J C E M A d v a n c e s

i n

O N L I N E

G e n e t i c s

Loss of Methylation at GNAS Exon A/B Is Associated With Increased Intrauterine Growth Anne-Claire Bréhin, Cindy Colson, Stéphanie Maupetit-Méhouas, Virginie Grybek, Nicolas Richard, Agnès Linglart, Marie-Laure Kottler, and Harald Jüppner Department of Genetics (A.-C.B., C.C., N.R., M.-L.K.), Centre Hospitalier Universitaire de Caen, Reference Centre for Rare Disorders of Calcium and Phosphorus Metabolism, F-14000 Caen, France; Paediatric Endocrinology and Diabetology (S.M.-M., V.G., A.L.), Reference Centre for Rare Disorders of the Mineral Metabolism, AP-HP Hôpital Bicêtre, le Kremlin-Bicêtre 94270, France; Faculté de Médecine, Université Paris Sud, le Kremlin-Bicêtre 94270, France; and Pediatric Nephrology Unit and Endocrine Unit (H.J.), Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114

Context: GNAS is one of few genetic loci that undergo allelic-specific methylation resulting in the parent-specific expression of at least four different transcripts. Due to monoallelic expression, heterozygous GNAS mutations affecting either paternally or maternally derived transcripts cause different forms of pseudohypoparathyroidism (PHP), including autosomal-dominant PHP type Ib (AD-PHP1B) associated with loss of methylation (LOM) at exon A/B alone or sporadic PHP1B (sporPHP1B) associated with broad GNAS methylation changes. Similar to effects other imprinted genes have on early development, we recently observed severe intrauterine growth retardation in newborns, later diagnosed with pseudopseudohypoparathyroidism (PPHP) because of paternal GNAS loss-of-function mutations. Objectives: This study aimed to determine whether GNAS methylation abnormalities affect intrauterine growth. Patients and Methods: Birth parameters were collected of patients who later developed sporPHP1B or AD-PHP1B, and of their healthy siblings. Comparisons were made to newborns affected by PPHP or PHP1A. Results: As newborns, AD-PHP1B patients were bigger than their healthy siblings and well above the reference average; increased sizes were particularly evident if the mothers were unaffected carriers of STX16 deletions. SporPHP1B newborns were slightly above average for weight and length, but their overgrowth was less pronounced than that of AD-PHP1B newborns from unaffected mothers. Conclusion: LOM at GNAS exon A/B due to maternal STX16 deletions and the resulting biallelic A/B expression are associated with enhanced fetal growth. These findings are distinctly different from those of PPHP patients with paternal GNAS exons 2–13 mutations, whose birth parameters are almost 4.5 z-scores below those of AD-PHP1B patients born to healthy mothers. (J Clin Endocrinol Metab 100: E623–E631, 2015)

G

NAS is a complex-imprinted locus on chromosome 20q13.3, which encodes several different transcripts that are either maternally, paternally, and/or bial-

lelically expressed (for reviews see Refs. 1–5). Three protein-coding transcripts share GNAS exons 2 through 13 (E2–E13), but use alternative first exons and promoters,

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2015 by the Endocrine Society Received November 10, 2014. Accepted January 14, 2015. First Published Online January 20, 2015

Abbreviations: AD-PHP1B, autosomal dominant PHP type Ib; AS, noncoding antisense; DMR, differentially methylated region; E, exon; GOM, gain of methylation; IUGR, intrauterine growth retardation; LOM, loss of methylation; patUPD20q, paternal uniparental disomy involving the long arm of chromosome 20; PHP, pseudohypoparathyroidism; PHP1A, PHP type Ia; PPHP, pseudopseudohypoparathyroidism; POH, Progressive osseous heteroplasia; sporPHP1B, sporadic PHP1B.

doi: 10.1210/jc.2014-4047

J Clin Endocrinol Metab, April 2015, 100(4):E623–E631

jcem.endojournals.org

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

E623

E624

Bréhin et al

Intrauterine Growth and GNAS Methylation

Figure 1. Scheme of GNAS locus. Diagram showing the GNAS locus in the chromosomal region 20q13.3 on the paternal and maternal allele. Methylated (silent) and unmethylated (active) exons and promoters are indicated by * and by allele-specific arrows, respectively. The unmethylated GNAS exon 1 promoter allows biallelic Gs␣ expression in most tissues, but it is derived in some tissues predominantly or exclusively from the maternal allele (indicated by solid arrow); for example, in proximal renal tubules, thyroid, and pituitary. STX16 shows biallelic expression.

namely NESP, XL, and E1, that produce NESP55 (chromogranin-like secretory protein), XL␣s (extra large form of Gs␣), and Gs␣, respectively (Figure 1). The GNAS cluster furthermore gives rise to at least two additional transcripts, namely the noncoding antisense (AS) transcript derived from the alternative first exon located centromeric of exon XL and the A/B transcript derived from the alternative first telomeric of exon XL. Exon A/B is spliced onto GNAS exon E2, which comprises an initiator ATG, and thus can give rise to an N-terminally truncated form of Gs␣ (6). With the exception of exon E1, the promoters for all alternative first GNAS exons are located within differentially methylated regions (DMRs) and transcripts are derived exclusively from the nonmethylated parental allele. Exons XL, A/B, and AS are methylated on the maternal allele and are thus transcribed only from the paternal allele. Conversely, exon NESP is methylated on the paternal allele and its transcript is therefore derived only from the maternal GNAS allele. The exon E1 promoter, which gives rise to the Gs␣ transcript is not methylated; accordingly, Gs␣ expression is biallelic in most tissues. However, it shows predominantly maternal expression in the renal proximal tubules, whereas variable degrees of paternal Gs␣ expression have been described in thyroid, pituitary, brown fat, and gonads (5, 7–9). The mechanism(s) through which paternal Gs␣ expression is silenced in a tissue-specific and time-dependent manner is unknown, but may require active transcription from the exon A/B promoter, since loss of methylation (LOM) at the A/B exon and consequently biallelic transcription from this promoter occurs in different PHP1B variants (10, 11). GNAS mutations are responsible for several phenotypically different, inherited disorders (4 –5). Maternally inherited, loss-of-function mutations affecting those GNAS exons that encode Gs␣ cause pseudohypoparathyroidism (PHP) type Ia (PHP1A) (Online Mendelian Inheritance in Man [OMIM] 103580), a disorder characterized by multiple hormonal resistance in combination with phenotypic features of Albright’s Hereditary Osteodystrophy. Meth-

