Compound deletion of Fgfr3 and Fgfr4 partially rescues the Hyp mouse phenotype

Abstract

Uncertainty exists regarding the physiologically relevant fibroblast growth factor (FGF) receptor (FGFR) for FGF23 in the kidney and the precise tubular segments that are targeted by FGF23. Current data suggest that FGF23 targets the FGFR1c-Klotho complex to coordinately regulate phosphate transport and 1,25-dihydroxyvitamin D [1,25(OH)₂D] production in the proximal tubule. In studies using the Hyp mouse model, which displays FGF23-mediated hypophosphatemia and aberrant vitamin D, deletion of Fgfr3 or Fgfr4 alone failed to correct the Hyp phenotype. To determine whether FGFR1 is sufficient to mediate the renal effects of FGF23, we deleted Fgfr3 and Fgfr4 in Hyp mice, leaving intact the FGFR1 pathway by transferring compound Fgfr3/Fgfr4-null mice on the Hyp background to create wild-type (WT), Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice. We found that deletion of Fgfr3 and Fgfr4 in Fgfr3^−/−/Fgfr4^−/− and Hyp/Fgfr3^−/−/Fgfr4^−/− mice induced an increase in 1,25(OH)₂D. In Hyp/Fgfr3^−/−/Fgfr4^−/− mice, it partially corrected the hypophosphatemia (P_i = 9.4 ± 0.9, 6.1 ± 0.2, 9.1 ± 0.4, and 8.0 ± 0.5 mg/dl in WT, Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice, respectively), increased Na-phosphate cotransporter Napi2a and Napi2c and Klotho mRNA expression in the kidney, and markedly increased serum FGF23 levels (107 ± 20, 3,680 ± 284, 167 ± 22, and 18,492 ± 1,547 pg/ml in WT, Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice, respectively), consistent with a compensatory response to the induction of end-organ resistance. Fgfr1 expression was unchanged in Hyp/Fgfr3^−/−/Fgfr4^−/− mice and was not sufficient to transduce the full effects of FGF23 in Hyp/Fgfr3^−/−/Fgfr4^−/− mice. These studies suggest that FGFR1, FGFR3, and FGFR4 act in concert to mediate FGF23 effects on the kidney and that loss of FGFR function leads to feedback stimulation of Fgf23 expression in bone.

fibroblast growth factor (FGF) 23 (FGF23), a circulating phosphaturic hormone produced by osteocytes in bone, targets the kidney to regulate Na-phosphate cotransporter and vitamin D metabolism. Excess FGF23 causes hypophosphatemia via inhibition of SLC34A1 [Na-dependent phosphate cotransporter (Npt) 2a (Npt2a)] and SLC34A2 (Npt2c) Na-dependent phosphate transport in the proximal tubule. FGF23 also suppresses 1,25-dihydroxyvitamin D [1,25(OH)₂D] production via more complex proximal tubular effects that involve the inhibition of cytochrome P-450, family 27, subfamily B, polypeptide 1 [Cyp27b1 (1α-OHase)] activity via transcriptional and posttranslational mechanisms and stimulation of cytochrome P-450, family 24 [Cyp24 (24-OHase)]-mediated degradation of 1,25(OH)₂D (1, 7, 17, 34, 40, 41, 52). The apparent physiological functions of FGF23 are to act as the counterregulatory hormone for 1,25(OH)₂D (23) and to coordinate renal phosphate handling with bone turnover (37). Phosphate loading in mice increases FGF23 levels, suggesting a physiological response of FGF23 to dietary phosphate intake (31); however, evidence of the importance of dietary phosphate in regulating FGF23 levels in humans is conflicting (9, 29). Elevated FGF23 is responsible for several acquired and hereditary hypophosphatemic ricketic disorders, such as X-linked hypophosphatemia (XLH) (32). The Hyp mouse, a well-characterized homolog of human XLH, has been used to study the effects of FGF23 (20). Hyp mice exhibit hypophosphatemia secondary to renal phosphate wasting, impaired vitamin D metabolism, and rickets/osteomalacia due to increased FGF23 levels caused by loss-of-function mutations in Phex (phosphate-regulating gene with homologies to endopeptidases on the X chromosome), which encodes an endopeptidase expressed in osteocytes (25, 43). In contrast, decreased FGF23 causes tumoral calcinosis, a disorder characterized by hyperphosphatemia, elevated 1,25(OH)₂D levels, and soft tissue calcification (32).

FGF23 activates FGFRs complexed with Klotho, a cell surface glucosidase that imparts tissue specificity to FGF23 (16, 46). The importance of Klotho in Fgf23 signaling is illustrated by human and mouse genetic disorders where loss of Klotho results in abnormalities that resemble Fgf23 deficiency (12–14, 38). FGF23 also decreases the expression of Klotho (13, 26), providing a mechanism for FGF23 receptor desensitization. Recent in vivo studies showing nonadditive phenotypes in combined Fgf23/Klotho-null mice and rescue of the Hyp phenotype in combined Hyp/Klotho-null mice indicate that FGF23 does not have a Klotho-independent role in the regulation of systemic phosphate and vitamin D homeostasis (4, 28).

