FGF23 Neutralizing Antibody Partially Improves Bone Mineralization Defect of HMWFGF2 Isoforms in Transgenic Female Mice

Abstract

Mice overexpressing high molecular weight FGF2 isoforms (HMWTg) in osteoblast lineage phenocopy human X-linked hypophosphatemic rickets (XLH) and a Hyp murine model of XLH demonstrating increased FGF23/FGF receptor signaling and hypophosphatemic rickets/osteomalacia. Because HMWFGF2 was upregulated in bones of Hyp mice and abnormal FGF23 signaling is important in XLH, HMWTg mice were used to examine the effect of the FGF23 neutralizing antibody (FGF23Ab). Eight-week-old female Vector control mice and HMWTg mice were treated with FGF23Ab or control IgG. A single injection of FGF23Ab rescued abnormal hypophosphatemia in HMWTg. The decreased type II sodium-dependent phosphate co-transporter (Npt2a) was rescued by FGF23Ab treatment. Inappropriately low serum 1,25(OH)₂D in HMWTg mice was normalized by FGF23Ab treatment, which is accompanied by increased anabolic vitamin D hydroxylase Cyp27b1 and decreased catabolicvitamin D hydroxylase Cyp24 mRNA in kidney. Long-term treatment with FGF23Ab normalized femur length and significantly increased vertebrae BMD and BMC, and femur BMC in HMWTg mice compared to IgG-treated HMWTg mice. Micro–computed tomography (μCT) revealed increased cortical porosity and decreased cortical apparent density in the HMWTg-IgG group compared with the Vector-IgG group; however, FGF23Ab treatment rescued defective cortical mineralization, decreased porosity, and increased apparent density in HMWTg mice. Bone histomorphometry analysis showed FGF23Ab treatment decreased osteoid volume, increased intra-label thickness, mineralization apposition rate, and bone formation rate in HMWTg mice. FGF23Ab improved disorganized double labeling in femurs from HMWTg mice. Quantitative real-time PCR analysis of tibia shafts showed FGF23Ab treatment normalized the osteocalcin (Ocn) mRNA expression in HMWTg mice, but further increased expression of SIBLING protein–related and pyrophosphate-related genes that are important in matrix mineralization, suggesting that HMWFGF2 modulates these genes independent of FGF23. We conclude that FGF23Ab partially rescued hypophosphatemic osteomalacia in HMWTg. However, long-term treatment with FGF23Ab further increased SIBLING protein–related genes and pyrophosphate-related genes in bone that could contribute to incomplete rescue of the mineralization defect. © 2018 American Society for Bone and Mineral Research.

Introduction

X-linked hypophosphatemia (XLH) is the most common form of heritable rickets/osteomalacia resulting from phosphate-regulating endopeptidase homolog X-linked (Phex) mutations and characterized by elevated levels of fibroblast growth factor 23 (FGF23) leading to a defect in inorganic phosphate (Pi) metabolism with consequent renal Pi wasting and impaired renal production of 1,25-dihydroxyvitamin D (1,25D).^(1-4) In adults with XLH, inappropriately elevated circulating FGF23 levels lead to persistent osteomalacia, musculoskeletal pain, stiffness, pseudofractures, osteoarthritis, enthesopathy, and muscle dysfunction.⁽³⁾ Oral calcitriol and Pi supplements are the main traditional therapeutic options. However, they offer limited efficacy, and require regular monitoring for potential toxicities.⁽³⁾ In addition, treating XLH with Pi and calcitriol was associated with concurrent increases in circulating FGF23 concentrations, which may diminish therapeutic effect or contribute to complications of therapy.⁽⁵⁾ Recently, FGF23 neutralizing antibody (FGF23Ab) has been in clinical trial to treat adults with XLH^(6-10) and has been shown to effectively restore serum Pi homeostasis, reduce stiffness, improve physical functioning, and increase markers of bone remodeling with consequent improved healing of fractures/pseudofractures. Although FGF23 plays a major role in hypophosphatemia and the associated bone mineralization defect in XLH,⁽¹⁾ the regulators of FGF23 production, the signaling pathway for FGF23-induced Pi wasting, and the cause of bone mineralization defects in these disorders are not fully defined.

Fibroblast growth factor 2 (FGF2) is another member of the FGF family of ligands that is mitogenic for cells including osteoblasts and chondrocytes.^(11-13) A single Fgf2 gene encodes multiple FGF2 isoforms. FGF2 high molecular weight (HMW) protein isoforms are expressed from unique CUG alternative translation start sites located 5′ to the classical AUG initiation codon for the low molecular weight (LMW) exported isoform.^(13,14) HMW FGF2 isoforms are usually not released from the cells, but have nuclear localization sequences and function in an intracrine manner.^(15-19) Our previous studies have shown differential effects of the LMW and the HMW isoforms on bone homeostasis in mice.^(20-22) We reported that transgenic mice which express the human LMW isoform of FGF2 under the control of the Col3.6 promoter (LMWTg) had increased bone mass with normal calcium and Pi homeostasis.⁽²¹⁾ Furthermore, there was no increase in FGF23 in serum of LMWTg mice relative to Vector control mice. In contrast to the increased bone mass found in the LMWTg mice, we reported that transgenic mice which express the human HMW FGF2 isoforms under the control of the Col3.6 promoter (HMWTg) phenocopy the Hyp mouse, a homologue of XLH, and display reduced bone mineral density (BMD), rickets/osteomalacia, defective mineralization, hypophosphatemia, and increased FGF23 in serum and bone.⁽²²⁾ Intriguingly, in the Hyp mouse we found that Fgf2 mRNA and HMW FGF2 protein isoforms were increased in bones and osteoblasts, and that FGF23 and HMW FGF2 protein co-localized in Hyp osteocytes.⁽²²⁾

Our previous in vitro studies showed that defective bone matrix mineralization in bone marrow stromal cell (BMSC) cultures from HMWTg mice is due in part to increased FGF23/FGFR signaling. Impaired mineralized bone nodule formation in BMSC cultures from HMWTg mice was only partial rescued with neutralizing FGF23Ab that are independent of its effects on in vivo Pi homeostasis.⁽²⁰⁾ The present in vivo study is to examine whether bone and Pi wasting phenotypes of HMWTg mice can be fully or partially rescued by neutralizing FGF23Ab that was shown to ameliorate rickets in the Hyp mouse model⁽²³⁾ and to examine whether the mineralization defect in HMWTg mice is totally dependent on FGF23/FGFR signaling.

