|Home | About | Journals | Submit | Contact Us | Français|
Patients with McCune-Albright syndrome (MAS), characterized primarily by hyperpigmented skin lesions, precocious puberty, and fibrous dyslasia of bone, carry postzygotic heterozygous mutations of GNAS causing constitutive cAMP signaling. GNAS encodes the α-subunit of the stimulatory G protein (Gsα), as well as a large variant (XLαs) derived from the paternal allele. The mutations causing MAS affect both GNAS products, but whether XLαs, like Gsα, can be involved in the pathogenesis remains unknown. Here, we investigated biopsy samples from four previously reported and eight new patients with MAS. Activating mutations of GNAS (Arg201 with respect to the amino acid sequence of Gsα) were present in all the previously reported and five of the new cases. The mutation was detected within the paternally expressed XLαs transcript in five and the maternally expressed NESP55 transcript in four cases. Tissues carrying paternal mutations appeared to have higher XLαs mRNA levels than maternal mutations. The human XLαs mutant analogous to Gsα-R201H (XLαs-R543H) showed markedly higher basal cAMP accumulation than wild-type XLαs in transfected cells. Wild-type XLαs demonstrated higher basal and isoproterenol-induced cAMP signaling than Gsα and co-purified with Gβ1γ2 in transduced cells. XLαs mRNA was measurable in mouse calvarial cells, with its level being significantly higher in undifferentiated cells than those expressing preosteoblastic markers osterix and alkaline phosphatase. XLαs mRNA was also expressed in murine bone marrow stromal cells and preosteoblastic MC3T3-E1 cells. Our findings are consistent with the possibility that constitutive XLαs activity adds to the molecular pathogenesis of MAS and fibrous dysplasia of bone.
The stimulatory guanine nucleotide binding protein mediates receptor-activated adenylyl cyclase stimulation and, thereby, intracellular cAMP accumulation [1-3]. The α-subunit of the stimulatory G protein (Gsα) is encoded by exons 1-13 of the human GNAS gene . GNAS is a complex locus giving rise to additional products including XLαs, XXLαs, NESP55 (neurosecretory protein 55), and the A/B transcripts (also known as 1A or 1′) [5-12]. In addition, a transcript that is antisense to NESP55 (AS) is derived from GNAS [13-15]. XLαs, XXLαs, A/B, and AS are expressed paternally, while NESP55 is expressed maternally in all investigated tissues. In contrast, Gsα expression is biallelic in most tissues, including the bone ; however, a small subset of cells and tissues, such as renal proximal tubules, pituitary, testis, and thyroid, express Gsα predominantly from the maternal allele [17-21]. Gsα and its variant XLαs are identical in the C-terminal portion encoded by exons 2-13 of GNAS but differ in their N-terminal domains encoded by different first exons [9, 10, 22]. Towards its C-terminal end, the N-terminal portion of XLαs encoded by its first exon (the XL domain) becomes increasingly homologous to the corresponding region of Gsα. Consistent with the high degree of homology between Gsα and XLαs, both rat and human XLαs can couple to typical Gsα-coupled receptors and can stimulate intracellular cAMP generation in transfected cells [23-25].
McCune-Albright syndrome (MAS) is a heterogeneous disease, in which affected individuals can present with various abnormalities, including fibrous dysplasia of bone, hyperpigmented (café-au-lait) skin lesions, sexual precocity and, sometimes, other endocrine dysfunctions (thyroid nodules, acromegaly, Cushing syndrome) [26, 27]. Fibrous dysplasia of bone (FD) is a disorder with characteristic bone lesions containing a distinctive pathologic fibrous tissue [28, 29]. FD can manifest as an isolated entity restricted to a single bone or it can affect multiple bones, typically as part of MAS. Both patients with MAS and those with isolated FD carry postzygotic missense mutations of GNAS that involve Arg201 [30-34] or, much less frequently, Gln227 of Gsα . The identified amino acid changes at these residues inhibit the intrinsic GTP hydrolase activity, thereby resulting in constitutive Gsα signaling . Arg201 and Gln227 are encoded by nucleotide sequences derived from exon 8 or exon 9, respectively , and are thus shared by Gsα and XLαs. The mutations in patients with MAS/FD are heterozygous and can occur on the maternal or the paternal allele . Therefore, these mutations are predicted to affect either Gsα alone or both Gsα and XLαs. However, the possibility that the mutations on the paternal allele enhance signaling due to constitutive XLαs activity, in addition to constitutive Gsα activity, and that this enhanced signaling contributes to the pathogenesis of MAS/FD have remained unexplored.
