Emerging role of a phosphatonin in mineral homeostasis and its derangements


  • Institute of Physiology, Zurich, Switzerland (B. Bielesz).

Dr Bernhard Bielesz, Medical University of Vienna, Department of Internal Medicine III, Division of Nephrology and Dialysis, Währinger Gürtel 18–20, 1090 Vienna, Austria. Tel.: +43-1404004358; fax: +43-1404004452; e-mail: bernd.bielesz@gmx.net


Background  The study of a distinct group of renal phosphate wasting disorders with bone disease which comprise X-linked hypophosphatemic rickets (XLH), autosomal dominant hypophosphatemic rickets (ADHR) and tumour-induced osteomalacia (TIO) gave rise to the identification of different hormone-like peptides, also known as phosphatonins. These factors are responsible for the major disease features that characterize XLH, ADHR and TIO. Recent reports on one of these phosphatonins, fibroblast growth factor-23 (FGF-23), point to a general role of this factor in mineral ion metabolism.

Objectives  The main focus regards recent evidence implicating FGF-23 in normal and disordered mineral homeostasis with special emphasis on chronic kidney disease. The interactions of FGF-23 with phosphate, parathyroid hormone and vitamin D are discussed in detail.

Summary  The FGF-23 has been shown to increase urinary phosphate excretion, inhibit bone mineralization and suppress 1,25-dihydroxy vitamin D3[1,25(OH)2D3], the main characteristics that XLH, ADHR and TIO have in common. Apart from its role in these phosphate wasting disorders serum FGF-23 is elevated in hypoparathyroidism and humoral hypercalcaemia of malignancy and responds to altered dietary phosphate and calcium supply in healthy subjects. The FGF-23 is also variably elevated in chronic kidney disease and associated secondary hyperparathyroidism where it correlates positively with serum phosphate and parathyroid hormone and negatively with 1,25(OH)2D3. Such relationships, along with data from experimental studies, raise the question of whether FGF-23 contributes to the pathophysiology of chronic kidney disease.


The regulation of overall body calcium and phosphate homeostasis is tightly coupled with the regulation of bone metabolism and turnover. Parathyroid hormone (PTH) and 1,25-dihydroxy vitamin D3[1,25(OH)2D3] are well-known key-players in these processes. A distinct group of hypophosphatemic disorders is characterized by isolated renal phosphate wasting and impaired bone mineralization leading to either rickets or osteomalacia. There is considerable scientific consensus that one or more hormone-like factors, also known as phosphatonins, that circulate in the blood stream play a central causative role in these diseases. Recently, an established candidate to represent such a phosphatonin, fibroblast growth factor-23 (FGF-23), has been intensively investigated in a context different from hypophosphatemic disorders. The FGF-23 can be variably elevated in chronic kidney disease and secondary hyperparathyroidism, humoral hypercalcaemia of malignancy, as well as by increases in both dietary calcium and phosphate intake. These findings suggest a global function of FGF-23 in normal and deranged body mineral metabolism. This review focuses on discussing the available evidence and, based on current knowledge, attempts to draw a picture of its functions and interactions with PTH and 1,25(OH)2D3.

The following paragraph outlines the above-mentioned phosphate wasting disorders, namely two hereditary forms of hypophosphatemic rickets and acquired tumour-induced osteomalacia. Although a detailed discussion of the clinical and pathophysiological aspects of these disorders is beyond the scope of this article, the reader is referred to comprehensive and recent reviews on this topic [1–8].

Hereditary hypophosphatemic rickets and tumour-induced osteomalacia

X-linked hypophosphatemic rickets (XLH), autosomal dominant hypophosphatemic rickets (ADHR) and tumour-induced osteomalacia (TIO) are similarly characterized by hypophosphatemia, decreased tubular maximum phosphate re-absorption per decilitre of glomerular filtration rate (TmP/GFR), phosphaturia and defective bone mineralization. The TmP/GFR defines the threshold of serum phosphate above which net phosphate excretion occurs. The reference value approximates normal serum phosphate levels [9]. Serum calcium is generally normal. The patients fail to adequately up-regulate serum 1,25(OH)2D3 in response to the prevailing hypophosphatemia. Thus, serum active vitamin D3 levels are either normal or decreased. The PTH serum levels are mostly normal but can also be elevated in XLH and TIO. The role of hyperparathyroidism in XLH is discussed in detail in reference [10].