J Clin Endocrinol Metab, April 2015, 100(4):E623–E631

ylation abnormalities within the GNAS cluster on the maternal allele are responsible for another PHP variant, namely PHP type Ib (PHP1B) (Online Mendelian Inheritance in Man [OMIM] 603233). Patients affected by this disorder show PTH-resistant hypocalcemia and hyperphosphatemia, just like patients affected by PHP1A, but obvious Albright’s Hereditary Osteodystrophy features are present in only few PHP1B cases (12–15). LOM at the exon A/B DMR, which leads to biallelic expression of the A/B transcript, thereby reducing Gs␣ expression, is invariably observed in PHP1B patients (10, 11). In the proximal renal tubules, Gs␣ is expressed predominantly from the maternal allele. Consequently, GNAS mutations affecting exons E1-E13 on this parental allele lead in that tissue to a loss or a severe reduction in maternal Gs␣ activity. The resulting PTH-resistance would be sufficient to explain the reduction in urinary phosphate excretion. Recent findings in mice lacking Gnas exon 1 on the maternal allele have furthermore suggested a mild reduction in 1␣-hydroxylase mRNA (9) that would help explain the reduction in 1,25(OH)2 vitamin D levels, previously noted in PHP1A patients (16, 17). Similarly reduced levels of this biologically active vitamin D metabolite could be encountered in PHP1B, thus leading to impaired intestinal calcium absorption and hypocalcemia. Silencing of paternal Gs␣ expression may occur also, albeit incompletely, in other tissues such as thyroid and pituitary. A considerable number of PHP1B and PHP1A patients therefore display detectable abnormalities in other endocrine systems (1–5). PHP1B can follow an autosomal-dominant trait (ADPHP1B) or it occurs as a sporadic disease (sporPHP1B). Most patients with AD-PHP1B exhibit maternally inherited deletions comprising the gene encoding syntaxin-16 (STX16) or the GNAS exon NESP, which result in an isolated LOM at the exon A/B DMR (11, 18 –21). In contrast, deletions comprising GNAS exon NESP and/or antisense exons 3 and 4 lead to AD-PHP1B through a loss of all maternal methylation imprints (22, 23). Similarly broad GNAS methylation changes are also observed in most patients affected by sporPHP1B, who show LOM at the DMRs for exons A/B, XL, and AS, and a gain of methylation (GOM) at the DMR of exon NESP, but no evidence for paternal uniparental disomy involving the long arm of chromosome 20 (patUPD20q) (24). In mice, the paternally expressed XL␣s and the maternally expressed Gs␣ seem to act antagonistically, given that both transcripts have opposite effects on body weight. Indeed, loss-of-function of paternally expressed XL␣s gives rise to mice that are growth restricted, lean, and hypermetabolic, whereas loss of the maternally expressed Gs␣ (⌬E1m mice) gives rise to mice that present with

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4047

jcem.endojournals.org

edema and no obesity at birth, but are postnatally obese and hypometabolic (25–29). Taken together, these findings suggest that several GNAS-derived transcripts can affect intrauterine development. Consistent with these findings in genetically manipulated mice, some patients with PHP1B due to patUPD20q are bigger at birth and subsequently show macrosomia, macrocephaly, and tall stature (24); their overgrowth may be related to biallelic expression of AS, XL␣s, and A/B, or to the lack of NESP55. The opposite, namely severe intrauterine growth retardation (IUGR) was recently documented for patients affected by pseudopseudohypoparathyroidism (PPHP) (or progressive osseous heteroplasia [POH], a variant of PPHP), particularly when this disorder is caused by loss-of-function mutations involving the paternal GNAS exons E2–E13, ie, those exons that are shared by Gs␣ and XL␣s. In fact, lack of the paternally derived XL␣s seems to be the most plausible cause of abnormal prenatal growth, given that PPHP patients with paternal mutations involving GNAS exon 1, which is specific for Gs␣, showed a milder degree of growth retardation (32). In addition, severe pre- and postnatal growth retardation is also observed in patients with maternal uniparental disomy of chromosome 20 (matUPD20), who are predicted to have biallelic NESP55 expression, yet lack expression of AS, XL, and A/B, and are likely to show increased Gs␣ levels (33, 34). Currently available data thus suggest that normal intrauterine growth and/or placental development requires a unique balance between different paternally and maternally expressed GNAS transcripts. To explore the mechanisms associated with intrauterine development further, we have now studied the birth records of a large number of PHP1B patients presenting with different GNAS epimutations. Particularly, AD-PHP1B patients showed a significant overgrowth phenotype at birth, which was most pronounced when the mutation was inherited from a healthy carrier mother. Table 1.

