Actions and Mode of Actions of FGF19 Subfamily Members

Endocrine Journal 2008, 55 (1), 23–31

REVIEW

SEIJI FUKUMOTO

Division of Nephrology & Endocrinology, Department of Medicine, University of Tokyo Hospital, Tokyo 113-8655, Japan

Abstract. Fibroblast growth factors (FGFs) are humoral factors with diverse biological functions. While most FGFs were shown to work as local factors regulating cell growth and differentiation, recent investigations indicated that FGF19 subfamily members, FGF15/19, FGF21 and FGF23 work as systemic factors. FGF15/19 produced by intestine inhibits bile acid synthesis and FGF21from liver is involved in carbohydrate and lipid metabolism. In addition, FGF23 was shown to be produced by bone and regulate phosphate and vitamin D metabolism. Furthermore, these FGFs require klotho or βklotho for their actions in addition to canonical FGF receptors. It is possible that these FGFs together with their receptor systems might be targets for novel therapeutic measures in the future.

Key words: Hypophosphatemia, Hyperphosphatemia, Klotho, FGF (Endocrine Journal 55: 23–31, 2008)

FIBROBLAST growth factors (FGFs) are humoral FGF23 factors with diverse biological functions. It has been proposed that there are 22 members of FGF family in Identification and action of FGF23 human (FGF1 to FGF14, and FGF16 to FGF23) and these family members are divided into several subfam- FGF23 is the latest member of FGF family. We ilies [1] (Table 1). However, 4 members, FGF11 to have reviewed FGF23 as a responsible factor for hypo- FGF14, have been shown to be unable to bind to and phosphatemic rickets/osteomalacia in 2001 [4]. Since activate FGF receptors (FGFRs) [2]. These factors are then, significant progress has been made in the under- also called as FGF homologous factors and discrimi- standing of physiological and pathophysiological im- nated from other members of FGF family by some re- portance of FGF23. FGF23 was identified almost searchers [3]. Most FGFs have been shown to work as simultaneously by three groups by different approaches. local factors regulating cell growth and differentiation. FGF23 was first cloned by homology to FGF15, a mu- In contrast, recent findings concerning actions and rine orthologue of human FGF19, in mice [5]. The mode of actions of FGF19 subfamily members clearly same group cloned human FGF23 as well. FGF23 expanded our knowledge about the biological signifi- was also cloned as a responsible gene for autosomal cance of FGF family members. In this review, charac- dominant hypophosphatemic rickets/osteomalacia teristics of FGF19 subfamily members are summarized (ADHR) by positional cloning [6] and as a causative with special emphasis on FGF23. humoral factor for tumor-induced rickets/osteomalacia (TIO) [7]. Rickets/osteomalacia is characterized by impaired mineralization of bone matrix. There are many causes for rickets/osteomalacia as shown in Ta- ble 2. Of these, several hypophosphatemic diseases Received: August 4, 2007 are known to share common clinical features. Those Accepted: August 5, 2007 Correspondence to: Dr. Seiji FUKUMOTO, Division of Neph- include ADHR, autosomal recessive hypophosphatemic rology & Endocrinology, Department of Medicine, University of rickets/osteomalacia (ARHR), X-linked hypophos- Tokyo Hospital, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, phatemic rickets/osteomalacia (XLH), TIO and hypo- Japan phosphatemic rickets/osteomalacia associated with 24 FUKUMOTO