Materials and Methods

Animals

All animal protocols were approved by the UConn Health Institute of Animal Care and Use Committee. Generation and initial characterization of the bone phenotype of the HMWTg mice; ie, mice expressing HMW isoforms of hFGF2 in a bone-specific manner, was previously reported.⁽²²⁾ Briefly, Col3.6-HMWFgf2 isoform-IRES-GFPsaph was built by replacing a CAT fragment in previously made Col3.6-CAT-IRES-GFPsaph with a human HMWFgf2 cDNA between AfeI and ScaI sites. This expression construct is capable of concurrently overexpressing the human HMW FGF2 isoforms⁽¹⁴⁾ and GFPsaph from a single bicistronic mRNA. Generation of the Fgf2 cDNAs was previously described.⁽¹⁴⁾ As control, a Col3.6-IRES/GFPsaph (Vector) construct was also generated and purified according to standard techniques. Microinjections into the pronucleus of fertilized oocytes were performed at the Gene Targeting & Transgenic Facility at UConn Health. Founder mice of the F2 (FVBN) strain were mated with wild-type mice to establish individual transgenic lines. Homozygote mice were generated by mating heterozygote males with heterozygote females. Female mice were used in this study.

FGF23 neutralizing antibody treatment

A rat anti-rat FGF23 neutralizing antibody (FGF23Ab, 10 mg/kg body weight; Amgen Inc., Thousand Oaks, CA, USA) or control IgG (rat-anti-NGFPb-3F8-raIgG2a; Amgen Inc.) was administered by intraperitoneal injection (i.p.). The dosage was based on the pilot experiments showing a single injection of FGF23Ab at 10 mg/kg was able to significantly increase serum Pi level in HMWTg mice at 24 hours (Fig. 1) or 96 hours (data not shown) postinjection. Mice were used at 8 weeks of age in the case of single-dose administrations. For long-term treatment over 6 weeks, dosing was initiated at 8 weeks of age with a schedule of two treatments per week. Vector mice were treated with control IgG only. We also performed additional long-term experiment that includes a Vector group subject to FGF23AB or control IgG treatment. In this experiment, mice were 3 weeks old at treatment initiation and received two treatments per week for 6 weeks. For serum tissue nonspecific alkaline phosphatase (TNAP) concentration and total alkaline phosphatase (ALP) activity measurement, starting at 18 days of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Mice were euthanized 48 hours after the last FGF23Ab administration for serum collection. Mice were euthanized by CO₂ for sample collection.

Fig. 1.

Effect of single injection of FGF23Ab on serum and urinary biochemical markers and renal Fgfr1c, Fgfr3c, Npt2a, Cyp27b1, and Cyp24 mRNA expression in Vector and HMWTg mice. Eight-week-old Vector or HMWTg female mice received a single intraperitoneal injection of the neutralizing FGF23Ab (10 mg/kg) or control IgG. Serum and spot urine, and kidneys were collected during euthanasia at 6 hours or 24 hours postinjection. (A) Serum calcium, (B) serum phosphate, (C) phosphate excretion index, and (D) serum 1,25D levels were measured. Kidney (E) Fgfr1c, (F) Fgfr3c, (G) Npt2a, (H) Cyp27b1, and (I) Cyp24 mRNA level was analyzed by qPCR. Data are means ± SE. n = 7 mice/group. ‘compared with Vector-IgG group, p < 0.05; #compared to corresponding IgG group, p < 0.05.

Faxitron X-ray

X-ray pictures of the excised femurs and tibias were taken using a SYSTEM MX 20 from Faxitron X-ray Corp. (Wheeling, IL, USA). X-ray images were taken under constant conditions (25 kV, 20-s exposure at magnification 4.5×).

DXA

DXA imaging was performed using the LunarPIXImus2 (GE Medical Systems, Inc., Waukesha, WI, USA) densitometer to measure bone mineral density (BMD) and bone mineral content (BMC) for both in vivo and excised bones.

Micro–computed tomography scanning of femurs

The mid-diaphysis cortical bones and metaphyseal cancellous bones of the distal femurs were used for analysis with microcomputed tomography (μCT) instrumentation (μCT20; Scanco Medical AG, Bassersdorf, Switzerland). Using two-dimensional (2D) data from scanned slices, three-dimensional (3D) analysis was conducted to calculate morphometric parameters defining microarchitecture, including cortical thickness (Ct.Th), periosteal perimeter, endosteal perimeter, cortical porosity, and cortical apparent density.

Bone histomorphometry

To analyze the bone turnover rate, intraperitoneal injection of calcein and xylenol orange were performed at 7 and 2 days, respectively, before euthanization. Femurs were isolated and fixed in 10% formalin then placed overnight in 30% sucrose dissolved in PBS and embedded in Cryomatrix. Seven-micrometer (7-μm) longitudinal central sections that include the central vein were collected on a cold adhesive tape Cryofilm type IIC (FINETEC Co. Ltd., Yanagawa, Japan). Unstained tapes with samples were soaked in PBS for half an hour and then mounted in 50% glycerol in PBS for dynamic parameter analysis. Additional sections were stained for tartrate-resistant acid phosphatase (TRAP) to visualize osteoclasts and counterstained with hematoxylin. Histomorphometric measurements were made in a blinded, nonbiased manner using the OsteoMeasure image analysis system (R&M Biometrics, Nashville, TN, USA) interfaced with a Nikon E400 microscope (Nikon Inc., Melville, NY, USA). The terminology and units used are those recommended by the Histomorphometry Nomenclature Committee of the American Society for Bone and Mineral Research.⁽²⁴⁾ BV/TV, Tb.N, Tb.Th, Tb.Sp, osteoclast number/bone surface (Oc.N/BS), osteoclast surface/bone surface (Oc.S/BS), osteoblast surface/bone surface (Ob.S/BS), osteoid volume/bone volume (OV/BV), interlabel thickness (Ir.L.Th), mineral apposition rate (MAR), mineralizing surface/bone surface (MS/BS), and the bone formation rate (BFR/BS) were measured. Osteoblasts were identified as cuboidal cells lining the trabecular bone. Osteoclasts were identified as multinucleated cells with more than three nuclei on the trabecular bone surface.