E625

Patients and Methods Investigated patients For this retrospective study, we enrolled a total of 164 individuals living in either Europe or North America (Table 1). Clinical, laboratory, genetic, and/or epigenetic phenotypes were consistent with PHP1B, PHP1A, or PPHP/POH, respectively; healthy siblings of PHP1B patients had no obvious abnormality in the regulation of mineral ion homeostasis and analysis of their genomic DNA showed a normal GNAS methylation pattern. Most patients were Caucasians of Western European origin; 4 patients were CaucasianHispanic, 1 patient was Asian, 1 was Asian-Caucasian, and for 11 patients the racial background was unknown. Birth parameters of new and previously investigated individuals are provided (Total: 114 patients; 60 males, 54 females) (Table 1 and Supplemental Table 1, A and B). Sixty-one ADPHP1B patients are carriers of the 3-kb or the 4.4-kb STX16 deletion, which is associated with LOM restricted to GNAS exon A/B; 28 of these individuals are from North America and 33 are from Europe. Twenty-five AD-PHP1B patients (16 males and 9 females) were born to affected mothers (STX16 deletion on the maternal allele), whereas 31 (13 males and 18 females) were born to mothers who are healthy carriers of the genetic defect (STX16 deletion on the paternal allele) and 5 (2 males and 3 females) were born to mothers with an unknown status. Twenty-eight individuals were healthy siblings of AD-PHP1B patients (15 males and 13 females) (Table 1 and Supplemental Table 2); 10 had a mother affected by PHP1B due to a maternally inherited STX16 deletion, whereas 18 had an unaffected mother carrying a paternally inherited STX16 deletion. Fifty-three sporPHP1B patients (8 from North America and 45 from Europe; 29 males and 24 females) showed LOM at exons A/B, XL, and AS, and a gain of methylation at exon NESP. Ten individuals were healthy siblings of sporPHP1B patients (3 males and 7 females) (Table 1 and Supplemental Table 2). Twelve individuals (8 males, 4 females) affected by PHP1A due to exon1/intron 1 mutations were born to mothers affected by PPHP/POH; 4 of these patients were previously reported (32); data for more recently investigated probands are provided (Table 1 and Supplemental Table 3). Data from 10 previously reported PPHP/POH patients with exon1/intron 1 mutations (3 males, 7 females) or with mutations in GNAS exons E2–E13 were used for comparison (Table 1 and Supplemental Table 4) (32).

Number and Sex of Investigated Individuals and Their Genetic and Epigenetic Findings n

Total PHP1B patients AD-PHP1B Total Affected mother Carrier mother Maternal status unknown Healthy siblings Total Affected mother Carrier Mother

STX16 Deletion

LOM

GOM

114 61

Males

Females

60

54

25 31 5

A/B A/B no ?

no no no no

31 16 13 2

30 9 18 3

10 18

A/B no

no no

15 6 9

13 4 9

28

SporPHP1B Healthy siblings

53 10

no no

AS, XL, A/B no

NESP no

29 3

24 7

PHP1A (exon/intron 1 mutation) PPHP (exon/intron 1 mutation)

12 10

no no

no no

no no

8 3

4 7

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

E626

Bréhin et al

Intrauterine Growth and GNAS Methylation

For each individual, written informed consent was obtained from the patients and/or their parents for collection of clinical and laboratory data, and for DNA collection to conduct molecular studies. Outcome of pregnancy and biometric data at birth were collected retrospectively using birth certificates, or records from hospitals or primary care physicians. Preterm birth was defined as delivery before 37 weeks of gestation and preterm babies less than 36 weeks were excluded. Gestational age, weight, length, and head circumference were compared with sexspecific French reference charts, which allowed calculation of z-scores by AUDIPOG, as described (32). Data are presented as mean ⫾ SD.

Molecular studies Genomic DNA was isolated from peripheral blood leukocytes using standard methods. GNAS methylation studies were performed by qPCR assays after digestion of genomic DNA with methylation-sensitive and -insensitive restriction enzymes (35) and by methylation-specific multiple ligation-dependent probe amplification, as described (32). Biochemical and endocrine studies, as well as DNA analyses were performed for parents to determine whether mother or father are affected, and to confirm that AD-PHP1B patients had a maternally inherited STX16 deletion.

Statistical analysis The d’Agostino and Pearson omnibus normality test was used for determining whether birth parameters showed a normal distribution. Differences between groups were calculated using the Mann-Whitney nonparametric U test. A two-tailed P ⬍ .01 was considered statistically significant (*, P ⱕ .01; ** P ⱕ .001; ***, P ⱕ .0001).

Results Individuals, who later developed PTH resistance because of PHP1B, are bigger at birth Patients affected by sporPHP1B or AD-PHP1B due to a STX16 deletion had at birth significantly increased weights, lengths, and head circumferences. These measurements followed for each cohort a Gaussian distribution equivalent to that of the reference population with K2

J Clin Endocrinol Metab, April 2015, 100(4):E623–E631

values of 0.23 (P ⫽ .89) for weight, 0.54 (P ⫽ .76) for length, and 0.07 (P ⫽ .97) for head circumference (Supplemental Figure 1, A–C). When analyzed as a single combined cohort, sporPHP1B and AD-PHP1B patients revealed z-scores of ⫹0.73 ⫾ 1.2 for weight (n ⫽ 114; P ⬍ .0001), ⫹0.92 ⫾ 1.3 for length (n ⫽ 74; P ⬍ .0001), and ⫹0.68 ⫾ 1.1 for head circumference (n ⫽ 40; P ⫽ .0004). To determine whether PHP1B patients from North America and Europe could be studied as a single group, we next compared the birth parameters of both groups of patients. There were no significant differences for weight or head circumference; the average z-scores were ⫹0.83 ⫾ 1.0 (n ⫽ 36) vs ⫹0.68 ⫾ 1.2 (n ⫽ 78) for weight (P ⫽ .465), and ⫹1.00 ⫾ 1.4 (n ⫽ 6) vs ⫹0.63 ⫾ 1.1 (n ⫽ 34) for head circumference (P ⫽ .636). However, North American PHP1B patients had in comparison with Europeans an increased length at birth; the average z-scores were ⫹1.56 ⫾ 1.4 (n ⫽ 25) vs ⫹0.60 ⫾ 1.2 (n ⫽ 49) (P ⫽ .007) (Supplemental Figure 2). No sex-dependent z-score differences were observed. The average z-scores were ⫹0.79 ⫾ 1.0 (60 males) vs ⫹0.66 ⫾ 1.3 (54 females) for weight (P ⫽ .43), and ⫹1.14 ⫾ 1.2 (38 males) vs ⫹0.69 ⫾ 1.4 (36 females) for length (P ⫽ .11), and ⫹0.76 ⫾ 1.0 (25 males) vs ⫹0.56 ⫾ 1.3 (15 females) for head circumference (P ⫽ .47) (Supplemental Figure 3, A–C). Cohorts of AD-PHP1B and sporPHP1B patients show indistinguishable birth parameters We next wanted to determine whether LOM of GNAS exon A/B alone (AD-PHP1B) or broad epigenetic GNAS changes such as LOM of AS and XL␣s, and GOM of NESP55 (sporPHP1B) contribute to fetal growth and development; both groups of PHP1B patients were therefore studied separately. SporPHP1B patients tended to have slightly lower birth weights and lengths than AD-PHP1B patients, but the difference was not significant when comparing both cohorts (Figure 2, A–C).

