| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Medicine, University of Toronto, and the Division of Endocrinology and Metabolism, St. Michael’s Hospital (T.M.M., L.G.R.), Toronto, Ontario, Canada M5B 1A6; and the Endocrine Unit (P.D., F.R.B.), Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114
Correspondence: Address all correspondence and requests for reprints to: Timothy M. Murray, M.D., F.R.C.P.(C), St. Michael’s Hospital, 38 Shuter Street, Suite 213, Toronto, Ontario M5B 1A6, Canada. E-mail: drtimothy.murray{at}utoronto.ca
 |
Abstract |
|---|
|
I. Introduction |
|---|
|
II. Structure of PTH |
|---|
). The intact, secreted form of the hormone is generated by cleavage of a prohormone (proPTH) containing a 6-amino acid N-terminal extension. proPTH, in turn, is produced by proteolysis of a larger, short-lived precursor, preproPTH, which is the initial translation product (16). PreproPTH incorporates an additional 25-residue N-terminal signal sequence, rich in hydrophobic amino acids, that, together with the 6-residue "pro" sequence, is necessary for efficient transport into the endoplasmic reticulum and subsequent cleavage to the mature hormonal form (17, 18, 19, 20). The C-terminal portion of PTH also is needed for efficient processing of the hormone. Thus, in cells transfected with cDNA encoding preproPTH(1–40) or preproPTH(1–52), the proPTH forms were not cleaved but, instead, were degraded intracellularly (21). These observations have suggested one possible role for the C terminus of PTH, i.e., to ensure high-fidelity processing and effective transport of the hormone through the parathyroid secretory apparatus.
|
). The homology is strongest in the N terminus of the molecule, where 32 of the first 38 residues are identical and even those differences that do occur are relatively conservative. The greatest evolutionary variation is evident in the middle region of the hormone, between residues 39 and 52, whereas several stretches of high homology can be found in the C-terminal region, i.e., PTH(53–84). For example, PTH(53–61) shows 100% identity among mammals except for conservative exchanges of acidic residues (Asp for Glu, or vice versa) at positions 56 and 61 in the rodent sequences. Several residues are conserved also in the chicken and zebrafish sequences. Furthermore, the 14-residue sequence PTH(65–78) shows variation at only three positions among mammals, and residues Lys80, Lys82, and Gln84 are invariant as well. The extensive sequence conservation at the PTH N terminus fits well with abundant evidence that this region of the molecule is both necessary and sufficient for full activation of the classical PTH1R (see Section III), whereas the high homology in portions of the C terminus of the hormone had remained enigmatic before the emergence of more recent evidence of the existence of CPTH-specific receptors.
Understanding of the structure of PTH, either alone in solution or, perhaps more importantly, in direct association with its receptor(s), remains incomplete. Early studies using dark field electron microscopy or computational schemes suggested that PTH in aqueous solution consists of two (N- and C-terminal) linked globular domains (22, 23), whereas nuclear magnetic resonance (NMR) and circular dichroism analyses revealed no evidence of extensive secondary structure (24, 25). Considerable effort has been focused on the structures of peptides comprising the N-terminal region of PTH required for PTH1R activation. Thus, numerous solution-phase NMR studies support the presence of two 
-helical domains, especially within the regions PTH(3–11) and PTH(21–30), that are joined by a more flexible "hinge" region, a structure permissive of a large number of overall hormone conformations in solution (24, 26, 27, 28, 29, 30, 31, 32, 33, 34). A similar secondary structure, involving N- and C-terminal -helices, was shown for hPTHrP(1–34) in solution by two-dimensional NMR (35, 36). X-ray crystallography of hPTH(1–34), however, predicted the presence of a single linear -helix extending from Ser3 to Asn33 (37). Analysis of the impact of chemical substitutions that constrain peptide confirmation is also most consistent with the concept that PTH(1–19) binds to the PTH1R as an extended -helix (38, 39, 40). Solution phase two- and three-dimensional NMR analysis of the structure of prolylhPTH(1–84) showed no evidence of secondary structure unless the solvent hydrophobicity was increased by addition of trifluoroethanol (up to 70%) (41). Under these conditions, extensive -helicity was detected between Ser3 and Gly38, and this was strongest between Met18 and Gln29. Interestingly, the PTH(39–53) region, least conserved at the level of primary sequence (Fig. 1), appeared devoid of secondary structure, whereas evidence of limited structure, i.e., turns and short helices, was found within the C-terminal hPTH(54–84) region. Furthermore, nuclear Overhauser effect analysis confirmed evidence of secondary structure in regions 3–10, 17–27, 30–37, and 57–62, but not within region 40–52.
It is important to emphasize that these biophysical measurements have been performed under experimental conditions with uncertain analogy to the local environment in which the ligand-receptor interaction actually takes place, i.e., at the interface between the extracellular fluid compartment and the cell surface. Given this proviso, one can conclude from available data that regions of PTH, especially those most highly conserved genetically, exhibit some evidence of ordered structure(s) that may be relevant to receptor interaction. More definitive information must await further technical advances, including the possible crystal structure of the active hormone/receptor complex.
|
III. Classical Actions of PTH and the PTH/PTHrP Receptor |
|---|
Although the natural hormone is 84 amino acids in length, it was found that a synthetic N-terminal fragment, bovine (b)PTH(1–34), could reproduce the major biological actions attributed to full-length bPTH, including activation of adenylyl cyclase in bone and kidney cells, increased urinary excretion of cAMP and phosphate in rats, and elevation of blood calcium in rats, dogs, and chicks (7, 69, 70, 71, 72, 73). Moreover, it was found that synthetic carboxyl fragments such as PTH(44–68), PTH(53–84), and PTH(39–84) did not compete for binding with PTH(1–34) radioligands, nor did they activate adenylyl cyclase in renal membranes or bone cells (74, 75, 76). These observations, together with the practical difficulties at that time in reliably producing large quantities of chemically pure PTH(1–84), led to widespread use of synthetic PTH(1–34) as a surrogate for intact PTH in investigations of hormone action in vitro and in vivo. Detailed structure-function analysis of the PTH(1–34) ligand demonstrated the importance of its extreme N terminus (especially residues 1 and 2) for activation of adenylyl cyclase (7, 77, 78, 79) and of its C terminus (i.e., residues 15–34) for high-affinity receptor binding (80, 81, 82, 83, 84). Other binding determinants must exist within the N terminus of PTH(1–34), however, because N-truncated peptides such as PTH(3–34) and PTH(7–34) bind with affinities considerably lower than that of PTH(1–34) (74, 80, 85).
