative first exons which encode novel proteins as well as exons that generate a regulatory antisense transcript (1, 2, 10-12). Utilization of these alternative first exons to generate additional transcripts is regulated by differentially methylated regions (DMRs) such that the extra-large variant of Gsα (XLαs), an amino terminus truncated form of Gsα utilizing exon A/B, and the antisense transcript (AS) are transcribed from the paternal GNAS allele while neuroendocrine secretory protein 55 (NESP55) is transcribed from the maternal allele (1, 2, 10-12). Although exon 1 does not have an associated DMR, transcription is subject to tissue-specific imprinting. G_sα is expressed from both parental alleles in most tissues but in some cells such as proximal renal tubular epithelium, pituitary somatotropes, gonads, thyroid epithelial cells, and regions of the central nervous system, transcription of G_sα is predominately from the maternal GNAS allele (13-19). Thus, tissue-specific paternal imprinting of G_sα accounts for the distinctive pattern of hormone resistance, obesity, and neurocognitive dysfunction in patients with PHP1A who carry maternal GNAS mutations (8, 9, 20, 21).

Epigenetic defects that reduce expression of G_sα from the maternal GNAS allele account for the development of parathyroid hormone (PTH) resistance in patients with pseudohypoparathyroidism type 1B (PHP1B) (10, 22). Subjects with PHP1B may also manifest mild thyrotropin (TSH) resistance, early-onset obesity, and brachydactyly, but typically lack the major features of AHO, especially heterotopic ossification (23-28). Most PHP1B cases are sporadic, and some can be explained by paternal uniparental disomy of chromosome 20q, which leads to global defects in methylation at all of GNAS DMRs (29). By contrast, the epigenetic defect in most cases of familial PHP1B is limited to loss of methylation at the exon A/B DMR, which is associated with heterozygous deletions within the maternal STX16 allele (30). A minority of familial PHP1B patients have global methylation defects at DMRs associated with the exons of NESP55, XLαs/AS, and A/B (ie, high–low–low methylation pattern) due to heterozygous deletions affecting the first exon of NESP55 and/or AS exons on the maternal GNAS allele (28, 31-33).

Since the initial identification in 1990 of heterozygous inactivating protein-coding mutations in G_sα in patients with PHP1A and PPHP (5-7), more than 200 pathogenic point mutations and small coding insertions/deletions (indels) have been described (reviewed in (34)). The majority of these mutations are nonsense or truncating mutations (loss of function), and intragenic deletions and rearrangements that affect GNAS exons are rarely reported as causes of PHP1A by Sanger sequencing and whole exome sequence (WES) analysis. Here we describe the application of whole genome sequencing (WGS) to identify the molecular basis for PHP1A/PPHP in 2 kindreds. We identified 2 novel deletions, including 1 with a breakpoint near exon A/B that also led to reduced methylation at exon A/B DMR.

Subjects and Methods

Patients

Patient 1 presented at age 21.4 years to the Medical Genetics Clinic at Spectrum Health for diagnostic evaluation of early-onset obesity, congenital hypothyroidism, brachydactyly (short first and fourth metatarsals and short fourth and fifth metacarpals), multiple subcutaneous ossifications, and mild learning difficulties. He was the second child of nonconsanguineous parents. He was born at 42 weeks’ gestation via C-section due to breech presentation. He required some respiratory resuscitation at birth but was discharged home at 3 days of life. The medical history was notable for postnatal growth failure, bilateral carpal tunnel syndrome that was treated by surgery, and right-sided hearing loss secondary to a cholesteatoma that was removed at age 8 years. At the time of evaluation his height was 153.3 cm (Z = −3.26) and weight was 81 kg (82% percentile), with body mass index of 34.1 (Z = 2.24). He has Horner syndrome and a history of asthma. His most recent laboratory findings while taking levothyroxine showed a normal TSH level of 0.29 µIU/mL (reference range: 0.27-4.2), a normal calcium level of 9.5 mg/dL (reference range: 8.6-10.3 mg/dL), and a significantly elevated level of intact PTH of 199 pg/mL (reference range: 15-65). He has an older sister (170 cm) and a younger brother (180 cm) who both have learning disorders but no endocrine defects or features of AHO. Father (191 cm) and mother (155 cm) were both normal. Sanger sequencing of GNAS exons 1-13 (Johns Hopkins DNA Diagnostic Laboratory) was normal. Genetic analysis using the high-resolution GGXChip + SNP v1.0 4-plex comparative genome hybridization (CGH) array revealed the patient to be 46XY with benign deletions noted at Xp22.31 (78.46 Kb), 3p12.3-p12.1 (7.61 Mb), and 5q14.3 (4.35 Mb). Methylation analysis of exon A/B DMR (Johns Hopkins DNA Diagnostic Laboratory) revealed normal methylation at 45% (normal >43%).