Figure 2. Birth parameters of patients with different forms of PHP1B. Comparison between patients with sporadic PHP1B (white circles) and patients with AD-PHP1B (black circles) for weight (A), length (B), and head circumference (C) at birth. Data for each subject are expressed as zscore. N, number of subjects per group. The horizontal and vertical bars for each group represent the mean ⫾ SD; statistically nonsignificant. No significant difference in birth parameters between patients with sporadic or familial forms of PHP1B.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4047

jcem.endojournals.org

E627

healthy carriers, ie, females with a STX16 deletion on the maternal or the paternal allele, respectively. AD-PHP1B patients were significantly heavier at birth when their mothers were healthy carriers of a STX16 deletion, as compared to newborns with the same mutation born to mothers who were themselves affected by AD-PHP1B, ie, carried the STX16 deletion on their maternal allele (z-scores, ⫹1.44 ⫾ 1.0 vs ⫹0.36 ⫾ 1.0; P ⫽ .0004) (FigFigure 3. Birth weight and length of AD-PHP1B patients and their healthy siblings. Comparison ure 3A). Birth lengths were, in of weight (A) and length (B) at birth of PHP1B patients, who were born to mothers carrying the STX16 deletion on either their paternal (unaffected carriers, black circles) or maternal allele comparison with the reference pop(affected, white circles). Also shown are the birth parameter of healthy siblings born to ulation, significantly higher for all unaffected (black diamonds) or affected mothers (white diamonds). Data for each subject are AD-PHP1B patients, particularly if expressed as z-score. N, number of subjects per group. The horizontal and vertical bars for each group represent the mean ⫾ SD. P values for comparison between groups and for comparison the mothers were unaffected carriers with reference population are indicated. (Figure 3B). z-scores of healthy newborns when born to healthy carriers AD-PHP1B patients born to healthy carriers of a (n ⫽ 18) or to affected mothers (n ⫽ 10) were not signifSTX16 deletion show the most obvious increase in icantly different from the reference population (P ⫽ .077 intrauterine growth and P ⫽ .202, respectively). However, healthy babies born To determine whether the carrier status of the mothers, to affected mothers were slightly smaller than healthy baie, STX16 deletions on the maternal or the paternal allele, bies born to unaffected carrier mothers (z-score, ⫹0.50 ⫾ affects fetal growth and development, we collected the 1.1 vs ⫺0.50 ⫾ 1.2; P ⫽ .044.). There was no difference birth parameters of AD-PHP1B patients and their healthy in the average length for healthy babies born to healthy or siblings, who were born by either affected mothers or by affected mothers (z-score, ⫹0.63 ⫾ 1.2 vs ⫹0.96 ⫾ 2.0; P ⫽ .088) and these data were not significantly different from the general population (P ⫽ .11 and P ⫽ .287, respectively). Patients with AD-PHP1B from unaffected mothers showed a more severe overgrowth phenotype at birth than patients with sporPHP1B (Figure 4, A and B), but no significant difference was observed when comparing head circumferences of both PHP1B cohorts (data not shown). Several of the investigated AD-PHP1B patients were from the Figure 4. Weight (A) and length (B) of patients affected by AD-PHP1B and sporadic PHP1B in same family, which could have concomparison with patients affected by PPHP/POH and PHP1A. Birth parameters of patients founded the birth parameters due to affected by either PPHP/POH (white squares) or PHP1A (black squares); these individuals have exon/intron 1 mutations on the paternal or the maternal GNAS allele, respectively; note that the the shared genetic background. Remothers of these PHP1A patients are affected by PPHP/POH and thus have no PTH resistance, or analysis of data was therefore rethe mutation of the newborn had occurred de novo on the maternal allele. AD-PHP1B patients born to mothers, who are healthy carriers of a STX16 deletion (black circles) and sporadic PHP1B stricted to birth parameters of the inpatients (white circles). Data for each subject are expressed as z-score on the x-axis. P-values are dex cases in each of the different indicated for comparison to the reference population and for comparisons between different AD-PHP1B kindreds; measurements groups. N, number of subjects per group. The horizontal and vertical bars for each group represent the mean ⫾ SD; statistically significant differences were observed between PHP1A and were compared with the sporPHP1B sporadic PHP1B, and between sporadic PHP1B and AD-PHP1B patients from nonaffected carrier patients (see Figure 2, A and B). For mothers.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

E628

Bréhin et al

Intrauterine Growth and GNAS Methylation

this subgroup of AD-PHP1B patients, the z-scores for birth weight and length were ⫹0.87 ⫾ 1.1 (n ⫽ 35) and ⫹1.12 ⫾ 1.3 (n ⫽ 22), respectively, which was indistinguishable from the entire AD-PHP1B cohort and remained highly significant in comparison with the reference population (P ⬍ .0001 and P ⫽ .0006, respectively). No apparent differences were observed when comparing the birth parameters of sporPHP1B patients (n ⫽ 53) to those of their healthy siblings (n ⫽ 10) (Supplemental Figure 4, A and B).