Histology

Femurs were isolated and fixed in 10% formalin. Femurs were then placed overnight in 30% sucrose dissolved in PBS and embedded in Cryomatrix. The Cryomatrix block containing each undecalcified femur was oriented in the block holder to obtain a 7-μm longitudinal central section that includes the central vein. Sections were collected on a special cold adhesive tape Cryofilm type IIC (FINETEC Co. Ltd.). Frozen sections were scanned to detect GFPsaph, calcein, and xylenol orange labeling of bone cells. GFP expression in cells was visualized using an Olympus IX50 inverted system microscope equipped with an IX-FLA inverted reflected light fluorescence (Olympus America, Inc., Melville, NY, USA). A specific excitation wavelength was obtained using filters for GFPsaph and recorded with a SPOT-camera (Diagnostic Instrument, Inc., Sterling Heights, MI, USA). Fluorescent images were taken with equal exposure times applied to bones derived from Vector and HMWTg mice. Masson Trichrome staining of femur sections was performed to detect osteoid.

For detection of TNAP staining in osteocytes on cortical bone, bone sections on tape were incubated with Vector Red ALP staining kit (Vector Laboratories, Burlingame, CA, USA) according to the manufacturer’s protocol. Fluorescent images of TNAP and DAPI were obtained using appropriate filter sets. Fluorescence measurements were obtained using Image J software (NIH, Bethesda, MD, USA; https://imagej.nih.gov/ij/). At least 100 osteocytes in cortical bone per mouse were chosen for the quantification of relative fluorescence in osteocytes. Relative fluorescence in osteocytes was normalized to cell number. Vector Blue ALP staining kit (Vector Laboratories) was used for TNAP staining of cancellous bone. Dark blue staining along the bone surface was quantified using OsteoMeasure for obtaining TNAP surface per bone surface of osteoblasts.

Frozen sections of femur were used for immunohistochemistry staining. The sections were washed in 1× PBS/1% fetal bovine serum (FBS) and permeabilized with 0.25% Triton X-100 in 1× PBS/1% FBS for 10 min, followed by blocking in 1:200 normal serum for 30 min, and then incubated with primary antibodies at 4°C overnight. The primary antibodies were a polyclonal anti-phospho-FGFR1 (Abnova, Taipei, Taiwan) utilized at 1:50 dilution, a monoclonal anti-phospho-ERK antibody (Cell Signaling Technology, Beverly, MA, USA) utilized at a 1:50 dilution, a polyclonal anti-ANK antibody (ABGENT, San Diego, CA, USA) used at a 1:50 dilution. The sections were washed three times for 5 min in PBS, and then a horseradish peroxidase–conjugated secondary antibody was applied for 30 min. Following three 5-min PBS washes, color was developed with ABC reagent for 30 min. After washing, sections were counterstained with hematoxylin.

RNA isolation and real-time PCR

Total RNA was extracted from kidney or flushed tibia shafts using Trizol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA). For real-time quantitative RT-PCR (qRT-PCR) analysis, RNA was reverse-transcribed using the RNA to cDNA EcoDry^™ Premix (Oligo dT). qPCR was carried out using the iTaq^™ Universal SYBR^® Green Supermix on a MyiQ^™ instrument (BIO-RAD Laboratories Inc., Hercules, CA, USA). β-Actin or Gapdh was used as an internal reference for each sample. mRNA was normalized to the β-Actin or Gapdh mRNA level and expressed as the fold-change relative to the first sample for each experimental group. Relative mRNA expression was calculated using a formula reported previously.⁽²⁵⁾ The primers for the genes of interest are listed in Supplemental Table 1.

Biochemistry

Serum was prepared from blood that was collected from euthanized animals by cardiac puncture. For long-term experiments, mice were euthanized at 48 hours after the last injection of the 6-week study. Serum and urine Pi were measured using the Phosphorus Liqui-UV (StanBio Laboratory, Boerne, TX, USA). Serum calcium was measured using a Calcium (CPC) reagent set (Eagle Diagnostics, Cedar Hill, TX, USA). Serum 1,25D was measured using 1,25-Dihydroxy Vitamin D EIA kit from IDS (Immunodiagnostic Systems, Gaithersburg, MD, USA). Spot urine samples were collected during euthanasia and renal phosphate excretion was determined. Renal phosphate excretion is expressed as phosphate excretion index (PEI) to minimize inaccuracies of urine collection and the effect of fluctuations in glomerular filtration rate with time because PEI relates to creatinine clearance. PEI is calculated by urine phosphate concentration × serum creatinine concentration, divided by serum phosphate concentration. Serum TNAP was measured using a bone ALP ELISA Kit (MyBioSource Inc, San Diego, CA, USA). Serum total ALP activity was measured using ALP Reagent (Thermo Scientific, Waltham, MA, USA) following the manufacture’s instruction.

Statistical analysis

Experimental values are reported as mean ± standard error (SE). ANOVA followed by least significant difference (LSD) for post hoc multiple comparisons and Student’s t test were used. SPSS software was used for statistical analysis (IBM Corp., Armonk, NY, USA), and the results were considered significantly different at p < 0.05.

Results

Single injection of neutralizing FGF23Ab in adult Vector and HMWTg mice

To initially assess the effect of FGF23Ab on FGF23 signaling in both Vector and HMWTg mice we performed a single-dose, short-term treatment study in which serum biochemical markers were analyzed at 6 hours and 24 hours postinjection to determine the immediate effects of FGF23 neutralizing before the onset of feedback regulation. As shown in Fig. 1A, short-term FGF23 neutralizing at 6 hours or 24 hours post-FGF23Ab treatment did not affect serum calcium level in either Vector or HMWTg mice. Consistent with our previous publication in male mice,⁽²²⁾ the current study in female mice showed that there was a significant decrease in serum Pi in the HMWTg-IgG group compared with the Vector-IgG group. FGF23Ab significantly increased serum Pi level in HMWTg mice at 24 hours postinjection (Fig. 1B). FGF23Ab treatment did not affect serum Pi level in Vector mice at 6 hours or 24 hours posttreatment. Increased PEI in HMWTg mice was suppressed at 24 hours by FGF23Ab administration (Fig. 1C).

It is known that FGF23 regulates serum Pi in part through modulation of 1,25-dihydroxyvitamin D (1,25D) production via 1-α-hydroxylase in the kidney.⁽²⁶⁾ Although we reported that serum FGF23 was significantly increased in HMWTg mice,⁽²²⁾ serum 1,25D in the HMW-IgG group was not significantly different compared to the Vector-IgG group (Fig. 1D). In HMWTg mice, FGF23Ab treatment resulted in a significant increase in serum 1,25D at 24 hours postdosing (Fig. 1D). Taken together, these results suggest that pharmacological neutralization of FGF23 counteracts FGF23 signaling mainly in HMWTg mice.