The XLH is the most common form of hereditary rickets (prevalence 1 : 20 000). Patients are growth retarded and exhibit either bowing or knock-knee deformity of the lower extremities. The incidence of ADHR is extremely rare and largely similar in terms of clinical presentation but variable penetrance has been reported [11]. In TIO, patients suffer from osteomalacic bone disease with bone and muscle pain. In most cases benign mesenchymal neoplasms underlie the disorder. These lesions are often small in size which makes them difficult to localize. The disease is effectively cured by complete surgical removal of the causative tumour. The TIO can also occur in association with malignancies such as prostate carcinoma and multiple myeloma, reviewed in reference [12]. The prevalence of TIO is difficult to assess, but in 2001 Drezner reported approximately 120 patients in a comprehensive review of published cases [12]. However, many additional case reports and studies on TIO have been published which suggests that clinical awareness is increasing.

The XLH is caused by mutations in the phosphate-regulating gene with homologies to endopeptidases on the X-chromosome (PHEX) [13], which is a membrane-bound enzyme which is predominantly expressed in bone, teeth and chondrocytes [14–19]. Homologies of PHEX to M13 zinc metallopeptidase genes suggest that PHEX functions as an endopeptidase which either activates or inactivates certain precursor peptides, reviewed in reference [20]. Thus far, the PHEX substrate(s) of physiological significance cannot be reliably identified and either direct or indirect interaction with the phosphatonins is proposed but more evidence is required.

The FGF-23 is mutated in ADHR [21], but only three mutations have been described to date that were found in four unrelated ADHR families. All these mutations prevent FGF-23 cleavage into an N- and C-terminal peptide of similar size by disrupting a consensus cleavage site of subtilisin-like proprotein convertases. According to current understanding, only full length FGF-23 is biologically active [22]. Consistent with that, mutations underlying ADHR have been shown to result in increased biological activity of FGF-23 [23,24]. Proprotein convertases represent a family of enzymes involved in the processing of peptide hormones, neuropeptides and growth factors, as reviewed in reference [25]. The FGF-23 is, among other tissues, expressed in normal bone [26–29].

Common pathophysiological concept – the phosphatonins

It is widely accepted that the renal phosphate wasting in XLH, ADHR and TIO is primarily induced by a PTH-independent mechanism. The term phosphatonin was introduced to describe one or a set of postulated factors which circulate in the blood stream that are responsible for the biochemical and skeletal defects in XLH and TIO [30,31]. Tumour tissue surgically removed from patients diagnosed with TIO has proved useful in identifying candidate substances that might act as a phosphatonin. Matrix extra-cellular phosphoglycoprotein (MEPE) and secreted frizzled related protein-4 (sFRP4) were both identified owing to the abundant expression in TIO causing tumours [32,33].

There is evidence that MEPE, which is expressed in normal bone and teeth [34–36], acts as a mineralization inhibitor [37,38]. In addition, a phosphaturic effect has been published after injection of MEPE into mice [37]. Elevated levels of a proposed bioactive fragment of MEPE (ASARM peptide) were reported in nine XLH patients but also in Hyp mice, which represented a murine XLH disease model. For a detailed review on the properties of MEPE and its proposed role in XLH and TIO, see reference [3]. Thus far, little is known about the general MEPE function in bone. The expression of the murine MEPE homologue, osteoblast/osteocyte factor-45, in osteoblast differentiation and mineralization was studied in both primary osteoblast cultures and different cell lines in mouse, rat and human with, in part, conflicting results [34,36,39,40]. However, immunohistochemistry and in situ hybridization studies on adult rat tibia and developing mandible, respectively, revealed predominant expression in mature bone matrix-associated osteoblasts and osteocytes [36,40]. Studies on bone fracture healing demonstrated the highest MEPE expression in the latest stages of bone healing, here again strongest in bone embedded osteocytes, which might be consistent with an inhibitory effect on mineralization [41]. The MEPE serum levels were shown to decrease with age and correlate with serum phosphate, PTH and bone mineral density of the hip, femur and neck [42].