Table 1. FGF family members Table 2. Causes of rickets/osteomalacia Causative genes were described in brackets. Diseases FGF1 subfamily caused by excess actions of FGF23 are underlined. FGF1 (Acidic FGF) Although McCune-Albright syndrome is not a herita- FGF2 (Basic FGF) ble disease, its genetic cause has been clarified. FGF4 subfamily FGF4 Genetic causes FGF5 X-linked hypophosphatemic rickets/osteomalacia (XLH) FGF6 [PHEX] FGF7 subfamily Autosomal recessive hypophosphatemic rickets/osteomalacia FGF3 (ARHR) [DMP1] FGF7 Autosomal dominant hypophosphatemic rickets/osteomalacia FGF10 (ADHR) [FGF23] FGF22 Hereditary hypophosphatemic rickets/osteomalacia with FGF8 subfamily hypercalciuria (HHRH) [SLC34A3] FGF8 FGF17 Vitamin D dependency type I [CYP27B1] FGF18 Vitamin D dependency type II [VDR] FGF9 subfamily Selective 25-hydroxyvitamin D deficiency [CYP2R1] FGF9 Hypophosphatasia [Alkaline phosphatase] FGF16 FGF20 Renal tubular diseases (Dent's disease [CLC5], distal renal FGF11 subfamily tubular acidosis [SLC4A1] ...) FGF11 (FHF*3) McCune-Albright syndrome/Fibrous dysplasia [GNAS] FGF12 (FHF1) Acquired causes FGF13 (FHF2) Vitamin D deficiency FGF14 (FHF4) Malabsorption (malnutrition, gastrointestinal diseases ...) FGF19 subfamily FGF19 Chronic renal failure FGF21 Drugs (anticonvulsant, aluminum ...) FGF23 Tumor-induced rickets/osteomalacia (TIO) * FHF: FGF homologous factor Renal tubular diseases PHEX: phosphate-regulating gene with homologies to endopeptidases on the X chromosome DMP1: dentin matrix protein 1 McCune-Albright syndrome (MAS)/fibrous dysplasia VDR: vitamin D receptor (FD). These diseases are characterized by rickets/os- CLC: chloride channel teomalacia and hypophosphatemia at least in part due to GNAS: guanine nucleotide-binding protein G-stimulatory impaired proximal tubular phosphate reabsorption. subunit alpha Hypophosphatemia is known to stimulate 1,25-dihydroxyvitamin D [1,25(OH)2D] production and in- more, FGF23 was shown to reduce the expression crease serum 1,25(OH)2D level. However, serum level of 25-hydorvitamin D-1α-hydroxylase which 1,25(OH)2D remains to be low to low normal indicating that both phosphate reabsorption and vitamin D mediates the synthesis of 1,25(OH)2D. In addition, metabolism are deranged in these hypophosphatemic FGF23 increases the expression of 25-hydorvitamin diseases. D-24-hydroxylase, one of whose function is to degrade After cloning of FGF23, actions of FGF23 were ex- 1,25(OH)2D into more hydrophilic metabolites [8]. amined in vivo using recombinant FGF23. FGF23 was By altering the expression levels of these enzymes shown to suppress expression levels of type 2a and 2c involved in vitamin D metabolism, FGF23 reduces sodium-phosphate cotransporters in proximal tubular serum 1,25(OH)2D level. Because 1,25(OH)2D stimu- cells [8]. These transporters mediate physiological lates intestinal phosphate absorption, FGF23 reduces phosphate reabsorption and proximal tubular phos- serum phosphate level by inhibiting proximal tubular phate reabsorption is the main determinant of serum phosphate reabsorption as well as suppressing intesti- phosphate level at least in a chronic state. Further- nal phosphate absorption through decreased circulatory level of 1,25(OH)2D. ENDOCRINE FIBROBLAST GROWTH FACTORS 25

FGF23 and hypophosphatemic rickets/osteomalacia controls [19]. These results could be explained by regulatory production of FGF23. Several studies indi- As expected from actions of FGF23 in vivo, ADHR, cated that FGF23 production is tightly regulated. For ARHR, XLH, TIO and hypophosphatemic rickets/os- example, low phosphate diet lowers and high phos- teomalacia in MAS/FD have been shown to be caused phate diet increases serum FGF23 [20]. 