Uncertainty exists regarding the physiologically relevant FGFR for FGF23 in the kidney and the precise tubular segments that are targeted by FGF23. Although FGF23 binds to FGFR3c, FGFR4, and FGFR1c, but not FGFR2c, in vitro (14–16), there is strong support for the FGFR1c-Klotho complex being the relevant target for FGF23 in the kidney. FGFR1-Klotho complexes have been identified as the principal binding partner for FGF23 (46); neither loss of Fgfr3 nor loss of Fgfr4 rescues the Hyp mouse phenotype (24), and the conditional deletion of Fgfr1 in the kidney is purported to block the phosphaturic effects of recombinant FGF23 administration (10). However, FGF23 levels were further elevated in Hyp/Fgfr3^−/− and Hyp/Fgfr4^−/− mice (24), consistent with end-organ resistance to FGF23 caused by loss of Fgfr3 and Fgfr4.

To further investigate the physiologically relevant FGFRs in the kidney, we examined the effects of ablating Fgfr3 and Fgfr4 in Hyp mice. We found that the remaining expression of Fgfr1 in combined Hyp/Fgfr3/Fgfr4-null mice is not sufficient to fully mediate the renal effects of FGF23. Rather, FGFR3 and FGFR4 work in concert with FGFR1 to differentially regulate phosphate transport and vitamin D metabolism. Thus, FGFR3 and FGFR4 have redundant roles in mediating the effects of FGF23 in the kidney.

MATERIALS AND METHODS

Generation of Fgfr3 and Fgfr4 double-homozygous mice.

Fgfr3 and Fgfr4 knockout mice were generated as previously reported (47) and provided by Dr. Weinstein (The Ohio State University, Columbus, OH). We first created double-heterozygous Hyp females (X^HypX/Fgfr3^+/−/Fgfr4^+/−) by crossing double-homozygous males (Fgfr3^−/−/Fgfr4^−/−) with hemizygous Hyp females (X^HypX). We then crossed double-heterozygous males (XY/Fgfr3^+/−/Fgfr4^+/−) with double-heterozygous Hyp females (X^HypX/Fgfr3^+/−/Fgfr4^+/−) and collected data from all genotypes. In addition, to optimize the production of double-mutant mice, we also bred XY/Fgfr3^−/−/Fgfr4^+/− male with X^HypX/Fgfr3^+/−/Fgfr4^−/− female mice. We found that double-heterozygous Fgfr3^+/−/Fgfr4^+/− mice were indistinguishable from wild-type (WT) mice. Therefore, we combined WT and heterozygous Fgfr3^+/−Fgfr4^+/− mice into a single control (Ctr) group. Studies focused on Ctr, Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− male mice at 6 wk of age on a mixed 129Sv and C57B6/J genetic background. We previously showed that the phenotype caused by excess FGF23 is not different in 129Sv and C57B6/J mice or their progeny. We studied only male mice because of the difficulty in genotyping female Hyp (X^HypX) mice. The mice were genotyped by PCR. Primers used to genotype the Phex mutation were as follows: 5′-GCTTGGGCTAGTTTGCTATCTC-3′ (forward) and 5′-TGAGTTGGTGCTATACACGGAG-3′ (reverse). A 250-bp PCR product is obtained in WT mice and absent in Hyp mice. Animal care and protocols were approved by the University of Tennessee Institutional Animal Care and Use Committee and carried out in accordance with National Institutes of Health guidelines [Guide for Care and Use of Laboratory Animals, Institute on Laboratory Animal Resources, National Research Council (Department of Health and Human Services Publication NIH 86-23, National Academy Press, 1996)].

Dual-energy X-ray absorptiometry.

At 6 wk of age, mice are anesthetized using a ketamine-xylazine solution (120 mg/kg and 80 mg/kg, respectively), and densitometry acquisition was performed using the PIXImus (Lunar, Madison, WI). After image acquisition, bone mineral density (BMD) of the femur was measured by adjustment of the region of interest according to the entire bone size.

Biochemistry.

Serum calcium was measured using a calcium CPC Liquicolor kit (Stanbio Laboratories, Boerne, TX), and serum phosphorus was measured using the phosphomolybdylate-ascorbic acid method, as described previously (19). Serum parathyroid hormone (PTH) levels were measured using a Mouse Intact PTH ELISA kit (Immutopics, Carlsbad, CA). Serum 1,25(OH)₂D was measured using a 1,25(OH)₂D enzyme immunoassay kit (Immunodiagnostic Systems, Fountain Hills, AZ). Serum FGF23 was measured using an FGF23 ELISA kit (Kainos Laboratories, Tokyo, Japan).

High-resolution three-dimensional micro-CT.