In this study, we addressed whether MAS mutations that occur on the paternal GNAS allele affect the XLαs transcript and investigated the expression of this transcript in biopsy samples from MAS patients. Furthermore, we compared the activities of Gsα and XLαs in cultured cells with respect to cAMP generation. We also investigated the expression of XLαs transcript in mouse multipotent cells and bone cell precursors.
Except for two (patients 9 and 11), all samples were obtained from patients with at least two of the clinical triad of MAS (café-au-lait spots, precocious puberty, and fibrous dysplasia) and, in some cases, additional endocrine abnormalities (cortisol excess due to adrenal hyperplasia, hyperthyroidism, or GH excess) (Table 1). Patient 9 presented with hyperthyroidism, absence of anti-TSH receptor antibodies, and an adenoma on scintigraphy, which was removed surgically. Patient 11 had café-au-lait spots and neonatal Cushing’s syndrome only; because he was only 6 months old (Table 1), fibrous dysplasia and precocious puberty could not be considered. Patients 3, 5, and 8 underwent ovarian surgery for the removal of a large cyst that could otherwise cause ovarian torsion; the surgery was not done for diagnosis. For all patients, the samples were obtained from biopsies during surgery, snap frozen in liquid nitrogen, and kept at −80°C until experiments were performed. Patients 1, 2, 3 and 6 correspond to patients 1, 3, 5 and 10, respectively in reference . All patients, or their parents in the case of children, gave their informed consent to the genetic analysis presented herein, and the study was approved by the local institutional review board.
Complementary DNA encoding wild-type(wt) Gsα was previously subcloned into pcDNA3.1 (Invitrogen), as described . To engineer cDNA encoding wt human XLαs, the XLαs portion was PCR-amplified from human genomic DNA with XbaI/AvrII linkers, and the portion encoded by exon 2-13 was PCR-amplified from wt Gsα cDNA with AvrII/HindIII linkers. Note that the AvrII site was introduced as a silent mutation to enable ligation of these two PCR products, which were subcloned into pcDNA3.1. The substitution of the Arg to His at position 201 in Gsα cDNA and the analogous substitution in XLαs cDNA (Arg to His at position 543) were carried out by using the QuickChange site-directed mutagenesis kit (Stratagene). Both plasmid constructs were confirmed by complete sequencing of cDNA inserts. The adenoviral constructs have been described previously .
Human osteoblastic sarcoma (HOS) cells were kindly provided by Dr. Pierre Lebon (Hospital Saint Vincent de Paul). Mouse embryonic cells null for Gsα and XLαs due to homologous disruption of Gnas exon 2 (GnasE2−/E2−) have been described . GnasE2−/E2− cells and HOS cells cultured in 12-well plates in DMEM (Life Technologies, Paisley, Scotland) with 10% fetal bovine serum were cotransfected with (i) 500 ng per well plasmid DNA encoding wt-Gsα, R201H-Gsα, wt-XLαs, R201H-XLαs or the empty vector, and ii) 50 ng β-Gal-pCRE-Luc plasmid (Promega), which is a luciferase reporter gene used to assess the efficacy of transfection. Transfections were performed by using Effectene (QIAgen, Courtaboeuf, France) at a DNA/Effectene ratio of 1/5. To compare Gsα and XLαs mediated cAMP accumulation, GnasE2−/E2− cells were transduced with adenoviral vectors in order to obtain more than 90% of cells expressing the transduced vectors . MC3T3-E1 cells were purchased from American Type Tissue Collection.