The abolition of bioactivity that accompanied progressive N-terminal truncation of PTH suggested a strategy for developing effective PTH antagonists, but initial efforts based on the use of PTH(3–34) analogs were thwarted when residual agonism, not readily apparent during in vitro analyses, was detected in vivo (86, 87). Additional truncation to analogs lacking up to six N-terminal amino acids allowed effective inhibition in vivo, as with [Tyr34]hPTH (7–34)NH2, although potency was limited by low binding affinity (88). Subsequent introduction of amino acid substitutions found to enhance binding of such shortened analogs has led to design of even more effective antagonists, such as [Nle8, D-Trp12,18, Tyr34]bPTH(7–34)NH2, [Leu11, D-Trp12]PTHrP(7–34)NH2, and [Ile5, Leu11, D-Trp12, Trp23]PTHrP(5–36)NH2 (77, 89, 90).
Use of cAMP generation as the exclusive measure of the intracellular action of PTH subsequently became untenable when studies of both bone- and renal-derived cells and tissues demonstrated that PTH(1–34), as well as PTH(1–84), could activate other signal transduction pathways, independently of adenylyl cyclase, including those involving phospholipase C (PLC), protein kinase C (PKC)(s), cytosolic free calcium (Ca2+), phospholipase D, and phospholipase A2 (91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110). PTH also regulates MAPKs, including p42/p44 ERKs, p38 and c-Jun N-terminal kinase subtypes, although the direction of this regulation and its mediation by more proximal effectors such as cAMP/PKA and PKC, especially in the case of p42/p44 ERKs, appears to depend on cell type and the concentration of PTH (111, 112, 113, 114, 115, 116, 117, 118, 119).
Such extensive signaling diversity initially raised the possibility that PTH might interact with more than one type of receptor in these target tissues, although subsequent studies with cloned PTH receptors (see Section III.B) showed that such multiple signaling, at least for adenylyl cyclase, PLC, cytosolic Ca2+, and PKC, can be mediated by a single receptor species. Of particular interest were observations that these cAMP-independent responses, notably PKC activation and elevation of cytosolic Ca2+, could be induced by N-truncated PTH fragments or analogs that were unable to effectively stimulate adenylyl cyclase (96, 103, 104, 105, 120, 121, 122, 123). In particular, a short sequence comprising residues 29–32 of hPTH was shown to be both necessary and sufficient for activation of membrane-associated PKC(s) in osteoblastic cells, whereas the fragment hPTH(1–31), lacking this domain, could not elicit this PKC response (122, 123). Such observations are of interest in the context of the present review for at least two reasons. First, they raise the possibility that biological effects reported for certain C-terminal PTH fragments, devoid of cAMP-stimulating activity but long enough to include the PTH(29–32) domain (see Section III.B), might involve activation of PKC via classical PTH1Rs, even if such fragments cannot be shown to compete effectively with N-terminal PTH radioligands for binding to target cells. Second, they could provide a mechanism to explain cAMP-independent biological actions (mitogenesis, regulation of creatine kinase, regulation of other genes) reported for certain midregional PTH fragments, such as PTH(25–39), PTH(28–47), PTH(28–48), and PTH(29–47) (124, 125, 126, 127, 128, 129, 130, 131). That these latter actions may reflect interaction of these midregional peptides with PTH1Rs is further supported by the observation that excess PTH(28–48) inhibits the cAMP response to PTH(1–84) (125) and that the PTH(28–42) and PTH(28–48) fragments can activate PKC in CHO cells expressing transfected PTH1R cDNA (132). On the other hand, the anabolic action of intermittently administered PTH via the PTH1R is not seen with intermittent administration of hPTH(28–48) (133). Thus, the in vitro mitogenic activity observed in response to this midregion fragment is not linked to stimulation of bone formation in vivo.
B. Identification, cloning, and signaling properties of the PTH/PTHrP receptor
The discovery that PTH elicits activation of adenylyl cyclase and production of cAMP predicted that the responsible receptor would be a member of the G protein-coupled receptor (GPCR) family. This concept was further supported by evidence that PTH action could be modulated by guanyl nucleotides (activation of adenylyl cyclase or PLC) or by pertussis toxin (99, 134, 135, 136, 137, 138). Direct radioligand binding analysis of the interaction of PTH with membranes or intact cells of skeletal or renal origin demonstrated saturable binding with an affinity constant in the low nanomolar range (74, 75, 137, 139, 140, 141, 142, 143, 144, 145), and chemical cross-linking studies indicated that the PTH receptor was likely to be a 60- to 80-kDa membrane glycoprotein (146, 147, 148, 149, 150, 151, 152). After the discovery of PTHrP as the cause of the humoral hypercalcemia of malignancy syndrome, the recognition of the high sequence homology of its N terminus with that of PTH, and the demonstration that the biological actions of N-terminal fragments of PTH and PTHrP were equivalent in many different bioassays (153, 154), it was shown that both PTH and PTHrP N-terminal fragments bind to the same receptor sites in kidney and bone cells (155, 156, 157).