Patient 2 presented at age 20 months to the Endocrine Clinic at the Children’s Hospital of Philadelphia for evaluation and management of PHP1A. He was the first pregnancy to a 32-year-old mother and 32-year-old father; pregnancy was facilitated by in vitro fertilization owing to poor sperm motility. The mother developed gestational diabetes, which was treated by diet, and took enoxaparin for thrombosis. The patient was delivered at full term by C-section with a birth weight of 2.72 kg and length of 47 cm. The newborn screen showed slightly elevated TSH with normal total serum T4 and he was subsequently begun on daily levothyroxine. Soon after birth he was noted to have multiple subcutaneous ossifications (right leg, right chest, and head, bottom of left foot). Postnatally, he grew along the 15th percentile for length and the 98th percentile for weight, with a body mass index of 21.7 kg/m² (>99.98% percentile; Z = 2.72 standard deviation score), until age 5.5 years, when his weight was Z = 2.8 standard deviation score and height at the 50% to 60% percentile. At age 5.5 years his physical examination was significant for subcutaneous ossifications but no brachydactyly; he had some speech delay and has received school-based enrichment and therapy. His serum levels of calcium, phosphorus, and intact PTH were normal until age 2.5 years, when his intact PTH increased to 218 pg/mL (reference range: 15-65) with corresponding serum levels of calcium of 8.8 mg/dL and phosphorus of 7.6 mg/dL. He was subsequently treated with calcitriol and calcium carbonate. He began treatment with growth hormone at age 3 years after clonidine/arginine-stimulated growth hormone testing showed a subnormal peak growth hormone value of 3.8 ng/mL. Molecular analyses were performed at the Johns Hopkins DNA Diagnostic Laboratory, where sequence analysis of GNAS exons 1-13 was normal and methylation analysis of the GNAS exon A/B DMR was reduced at 34% (normal >43%). A single nucleotide polymorphism microarray was negative for causative copy number variations. His bone age has been normal. Research whole exome sequencing had failed to disclose a causative mutation (data not shown).

The proband’s mother had been diagnosed with AHO/PPHP at age 17 years, and was noted to have brachydactyly at age 7 to 8 years. She has multiple subcutaneous ossifications and intermittently mildly elevated TSH, for which she had taken levothyroxine. Her serum levels of PTH, calcium, and phosphorus were normal. She had menarche at age 12 years and has regular menses. The father and a 13-month-old sister are healthy.

These studies were approved by the Institutional Review Boards of the participating hospitals and all subjects provided written informed consent/assent for participation in the study.

CpG methylation analysis

CpG methylation analysis was determined using genomic deoxyribonucleic acid (DNA) that had been purified from peripheral blood mononuclear cells. Methylation status of CpG dinucleotides was determined by PCR amplification of bisulfite-modified, RNase-treated genomic DNA, followed by pyrosequencing using assays designed by EpigenDx (Worcester, MA). Nucleotides sequences in GNAS (20q13.32) that were analyzed were in the DMRs for NESP55 (ADS471; 20(GRCh37):g.57415807-57415853), XLαs (ADS470; 20(GRCh37):g.57429235-57429362), and exon A/B (ADS464; 20(GRCh37):g.57464773-57464927). We also used assay ADS1410 to analyze methylation of the DMR of the imprinted gene NNAT encoding neuronatan on chromosome at 20q11.23 (20(GRCh37):g.36149607-36152092). Experimental conditions are available upon request. Bisulfite pyrosequencing is a sequencing-by-synthesis method used to quantitatively determine the methylation of individual CG cytosines from PCR amplicons of a region up to 115 bases in length. The procedure relies on prior bisulfite conversion of all potentially methylated CpG cytosines to either cytosine (methylated) or thymine (unmethylated) and involves the stepwise incorporation of deoxynucleotide triphosphates into the growing strand of nascent DNA. The incorporation of these deoxyribonucleotides (dNTPs) results in the proportional release of pyrophosphate, which is then converted to adenosine triphosphate. The procedure is quantitative, and is based upon adenosine triphosphate facilitating a catalytic reaction by firefly luciferase in which luciferin is converted to oxyluciferin, and the amount of light released is proportional to the number of nucleotides incorporated. Methylation is expressed as the percentage of cytosines within the CpG sequences that are methylated.