Femurs from 6-wk-old mice were collected and fixed in 70% ethanol. Bone growth was evaluated by measurement of the length of the femur with a slide caliper. High-resolution micro-CT (μCT40, Scanco Medical, Basserdorf, Switzerland) was used to scan and evaluate the metaphyseal trabecular bone microarchitecture and the midshaft cortical bone parameters. The entire femurs were scanned in a 12.3-mm-diameter sample holder at 6-μm resolution at an energy level of 55 keV and intensity of 145 μA. The trabecular bone volume was measured within the secondary spongiosa on a set of 50 sections (0.6 mm) beneath the growth plate at a threshold of 200. Trabecular thickness, trabecular number, trabecular separation, and structure model index were calculated without assumption of a constant model, as previously described (27). The cortical bone structure was analyzed from 100 sections chosen at the midshaft of each femur at a threshold of 350.

RNA isolation and RT-PCR.

Total RNA was extracted from homogenized kidney and bone using TRI Reagent (Molecular Research Center, Cincinnati, OH) and then treated with RNase-free DNase (Qiagen, Valencia, CA). First-strand cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). Total RNA (1 μg) was used in each 20-μl RT reaction. The iCycler iQ real-time PCR detection system and iQ SYBR Green Supermix (Bio-Rad) were used for real-time quantitative PCR analysis. Values were normalized by GAPDH in the same sample and expressed as 100% of the control (WT). Sequences of primers used for real-time quantitative RT-PCR are listed in Table 1.

Table 1.

Sequences of primers used for RT-PCR

	Primer
Target Gene	Forward	Reverse
Fgfr1	AAC CTC TAA CCG CAG AAC	GAG ACT CCA CTT CCA CAG
mKlotho	AGC GAT AGT TAC AAC AAC	GCA TTC TCT GAT ATT ATA GTC
Npt2a	ATG CTG GCT TTC CTT TAC	CCA CAA TGT TCA TGC CTT CT
Npt2c	CGT GCG GAC TGT TAT CAA TG	TAC TGG GCA GTC AGG TTT CC
Cyp27b1	ACA CTT CGC ACA GTT TAC G	TTA GCA ATC CGC AAG CAC
Cyp24a1	GTT CTG TCC ACG GTA GGC	CCA GTC TTC GCA GTT GTC C

Fgfr, fibroblast growth factor receptor; Npt, Na-dependent phosphate cotransporter; Cyp27B1, cytochrome P-450, family 27, subfamily b, polypeptide 1; Cyp24A1, cytochrome P-450, family 24, subfamily a, polypeptide 1.

Immunohistochemistry.

Kidneys collected from WT and mutant mice were fixed in 4% paraformaldehyde and embedded in paraffin. Sections (5 μm thick) were cut, dried overnight, deparaffinized, and rehydrated. After antigen retrieval by incubation in citric acid buffer (10 mM, pH 3) for 60 min at 37°C, nonspecific sites were blocked with 1× animal-free blocker (Vector Laboratories, Burlingame, CA); then sections were incubated with specific primary antibodies for 1 h. An immunohistological Vectastain ABC kit (Vector Laboratories) is routinely used, and slides were counterstained with 4′,6-diamidino-2-phenylindole or methyl green, dehydrated, and mounted with Entellan medium. Rabbit-raised anti-mouse FGF23, anti-human FGFR1, and goat-raised anti-human Klotho, anti-mouse phosphorylated (Thr²⁰²/Tyr²⁰⁴) ERK, and Npt2a polyclonal primary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The Vector Red alkaline phosphatase substrate kit (Vector Laboratories) produces an intense red reaction product that can be visualized using bright-field microscopy or fluorescence (IX71, Olympus Latin, Miami, FL).

Statistical analysis.

Statistical analyses were performed using commercially available statistical software (STATISTICA, Statsoft, Tulsa, OK). We evaluated differences between groups by one-way ANOVA for multiple-group comparison and two-tailed t-test for two-group comparison. Values are means ± SE. Differences were considered statistically significant at P < 0.05.

RESULTS

Effects of compound Fgfr3 and Fgfr4 deletion on gross appearance of Hyp mice.

We transferred combined Fgfr3 and Fgfr4 deficiency onto the Hyp mouse background (Fig. 1A). Mice were born at the expected Mendelian frequency, and compound-mutant mice had survival rates identical to Ctr mice over the duration of the observation period. Hyp mice displayed growth retardation. Fgfr3^−/−/Fgfr4^−/− mice were smaller than Ctr mice and showed kyphosis and wavy, elongated tails, as previously described (47). Superimposing the Fgfr3^−/−/Fgfr4^−/− phenotype on Hyp mice (X^HypY/Fgfr3^−/−/Fgfr4^−/−) resulted in persistent kyphosis and tail distortions (Fig. 1B) and a further reduction in body length and weight compared with the Ctr and Hyp littermates (Table 2). In single-homozygous Fgfr3^−/− and Fgfr4^−/− mice on the Hyp background, there was no effect on the Hyp phenotype, as we previously reported (24).

Fig. 1.

Generation of combined phosphate-regulating gene with homologies to endopeptidases on the X chromosome (Phex)- and fibroblast growth factor receptor 3 and 4 (Fgfr3 and Fgfr4)-deficient mice.

