Introduction

Parathyroid hormone-related protein (PTHrP) was discovered in the search to identify mechanisms by which certain solid cancers cause hypercalcemia without metastasizing to bone [1, 2]. This syndrome had for a long time been ascribed to inappropriate production by tumors of parathyroid hormone (PTH), a concept that arose from a suggestion by Albright [3] in discussing a patient with renal cancer and elevated calcium and lowered phosphorus. Since this patient had no apparent parathyroid cancer, Albright correctly theorized that the high plasma calcium and low phosphorus were the result of tumor elaboration of a PTH-like substance. This novel idea appeared to be supported two decades later with the development of the first PTH radioimmunoassay that identified elevated plasma PTH in patients with bronchogenic carcinoma [4]. The concept of “ectopic PTH production” by cancers was accepted through the 1960s (e.g., [5]), but the idea began to lose favor by the 1970s.

Some doubt about this view began to emerge when more specific radioimmunoassays were developed. Two groups independently identified that the circulating PTH in hypercalcemic cancer patients was immunologically distinguishable from human PTH [6–8] and that the plasma levels were lower than in patients with primary hyperparathyroidism. In addition, breast tumor extracts from a hypercalcemic patient contained activity that reacted with a specific PTH antiserum, but which was non-parallel to normal PTH standards [9], suggesting a related, but distinct antigen. When several different anti-PTH sera were used that recognized various epitopes, PTH could not be identified in a series of plasma and tumor samples from patients with hypercalcemic cancers [10]. Thus, it was concluded that the cancer-associated hypercalcemia was caused not by PTH, but by a different and distinct entity with many PTH-like effects, giving rise to the term “humoral hypercalcemia of malignancy” (HHM) [11]. Affirmation of this came from a number of excellent clinical studies showing that patients with hypercalcemia in cancers that had not metastasized had low plasma phosphate levels, increased nephrogenous cyclic AMP and suppressed PTH [12–14].

It had become clear that certain crucial biochemical actions were shared by PTH and the unknown related substance. These included promotion of bone resorption sufficient to cause hypercalcemia, as well as promotion of the renal excretion of phosphorus and of the generation of nephrogenous cyclic AMP. Based on the latter response, rapid, sensitive, robust biological assays of PTH activity had been developed [15–17]. These were applied to the identification of PTH-like activity in tumor extracts from HHM patients [18], in tissue culture medium [19] and in animal models of HHM [20]. A sensitive cytochemical assay was also used to show elevated levels of PTH-like activity in the plasma of HHM patients that was distinct from PTH [21].

Identification of PTHrP and Its Gene

Driven by the clinical studies described above, and the availability of biological assays with high sensitivity and specificity, and capable of high throughput, the efforts to isolate and clone of the PTH-like activity followed. Certain criteria were applied to identification of PTH-like activity in tumor extracts and media. These were (1) it should promote adenylyl cyclase activity strictly in PTH targets, (2) that action should be inhibited by peptide receptor antagonists of PTH–receptor interaction, but not by anti-PTH neutralizing antiserum, and (3) PTH should be undetectable by radioimmunoassay in extracts or media. These stringent requirements were met in several early studies of PTH-related activity identified in tumor extracts from patients and animals with HHM [18, 20, 21] and in culture media from one such tumor line [19].

Using a cell line from a patient with HHM in association with a squamous cell carcinoma of the lung, PTHrP was purified and sequenced through the first 16 amino acid residues, revealing that 8 of the first 13 residues were identical with those in PTH, but no PTH was expressed by the cells. Anti-peptide antibodies based on this short sequence were able to detect PTHrP but did not cross-react with synthetic PTH(1–34) [1]. Further purification allowed sequencing of the first 50 residues to be obtained [2]. Purification was also achieved from a breast cancer extract [22] and a renal cancer cell line [23]. Cloning revealed a pre-pro-peptide of 36 amino acids and a mature protein of 141 amino acids with significant homology with PTH about the N-terminal region [2, 24], but no more identities within the remainder of the sequence other than those expected by chance. The biological activity of PTHrP at the PTH receptor was contained within the first 34 residues, and study of the efficacies of several truncated peptides showed that the requirements for biological activity within the amino-terminal region of this new protein were very similar to those within PTH [25].

Data rapidly accumulated to show that PTHrP was expressed by squamous cell cancers regardless of origin, as well as in a number of other tumors, including renal cortical carcinoma, breast cancer and neuroendocrine tumors. The part played by PTHrP in these cancers and in bone metastases will be discussed by Guise and co-authors in this volume. PTHrP was identified by immunohistochemistry in several non-malignant tissues, including keratinocytes, endothelial cells, smooth muscle, bone, lactating mammary tissue and neuroendocrine tissues [26], as well as in mammalian embryonic tissues [27].