1,25(OH)2D by excess actions of FGF23 (Table 2). Actually, the also increases circulatory FGF23 [21] and hyperphos- establishment of enzyme-linked immunosorbent assay phatemia was shown to be associated with high FGF23 for FGF23 indicated that circulatory levels of FGF23 [22]. Therefore, it may not be surprising that FGF23 are high in most of these patients [9–11]. The respon- levels are not high in patients with ADHR because hy- sible genes for XLH and ARHR are phosphate-regu- pophosphatemia or other associated metabolic changes lating gene with homologies to endopeptidases on the suppress expression of FGF23 even though mutant X chromosome (PHEX) [12] and dentin matrix protein FGF23 protein is resistant to the processing. In con- 1 (DMP1) [13, 14], respectively. Although the precise trast, it has been also shown that serum phosphate mechanism is unknown, model mice for XLH and level changes with time and high FGF23 level is asso- DMP1 knockout mice show enhanced expression of ciated with hypophosphatemia in patients with ADHR FGF23 in bone [13–15]. These results suggest that [19]. Therefore, it is likely that regulatory mecha- PHEX and DMP1 proteins are involved in the regula- nisms of FGF23 production are somehow deranged in tion of FGF23 expression in bone. Similarly, overex- these patients although the relationship between muta- pression of FGF23 in bone including regions affected tions in FGF23 gene and the regulation of FGF23 pro- by FD has been reported in patients with MAS/FD, duction is unclear. although the relationship between mutations in GNAS gene and overproduction of FGF23 remains unclear Physiological role of FGF23 [10]. TIO has been shown to be caused by overproduction of FGF23 in tumors responsible for TIO [7]. In addition to participate in the development of sev- Therefore, ARHR, XLH, hypophosphatemic rickets/ eral hypophosphatemic diseases, FGF23 is a physio- osteomalacia in MAS/FD and TIO are caused by ex- logical regulator of serum phosphate and 1,25(OH)2D cess production of FGF23 either in bone or tumors levels. FGF23 knockout mice show hyperphos- causing TIO. On the other hand, several studies have phatemia, enhanced proximal tubular phosphate reab- indicated that bone, especially osteocyte, is a physio- sorption, increased 1,25(OH)2D production and high logical source of FGF23 in addition to making excess 1,25(OH)2D levels [16, 23]. These features are mirror amount of FGF23 in several hypophosphatemic dis- images of those in hypophosphatemic diseases caused eases [16]. by excess actions of FGF23. Furthermore, there is a ADHR was shown to be caused by missense muta- human disease which shows similar features to those tions in FGF23 gene [6]. FGF23 gene produces a proof FGF23 knockout mice. Tumoral calcinosis is char- tein with 251 amino acids including a signal peptide acterized by ectopic calcification especially around with 24 amino acids. A part of FGF23 protein is pro- large joints. Tumoral calcinosis is most frequently teolytically cleaved between Arg179 and Ser180 before seen in patients with chronic renal failure undergoing secretion and these cleaved fragments do not have bio- dialysis. In addition, there is another form of tumoral logical activity to reduce serum phosphate level [17]. calcinosis from a genetic origin. Familial hyperphos- This cleavage is believed to be mediated by furin that phatemic tumoral calcinosis is characterized by hyper- 176 177 178 179 recognizes Arg -X -X -Arg sequence just be- phosphatemia and rather high 1,25(OH)2D levels [24]. fore the processing site [18]. Mutations in FGF23 Two genes, GALNT3 and FGF23, have been shown to gene causing ADHR are observed either in Arg176 or be responsible for this disease and several homozy- Arg179 in the sequence recognized by furin. Therefore gous mutations are reported in these genes [25–28]. it has been postulated that mutations in FGF23 gene GALNT3 gene encodes a protein called UDP-N-acetyl- result in impaired processing of FGF23 protein and in- alpha-D-galactosamine: polypeptide N-acetylgalac- crease circulatory full-length FGF23. However, recent tosaminyltransferase 3 (ppGaNTase-T3) which attaches investigation indicated that average FGF23 levels in N-acetylgalactosamine to serine or threonine residue patients with ADHR are not different from those of as an initial sugar of mucin-type O-glycosylation. 26 FUKUMOTO