A: genotyping of mice by PCR. Representative PCR analysis of genomic DNA for Fgfr3 and Fgfr4 genes, neomycin (Neo) cassette, and Phex gene in wild-type (WT), Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice. B: gross appearance of 6-wk-old control (Ctr), Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice.

Table 2.

Body parameters of Ctr, Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice

	Ctr	Hyp	Fgfr3−/−/Fgfr4−/−	Hyp/Fgfr3−/−/Fgfr4−/−
Length, cm	15.6 ± 1.3	14.0 ± 0.5	14.1 ± 0.8	12.6 ± 1.3a
Weight, g	20.5 ± 2.4	16.5 ± 2.6	12.4 ± 2.6a	10.5 ± 2.5a,b

Values are means ± SE; n ≥ 8 mice per group. Ctr, control. Comparisons were performed using 1-way ANOVA and t-test:

^aP < 0.05 vs. Ctr;

^bP < 0.05 vs. Hyp.

Compound deletion of Fgfr3 and Fgfr4 in Hyp mice results in a partial rescue of the hypophosphatemic phenotype.

A highly significant (>30-fold) elevation in circulating FGF23 levels in Hyp mice (Table 3) was associated with hypophosphatemia and inhibition of Npt2 message and protein expression in the proximal tubule (Fig. 2). Klotho mRNA was decreased in the kidney of Hyp mice. Serum FGF23 levels (Table 3) of Fgfr3^−/−/Fgfr4^−/− mice were identical to those of Ctr mice. The absence of Fgfr3 and Fgfr4 also did not affect serum phosphate levels or Npt2a or Klotho expression in the kidney. In contrast, compound Hyp/Fgfr3^−/−/Fgfr4^−/− mice displayed a partial rescue of the Hyp phenotype, as evidenced by a significant increase in serum phosphate concentrations in Hyp/Fgfr3^−/−/Fgfr4^−/− compared with Hyp mice, reaching a value of only 11% less than in Ctr mice. The increase in serum phosphate was associated with a concomitant increase in Npt2a protein and message expression in the kidney of Hyp/Fgfr3^−/−/Fgfr4^−/− mice (Fig. 2, Table 4). There was also a partial correction of FGF23-mediated suppression of Klotho in the kidney of compound Hyp/Fgfr3^−/−/Fgfr4^−/− mice. Klotho message expression increased to within 15% of normal values in combined Hyp/Fgfr3^−/−/Fgfr4^−/− mice.

Table 3.

Serum biochemistry of 6-wk-old Ctr, Hyp, Fgfr3^−/−/Fgfr4^−/−, and Hyp/Fgfr3^−/−/Fgfr4^−/− mice

	Ctr	Hyp	Fgfr3−/−/Fgfr4−/−	Hyp/Fgfr3−/−/Fgfr4−/−
Pi, mg/dl	9.4 ± 0.9	6.1 ± 0.2a	9.1 ± 0.4	8.0 ± 0.5b
Ca2+, mg/dl	8.2 ± 0.3	8.1 ± 0.6	8.3 ± 0.3	8.6 ± 0.4
1,25(OH)2D, pM	233 ± 21	186 ± 7	313 ± 24a	477 ± 39a,b
PTH, pg/ml	38 ± 4	58 ± 6a	42 ± 4	93 ± 7a,b,c
FGF23, pg/ml	107 ± 20	3,680 ± 284a	167 ± 22	18,492 ± 1,547a,b,c

Values are means ± SE; n ≥ 8 mice per group. 1,25(OH)₂D, 1,25-dihydroxyvitamin D; PTH, parathyroid hormone; FGF23, fibroblast growth factor 23. Comparisons were performed using 1-way ANOVA and t-test:

^aP < 0.05 vs. Ctr;

^bP < 0.05 vs. Hyp;

^cP < 0.05 vs. Fgfr3^−/−/Fgfr4^−/−.

Fig. 2.

Kidney immunostaining of Na-phosphate cotransporter (Npt) type 2a (Npt2a), membrane Klotho, FGFR1, and phosphorylated ERK (pERK) in 6-wk-old control (Ctr) and mutant mice.

Table 4.

Kidney- and bone-related gene expressions measured by RT-PCR

	Ctr	Hyp	Fgfr3−/−/Fgfr4−/−	Hyp/Fgfr3−/−/Fgfr4−/−
Kidney
Npt2a	100 ± 3.8	58 ± 9a	104 ± 13	80 ± 3b
Npt2c	100 ± 12	81 ± 14	98 ± 7	89 ± 8
Cyp27b1	100 ± 52	213 ± 43a	288 ± 147a	307 ± 184a
Cyp24a1	100 ± 38	386 ± 61a	94 ± 23	151 ± 2b
mKlotho	100 ± 3	66 ± 7a	94 ± 7	85 ± 7b
Bone
Fgf23	100 ± 15	2,236 ± 235a	181 ± 23	17,496 ± 2,980a,b,c
Fgfr1	100 ± 15	174 ± 12	129 ± 18	344 ± 41a,b,c

Values are means ± SE, expressed as percentage of wild-type value, which was set at 100%; n = 5 mice per group, except n = 7 for Cyp24a1 Ctr. Comparisons were performed using 1-way ANOVA and t-test:

^aP < 0.05 vs. Ctr;

^bP < 0.05 vs. Hyp;

^cP < 0.05 vs. Fgfr3^−/−/Fgfr4^−/−.