The sFRP4 is a very recently identified protein highly expressed in TIO tumours [33]. Moreover, sFRP4 causes phosphaturia and hypophosphatemia in vivo and it can be measured in serum [43]. The secreted frizzled-related proteins are extra-cellular inhibitors of Wnt signalling, reviewed in reference [44]. Although Wnt signalling is involved in bone turnover and osteoblast function [45], there are currently no data about a physiologic function of sFRP4 in bone.

Shortly after mutations in FGF-23 were recognized as being responsible for ADHR, its high expression in TIO tumours was reported by several authors [46–48]. Significant evidence has been accumulated showing that this factor has the properties to account for the major disease features that characterize the phosphate wasting disorders. Injecting recombinant FGF-23 into rodents causes hypophosphatemia, renal phosphate loss independent of PTH and suppresses serum 1,25(OH)2D3, but leaves serum calcium unchanged [49,50]. The FGF-23 transgenic mice largely reproduce the biochemical abnormalities and defective bone mineralization [51,52]. It was hypothesized that in XLH, mutated PHEX might fail to cleave FGF-23 and the increased concentrations would then cause the abnormalities typical for the disease. Although a possible cleavage of FGF-23 by PHEX was shown in some experimental settings [47,53], other studies argue against this possibility [29,54,55]. However, FGF-23 serum levels are increased in a subset of TIO, XLH and also fibrous dysplasia patients [26,56,57]. Weber et al. [58] could not substantiate significantly elevated FGF-23 concentrations in their group of XLH patients but they established a significant negative correlation of serum FGF-23 with both serum phosphate and calcium phosphate product in XLH patients but not in healthy controls. Comparable correlations were found in fibrous dysplasia patients with renal phosphate wasting [26]. The increase in urinary phosphate excretion by FGF-23 was associated with reduced expression of the sodium-dependent phosphate transporter NaPi IIa, which is the major key-player in renal phosphate re-absorption, reviewed in references [50,59]. However, it was not known whether this was a directly mediated effect of FGF-23 on the kidney.

When our understanding of phosphate wasting disorders pointed to FGF-23 being important in phosphate homeostasis, many research groups were interested whether FGF-23 could also be involved in other conditions with defective phosphate regulation. End stage renal disease (ESRD) is associated with complex derangements of calcium and phosphate balance. Patients suffer not only from different forms of bone disease commonly referred to as renal osteodystrophy but also tend to develop hyperphosphatemia and secondary hyperparathyroidism. The following summarizes recent emerging evidence that FGF-23 might take part in the complex pathophysiology of these disorders.

Fibroblast growth factor-23 in health and disease

Years before any phosphatonin candidate was cloned, it had been shown in vitro that dialysates derived from ESRD patients contained a phosphaturic factor which acted differently from that of PTH [60]; however, the identity of this factor remains unknown.