Total RNA was extracted from tissues using Trizol® Reagent (Life Technologies, Paisley, Scotland) followed by a treatment in the RNeasy Mini Kit (QIAgen) according to the supplied protocol and eluted in sterile water. One μg total RNA was reverse transcribed using the kit Superscript III First Strand Synthesis System (Invitrogen) according to the supplied protocol and then subjected to PCR. Each PCR reaction mixture contained 2 μL of cDNA, 1 X PCR buffer, 10 mM dNTP mixture, 50 mM MgCl2, 10 μM of each primer, and 1 U of Taq DNA Polymerase (Invitrogen) in a final reaction volume of 50 μL. PCR samples were amplified for 1 cycle at 94°C for 2 min, followed by 40 cycles of 94°C for 30 s, 58°C for 30 s and 72°C for 30 s, and a final extension at 72°C for 10 min. PCR products were purified and sent to be sequenced. The sequences of the primers used were as follows: Gsα forward primer 5′-CGTGAGGCCAACAAAAAGAT-3′ and reverse 5′-ATGGCAGTCACATCGTTGAA-3′ (674 bp); NESP forward primer 5′-CTTCCAAAAAGGGACCCATC-3′ and reverse 5′-ATGGCAGTCACATCGTTGAA-3′ (702 bp); XLαs forward primer 5′-CGCAGTAAGCTCATCGACAA-3′ and reverse 5′-ATGGCAGTCACATCGTTGAA-3′ (668 bp).
Real-time PCR experiments were done with the LightCycler LC480 device (Roche, Meylan, France) using the LightCycler® 480 SYBR Green I Master. The reaction mixture for the real-time PCR contained 1 μl of cDNA solution in a total volume of 10 μl containing 0.3 μM of each PCR primer, and master mix in accordance with the manufacturer’s advices. The annealing temperature for PCR was 53°C (to amplify XLαs and GAPDH) or 54°C (to amplify Gsα). Specificity of amplified products was verified by performing a melting curve analysis at the end of amplification (by stepwise increase of the temperature from 55 to 95 °C).
The primers and the product sizes were: for detecting Gsα F (5′ AAGGACAAGCAGGTCTACCG 3′) and R (5′ TCTGCTTCACAATGGTGCTT 3′) 121 bp, for detecting XLαs F (5′ GGACGAAAAGATGGGCTACA 3′) and R (5′ TCTGCTTCACAATGGTGCTT 3′) 115 bp. The primers for the “house-keeping” gene, GAPDH, chosen as an endogenous internal control, were: F (5′ GAGAAGGCTGGGGCTCAT 3′) and R (5′ TGCTGATGATCTTGAGGCTG 3′); product size was 130 bp. For all quantitative RT-PCR assays involving GAPDH amplification, the following controls were always undertaken: (1) human GAPDH cDNA was used as a positive control for the effectiveness of RNA extraction and RT-PCR reactions; (2) to control for false positive PCR amplification of contaminating genomic DNA, some samples did not include reverse transcriptase; (3) water was added instead of template to test for contamination with extraneous DNA. The XLαs and Gsα transcripts in mouse bone marrow stromal and preosteoblastic cells were quantified by using the β-actin transcript as internal control. Each product was amplified with an efficiency of ~2. Primers used for the amplification of these transcripts were previously described .
After transfected cells were washed once with phosphate buffered saline (PBS), IBMX buffer (DMEM medium supplemented with 35 mM Hepes pH7.4, 2 mM IBMX, 0,1% BSA) was added into the well, and cells were incubated at 37°C for 15 min. Upon removal of the buffer, cells were subjected to one round of freeze-thaw lysis, before adding 0.5 ml 50 mM HCl into each well. Samples were kept at −20°C until the quantification of cAMP, which was accomplished by using the DELFIA cAMP kit (Perkin Elmer) according to the manufacturer protocol. In adenovirally transduced GnasE2−/E2− cells, basal cAMP accumulation and isoproterenol stimulation was carried out the same way, but cAMP concentration in each well was determined by a radioimmunoassay, as previously described . The concentration of isoproterenol was 10−5M, which has been shown to be saturating with respect to cAMP accumulation [24, 39].