cDNA encoding the PTH1R was successfully isolated in 1991 by using photoemulsion autoradiography to screen for binding of 125I-hPTHrP(1–36) radioligand by COS-7 cells that were transfected with pcDNA1 plasmid pools from an opossum kidney cell cDNA library (8). Subsequent comparison of cDNAs encoding the opossum, rat, human, and porcine PTH1Rs (8, 9, 158, 159) demonstrated that these specify highly homologous (80–95% amino acid identity) single-chain polypeptides approximately 590 amino acids in length, each featuring an extended N-terminal extracellular domain, the anticipated seven hydrophobic helical transmembrane domains (TMDs), and an intracellular cytoplasmic "tail" containing a number of serine residues that undergo phosphorylation upon agonist interaction (160, 161, 162, 163). The N-terminal domain is glycosylated at four asparagine residues clustered near the junction with the first TMD (164), and includes three disulfide bonds involving six highly conserved cysteines (165). The PTH1R is activated equivalently by intact and N-terminal PTH and PTHrP peptides and, like other members of the class II family of GPCRs (including secretin, calcitonin, vasoactive intestinal peptide, glucagon-like peptide-1, GHRH, corticotropin-releasing factor, and glucagon, among others), it is capable of coupling to several different G proteins, thereby activating multiple signaling pathways concurrently, including adenylyl cyclase/cAMP, PLC, cytoplasmic Ca2+, and PKC, when expressed in heterologous cell systems (9, 159, 166, 167, 168, 169, 170). The intracellular tail of the receptor seems also to be important for coupling to a pertussis toxin-sensitive Gi protein that inhibits adenylyl cyclase (171). The receptor tail recently was also shown to include a highly conserved PDZ interaction domain required for binding to the sodium/hydrogen exchanger regulator factor (NHERF) family of adapter/regulatory proteins, which may govern the balance between the PLC of the receptor vs. adenylyl cyclase signaling, at least in some cells (172).
The PTH1R is highly expressed in bone and kidney, but is found also in a variety of tissues not regarded as classical PTH target tissues (173, 174). This likely reflects the widespread local paracrine role of PTHrP postulated in tissues such as breast, skin, heart, blood vessels, pancreas, and others (154, 174). Ablation of the PTH1R gene in mice [and inactivating mutations in humans (175)] results in neonatal lethality and a severe defect in endochondral bone formation characterized by impaired proliferation and accelerated chondrocyte maturation and mineralization (176). Such mice have been "rescued" via chondrocyte-specific expression of constitutively active PTH1Rs (177), but these animals display other abnormalities in tooth development and bone, and detailed analysis of mineral ion homeostasis has not yet been possible. Because a phenotype similar to that of the receptor-null animals results from PTHrP gene ablation (178), the predominant endochondral defect likely reflects interruption of the critical local paracrine role of PTHrP in the growth plate (179). Ablation of the PTH gene (180, 181) does lead to hypocalcemia and hyperphosphatemia, whereas activating mutations in the PTH1R cause hypercalcemia and hypophosphatemia (182, 183), confirming the primary role of PTH per se, and of the PTH1R, in maintaining normal mineral ion homeostasis. Mice lacking the gene for PTH also exhibit abnormalities in mineralization and formation of primary spongiosa of long bones, which are not seen in PTHrP-null animals and are presumed to reflect loss of PTH-specific actions in bone, although the receptors involved have not been defined (180, 181).
The manner in which the PTH ligand interacts with the PTH1R has been deduced from an extensive series of studies by several groups involving mutagenesis of both the receptor and the ligand, use of hybrid receptors and ligands, and direct chemical cross-linking of ligands bearing photoreactive groups at specific locations within the peptide chain (84, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194). These analyses indicate that interaction of the PTH(1–34) ligand with the PTH1R involves high-affinity binding of the C terminus of the ligand with portions of the receptor’s extracellular N-terminal domain and the extracellular loops that connect the TMDs. The N terminus of the ligand then interacts with the TMD domains to catalyze the G protein activation(s) required for signal transduction (195, 196). The critical role in receptor activation of the juxtamembrane ("J") domain, comprising the seven TMDs and their connecting intra- and extracellular loops, is highlighted by the fact that the several mutations identified as causing constitutive (i.e., ligand-independent) receptor activation (Jansen’s metaphyseal chondrodysplasia) are located within the J domain (182). Moreover, N-terminal PTH peptides as short as PTH(1–14) or PTH(1–11) that incorporate substitutions designed to promote -helical structure, unlike PTH(1–34), can activate truncated receptors lacking the N-terminal extracellular domain as effectively as wild-type receptors (38, 197).
Studies of the signaling properties of clonal PTH1Rs expressed in heterologous cell lines, such as LLC-PK1, COS-7, or HEK 293 kidney cells or CHO cells, have shown that this single receptor type can mediate activation of adenylyl cyclase, PLC, phospholipase D, PKC, and MAPK and can increase the concentration of cytoplasmic Ca2+ (9, 166, 169, 198, 199, 200, 201). Thus, it is not necessary to invoke the existence of other types of PTH receptors to explain the diversity of signaling events that N-terminal fragments of PTH or PTHrP can elicit in various target cells, although this possibility has not been excluded completely. Mutational analysis has indicated that different G proteins likely do not interact with the PTH1R identically. For example, a clustered mutation (EKKY
DSEL) in the second intracellular loop of the rat receptor completely abrogates PLC signaling without affecting adenylyl cyclase activation (202), and mice expressing only this mutant receptor display subtle developmental defects in endochondral bone formation (203). Although it is clear that an intact N terminus of PTH or PTHrP is needed for effective adenylyl cyclase activation via PTH1Rs (76, 132, 169, 204, 205), this does not appear to be true for activation of PKC. Thus, as noted earlier, various N-truncated PTH peptides have been shown to activate PKC(s) in cells expressing endogenous or transfected recombinant PTH1Rs (96, 103, 104, 105, 120, 121, 122, 123, 132, 206). At the same time, studies of cells stably expressing the transfected PTH1R indicate that activation of PLC, which can lead to activation of PKC via generation of inositol trisphosphate and diacylglycerols, requires that the N terminus of the ligand be intact (205). Also, PTH(1–31), found to lack PKC activation in some systems (122, 123), can nevertheless activate PKC, presumably via PLC, in others (206, 207, 208). Although available data are not fully congruent, one interpretation is that PTH1Rs can activate PKC(s) via at least two different mechanisms, one of which involves PLC and requires that the ligand have an intact N terminus, whereas the other, a PLC-independent mechanism, is triggered by more carboxyl ligand determinants, such as the region PTH(29–32) [or PTHrP(25–34) (209)]. These mechanisms may not both be active in all target cells.