The FGF-23 is variably elevated in chronic kidney disease (CKD) and ESRD [58,61–64]. The gradual decrease in renal function is associated with a corresponding increase in serum FGF-23 as indicated by documented correlations of FGF-23 with either serum creatinine or creatinine clearance [61,63]. Moreover, these studies reported significant relationships of FGF-23 to phosphate, calcium phosphate product, PTH and 1,25(OH)2D3 in declining renal function. Before discussing these studies in more detail a methodological aspect should be mentioned: as previously described, bio-active full length FGF-23 (30–32 kDa) is cleaved into an N-terminal (16 kDa) and a C-terminal (10–12 kDa) fragment, both presumably inactive. Three studies [58,61,62] employed a FGF-23 ELISA which recognized a C-terminal epitope. Accordingly, both full length and C-terminal fragments were measured. Fukagawa et al. used a different ELISA which required the simultaneous presence of both the N-terminal and C-terminal ends (full length FGF-23) [63,64]. Weber et al. [58] estimated by Western blots of concentrated serum that the C-terminal fragment accounts for the majority of total FGF-23 present in serum of ESRD patients, which is supported by Imanishi et al. [62], who compared the results of both assays in a subset of ESRD patients and healthy controls. They measured a seven to tenfold higher abundance of C-terminal FGF-23 compared with full length protein, a proportion which did not differ significantly in both groups. This may explain why the major findings were largely comparable in these studies, regardless of which type of ELISA was used. This simplification might at least apply in the setting of progressive renal failure.

In a small analysis of ESRD patients on haemodialysis both phosphate and calcium × phosphate product were significantly positively correlated with FGF-23 [58]. In multiple regression analysis the calcium × phosphate product was the best predictor of serum FGF-23. None of these correlations were evident in normal subjects. Imanishi et al. [62] found positive correlations for phosphate, calcium and also PTH in a larger sample including ESRD, nondialyzed uraemic patients and healthy controls. These factors were also independently associated with FGF-23 in multiple regression. Sato et al. [64] studied an ESRD patient group with marked secondary hyperparathyroidism. Their findings are consistent with the previously mentioned, with the strongest correlation of calcium × phosphate product with FGF-23.

Larsson et al. [61] analyzed three patient groups with different severity of renal function impairment. In the whole study population FGF-23 correlated positively with phosphate, calcium, PTH and creatinine. Moreover, logarithmically transformed FGF-23 correlated negatively also with 1,25(OH)2D3. Within the group with the best renal function in the study population (kidney transplant recipients) no significant correlations with any parameter were evident, comparable to findings in healthy controls [58]. A negative correlation of 1,25(OH)2D3 and serum FGF-23 was also noted by other authors [63]. This group also demonstrated that, in spite of worsening of renal function with impaired total urinary phosphate excretion, Tmp/GFR was decreased in the presence of high FGF-23 [63].

What information can be drawn concerning the role of FGF-23 in CKD from this outline of several studies with, in large, similar findings? Are elevated serum levels in CKD part of a physiological response to altered bone and mineral homeostasis, or do these changes simply reflect the result of decreased clearance of either full length FGF-23 or its fragments when the renal function deteriorates? Indeed, accumulation appears to be relevant because it has been shown that urine from an ESRD patient on dialysis contains FGF-23 immunoreactive material [61]. Faint bands have also been observed in healthy controls indicating that FGF-23 is cleared by the kidney. However, from the following evidence it will be shown that serum FGF-23 varies under defined conditions in which renal function impairment is absent. This renders decreased FGF-23 clearance unlikely to be the sole cause for its rise in CKD serum.

Fibroblast growth factor-23 and parathyroid hormone

Secondary hyperparathyroidism is commonly found in XLH patients. Phosphate supplementation is known to be a major cause, but this complication may also arise in untreated patients – reviewed in reference [10]. Is FGF-23 capable of causing secondary hyperparathyroidism? Elevated serum PTH and hyperplastic parathyroid glands were observed in mice over-expressing a mutated form of FGF-23 which had been previously described in ADHR patients and is known to exhibit increased biological activity (see above). This was reproducible by using two different methods of gene over-expression [24,65]. In one of these models serum calcium was also mildly decreased compared with wild-type mice. Both increased serum PTH and parathyroid gland hyperplasia but unchanged serum calcium was found in one study using wild-type FGF-23 transgenic mice [52]. Other authors have reported opposite findings in terms of serum PTH, using either a wild-type FGF-23 transgene [51] or naked DNA injection [66]. The reason for the opposing findings is unclear. It is possible, however, that the quantitative amount of FGF-23 that had been induced in the different experimental approaches and increased bio-activity of mutant FGF-23 play some role.