Compound deletion of Fgfr3 and Fgfr4 overcorrects abnormal vitamin D metabolism in Hyp mice.

Hyp mice exhibited low circulating 1,25(OH)₂D levels for the degree of hypophosphatemia, and Cyp27b1 and Cyp24 message levels in the kidney were significantly increased (Table 3). Hyp mice had normal calcium levels and slightly increased serum PTH levels (Table 3). The serum biochemical profile of Fgfr3^−/−/Fgfr4^−/− mice was similar to that of Ctr littermates, except for alterations in vitamin D metabolism. Expression of neither Cyp24 nor Klotho in the kidney was altered in Fgfr3^−/−/Fgfr4^−/− mice. The absence of Fgfr3 and Fgfr4 also did not affect serum calcium or PTH levels in the setting of normal Phex. Compound Fgfr3^−/−/Fgfr4^−/− mice, however, displayed increased expression of Cyp27b1 message in the kidney that was associated with a significant increase in serum 1,25(OH)₂D levels. The increments in Cyp27b1 required the loss of Fgfr3 and Fgfr4, since single-homozygous Fgfr3^−/− and Fgfr4^−/− mice have normal 1,25(OH)₂D levels (24). The effects of Fgfr3 and Fgfr4 on vitamin D metabolism where further accentuated in Hyp mice. Indeed, Hyp/Fgfr3^−/−/Fgfr4^−/− mice exhibited an overcorrection of serum 1,25(OH)₂D levels, resulting in 1,25(OH)₂D levels in Hyp/Fgfr3^−/−/Fgfr4^−/− mice that significantly exceeded levels in Ctr mice. The overcorrection of serum 1,25(OH)₂D was associated with a nearly threefold increase in Cyp27b1 expression in the kidney of Hyp/Fgfr3^−/−/Fgfr4^−/− compared with Ctr mice (Table 4). We also observed a further increase in serum PTH levels in Hyp/Fgfr3^−/−/Fgfr4^−/− mice, despite the increased 1,25(OH)₂D levels and the presence of normal serum calcium levels.

Compound deletion of Fgfr3 and Fgfr4 leads to further elevations of FGF23 in Hyp mice.

In accord with the increased circulating levels of FGF23, Hyp mice had a 22-fold increase in the expression of Fgf23 mRNA levels in bone and evidence for increased FGF23 protein expression in osteocytes in cortical bone compared with Ctr mice (Tables 3 and 4, Fig. 3), consistent with Phex mutations leading to increased FGF23 production by osteocytes in bone. Fgf23 mRNA expression in bone of Fgfr3^−/−/Fgfr4^−/− mice was identical to that in bone of Ctr mice, and limited amounts of FGF23 protein were detectable by immunohistochemistry in Fgfr3^−/−/Fgfr4^−/− or Ctr mice (Fig. 3), indicating that, in the presence of Phex, loss of Fgfr3 or Fgfr4 does not regulate Fgf23 expression. In contrast, in compound Hyp/Fgfr3^−/−/Fgfr4^−/− mutant mice, a further fivefold increase in serum FGF23 levels was associated with a ninefold increase in bone expression of Fgf23 message and a concomitant increase in FGF23 protein expression in cortical and trabecular bone (Fig. 3) compared with Hyp mice. These findings are consistent with loss of Fgfr3 and Fgfr4, leading to end-organ resistance to FGF23 in the kidney and compensatory increments in Fgf23 expression in bone through a yet-to-be-defined feedback loop. We also assessed the level of Fgfr1 expression in bone in compound Fgfr3^−/−/Fgfr4^−/− and combined Hyp/Fgfr3^−/−/Fgfr4^−/− mice. Interestingly, Fgfr1 message expression was increased twofold in Hyp mice and threefold in combined Hyp/Fgfr3^−/−/Fgfr4^−/− mice.

Fig. 3.

Bone immunostaining of Fgfr23 in 6-wk-old Ctr and mutant mice.

Effect of Fgfr3 and Fgfr4 double deletion on bone abnormalities in Hyp mice.

Hyp mice displayed classical features of rickets, such as splaying of the ends of the long bones, widening of the midshafts, and widening of the growth plate, as observed by dual-energy X-ray absorptiometry and micro-CT analysis (Fig. 4), associated with significant growth retardation (Fig. 1), consistent with prior reports (49). An overall 50% reduction in BMD in Hyp mice reflects the underlying osteomalacia (25).

Fig. 4.

A–C: femur bone mineral density (BMD, assessed by dual-energy X-ray absorptiometry), femur length, and 3-dimensional micro-CT representation of entire femurs in control (WT) and mutant mice at 6 wk of age. Values are means ± SE from ≥7 mice per group. Comparisons were performed using 1-way ANOVA and t-test: a vs. b, P ≤ 0.05.

