Portal de Periódicos da CAPES

Pseudohypoparathyroidism: From Bedside to Bench and Back

1999; Oxford University Press; Volume: 14; Issue: 8 Linguagem: Inglês

10.1359/jbmr.1999.14.8.1255

1523-4681

Michael A. Levine,

Metabolism, Diabetes, and Cancer

One approach to the investigation of metabolic disorders relies upon transfer of the clinical problem to the laboratory, where both insight and technology can be rigorously applied in a controlled environment. For the medical scientist, ultimate resolution of the metabolic disorder will of course require the successful translation of the laboratory discovery to clinical usefulness, a process often described as "bench to bedside." This approach has been applied to the study of many endocrine disorders, including pseudohypoparathyroidism (PHP), an unusual condition in which biochemical hypoparathyroidism (i.e., hypocalcemia and hyperphosphatemia) arises from the failure of target tissues to respond appropriately to the biological actions of parathyroid hormone (PTH). Thus the pathophysiology of PHP differs fundamentally from true hypoparathyroidism, in which PTH secretion rather than PTH responsiveness is defective. More than 50 years have passed since Fuller Albright and his associates first described the failure of patients with PHP to show either a calcemic or a phosphaturic response to administered parathyroid extract.1 These early clinical studies led to the hypothesis that PHP is due to a defect in PTH signaling but could not even begin to anticipate the complexity of that signaling process! The PTH signal transduction cascade consists of at least three membrane-bound components: receptors that detect ligand in extracellular milieu, signal generating enzymes such as adenylyl cyclase, and the heterotrimeric (α, β, γ) G proteins (reviewed elsewhere2). The classical PTH receptor that is expressed in bone and kidney is a ∼75 kDa glycoprotein that is often referred to as the PTH/parathyroid hormone receptor protein (PTHrP) or type 1 PTH receptor (PTHr).3 The type 1 PTHr binds both PTH and PTHrP, a factor made by diverse tumors that cause humorally mediated hypercalcemia, with equivalent affinity, which accounts for the similar activities of both hormones. By contrast, a second PTH receptor, termed the type 2 receptor protein, is not expressed in conventional PTH target tissues (i.e., bone and kidney) and interacts with PTH but not PTHrP.4, 5 Both PTHrs are members of a large family of receptors that can bind hormones, neurotransmitters, cytokines, light photons, taste and odor molecules. These receptors consist of a single polypeptide chain that is predicted by hydrophobicity plots to span the plasma membrane seven times (i.e., heptahelical), forming three extracellular and three or four intracellular loops and a cytoplasmic carboxy-terminal tail. The heptahelical receptors are coupled by G proteins2 to signal effector molecules localized to the inner surface of the plasma membrane. Highly specific associations among at least 20 α, 6 β, and 12 γ chains generate a diversity of heterotrimeric G proteins that have the ability to discriminate among a multitude of receptor and effector molecules. Hormone binding to a receptor facilitates activation of the G protein, a process in which the α chain exchanges bound GDP for GTP and dissociates from the βγ dimer and the receptor. The free, GTP-bound form of the α chain is the primary modulator of relevant effector molecules, although βγ dimers can also influence activity of many effectors (e.g., some forms of adenylyl cyclase and phospholipase C). An intrinsic GTPase associated with the α chain acts as a molecular timing mechanism, and after a predetermined interval, GTP is hydrolyzed to GDP. The inactive GDP-bound α chain reassociates with a βγ dimer, and the heterotrimeric G protein is ready for another cycle of hormone activation. Interaction of PTH with its receptor activates intracellular signal effector systems that generate the second messengers cAMP,6, 7 inositol 1,4,5-trisphosphate and diacylglycerol,8, 9 and cytosolic calcium.10-13 The best characterized mediator of PTH action is cAMP, which rapidly activates protein kinase A.14 The relevant target proteins that are phosphorylated by protein kinase A and the precise mode(s) of action of these proteins remain uncharacterized, though proteins that activate genes responsive to cAMP and ion channel proteins are strong candidates. In contrast to the well recognized biologic effects of cAMP in PTH target tissues, the physiological importance of metabolites of phosphotidylinositol hydrolysis and intracellular calcium as PTH-induced second messengers has not yet been established. Studies of the expressed type 1 PTHr have revealed