In CKD patients PTH positively correlates with FGF-23 (see above). This is especially true when secondary hyperparathyroidism had developed, in which FGF-23 was considerably elevated [63]. Further support for the pathophysiological link with PTH comes from the observation that elevated FGF-23 in secondary hyperparathyroidism declines following parathyroidectomy [64]. In humoral hypercalcaemia of malignancy patients, who exhibited increased serum parathyroid-hormone related protein concentrations, significant elevations in FGF-23 were reported [67]. Several studies assessed a possible role of FGF-23 in primary hyperparathyroidism [67–70]. In these patients FGF-23 was neither significantly elevated nor any correlation to calcium, phosphate, PTH or 1,25(OH)2D3 could be found. However, it has to be considered that the increases in PTH that have been reported are quantitatively small compared with those that occur in renal secondary hyperparathyroidism (approximately twofold in primary hyperparathyroidism versus tenfold and higher, in secondary hyperparathyroidism). Remarkably, patients suffering from hypoparathyroidism who have increased serum phosphate levels exhibited elevated FGF-23 in one study [70]. This was in accordance with animal experiments in rats [71]. Although thyroparathyroidectomy (TPTX) decreased serum FGF-23 to a lower baseline level compared with intact rats, 1,25(OH)2D3 injections raised FGF-23 in both intact and TPTX rats. Summarizing this evidence, it can be concluded that FGF-23 might be regulated in concert with PTH, but elevated PTH is apparently not a prerequisite for elevated FGF-23.

Fibroblast growth factor-23 and 1,25(OH)2 vitamin D3

As mentioned in the introduction, FGF-23 not only increases renal phosphate excretion but also suppresses serum 1,25(OH)2D3. Both transcriptional and post-transcriptional events are implicated in exerting the latter effect [50,65]. It has been specified that the suppressive effect on 1,25(OH)2D3 after injection of recombinant FGF-23 precedes the phosphaturic response by several hours. Moreover, the doses required to elicit phosphaturia were higher than those that lowered 1,25(OH)2D3[50]. Consistent with these observations that come from animal experiments are the negative correlations of 1,25(OH)2D3 and FGF-23 in CKD, as outlined above.

The FGF-23 not only down-regulates serum 1,25(OH)2D3 but, in turn, itself appears to be under the control of 1,25(OH)2D3. After injection of 1,25(OH)2D3 into mice dose-dependent increases in serum FGF-23 were observed [50]. This rise was accompanied by elevations in serum phosphate. Such interplay might be regarded as a negative feedback loop. The physiologic relationship of vitamin D and FGF-23 has recently been studied in rats [71]. Intravenous injection of 1,25(OH)2D3 raised serum FGF-23 in both intact and thyroparathyroidectomized rats. In the intact rats, increasing FGF-23 after 1,25(OH)2D3 injections were accompanied by increases in serum phosphate. In equally treated TPTX rats, which are hyperphosphatemic and hypocalcaemic at baseline, serum phosphate dropped and calcium increased in the face of rising FGF-23. This indicated that 1,25(OH)2D3 was capable of regulating FGF-23 independently of serum phosphate. Saito et al. [71] affirmed the possible role of vitamin D by their studies on the vitamin D receptor knockout mouse (VDRKO), which on a high calcium and phosphate diet had very low serum FGF-23, which remained unaffected by 1,25(OH)2D3 injections. These findings were, in part, also confirmed by others [72].

As low serum 1,25(OH)2D3 levels, which occur in ESRD, do not favour low FGF-23 levels in humans, the FGF-23–vitamin D interaction cannot be considered a simple mutual regulatory axis. The lack of such an interaction will be discussed in the following sections.