There are several methods for measuring FGF23. The intact assay uses two monoclonal antibodies that recognize N-terminal and C-terminal portion of the processing site of FGF23 and measures only full- length FGF23 [11]. On the contrary, C-terminal assay uses two kinds of polyclonal antibodies against C- terminal portion of the processing site of FGF23 [9]. This C-terminal assay seems to detect both full-length and processed C-terminal fragment of FGF23. It has been reported that FGF23 levels evaluated by these two assays correlate well in patients with XLH [29]. However, it was shown that these two assays produce clearly discrepant FGF23 values in patients with hyperphosphatemic tumoral calcinosis caused by mutations of either GALNT3 or FGF23 gene [25, 30]. Fig. 1. Pathogenesis of familial hyperphosphatemic tumoral While FGF23 measured by C-terminal assay was quite calcinosis. high, FGF23 was low to low normal when assessed by Mutations in either GALNT3 or FGF23 gene induce intact assay. These results suggested that there is only enhanced processing of full-length FGF23 and resultant small amount of, if any, full-length FGF23 in the face rather low full-length FGF23 causes hyperphosphatemia of large amount of processed C-terminal fragment, and and high 1,25(OH)2D. Although the precise regulatory mechanisms of FGF23 production remain to be clari- therefore the processing and the production of FGF23 fied, these metabolic changes seem to stimulate FGF23 protein are enhanced in these patients. Actually, production and together with enhanced processing of FGF23 has three O-linked glycan chains and the at- FGF23 protein increase circulatory level of processed tachment of O-glycan to Thr178 was shown to be medi- C-terminal fragment of FGF23. ated specifically by ppGaNTase-T3 [30, 31]. The attachment of this O-glycan in the vicinity of the pro- though mucin-type O-glycans are involved in diverse cessing site seems to impair the cleavage of FGF23 functions of various proteins, patients with mutations protein probably by preventing furin from accessing in GALNT3 gene show phenotypes mainly restricted to the processing site. Detailed mechanisms how muta- abnormal bone and mineral metabolism. This could be tions in FGF23 gene cause enhanced processing of explained by redundant substrate specificity of various FGF23 protein are unknown. From these data it has ppGaNTase proteins in the process of O-glycan syn- been proposed that mutations in either GALNT3 or theses for proteins other than FGF23. However, it is FGF23 gene induce enhanced processing of full-length unknown why the attachment of O-glycan to Thr178 FGF23 and resultant rather low full-length FGF23 can not be mediated by other ppGaNTases. causes hyperphosphatemia and high 1,25(OH)2D. These metabolic changes stimulate FGF23 production Mode of actions of FGF23 and together with enhanced processing of FGF23 protein increase circulatory level of processed C-terminal FGF23 is produced by bone and effects of FGF23 fragment of FGF23 (Fig. 1). These results indicate are observed in kidney. Therefore, there must be a that FGF23 action is essential for the maintenance of specific receptor system for FGF23 in kidney. In an serum phosphate and 1,25(OH)2D levels in human. attempt to find molecules that bind to FGF23 in kid- There are more than twenty ppGaNTase proteins. ney, klotho was identified as a major binding protein Familial hyperphosphatemic tumoral calcinosis is so to FGF23 [33]. Klotho is a type I membrane protein far the only disease known to be caused by mutations and has a single membrane-spanning region with a encoding these ppGaNTases. Furthermore, while short intracellular tail [34]. The expression of klotho there are several diseases called congenital disorders is observed in several restricted tissues including kid- of glycosylation (CDG), familial hyperphosphatemic ney, parathyroid and pituitary glands. Klotho has a tumoral calcinosis is the only disease caused by im- homologous protein called βklotho [35]. Klotho and paired synthesis of mucin-type O-glycan [32]. Al- βklotho have two β-glucosidase-like domains in the ENDOCRINE FIBROBLAST GROWTH FACTORS 27 extracellular regions. However, it has been unclear whether klotho and βklotho actually work as β-glucosidase. Klotho mice were created by transgenic method and have been shown to have severely reduced expression level of klotho protein. Klotho mice show several phenotypes resembling senescence such as short life span, growth retardation and atrophy of skin [34]. Interestingly, these features are also observed in FGF23 knockout mice. Furthermore, klotho mice were shown to have abnormal mineral metabolism. Serum phosphate and 1,25(OH)2D levels are high in Fig. 2. Mode of actions of FGF19 subfamily members. klotho mice as in FGF23 knockout mice [36]. In addi- FGF19 subfamily members work as endocrine FGFs. tion, FGF23 level in klotho mice was shown to be ex- These members require klotho or βklotho in addition to tremely high [33]. These results suggested that FGF23 canonical FGFRs for their actions. Representative target tissues and FGFRs are described. and klotho work in the same signal transduction pathway and klotho is in the downstream of FGF23. Several in vivo and in vitro studies indicated that FGF19 klotho actually binds to FGF23 and FGF23 can induce second messenger such as phosphorylation of extracel- FGF19 is a human orthologue of murine FGF15 lular signal regulated kinase (ERK) and induction of and was cloned by homology to mouse FGF15 [38]. early growth responsive-1 (Egr-1) in the presence of FGF19 gene produces a peptide with 216 amino acids klotho [33, 37]. However, because intracellular do- including a signal peptide with 22 amino acids. Initial main of klotho is very short, it was unclear whether study indicated that FGF19 is expressed only in brain klotho alone could transduce intracellular signals. Ac- [38]. However, subsequent study showed that FGF15/ tually, FGF23 could not invoke second messenger in 19 is produced by