that ligand activation of these diverse second messengers derives from the ability of the receptor to interact with several different G proteins. The agonist-bound type 1 PTHr can activate members of the Gq/11 family, and thereby stimulate phospholipase C, and can activate Gs to stimulate adenylyl cyclase.15, 16 These studies have revealed that the number of PTHrs expressed, as well as the concentration of G protein and PTH, cooperate to determine the precise signal response. In addition to PTH resistance, most patients with PHP also manifest a peculiar constellation of developmental and somatic defects that are collectively termed Albright's hereditary osteodystrophy (AHO).1 Subjects with classical AHO have short stature, round faces, obesity, brachydactyly, and subcutaneous ossifications, but in some patients many of these features may be absent or subtle. The subsequent identification of individuals with AHO who lacked apparent hormone resistance led Albright to propose the rather awkward term pseudopseudohypoparathyroidism (pseudoPHP) to describe this normocalcemic variant of PHP.17 Characterization of the molecular basis for PHP commenced with the observation that cAMP mediates many of the actions of PTH on kidney and bone, and that administration of biologically active PTH to normal subjects leads to a significant increase in the urinary excretion of nephrogenous cAMP.18 The PTH infusion test remains the most reliable test available for the diagnosis of PHP and enables distinction between the several variants of the syndrome. Thus, patients with PHP type I fail to show an appropriate increase in urinary excretion of both cAMP and phosphate,18 while subjects with the less common type II form show a normal increase in urinary cAMP excretion but have an impaired phosphaturic response.19 Subjects with pseudoPHP have a normal urinary cAMP response to PTH,18, 20 which distinguishes them from occasional patients with PHP type Ia who maintain normal serum calcium levels without treatment.21 Albright's original description of PHP emphasized PTH resistance as the biochemical hallmark of this disorder. Resistance to PTH alone would be consistent with a defect in the cell surface receptor specific for PTH. However, some patients with PHP type I display resistance to multiple hormones, including PTH, thyroid-stimulating hormone (TSH), gonadotropins, and glucagon, whose effects are mediated by cAMP.22 Cell membranes from most of these patients have an ∼50% reduction in expression or activity of Gsα protein, a property that defines this condition as PHP type Ia. In a given cell type, the responsiveness to stimulus, as well as the character of the response, is determined by not only the complement of receptors, but also by the availability of G proteins. Thus, a generalized deficiency of Gsα may reduce the ability of many hormones and neurotransmitters to activate adenylyl cyclase, and could thereby cause hormone resistance. PseudoPHP is genetically related to PHP type Ia. Within a given kindred, some affected members will have only AHO (i.e., pseudoPHP) while others will have hormone resistance as well (i.e., PHP type Ia), despite equivalent functional deficiency of Gsα in tissues that have been analyzed.20 Gsα deficiency in patients with AHO results from heterozygous inactivating mutations in the GNAS1 gene, a complex gene that maps to 20q13.2.23 The human GNAS1 gene24 is comprised of at least 17 exons, including three alternative first exons.25, 26 Alternative splicing of nascent transcripts derived from exons 1–13 generates four mRNAs that encode two Gsα proteins with apparent molecular weights of 45 kDa (exclusion of exon 3) and two isoforms of apparent molecular weights of 52 kDa24 that exhibit specific patterns of tissue expression.27 Both long and short forms of Gsα can stimulate adenylyl cyclase and open calcium channels,28 but biochemical characterization of these isoforms has revealed subtle differences in the binding constant for GDP, the rate at which the forms are activated by agonist binding, efficiency of adenylyl cyclase stimulation, and the rate of GTP hydrolysis that are of uncertain significance.28-31 Additional complexity in the processing of the GNAS1 gene derives from the use of alternative first exons that generate novel transcripts. Because these alternative forms of Gsα lack amino acid sequences encoded by exon 1, which are required for interaction with Gβγ and attachment to the plasma membrane, it is unlikely that these proteins can function as transmembrane signal transducers. In one case, a Gsα transcript is produced with an alternative first exon that lacks an initiator ATG; thus, a truncated, nonfunctional Gsα protein is translated from an inframe ATG