It is well known that FGF23 activates FGF receptors in kidney which are necessary for the transduction of the FGF23 signal, which results in downregulation of type II sodium-dependent phosphate co-transporter (Npt2a). We therefore examined the effects of FGF23Ab on renal Fgfr1c, Fgfr3c, and Npt2a mRNA expression. Consistent with our published data in male mice, renal Fgfr1c mRNA expression was significantly increased in HMWTg-IgG female mice compared with the Vector-IgG group. At 24 hours posttreatment FGF23Ab decreased Fgfr1c mRNA expression in HMWTg, but increased Fgfr1c mRNA expression in Vector mice (Fig. 1E). Similarly, increased renal Fgfr3c mRNA level in HMWTg mice was rescued with FGF23Ab at 24 hours posttreatment (Fig. 1F). Decreased renal Npt2a mRNA level in HMWTg mice was rescued with FGF23Ab treatment at 24 hours (Fig. 1G), which was accompanied by a recovery of urine PEI (Fig. 1C).

Because a single injection of FGF23Ab resulted in a significant increase in serum 1,25D level at 24 hours postadministration, we examined whether this change in serum 1,25D level was due to alteration in the renal expressions of two key vitamin D-metabolizing enzymes: 25-hydroxyvitamin-D-1a-hydroxylase (1aOHase, Cyp27b1) that metabolize 25 hydroxyvitamin D into active 1,25D in kidney, and 24-hydroxylase (24OHase, Cyp24) that catabolizes 1,25D into its inactive form. As shown in Fig. 1H, there was no difference in renal Cyp27b1 mRNA levels between the Vector-IgG and HMWTG-IgG group. The FGF23Ab treatment significantly increased Cyp27b1 mRNA expression in HMWTg at 24 hours posttreatment. Cyp24 mRNA expression was significantly increased in the HMWTg-IgG group compared with the Vector-IgG group. FGF23Ab treatment restored Cyp24 mRNA expression to normal levels at 6 hours and 24 hours posttreatment (Fig. 1I).

Repeated long-term injections of FGF23Ab in adult HMWTg mice

To determine whether long-term neutralization of FGF23 modulates body weight and phosphate/calcium homeostasis in HMWTg mice, HMWTg mice at 2 months of age were treated with control IgG or FGF23Ab for 6 weeks. As shown in Fig. 2A, body weight was decreased in both HMWTg-IgG and HMWTg-FGF23Ab groups compared with the Vector-IgG group. FGF23Ab treatment did not affect body weight in HMWTg mice during 6 weeks of treatment. Serum calcium and Pi was measured at the end of the 6-week study. As shown in Fig. 2B, there was no difference in serum calcium among groups. Decreased serum Pi level (Fig. 2C) and increased urine Pi wasting (Fig. 2D) in HMWTg mice were completely rescued by FGF23Ab treatment.

Fig. 2.

Effects of long-term FGF23Ab treatment on body weight and mineral ion homeostasis in HMWTg mice. Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Body weight was measured biweekly. Serum and spot urine were collected during euthanasia. Mice were euthanized 48 hours after the last FGF23Ab administration. (A) Body weight, (B) serum calcium, (C) serum phosphate, and (D) phosphate excretion index were measured. Data are means ± SE. n = 5 mice/group. *Vector-IgG versus HMW-IgG, p < 0.05; #Vector-IgG versus HMW-FGF23Ab, p < 0.05; @HMW-IgG versus HMW-FGF23Ab, p < 0.05.

To determine whether long-term neutralization of FGF23 regulated BMD and BMC in HMWTg mice, we measured BMD and BMC before treatment and at 2 weeks and 4 weeks of treatment in vivo (Fig. 3A-D). We also measured BMD and BMC on excised bone at 6 weeks of treatment (Fig. 3E-H). In vivo DXA analysis showed that vertebrae BMD was decreased in both the HMWTg-IgG and HMWTg-FGF23Ab group compared to the Vector-IgG group before treatment and at 2 weeks and 4 weeks of treatment (Fig. 3A). Vertebrae BMC was decreased in the HMWTg-IgG group compared to the Vector-IgG group before treatment and at 2 weeks of treatment. However after 4 weeks of treatment, decreased vertebrae BMC in HMWTg mice was partially rescued with FGF23Ab treatment (Fig. 3B). In vivo DXA analysis showed that femur BMD was decreased in both the HMWTg-IgG and HMWTg-FGF23Ab group compared to the Vector-IgG group before treatment and at 2 weeks and 4 weeks of treatment. FGF23Ab did not rescue decreased femur BMD at these time points (Fig. 3C). Femur BMC was also decreased in both the HMWTg-IgG and HMWTg-FGF23Ab group compared to the Vector-IgG group before treatment and at 2 weeks and 4 weeks of treatment (Fig. 3D). Ex vivo DXA analysis on excised bones showed that at 6 weeks of FGF23Ab treatment, decreased vertebrae BMD and BMC in HMWTg mice was rescued (Fig. 3E, F). Femur BMD was decreased in the HMWTg-IgG compared to Vector-IgG group at 6 weeks treatment that was not rescued by FGF23Ab treatment (Fig. 3G). However, after 6 weeks decreased femur BMC in HMWTg mice was partially rescued with FGF23Ab treatment (Fig. 3H).

Fig. 3.

Effects of long-term FGF23Ab treatment on vertebrae and femur BMD and BMC in HMWTg mice. Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. In vivo BMD and BMC were measured before treatment and at 2 weeks and 4 weeks of treatment. Ex vivo BMD and BMC were measured on excised bone at 6 weeks of treatment. In vivo (A) vertebrae BMD, (B) vertebrae BMC, (C) femur BMD, and (D) femur BMC before treatment and at 2 weeks and 4 weeks of treatment. Ex vivo (E) vertebrae BMD, (F) vertebrae BMC, (G) femur BMD, (H) femur BMC at 6 weeks of treatment on excised bones. Data are means ± SE. n = 5 mice/group. *Vector-IgG versus HMW-IgG, p < 0.05; #Vector-IgG versus HMW-FGF23Ab, p < 0.05; @HMW-IgG versus HMW-FGF23Ab, p < 0.05.

To determine the effect of FGF23Ab on Vector mice, 3-week-old Vector mice were treated with control IgG or FGF23Ab for 6 weeks. As shown in Supplemental Fig. 1, body weight, tail length, long-bone growth, femur BMD, and BMC were similar between control and FGF23Ab-treated Vector mice.