Like other GPCRs, activated PTH1Rs appear to be phosphorylated by specific GPCR kinases, which then facilitate association with ß-arrestin proteins (161, 163, 210, 211, 212, 213, 215). ß-Arrestins, in turn, may terminate receptor-G protein coupling and promote receptor endocytosis, although in the case of the PTH1R the specific role(s) of ß-arrestin in receptor endocytosis and either degradation or recycling back to the surface remains unsettled and may differ in cells of different types (161, 162, 163, 212, 213, 215, 216, 217). ß-Arrestins, when associated with certain GPCRs, also may support signaling functions, independent of classical G protein-mediated second-messenger generation, by serving as molecular scaffolds to assemble and activate kinases such as MAPKs and nonreceptor tyrosine kinases (218). It is not yet known whether such ß-arrestin-mediated signaling can occur via PTH1Rs and contribute, for example, to the activation of MAPK observed in some PTH target cells (112, 113, 114, 115, 117). Of particular interest are very recent observations that PTH1R internalization can be dissociated from receptor activation and that, in some cells, N-truncated PTH peptides such as PTH(7–34) and PTH(7–84) may promote PTH1R endocytosis via a ß-arrestin-independent, dynamin-dependent mechanism that is regulated (blocked) by interaction of the adapter protein NHERF1 with the cytoplasmic domain of the PTH1R (219). These novel findings raise the intriguing possibility that certain N-truncated PTH (CPTH) peptides, incapable of classical PTH1R second-messenger signaling but long enough to bind in some way to the PTH1R, could antagonize the actions of PTH1R agonists by inducing rapid receptor down-regulation, at least in cells lacking NHERF1. Although the role that this phenomenon plays in modulating PTH action in vivo remains to be determined, these observations suggest the possibility of a distinct cellular mechanism of action for at least some CPTH peptides, in addition to activation of CPTHRs.
C. Other members of the PTH/PTHrP receptor family
In 1995, Usdin et al. (220) reported that homology screening of a human brain cDNA library for other members of the class II GPCR family had led to isolation of a novel receptor, closely related to the PTH1R, which they named the "PTH2 receptor." Homologs of this receptor subsequently were identified in rat and zebrafish (221, 222). In rats, this receptor is expressed in discrete areas of the central nervous system, including hypothalamic, limbic, and sensory areas, especially in the spinal cord; parafollicular cells of the thyroid; peptide-secreting cells of the gastrointestinal tract; somatostatin-rich pancreatic islet cells; pancreatic exocrine cells; cardiac and vascular endothelium; vascular smooth muscle; lung; placenta; testis; and the vascular pole of the renal glomeruli but, unlike the PTH1R, not in renal tubules or bone (220, 223, 224). Pharmacologically, the PTH2 receptor (PTH2R) also differs strikingly from the PTH/PTHrP receptor, now referred to also as the type-1 PTH receptor (PTH1R), in that it is activated by PTH but not by PTHrP (90, 220, 221, 225, 226, 227, 228). This PTH selectivity mapped to differences at position 5 of the ligand (Ile5 in hPTH vs. His5 in hPTHrP) and position 23 (Trp23 in hPTH vs. Phe23 in hPTHrP), which affected activation and binding, respectively (225, 226). Thus when the PTH-specific residues at these two positions were substituted into hPTHrP(1–36), activity at the PTH2R was reconstituted. Photoaffinity cross-linking analyses suggest that the overall orientation of the ligand relative to the receptor protein is similar for PTH binding to PTH1Rs and PTH2Rs, although specific residues within the N terminus of the ligand may play different roles in activating one receptor vs. the other (90, 229).
Like the PTH1R, the PTH2R exhibits dual signaling in response to PTH(1–34), with generation of both cAMP and cytoplasmic Ca2+ transients (227, 230). PTH(1–34) is a relatively potent agonist for the hPTH2R, at least when this receptor is expressed at high levels in cultured cells, but this is not true of the rat PTH2R (221, 230). This suggested that PTH may not be the endogenous ligand for the PTH2R. Subsequent demonstration of a potent PTH2R-selective activating factor in bovine hypothalamic extracts (231) was followed by the isolation and identification of a 39-residue peptide termed tuberoinfundibular peptide of 39 residues, or TIP39, that shows limited amino acid sequence homology to bPTH and activates PTH2Rs but not PTH1Rs (232). Later isolation of human and mouse TIP39 genomic DNA and tissue expression analysis in the mouse confirmed that this is a secreted peptide that is highly expressed in testis and, at lower levels, in various central nervous system nuclei, liver, and kidney (233). On the basis of the localization of PTH2Rs and TIP39 in the central nervous system and recent neurobehavioral studies, it appears likely that one of the important actions of TIP39 is to facilitate the response to painful stimuli (234). The selectivity of PTH2Rs vs. PTH1Rs for TIP39 is dictated by interaction of the first six amino acids of TIP39 with the J domain of the PTH2R, because chimeric receptors consisting of the PTH2R J domain and the PTH1R N-terminal extracellular domain, but not the reciprocal chimera, mediate binding and activation by TIP(1–39), whereas an analog, TIP(7–39), binds poorly to the PTH2R but well to the PTH1R (235). In fact, TIP(7–39) and TIP(9–39) are highly effective PTH1R antagonists (233, 235, 236). These surprising results indicate that the C-terminal portion of TIP39 can bind well to the PTH1R, presumably via interaction with its extracellular domain (237), but that this affinity is overridden by a conformational incompatibility between the J domain of the PTH1R and the N-terminal six residues of TIP39.
A third type of PTH receptor, termed the "type-3 zPTH receptor" or "zPTH3R", was cloned from a zebrafish cDNA library (222). At the level of amino acid sequence, this receptor is more closely related to mammalian PTH1Rs than PTH2Rs yet clearly is different from the zebrafish PTH1R, which was isolated concurrently (222). When expressed in mammalian (COS-7) cells, the zPTH3R activates adenylyl cyclase but not PLC and exhibits 20-fold higher affinity and potency with hPTHrP(1–36) than hPTH(1–34). Subsequent analysis, however, indicated that rat PTH activates the zPTH3R with higher potency than PTHrP, suggesting that this receptor probably is not preferentially responsive to PTHrP peptides (238). The importance of the PTH3R to human physiology is uncertain, however, because no evidence has been produced to date for the existence of a mammalian homolog of this receptor.