intestine and is involved in bile acid L6 cells expressing klotho indicating that klotho is not synthesis and gallbladder filling. FGF15 knockout sufficient for FGF23 signaling [33]. This L6 cell line mice was reported to show enhanced bile acid synthe- is well known to be devoid of FGFRs. Therefore, the sis due to high expression level of cholesterol 7alpha- involvement of FGFRs in FGF23 signaling was exam- hydroxylase (CYP7A1) in liver, the key rate-limiting ined. Although somewhat different results concerning enzyme for the synthetic pathway of bile acid [39]. the specificity of FGFRs have been reported, klotho These results indicate that FGF15 inhibits the expres- was found to bind to some FGFRs and convert canoni- sion of CYP7A1 and thereby bile acid synthesis. In cal FGFR into a specific receptor for FGF23. There contrast, bile acid stimulates the expression of FGF15 are four genes encoding FGFRs, FGFR1 to FGFR4, in distal intestine. Therefore, there seems to be a nega- and alternative splicing form these genes produce sev- tive feedback system for bile acid synthesis through eral subtypes of FGFRs. Our results indicated that FGF15. It has been also reported that FGFR4 knock- FGF23 acts on klotho-FGFR1c [33], but another report out mice show similar phenotypes to those of FGF15 showed that klotho can interact with FGFR3c and 4 in knockout mice [39]. FGFR4 knockout mice exhibit addition to FGFR1c [37]. Therefore, klotho works as enhanced expression of CYP7A1 and bile acid synthe- a co-receptor for FGF23 together with FGFR in any sis. Therefore, it was postulated that FGF15 produced case (Fig. 2). These results explain the similarity of by intestine acts on liver through FGFR4. Interesting- phenotypes of klotho and FGF23 knockout mice and ly, it was reported that βklotho knockout mice also the specific effects of FGF23 on kidney at least in part. show enhanced bile acid synthesis suggesting that However, an important question remains. Klotho is βklotho is involved in FGF15/19 action [40]. Actu- expressed in distal convoluted tubules [34] and effects ally, it has bee shown that βklotho works as a co- of FGF23 are observed in proximal tubules. It is cur- receptor for FGF15/19 together with FGFR4 [41, 42] rently unknown how klotho-FGFR in distal convoluted (Fig. 2). tubules modifies phosphate reabsorption and vitamin FGF15 knockout mice also show contracted gall- D metabolism in proximal tubules. bladder and FGF15/FGF19 increases gallbladder vol- 28 FUKUMOTO ume [43]. In vitro study indicated that FGF15/FGF19 Actually, it has been shown that FGF19 subfamily relaxes smooth muscle in gallbladder. These results members have lower affinity to heparin than classical clearly indicate that FGF15/19 also works as a system- FGF members [49]. This low affinity to heparin prob- ic factor regulating bile acid synthesis and gallbladder ably helps these FGF19 subfamily members to be se- filling. creted to the systemic circulation. However, heparin enhances the interaction between classical FGF family members with FGFRs. Therefore, it is likely FGF19 FGF21 subfamily members need another factors that facilitate the interaction with FGFRs. Furthermore, there must Cloning of FGF21 indicated that FGF21 protein be specific response machineries to systemic factors. consists of 209 amino acids with a signal peptide with Klotho and βklotho seem to work to attain these two 28 amino acids [44]. FGF21 is mainly expressed in objectives. Although heparin actually stabilizes liver. FGF21 was shown to be involved in carbohy- FGF23-klotho-FGFR1c complex [33], FG23 binds to drate and fat metabolism. For example, injection of FGFRs very weakly without klotho. Tissue-specific FGF21 reduces blood glucose and triglyceride levels expression of klotho and βklotho together with selec- in ob/ob and db/db mice [45]. FGF15 transgenic mice tive usage of FGFR subtypes seem to determine tissue- exhibit lower fasting glucose level and are resistant to selective effects of FGF19 subfamily members. There diet-induced obesity. Furthermore, it has been recent- are several important reports describing biological ef- ly shown that FGF21 is involved in the response to fects of klotho [50–52]. However, it is reasonable to fasting [46, 47]. FGF23 induces several metabolic speculate that at least one of the main physiological changes including lipolysis, torpor and so on in re- functions of klotho is to work as a co-receptor for sponse to fasting mainly acting on fat tissues. These FGF23 considering phenotypic similarities of klotho effects of FGF21 were again reported to need βklotho and FGF23 knockout mice, and the involvement of and FGFR1c [41, 48] (Fig. 2). Therefore, it has been βklotho in actions of FGF19 and FGF21. proposed that tissue-specific expression of FGFR sub- Already two diseases are known to be caused by types together with limited expression of βklotho de- mutations in FGF23 gene, ADHR and familial hyper- termine the specific effects of FGF15/19 and 21 [41]. phosphatemic tumoral calcinosis. Given diverse effects of FGF19 subfamily members, it is not surprising to see other diseases caused by deranged actions of these Summary and future perspectives subfamily members in the future. More importantly, FGF19 subfamily members together with their recep- Results reported in the past several years indicate tor systems might be targets for novel therapeutic that FGF19 subfamily members act as systemic fac- measures. The discovery of actions and mode of ac- tors. These “endocrine” effects are clearly different tions of FGF19 subfamily members not only increased from classical local effects of other FGF family mem- the number of existing hormones and endocrine dis- bers. In order to work as systemic factors, these FGF eases but also deepened our understanding about the family members must have several characteristics. diversity of hormone actions and the mechanism of First, these factors can not be trapped in the extracellu- hormone actions. lar matrix surrounding cells producing themselves.