PTHrP as a Multifunctional Cytokine

Although PTHrP was discovered as a hormone produced in excess by certain solid tumors, its physiological functions result predominantly from its action as a paracrine, and perhaps as an autocrine or even intracrine manner in multiple tissues. Only in three circumstances has PTHrP been shown convincingly to be measurable present in the circulation and acting in an endocrine fashion. These are (1) the HHM syndrome, in which PTHrP is secreted by tumors and acting systemically, (2) lactation, in which PTHrP is made in the breast and reaches the circulation, and (3) fetal life, where PTHrP regulates maternal-to-fetal placental calcium transport. Since PTHLH mRNA is highly unstable and its expression regulated by multiple growth factors and cytokines with which it is often co-expressed (during development and in various tissues), it rapidly became regarded as a cytokine.

Other chapters will engage in detail with the local roles of PTHrP in many tissues, including bone and cartilage, vascular smooth muscle, placenta, breast development and pancreatic islets. Introductory comments are provided here to consider how PTHrP might act locally in mammalian tissues, before discussing its identification and possible roles in non-mammalian species.

With PTH as a circulating hormone, and since the parathyroid glands only develop after mesenchymal condensations are formed at sites of skeletal development, it is the paracrine/autocrine factor PTHrP, secreted locally by chondrocytic cells, that plays the prime role in endochondral bone formation. Deletion of Pthlh in mice resulted in perinatal death from respiratory failure due to defective rib development [28, 29]. Rescue of these mice by directing PTHrP production to cartilage using the collagen II promoter allowed study of the PTHrP-null phenotype in several other organs. Thus, the central role of PTHrP in endochondral bone formation was discovered [30, 31], with PTHrP exerting a paracrine role, produced by the distal perichondrium and interacting with PTH/PTHrP receptor (PTH1R) in the proliferative and hypertrophic zones of the growth plate, revealing it to have a major developmental role.

The predominant role of PTHrP in bone metabolism was demonstrated in mice deficient in PTHrP. Despite the neonatal lethality of Pthlh (−/−) mice [29], haploinsufficient Pthlh (±) mice, that are phenotypically normal at birth, by 3 months of age have low bone mass, with a marked decrease in trabecular thickness and connectivity, and an abnormally high number of adipocytes in the bone marrow compared to wild-type littermates [32]. The critical role of osteoblast-derived PTHrP in the process of bone formation was confirmed when this phenotype was recapitulated in transgenic mice with osteoblast-specific deletion of Pthlh [33]. Importantly, mice with Pthlh-deficient osteoblasts were normocalcemic, emphasizing the fact that PTHrP action in bone is not required for the maintenance of calcium homeostasis.

One of the most clearly defined physiological actions of PTHrP has been its role as a local paracrine regulator of smooth muscle tone. It had long been recognized that injection of PTH (or parathyroid extract in the early years) rapidly increased blood flow through a range of vascular beds, accompanied by decreased blood pressure [34–36] (reviewed in [37]). PTHrP was found to be produced in vascular smooth muscle cells and to be rapidly up-regulated in response to mechanical distension of the arterial wall, to increases in arterial pressure, and to vasoconstrictors such as angiotensin II [38, 39]. It interacts with the PTH1R, acting as a potent relaxing agent for cardiac and smooth muscle from many locations such as blood vessels, uterus and the gastrointestinal tract [37, 40, 41]. Thus, the pharmacological effects previously observed with PTH resulted from widespread activation by PTHrP of PTH1R receptors in vascular beds [42, 43].

Although there is ample evidence for the cardiac, blood vessel and other smooth muscle actions of PTHrP, the Pthlh-null mutant mice develop normal cardiovascular systems [28], indicating that PTHrP is not essential for cardiovascular development. However, because the animals die at birth of skeletal and perhaps other defects, the roles of PTHrP in vascular remodeling, local regulation of vascular flow and the vascular response to injury in adult animals remain undefined.

Nuclear Import and Intracrine Function of PTHrP

PTHrP appears so far to be the only protein classed at least in some circumstances as a hormone, which possesses a candidate nucleus (CcN) motif and displays differential cellular localization (nuclear/nucleolar versus cytoplasm). The functions served by nuclear/nucleolar localization of PTHrP remain obscure, but it is such a striking property that it needs to be borne in mind in considering what roles PTHrP might play as an autocrine or intracrine mediator. This might become all the more important in investigating in due course how PTHrP functions physiologically in fish and in even more primitive species, where it seems likely that it will exert its action more as a local than as a humoral mediator.