Effects of long-term FGF23Ab treatment on femur structure in HMWTg mice

Because a single dose of FGF23Ab rescued the hypophosphatemia of HMWTg mice, we assessed whether long-term neutralization of FGF23 could rescue the osteomalacia bone phenotype in adult HMWTg mice. As shown by radiography in Fig. 4A, B, femur length that was significantly decreased in HMWTg-IgG compared with Vector-IgG mice was partially normalized with FGF23Ab treatment.

Fig. 4.

μCT analysis of femur after long-term administration of FGF23Ab in HMWTg mice. Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. (A) Representative X-ray images of femur at 6 weeks posttreatment. (B) Quantification of femoral length. (C–H) μCT showed FGF23 neutralization improves integrity of femoral cortical bone of HMWTg mice. (C) Representative μCT images of femoral cancellous bone (upper panel) and cortical bone (bottom panel) of Vector and HMWTg treated with FGF23Ab or control IgG. Note trabecular area of femur revealed abnormal aberrant bubbling regions (arrowhead) in HMW-Vehicle group. These structural abnormalities were partially rescued by FGF23Ab treatment. In addition, radiolucency (arrow) in cortical of HMWTg-IgG mice was normalized with FGF23Ab treatment. Quantification of (D) cortical thickness, (E) periosteal perimeter, (F) endosteal perimeter, (G) cortical porosity, (H) cortical apparent density. Data are means ± SE. n = 5 mice/group. *Vector-IgG versus HMW-IgG, p < 0.05; #Vector-IgG versus HMW-FGF23Ab, p < 0.05; @HMW-IgG versus HMW-FGF23Ab, p < 0.05.

The bone structure of excised femurs was assessed in control IgG-treated or FGF23Ab-treated HMWTg mice by μCT analysis at 6 weeks posttreatment. Trabecular area of femur revealed abnormal aberrant bubbling regions as shown by the arrowhead in Fig. 4C in the HMWTg-Vehicle group. These structural abnormalities were partially rescued by FGF23Ab treatment. μCT scanning of femurs showed radiolucency in the cortical bone of HMWTg-IgG mice (Fig. 4C, arrow) that was normalized by 6 weeks of FGF23Ab treatment. There were no differences in cortical thickness among groups at 6 weeks of treatment (Fig. 4D). Periosteal perimeter and endosteal perimeter were increased in the HMWTg-FGF23Ab group compared with the Vector-IgG group (Fig.4 E, F). μCT analysis revealed that cortical porosity in HMWTg-IgG mice was more than twofold higher compared with the Vector-IgG group, which was partially normalized with FGF23Ab treatment (Fig. 4G). There was a significant decrease in cortical apparent density in HMWTg mice, which was not normalized with FGF23Ab treatment (Fig. 4H).

Long-term FGF23Ab treatment ameliorates mineralization defect in HMWTg mice

To examine the long-term effect of neutralization of FGF23 on bone structure in more detail, bone histomorphometric analysis was performed on femurs harvested at the end of 6 weeks of treatment. Histomorphometric analysis showed that decreased BV/TV in HMWTg mice was partially rescued with FGF23Ab treatment (Fig. 5A). Decreased trabecular number and increased trabecular spacing was fully rescued with FGF23Ab treatment (Fig. 5B, D). There was no significant difference in trabecular thickness among groups (Fig. 5C). There were no significant differences in osteoclast number and osteoclast surface among the groups (Fig. 5E, F). There was no difference in osteoblast surface among the groups (Fig. 5G). However, osteoid volume was significantly increased in the HMWTg-IgG group compared with the Vector-IgG group. The increased osteoid volume in HMWTg mice was partially rescued with FGF23Ab treatment (Fig. 5H). To examine osteoblast function at the cellular level, dynamic bone histomorphometry was performed. As shown in Fig. 5I, intralabel thickness was decreased in the HMWTg-IgG group compared with the Vector-IgG group and was completely rescued with FGF23Ab treatment. MAR, an indicator of osteoblast activity, was significantly decreased in the HMWTg-IgG group, and was fully rescued with FGF23Ab treatment (Fig. 5J). There was no significant difference in mineral surface per bone surface among groups (Fig. 5K). Decreased bone formation rate in the HMWTg-IgG group was fully rescued with FGF23Ab treatment (Fig. 5L). Calcein and Xylenol orange labeling showed disorganized double labeling in HMWTg-IgG mice was partially rescued with FGF23Ab treatment (Fig. 5M, white arrow). Masson’s trichrome staining showed mineralization defect in HMWTg-IgG mice was partially rescued with FGF23Ab treatment (Fig. 5N, black arrow). These data suggest that FGF23 neutralization significantly reduced the mineralization defects and abnormalities in histomorphometric indices in HMWTg mice.

Fig. 5.

Bone histomorphometry analysis of femur metaphysis in HMWTg mice with long-term FGF23Ab treatment. Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Bone histomorphometry was performed on metaphysis of excised femur. (A) BV/TV, (B) trabecular number, (C) trabecular thickness, (D) trabecular spacing, (E) osteoclast number/bone surface, (F) osteoclast surface/bone surface, (G) osteoblast surface/bone surface, (H) osteoid volume/bone volume, (I) interlabel thickness, (J) mineral apposition rate, (K) mineral surface/bone surface, (L) bone formation rate/bone surface. (M) Calcein and Xylenol orange labeling showed disorganized double labeling in HMWTg-IgG mice was partially rescued with FGF23Ab treatment (white arrow). (N) Masson’s trichrome staining showed mineralization defect in HMWTg-IgG mice that was partially rescued with FGF23Ab treatment (black arrow). Data are means ± SE. n = 5 mice/group. *Vector-IgG versus HMW-IgG, p < 0.05; #Vector-IgG versus HMW-FGF23Ab, p < 0.05; @HMW-IgG versus HMW-FGF23Ab, p < 0.05.

