Division of Endocrinology and Metabolism, Department of Medicine (S.M.J.), The Ilyssa Center for Molecular and Cellular Endocrinology (S.M.J., M.A.L.), and Division of Pediatric Endocrinology, Department of Pediatrics (M.A.L.), The Johns Hopkins University School of Medicine, Baltimore, Maryland 21287
Address all correspondence and requests for reprints to: Suzanne M. Jan de Beur, M.D., Assistant Professor of Medicine, 1830 East Monument Street, Suite 333, Baltimore, Maryland 21287. E-mail: . sjandebe{at}jhmi.edu
The past several years have brought dramatic advances in our understanding of the molecular and biochemical bases of inherited and acquired hypophosphatemic disorders. In large part, these advances can be attributed to the development of new molecular techniques and improved genomic databases. The identification of the genetic defects in X-linked and autosomal dominant forms of hypophosphatemic rickets by positional cloning, and the discovery of novel phosphate-regulatory genes in tumors associated with oncogenic osteomalacia by gene expression profiling, have begun to elucidate novel phosphate homeostatic pathways. It is, therefore, particularly appropriate to include this perspective on the molecular pathogenesis of hypophosphatemic syndromes in recognition of the "Genes and Genomics in Endocrinology" theme for this months Annual Meeting of The Endocrine Society.
Phosphate homeostasis: our current understanding
Phosphorus plays a critical role in skeletal development, mineral metabolism, and diverse cellular functions involving intermediary metabolism and energy-transfer mechanisms. It is a vital component of bone mineralization, phospholipids in membranes, nucleotides that provide energy and serve as components of DNA and RNA, and phosphorylated intermediates in cellular signaling.
Phosphorus exists in the plasma in an inorganic form and an organic form. Approximately 20% of the plasma inorganic phosphorus is protein bound, and the remainder circulates as free phosphate ions HPO42 - or H2PO4-. Phosphorus in the form of the phosphate ions circulates in the blood and is filtered at the glomerulus. However, measurement of the plasma and urine phosphate content is expressed in terms of elemental phosphorus. Thus arises the interchangeable use of phosphorus and phosphate concentrations.
Similar to calcium, the serum phosphate level is maintained within a narrow range through a complex interplay between intestinal absorption, exchange with intracellular and bone storage pools, and renal tubular reabsorption. Hypophosphatemia stimulates calcitriol synthesis via the 25(OH)D-1-hydroxylase in the kidney, leading to increased calcium and phosphorus absorption in the intestine and enhanced mobilization of calcium and phosphorus from bone. In addition, hypophosphatemia is a potent stimulator of an increase in maximal tubular reabsorption of phosphate (TmP/GFR). The resultant increased serum calcium inhibits PTH secretion with a subsequent increase in urinary calcium excretion and increases tubular reabsorption of phosphate. Thus, normal serum calcium levels are maintained and serum phosphorus levels are returned to normal.
The principal organ that regulates phosphate homeostasis is the kidney. Serum inorganic phosphorus (Pi) is filtered by the glomerulus, and 80% of the filtered load is reabsorbed predominantly along the proximal nephron. Regulation of proximal renal tubular reabsorption of phosphate is achieved through changes in the activity, number, and intracellular location of the brush border membrane type IIa sodium-phosphate co-transporter (NPT2). In mice, Npt2 is the major sodium-phosphate (Na+/Pi) co-transporter in the renal proximal tubule and comprises 84% of the sodium phosphate co-transporters (1). Transgenic mice, in which both copies of Npt2 have been disrupted, develop marked renal phosphate wasting that is associated with an 80% loss in proximal renal tubular absorptive capacity (2).
Many hormones and cytokines influence Pi reabsorption in the proximal renal tubule. GH, IGF-I, insulin, epidermal growth factor, thyroid hormone, calcitriol, and dietary phosphate depletion stimulate renal Pi reabsorption. PTH, PTH-related protein (PTHrp), calcitonin, atrial naturetic factor, TGF and ß, and glucocorticoids all inhibit renal Pi reabsorption.