Whole genome sequencing and bioinformatic analysis

We performed WGS on the 2 unrelated patients with PHP1A using Illumina HiSeq X Ten with paired-end 150 bp reads. Libraries were generated from genomic DNA using the Illumina TruSeq DNA PCR-Free Library Prep Kit (Illumina, San Diego, CA). All the raw reads were aligned to the reference human genome using the Burrows–Wheeler Aligner (BWA-Mem) (35) and single-nucleotide variants and small indels were captured using the Genome Analysis Tool Kit (36). ANNOVAR (37) and SnpEff (38) were subsequently used to functionally annotate the variants. BAM files generated by the Genome Analysis Tool Kit were fed to multiple short-read structural variant callers, including BreakDancer (39), Manta (40), Wham (41), and CNVnator (42), to capture structural variations (SVs) with default parameters. Similarly, the split and discordant reads files were generated by SpeedSeq (43) and were provided as inputs to Lumpy (44), another SV calling program.

Statistics

Pyrosequencing results are presented as the mean ± SD, and comparisons between groups were analyzed by analysis of variance using Tukey–Kramer multiple comparison tests using GraphPad InStat 3 (San Diego, CA).

Results

The average WGS coverage depth for patient 1 and patient 2 was 32- and 33-fold, respectively. Analyses of single-nucleotide variants and small indels did not reveal any rare disease-causing variants in GNAS exons 1-13 and exon–intron junctions in the 2 patients, which confirmed the previously uninformative results by Sanger sequencing and WES. Similarly, prior comparative genome hybridization had failed to identify a pathogenic deletion. We then focused on structural variation analysis at 20q13.3, where GNAS is located (7). For patient 1, a novel heterozygous 6.4-kb intragenic deletion that includes exon 2 of GNAS was captured by Manta and Lumpy with the support of BreakDancer and CNVnator. With Sanger sequencing and subsequent bioinformatic analyses to infer the underlying molecular mechanism of deletion formation, we were able to map the 2 breakpoints within 20(GRCh37):g.57,467,466-57,467,470 (20:58,892,411-58,892,415, hg38 version) and 20(GRCh37):g.57,473,859-57,473,863 (20:58,898,804-58,898,808, hg38 version) that are associated with a direct 5 bp microhomology (GGAGG; Fig. 1A). Searching low copy repeats (LCRs), long interspersed nuclear elements, and short interspersed nuclear elements/Alus flanking both breakpoints of the deletion by using RepeatMasker (http://www.repeatmasker.org/cgi-bin/WEBRepeatMasker) did not reveal any pairs of elements consisting of both breakpoints. Taken together these results suggest that the deletion may have arisen by a template switching replicative mechanism such as fork stalling and template switching (FoSTeS) or microhomology-mediated break-induced replication (MMBIR) (45, 46). DNA was not available from other members of this proband’s family.

Figure 1.

Intragenic GNAS deletion analysis. (A) DNA sequences of the deletion junctions, including 5 and 9 bp microhomologies in shadow, showing the locations of deletion breakpoints in between. (B) The GNAS locus showing the positions of the forward (F1 and F2) and reverse (R1 and R2) primers used for polymerase chain reactions to demonstrate the microdeletions. (C) Gel showing the presence of amplified DNA segments corresponding to the presence of the GNAS deletions present in proband 1 and proband 2; a short extension time (20 seconds) of PCR was used to eliminate amplification of the wild-type alleles.

Similarly, WGS of DNA from proband 2 and his mother with PPHP, followed by bioinformatic analyses and Sanger sequencing, revealed that both patients had a small, heterozygous deletion (1438 bp) that involves partial loss of exon 1 and partial loss of the intronic region between exon A/B and exon 1 (Fig. 1A). His father and unaffected sister had wild-type alleles (Fig. 1B). The 1438 bp deletion was supported by all 5 SV detection programs. The proximal breakpoint of the deletion maps within the (GCC)n repeat that is located between 20(GRCh37):g.57,465,300 and 20(GRCh37):g.57,465,392 and the distal breakpoint maps within another (GCC)n repeat between 20(GRCh37):g.57,466,735 and 20(GRCh37):g.57,466,781, which allowed us to determine this deletion is likely mediated by the simple repeats, nonallelic homologous recombination (NAHR) mechanism (47), with deletion breakpoints located at 20(GRCh37):g.57,465,324-57,465,332 and 20(GRCh37):g.57,466,762-57,466,770. However, we observed a direct 9 bp microhomology (CGGCGGCGG; Fig. 1A) in the breakpoints, therefore FoSTeS or MMBIR represent alternative mechanisms for the deletion formation.