J Clin Endocrinol Metab, April 2015, 100(4):E623–E631

ences when comparing both groups (P ⫽ .021 for weight and P ⫽ .028 for length). However, in comparison with the reference population, z-scores for weight and length were ⫺1.92 ⫾ 1.2 (n ⫽ 10; P ⫽ .0005) and ⫺2.20 ⫾ 0.8 (n ⫽ 9; P ⬍ .0001), respectively, for PPHP/POH patients, and ⫺0.68 ⫾ 1.1 (n ⫽ 12; P ⫽ .045) and ⫺1.20 ⫾ 1.2 (n ⫽ 11; P ⫽ .009), respectively, for PHP1A patients (Figure 4, A and B).

Discussion

Weight and length of newborns affected by either PPHP/POH or PHP1A with exon/intron 1 mutations in comparison with PHP1B patients We next compared the birth parameters of PHP1A and PPHP/POH patients with GNAS mutations on the maternal or paternal allele, respectively, involving only exon/ intron 1, ie, mutations that affect only Gs␣ function; some of these data were previously reported (Supplemental Tables 3 and 4). To allow appropriate comparisons between PPHP/POH and PHP1A patients, we included in the latter group only patients born to mothers without hormonal resistance, ie, PHP1A patients, who had either de novo mutations on the maternal allele or were born to mothers affected by PPHP/POH. There were no significant differ-

We have recently shown that patients with loss-of-function mutation in GNAS on the paternal allele have birth weights and lengths that are more than two SDs below the 50th percentile, when analyzing this cohort as a whole (z-score for weight, ⫺2.69 ⫾ 0.24) (32). Importantly, paternal mutations involving GNAS exons E2–E13, ie, those exons that are shared between Gs␣ and XL␣s, resulted in the most severe growth retardation (z-score for weight: ⫺3.01 ⫾ 0.27), raising the possibility that XL␣s has a particularly important role in normal fetal development (Figure 5). In the present study, we have now shown that patients with epimutations at the GNAS locus, which reduce Gs␣ expression from the maternal allele, display enhanced fetal growth resulting on average in significantly higher weights and increased lengths at birth. These changes were observed in patients who later developed AD-PHP1B and, to a lesser extent, in patients later affected by sporPHP1B. The overgrowth phenotype was indistinguishably observed in both sexes and it occurred equally in PHP1B patients from North America and Europe, thus allowing comparisons between relatively large cohorts of these rare diseases. Our findings were particularly striking when comparing AD-PHP1B patients born to healthy carrier mothers to those of PPHP/POH patients with GNAS mutations in exons E2– E13 (32), because these two groups revealed differences in birth weights of Figure 5. Schematic presentation of parent-specific expression of Gs␣, A/B, and XL␣s and the effect on growth in healthy controls, and in patients with PPHP caused by mutations in GNAS almost 4.5 z-scores. exons E2–E13, PHP1A, sporPHP1B due to epigenetic changes without a defined genetic The health status of the mother, mutation, or PHP1B due to a maternal STX16 deletion; note that only patients of mothers with ie, normal regulation of mineral ion normal mineral ion homeostasis are depicted. Blue ⫹ or ⫺, paternal expression; orange ⫹, maternal expression; nl, normal; solid blue cross, inactivating mutations in paternal GNAS exons homeostasis (mother with STX16 E2–E13; solid red cross, inactivating mutations in GNAS exons 1–13; stippled red cross, lack of deletion on the paternal allele), or maternal Gs␣ expression due to LOM at GNAS exon A/B; arrows depict decreased or increased hormonal resistance (mother with intrauterine growth.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4047

STX16 deletion on the maternal allele) seemed to contribute not only to intrauterine growth of AD-PHP1B patients, but also of their healthy siblings. This implies that growth is attenuated not only for AD-PHP1B fetuses, but also for healthy offspring, of mothers affected by ADPHP1B. Whether this negative effect is related to inadequate medical treatment and thus persistent maternal abnormalities in the PTH-dependent regulation of mineral ion homeostasis, or to abnormalities in other Gs␣-dependent systems of the mother and/or the placenta, remains to be determined. In contrast with these findings in AD-PHP1B, newborns affected by PHP1A did not seem to be influenced by the status of the mother, because the latter babies displayed birth parameters that were slightly below average, irrespective whether they were born to mothers affected by PPHP or PHP1A [Figure 4 and Richard et al (32)]. This makes it plausible that a factor(s) other than fetal or maternal Gs␣ expression is an important contributor to intrauterine growth. Newborns, who later developed AD-PHP1B due to a STX16 deletion inherited from a healthy carrier mother, showed the most obvious overgrowth. These mutations cause a loss of GNAS exon A/B methylation on the maternal allele leading to biallelic expression of A/B transcripts and little or no Gs␣ protein in those tissues in which paternal Gs␣ expression is silenced. However, newborns affected by PHP1A, ie, carriers of inactivating mutations involving the maternal GNAS exons E1–E13, showed no overgrowth. This raises the possibility that biallelic A/B expression, as observed in AD-PHP1B patients with STX16 deletions, contributes importantly to enhanced intrauterine growth through as-of-yet unknown regulatory functions. Our cases with sporPHP1B, ie, patients with biallelic expression of A/B as well as XL␣s, and AS, yet no expression of NESP55, showed birth parameters that were only slightly above the mean of the reference population. Although weight and length were available only for a small number of healthy siblings, their measurements were indistinguishable from those of sporPHP1B patients. Although it will be important to obtain additional data, it seems unlikely that the birth parameters of unaffected siblings of sporPHP1B patients is influenced by their healthy mothers. With the exception of patUPD20q, little is known about molecular causes of the different epigenetic changes at the GNAS locus. The differences in birth parameters observed between newborns, who later developed ADPHP1B or sporPHP1B suggests a complex interplay between the different GNAS-derived transcripts. In fact, the paternal XL␣s and A/B transcripts seem to enhance