Nuclear location of immunoreactive PTHrP was observed in 5–10 % of transiently transfected smooth muscle cells [44], and in chondrocytes, its nuclear localization was associated with enhanced chondrocyte survival following prolonged periods of serum starvation [45]. The localization was associated with a highly basic region within PTHrP (67–93). Studies in both smooth muscle cells [46] and keratinocytes [47] have demonstrated PTHrP expression to be cell-cycle-dependent. The highest levels of PTHrP mRNA appeared to be a response to mitogenic factors only at the G1 phase of the cell cycle during which PTHrP localized to the nucleolus. PTHrP also localizes to nucleoli where its role is still not understood [48]. This is likely mediated by a nucleolar localizing sequence from (94–108), immediately after the NLS [44]. Nucleolar localization has been shown to be cell type and cell cycle dependent, and may be a component of cell cycle control. For instance, it is increased in response to smooth muscle cell proliferation [49].

Within the PTHrP sequence, there are CK2-cdc2-NLS (CcN) and nucleolus localization motifs, with the former being similar to that described for the archetypal CcN-containing protein, SV40 T-antigen [50], comprising also consensus protein kinase CK2 and confirmed cyclin-dependent kinase phosphorylation sites. This similarity in structure led to experiments showing specific import of PTHrP into the nucleus, brought about by binding to importin β, which together with the GTP-binding protein Ran is able to mediate efficient nuclear accumulation in the absence of importin α, whereas addition of nuclear transport factor 2 (NTF2) reduced transport [51]. In support of these observations, the crystal structure showing importin β bound to the nuclear localizing sequence of PTHrP has been determined [52]. This novel nuclear import mechanism for PTHrP may be the result of either an intracrine pathway in a cell expressing PTHrP, or in a target cell as a result of binding of PTHrP to the PTH1R and internalization of the complex, leading to nuclear import; there is evidence supporting each of these potential pathways.

First, the idea that intracrine action is a feasible prospect comes from studies showing altered translation initiated from sites downstream of the initiator methionine of PTHrP [48]. This would give rise to forms that bind weakly or not at all to endoplasmic reticulum, and thus be available for transport to the nucleus [48, 49, 53]. The ability of PTHLH to be translated from start sites other than AUG codons was the first to be described in eukaryotes.

Second, internalization of PTHrP (1–108) using a fluorescein isothiocyanate label could be observed only in cells expressing the PTH1R [54]. Moreover, transport of PTHrP(1–141) to the nucleus after binding to the cell surface required an intact NLS [44]. The latter effect was abrogated by pre-treatment with PTHrP(74–113) alone, indicating the involvement of another cell surface receptor in endocytosis-dependent nucleolar translocation. Such receptor-dependent nuclear translocation of PTHrP, similar to mechanisms operating with interleukin-5 and growth hormone [55], as well as vascular endothelial growth factor [56], implied that the PTH1R should be found in the nucleus. There is some evidence for that, with the PTH1R being found by immunostaining in the nucleus of cell lines and rat tissues [57, 58]. Furthermore, the cytoplasmic tail of the PTH1R at residues (471–487) has a putative NLS similar to that of PTHrP [58].

PTHrP Post-translational Processing: Active Domains

In addition to the complex transcriptional control of PTHLH gene expression, the individual isoforms themselves are subject to a variety of critical post-translational modifications and enzymatic processing. The first 36 amino acids (−36 to −1) code for the intracellular “prepro” and “pro” precursors of the mature peptide which are important for intracellular trafficking and secretion of polypeptide hormones [2]. The 14–36 region of PTHrP which has almost no homology with the primary amino acid sequence of PTH (Fig. 1), nevertheless appears to be critical for the binding of PTHrP to the PTH1R receptor and for its subsequent activation [25, 59].

Fig. 1
figure 1

A comparison of the gene structure of elephant shark, Fugu, chicken, mouse, rat and human PTHLH/pthrp and elephant shark, Fugu, chicken, mouse, rat and human PTH/pth (adapted from Liu et al. [104])

Full size image

The importance of the next region extending from amino acids 36–139 (Fig. 1) is less well established. This region is encoded by all three isoforms of human PTHrP mRNA, and the primary sequence is highly conserved across species through to amino acid 111, indicating that important functions are potentially contained within this part of the molecule, though few have been definitively and repeatedly demonstrated. The “mid-molecule” portion, between residues 35–84, is responsible for promoting calcium transport across the placenta, making calcium available for fetal skeletal development [60–62]. PTHrP(107–139) has been reported to inhibit osteoclast activity and bone resorption in vitro [63] and in vivo [64], and to promote osteoblast proliferation and function [65–67], perhaps by down-regulation of sclerostin [68]. Although no receptor or signal transduction mechanism has been identified, preliminary evidence has suggested an action of PTHrP(107–139) through a Ca2+-coupled receptor of unknown identity in osteoblasts [69]. These experiments with the mid-region and distal regions of PTHrP are reminiscent of similar studies with carboxyl fragments of PTH, where biological effects have been identified [70], but no mechanism(s) ascertained.