Total cellular extracts were prepared from transfected cells, which were pelleted, washed twice in PBS, and lysed in a buffer containing 25 mM Hepes, 150 mM NaCl, 5 mM EDTA, 10% Glycerol, 0.02% sodium azide, 10 mM β-glycerophosphate, 10 mM sodium fluoride, and a protease inhibitor cocktail (Sigma, St. Louis, MO). Following separation by 9% SDS-PAGE, proteins were blotted onto HyBond C membranes (Amersham), and reacted to specific antisera in a Tris-HCl buffered saline solution containing 0.1% Tween-20. An antibody against the common C-terminus of Gsα and XLαs (sc-383) were used to identify both of these proteins. An antibody against β-actin (sc-8432) was used as a gel loading control. Bound antibodies were detected by horseradish peroxidase-conjugated secondary antibodies, followed by chemiluminescence (ECL, Amersham). Primary antibodies were purchased from Santa Cruz Biotechnology, and the antibody against rabbit IgG (#4010-05) from Southern Biotech.
COS7 cells were triply transduced with adenovirus encoding XLαs or Gsα together with adenovirus encoding Gγ2 and his6-tagged Gβ1. Forty-eight hours after transduction, cells were harvested and membrane extracts were prepared by gentle lysis with hypotonic saline solution, followed by homogenization with a Dounce homogenizer. Following further extraction with 1% Na-cholate, proteins were purified through the use of immobilized nickel affinity chromatography, as previously described . Eluted proteins were separated by 9% SDS-PAGE and Western blot was performed by using either the aforementioned anti-Gsα C-terminal antibody (for Gsα), an anti-human XLαs antiserum directed against an epitope within the XL domain of human XLαs (for XLαs) , or an anti-pan Gβ antibody (Santa Cruz Biotechnology, sc-378).
Experiments using the mouse were done in compliance with the International Animal Care and Use Convention and approved by Massachusetts General Hospital (MGH) subcommittee on research animal care. Primary calvarial osteoblastic cells were isolated from wild-type mice at postnatal day 6 by serial collagenase digestion . Adherent cells were plated at 12,500 cells/cm2 and osteogenic differentiation induced with 10 mM β-glycerol phosphate and 50 μg/ml ascorbic acid. At specified time points some wells were harvested for total RNA isolation, while others were stained with 1% alizarin red to assess mineralization. Alizarin red deposition was then quantified by extraction with cetylpyridinium chloride . Mouse bone marrow stromal cells were isolated from four-week old mice upon flushing the marrow space by Hank’s buffered saline solution (Invitrogen). Adherent stromal cells were cultured in α-MEM medium (Invitrogen) containing 10% fetal bovine serum for two days before RNA isolation.
Osteoblastic cells were harvested from neonatal calvariae of Osx-GFP/+ mice (kindly provided by Stephen Rodda and Andrew McMahon, Harvard Medical School)  by serial collagenase digestion. Fractions 3–6 were pooled and resuspended in phosphate-buffered saline (PBS) with 2% fetal bovine serum. Cells were incubated with a biotinylated antibody directed against alkaline phosphatase (ALP) (kindly provided by Dr. Mark Horowitz, Yale University) followed by APC-conjugated streptavidin (eBioscience). Cells were isolated based on expression of Osx-GFP and ALP using a Dako MoFlo high-speed sorter.
Experiments measuring cAMP accumulation in transfected GnasE2−/E2− and HOS cells were analyzed by taking the mean and standard error (S.E.M) of independent experiments. Statistical significance of observed differences was determined by using Student’s t-test for unpaired samples. A P value of less than 0.05 was considered significant.