in exon 2.32 In two other instances, unique transcripts are generated using additional coding exons that are present upstream of the exon 1 used to generate Gsα. The more 5′ of these exons encodes the neuroendocrine secretory protein NESP55, a chromogranin-like protein that contains sequences derived from exon 2 of GNAS1 in the 3′ uncoded region.33, 34 The more downstream alternative exon encodes a 51 kDa protein, and when spliced to exons 2–13 results in a transcript that encodes a larger Gsα isoform (XLαs).35 Molecular studies of DNA from subjects with AHO have disclosed inactivating mutations in the GNAS1 gene36-49 that account for a 50% reduction in expression or function of Gsα protein. All patients are heterozygous and have one normal GNAS1 allele and one defective allele. A large variety of mutations in the GNAS1 gene have been identified, including missense mutations,39, 41-43, 48 point mutations in sequences required for efficient splicing,40 and small deletions.40, 41, 44, 50 Although novel mutations have been found in nearly all of the kindreds studied, a 4-base deletion in exon 7 has been detected in multiple families,49-51 and an unusual missense mutation in exon 13 (A366S; see below) has been identified in two unrelated young boys,37 suggesting that these two regions may be genetic "hot spots." These studies confirm the molecular defect in AHO, but they do not explain the striking variability in biochemical and clinical phenotype. Why do some Gsα-coupled pathways show reduced hormone responsiveness (e.g., PTH, TSH, gonadotropins) whereas other pathways are clinically unaffected (adrenocorticotropic hormone in the adrenal and vasopressin in the renal medulla)? One possible interpretation of variable hormonal responsiveness is that haploinsufficiency of Gsα is tissue specific; that is, in some tissues a 50% reduction in Gsα is still sufficient to facilitate normal signal transduction. However, this explanation leaves unanswered the even more intriguing paradox of why some subjects with Gsα deficiency have hormone resistance (PHP type Ia), whereas others have apparently normal hormone responsiveness (pseudoPHP). Analysis of published pedigrees has indicated that in most cases maternal transmission of Gsα deficiency leads to PHP type Ia, whereas paternal transmission of the defect leads to pseudoPHP,20, 47, 52, 53 findings that have implicated genomic imprinting of the GNAS1 gene as a possible regulatory mechanism.53 Recent studies have indeed confirmed that the GNAS1 gene is imprinted, but in a far more complex manner than had been anticipated. Two upstream promoters, each associated with a large coding exon, lie 35 kb upstream of GNAS1 exon 1. These promoters are only 11 kb apart, yet show opposite patterns of allele-specific methylation and monoallelic transcription. The more 5′ of these exons encodes NESP55, which is expressed exclusively from the maternal allele. By contrast, the XLαs exon is paternally expressed.25, 26 Despite the simultaneous imprinting in both the paternal and maternal directions of the GNAS1 gene, expression of Gsα appears to be biallelic in all human tissues examined.25, 26, 54 The lack of access to relevant tissues in patients with PHP type Ia has hindered studies of Gsα expression and stimulated attempts to develop suitable animal models. Recently two groups have succeeded in developing mice in which one Gnas gene is disrupted, thereby generating murine models of PHP type Ia.55, 56 Although these mice have reduced levels of Gsα protein, they lack many of the features of the human disorder. Biochemical analyses of these heterozygous Gnas knockout mice suggest that Gsα expression may derive from only the maternal allele in some tissues (e.g., renal cortex) and from both alleles in other tissues (e.g., renal medulla). Accordingly, mice that inherit the defective Gnas gene maternally express only that allele in imprinted tissues, such as the PTH-sensitive renal proximal tubule, in which there is no functional Gsα protein. By contrast, the 50% reduction in Gsα expression that occurs in nonimprinted tissues, which express both Gnas alleles, may account for more variable and moderate hormone resistance in these sites (e.g., the thyroid). Confirmation of this proposed mechanism in patients with AHO will require demonstration that the human Gsα transcript is paternally imprinted in the renal cortex. Despite these exciting molecular insights, many of the endocrine manifestations of Gsα deficiency in subjects with PHP type Ia are still incompletely explained. Hypocalcemia and hyperphosphatemia typically are not present until after age 5 years, but elevations in serum PTH can be documented much earlier and sometimes are associated with mild