Genetic experiments in mice support the likelihood that other biological activities may be exerted by domains within PTHrP that are released by proteolytic cleavage. The knock-in of PTHrP (1–84) lacking both the nuclear localization sequence (NLS) and C-terminal region while retaining the bioactive amino-terminal region, resulted in multiple abnormalities and early lethality in mice [71]. Homozygous mice exhibited skeletal growth retardation and osteopenia associated with reduced proliferation and increased apoptosis of osteoblasts, and early senescence. In a second model, the knock-in of PTHrP(1–66), which excludes a significant portion of the mid-region, resulted in a similar, but even more severe phenotype [72]. These mouse studies convincingly demonstrate that many of the actions of PTHrP are not mediated solely by the amino-terminal region, and that the absence of the mid-region, NLS or C-terminal region of PTHrP, result in greatly impaired commitment and survival of both osteogenic and hemopoietic precursors. Additional studies are required to ascertain the mechanisms of action of these important but less well-defined PTHrP regions. The final C-terminal “tail” region of PTHrP, from amino acids 142 to 173, is encoded by only one of the three isoforms of human PTHrP mRNA and is not represented at all in other species. The significance of this isoform remains unknown, although PTHrP(141–173) immunoreactivity has been identified in plasma [73], and PTHrP(140–173) immunoreactivity is present in human amnion [74]. Thus, while might be presumed that the actions ascribed to regions of the molecule beyond the N-terminal 34 amino acids are mediated by unique receptors, these have yet to be discovered.

Potential enzymatic processing sites abound in the PTHrP primary sequence, and it seems likely that processing of PTHrP takes place in an isoform- and tissue-specific manner. The physiological implications of post-translational processing of PTHrP are still poorly understood [75]. However, it is clear that the secreted forms of mid- and COOH-terminal regions of PTHrP have a number of bioactivities. This implicates PTHrP as a polyfunctional protein, as is the case with many neuroendocrine proteins that give rise to families of mature secretory peptides [76]. These forms may have their own receptors to elicit signaling pathways that are in part regulated by isoform- and cell-specific processing and secretion. It may also be that the great susceptibility of PTHrP to endoproteolytic cleavage is the result of evolutionary pressure to fit it so well as a paracrine and autocrine regulator with a very short tissue half-life and at the same time be a source of peptide products that serve other purposes, either as circulating hormones or as local agents.

The Mammalian PTHrP Gene and Its Products

The finding of multiple RNA transcripts on Northern analysis suggested that alternative splice products were likely. This was established with data showing two further cDNA species, predicting PTHrP variants of 139 and 173 amino acids in length [77–79]. When the genomic structure was identified shortly after [79–81], all three promoter regions of the complex PTHLH gene were defined and their specific splicing patterns resolved and confirmed. In normal human cells, PTHrP is encoded by a single copy PTHLH gene located on chromosome 12p11.2–p12.1 [77, 81]. The PTHLH locus is distinct from that of the PTH gene, which is located on chromosome 11p15.4. Chromosomes 11 and 12 are thought to have arisen through duplication events from a single common ancestral chromosome [82]. In addition to PTH and PTHLH, there are several examples of related genes, one localized to chromosome 11 and a related gene present on chromosome 12 [82, 83].

The PTHLH and PTH genes are similar in their overall genomic organizations, having similar exon–intron boundaries (Fig. 1). The human PTHLH gene is composed of nine exons spanning approximately 15 kb. The gene is transcribed by three functionally distinct promoters, and primary mRNA transcripts are alternatively spliced both at the 5′ and 3′ ends (Fig. 1) [79–81, 84]. It should be noted that with this nomenclature, there is no intron between exons 6 and 7 in the human PTHLH gene, although the sequence of this junction represents a splice donor site consensus sequence, and the observed splicing involving this site and exons VIII or IX shows its operation as a splice donor site.

In their structural organizations, the chicken, mouse and rat Pthlh genes share substantial homology with the human PTHLH gene, although the mouse and rat genes are considerably simpler than the human gene. The greater complexity of the human PTHLH gene appears to be a late evolutionary event, especially in light of the relative simplicity of the mouse and rat genes (Fig. 1). Both have exons equivalent to exons 4, 5, 6 and 9 of the human PTHLH gene and do not undergo alternative 3′ splicing, thereby producing only one isoform of PTHrP (Fig. 1). This conservation suggests that these exons constitute a minimal PTHLH gene structure. There is no evidence to suggest that the mouse or rat have exons equivalent to exons 1 or 2 of the human PTHLH gene. To date only one promoter region has been defined for the mouse and rat genes, equivalent to the P3 promoter of the human gene. Nucleotide sequence comparison of the mouse and rat genes with the human gene demonstrated substantial homology over the GC-rich promoter region (P2) and exon 3 of the human gene, suggesting that this region is also transcribed in rodents. Confirmation of this idea was achieved when transcripts corresponding to human exon 3 were detected in a clonal rat line (PT-r) [85] and an HHM-associated rat Leydig cell tumor (H500) [86]. However, there is no evidence to suggest that the mouse or rat have exons equivalent to exons 1 or 2 of the human PTHLH gene.