Evidence exists for other types of receptors that recognize N-terminal peptides of PTH or PTHrP but are less well characterized. Thus, Orloff et al. (239) reported sensitive cytosolic Ca2+ responses to both hPTHrP(1–36) and hPTH(1–34) (EC50 = 50–80 pM), without corresponding activation of adenylyl cyclase, in a series of human squamous cell carcinoma and keratinocyte cell lines. Specific binding of radioiodinated [Tyr36]hPTHrP(1–36), competed similarly by hPTHrP(1–36), hPTHrP(1–74), hPTH(1–34), and bPTH(1–84) and of relatively low affinity (IC50 = 100–300 nM), was observed also, but only in some of the squamous cell lines. Efforts to identify receptor mRNA transcripts using PTH1R probes revealed atypical hybridizing bands but no clear evidence for PTH1R expression. Moreover, PTH1Rs transfected into squamous cells did elicit the expected cAMP response (240). Similar findings were obtained using a rat insulinoma cell line (241). Evidence of a brain receptor specific for PTHrP has been provided by Yamamoto et al. (242, 243), who observed that PTHrP(1–34), but not rat or human PTH(1–34), PTHrP(7–34), or hPTH(13–34), stimulated release of arginine vasopressin from slices of rat supraoptic nucleus. This effect was dose dependent (0.1–1000 nM) and associated with a modest increase in cAMP [not seen with PTH(1–34)]. Both responses were blocked by PTHrP(7–37). PTHrP, but not PTH, showed specific binding to membranes that was of relatively low affinity (IC50 = 100 nM) and was competed by PTHrP(7–37) but not by PTH. None of these responses are readily explained by the known properties of the cloned PTH receptors isolated to date.
|
IV. PTH Secretion and Metabolism |
|---|
A different question was raised by the detection, in 1973, of N-terminal fragments in the circulation that were biologically active, as determined by monitoring renal adenylate cyclase activity (252). This and other immunochemical evidence for such circulating, short-lived N-terminal fragments (249, 250, 253, 254) led to the hypothesis that secreted PTH might have to be cleaved into fragments in the periphery before its ultimate actions in target tissues. A similar conclusion had been reached by Parsons and Robinson (255) in their studies of perifused feline bone. This theory subsequently was disproven, however. Thus, Goltzman et al. (256) demonstrated that intact bPTH could activate adenylate cyclase in renal cortical membranes and fetal rabbit calvarial bone without prior cleavage to fragments (73). The same was later shown with respect to release of cAMP and osteocalcin by intact bPTH from isolated perfused rat hindquarters (257). Intact biologically active radioiodinated bPTH also was shown to bind to canine renal membranes (140) and rat osteosarcoma cells (258) without prior cleavage to fragments. Thus, although synthetic N-terminal fragments such as hPTH(1–34) are highly active at PTH1Rs, conversion of PTH to such smaller fragments is not required for activity, nor is there any direct evidence that such fragments are produced in vivo by metabolic cleavage or gland secretion (see Section IV.E.1). Thus, such N-terminal fragments are not physiological agonists of the PTH1R.
B. Sources of circulating C-terminal PTH fragments
1. Secretion by the parathyroid glands.
It has been firmly established that the parathyroids secrete C-terminal fragments of PTH as well as the intact hormone. The initial evidence for this came from use of assays specific for different regions of the hormone molecule to demonstrate C-terminal PTH fragments in venous effluent of parathyroid tumors (246, 259). Subsequently, Mayer et al. (260) determined by direct sampling of parathyroid venous blood of calves that, similar to the finding in patients with hyperparathyroidism, venous effluent from normal calves contained not only PTH, but also a major amount of C-terminal fragments. Indeed, these CPTH fragments were secreted in greater amounts than intact hormone, and the relative amounts of C-fragments increased with induced hypercalcemia (260). These results were consistent with earlier in vitro studies showing that most initially synthesized PTH is degraded intracellularly (261, 262), although neither PTH mRNA translation nor conversion of prePTH to proPTH is regulated by calcium (261, 263, 264, 265). Studies of cultured or perifused bovine parathyroid tissue in vitro also showed that CPTH fragments are present in parathyroid cells and are secreted directly from the glands (266, 267, 268, 269, 270, 271, 272). This was further supported by extensive chemical analysis of specific CPTH fragments produced by organ-cultured parathyroid tissue (273, 274). The ability of high extracellular calcium to augment release of CPTH fragments relative to that of intact PTH also has been repeatedly confirmed (272, 273, 275, 276, 277, 278).
2. Hepatic proteolysis of intact PTH.
Early work by Fang and Tashjian (279) showed that the liver contributes substantially to the clearance of circulating intact PTH, and this was confirmed by numerous subsequent analyses in several species, including humans (280, 281, 282, 283, 284, 285). The liver can extract biologically active PTH(1–34) as well (286, 287, 288), when this peptide is administered exogenously. PTH has been shown to activate adenylate cyclase in liver (286, 289, 290), but direct hepatic actions of PTH are not known to be involved in systemic calcium homeostasis. Such hepatic actions of PTH, presumably mediated by the PTH1R, could reflect physiological autocrine actions of PTHrP. A more likely hepatic role in calcium and bone homeostasis, for which there now is considerable supporting evidence, relates to the uptake of intact PTH and its proteolysis by Kupffer cells to generate various circulating CPTH fragments (284, 291, 292, 293). Hepatic production of CPTH fragments (282) requires initial uptake of intact hormone by a mechanism that recognizes determinants present within PTH(28–48) but not PTH(1–34) (294). This fits with autoradiographic studies showing that intact PTH(1–84) binds to Kupffer cells, whereas PTH(1–34) does not (295). On the other hand, both PTH(1–84) and PTH(1–34) do bind to hepatocytes and sinusoidal cells (295). Some of the CPTH fragments generated by Kupffer cells are released back into the bloodstream, where they are not subject to further hepatic clearance (294) but rather are removed mainly by the kidneys (282) (see Section IV.C). Studies in hepatectomized vs. nephrectomized rats showed that the liver is the principal source of circulating CPTH fragments that result from peripheral metabolism of intact PTH (296).