References

1. Itoh N, Ornitz DM (2004) Evolution of the Fgf and 3. Goldfarb M (2005) Fibroblast growth factor homolo- Fgfr gene families. Trends Genet 20: 563–569. gous factors: evolution, structure, and function. Cytok- 2. Olsen SK, Garbi M, Zampieri N, Eliseenkova AV, ine Growth Factor Rev 16: 215–220. Ornitz DM, Goldfarb M, Mohammadi M (2003) 4. Fukumoto S, Yamashita T (2001) Fibroblast growth Fibroblast growth factor (FGF) homologous factors factor (FGF)-23 and hypophosphatemic rickets/osteo- share structural but not functional homology with malacia. Endocr J 48: 603–610. FGFs. J Biol Chem 278: 34226–34236. 5. Yamashita T, Yoshioka M, Itoh N (2000) Identification ENDOCRINE FIBROBLAST GROWTH FACTORS 29

of a novel fibroblast growth factor, FGF-23, preferen- Quarles LD (2003) Regulation of fibroblastic growth tially expressed in the ventrolateral thalamic nucleus of factor 23 expression but not degradation by PHEX. J the brain. Biochem Biophys Res Commun 277: 494– Biol Chem 278: 37419–37426. 498. 16. Sitara D, Razzaque MS, Hesse M, Yoganathan S, 6. ADHR Consortium (2000) Autosomal dominant hypo- Taguchi T, Erben RG, Juppner H, Lanske B (2004) phosphataemic rickets is associated with mutations in Homozygous ablation of fibroblast growth factor-23 FGF23. The ADHR Consortium. Nat Genet 26: 345– results in hyperphosphatemia and impaired skeletogen- 348. esis, and reverses hypophosphatemia in Phex-deficient 7. Shimada T, Mizutani S, Muto T, Yoneya T, Hino R, mice. Matrix Biol 23: 421–432. Takeda S, Takeuchi Y, Fujita T, Fukumoto S, Yamashita 17. Shimada T, Muto T, Urakawa I, Yoneya T, Yamazaki T (2001) Cloning and characterization of FGF23 as a Y, Okawa K, Takeuchi Y, Fujita T, Fukumoto S, causative factor of tumor-induced osteomalacia. Proc Yamashita T (2002) Mutant FGF-23 responsible for Natl Acad Sci U S A 98: 6500–6505. autosomal dominant hypophosphatemic rickets is re- 8. Shimada T, Hasegawa H, Yamazaki Y, Muto T, Hino sistant to proteolytic cleavage and causes hypophos- R, Takeuchi Y, Fujita T, Nakahara K, Fukumoto S, phatemia in vivo. Endocrinology 143: 3179–3182. Yamashita T (2004) FGF-23 is a potent regulator of 18. Benet-Pages A, Lorenz-Depiereux B, Zischka H, vitamin D metabolism and phosphate homeostasis. J White KE, Econs MJ, Strom TM (2004) FGF23 is Bone Miner Res 19: 429–435. processed by proprotein convertases but not by PHEX. 9. Jonsson KB, Zahradnik R, Larsson T, White KE, Bone 35: 455–462. Sugimoto T, Imanishi Y, Yamamoto T, Hampson G, 19. Imel EA, Hui SL, Econs MJ (2007) FGF23 concentra- Koshiyama H, Ljunggren O, Oba K, Yang IM, Miyauchi tions vary with disease status in autosomal dominant A, Econs MJ, Lavigne J, Juppner H (2003) Fibroblast hypophosphatemic rickets. J Bone Miner Res 22: 520– growth factor 23 in oncogenic osteomalacia and X- 526. linked hypophosphatemia. N Engl J Med 348: 1656– 20. Ferrari SL, Bonjour JP, Rizzoli R (2005) Fibroblast 1663. growth factor-23 relationship to dietary phosphate and 10. Riminucci M, Collins MT, Fedarko NS, Cherman N, renal phosphate handling in healthy young men. J Clin Corsi A, White KE, Waguespack S, Gupta A, Hannon Endocrinol Metab 90: 1519–1524. T, Econs MJ, Bianco P, Gehron Robey P (2003) FGF- 21. Saito H, Maeda A, Ohtomo S, Hirata M, Kusano K, 23 in fibrous dysplasia of bone and its relationship to Kato S, Ogata E, Segawa H, Miyamoto K, Fukushima renal phosphate wasting. J Clin Invest 112: 683–692. N (2005) Circulating FGF-23 is regulated by 1alpha, 11. Yamazaki Y, Okazaki R, Shibata M, Hasegawa Y, 25-dihydroxyvitamin D3 and phosphorus in vivo. J Satoh K, Tajima T, Takeuchi Y, Fujita T, Nakahara K, Biol Chem 280: 2543–2549. Yamashita T, Fukumoto S (2002) Increased circulatory 22. Gupta A, Winer K, Econs MJ, Marx SJ, Collins MT level of biologically active full-length FGF-23 in (2004) FGF-23 is elevated by chronic hyperphos- patients with hypophosphatemic rickets/osteomalacia. phatemia. J Clin Endocrinol Metab 89: 4489–4492. J Clin Endocrinol Metab 87: 4957–4960. 23. Shimada T, Kakitani M, Yamazaki Y, Hasegawa H, 12. The HYP Consortium (1995) A gene (PEX) with ho- Takeuchi Y, Fujita T, Fukumoto S, Tomizuka K, mologies to endopeptidases is mutated in patients Yamashita T (2004) Targeted ablation of Fgf23 dem- with X-linked hypophosphatemic rickets. Nat Genet onstrates an essential physiological role of FGF-23 in 11: 130–136. phosphate and vitamin D metabolism. J Clin Invest 13. Feng JQ, Ward LM, Liu S, Lu Y, Xie Y, Yuan B, Yu 113: 561–568. X, Rauch F, I. DS, Zhang S, Rios H, Drezner MK, 24. Lyles KW, Halsey DL, Friedman NE, Lobauch B Quarles LD, F. BL, White KE (2006) Loss of DMP1 (1988) Correlations of serum concentrations of 1,25- causes rickets and osteomalacia and identifies a role dihydroxyvitamin D, phosphorus, and parathyroid hor- for osteocytes in mineral metabolism. Nat Genet 38: mone in tumoral calcinosis. J Clin Endocrinol Metab 1310–1315. 67: 88–92. 14. Lorenz-Depiereux B, Bastepe M, Benet-Pages A, 25. Araya K, Fukumoto S, Backenroth R, Takeuchi Y, Amyere M, Wagenstaller J, Muller-Barth U, Badenhoop Nakayama K, Ito N, Yoshii N, Yamazaki Y, Yamashita K, Kaiser SM, Rittmaster RS, Shlossberg AH, Olivares T, Silver J, Igarashi T, Fujita T (2005) A novel muta- JL, Loris C, Ramos FJ, Glorieux F, Vikkula M, tion in fibroblast growth factor 23 gene as a cause of Juppner H, Strom TM (2006) DMP1 mutations in auto- tumoral calcinosis. J Clin Endocrinol Metab 90: 5523– somal recessive hypophosphatemia implicate a bone 5527. matrix protein in the regulation of phosphate homeo- 26. Benet-Pages A, Orlik P, Strom TM, Lorenz-Depiereux stasis. Nat Genet 38: 1248–1250. B (2005) An FGF23 missense mutation causes familial 15. Liu S, Guo R, Simpson LG, Xiao ZS, Burnham CE, tumoral calcinosis with hyperphosphatemia. Hum Mol 30 FUKUMOTO