Effects of long-term FGF23Ab treatment on expression of bone-related genes in HMWTg mice

To determine whether improved mineralization in FGF23Ab-treated HMWTg mice is associated with enhanced expression of genes involved in bone homeostasis, RNA was extracted from the flushed tibia shafts of long-term-treated Vector and HMWTg mice and RT-qPCR was performed. Fgf2 gene expression was high in HMWTg mice compared with Vector mice and was decreased by FGF23Ab treatment (Fig. 6A). FGFR1c mRNA was increased in the HMWTg-IgG group compared with the Vector-IgG group mice and there was a further increase by FGF23Ab treatment (Fig. 6B). There was no difference in FGFR3c mRNA level among groups (data not shown). Osteoclast-related gene Rankl mRNA was decreased in the HMWTg-IgG and HMWTG-FGF23Ab group compared with the Vector-IgG group (Fig. 6C) while Osteoprotegerin (Opg) mRNA level was increased in the HMWTg-FGF23Ab group compared with the Vector-IgG group (Fig. 6D). Rankl/Opg ratio, an important determinant of osteoclast formation, was decreased in the HMWTg-IgG group compared with the Vector-IgG group. FGF23Ab treatment did not alter the Rankl/Opg mRNA ratio in HMWTg mice (Fig. 6E). Runx2, a transcript factor that is important for osteoblast differentiation, was increased with FGF23Ab treatment in HMWTg mice (Fig. 6F). Col1a1 mRNA expression was increased in the HMWTg-IgG group compared with the Vector-IgG group, and there was a further increase in Col1al mRNA expression with FGF23Ab treatment in HMWTg mice (Fig. 6G). Osteocalcin (Ocn), a terminal osteoblast marker gene, was decreased in the HMWTg-IgG group compared with the Vector-IgG group and was rescued with FGF23Ab treatment (Fig. 6H).

Fig. 6.

Effect of long-term FGF23Ab treatment on expression of bone-related genes in flushed tibias from HMWTg mice. Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Mice were euthanized 48 hours after the last FGF23Ab administration. Flushed tibia bone shafts were used for RNA extraction. Quantitative real-time PCR analysis was performed for the following gene expression: (A) Fgf2, (B) Fgfr1c, (C) Rankl, (D) Opg, (E) Rankl/Opg, (F) Runx2, (G) Col1a1, (H) Ocn, (I) Opn, (J) Mgp, (K) Phex, (L) Mepe, (M) Dmp1, (N) Dmp4, (O) Enpp1, (P) Ank, and (Q) Slc20a1. Data are means ± SE. n = 5 mice/group. *Vector-IgG versus HMW-IgG, p < 0.05; #Vector-IgG versus HMW-FGF23Ab, p < 0.05; @HMW-IgG versus HMW-FGF23Ab, p < 0.05.

Because we observed partial but not complete rescue of the mineralization defect in response to long-term FGF23Ab treatment, the expression of matrix-related genes including osteopontin (OPN), matrix gla protein (Mgp), Phex, matrix extracellular phosphoglycoprotein (MEPE), dentin matrix protein-1 (DMP1), and dentin matrix protein-4 (DMP4), were determined. These genes are involved in mineralization of bone and teeth.^(27-29) Gene expression analysis showed that Opn mRNA level was similar between the Vector-IgG and HMWTg-IgG group. However, FGF23Ab treatment significantly increased Opn mRNA expression in HMWTg mice (Fig. 6I). The expression of Mgp, an inhibitor of mineralization, was significantly increased in the HMWTg-IgG and HMWTg-FGF23Ab groups (Fig. 6J). Phex and Mepe mRNA was increased in the HMWTg-IgG group compared with the Vector-IgG group, and the Phex and Mepe mRNA level was further increased with FGF23Ab treatment in HMWTg mice (Fig. 6K, L).

Dmp1 mRNA levels were decreased in both the HMWTg-IgG and HMWTg-FGF23Ab groups compared with the Vector-IgG group (Fig. 6M). Dmp4 mRNA was increased in the HMWTg-IgG group compared with the Vector-IgG group, and was further increased with FGF23Ab treatment in HMWTg mice (Fig. 6N).

Inorganic pyrophosphate (PPi) is a known inhibitor of matrix mineralization. Because FGF23Ab did not fully rescue the mineralizing defect in vivo, we assessed whether there were changes in genes that regulated PPi. A critical requirement for mineralization is the elaboration and transport of PPi into the extracellular matrix and its hydrolysis to Pi.^(30-32) The nucleoside triphosphate pyrophosphate hydrolase (ENPP1/PC-1) catalyzes ATP to generate PPi. The transporter ANK decreases intracellular PPi and increases extracellular PPi, while TNAP hydrolyses PPi to Pi that is necessary for crystal generation. In addition we examined the type III sodium-dependent Pi co-transporter Slc20a1 (Pit-1) that regulates Pi homeostasis in bone.⁽³³⁾ As shown in Fig. 6O and Q, there were significant increases in mRNA for Enpp1 and Slc2a1 in tibia shaft of the HMWTg-IgG group compared with the Vector-IgG group, and there was a further increase with FGF23Ab treatment in HMWTg mice. FGF23Ab treatment also significantly increased Ank mRNA expression in HMWTg mice (Fig. 6P). These data suggest that HMWFGF2 modulates these genes independent of FGF23.

Because TNAP hydrolyses PPi to Pi that is important for mineralization, we examined TNAP expression in bones of HMWTg mice. TNAP mRNA expression in tibia shaft was similar among groups (data not shown). We also analyzed TNAP protein expression in osteocytes and osteoblasts of femurs. Analysis of TNAP enzyme activity in femur sections showed decreased TNAP enzyme activity in osteocytes of cortical bone from HMWTg mice. Decreased TNAP in osteocytes of cortical bone was partially rescued with FGF23Ab treatment (Fig. 7A-C, E). However, TNAP activity (blue color) was similar in osteoblasts at the trabecular surface among groups (Fig. 7D, F). Serum TNAP concentration was decreased in HMWTg mice but was rescued with FGF23Ab treatment (Fig. 7G). Decreased TNAP enzyme activity in osteocytes of cortical bone from HMWTg mice could contribute to decreased serum TNAP level of HMWTg mice. Serum total ALP activity was significantly increased in HMWTg mice and was reduced by FGF23Ab treatment (Fig. 7H).

Fig. 7.

Effect of long-term FGF23Ab treatment on TANP activity of osteocytes and osteoblasts, serum TNAP protein concentration, and total ALP activity from HMWTg mice. (A–F) Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Mice were euthanized 48 hours after the last FGF23Ab administration. (A) Histochemical TNAP staining, (B) DAPI staining, and (C) merged TNAP and DAPI staining in osteocytes of cortical bone of femur sections. (D) Histochemical TNAP staining in osteoblasts on trabecular surface of femur sections. (E) Quantification of relative fluorescence in osteocytes. n = 3 mice/group. (F) Quantification of TNAP surface per bone surface on trabecular bone. n = 3 mice/group. (G, H) Starting at 18 days of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Mice were euthanized 48 hours after the last FGF23Ab administration for serum collection. (G) Serum TNAP protein concentration was measured by ELISA. (H) Serum total ALP activity was measured by ALP reagent. n = 8 mice/group. Data are means ± SE. *Vector-IgG versus HMW-IgG, p < 0.05; #Vector-IgG versus HMW-FGF23Ab, p < 0.05; @HMW-IgG versus HMW-FGF23Ab, p < 0.05.