Canterbury et al. (297) were the first to study liver metabolism of intact PTH in perfused rat liver, and they detected generation of N-terminal PTH fragments by using RIA after Biogel P-10 gel filtration. A fragment peak with a molecular weight of approximately 3500 Da was detected that was relatively enriched in N-terminal PTH immunoreactivity and could activate adenylate cyclase, similar to eluted peaks of intact bPTH or bPTH(1–34) (297). Subsequently, however, Daugaard et al. (298, 299) found in similar experiments that only C-terminal, biologically inactive fragments were generated during liver perfusion, as analyzed using HPLC fractionation. Daugaard suggested that differences in the purity of hormone preparations and the methods for analysis of fragments may have been responsible for the discrepant results. As discussed further below (Section IV.E.1), available evidence indicates that the N-terminal portion of the hormone is degraded locally by Kupffer cells and that N-PTH fragments do not reemerge from the liver into the circulation (292, 300, 301).
C. Renal clearance of PTH and PTH fragments
PTH immunoreactivity disappears more slowly from blood in humans with renal insufficiency (244, 302, 303, 304) and in nephrectomized animals (282, 284, 305, 306), and direct analysis of the fate of radiolabeled or immunoreactive PTH has confirmed the crucial role of the kidneys in clearance from blood of both intact hormone and, especially, CPTH fragments (248, 249, 282, 302, 306, 307, 308, 309, 310). A portion of intact PTH is cleared from blood by renal mechanisms that do not involve glomerular filtration, termed "peritubular uptake" by Martin et al. (306). This route is selective for PTH or PTH fragments capable of binding to the PTH1R (306, 311) and may involve receptor-triggered endocytosis at the basolateral surfaces of renal epithelial cells (145). The bulk of the hormone, however, is cleared by glomerular filtration and then is actively reabsorbed by the tubules (298, 306, 308). Recent studies implicate the megalin/cubilin endocytic system in the apical renal tubular epithelial clearance of PTH from tubular urine (312); this system is an important renal tubular scavenger receptor mechanism responsible for limiting renal clearance of low-molecular weight proteins. This PTH1R-independent mechanism is consistent with observations that overall clearance of biologically active PTH preparations does not differ from that of biologically inactive preparations (313, 314).
Although the involvement of the kidneys in clearance of PTH from the circulation is unequivocal and explains the disproportionate elevation of CPTH fragments observed in renal failure, it is unlikely that the kidneys are an important source of circulating CPTH fragments. Analysis of PTH metabolism by isolated perfused kidneys has produced conflicting results on this point (298, 299, 309, 315), but studies of acutely hepatectomized vs. nephrectomized rats demonstrated that the liver, not the kidneys, is the principal source of CPTH fragments in blood (296, 300).
Thus, both the liver and kidneys participate in clearance of circulating intact PTH, but only the liver generates, via the action of endopeptidases expressed by Kupffer cells, CPTH fragments that can then reenter the circulation (without accompanying N-PTH fragments). These CPTH fragments, like those secreted from the parathyroid glands, then undergo predominantly renal clearance.
D. Regulation of circulating C-terminal PTH fragment concentrations by serum calcium
By 1979 it was known from experiments in animals and humans that the relative contributions of intact PTH and hormonal fragments to PTH immunoreactivity in blood are regulated by calcium and influenced by renal failure (260, 304). Mayer et al. (260) provided the most definitive early data regarding the effect of blood calcium on secretion of PTH fragments by studying catheterized calves, where parathyroid effluent could be directly sampled under highly controlled conditions. These workers showed that C-terminal fragment secretion was responsible for the major portion of PTH immunoreactivity in hypercalcemia, whereas intact hormone was by far the major species of glandular output in hypocalcemia. As already noted above, this was subsequently confirmed by direct in vitro analysis of PTH peptide content and secretion from cultured or perifused parathyroid tissue (272, 273, 275, 276, 278).
At the same time, immunochemical analysis of circulating PTH in humans also documented that the ratio of CPTH fragments to intact hormone in peripheral blood is directly related to the calcium concentration (304, 316). Subsequent detailed analyses, using region-specific antibodies, have been carried out for the most part by the laboratory of D’Amour (317, 318, 319, 320, 321, 322, 323). In studies of normal human subjects in which parathyroid function was stimulated acutely by EDTA-induced hypocalcemia or suppressed by calcium infusion, these investigators, using RIAs specific for intact hormone vs. mid- or late-carboxyl PTH fragments, found that the regulation of intact hormone and hormonal fragments during hypo- and hypercalcemia differed (317, 318, 319). During acute hypocalcemia, intact PTH increased in serum 5- to 6-fold, and mid- and late-CPTH to a lesser extent, but these same CPTH fragments remained the predominant forms of PTH in blood. In response to acute hypercalcemia, intact PTH was suppressed 4- to 5-fold, whereas mid- and late-CPTH fragments declined only 30–50%, such that the latter became even more predominant relative to intact PTH (i.e., 10-fold or higher in relative molar concentrations). Directionally similar calcium-dependent changes in ratios of mid- and late-CPTH fragments to intact hormone were observed in patients with primary hyperparathyroidism, but the apparent set point for calcium regulation was higher, such that higher serum calcium levels were required than in normal subjects to achieve the same ratios of CPTH/intact PTH (320). One notable aspect of this work was the finding that the ratio of mid-CPTH fragments to late-CPTH fragments was directly related to blood calcium concentration (318). This suggests that calcium may regulate the pattern of PTH proteolysis to produce relatively more CPTH fragments that are truncated at both their N and C termini. Experiments performed in dogs, designed to apply chronic stimulation or suppression of parathyroid function, have shown that CPTH fragments are more readily generated during hypercalcemic challenge after the glands have adapted to a chronic suppressive influence [such as 1,25-(OH)2D3 administration] (321, 322). Conversely, CPTH fragments are less easily produced, following the same acute hypocalcemic challenge, after a prolonged interval of parathyroid stimulation (as by calcium and vitamin D deficiency or after partial parathyroidectomy), consistent with time-dependent adaptation of intraparathyroidal peptidase activity (322, 323). In all of these investigations, a component of nonsuppressible intact PTH was noted during hypercalcemia, although it now is clear that immunoassays previously thought to be specific for the intact hormone may also detect long CPTH fragments [such as PTH(7–84)] (see Section IV.E). Thus, calcium-dependent excursions in ratios of secreted CPTH fragments:intact PTH may be even greater than suggested by these studies.