To investigate whether disease-causing GNAS mutations that lead to constitutive cAMP signaling affect not only Gsα but also XLαs, we examined twelve patients, eleven of whom presented with clinical features of MAS and one with toxic thyroid adenoma. An activating GNAS mutation, along with its allelic origin, had been revealed in four of the cases . To find out whether the eight previously unreported patients carried missense Gsα mutations at either Arg201 or Gln227, the two residues at which activating mutations have been previously identified, we isolated total RNA from biopsy samples and amplified, by conventional RT-PCR, three different GNAS transcripts with documented biallelic (Gsα), paternal (XLαs), or maternal (NESP55) expression. Direct nucleotide sequence analysis of the RT-PCR products indicated that five of the 8 cases carried mutations (patients 4, 9, 10, 11 and 12) at position 201 (according to Gsα) (Table 1). The remaining 3 cases demonstrated the wild-type sequence both at this residue and at position 227 (patients 5, 7 and 8). The mutation was detected in the XLαs transcript in patient 9, 10, 11 and 12, indicating a paternal origin. In contrast, the NESP55 transcript carried the mutation in patient 4, indicating a maternal origin (Table 1). The allelic origin of the mutation in biopsy samples from the previously reported cases was also confirmed, and thus, the number of patients with paternal or maternal mutations was five and four, respectively (Table 1).
To quantify the relative levels of XLαs transcript in the biopsy samples, we performed real-time RT-PCR, using GAPDH as a reference control. Although this analysis could not be performed for patients 11 and 12 due to insufficient amount of total RNA from the biopsy material, analysis of the remaining samples revealed that the level of XLαs transcript was higher in cases with paternal mutations than maternal mutations (Fig. 1). For example, whereas XLαs transcript could not be amplified in the bone biopsy samples from patients 2 and 4, who had the mutations on the maternal allele, it could easily be amplified in the bone biopsy sample from patient 10, who had the mutation on the paternal allele. In contrast, no similar correlation was evident between the level of Gsα expression and the allelic origin of the mutation (Fig. 1).
A rat XLαs mutant corresponding to the Gsα-Q227L mutant has been shown to display constitutive activity in transfected cells . To determine whether mutations affecting Arg201, which are exceedingly more frequent in patients with MAS/FD, also result in constitutive activity when placed in the human XLαs backbone, we examined the basal signaling activity through XLαs-R543H, which is analogous to Gsα-R201H. Agonist-independent basal cAMP accumulation in GnasE2−/E2− cells (mouse embryonic fibroblasts null for endogenous Gsα and XLαs) transiently expressing XLαs-R543H was significantly higher than cells transiently expressing wild-type XLαs (3.5±1 fold), and this finding was similar to that observed for Gsα-R201H compared to wild-type Gsα (1.7±0.2 fold) (Fig. 2A). Likewise, HOS cells (a cell line derived from human osteosarcoma) transiently transfected with cDNA encoding either wild-type or mutant forms of XLαs and Gsα demonstrated that XLαs-R543H, like Gsα-R201H, is constitutively active. Basal cAMP accumulation in HOS cells transiently expressing XLαs-R543H or Gsα-R201H was 19±3.9 fold or 8.6±4.9 fold (p<0.05) higher than in those expressing native forms of these proteins, respectively (Fig. 2B). Western blot analyses showed that in GnasE2−/E2− cells both constitutively active XLαs and Gsα mutants were expressed slightly lower than wild-type XLαs and Gsα (Fig. 2C), confirming that the higher cAMP accumulation through the mutants was not due to their higher expression levels than wild-type proteins. The expression level of the constitutively active XLαs mutant attained in transfected HOS cells appeared to be slightly higher than other transiently expressed proteins (Fig. 2C). However, this expression difference was not sufficient to account for the enhanced basal signaling through the XLαs mutant, because when normalized for expression levels, XLαs-R543H showed ≈11-fold higher basal cAMP than wild-type XLαs.