PTH is the best-characterized physiological regulator of phosphate reabsorption, but its prinicipal function is to maintain calcium homeostasis. PTH increases urinary phosphate excretion via cAMP-dependent inhibition of NPT2 expression. This effect is rapid and is achieved by endocytic retrieval of NPT2 molecules from the brush border membrane and enhanced lysosomal degradation (3, 4). By contrast, both acute and chronic Pi deprivation initiates an adaptive increase in brush border membrane sodium phosphate transport through microtubule-dependent recruitment of Npt2 protein to the apical membrane surface (5).
Growth and development of the skeleton requires adequate supplies of calcium and phosphate, and deficiency of these minerals results in impaired bone mineralization. This mineralization defect results in rickets in growing children with open epiphyses and osteomalacia in adults. Hypophosphatemic rickets, as occurs in renal phosphate wasting disorders, has many clinical and radiographic features that are similar to those that occur in calcium deficiency, but the accumulation of excess osteoid is not associated with secondary hyperparathyroidism or increased bone reabsorption. Genetic defects that lead to decreased renal tubular reabsorption of phosphate and chronic hypophosphatemia are the most common cause of inherited rickets. In addition, tumor-induced hypophosphatemia (i.e. oncogenic osteomalacia) is often an occult cause of rickets and is accompanied by considerable morbidity. The elucidation of the pathogenesis of acquired and inherited hypophosphatemic syndromes has rapidly advanced our understanding of not only the syndromes themselves but the general mechanism of phosphate homeostasis.
Inherited and acquired hypophosphatemic disorders: is there a common molecular pathophysiology?
X-linked hypophosphatemic rickets (XLH).
XLH (OMIM no. 307800) is the most common inherited hypophosphatemic disorder, accounting for 80% of cases of familial phosphate wasting. In XLH, hypophosphatemia occurs as a consequence of decreased renal tubular Pi reabsorption. Serum levels of phosphate are reduced, and serum calcitriol levels are reduced or inappropriately normal while serum calcium and PTH levels are normal (Table 1). Aberrant regulation of the 25(OH)D-1-
-hydroxylase accounts for the paradoxical occurrence of inappropriately normal or reduced calcitirol levels in the face of hypophosphatemia. In its fullest expression, XLH is associated with rickets/osteomalacia, lower extremity deformities, short stature, bone pain, enthesopathy, and dental abscesses (6). By contrast, patients with XLH do not have the muscle weakness or pain that is typically experienced by patients with acquired hypophosphatemia. Although females with XLH are heterozygous, there is little, if any, difference in the severity or extent of the disorder in affected males and females (7, 8).
|
A broad range of PHEX gene defects has been identified in humans (Refs. 7 , 11, 12, 13 and Phexdatabase). XLH is an X-linked dominant disorder in which one defective PHEX allele results in disease. Extensive analysis of phenotype/genotype relationships showed that there was no correlation between severity of disease and type or location of PHEX mutation. Because PHEX is located on the X chromosome, random X chromosome inactivation in females is predicted to result in a less severe phenotype compared with males who are completely PHEX deficient. However, males and females seem to be affected equally with similar biochemical indices and skeletal manifestations, although in one study postpubertal males tended to have more dental disease (7). X chromosome inactivation analysis of the PHEX gene has demonstrated that the PHEX gene is randomly inactivated. Thus, preferential inactivation of the normal PHEX allele is excluded as a potential basis for the similar phenotype observed between males and females with XLH.
PHEX gene defects have been identified in mouse models of XLH. The hyp mouse has a large 3' deletion in the Phex gene (10). A similar XLH murine homolog, the Gy mouse, exhibits the hyp phenotype plus a gyro-rotating behavior. The Gy mouse has a large deletion that includes the 5' Phex gene and spermine synthase gene (14).