It seems clear that the secretion of CPTH fragments relative to intact PTH is regulated positively by extracellular calcium and that this likely contributes to the altered patterns of immunoreactive PTH peptides present in peripheral blood at different levels of blood calcium. However, the possibility that peripheral metabolism of PTH also may be regulated by changes in blood calcium remains unsettled. Work in rats and dogs has shown that the overall clearance rate of exogenously administered PTH is not affected by blood calcium (301, 314, 324, 325). Similarly, Oldham et al. (283) found no relation between serum calcium and the transhepatic arteriovenous gradient of PTH immunoreactivity in a small group of patients with primary hyperparathyroidism, although there was evidence of a calcium-dependent increase in renal extraction. Daugaard et al. (326), on the other hand, reported that hepatic extraction of intact PTH was accelerated 60% by increasing calcium concentration in the perfused rat liver system, but there was no evidence of a change in the efficiency of proteolysis to CPTH fragments. Earlier experiments performed by Canterbury et al. (297), in contrast, had indicated that the rate of cleavage of PTH by perfused rat livers was accelerated at low perfusate calcium concentrations. In intact anesthetized dogs, D’Amour et al. (327) found evidence of hepatic extraction of CPTH fragments that was suppressed by induced hypercalcemia. These authors concluded, however, that the elevated ratio of CPTH to intact PTH seen during hypercalcemia was due mainly to corresponding differences in rates of secretion rather than differences in metabolic clearance. With respect to renal clearance and proteolysis of PTH, Daugaard et al. (326) observed no relation between perfusate calcium and extraction of PTH by isolated perfused rat kidneys, nor were any CPTH fragments delivered into the perfusate. Hruska et al. (309), in contrast, found increased fragment production at low calcium in perfused canine kidneys. As noted by Daugaard, however, the latter observations may have been due to decreased glomerular filtration rates at higher calcium levels. Regulation of renal PTH proteolysis is of uncertain significance, however, because renal metabolism of PTH does not contribute significantly to the circulating pool of CPTH fragments (296, 300), which are derived principally from hepatic cleavage and parathyroid secretion.
In summary, whereas total immunoreactive PTH concentrations decline during hypercalcemia, the ratio of CPTH fragments to PTH in blood is positively correlated with serum calcium. This is associated with a calcium-dependent increase in secretion by the parathyroid glands of CPTH fragments relative to intact PTH. Although the overall metabolic clearance rate of PTH is not altered by changes in blood calcium, the possibility that calcium may regulate the rate or nature of hormonal proteolysis in the liver, and thereby contribute to an altered pattern of circulating PTH fragments, has not been excluded.
E. Nature of PTH fragments in blood
1. Circulating forms of N-terminal PTH.
The almost exclusive form of circulating PTH capable of activating PTH1Rs and thereby exerting the classical actions on calcium homeostasis in kidney and bone is the intact hormone, as shown by experiments in humans, rat, and bovine species. Some studies, involving the use of region-specific immunoassays, chromatography of serum, or both, have pointed to the presence of low levels of circulating N-terminal PTH fragments, mainly in subjects with hyperparathyroidism, renal insufficiency, or both (249, 250, 252, 253, 254, 304, 305, 328, 329). Small amounts of immunoreactive or bioactive PTH N-fragments have been reported also in parathyroid venous effluent or perfusate (246, 259, 260, 268, 328), again mainly with adenomatous or hyperplastic parathyroid tissue. Others, however, using direct chemical or radiochemical methods, have observed no secretion of N-terminal PTH fragments from parathyroid tissue (273, 274). All of these analyses have been plagued, to a greater or lesser extent, by issues of assay sensitivity and specificity, and in most cases the possibility of postcollection proteolysis ex vivo was not rigorously excluded. Thus, small PTH fragments reported in one study were observed also in hypoparathyroid serum in which intact PTH was entirely absent (304). In another well-controlled study involving use of a renal cytochemical bioassay and an immunoassay with predominant (but not exclusive) N-terminal specificity, low-molecular weight bioactivity was detected in uremic (but not normal) plasma and in parathyroid venous effluent, but there was no coeluting PTH immunoreactivity (328). The latter study did indicate that in normal subjects, at least 85% of plasma bioactivity coeluted with intact PTH.
The possibility that circulating bioactive N-terminal PTH fragments might result from postsecretory cleavage in peripheral tissues was raised by studies showing production of such fragments by perfused liver (297) or kidney (309), although others have not successfully replicated these results (299, 326, 330). Using bPTH(1–84) labeled to high specific activity at N-terminal methionines (positions 8 and 18) and HPLC resolution of radioactive peptides, Bringhurst et al. (292) showed that N-terminal fragments are produced during the endopeptidic cleavage of PTH by isolated rat Kupffer cells, but, unlike the corresponding CPTH fragments, are rapidly degraded. Moreover, subsequent studies in vivo using the same tracer, administered by continuous iv infusion to steady-state plasma concentrations, showed no accumulation of 35S-labeled N-terminal PTH fragments in blood of normal, nephrectomized, hepatectomized, nephrectomized/hepatectomized, thyroparathyroidectomized/hypocalcemic or vitamin D-intoxicated/hypercalcemic rats, or in rats chronically maintained on either low- or high-calcium diets, under circumstances where the limit of detection of such fragments was 0.1 pM (300). These authors concluded that peripheral metabolism of PTH does not result in formation of measurable quantities of circulating PTH N-fragments under physiological or pathological circumstances. Thus, although yet to be proven definitively, it seems likely that, under physiological conditions, the holohormone, PTH(1–84), is the only circulating form of PTH with an intact N terminus and PTH1R bioactivity, as assessed by adenylate cyclase or cytochemical bioassay, or by known structural PTH sequences that interact with the N-terminal receptor. In particular, there is no direct evidence that N-fragments such as the PTH(1–34) peptide, widely used as a laboratory agonist for the PTH1R and as a pharmaceutical therapy for osteoporosis, exist naturally in vivo. In renal failure, especially with concomitant hyperparathyroidism, it is possible that low levels of N-terminal PTH fragments are produced and persist in the circulation in association with delayed renal clearance of hormone.