The bone phenotype of Fgfr3^−/−/Fgfr4^−/− mice displayed bone dysplasia, characterized by increased bone length, curvature of the femur, and crooked tails, which is reported to be due to the loss of Fgfr3 function in the hypertrophic zone of growth plate cartilage that leads to overgrowth and bending of long bones without affecting the mineralization of osteoid (6, 24). The absence of Fgfr3 and Fgfr4 was associated with an overall 30% reduction in femur BMD mainly due to a 15% decrease in the trabecular bone volume and a significant thinning of the cortical bone (Table 5). Analysis of double-homozygous Fgfr3^−/−/Fgfr4^−/− mice did not reveal any growth defect, as evidenced by a nonsignificantly increased femur length compared with the Ctr group, suggesting that the observed trend for a shorter body length comes from the curvature of the spine, rather than a growth defect per se. Femurs displayed a slight curvature and an increased calcification of the primary spongiosa of the distal metaphysis as measured by micro-CT.

Table 5.

Results from three-dimensional micro-CT analysis of femoral distal metaphysis trabecular structure and midshaft cortical envelope

	Ctr	Hyp	Fgfr3−/−/Fgfr4−/−	Hyp/Fgfr3−/−/Fgfr4−/−
Trabecular bone parameters
Tb.BMD, mg HA/cm3	259.3 ± 33.0	36.8 ± 7.8a	138.0 ± 31.2a	46.0 ± 10.2a,c
BV/TV, %	34.98 ± 5.7	4.43 ± 0.98a	19.19 ± 2.84a	4.18 ± 0.92a,c
Tb.N, mm−1	7.14 ± 0.64	4.00 ± 0.72a	5.46 ± 0.37a	3.7 ± 0.39a,c
Tb.Th, μm	64.1 ± 6.35	46.7 ± 4.52a	50.77 ± 5.76	50.27 ± 5.79
Tb.Sp, μm	133.2 ± 12.1	278.7 ± 34.3a	181.85 ± 19.42a	309.6 ± 58.2a,c
Conn.Dens	319.9 ± 25.3	9.43 ± 0.7a	165.6 ± 28.8a	16.52 ± 2.52a,c
SMI	1.13 ± 0.26	2.6 ± 0.33a	2.35 ± 0.25a	3.41 ± 0.38a,b,c
DA	1.85 ± 0.1	2.72 ± 0.34a	1.72 ± 0.1	1.68 ± 0.17b
Cortical bone parameters
Ct.BMD, mg HA/cm3	1,069 ± 24	878 ± 14a	1,011 ± 34	868.7 ± 46a,c
Ct.Th, μm	162.4 ± 18.2	103.6 ± 4.7a	115.2 ± 17.6a	93.8 ± 20.0a
CSA, mm2	2.8 ± 0.3	3.1 ± 0.1	2.6 ± 0.2	2.7 ± 0.3
Ct.Ar, mm2	0.7 ± 0.1	0.5 ± 0.1a	0.5 ± 0.1a	0.4 ± 0.1a
Ma.Ar, mm2	2.1 ± 0.2	2.6 ± 0.3	2.1 ± 0.2	2.2 ± 0.2

Values are means ± SE; n = 5 mice per group. Tb.BMD, trabecular bone mineral density (BMD); HA, hydroxyapatite; BV, bone volume; TV, trabecular volume; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tb.Sp, trabecular separation; Conn. Dens, connectivity density; SMI, structure model index; DA, degree of anisotropy; Ct.BMD, cortical BMD; Ct.Th, cortical thickness; CSA, cross-sectional area; Ct.Ar, cortical area; Ma.Ar, matrix area. Comparisons were performed using 1-way ANOVA and t-test:

^aP < 0.05 vs. Ctr;

^bP < 0.05 vs. Hyp;

^cP < 0.05 vs. Fgfr3^−/−/Fgfr4^−/−.

The combined Hyp/Fgfr3^−/−/Fgfr4^−/− mice showed correction of some of the manifestations of rickets, with an overall improved bone appearance resembling that of Fgfr3^−/−/Fgfr4^−/− mice, except the growth defect was persistent in Hyp/Fgfr3^−/−/Fgfr4^−/− mice (Figs. 4 and 5). In addition, Hyp/Fgfr3^−/−/Fgfr4^−/− mice developed a more pronounced increased density of the primary spongiosa in the distal metaphyseal bone just beneath the growth plate, similar to previous findings of osteosclerosis in Hyp mice lacking Fgf23 or Klotho (4, 25). However, in the secondary spongiosa, the trabecular bone volume and density were identical in Hyp/Fgfr3^−/−/Fgfr4^−/− and Hyp bones, despite an improvement of the trabecular thickness. Finally, the variations in cortical bone parameters were also similar between Hyp and Hyp/Fgfr3^−/−/Fgfr4^−/− mice, but there was a trend toward improvement of the diaphysis shape, as evidenced by a smaller cross-sectional area (CSA) in Hyp/Fgfr3^−/−/Fgfr4^−/− mice (Table 5, Fig. 5).