Although mutational analysis provides compelling evidence for the causative role of PHEX defects in XLH, data demonstrating correction of the hypophosphatemia, osteomalacia, and rickets with gene rescue are conflicting. Miyamura et al. (15) performed syngenic bone marrow transplant in hyp mice and demonstrated a significant increase in serum phosphorous, increase in Npt2 gene expression, increase in bone density, reduction in alkaline phosphatase, and reduced vitamin D catabolism that correlated with the percentage of engrafted cells. Further supporting the role of PHEX in defective bone mineralization and renal phosphate wasting, Shih et al. (16) stably transfected a human osteoblast cell line with a PHEX antisense vector and decreased PHEX mRNA and protein. This resulted in impaired mineralization of the osteoblasts in vitro. Conditioned medium from the cultured PHEX antisense osteoblasts inhibited 45Ca incorporation and nodule formation by wild-type osteoblasts and inhibited 32P uptake in opossum kidney cells (16). Recent work by Liu et al. (17), however, failed to demonstrate correction of hypophosphatemia, rickets, or osteomalacia in transgenic hyp mice in which overexpression of Phex cDNA was targeted to osteoblasts under the control of the osteocalcin promoter.
The PHEX gene encodes a 749 amino acid protein that is a member of the M13 family of membrane-bound metalloproteases (9). By analogy to other members of this family, such as neutral endopeptidase 24.11, that have been shown to inactivate hormones (18) and endothelin-converting enzyme that process hormones to a mature form (19), PHEX is likely to proteolytically cleave a peptide hormone, thereby either activating or inactivating its substrate (20). How the loss of PHEX function leads to renal phosphate wasting, defective bone mineralization, and reduced calcitriol synthesis is not apparent. However, several observations have offered insights into the function of PHEX. Levels of Npt2 protein and mRNA are reduced in hyp mice to 57% of normal littermates, suggesting a relationship between loss of Phex and transcriptional down-regulation of Npt2 (21), thus implicating Phex in a regulatory role in renal phosphate handling. Furthermore, there appears to be a dual pathophysiology in XLH: a primary osteoblast defect that may account for the defective bone mineralization observed in this disorder and a circulating phosphaturic factor that may result from failure of PHEX to inactivate its substrate.
The hyp and gy mice, murine homologs of XLH, have provided useful models to study the basis for hypophosphatemia in this disorder. Early studies of parabiotic union of hyp and normal mice showed development of hypophosphatemia and phosphaturia in the normal mouse rather than correction of the phosphaturic defect in the hyp mouse (22) and provided the initial evidence implicating a humoral factor as the cause of renal phosphate wasting. Subsequent work using renal cross-transplantation between normal and hyp mice showed that the mutant phenotype was not transferred by transplantation of hyp kidneys into normal mice, nor was the hyp defect corrected by transplantation of normal kidneys into hyp mice (23). These studies confirmed earlier suggestions that hyp mice produce a circulating factor that inhibits the Npt2 co-transporter and that is not inactivated by circulation through a normal mouse (24). The humoral factor in the hyp mouse was further characterized in studies that demonstrated that hyp mouse serum could inhibit phosphate uptake in primary cultures of mouse proximal renal tubular cells in a dose-dependent manner (25). A similar pathogenetic mechanism appears operative in XLH, because transplantation of an unaffected sisters kidney into a brother with XLH led to renal phosphate wasting by the normal kidney (26). The source of the circulating factor is presently unknown, but the observation that transplantation of hyp bone nodules into normal mice fails to completely correct the mineralization defect has suggested that osteoblasts may produce the factor (27).
In addition to hypophosphatemia, an intrinsic osteoblast defect also contributes to the bone disease in XLH. Even after adequate treatment with phosphorus and calcitriol, XLH patients continue to have hypomineralized periosteocytic lesions in bone (28). Moreover, explants from hyp mouse calvaria fail to mineralize appropriately even in physiologically normal environments (29) and the mineralization defect is transferable to normal cells in co-culture experiments (29). Further evidence suggests that a mutant Phex gene results in abnormal bone matrix protein expression and deposition (30). The dual defect in the osteoblast and evidence for a humoral factor has led to the hypothesis that the intrinsic bone defect leads to release of humoral factors that effect bone mineralization, calcitriol synthesis, and proximal tubular Pi absorption.