Because the proximal breakpoint of the second deletion is close to the DMR at exon A/B, but does not extend into these CpG sequences, we performed pyrosequencing to investigate the potential effect of either deletion on the imprinting status of the GNAS DMRs. Pyrosequencing analyses showed that control PHP1B patients with STX16 deletions had significantly (P < .01) less methylation at the exon A/B DMR than either normal subjects or patients with PHP1A (Table 1). By contrast, there were no differences between the groups for methylation at the other DMRs. Patient 2 showed a slightly reduced average methylation level (32%) at CpGs within exon A/B of GNAS (Fig. 1A and Table 1) compared with normal subjects and patients with PHP1A, but greater than in PHP1B patients who carry an STX16 deletion (Table 1). By contrast, we observed a normal methylation level (57%) in patient 1 at exon A/B of GNAS (Fig. 1A). For both patients, the level of methylation was similar at each CpG that we assayed (Table 2), either normal (patient 1) or reduced (patient 2). The methylation status of both NESP55- and XLαs-DMRs were normal in both patients (Table 1). In addition, methylation status of the nearby imprinted gene NNAT was also normal in all subjects (data not shown).

Table 1.

Methylation analyses of GNAS DMRs

	Methylation status of DMR, %a
	Exon A/B (Johns Hopkins)a	Exon A/Bb	XLαs	NESP55
Patient 1	45	57	38	41
Patient 2	34	32	38	40
Normal	>43
Normal n = 7		64 ± 5	42 ± 2	40 ± 16
PHP1A, n = 3		63 ± 4	42 ± 5	39 ± 2
PHP1B (STX16 del) n = 9		3 ± 5	43 ± 4	43 ± 5

^aJohns Hopkins assesses cytosine methylation at CpG at chr20:57,464,763-57,464,814 on hg19. See Methods.

^bCHOP pyrosequencing assesses cytosine methylation at CpG at chr20:57,464,773-57,464,927 on hg19;

Table 2.

Human GNAS methylation analysis

	ADS464FS1																		ADS464FS2
From ATG	49612	49619	49627	49629	49634	49636	49642	49650	49652	49655	49662	49677	49682	49687	49691	49698	49707	49713	49732	49741	49743	49748	49757	49766
From TSS	50001	50008	50016	50018	50023	50025	50031	50039	50041	50044	50051	50066	50071	50076	50080	50087	50096	50102	50121	50130	50132	50137	50146	50155
GRCh38/hg38 Chrom 20	58889718	58889725	58889733	58889735	58889740	58889742	58889748	58889756	58889758	58889761	58889768	58889783	58889788	58889793	58889797	58889804	58889813	58889819	58889838	58889847	58889849	58889854	58889863	58889872
GRCh37/hg19 Chrom 20	57464773	57464780	57464788	57464790	57464795	57464797	57464803	57464811	57464813	57464816	57464823	57464838	57464843	57464848	57464852	57464859	57464868	57464874	57464893	57464902	57464904	57464909	57464918	57464927
Patient 1	55.7	52.1	57.1	60.4	55.9	53.9	59.3	57.4	52.5	55.3	56.7	56.2	53.6	52.9	56.8	54.5	58.1	59.8	56.7	60.3	56.9	61.0	58.5	56.7
Patient 2	36.8	33.9	37.2	37.2	37.4	31.9	38.1	35.7	34.4	33.3	34.0	32.5	24.8	30.8	29.4	21.5	25.3	31.9	29.1	28.8	27.1	28.8	26.6	36.7
Low	7.4	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	5.6	0.0	0.0	0.0	6.7	0.0	12.4	5.3	5.6	2.3	0.0	0.0
Med	80.8	75.1	79.9	77.0	74.5	72.5	85.1	79.2	72.7	71.9	85.3	81.9	70.9	72.6	77.4	64.0	83.8	85.0	81.5	78.5	76.6	83.8	68.2	80.0
High	90.7	86.7	99.5	93.0	90.1	93.6	98.4	93.1	89.7	87.9	99.1	100	92.7	90.7	91.3	94.3	93.3	91.0	94.9	92.5	95.6	93.0	89.9	98.4

Results in % methylation.