Fig. 5.

Three-dimensional micro-CT representation of longitudinal section of distal femur and cross section of the femur midshaft from WT (control) and mutant mice. Degree of mineralization of the cortical bone is represented by a heat scale, with red color showing the more mineralized areas.

Effect of Fgfr3 and Fgfr4 double deletion on Fgf23 signaling in the kidney of Hyp mice.

The FGF23-mediated effects on the kidney of compound Hyp/Fgfr3^−/−/Fgfr4^−/− mutant mice are due to actions of Fgfr1, since this receptor is the only remaining target for FGF23 in the kidney of these mice. To examine whether loss of Fgfr3 and Fgfr4 affected the expression and function of the remaining Fgfr1 target in the kidney, we assessed Fgfr1 and phosphorylated ERK protein levels, respectively. We observed an apparent decrease in FGFR1 protein expression in the kidney of Hyp/Fgfr3^−/−/Fgfr4^−/− mice. Phosphorylated ERK expression was present in the distal tubules of Hyp and Ctr mice (Fig. 2). Loss of Fgfr3 and Fgfr4 resulted in a decrease in phosphorylated ERK levels in Fgfr3^−/−/Fgfr4^−/− and Hyp/Fgfr3^−/−/Fgfr4^−/− mice, consistent with the inability of the residual Fgfr1 to compensate for the loss of Fgfr3 and Fgfr4 (Fig. 2).

DISCUSSION

In the current study, we found that FGFR3 and FGFR4 act in a cooperative fashion to mediate the effects of FGF23 on the kidney and that FGFR1 is not sufficient to mediate the full effects of FGF23 on phosphate and vitamin D metabolism. Combined loss of Fgfr3 and Fgfr4 in Hyp mice partially corrected the hypophosphatemia and Npt2-dependent transport defect in the proximal tubule and derepressed Klotho expression in the distal tubule, despite further increments in circulating FGF23 produced in response to end-organ resistance and upregulation of Klotho expression in the kidney. Our findings are in contrast with a previously published study in which deletion of Fgfr3 or Fgfr4 alone did not prevent phosphate wasting in Hyp mice (24). The ∼5-fold increase in serum FGF23 in Hyp/Fgfr3^−/−/Fgfr4^−/− compared with Hyp mice represents the combined effects of the 2.3- and 2.1-fold increase in circulating FGF23 in Hyp/Fgfr3^−/− and Hyp/Fgfr4^−/− mice, respectively (24). The failure to fully rescue the hypophosphatemia and decreased Klotho expression in Hyp/Fgfr3^−/−/Fgfr4^−/− mice defines the contribution of Fgfr1 to proximal and distal tubular functions, since Fgfr2 is not a target for FGF23 (46). Thus, Fgfr1c, Fgfr3c, and Fgfr4 have redundant roles in mediating the effects of FGF23 on the kidney, consistent with the redundant and overlapping functions of Fgfr1, Fgfr3, and Fgfr4 in the presence of the coreceptor Klotho in vitro (15).

The finding of redundant functions of Fgfr1, Fgfr3, and Fgfr4 in the kidney challenges the widely held view that Fgfr1 is the physiologically relevant receptor mediating FGF23 effects in the kidney and has important implications in understanding the ability of FGF23 to differentially regulate specific gene products and tubular segments. For example, we found that loss of Fgfr3 and Fgfr4 had a greater impact on vitamin D metabolism than phosphate transport. In this regard, basal 1,25(OH)₂D levels and Cyp27b1 expression, both representing proximal tubular functions, were increased above normal levels in Fgfr3^−/−/Fgfr4^−/− mice. This suggests that Fgfr3 and Fgfr4 function cooperatively under basal, nonstimulated conditions to suppress 1,25(OH)₂D production without affecting phosphate transport. The increments in Cyp27b1 required the loss of Fgfr3 and Fgfr4, since single-homozygous Fgfr3^−/− and Fgfr4^−/− mice have normal 1,25(OH)₂D levels (24). Interestingly, aberrant production of 1,25(OH)₂D and regulation of Cyp27b1 in Hyp mice were overcorrected in Hyp/Fgfr3^−/−/Fgfr4^−/− mice, leading to elevations of circulating 1,25(OH)₂D levels and Cyp27b1 expression. Fgf23^−/− and Klotho^−/− mice have unusually high serum levels of 1,25(OH)₂D, suggesting that removal of FGF23 effects on the kidney leads to markedly increased 1,25(OH)₂D production (39, 42, 45). Our studies indicate that Fgfr3 and Fgfr4, rather than Fgfr1, are mediating this effect. In contrast, hypophosphatemia was only partially corrected in Hyp/Fgfr3^−/−/Fgfr4^−/− mice, implying that FGFR1 has an important role in phosphate transport. In addition, we found that combined loss of Fgfr3 and Fgfr4 in Hyp mice had a greater impact on Npt2a than Npt2c expression, which is necessary for the full phosphaturic activity of FGF23 (44). Interestingly, deletion of Fgfr3 and Fgfr4 in the Fgfr3^−/−/Fgfr4^−/− mice did not alter the phosphate levels. This finding suggests that FGFR1 is sufficient for normal phosphate homeostasis and that Fgfr3 and Fgfr4 respond when nonphysiological levels of FGF23 occur.