Evolutionary Aspects of PTHrP: Identification in Fish

From the time of the discovery of PTHrP, an obvious question was posed—if PTH and PTHrP were derived from a common ancestral gene, evolving through gene duplication, which came first, and was one a copy of the other? The search for answers to these questions began in fish. At the time of the discovery of PTHrP, stanniocalcin had recently been sequenced in the eel [87]. Stanniocalcin is a hypocalcemic hormone which was first identified in fish [87] and subsequently isolated and cloned in mammals [88]. When antiserum against hPTHrP(1–14) was used on sections of eel corpuscles of Stannius, the corpuscles were negative but the adjacent kidney tubules stained positively. The corpuscles from a number of fish species were also examined immunohistochemically for PTHrP, and all were found to be negative [89]. PTHrP immunoreactivity was also found in the corpuscles of Stannius in the euryhaline flounder (Platichthys flesus) [90] and possibly related to the fish moving from fresh to saltwater as part of its lifecycle.

The positive PTHrP staining in kidney tubules prompted a more extensive examination resulting in the demonstration of PTHrP in a number of fish tissues. In addition to the kidney tubules, PTHrP was localized to two areas of the fish pituitary (Sparus aurata), primarily located in the pars intermedia, and there were high circulating levels in sea bream as measured by N-terminally directed radioimmunoassay [91]. The presence of the antigen in the pituitary was surprising, but that the late Parsons [92] had suggested at a conference in 1979 that cod pituitary contained a factor similar to, but distinct from PTH. Parsons had data that this factor was hypercalcemic and contained a region that was “immunologically closely similar to the amino-terminal” of mammalian PTH. He predicted that this region would be relatively invariant between species and also speculated that this fish hypercalcemic factor might be “primitive” and the “evolutionary history of PTH may be hundreds of millions longer than previously believed”. This prescient statement was followed by the speculation that this hypercalcemic factor evolved as an anabolic agent because fish do not resorb their skeleton [92].

There were a number of studies that followed indicating that there was PTH or PTH-like activity in brain and pituitary of some bony fish [93, 94]. Other studies focused on tissues of fish and shark species and found that PTHrP localization reflected the picture seen in humans. Using antisera raised to amino-terminal hPTHrP, PTHrP antigen was demonstrated in the skin, kidney and muscle of lamprey (Geotria australis), lungfish (Neoceratodus fosteri), seven species of elasmobranchs (sharks) and rainbow trout (Oncorhynchus mykiss) [95, 96]. This raised the possibility that PTHrP might have role(s) similar to those in mammals, probably acting in an autocrine/paracrine manner and influencing cell growth and differentiation. This was supported by the demonstration of both the gene and protein in the developing denticle in one of the shark species (Mustelus antarcticus) [95]. Elasmobranchs have a cartilaginous skeleton and include sharks, skates and rays. It has been asserted by a number of researchers that both modern bony and cartilaginous fish have a common bony (armored) ancestor and that remnants of this armor exist in the scales on bony fish and the dermal denticles found on the surface of sharks.

Using parallel immunostaining and in situ hybridization, expression of the pthrp gene was also localized in the same tissues of the lamprey, elasmobranchs and rainbow trout [95]. Both protein and mRNA were localized in liver of the lungfish, elasmobranchs and rainbow trout as with the rectal gland of the elasmobranch. Identification of Pthrp in the rectal gland implied that it might have a role in osmoregulation [95, 97, 98]. This study also showed that the Pthrp protein was present in the pituitary of the lungfish, elasmobranch and rainbow trout, and the gene was present in the elasmobranch and rainbow trout pituitaries. The protein was detected in the nerve chord of the bony and cartilaginous fish as well [95].

Cloning of Fish PTHrP Genes

Once whole genome sequencing of the bony fish Fugu rubripes began in early 1990s, efforts began to isolate and clone the fish Pthrp gene. These studies revealed that gene structure (Fig. 1) was simpler than that of chicken, mouse or human with only three exons [96]. This structure was more like the highly conserved three exon PTH gene and was an exciting finding because it suggested that we might be closer to the point in evolution that indicates the duplication event explaining the similarity between PTH and PTHrP. This had been predicted because these two genes are on short arms of chromosomes 12 and 11, respectively, and the gene order and synteny is highly conserved among vertebrate species. When the first fish pth gene was identified in Fugu [99], this was counter to the accepted dogma that amphibians were the first animals to have PTH as they are the most primitive species to possess a parathyroid gland. Indeed, it dispensed with the idea that a parathyroid gland was needed to produce PTH.