Discussion

We describe here 3 patients with PHP1A/PPHP from 2 families whom we found to carry small heterozygous GNAS intragenic deletions that each affect one exon. Patient 1 carries a 6393 bp deletion that includes all of exon 2 (73 bp), and is predicted to result in a truncated, nonfunctional G_sα protein (p.G49Afs*27) that leads to PHP1A. The second heterozygous deletion, 20(GRCh37):g.57,465,324-57,466,761 (or extends to 20(GRCh37):g.57,465,333-57,466,770 because of repeat sequences on the both ends), removes the sequences in regulatory region of G_sα (including the putative promoter region and 5′ untranslated region) but leaves the reading frame intact without deleting any coding sequences. In general, no transcript is produced if the promoter is missing. We propose that this deletion impairs G_sα expression and results in PHP1A or PPHP.

It has been increasingly acknowledged that 4 major mechanisms (45-49) can account for the formation of structural gene variants, including NAHR, replicative mechanism (eg, FoSTeS and MMBIR), nonhomologous end-joining, and long interspersed nuclear element-mediated retrotransposition or mobile element insertions. It has been appreciated that NAHR often causes recurrent deletion where the deletion is flanked by LCRs or Alu sequences, whereas the replicative mechanism is associated with nonrecurrent deletion (46, 49). For the 6.4 kb deletion, the presence of 5 bp microhomologies at the breakpoint junctions and absence of LCRs and Alu elements highly suggest that a replicative mechanism is responsible for the intragenic deletion. Due to repetitive sequences in the breakpoint junctions of the second deletion (1.4 kb), one may argue it could also be mediated by FoSTeS/MMBIR mechanism, as there are multiple microhomology sequences (eg, CCGCCGCCGCCGC or GCCGCC) in or near the breakpoints. However, considering the replication-based mechanism, break-induced replication and repair would generate more GCC repeats deletion in both ends and that would not match the Sanger sequencing results as shown in Fig. 1A. Thus, we speculate the LCRs flanking the breakpoint junctions are more likely the basis for NAHR-mediated deletion.

A variety of pathogenic GNAS mutations have been identified in patients with PHP1A/PPHP (34, 50, 51), including point mutations, small coding indels, exon inversion and deletions (52, 53), and a few chromosomal rearrangements (52, 54-56) (ie, big deletions covering the entire GNAS locus). However, so far, only 3 smaller intragenic deletions (eg, bigger than 1 kb but smaller than 50 kb) disrupting the exon A/B or being close to the exon A/B have been identified by CGH array and multiplex ligation-dependent probe amplification, to cause PHP1A (Fig. 2) (52, 57, 58). Specifically, a deletion of 30 431 bp starting from the intronic region between exons XL and A/B to intron 5 inherited from maternal allele, which as expected, led to an apparent loss of methylation at exon A/B DMR (57), a second deletion comprising exon A/B to exon 1 resulted in a methylation defect limited to the exon A/B DMR (52), and a third 2015-bp deletion consisting of the entire exon 1 and the flanking intronic regions that caused no abnormal methylation at exon A/B DMR (58). The loss of methylation within the exon A/B DMR in patients carrying the first 2 deletions likely reflects the inclusion of the assayed CpG sequences within the deleted region of genomic DNA. In this situation, only the unmethylated CpGs of the paternal exon A/B DMR are present, which gives an impression of an epigenetic loss of methylation when in fact there is actually a physical loss of methylated DNA. One can implicate a similar but reciprocal mechanism for the gain of methylation at upstream DMRs in some patients with PHP1B who have deletions in the NESP, XLαs, and NESPAS exons of GNAS that include DMRs (28, 31-33). By contrast, the third previously published deletion, which comprises only the E1 exon and the surrounding intronic region, did not lead to a methylation change in the exon A/B DMR (58). Hence, it is remarkable that the deletion that we identified in patient 2, which also does not include exon A/B, does in fact lead to a moderate reduction in methylation of the exon A/B DMR. Although PHP1A is not thought to be an imprinting disorder, deletions inherited on the maternal allele that physically disrupt exon A/B, the corresponding DMR, or, as we demonstrate here, a putative control region, can cause an apparent or real loss of methylation at exon A/B that might potentially lead to a misdiagnosis of PHP1B. Nevertheless, we would note that most so-called PHP1A patients with GNAS imprinting defects show a global imprinting alteration at all DMRs rather than an isolated A/B loss of methylation.