jcem.endojournals.org

E629

growth, whereas others such maternal NESP55 (or perhaps paternal AS and ALEX) may restrain growth, as suggested by findings in mice with a maternal truncating mutation in exon 1A (⌬E1A-Tm) that seems to impair Nesp55 expression leading to a small overgrowth phenotype; note that murine exon 1A is equivalent to human exon A/B (30, 31). Imprinted genes, such as GNAS, account for only 0.1– 0.5% of the genome but they have a disproportionately important influence on early mammalian development (36 –38). Their dysregulated expression has been implicated in several diseases, including sporadic, inherited, and environmentally induced growth disorders, such as Beckwith-Wiedeman and Angelman syndrome (39 – 41). Conflict between maternal and paternal genomes in allocating maternal resources to the developing embryo is believed to provide one of several possible explanations for the emergence of imprinting mechanisms during evolution. According to this hypothesis, genes expressed from the paternal allele tend to stimulate intrauterine growth while genes expressed from the maternal allele have the opposite effect. Consistent with this parental conflict theory, PPHP/POH patients, particularly individuals with paternal GNAS mutations involving exons E2-E13, show severe IUGR, whereas patients with AD-PHP1B due to maternal STX16 deletions and the associated LOM at GNAS exon A/B alone show enhanced fetal growth. In conclusion, newborns who later developed ADPHP1B due to STX16 deletions inherited from a healthy carrier mother displayed a significant increase in intrauterine growth, which is vastly different from the IUGR observed in newborns who later developed PPHP/POH due to paternal mutations affecting GNAS exons E2-E13. It is conceivable that LOM at GNAS exon A/B and the resulting biallelic expression of A/B transcripts contributes to enhanced intrauterine growth, but it remains to be determined whether this overgrowth phenotype persists postnatally.

Acknowledgments We thank all patients and their parents for their willingness to participate in our studies, and the staff of all units involved in collecting data. We are furthermore grateful to Ms T. Marie for her excellent contribution to the recording of data and to Dr Stephen Alexander for helpful advice. Address all correspondence to: Dr Harald Jueppner, MD, Massachusetts General Hospital, Endocrine Unit, 50 Blossom Street, Boston MA 02114. E-mail: [email protected]. The physicians who contributed clinical and laboratory information and DNA samples: Grant Ayling, Family and General Practice Westcoast Family Practice Centre, Vancouver, Canada;

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

E630

Bréhin et al

Intrauterine Growth and GNAS Methylation

Justine Bacchetta, Pediatric Endocrinology, CHU de Lyon-GH Est, Hospices Civils de Lyon, Lyon, France; Sabine Barin, Department of Endocrinology, CHU de Nantes Hôpital Mère-Enfant, Nantes, France; Helene Bony-Trifunovic, Department of Pediatrics, Centre Hospitalier Universitaire Amiens, Amiens, France; Catherine Cardot-Bauters, Endocrinology, Hôpital Claude Huriez, Lille, France; Stuart Carr, University of Alberta, Edmonton, Canada; Clémence Carre, Department of Genetics, Hôpital J. de Flandre CHRU, Lille, France; Christine Cessans, Pédiatrie 13 Rue des Mottes, Nieul Sur Mer, France; Anne Coeslier Dieux, Department of Genetics, Hôpital J. de Flandre, Lille, France; Patrick Collignon, Department of Genetics, C.H.I.T.S. Hôpital Sainte Musse, Toulon, France; Fabienne Dalla-Vale, Pediatrics Hôpital Arnaud de Villeneuse, Montpellier, France; Terry Joe Declue, Pediatric Endocrinology, St. Joseph Hospital, Tampa, FL; Rachel Dessailloud, Endocrinology Diabetologia and Nutrition, CHU Avenue René Laennec Salloüel, Amiens, France; Deborah V. Edidin, Northwestern University Feinberg School of Medicine, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago IL; Melinda Fawbush, Baptist Medical Center, Jacksonville, FL; Christine Francannet, Department of Genetics, CHU Estaing, Clermont-Ferrand, France; Mitchell Geffner, Children’s Hospital Los Angeles, The Saban Research Institute, Los Angeles, CA; Bruno Gombert, Department of Genetics, Groupe Hospitalier La Rochelle, La Rochelle, France; Parlier Guilhem, Department of Pediatrics, CH St Nazaire BP, St Nazaire, France; Muriel Holder, Department of Genetics, Hôpital J. de Flandre, Lille, France; Jessica Jaillet, Hôpital de Tarare, Tarare, France; Fanny Laffargue, Department of Genetics, CHU Estaing, Clermont-Ferrand, France; Béatrice Lebon-Labich, Department of Pediatrics, Hôpital d’enfants, Vandoeuvre les Nancy, France; Juliane Leger, Pediatric Endocrinology, Hôpital Robert Debré, Paris, France; Peter Lyons, Andrews University, Berrien Springs, MI; Christoph Mache, Department of Pediatrics, Division of Pediatric Nephrology, Medical University Graz, Graz, Austria; Christine Martel, Pédiatrics, Centre Hospitalier des Pays de Morlaix, Morlaix, France; Jukka Moilanen, Department of Clinical Genetics, Oulu University Hospital, University of Oulu, Oulu, Finland; Myriam Oliel, Division of Pediatric Endocrinology, Centre Hospitalier Lyon Sud, Lyon, France; Graziella Pinto, Pediatric Endocrinology, Hôpital Necker Enfants Malades, Paris, France; Michel Polak, Pediatric Endocrinology, Hôpital Necker Enfants Malades, Paris, France; Annie Pouvreau, Department of Pediatrics, Centre Hospitalier, Saintes, France; Michel Pugeat, Endocrinology, CHU de Lyon-GH Est, Hospices Civils de Lyon, Lyon, France; Virginie Ribault, Pédiatrics, CHU Caen, Caen, France; William Russell, Division of Pediatric Endocrinology and Diabetes, Vanderbilt University School of Medicine, Nashville, TN; Gary Saito, Arizona Kidney Disease and Hypertension Centers, Tucson Access Center Tucson, AZ; Jean-Pierre Salles, Pediatric Endocrinology, Génétique et Gynécologie Médicale, Pôle Enfants, Hôpital des Enfants, Toulouse, France; Lars Savendahl, Pediatric Endocrinology Unit, Department of Women and Children’s Health, Karolinska Institutet, Stockholm, Sweden; Amita Sharma, Massachusetts General Hospital, Clinical Director, Pediatric Nephrology, Boston, MA; Dolores Shoback, University of California San Francisco School of Medicine, San Francisco, CA; Caroline Thalassinos, Pediatric Endocrinology, Hôpital Necker Enfants Malades, Paris, France; Agathoklis Tsatsoulis, Department of Medicine, Division of Endocrinology, University of Ioannina, Ioannina, Greece; Serap Turan, Pediatric Endocrinology, Mar-