Subsequently, several other copies of fish pthrp and pth genes were located in the Fugu genome [100]. These authors identified two pth genes (pthA and pthB), two pthrp genes (pthrpa and pthrpb) and a Pth-like ligand (pth-L) (Table 1). They were found to have hypercalcemic activity in fish but these extra copies of both pth and pthrp genes are not present in the human genome [101]. Although pth-L has not been identified in mammals, it has been found in the genomes of frogs and chickens [101] (Table 1). These in silico experiments have suggested that gene structure of Pth-L is more like pth than pthrp.

Table 1 The parathyroid hormone family genes identified in a range of vertebrate and invertebrate species
Full size table

It should be emphasized that bony fish have undergone a “teleost-specific” genome duplication. Genome duplication has long been accepted as the principal method by which new genes arise, resulting in new genes with different functions [102]. Some evidence implies that genome duplication allows the evolution of diversity [103]. The vertebrate genome has undergone several rounds of gene duplication, two before the jawless fish (agnathans), and then after bony fish split off from other vertebrates they underwent their “teleost-specific” genome duplication. This helps to account for the number of fish species worldwide and their ability to fill very different environmental niches. To further complicate matters, salmonids underwent another genome-wide duplication on their own, resulting in a large genome.

Both pthrp and pth genes were identified in the genome of the elephant shark (Callorhinchus milii) [104]. Two pth genes were present but only one pthrp gene (Fig. 1; Table 1). Elephant sharks have a cartilaginous skeleton-like elasmobranchs but belong to a sister group, the chimeras. They are the oldest group of living jawed vertebrates, and their genome has been termed the “vertebrate reference genome” [105]. The elephant shark pthrp gene was expressed in a wide range of tissues including kidney, brain, eye, heart, muscle, ovary, testis and uterus, reflecting the patterns seen in mammalian tissues [104]. The elephant shark Pthrp protein was localized in a broad range of tissues mirroring not only the expression but also immunohistochemical findings in other studies in fish and humans. This conservation of localization through hundreds of millions of years indicates that the function has probably been conserved as well. Certainly elephant shark Pthrp(1–34) peptide activated the common PTH1R in the rat osteosarcoma cell line, UMR106 [104]. It had been shown previously that Fugu pthrp(1–34) could activate the Pth1r in fish scale assays [100]. Recent work by Yan et al. [106] showed that the knockdown of zpthrpa gene causes premature bone mineralization during craniofacial development in zebrafish.

Interspecies comparison among the analogous PTHLH/Pthrp sequences reveals interesting similarities and offers rudimentary insight into function. Figure 2 depicts PTHLH and approximately 20 kilobase pairs (kbp) 5′ in the human gene [107]. Homologous segments of the mouse, chicken, Xenopus and Fugu DNA are drawn for comparison [108]. A graph of the Vertebrate Multiz Alignment (VMA) plot, a score of base-wise conservation across 28 vertebrate species, affords a more global perspective on conservation [109] (Fig. 2). The Encode H3k27Ac signals from the region are also plotted as an indication of likely enhancer and promoter activity locations [110] (Fig. 2).

Fig. 2
figure 2

Comparative genomics of the PTHLH/pthrp gene and its 20-kbp upstream region mapped to the HG18 genome (see text for details). The human PTHLH gene is featured in the top row, mirrored horizontally as it is an antisense gene. The second to fourth rows feature the genes for the mouse, chicken, Xenopus and Fugu. Where stretches of DNA comparative to the human gene exist, their conservation is graphed, leaving gaps where no comparable sequence exists. The Vertebrate Multiz Alignment graphs the relative conservation of each base pair among 28 vertebrate species. The bottom row features the results of longitudinal Encode H327Ac assay performed on multiple cell types

Full size image

It should be noted that the segment of DNA is reversed in both Xenopus and human, PTHLH being an antisense gene, but not the others. This indicates a transposition or breakage must have occurred in speciation. At least the 20 kbp 5′ is generally considered regulatory of the adjacent gene. The mouse shares significant homology with the human in this region, while the chicken shares shorter segments and is similar to the VMA, a consensus sequence based on 28 vertebrate species. This relationship continues through the intronic regions of PTHLH proper, which are likewise regulatory.

Considering the gene in its functional divisions, zone 1 in Fig. 2 contains exons 1–3. It also includes the 5′ un-translated region (UTR) which is likely important for post-transcriptional trafficking and regulation. The bulk of this segment is shared in its entirety by the human and mouse, but not the others featured. The VMA shows that the adjacent UTR cap forming the start point of the gene is moderately conserved. The intronic segment immediately preceding exon 2, however, is particularly highly conserved throughout the VMA species, along with the two segments of UTR, which are only slightly less conserved. This intron is likely a promoter region, which is supported by the high Encode H3k27ac assay signal intensity.