A previous study by our group has shown markedly higher basal cAMP accumulation in OK cells transiently expressing XLαs compared to Gsα , suggesting that wild-type XLαs and wild-type Gsα may differ from each other in their abilities to stimulate cAMP formation. To address this question, we transduced GnasE2−/E2− cells with increasing multiplicity of infection of adenovirus encoding Gsα or XLαs and, subsequently, determined basal cAMP accumulation. In XLαs-expressing cells, basal cAMP accumulation was approximately 2-fold higher than in Gsα-expressing cells despite equivalent MOI used for the transductions (Fig. 3A). When these cells were stimulated by a saturating concentration of isoproterenol – acting on the endogenously expressed β-adrenergic receptor – cAMP accumulation was also higher in XLαs expressing cells than Gsα expressing cells (Fig. 3A). The observed differences between XLαs and Gsα mediated cAMP accumulation did not simply reflect the relative expression levels of these proteins, as Western blot (Fig. 3B) and real-time RT-PCR (Fig. 3C) analyses demonstrated that XLαs was not expressed at higher levels than Gsα in the transduced GnasE2−/E2− cells.
Gβγ subunits are both involved in controlling the inactive state of Gα subunits and ensuring effective Gα activation by receptors (see references [44, 45] for review). We thus asked whether XLαs, like Gsα, is able to form a stable heterotrimeric complex with Gβγ. To address this question, we performed immobilized metal affinity chromatography using membrane proteins of adenovirally transduced COS7 cells transiently expressing either Gsα or XLαs in the presence or absence of his6-tagged Gβ1 and Gγ2. On analysis of the purified samples by Western blots, Gsα and XLαs immunoreactivity were detected in samples from cells expressing his6-tagged Gβ1 and Gγ2 in combination with either Gsα or XLαs, but not from those expressing either Gsα or XLαs alone (Fig. 4). These results thus indicated that XLαs can interact with Gβ1γ2.
To explore the expression of XLαs in multipotent and preosteoblastic cells, we isolated mouse calvarial cells. RT-PCR experiments could detect XLαs as well as Gsα expression in these cells both before and after osteogenic differentiation, although no significant alterations were detected in the levels of these transcripts during differentiation (Fig. 5A). However, real-time quantitative RT-PCR analysis of mouse calvarial cells upon fluorescent activated cell sorting revealed that XLαs mRNA is more abundant in cells that lack the expression of both osterix and alkaline phosphatase compared to those that express these preosteoblastic markers (Fig. 5B). This trend was different from Gsα levels, which appeared to be similar in both cell populations (Fig. 5B). In all groups of sorted calvarial cells, however, XLαs expression appeared to be markedly lower than Gsα expression, being 0.26 ± 0.04% of Gsα expression in cells lacking both osterix and alkaline phosphatase expression and 0.05 ± 0.02% of Gsα expression in cells expressing both of these markers. In addition to mouse calvarial cells, we were able to amplify the XLαs transcript by using RT-PCR and total RNA from undifferentiated murine bone marrow stromal cells and the preosteoblastic MC3T3-E1 cells (Fig. 5C). Our attempts to detect endogenous XLαs protein by Western blots using the aforementioned antibody against the common C-terminus of Gsα and XLαs did not yield specific immunoreactive bands (data not shown).
In this study, we determined that constitutively activating GNAS mutations found in patients with MAS can occur on the paternal allele, thus affecting the XLαs transcript. Our measurement of XLαs transcript levels in biopsy samples revealed higher XLαs expression in those tissues with paternal GNAS mutations than maternal mutations. Moreover, we showed that MAS mutations can lead to constitutive cAMP accumulation not only through Gsα but also through XLαs in transfected cells, that XLαs is more potent than Gsα in mediating basal and isoproterenol-stimulated cAMP accumulation, and that XLαs can interact with Gβ1γ2. Finally, we demonstrated that XLαs mRNA is expressed in mouse calvarial osteoblasts, bone marrow stromal cells, and a pre-osteoblastic cell line.
We identified constitutively active GNAS mutations in nine of our 12 patients. No mutation at either Arg201 or Gln227 was detected in the remaining three cases. It is likely that the abundance of the mutant cells relative to wild-type cells is markedly low in the biopsy samples from these three patients, as conventional PCR approaches can sometimes fail to amplify the mutant allele in cases with low degree of mosaicism [46, 47]. Analyzing the Gsα transcript, we also detected no activating mutations in patients 3 and 9, but the paternally expressed XLαs transcript appeared to bear such a mutation (R201H) in each case. This finding is consistent with the tissue-specific paternal silencing of Gsα in the ovary (patient 3) and the thyroid (patient 9) [19, 21].