hypercalcemia! Primary hypothyroidism, due to impaired responsiveness of the thyroid gland to TSH, frequently precedes the development of functional hypoparathyroidism, and often is detected during neonatal screening. As expected, PHP type Ia patients do not develop goiters despite elevated levels of serum TSH, and antithyroid antibodies are absent. Hypogonadism or delayed development of secondary sexual characteristics is very common, particularly in young girls, but levels of gonadotropins are generally not elevated.57 These findings, more plus recent observations,58 suggest that function of the pituitary or hypothalamus may be directly, or indirectly, affected in patients with PHP type Ia. In AHO, inherited Gsα gene mutations reduce expression or function of Gsα protein. By contrast, in the McCune–Albright syndrome, somatic mutations in the Gsα gene enhance activity of the protein.59 These mutations lead to constitutive activation of adenylyl cyclase and produce proliferation and autonomous hyperfunction of hormonally responsive cells. The clinical significance of Gsα activity as a determinant of hormone action is emphasized by the description by Iiri et al.37 of two males with both precocious puberty and PHP type Ia. These two unrelated boys had identical GNAS1 gene mutations (missense mutation in exon 13, A366S) that resulted in a temperature-sensitive Gsα that is constitutively activated in the cooler environment of the testis, while being rapidly degraded in other tissues at normal body temperature. Thus, different tissues in these two individuals could show hormone resistance (to PTH and TSH), hormone responsiveness (to adrenocorticortropic hormone), or hormone independence (to luteinizing hormone). What about subjects with PHP type I who lack features of AHO and have hormone resistance that is limited to PTH target tissues? Accessible cells from these patients have normal Gsα activity, a biochemical hallmark that defines this variant as PHP type Ib.22 As in PHP type Ia, subjects with PHP type Ib have a defective nephrogenous cAMP response to PTH. However, subjects with PHP type Ib who have elevated levels of PTH often manifest skeletal lesions similar to those that occur in patients with hyperparathyroidism.60 Specific resistance of target tissues to PTH, and normal activity of Gsα, had implicated decreased expression or function of the PTH/PTHrP receptor as the cause for hormone resistance. Fibroblasts from some, but not all, PHP type Ib patients accumulate less cAMP in response to PTH61 and contain decreased levels of mRNA encoding the PTH/PTHrP receptor.62 Several lines of evidence suggest that the primary defect in PHP type Ib is not in the gene encoding the PTH/PTHrP receptor, however. First, in most cases of PHP type Ib, pretreatment of cultured fibroblasts with dexamethasone normalizes the defective PTH-induced cAMP response and increases expression of PTH/PTHrP receptor mRNA.62 Second, molecular studies have failed to disclose mutations in the coding exons63 and promoter regions64 of the PTH/PTHrP receptor gene or its mRNA.65 Third, mice66 and humans67 that are heterozygous for inactivation of the gene encoding the PTH/PTHrP receptor do not manifest PTH resistance or hypocalcemia. Finally, inheritance of two defective type PTH/PTHrP receptor genes results in Blomstrand chondrodysplasia, a lethal genetic disorder characterized by advanced endochondral bone maturation.67 Most cases of PHP type Ib appear to be sporadic, but familial cases have been described in which transmission of the defect is most consistent with an autosomal dominant pattern.68, 69 Recent studies have used gene mapping to identify the molecular defect in PHP type Ib.70 In one study the unknown gene was mapped to a small region of chromosome 20q13.3 near the GNAS1 gene, thus raising the possibility that some patients with PHP type Ib have inherited a defective promoter or enhancer that regulates expression of Gsα in the kidney70 Although PHP type Ia and type Ib are uncommon syndromes, study of these unusual patients continues to provide us with novel and important insights about signal transduction that enhance our understanding of all endocrine diseases. The ability to study the consequences and pathophysiology of Gsα deficiency in newly created animal models of PHP type Ia now affords us the opportunity to uncover the more mysterious nuances of this human disorder. Continued gene mapping studies will no doubt soon disclose the basis for PHP type Ib, and may provide even greater surprises. With time, and great patience, some of these discoveries will advance our ability to care for subjects with PHP and other disorders of hormone signaling, and will bring our focus back from bench to bedside.