J Clin Endocrinol Metab, April 2015, 100(4):E623–E631

mara University School of Medicine Hospital, Istanbul, Turkey; Sheila Unger, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland. This work was supported by the National Institutes of Health (RO1 DK46718-20 to H.J.), by the Société Française D’endocrinologie (fellowship grant to S.M.-M.), and by the Société Française d’Endocrinologie et Diabétologie Pédiatrique (fellowship grant to V.G.). Disclosure Summary: The authors have nothing to disclose.

References 1. Levine M.2005 Hypoparathyroidism and pseudohypoparathyroidism. In: DeGroot L, Jameson J eds. Endocrinology. Philadelphia, PA: W.B. Saunders Company; 1611–1636. 2. Weinstein LS, Xie T, Zhang QH, Chen M. Studies of the regulation and function of the Gs alpha gene Gnas using gene targeting technology. Pharmacol Ther. 2007;115:271–291. 3. Plagge A, Kelsey G, Germain-Lee EL. Physiological functions of the imprinted Gnas locus and its protein variants Galpha(s) and XLalpha(s) in human and mouse. J Endocrinol. 2008;196:193–214. 4. Bastepe M, Jüppner H. 2010 Pseudohypoparathyroidism, Albright’s hereditary osteodystrophy, and progressive osseous heteroplasia: disorders caused by inactivating GNAS mutations. In: DeGroot LJ, Jameson JL, eds. Endocrinology. 6th ed. Philadelphia, PA: W.B. Saunders Company; 1223–1235. 5. Mantovani G. Clinical review: Pseudohypoparathyroidism: diagnosis and treatment. J Clin Endocrinol Metab. 2011;96:3020 –3030. 6. Puzhko S, Goodyer CG, Kerachian M, et al. Parathyroid hormone signaling via G␣s is selectively inhibited by an NH2-terminally truncated G␣s: Implications for pseudohypoparathyroidism. J Bone Miner Res. 2011;26:2473–2485. 7. Hayward BE, Barlier A, Korbonits M, et al. Imprinting of the G(s)alpha gene GNAS1 in the pathogenesis of acromegaly. J Clin Invest. 2001;107:R31–R36. 8. Liu J, Erlichman B, Weinstein L. The stimulatory G protein alphasubunit Gs alpha is imprinted in human thyroid glands: implications for thyroid function in pseudohypoparathyroidism types 1A and 1B. J Clin Endocrinol Metab. 2003;88:4336 – 4341. 9. Turan S, Fernandez-Rebollo E, Aydin C, et al. Postnatal establishment of allelic G␣s silencing as a plausible explanation for delayed onset of parathyroid hormone resistance owing to heterozygous G␣s disruption. J Bone Miner Res. 2014;29:749 –760. 10. Liu J, Litman D, Rosenberg MJ, Yu S, Biesecker LG, Weinstein L. A GNAS1 imprinting defect in pseudohypoparathyroidism type IB. J Clin Invest. 2000;106:1167–1174. 11. Turan S, Ignatius J, Moilanen J, et al. De novo STX16 deletions: An infrequent cause of pseudohypoparathyroidism type Ib that should be excluded in sporadic cases. J Clin Endocrinol Metab. 2012; 97(12):E2314 –E2319. 12. de Nanclares GP, Fernández-Rebollo E, Santin I, et al. Epigenetic defects of GNAS in patients with pseudohypoparathyroidism and mild features of Albright’s hereditary osteodystrophy. J Clin Endocrinol Metab. 2007;92:2370 –2373. 13. Mariot V, Maupetit-Mehouas S, Sinding C, Kottler ML, Linglart A. A maternal epimutation of GNAS leads to Albright osteodystrophy and parathyroid hormone resistance. J Clin Endocrinol Metab. 2008;93:661– 665. 14. Unluturk U, Harmanci A, Babaoglu M, Yasar U, Varli K, Bastepe M, Bayraktar M. Molecular diagnosis and clinical characterization of pseudohypoparathyroidism type-Ib in a patient with mild Albright’s hereditary osteodystrophy-like features, epileptic seizures, and defective renal handling of uric acid. Am J Med Sci. 2008;336: 84 –90. 15. Mantovani G, de Sanctis L, Barbieri AM, et al. Pseudohypopara-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4047