The mechanisms underlying the differential effects of FGFRs to regulate proximal tubular phosphate and vitamin D metabolism are not clear, since FGF23 binding and signaling pathways do not differ between FGFR1, FGFR3, and FGFR4 (16), and Fgfr3 is expressed in the proximal tubule (5, 24), whereas Fgfr1, Fgfr3, and Ffgr4 are expressed in the distal tubules (24). Klotho is also predominantly expressed in the distal tubule (18), although recent studies show the presence of Klotho in proximal tubule cell cultures (11). Nevertheless, a distal-to-proximal paracrine feedback mechanism may mediate some of the effects of FGF23 (22). In support of this idea, Klotho, which can be released from the distal tubule by ectodomain shedding, is capable of regulating the insertion of Npt2 into the brush border membrane of the proximal tubule (11). Ex vivo studies of proximal tubular segments and cell lines have demonstrated variable effects of exogenously added FGF23 to inhibit Na-dependent phosphate transport (35, 40). Cell culture studies are confounded by the use of nonphysiological amounts of FGF23 and the authenticity of the proximal tubular phenotype in cell culture models, which may be contaminated with distal tubular cells (3, 35, 36, 50, 51). FGF23 regulates gene expression in the distal tubule, as evidenced by FGF23-mediated decreases in Klotho expression and the observation that the distal tubule is the initial tubular segment to show a signaling response to FGF23 (8).

Analysis of our data in vivo is potentially confounded by alterations in serum phosphate and PTH levels, which may independently affect tubular functions. FGF23 is known to stimulate Cyp27b1 mRNA but reduces protein and enzymatic activity in Hyp mice due to posttranslational effects that lead to decreased synthesis of 1,25(OH)₂D through pathways that are influenced by renal tubular phosphate transport (52). In Hyp/Fgfr3^−/−/Fgfr4^−/− mice, increased proximal tubular transport of phosphate caused by the absence of Fgfr3 and Fgfr4 may have corrected the translational defect in Cyp27b1, thereby allowing the increments in FGF23 to further stimulate Cyp27b1 transcription, leading to a paradoxical increase in 1,25(OH)₂D production in Hyp/Fgfr3^−/−/Fgfr4^−/− mice. We also observed increased serum PTH in Hyp/Fgfr3^−/−/Fgfr4^−/− mice, which could also stimulate Cyp27b1 expression in the kidney.

The mechanism for the increase in PTH in Hyp/Fgfr3^−/−/Fgfr4^−/− mice is not clear. Serum calcium levels are normal and 1,25(OH)₂D levels are elevated in Hyp/Fgfr3^−/−/Fgfr4^−/− mice. Fgfrs and Klotho are expressed in the parathyroid gland (14, 18, 46), but existing data indicate that FGF23 directly suppresses PTH mRNA expression in vitro and decreases serum PTH in vivo (2). On the other hand, elevated FGF23 does not prevent the development of hyperparathyroidism in Hyp mice (25), and there is a strong association between elevated FGF23 levels and elevated PTH levels in chronic kidney disease (48), suggesting that excess FGF23, particularly at levels observed in the present study, may promote the development of hyperparathyroidism. Further studies are needed to clarify the potential direct and indirect effects of Fgfr3 and Fgfr4 ablation on parathyroid gland function.

Finally, we observed that compound deletion of Fgfr3 and Fgfr4 likely leads to further elevations of FGF23 in Hyp mice and an alteration in distribution of Fgf23 expression from cortical osteocytes to metaphyseal trabecular bone. At least three mechanisms explain the increase in FGF23 in Hyp/Fgfr3^−/−/Fgfr4^−/− mice: 1) increase of 1,25(OH)₂D, which is a potent stimulator of Fgf23 expression (23), 2) increase of Fgfr1 expression in bone, which has been implicated in regulation of Fgf23 expression in osteocytes (21, 33), and 3) end-organ resistance to FGF23, leading to production of a yet-to-be-identified renal factor that stimulates Fgf23 expression in bone. The fact that loss of FGF23 responsiveness in the kidney by ablation of Klotho also results in elevations of Fgf23 expression in bone (46), in association with elevated 1,25(OH)₂D levels, and that ablation of 1α-OHase in Klotho^−/− mice reverses the elevated serum level of FGF23 (30) implies that Fgf23 expression is regulated by 1,25(OH)₂D.

In conclusion, FGFR3 and FGFR4 function in a cooperative manner to mediate FGF23 effects on the kidney in Hyp mice. Thus, FGFR1, FGFR3, and FGFR4 are physiologically relevant FGFRs for FGF23 in the kidney. The ability of FGF23 to target multiple receptors provides redundancy and may allow for selective regulation of separate functions in different tubular segments.

GRANTS

This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant 7R56 AR-045955.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.