Abnormalities in the metabolism of vitamin D are characteristic of individuals with XLH. Stimulators of calcitriol synthesis such as PTH or low phosphate diets fail to increase calcitriol synthesis in patients with XLH. hyp mice demonstrate reduced 25(OH)D-1-hydroxylase activity after provocative manipulations such calcium, vitamin D, Pi deprivation, and PTH infusion (31). Furthermore, 24-hydroxylation, the initial step in calcitriol catabolism, is increased in hyp mice, contributing to the calcitriol synthetic defect (32). The pathogenesis of the renal phosphate wasting is dissociable from the vitamin D synthetic defect. Npt2 null mice have renal phosphate wasting but exhibit an appropriate increase in calitriol levels and synthesis. In contrast to hyp mice, osteomalacia and rickets are absent in the Npt2 null mice. This suggests that the circulating phosphaturic factor implicated in XLH and other hypophosphatemic syndromes has two independent effectsone inhibits renal phosphate conservation and the other impairs renal calcitirol synthesis.
Autosomal hypophosphatemic rickets.
Autosomal dominant hypophosphatemic rickets (ADHR; OMIM no. 193100) is characterized by low serum phosphorous concentration, phosphaturia, inappropriately low or normal 1,25(OH)2D levels, and bone mineralization defects that result in rickets, osteomalacia with bone pain, lower extremity deformities, and muscle weakness. The adult onset form of the disorder presents with osteomalacia with bone pain, weakness, and fractures, but no deformity. Those with childhood onset look phenotypically like XLH. Dental abscesses are a prominent feature of the syndrome. The pattern of inheritance with evidence of male-to-male transmission and the phenotypic variability distinguish this disorder from XLH. ADHR is a phenotypically variable disorder with incomplete penetrance, delayed onset (33), and, in several kindreds, postpubertal reversal of renal phosphate wasting.
Linkage studies of one large ADHR kindred demonstrated linkage of the ADHR trait to chromosome 12p13 (34). Subsequently, specific mutations in fibroblast growth factor (FGF)-23 were identified in several kindreds with ADHR (35). Composed of a 10-kb genomic sequence, three exons, and encoded by a 2.3-kb cDNA, FGF-23 is expressed at low levels in brain, thymus, small intestine, heart, liver, lymph node, thyroid/parathyroid, and bone marrow (35, 36, 37) and highly expressed in tumor-induced osteomalacia (TIO) tumors with 3 kb and 1.3 kb transcripts easily detected by RNA blot analysis (38, 39). Its expression is notably absent from bone.
FGF-23, the largest member of the FGF family, contains 251 amino acids, including a 24 amino acid hydrophobic amino terminus that is a signal sequence (36). FGF-23 has a distinct carboxy terminus that is not homologous to other members of the FGF family. FGF-23 is most similar to FGF-21 and FGF-19 (24% and 22% amino acid identities, respectively; Ref. 36). FGF-23 lacks several heparin binding residues that are conserved in other FGFs and, thus, FGF-23 may be more soluble and more likely to circulate than other FGFs (40).
Missense mutations in one of two arginine residues at positions 176 or 179 have been identified in affected members of four unrelated ADHR families (35). This clustering of missense mutations suggests that they are activating mutations. Furthermore, the mutated arginine residues, located in the consensus proteolytic cleavage RXXR motif, prevent the degradation of FGF-23 and, thus, may result in prolonged or enhanced FGF-23 action (38, 41). Thus far, FGF-23 is the only known FGF that is associated with a human disease.
Preliminary reports (42) demonstrate that FGF-23 binds to the FGFR2 and FGFR4 receptors, both of which are expressed in the kidney. Interestingly, the FGF-23 (R179Q) mutant binds preferentially to FGFR4. Identifying a specific receptor for FGF-23 is an area of intense interest, and although the entire FGF family shares four receptors, the distinctive structural properties of FGF-23 suggest that it may have a unique receptor.
FGF-23 has been proposed to have either paracrine and/or endocrine roles. Decreased FGF-23 secretion could represent an endocrine response to dietary phosphate restriction, and, thus, circulating FGF-23 might provide a closed feedback loop between a putative phosphate sensor and renal phosphate reabsorption (40). By analogy to PTHrp, a paracrine factor that was initially recognized as the humoral mediator of malignancy-associated hypercalcemia, FGF-23 may function physiologically as a locally-acting factor, but may also function pathophysiologically when secreted in excess into the circulation, where it can cause marked renal loss of Pi.
Hereditary hypophosphatemic rickets with hypercalciuria (HHRH).