The second zone includes several introns as well as exons 4–5. The first intron precedes exon 5 and is well conserved in the mouse and chicken, but not Fugu, Xenopus or the VMA group. It likewise shows signs of regulatory potential on the H3k27Ac assay. Exon 5 contains further a UTR as well as the pre-pro sequence (−36 to −1) for entry to the endoplasmic reticulum and post-translational processing. This segment is maximally conserved across all featured species and the VMA group, suggesting there may be mechanisms common to multiple species for control of PTHrP production in the post-translational phase. A long intron separates exons 5 and 6. While this intron is almost wholly conserved in the mouse, only a short segment is shared in the chicken, and no conservation is seen in the Fugu or Xenopus gene. Interestingly, the 5′ half of the intron features H3k27ac signal which peters out through the segment; this is correlated with patchy VMA consensus, indicating this may be a conserved regulatory mechanism.

Zone 3 contains exons 6–8 and comprises the bulk of the PTHrP protein product. Exon 6 codes 1–139 and is well conserved in both the VMA group and the species depicted. It has high conservation across the VMA’s 28 species, suggesting it may code for a shared core function of PTHrP. Exons 7 and 8 and the intron between them are well conserved in both the rat and mouse, but not in vertebrates or the other species. Exon 7 codes an exon sandwiched between the 139 and 173 amino acid isoforms.

Zone 4 contains a long intron, the 5′ section of which is only conserved in the mouse, but the 3′ end of which is conserved in mouse, chicken and Fugu, but not in the VMA 28 species group. This precedes exon 9, which is highly conserved in mouse, chicken, Fugu and across the VMA species, but is absent from Xenopus.

The receptors for fish PTHrP and PTH were identified in zebrafish [111, 112]. In addition to identification of fish receptors equivalent to hPTH2R and hPTH1R, a third receptor was isolated and cloned [111]. The preferred ligand for this novel receptor (zPth3r) was hPTHrP(1–34). It was already known that the ligand for hPTH2R was tuberoinfundibular peptide 39 (TIP39), and its location was restricted to the brain [113], so it was inferred that fish Pthrp could bind to zPth1r and zPth3r [112].

Since PTHrP is produced by a number of mammalian epithelial tumors, including those of skin [114] and breast [115], the presence of the PTHrP antigen was examined in tumors from a number of lower vertebrates. Fish rarely develop true malignant tumors [116], but their neoplasms are usually benign and are delineated by their distinctive characteristic of progressive proliferation, which is uncoordinated with the surrounding tissue. A range of 51 neoplasms and tumors from bony and cartilaginous fish were examined for the presence of PTHrP, by immunohistochemistry using antisera to human PTHrP (1–14) [116]. The squamous carcinomata, renal adenocarcinoma and papillomata were positive for PTHrP, while the hepatocellular carcinomata, rhabdomyoma and myxoma were negative (Table 2, [116]). The finding of PTHrP in neoplasms of epithelial origin correlates well with the findings in mammalian tumors and indicates that PTHrP may be a marker of epithelial neoplasia in all vertebrates.

Table 2 Results of immunohistochemistry performed with antisera to hPTHrP(1–14) on sections of fish tumors submitted to the Registry of Tumors in Lower Animals, George Washington University, Washington DC
Full size table

Amphibians and Birds

Amphibians were not only the first animals to have a distinct parathyroid gland but they were also the first to move out of water and onto land. The process of metamorphosis allows the aquatic larval stages to assume a tetrapod form and terrestrial physiology, which assists them to survive in a dry environment and without the support of water, which is relatively dense. Thus, once Pthrp had been identified in fish, this species was the next to investigate.

Immunohistochemistry was carried with antibodies raised to hPTHrP(1–14), and in situ hybridization used probes made to the chicken PTHrP gene sequence [117]. The animals (Rana temporaria) that were examined were those that were just emerging from water and so should have the tetrapod physiology that would allow them to control their calcium metabolism via their skeleton. The tissues that expressed the gene and possessed the antigen were skin, kidney, muscle, pituitary and the olfactory lobe of the brain. The lobes of pituitary did not have concordant results. The expression of the PTHrP mRNA was present in the pars intermedia only but the protein was present in the pars distalis, neural lobe, infundibular stalk and median eminence. This finding indicated that the pars intermedia was the site of production and translation of PTHrP mRNA but that the protein is transported to the pars distalis and then possibly by retrograde transport to the median eminence and infundibular stalk [117].

At the same time, the two forms of PTH1R were identified in Xenopus laevis [118] (Table 1). In adult frogs, one of the receptor isoforms (xppr-a) was found to be expressed in skin, lung, brain and small bowel while the other (x-ppr-b) was highest in lung, brain and heart. xppr-a expression was also detected in a small population of possible osteoblasts lining the perichondral and diaphyseal bone trabeculae. These two studies reinforce the view that there is a significant evolutionary pressure to conserve the localization of the ligand and the receptor, indicating that PTHrP functions are basic and fundamental in all vertebrates.