A previous study  investigated the origin of mutations in different tissues from patients with MAS, identifying an almost equal allelic distribution (4 maternal and 5 paternal). In the five new patients with MAS that we describe here, the activating mutation is on the maternal allele in one patient, and on the paternal allele in four patients. Since these paternal mutations affect both Gsα and XLαs, and since both of these GNAS products show imprinting (albeit tissue-specific with respect to Gsα), it is possible that certain features of MAS may develop depending on the allelic origin of the mutation. Indeed, cases with acromegaly have been associated with maternal mutations [18, 37], which likely reflects the paternal silencing of Gsα expression in the pituitary . Data from our study are consistent with those findings. None of the patients with paternal mutations developed clinical features of acromegaly, whereas patient 4, in addition to the previously reported patient 6 , who both carried maternal mutations, showed clinical features consistent with excess growth hormone. Although previous investigations have shown that XLαs expression is restricted to the paternal GNAS allele, it is currently unknown whether any MAS features develop exclusively or predominantly in patients with paternal mutations. In our study group, hyperthyroidism was present only in patients 3 and 9, both of whom carried a paternal mutation. However, a maternal constitutively activating GNAS mutation has been previously documented in a case with autonomous thyroid adenoma , thus arguing against a direct correlation between paternal mutations and the involvement of thyroid in the MAS phenotype. Careful investigations of additional patients may reveal additional correlations between specific MAS features and the allelic origin of GNAS mutations, and may identify a role for constitutive XLαs activity in the pathogenesis of this disorder. In fact, findings in patients 3 and 9 are rather interesting. Gsα is partially silenced from the paternal allele in the thyroid [19-21], and therefore, one might expect that the effect of a paternal mutation is more significant on XLαs than Gsα in that tissue. It is thus tempting to speculate that constitutive XLαs activity had an important role in the development of hyperthyroidism in these two patients.
A single study has shown that XLαs is biallelic, although with some variation, in both normal and mutation-carrying bone marrow stromal cells of patients with fibrous dysplasia , raising the possibility that constitutively active XLαs mutant is expressed regardless of the origin of the mutations causing MAS and fibrous dysplasia. Further investigations are needed to confirm this intriguing finding and address this possibility. Note, however, that our amplification of Gsα, XLαs, and NESP55 transcripts from MAS biopsy samples yielded data that are consistent with monoallelic expression of XLαs in bone (see Table 1). The discrepancy in these results may reflect technical differences. In the study showing biallelic XLαs expression , investigations were performed upon cloning the cells, and this may have resulted in a loss of XLαs imprinting. Disruption of allele-specific methylation as a result of cloning has been previously described .
Our findings are consistent with previous reports with respect to tissue-specific Gsα imprinting [19-21]. Only the maternal Gsα transcript was amplified from the biopsies of ovary and thyroid, whereas both wild-type and mutant Gsα transcripts were detectable in adrenals and bone. In fact, we also detected a heterozygous single nucleotide polymorphism (rs7121) in the Gsα transcript amplified from the healthy bone tissue of patient 10 (data not shown), further indicating biallelic Gsα expression in this tissue.
We could amplify the XLαs transcript more readily from biopsy samples with paternal mutations than maternal mutations. Because the sequence of the RT-PCR products from paternally mutated samples did not indicate amplification of wild-type transcripts, it appears unlikely that the amplicons reflect both the wild-type and the mutant XLαs transcript. As such, there may be a tendency of mutant XLαs transcripts to be more abundant compared to wild-type XLαs transcripts. However, it is possible that the difference between transcript levels from paternal and maternal mutations reflect the small sample size in our study. Analyses of additional biopsy samples will help address these possibilities, and if the difference between mutant and wild-type XLαs transcript levels is confirmed, further investigations will be necessary in order to elucidate the underlying molecular mechanisms.