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

thyroidism and GNAS epigenetic defects: clinical evaluation of Albright hereditary osteodystrophy and molecular analysis in 40 patients. J Clin Endocrinol Metab. 2010;95:651– 658. Drezner MK, Neelon FA, Haussler M, McPherson HT, Lebovitz HE. 1,25-Dihydroxycholecalciferol deficiency: The probable cause of hypocalcemia and metabolic bone disease in pseudohypoparathyroidism. J Clin Endocrinol Metab. 1976;42:621– 628. Lambert PW, Hollis BW, Bell NH, Epstein S. Demonstration of a lack of change in serum 1 alpha,25-dihydroxyvitamin D in response to parathyroid extract in pseudohypoparathyroidism. J Clin Invest. 1980;66:782–791. Elli FM, de Sanctis L, Peverelli E, et al. Autosomal dominant pseudohypoparathyroidism type Ib: A novel inherited deletion ablating STX16 causes loss of Imprinting at the A/B DMR. J Clin Endocrinol Metab. 2014:jc20133704. Linglart A, Gensure RC, Olney RC, Jüppner H, Bastepe M. A novel STX16 deletion in autosomal dominant pseudohypoparathyroidism type Ib redefines the boundaries of a cis-acting imprinting control element of GNAS. Am J Hum Genet. 2005;76:804 – 814. Liu J, Nealon J, Weinstein L. Distinct patterns of abnormal GNAS imprinting in familial and sporadic pseudohypoparathyroidism type IB. Hum Mol Genet. 2005;14:95–102. Richard N, Abeguile G, Coudray N, et al. A new deletion ablating NESP55 causes loss of maternal imprint of A/B GNAS and autosomal dominant pseudohypoparathyroidism type Ib. J Clin Endocrinol Metab. 2012;97:E863–E867. Bastepe M, Fröhlich LF, Linglart A, et al. Deletion of the NESP55 differentially methylated region causes loss of maternal GNAS imprints and pseudohypoparathyroidism type Ib. Nat Genet. 2005; 37:25–27. Chillambhi S, Turan S, Hwang DY, Chen HC, Jüppner H, Bastepe M. Deletion of the noncoding GNAS antisense transcript causes pseudohypoparathyroidism type Ib and biparental defects of GNAS methylation in cis. J Clin Endocrinol Metab. 2010;95:3993– 4002. Dixit A, Chandler KE, Lever M, et al. Pseudohypoparathyroidism type 1b due to paternal uniparental disomy of chromosome 20q. J Clin Endocrinol Metab. 2013;98:E103–E108. Plagge A, Gordon E, Dean W, et al. The imprinted signaling protein XL alpha s is required for postnatal adaptation to feeding. Nat Genet. 2004;36:818 – 826. Germain-Lee EL, Schwindinger W, Crane JL, et al. A mouse model of Albright hereditary osteodystrophy generated by targeted disruption of exon 1 of the Gnas gene. Endocrinology. 2005;146: 4697– 4709. Chen M, Gavrilova O, Liu J, et al. Alternative Gnas gene products

jcem.endojournals.org

28.

29. 30.

31.

32.

33.

34.

35.

36. 37.

38. 39. 40.

41.

E631

have opposite effects on glucose and lipid metabolism. Proc Natl Acad Sci U S A. 2005;102:7386 –7391. Kelly ML, Moir L, Jones L, et al. A missense mutation in the nonneural G-protein alpha-subunit isoforms modulates susceptibility to obesity. Int J Obes (Lond). 2009;33:507–518. Ball ST, Kelly ML, Robson JE, et al. Gene Dosage effects at the imprinted cluster. PLoS One. 2013;8:e65639. Eaton SA, Williamson CM, Ball ST, et al. New mutations at the imprinted Gnas cluster show gene dosage effects of Gsalpha in postnatal growth and implicate XLalphas in bone and fat metabolism but not in suckling. Mol Cell Biol. 2012;32:1017–1029. Eaton SA, Hough T, Fischer-Colbrie R, Peters J. Maternal inheritance of the Gnas cluster mutation Ex1A-T affects size, implicating NESP55 in growth. Mamm Genome. 2013;24:276 –285. Richard N, Molin A, Coudray N, Rault-Guillaume P, Jüppner H, Kottler ML. Paternal GNAS mutations lead to severe intrauterine growth retardation (IUGR) and provide evidence for a role of XLalphas in fetal development. J Clin Endocrinol Metab. 2013;98: E1549 –E1556. Chudoba I, Franke Y, Senger G, et al. Maternal UPD 20 in a hyperactive child with severe growth retardation. Eur J Hum Genet. 1999;7:533–540. Eggermann T, Mergenthaler S, Eggermann K, et al. Identification of interstitial maternal uniparental disomy (UPD) (14) and complete maternal UPD(20) in a cohort of growth retarded patients. J Med Genet. 2001;38:86 – 89. Bastepe M, Fröhlich LF, Hendy GN, et al. Autosomal dominant pseudohypoparathyroidism type Ib is associated with a heterozygous microdeletion that likely disrupts a putative imprinting control element of GNAS. J Clin Invest. 2003;112:1255–1263. Tycko B, Morison IM. Physiological functions of imprinted genes. J Cell Physiol. 2002;192:245–258. Moore T. Genetic conflict, genomic imprinting and establishment of the epigenotype in relation to growth. Reproduction. 2001;122: 185–193. Miozzo M, Simoni G. The role of imprinted genes in fetal growth. Biol Neonate. 2002;81:217–228. Moore T, Haig D. Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet. 1991;7:45– 49. Abu-Amero S, Monk D, Apostolidou S, Stanier P, Moore G. Imprinted genes and their role in human fetal growth. Cytogenet Genome Res. 2006;113:262–270. Smith AC, Rubin T, Shuman C, et al. New chromosome 11p15 epigenotypes identified in male monozygotic twins with BeckwithWiedemann syndrome. Cytogenet Genome Res. 2006;113:313– 317.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 May 2015. at 23:36 For personal use only. No other uses without permission. . All rights reserved.

B is associated with increased intrauterine growth.

GNAS is one of few genetic loci that undergo allelic-specific methylation resulting in the parent-specific expression of at least four different trans...
734KB Sizes 0 Downloads 7 Views