Genet 14: 385–390. A, Rosenblatt KP, Baum MG, Schiavi S, Hu MC, Moe 27. Larsson T, Yu X, Davis SI, Draman MS, Mooney SD, OW, Kuro-o M (2006) Regulation of fibroblast growth Cullen MJ, White KE (2005) A novel recessive muta- factor-23 signaling by klotho. J Biol Chem 281: 6120– tion in fibroblast growth factor-23 causes familial tu- 6123. moral calcinosis. J Clin Endocrinol Metab 90: 2424– 38. Nishimura T, Utsunomiya Y, Hoshikawa M, Ohuchi H, 2427. Itoh N (1999) Structure and expresseion of a novel 28. Topaz O, Shurman DL, Bergman R, Indelman M, human FGF, FGF-19, expressed in the fetal brain. Ratajczak P, Mizrachi M, Khamaysi Z, Behar D, Biochim Biophys Acta 1444: 148–151. Petronius D, Friedman V, Zelikovic I, Raimer S, 39. Inagaki T, Choi M, Moschetta A, Peng L, Cummins Metzker A, Richard G, Sprecher E (2004) Mutations in CL, McDonald JG, Luo G, Jones SA, Goodwin B, GALNT3, encoding a protein involved in O-linked Richardson JA, Gerard RD, Repa JJ, Mangelsdorf glycosylation, cause familial tumoral calcinosis. Nat DJ, Kliewer SA (2005) Fibroblast growth factor 15 Genet 36: 579–581. functions as an enterohepatic signal to regulate bile 29. Ito N, Fukumoto S, Takeuchi Y, Yasuda T, Hasegawa acid homeostasis. Cell Metab 2: 217–225. Y, Takemoto F, Tajima T, Dobashi K, Yamazaki Y, 40. Ito S, Fujimori T, Furuya A, Satoh J, Nabeshima Y, Yamashita T, Fujita T (2005) Comparison of two assays Nabeshima Y (2005) Impaired negative feedback sup- for fibroblast growth factor (FGF)-23. J Bone Miner pression of bile acid synthesis in mice lacking betaK- Metab 23: 435–440. lotho. J Clin Invest 115: 2202–2208. 30. Frishberg Y, Ito N, Rinat C, Yamazaki Y, Feinstein S, 41. Kurosu H, Choi M, Ogawa Y, Dickson AS, Goetz R, Urakawa I, Navon-Elkan P, Becher-Cohen R, Yamashi- Eliseenkova AV, Mohammadi M, Rosenblatt KP, ta T, Araya K, Igarashi T, Fujita T, Fukumoto S (2007) Kliewer SA, Kuro-o M (2007) Tissue-specific expres- Hyperostosis—hyperphosphatemia syndrome: A con- sion of beta KLOTHO and fibroblast growth factor genital disorder of O-glycosylation associated with receptor isoforms determines metabolic activity of augmented processing of fibroblast growth factor 23. J FGF19 and FGF21. J Biol Chem 282: 26687–26695. Bone Miner Res 22: 235–242. 42. Lin BC, Wang M, Blackmore C, Desnoyers LR (2007) 31. Kato K, Jeanneau C, Tarp MA, Benet-Pages A, Liver specific activities of FGF19 require KLOTHO Lorenz-Depiereux B, Bennett EP, Mandel U, Strom beta. J Biol Chem 282: 27277–27284. TM, Clausen H (2006) Polypeptide GalNAc-trans- 43. Choi M, Moschetta A, Bookout AL, Peng L, Umetani ferase T3 and familial tumoral calcinosis. Secretion of M, Holmstrom SR, Suino-Powell K, Xu HE, Richardson fibroblast growth factor 23 requires O-glycosylation. J JA, Gerard RD, Mangelsdorf DJ, Kliewer SA (2006) Biol Chem 281: 18370–18377. Identification of a hormonal basis for gallbladder fill- 32. Jaeken J, Matthijs G (2007) Congenital disorders of ing. Nat Med 12: 1253–1255. glycosylation: A rapidly expanding disease family. 44. Nishimura T, Nakatake Y, Konishi M, Itoh N (2000) Annu Rev Genomics Hum Genet 8: 261–278. Identification of a novel FGF, FGF-21, preferentially 33. Urakawa I, Yamazaki Y, Shimada T, Iijima K, Hasegawa expressed in the liver. Biochim Biophys Acta 1492: H, Okawa K, Fujita T, Fukumoto S, Yamashita T 203–206. (2006) Klotho converts canonical FGF receptor into a 45. Kharitonenkov A, Shiyanova TL, Koester A, Ford AM, specific receptor for FGF23. Nature 444: 770–774. Micanovic R, Galbreath EJ, Sandusky GE, Hammond 34. Kuro-o M, Matsumura Y, Aizawa H, Kawaguchi H, LJ, Moyers JS, Owens RA, Gromada J, Brozinick JT, Suga T, Utsugi T, Ohyama Y, Kurabayashi M, Kaname Hawkins ED, Wroblewski VJ, Li DS, Mehrbod F, T, Kume E, Iwasaki H, Iida A, Shiraki-Iida T, Nishikawa Jaskunas SR, Shanafelt AB (2005) FGF-21 as a novel S, Nagai R, Nabeshima YI (1997) Mutation of the metabolic regulator. J Clin Invest 115: 1627–1635. mouse klotho gene leads to a syndrome resembling 46. Badman MK, Pissios P, Kennedy AR, Koukos G, Flier ageing. Nature 390: 45–51. JS, Maratos-Flier E (2007) Hepatic fibroblast growth 35. Ito S, Kinoshita S, Shiraishi N, Nakagawa S, Sekine S, factor 21 is regulated by PPARalpha and is a key medi- Fujimori T, Nabeshima Y (2000) Molecular cloning ator of hepatic lipid metabolism in ketotic states. Cell and expression analyses of mouse βklotho, which en- Metab 5: 426–437. codes a novel Klotho family protin. Mech Dev 98: 115– 47. Inagaki T, Dutchak P, Zhao G, Ding X, Gautron L, 119. Parameswara V, Li Y, Goetz R, Mohammadi M, Esser 36. Yoshida T, Fujimori T, Nabeshima Y (2002) Media- V, Elmquist JK, Gerard RD, Burgess SC, Hammer RE, tion of unusually high concentrations of 1,25-dihydrox- Mangelsdorf DJ, Kliewer SA (2007) Endocrine regula- yvitamin D in homozygous klotho mutant mice by tion of the fasting response by PPARalpha-mediated in- increased expression of renal 1alpha-hydroxylase gene. duction of fibroblast growth factor 21. Cell Metab 5: Endocrinology 143: 683–689. 415–425. 37. Kurosu H, Ogawa Y, Miyoshi M, Yamamoto M, Nandi 48. Ogawa Y, Kurosu H, Yamamoto M, Nandi A, ENDOCRINE FIBROBLAST GROWTH FACTORS 31