Our findings regarding the activity of wild-type human XLαs are in keeping with the findings of a recent study, which showed, through the use of transient and stable transfections, that rat XLαs has higher basal and isoproterenol-stimulated activity than Gsα . The increased activity of XLαs could possibly reflect a deficiency in the interaction between XLαs and Gβγ subunits, but our findings argue against a prominent defect between XLαs-Gβγ coupling. Supporting our findings are the results of a previous study , which provided indirect evidence suggesting that rat XLαs forms a heterotrimeric complex with Gβ1γ2. On the other hand, although overt differences between XLαs and Gsα regarding Gβγ interaction can thus be ruled out (at least regarding the interaction with Gβ1γ2), it is theoretically possible that subtle differences in the efficiency of interaction contribute to the differences between the potencies of XLαs and Gsα in stimulating cAMP generation. Additionally, XLαs may be unable to interact with certain other Gβγ combinations. This, however, is less likely to provide an explanation as to why XLαs is more potent than Gsα, given that a variety of cell lines (mouse embryonic fibroblasts, HOS cells, HEK293 cells , and OK cells ), which likely differ from one another with respect to the repertoire and expression levels of the different Gβγ subunits, appear to be permissive for the relatively enhanced activity of XLαs. Alternatively, Gsα and XLαs may be subject to different regulatory constraints upon activation. For example, Gsα loses its membrane avidity upon activation [51-53], and this mechanism may not apply to XLαs, thus allowing the latter to activate adenylyl cyclase in a sustained manner. Consistent with this possibility, a recent study have shown that, whereas a constitutively active Gsα mutant is localized predominantly to the cytoplasm, the cognate XLαs mutant is localized to the plasma membrane . Finally, it is also possible that the intrinsic ability of XLαs to stimulate adenylyl cyclase is higher than that of Gsα. It will be important to dissect these possibilities and elucidate the differences between the cellular activities of XLαs and Gsα as stimulatory G protein α-subunits.
The XLαs transcript is expressed in cells that can differentiate into osteoblasts, albeit at much lower levels than the Gsα transcript. Although the expression of neither of these transcripts seemed to alter dramatically during osteogenic differentiation in vitro, the level of XLαs, but not Gsα, was reduced in calvarial cells committed to the osteoblast lineage compared to others. This finding may simply indicate a correlation between the expression level of XLαs and the degree of cellular differentiation in mesenchymal cells. In agreement with this explanation, XLαs expression in white adipose tissue has been shown to decline shortly after birth in mice . Alternatively, our finding may suggest that differentiation into the osteoblastic lineage requires XLαs levels to be suppressed, a hypothesis that is consistent with the potential role of constitutive XLαs signaling in the pathogenesis of FD. Future investigations are necessary to elucidate the role of XLαs in osteogenic cell fate and early osteoblastic differentiation.
In summary, we have shown that some patients with MAS/FD have paternal activating mutations within GNAS affecting both Gsα and XLαs. Our results also indicate that XLαs, which is expressed in multipotent stromal cells and preosteoblasts, has higher activity than Gsα with respect to adenylyl cyclase stimulation. These findings are consistent with the hypothesis that XLαs-mediated constitutive signaling may contribute to the pathogenesis of MAS and FD.
We thank Dr. H. Jüppner for helpful discussions and critically reviewing the manuscript. We also thank Drs. S. Rodda and A. McMahon for providing the mice with Osterix driven GFP expression and Dr. M. Horowitz for providing the biotinylated alkaline phosphatase antibody. This work was supported in part by grants from the National Institute of Diabetes and Digestive and Kidney Diseases (R01DK073911 to M.B.) and the National Institute of Arthritis and Musculoskeletal and Skin Diseases (K08AR054741 to J.Y.W), as well as by an AIRC (Associazione Italiana Ricerca Cancro, Milan, Italy) grant (to G.M.). C.A. received a research fellowship award from Gülhane Military Medical Academy, Turkish General Staff, Ankara, Turkey; V.M. received a grant from the French Society of Endocrinology; A.L. received a research fellowship grant from the European Society of Pediatric Endocrinology.
Funding: the National Institutes of Health
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest: All authors have no conflicts of interest.