Expression of pFGFR1, pERK, and ANK protein was also determined by immunohistochemistry staining of femurs. As shown in Fig. 8, there was increased pFGFR1 and pERK expression in femur cortical bone from the HMWTg-IgG group compared with the Vector-IgG group. FGF23Ab treatment reduced increased pFGFR1 and pERK in HMWTg mice. Increased ANK staining in osteocytes and osteoblasts in femurs of the HMWTg-IgG group was unchanged with FGF23Ab treatment.

Fig. 8.

Effect of long-term FGF23Ab treatment on expression of pFGFR1, pERK, and ANK protein expression in femur from HMWTg mice. Starting at 8 weeks of age, HMWTg female mice were treated with the FGF23Ab (10 mg/kg, intraperitoneal injection) or control IgG, two times/week for 6 weeks. Vector mice were treated with control IgG only. Mice were euthanized 48 hours after the last FGF23Ab administration. (A) pFGFR1, (B) pERK, and (C) ANK protein expression in femurs were determined by immunohistochemistry staining using undecalcified frozen section of femurs. There was increased pFGFR1 and pERK expression in femur cortical bone from HMWTg-FGF23Ab group compared with HMWTg-IgG group. Increased pFGFR1 and pERK expression in HMWTg mice was decreased with FGF23Ab treatment. There was increased ANK staining in osteocytes and osteoblast in femurs of HMWTg-IgG and HMWTg-FGF23Ab group.

Discussion

The findings of this in vivo study show that the osteomalacic bone and Pi wasting phenotypes of transgenic mice overexpressing HMWFGF2 isoforms in osteoblast lineage cells is due in part to increased FGF23/FGFR signaling that can be partially rescued by neutralizing FGF23Ab, which was shown to ameliorate rickets/osteomalacia and hypophosphatemia in the Hyp mice model⁽²³⁾ and XLH patients.⁽⁶⁾ In addition, overexpression of HMWFGF2 may also impair bone mineralization via upregulating PPi that is independent of FGF23/FGFR signaling.

Consistent with the studies using Hyp mice⁽²³⁾ that, interestingly, were reported to overexpress HMWFGF2 in osteocytes,⁽²²⁾ we observed that single-dose treatment with FGF23Ab increased serum Pi and reduced urine Pi excretion in HMWTg mice. This observation is consistent with our recent studies showing that FGF23Ab rescued Pi wasting via increased expression of Npt2a in kidneys of HMWTg male mice.⁽³⁴⁾ Consistent with our findings FGF23Ab also increased renal expression of Npt2a and decreased fractional excretion of Pi in Hyp mice.⁽²³⁾ Our studies show that Npt2a expression in the kidney was increased in HMWTg after single-dose FGF23Ab treatment. This suggests that the improvement in serum Pi levels was due to increased renal Pi reabsorption via Npt2a.

Because FGF23 is a potent regulator of kidney 1,25D production, we measured serum 1,25D levels after single-dose FGF23Ab treatment. Consistent with studies by Aono and colleagues⁽²³⁾ which showed a single dose of the FGF23Ab injection at 4 to 16 mg/kg increased serum 1,25D level in Hyp mice, we found that FGF23Ab caused a significant increase in 1,25D in serum of HMWTg mice that may contribute to increased intestinal absorption of Pi facilitating the correction of the hypophosphatemia in HMWTg mice. Similar to the effects of long-term treatment with FGF23Ab in Hyp mice,⁽²³⁾ long-term treatment with FGF23Ab completely normalized serum and urine Pi in HMWTg mice.

In contrast to the increase in body weight observed in Hyp mice with long-term FGF23Ab treatment,⁽²³⁾ FGF23Ab did not increase body weight in HMWTg mice. However, similar to its effect in Hyp mice, we observed a significant increase in longitudinal bone growth as well as a significant reduction in osteoid due to partial improvement of bone mineralization in HMWTg mice. FGF23Ab treatment of Vector mice did not affect body weight, skeletal growth, or BMD. Thus, in this preclinical study, FGF23Ab-mediated effects were mainly due to FGF23 neutralization, with no off-target effects being observed.

Because long-term treatment with FGF23Ab only partially improved bone phenotype in HMWTg mice, it is interesting to speculate that HMWFGF2 may be contributing to the failure of FGF23Ab to fully restore bone phenotype that is independent of its effects on FGF23 production. In support of additional effects of HMWFGF2 on matrix mineralization that are independent of FGF23, we previously conducted in vitro studies to assess the mechanism of the inhibitory effects of HMWFGF2 on bone marrow stromal cell differentiation and matrix mineralization and reported that neutralizing FGF23Ab and the FGFR inhibitor SU5402 only partially rescued the reduced mineralized bone nodule formation in HMWTg bone marrow stromal cell cultures.⁽²⁰⁾ Although it could be argued the partial correction of bone phenotypes was due to inadequate duration of the antibody administration, other investigators⁽³⁵⁾ showed only partial correction of bone phenotypes in Hyp mouse that were treated with FGF23Ab at a higher dose, more frequently, and for longer duration (35 mg/kg, three times per week for 10 weeks).⁽³⁵⁾

FGF23 is known to have effects on bone matrix mineralization through downregulation of the renal sodium/Pi co-transporter NPT2a resulting in reduced tubular reabsorption of Pi.⁽³⁶⁾ FGF23 is also known to have effects on bone matrix mineralization independent of its effects on serum Pi.^(37,38) This is supported by our previous publication that showed overexpression of HMWFGF2 increased Fgfr1c mRNA and nuclear accumulation of FGFR1 protein and increased p-ERK in cultured bone marrow stromal stem cells from HMWTg mice, and neutralizing FGF23Ab or MAPK inhibitor PD98056 partially rescued impaired bone nodule formation.⁽²⁰⁾ Studies also showed that FGF23 can phosphorylate and activate FGFR1 in osteoblasts.⁽³⁹⁾ Consistent with this in vitro data, in the current in vivo study we observed increased Fgfr1c mRNA expression in bones of HMWTg mice. However, treatment with FGF23Ab reduced increased pFGFR1 and pERK expression in femur cortical bone from HMWTg mice.