Rosenblatt KP, Goetz R, Eliseenkova AV, Mohammadi channel. Science 310: 490–493. M, Kuro-o M (2007) BetaKlotho is required for meta- 51. Imura A, Tsuji Y, Murata M, Maeda R, Kubota K, bolic activity of fibroblast growth factor 21. Proc Natl Iwano A, Obuse C, Togashi K, Tominaga M, Kita N, Acad Sci U S A 104: 7432–7437. Tomiyama K, Iijima J, Nabeshima Y, Fujioka M, 49. Goetz R, Beenken A, Ibrahimi OA, Kalinina J, Olsen Asato R, Tanaka S, Kojima K, Ito J, Nozaki K, SK, Eliseenkova AV, Xu C, Neubert TA, Zhang F, Hashimoto N, Ito T, Nishio T, Uchiyama T, Fujimori Linhardt RJ, Yu X, White KE, Inagaki T, Kliewer SA, T, Nabeshima Y (2007) alpha-Klotho as a regulator of Yamamoto M, Kurosu H, Ogawa Y, Kuro-o M, Lanske calcium homeostasis. Science 316: 1615–1618. B, Razzaque MS, Mohammadi M (2007) Molecular 52. Kurosu H, Yamamoto M, Clark JD, Pastor JV, Nandi insights into the klotho-dependent, endocrine mode of A, Gurnani P, McGuinness OP, Chikuda H, Yamaguchi action of fibroblast growth factor 19 subfamily mem- M, Kawaguchi H, Shimomura I, Takayama Y, Herz J, bers. Mol Cell Biol 27: 3417–3428. Kahn CR, Rosenblatt KP, Kuro-o M (2005) Suppres- 50. Chang Q, Hoefs S, van der Kemp AW, Topala CN, sion of aging in mice by the hormone Klotho. Science Bindels RJ, Hoenderop JG (2005) The beta-glucu- 309: 1829–1833. ronidase klotho hydrolyzes and activates the TRPV5 note added in proof

After the acceptance of this review, it has been reported that tion of FGF23. a homozygous mutation in klotho gene causes hyperphosphatemic tumoral calcinosis. These results underscore the Ichikawa S, Imel EA, Kreiter ML, Yu X, Mackenzie DS, So- importance of FGF23 and klotho in the regulation and phos- renson AH, Goetz R, Mohammadi M, White KE, phate and vitamin D metabolism. Circulatory FGF23 is Econs MJ (2007) A homozygous missense mutation in clearly high in this patient by both full-length and C-termi- human KLOTHO causes severe tumoral calcinosis. J nal assay for FGF23 again indicating the regulatory produc- Clin Invest 117: 2684–2691.