Figure 2.

Breakpoint structures of smaller intragenic deletions disrupting the exon A/B or being close to the exon A/B in PHP1A patients. The 3 reported deletions were shown above the schematic representation of GNAS locus and the 2 deletions at the bottom were identified in this study. Genomics coordinates for each of the alternative first exons of GNAS and DMR regions that were assayed: *Johns Hopkins DMR: 20(GRCh37):g.57,464,763-57,464,814; ^#CHOP DMR: 20(GRCh37):g.57,464,773-57,464,927; NESP: 20(GRCh37):g.57,414,795-57,415,941; XLαs: 20(GRCh37):g.57,428,036-57,430,388; A/B: 20(GRCh37):g.57,464,180-57,464,408; E1: 20(GRCh37):g.57,466,426-57,466,920.

Exon A/B is located within a CpG island that extends into exon 1 of GNAS. While exon A/B and the flanking intronic sequences constitute a DMR in which the maternal allele is preferentially methylated, the CpG sequences within exon 1 of GNAS do not exhibit a methylation profile that is based on the parent-of-origin of the allele (8, 9, 59). The deletion in patient 2 now identifies sequences 924 bp (or 915 bp) upstream (centromeric) of the deletion described by Reyes et al. (58), (20(GRCh37):g.57,466,248-57,468,262) that may be part of a region that, together with sequences in STX16, appear to be critical for establishing or maintaining parent-specific methylation at GNAS exon A/B. An alternative possibility is that this deletion might exert an indirect effect related to its close physical position to the critical region. Additional studies will be necessary to determine the exact boundaries of this region in GNAS, but the deletion that we have identified certainly establishes its existence.

Our results also demonstrate the usefulness of WGS in the investigation of patients with PHP1A. WGS can detect relatively small deletions with greater sensitivity than conventional CGH or single nucleotide polymorphism microarray, as well as WES, in which copy number variation calling often needs at least three contiguous exons deleted or duplicated (60). Exon 1 of GNAS, as well exon 1 for many other genes (if not all), is more difficult to analyze in WES, multiplex ligation-dependent probe amplification, and other PCR-based methods due to a very high GC content. Accordingly, WGS or targeted next generation sequencing of chromosome 20q might be useful as a single test to capture nearly all genetic variations and novel intragenic small deletions in GNAS in patients with PPHP/PHP1A as well as progressive osseous heteroplasia (61, 62). Similarly, WGS (63) or targeted next generation sequencing of the GNAS locus (28) may also be useful in identifying small genetic defects that lead to disturbed imprinting of GNAS in patients with PHP1B.

Our work has several strengths, including the use of pyrosequencing to evaluate methylation of GNAS DMRs in a highly quantitative manner (64) and the application of WGS as an unbiased approach to molecular diagnosis. Our study also has several weaknesses, specifically the inability to ascertain all subjects in both families and the lack of protein samples from affected subjects to confirm the effects of the mutations on synthesis of G_sα,

In summary, we have studied 3 patients in 2 families with PHP1A/PPHP in whom we used WGS to identify novel deletions that were not detected by conventional analyses. In the first case, a deletion in the G_sα promoter leads to PHP1A, presumably by decreasing transcription of G_sα. In the second case, a small deletion (1.4 kb) defines a region that may be critical for normal imprinting of the exon A/B DMR. Our work highlights the complexity of the control mechanism for imprinting of GNAS, and suggests a basis for the reduced methylation of the exon A/B DMR in some patients with PHP1A. Future studies will be necessary to define this control region completely, and to identify the putative proteins that bind to this region.

Glossary

Abbreviations

AHO

Albright hereditary osteodystrophy

AS

antisense transcript

CGH

comparative genome hybridization

CGH

comparative genome hybridization

DMR

differentially methylated regions

DNA

deoxyribonucleic acid

indel

insertion/deletion

LCR

low copy repeat

NAHR

nonallelic homologous recombination

NESP55

neuroendocrine secretory protein 55

PHP1A

pseudohypoparathyroidism type 1A

PPHP

pseudopseudohypoparathyroidism

PTH

parathyroid hormone

SV

structural variation

TSH

thyrotropin

WES

whole exome sequence

WGS

whole genome sequencing

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.