Insights gained from a third inherited hypophosphatemic disorder implicate a role for an as yet unidentified modulator of NPT2 expression or activity. HHRH (OMIM no. 241530) was first described in an inbred Bedouin kindred (43), and subsequently a few additional familial or sporadic cases have been identified. Because some members of the initial Bedouin family displayed only hypercalcuria whereas others displayed all the features of HHRH, it was hypothesized that those with only hypercalciuria were heterozygous for the mutant gene and those with both hypophosphatemia and hypercalcuria were homozygous for the mutation. Thus, the mode of inheritance is thought to be autosomal recessive. The biochemical characteristics of this syndrome are hypophosphatemia owing to decreased renal Pi reabsorption, and elevated levels of calcitriol that account for hypercalciuria. Circulating levels of PTH are not increased, suggesting that an elevated calcitriol level represents the normal response to hypophosphatemia. Patients with HHRH have bone pain, rickets and osteomalacia, muscle weakness, and growth retardation. The presence of elevated serum levels of calcitriol and hypercalciuria, as well as muscle weakness, distinguish patients with HHRH from patients with either XLH or ADHR. The striking similarity between human HHRH and the phenotype of Npt2 null mice lead to the hypothesis that HHRH is, indeed, a primary disorder of renal phosphate reabsorption. Homozygous Npt2 mutant mice exhibit increased urinary Pi excretion, hypophosphatemia, appropriately elevated calcitriol levels with attendant hypercalcemia, hypercalciuria, and suppressed PTH (2). Heterozygote Npt2 mutant mice display normal serum phosphorus but intermediate urine phosphate excretion and hypercalcemia. In contrast to humans with HHRH, Npt2 null mice do not have rickets or osteomalacia; rather, they exhibit retarded secondary ossification at weaning, followed by reversal of the early bone phenotype with age (2). Based on the observations in the Npt2 null mice, NPT2 was considered the leading candidate gene for HHRH. The NPT2 gene exons and intron/exon boundaries were screened for mutations in affected individuals from the original Bedouin kindred and four additional small families. No putative disease-associated mutations were identified. In addition, linkage analysis was performed using two NPT2 single nucleotide repeats and five flanking microsatellite markers in the NPT2 region on chromosome 5q35. These markers were not linked to HHRH. Furthermore, analysis of allele-sharing among affected individuals with HHRH using microsatellite markers in a 500-kb region including NPT1 on chromosome 6p22 failed to provide evidence of linkage to NPT1 (44). Candidate genes for HHRH include genes that encode other renal Pi transporters or accessory proteins that regulate NPT2 expression or function. To definitively determine the genetic basis of HHRH, additional candidate genes will need to be investigated or a genome-wide linkage scan will need to be performed.
TIO/oncogenic osteomalacia.
Oncogenic osteomalacia, or TIO, is an acquired hypophosphatemic syndrome that shares many features with the genetic forms of hypophosphatemic rickets. TIO is typically caused by a variety of benign, primitive mesenchymal tumors that secrete factors, collectively termed "phosphatonin" (45), that can inhibit proximal renal tubular Pi reabsorption and impair synthesis of calcitriol. The resulting hypophosphatemia impairs skeletal mineralization and causes rickets or osteomalacia. Experimental studies have provided evidence that phosphatonin can regulate phosphate homeostasis. Tumor extracts and conditioned media from cultured tumors can inhibit phosphate transport by renal cell lines in vitro (46, 47, 48, 49, 50), and can induce phosphaturia and hypophosphatemia when administered in vivo to mice (51). Furthermore, surgical removal of tumor tissue results in normalization of serum phosphate and calcitriol, reversal of phosphaturia, and eventual mineralization of bone.