Neutralization of serum FGF23 with long-term FGF23Ab treatment in HMWTg mice did not fully rescue the bone mineralization defect, indicating that the bone abnormalities in HMWTg mice could not be solely attributed to local production of FGF23. In view of this, we reasoned that other matrix modulators might play a role in the bone phenotype of HMWTg mice. Small integrin-binding ligand N-linked glycoproteins (SIBLINGs), the family of non-collagen glycophosphoproteins comprising OPN, MGP, BSP, DMP1, MEPE, dentin sialophosphoprotein, and matrix extracellular phosphoglycoprotein, share many structural characteristics, are primarily located in bone and dentin, and play important roles in matrix mineralization.⁽⁴⁰⁾ Overexpression of MEPE in mice resulted in a growth and mineralization defect due to decrease in bone remodeling.⁽⁴⁰⁾ Matrix Gla protein (MGP), an extracellular matrix mineralization protein, is a potent inhibitor of mineralization.⁽⁴¹⁾ In the current study, long-term FGF23Ab-treatment mineralization defect improvement was accompanied by further increases in the expression of Mgp, Phex, Mepe, and Dmp4 mRNA, suggesting that FGF23Ab-improved mineralization in HMWTg mice is not through modulation of the SIBLING proteins. The observed increased levels of SIBLING expression may be a compensatory mechanism due to low FGF23 signaling resulting in incomplete rescue.

Another potential reason for the failure of FGF23Ab to fully rescue impaired bone mineralization in HMWTg mice may be due to further increased PPi. PPi is a known inhibitor of mineralization. A critical requirement for mineralization is the elaboration and transport of PPi into the extracellular matrix and its hydrolysis to Pi.^(30-32) The nucleoside triphosphate pyrophosphate hydrolase (ENPP1/PC-1) catalyzes ATP to generate PPi. The transporter ANK decreases intracellular PPi and increases extracellular PPi, while tissue nonspecific alkaline phosphatase (TNAP) hydrolyses PPi to Pi that is necessary for crystal generation. Studies using cell culture showed that FGF2 induces expression of ENPP1 that is mediated by RUNX2 and Msx2.^(42-45) FGF2 induced expression of ENPP1 and ANK while inhibiting expression of TNAP in calvarial osteoblast cell line MC3T3E1 (C4).⁽³⁰⁾ Other studies showed that FGF2 upregulated Ank, Enpp1, Mgp, Slc20a1, and Dmp1 in osteocytic MLO-Y4 cells.⁽³²⁾ Consistent with data published by Murali and colleagues⁽⁴⁶⁾ showing that TNAP expression is decreased in osteocytes of Hyp mice, in the current study our data showed that TNAP protein expression is decreased in osteocytes but not in osteoblasts of the HMW-IgG group compared with the Vector-IgG group in vivo which was partially rescued with FGF23Ab treatment. This is further confirmed by reduced serum TNAP concentration. There were no differences in Tnap mRNA among the groups which could be because the RNA was extracted from flushed tibia shaft that includes a mixture of osteocytes and osteoblasts. Interestingly, in this study decreased serum TNAP concentration was accompanied with increased serum total ALP activity in HMWTg mice. Consistent with earlier studies in the Hyp mouse,^(46,47) we observed a significant increase in serum total ALP activity using an ALP reagent kit. In these earlier studies in Hyp mouse total ALP but not TNAP was measured. However, we also utilized a TNAP isoenzyme-specific ELISA kit that demonstrated decreased serum TNAP concentration consistent with the observed decrease in TNAP in the osteocytes of the HMWTg mice. To our knowledge, there are no published data on serum TNAP levels in Hyp mice. Thus our data are the first report of decreased TNAP in HMWTg mice. Our in vivo study shows that there were significant increases in mRNA for Enpp1, Ank, and Slc20a in the HMWTg-IgG group, and long-term FGF23Ab treatment further increased Enpp1, Ank, and Slc20a expression, indicating potential increased extracellular PPi accumulation that inhibits mineralization in HMWTg mice. Furthermore, these data suggest that HMWFGF2 may modulate these genes independent of FGF23.

Based on the results from the current and studies by others,⁽⁴⁸⁾ we have proposed a model of how overexpression of HMWFGF2 impairs bone mineralization through FGF23 dependent and independent mechanisms, as shown in Fig. 9. HMWFGF2 stimulates FGF23 promoter activity through binding of integrative FGFR1 and cAMP-response element-binding protein (CREB) to a conserved cAMP response element (CRE) in the FGF23 promoter⁽⁴⁸⁾ that results in increased FGF23 in circulation. Increased FGF23 leads to impaired bone mineralization due to hypophosphatemia,⁽²³⁾ inhibition of osteoblast terminal differentiation and mineralization–related genes through cell-surface FGFR1,^(20,39) and inhibition of TNAP expression⁽⁴⁹⁾ that results in PPi accumulation. Overexpression of HMWFGF2 also inhibits bone mineralization independent of FGF23 by upregulating ENPP1, ANK, and SLC20A potentially through binding of HMWFGF2 on their promoter region that results in further PPi accumulation because full CRE and half CRE sites were predicted in the promoter region of these genes using Ensembl.

Fig. 9.

Schematic model depicting how overexpression of HMWFGF2 impairs bone mineralization through FGF23 dependent and independent mechanisms. HMWFGF2 stimulates FGF23 promoter activity through binding of integrative FGFR1 and CREB to a conserved CRE in the FGF23 promoter⁽⁴⁸⁾ that results in increased FGF23 in the circulation. Increased FGF23 leads to impaired bone mineralization due to hypophosphatemia,⁽²³⁾ inhibition of osteoblast terminal differentiation–related and mineralization-related genes through cell surface FGFR1,^(20,39) and inhibition of TNAP expression⁽⁴⁹⁾ that results in PPi accumulation. Overexpression of HMWFGF2 also inhibits bone mineralization independent of FGF23 by upregulating ENPP1, ANK, and SLC20A potentially through binding of HMWFGF2 on their promoter region that results in further PPi accumulation since full CRE or half CRE sites were predicted in the promoter region of these genes using Ensembl. CREB = cAMP-response element-binding protein; CRE = cAMP response element.

In summary, our studies demonstrate partially improved bone mineralization in HMWTg mice with neutralizing FGF23Ab treatment. This suggests that FGF23Ab attenuates FGF23-dependent consequences of HMWFGF2 overexpression that may be critical determinants of skeletal pathology in HMWTg mice. Further study is warranted to examine whether a combination of neutralizing FGF23 and blocking accumulation of PPi will be better to improve mineralization in HMWTg that could also be of translational relevance to mineralization abnormalities in Hyp mouse or XLH subjects.