Fibroblast growth factor-23, phosphate and calcium

Healthy controls were consecutively subjected to equally directed changes in both the phosphate and calcium diets [73]. Restriction of these minerals led to decreases, whereas increased ingestion elicited a rise in serum FGF-23 levels, confirming that changes in FGF-23 are a physiologic response to altered mineral ion supply. Notably, PTH did not change throughout the experiment. The FGF-23 correlated negatively with TmP/GFR and 1,25(OH)2D3. Although diet-induced changes in serum 1,25(OH)2D3 were too low in the experiments by Ferrari et al. [73] to attain statistical significance, it is well known that low serum calcium and phosphate stimulate 1,25(OH)2D3 synthesis. In this setting, however, FGF-23 was decreasing, which suggests that phosphate and perhaps calcium have an important impact on FGF-23 regulation. Of note, sole phosphate deprivation and loading did not significantly change either PTH or serum FGF-23 in a small study on healthy controls, although the kidney readily adapted phosphate re-absorption to keep serum phosphate constant [61]. This indicates that adaptation to isolated changes in phosphate diet in human subjects do not depend on FGF-23. In mice, however, changes in phosphate diet were reported to alter serum FGF-23 concomitantly with corresponding changes in serum phosphate [72].

Impaired renal function in 5/6-nephrectomized rats relates serum phosphate closely to changes in dietary phosphate. In these animals, high dietary phosphate considerably increased serum FGF-23, along with elevated serum phosphate and decreased serum calcium [71]. The 1,25(OH)2D3 was not measured in this setting but can be expected to be lower on a high phosphate diet. Superimposed 1,25(OH)2D3 injections into these rats dramatically increased serum FGF-23. Phosphate and calcium in serum were not measured [71]. Experimental support for phosphate as a possible regulator of FGF-23 comes from cell culture studies on fetal bone cells [27]. Increased phosphate concentration in culture media stimulated FGF-23 mRNA expression, suggesting phosphorus to be a direct regulator of FGF-23 synthesis in bone. Additional evidence for the direct effect of both phosphate and Vitamin D on FGF-23 gene transcription is provided by studies on the FGF-23 promoter region [72].

Injection of recombinant FGF-23 did not alter serum calcium [49,50]. Slight hypocalcaemia is described in some transgenic mice carrying wild-type FGF-23 and a variant form containing an ADHR mutation [51,65]. Conversely, hypercalcaemia is reported in the FGF-23 null mouse [74]. These findings might be related to corresponding changes in serum 1,25(OH)2D3 that occurred in these experiments. As mentioned earlier, correlations of calcium or calcium × phosphate product and FGF-23 in CKD have been reported. It is possible that this is owing to the association of high calcium with secondary hyperparathyroidism. However, it cannot be fully excluded that calcium modulates the impact of phosphate and vitamin D on FGF-23 serum levels.

Fibroblast growth factor-23 – a role in chronic renal failure?

It is known that FGF-23 serum levels rise with a gradual decrease in renal function. From experimental studies, a lot has been learned about the biological effects that FGF-23 can exert, but is this of pathophysiological significance in chronic renal failure? If so, what are the consequences that can be anticipated? The notion that FGF-23 is a substance that causes phosphaturia suggests that increasing FGF-23 is beneficial in counteracting phosphate overload in CKD. The FGF-23, on the other hand, lowers 1,25(OH)2D3, an abnormality that is frequently found in CKD. Additionally, several groups have established that over-expression of FGF-23 leads to defective bone mineralization [24,51,52,65]. Although the mechanism is unknown, it is reasonable to assume that elevated FGF-23 in CKD could play a yet undefined role in renal osteodystrophy. As discussed above, FGF-23 is perhaps also implicated in the development of secondary hyperparathyroidism. Taken together, these mechanisms would rather suggest FGF-23 to act as a uraemic toxin. Moreover, it has been shown that an injection of 1,25(OH)2D3 into 5/6-nephrectomized rats increased serum FGF-23 [71]. Such increases have also been reported to occur in vitamin D-treated dialysis patients [75]. Is this an adverse effect of vitamin D treatment? Recently, two studies proposed FGF-23 as a possible clinical marker in CKD patients [75,76]. Determining serum FGF-23, especially in combination with serum PTH, might aid in assessing the severity of secondary hyperparathyroidism and in identifying those patients who are more likely to be refractory to treatment with vitamin D. However, more studies will be necessary to clearly define the potential benefit from FGF-23 measurements in CKD (Fig. 1).