The slow growth of cultured tumor cells and the frequent loss of phosphate-inhibitory activity of cultured or transplanted tumors has hampered the biochemical characterization or purification of the phosphaturic substance(s) produced by these tumors. However, molecular analyses of genes that are highly expressed in TIO tumors have led to the identification of several candidate genes for the phosphaturic substance (37, 38, 52, 53). These studies have demonstrated overexpression of FGF-23 in all TIO tumors in contrast to normal tissues where it is expressed at low levels, thus confirming the important role for this factor as a phosphate-regulatory protein first suggested by studies of the FGF-23 gene in patients with ADHR (35). Conditioned media from cells transfected with FGF-23, proteolytically-resistant FGF-23 mutants (R179Q, R176Q, R179W), purified recombinant human FGF-23, and recombinant FGF-23 R179Q all inhibit phosphate transport in cultures opossum kidney cells, a standard model for renal proximal tubular epithelium (38). When infused into mice, FGF-23 reduces serum phosphate and increases fractional excretion of Pi (37), and mice that are chronically exposed to elevated levels of recombinant FGF-23 produced by xenografts of stably-transfected Chinese hamster ovary cells become hypophosphatemic with increased renal phosphate clearance and demonstrate reduced bone mineralization. Moreover, these mice develop low circulating levels of calcitriol as a consequence of reduced expression of renal 1-hydroxylase (37). It is unknown whether FGF-23 directly inhibits phosphate transport in the kidney and, if so, the mechanism by which it exerts its effects. Furthermore, it is yet to be demonstrated whether serum levels of FGF-23 are elevated in individuals with TIO. Because several other genes are highly expressed in these tumors, including MEPE, Dentin matrix protein-1, osteopontin, and Frizzled-related protein-4 (37, 52, 53, 54), it remains possible that multiple secreted factors are involved in the pathogenesis of the syndrome.
The striking biochemical and clinical similarities between of TIO, XLH, and ADHR (Table 1) have led to speculation that a common circulating factor plays a pathogenic role in these disorders. The discovery of PHEX as the molecular defect in XLH and its homologies with other endopeptidases, has invited speculation concerning potential interactions between the PHEX metallopeptidase and the circulating phosphaturic factor in TIO. In an in vitro translation system, recombinant PHEX cleaved FGF-23 but not ADHR mutant FGF-23 (R179Q) (38). Studies of a recombinant, soluble PHEX revealed that: 1) PTHrp (107139) is a substrate for PHEX, and 2) osteocalcin, pyrophosphate, and Pi inhibited soluble PHEX activity (55). FGF-23, however, was not tested. The ability of mouse recombinant Phex to cleave a number of substrates, including an FGF-23 fragment encompassing the ADHR mutation site (amino acids 172186) was tested (55), and only [Leu]enkephalin was cleaved by Phex. [Leu]enkephalin, however, was not demonstrated to be a substrate for soluble PHEX (56).
Unifying hypothesis.
The study of acquired and inherited hypophosphatemic disorders has lead to a new understanding of the hormonal regulators of phosphate homeostasis. TIO, XLH, and ADHR have overlapping phenotypical features but all share hypophosphatemia due to decreased renal Pi reabsorption. Current evidence supports the notion that under normal physiologic states, the concentration of FGF-23 in serum (and possibly other tissues) is regulated via PHEX-dependent proteolysis (Fig. 1). However, conditions that lead to excess circulating FGF-23 concentrations or activity are associated with a marked depression in proximal renal tubular reabsorption of phosphate and hypophosphatemia. When PHEX is inactiveas in patients with XLHFGF-23 is not degraded and accumulates in the circulation. In ADHR, missense mutations replace key amino acids in FGF-23 and render the protein resistant to proteolysis, thereby leading to enhanced circulating FGF-23 concentrations or action. Similarly, ectopic overproduction of FGF-23 by TIO tumors may saturate the capacity of endogenous proteolytic enzymes such as PHEX to degrade FGF-23. It is clear that FGF-23 plays a central role in the pathophysiology of these disorders; however, its role in normal physiology and the mechanism(s) by which it regulates renal phosphate handling or calcitriol synthesis is unknown.
|
Acknowledgments
Footnotes
Abbreviations: ADHR, Autosomal dominant hypophosphatemic rickets; FGF, fibroblast growth factor; HHRH, hereditary hypophosphatemic rickets with hypercalciuria; Pi, inorganic phosphorus; PTHrp, PTH-related protein; TIO, tumor-induced osteomalacia; XLH, X-linked hypophosphatemic rickets.
Received April 3, 2002.
Accepted April 3, 2002.
References