2. Circulating C-terminal fragments of PTH.
As already reviewed, numerous analyses of human, rat, bovine, and porcine plasma have indicated that relatively high concentrations of circulating heterogenous C-terminal fragments exist under steady-state conditions. These fragments do not interact with the PTH1R and are therefore inactive in classical terms. It is likely that many of these fragments do possess biological activity, however, as discussed below.
Most available data concerning the nature of circulating CPTH fragments have been obtained using region-specific immunoassays (with or without preliminary gel-filtration chromatography) for which the reactive epitope(s) within the PTH molecule have been only crudely characterized. Efforts to more precisely define the structures of CPTH fragments in blood were initiated by Segre et al. (331, 332), who administered bPTH(1–84), radioiodinated at Tyr43, to dogs and rats, isolated the resulting large radiolabeled CPTH fragments by gel filtration of plasma, and subjected these fractions to automated sequential Edman degradation to define the N termini of these CPTH peptides. These experiments demonstrated the presence in blood and liver (284) of multiple CPTH fragments, initially corresponding to those with N termini at positions 34 and 37 and followed by the appearance of additional fragments with N termini at positions 38, 40, and 43 of the bPTH sequence. Using organ ablation, this group also showed that these fragments arose exclusively from the liver and not from the kidneys, the other major site of clearance of PTH (296, 300). Although the 125I-bPTH used by Segre et al. was biologically inactive (due to oxidation of N-terminal methionines during the iodination reaction), subsequent studies, using biologically active [3H-Tyr43]bPTH(1–84), documented major large circulating CPTH fragments with N termini at positions 34, 37, 41, and 43, with small amounts of additional fragments ending at positions 35 and 38 (300). Analogous fragments were found in rat liver extracts and could be produced by incubation of 125I-bPTH or [3H-Tyr43]bPTH with isolated hepatic Kupffer cells in vitro (284, 291, 292). It was clear from the studies with Kupffer cells that these N termini corresponded in most cases to more than one peptide structure, because, for example, peptides with the same N termini (at positions 34, 35, 37, 38, and 41) could be found in different, widely separated HPLC fractions, the most likely explanation for which would be CPTH fragments with the same N termini but with different C termini (300). Other workers, using biologically active 125I-bPTH and Kupffer cells prepared by different methods, identified the major CPTH fragments generated as having N termini at positions 35 and 38 (293), fragments which they also showed could be generated by incubation of the labeled PTH with purified cathepsin D (333). Clearance of CPTH fragments occurs mainly via glomerular filtration, as first inferred from very early studies (244) and later shown using purified 125I-CPTH fragments (310). The clearance of C-terminal fragments from plasma has been studied in the rat (310). In normal rats radioiodinated C-terminal fragments were extracted by kidneys (33%), muscle (16%), bone (7%), liver (<3%), and other tissues (<1%). In nephrectomized rats, 25% of C-terminal fragments were cleared in muscle, 10% in bone, and 7% in liver, with less than 1% in other tissues. Thus, nonrenal tissues can increase their ability to remove C-terminal PTH fragments in renal failure (310).
Analogous studies of the chemical nature of CPTH peptides secreted by porcine, bovine, or human parathyroid tissue have identified, remarkably, very similar N termini to those produced by hepatic proteolysis of intact PTH. Thus, Morrissey et al. (273), using a combination of microsequencing of radiolabeled peptides and tryptic peptide analysis, identified porcine (p)PTH(34–84) and pPTH(37–84) as secretory products of cultured porcine parathyroid cells, with pPTH(37–84) as the major moiety (2:1 molar ratio). Using N-terminal radiosequencing, MacGregor et al. (274) subsequently reported production by cultured bovine parathyroid cells of CPTH fragments having N termini at positions 24, 28, 34, 37, and 43 of bPTH, of which the putative cleavage at 23–24 was the earliest observed (within minutes) and the least suppressed by high medium calcium concentrations. The same group also reported secretion by human parathyroid cells of CPTH fragments with N termini at positions 24, 28, and 34 (334).
Thus, work performed more than 10 yr ago had identified the N termini of the principal CPTH fragments secreted by the parathyroid glands and shown that they were very similar, if not identical, to those generated by hepatic metabolism of circulating intact hormone and released back into the blood. It is important to recognize, however, that the precise structures of all circulating CPTH fragments have not yet been fully defined in any species. Differences in immunoreactivity registered in assays with predominantly mid- or late-carboxyl reactivity strongly suggest the presence of multiple CPTH fragments, some of which may extend to, or close to, the C terminus of intact PTH, whereas others may have undergone substantial C-terminal cleavage to produce bitruncated "mid-carboxyl" fragments (270, 318). Although most of the CPTH fragments characterized to date are at least 50–70% as large as PTH(1–84), some small "late carboxyl-terminal" fragments also may exist in human plasma (335).
More recently, strong evidence emerged for the presence in human plasma of N-truncated CPTH fragments that are long enough to register in conventional two-site immunoassays for intact PTH but that lack the N-terminal serine necessary for both full bioactivity at PTH1Rs and reactivity in a novel two-site immunoassay that stringently requires an intact N terminus (336, 337, 338). These fragments likely were not detected in previous studies because of poor chromatographic resolution from intact PTH and the limited sensitivity of repetitive Edman degradation when the N terminus is far removed from the radiolabel being monitored. They accumulate disproportionately to intact PTH in renal failure where they may constitute up to 50% or more of total intact PTH immunoreactivity, vs. 15–20% in normal subjects (336). Neither the sizes nor precise structures of these fragments have yet been ascertained directly. They do elute from reverse-phase HPLC columns just before PTH, in the same position as hPTH(7–84), which has been used to model their possible biological properties (see Section VII.D). It is known that these extended CPTH fragments can arise both from peripheral metabolism of iv administered hPTH(1–84) in rats and via secretion from human parathyroid adenomas in vitro (339), properties they share with other CPTH fragments previously described. The ability to measure intact PTH separately from these long CPTH fragments that interfere in conventional two-site assays may be of increased clinical value in assessing parathyroid and bone status in patients with renal failure (340, 341).
|
V. Nonclassical Actions of PTH |
|---|
However, in addition to the classical actions of PTH on kidney and bone, PTH may also act, at least in pathological or pharmacological situations, to
Top
Abstract
I. Introduction