Figure 1.

Fibroblast growth factor-23 (FGF-23) in chronic kidney disease. A hypothetical implementation of FGF-23 in pathophysiological aspects of chronic kidney disease and secondary hyperparathyroidism is shown, where FGF-23 is elevated owing to its decreased renal clearance and as a response to phosphate retention. Elevated FGF-23 contributes to suppressed 1,25(OH)2D3 serum levels and perhaps also directly to osteomalacia in renal osteodystrophy. Vitamin D treatment improves the elevated PTH but further increases FGF-23. PTH, parathyroid hormone; Pi, serum phosphate; FGF-23, fibroblast growth factor-23; 1,25(OH)2D3, 1,25-dihydroxy vitamin D3.

New loss of function mutation in fibroblast growth factor-23

As described earlier, particular FGF-23 mutations cause ADHR. These mutations are known to increase bio-activity of the protein. Recently, another FGF-23 mutation has been described in a disease with biochemical and clinical features which are diametrically opposed to those found in ADHR. Two studies reported the occurrence of a homozygous missense mutation in patients suffering from familial tumoural calcinosis (FTC) [77,78]. Patients with FTC exhibit metastatic calcifications that form periarticular tumoural masses around the shoulders, hips, or ankles. Biochemical findings comprise hyperphosphatemia, elevated serum 1,25(OH)2D3 levels but usually normal serum calcium and PTH levels. Experimental studies indicate that the described FGF-23 mutation most likely leads to trapping of mutant FGF-23 protein in the Golgi complex with predominant secretion of a C-terminal FGF-23 fragment [78]. Measurement of serum FGF-23 by means of an ELISA recognizing a C-terminal epitope revealed highly elevated FGF-23 levels. Another group that reported the same mutation used both ELISA systems that recognized either C-terminal and intact, or, only intact FGF-23 levels [77]. The FGF-23 was low normal when measured with the intact assay, whereas the C-terminal assay detected markedly elevated FGF-23. This might indicate up-regulated FGF-23 synthesis of a presumably nonfunctional protein which is not recognized by the intact assay. Familial tumoural calcinosis, which is an autosomal recessive disorder, can also be caused by mutations in the GALNT3 gene coding for UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyl-transferase 3, an enzyme involved in O-glycosylation, a form of post-translational modification [79,80]. The FGF-23 was found to be elevated in some of these FTC patients, although the kind of ELISA used was not specified by the authors [79].

Concluding remarks

Although this review focused on recently established data on FGF-23 and its possible role apart from renal phosphate wasting disorders, it is evident that FGF-23 is not the only factor that may act as a phosphatonin. Both TIO and XLH patients can have normal FGF-23 serum levels. The physiological role of MEPE and sFRP4 (see above) is not yet clarified to an extent that is comparable to FGF-23. The diversity is illustrated by the recent identification of another phosphatonin candidate, fibroblast growth factor-7 (FGF-7), which also derives from TIO tumour tissue. Conditioned media from FGF-7 expressing tumour cells inhibited phosphate transport in opossum kidney cells but only contained low levels of FGF-23 [81]. The PHEX, an endopeptidase mutated in XLH patients, has also been shown to be higher expressed in the 5/6-nephrectomy rat model. In contrast, 1,25(OH)2D3 had the opposite effect [82]. All these findings underline the complexity of bone and mineral homeostasis and their tight interaction. The renal phosphate wasting disorders will in the future bring us closer to a full picture of the underlying biological events that are difficult to conceive owing to the diverse interactions of both known and unknown factors.


The author wishes to express his gratitude to the Swiss National Fund for supporting this post-graduate course in experimental medicine and biology, the University of Zürich, Switzerland, and Jürgen Zapf, his supervisor.