The gene sequence for X. laevis pthrp gene (Table 1) was deduced recently by an in silico study [119]. The gene structure is composed of at least ten exons, and the putative protein is 131 amino acids in length. There appears to be alternative splicing, and it was reported that the gene encodes for the largest 3′UTR region seen in vertebrates. The xPthrp protein has the ability to stimulate calcium uptake across frog skin [101], thus one of PTHrP roles in Xenopus is probably as a modulator of calcium metabolism.

Because of their egg-laying capacity, birds have to produce readily available calcium supply for eggshell production on a daily or seasonal basis. Most studies have been carried out on chickens because of their agricultural importance. Gay [120] demonstrated that the calcium metabolism in chickens was amplified compared to humans because of this very rapid need for calcium. PTH has been shown to elevate calcium levels in chickens [121] by reduced bone uptake [122] and its actions on the kidney. Thiede et al. [123] demonstrated chicken PTHrP expression in the avian oviduct and proposed that PTHrP might play a role as a modulator of the oviduct smooth muscle tension that is required during egg laying.

The gene for chicken PTHrP [124, 125] (Fig. 1) and the protein have been sequenced, and the protein is highly homologous to human PTHrP. Recent in silico studies predicted that chicken PTHrP gene consists of six exons, and the mature protein is 139 amino acids in length [119], although earlier work [125, 126] showing a five exon structure had predicted alternative splicing resulting in a protein of 139 amino acids or a protein of 141 amino acids (Fig. 1).

The PTHrP gene is expressed during skeletal histogenesis in the chick mandible [127]. Both PTHrP and the PTH1R were present in the chick embryo before the appearance of bone and were expressed by cells in the ectoderm, skeletal muscle, peripheral nerve and mesenchyme. Previously, chicken PTHrP and PTH1R gene expression had been demonstrated in the chicken epiphyseal growth plate [128]. Both these studies support the role of PTHrP in the regulatory loop that is present in the chicken prehypertrophic and hypertrophic zone of the postnatal growth plate, thus indicating PTHrP involvement in bone formation in birds. Schermer and co-authors [125] found that PTHrP was expressed in adult chick brain, intestine, liver, skeletal muscle, heart, lung and gizzard. The expression of the gene in these tissues again reflects the situation in other vertebrates. The expression in the chick embryo suggested that at least part of the function of PTHrP in the chick is to influence cell growth and differentiation, reflecting its role in mammals.

Genescan dataset and a number of common tools such BLASTP and TBLASTN were used to examine the G protein-coupled receptors in the chicken genome [120]. It was hypothesized that the cPTH1R gene was a pseudogene as the fourth exon contained a stop codon. Given the role that PTH plays in avian calcium metabolism [121, 122], it seems very unlikely that cPTH1R is not functional. The fact that chickens do not have the PTH2R [129] or its ligand makes it unlikely that their PTH1R is inoperative. But the chickens do have cPTH3R which humans lack [119] (Table 1). This receptor was expressed predominantly in the hindgut, lung and cartilage of the chicken, while cPTH1R had the more ubiquitous expression profile that is seen in humans.

Conclusions

The recent sequencing of the whole lamprey genome (Lethenteron japonicum, [130]) is instructive. Agnathans (jawless fish) have a pivotal position in evolutionary history, having undergone two whole genome duplications when compared to invertebrates. They have a cartilaginous skeleton, and PTHrP has been localized in lamprey (G. australis) tissues using antisera raised against human PTHrP(1–34) [96]. This showed that the tissue localization was conserved among vertebrates and agnathans. Both pth1r and pth2r genes have been putatively identified in one of agnathan genome (Petromyzon marinus) [119] (Table 1).

Many questions remain unanswered. When did the gene duplication take place and which came first, PTH or PTHrP? Neither the PTH nor the PTHrP protein or gene has been identified in invertebrates. Interestingly, when hPTHrP (1–14) antiserum was used in immunohistochemistry on sections of an ascidian (Pyura praeputialis), the epithelia and the neural complex were positively stained [131]. This species is a sessile and solitary ascidian. Hull and co-authors [132] detected a PTH-like protein in snail brain but were unable to detect PTH immunoreactivity in hagfish, squid, cuttlefish or starfish. Thus, the presence of PTH or PTHrP in invertebrates is yet to be finally determined but two PTH receptors have been identified in Ciona intestinalis genome [133] (Table 1), indicating that these receptors pre-dated the formation of cartilage and bone.

Comparative endocrinology has established that PTHrP has a long evolutionary history, and the protein and its tissue localization are very highly conserved from fish to man. This strong evolutionary pressure to preserve PTHrP implies that PTHrP role(s) are basic and fundamental in all vertebrates. PTHrP clearly has roles in skeletal formation and cell proliferation and differentiation throughout evolution.