Increased PHOSPHO1 expression mediates cortical bone mineral density in renal osteodystrophy

Patients with advanced chronic kidney disease (CKD) often present with skeletal abnormalities, a condition known as renal osteodystrophy (ROD). While tissue non-specific alkaline phosphatase (TNAP) and PHOSPHO1 are critical for bone mineralization, their role in the etiology of ROD is unclear. To address this, ROD was induced in both WT and Phospho1 knockout (P1KO) mice through dietary adenine supplementation. The mice presented with hyperphosphatemia, hyperparathyroidism, and elevated levels of FGF23 and bone turnover markers. In particular, we noted that in CKD mice, bone mineral density (BMD) was increased in cortical bone (P  < 0.05) but decreased in trabecular bone (P  < 0.05). These changes were accompanied by decreased TNAP (P  < 0.01) and increased PHOSPHO1 (P  < 0.001) expression in WT CKD bones. In P1KO CKD mice, the cortical BMD phenotype was rescued, suggesting that the increased cortical BMD of CKD mice was driven by increased PHOSPHO1 expression. Other structural parameters were also improved in P1KO CKD mice. We further investigated the driver of the mineralization defects, by studying the effects of FGF23, PTH, and phosphate administration on PHOSPHO1 and TNAP expression by primary murine osteoblasts. We found both PHOSPHO1 and TNAP expressions to be downregulated in response to phosphate and PTH. The in vitro data suggest that the TNAP reduction in CKD-MBD is driven by the hyperphosphatemia and/or hyperparathyroidism noted in these mice, while the higher PHOSPHO1 expression may be a compensatory mechanism. Increased PHOSPHO1 expression in ROD may contribute to the disordered skeletal mineralization characteristic of this progressive disorder.


Introduction
Chronic kidney disease (CKD) is a disorder characterized by progressive loss of kidney function over time. Patients with advanced CKD frequently develop disturbances of mineral and bone metabolism and fail to maintain normal systemic levels of calcium (Ca), inorganic phosphate (Pi), parathyroid hormone (PTH), and fibroblastic growth factor-23 (FGF23) (Moe et al. 2006). Hyperphosphatemia, hyperparathyroidism, and elevated FGF-23 are the primary indicators for the diagnosis of CKD-mineral bone disorder (CKD-MBD) which develops in the early stages of CKD and disease progression can result in cardiovascular disease and renal osteodystrophy (ROD) -the skeletal pathology component of the CKD-MBD syndrome (Fang et al. 2014). The current classification system and treatment strategy for ROD are based on changes to bone turnover, mineralization, and volume (Kazama et al. 2013). A decrease in bone mineral density (BMD) is particularly common in patients with late-stage disease (Nickolas et al. 2013) but animal models have shown a more varied response (Bajwa et al. 2018;Lau et al. 2013a;Metzger et al. 2021). The other ROD-associated skeletal pathologies, have been attributed to CKD-related metabolic and hormonal disturbances (Zheng et al. 2016).
Although the precise mechanisms responsible for the impaired skeletal mineralization observed in ROD remain unclear, the origins are likely to involve a complex interplay between bone and the altered endocrine milieu. Phosphorus retention, due to the failing kidney, leads to chronically elevated concentrations of circulating FGF23 in an attempt to normalize serum Pi levels through enhanced urinary secretion and decreased intestinal absorption (Mirza et al. 2009). This is achieved by the inhibition of renal 1α-hydroxylase and stimulation of 24-hydroxylase but the resulting reduction in circulating levels of 1,25(OH) 2 D 3 contributes to hypocalcemia and secondary hyperparathyroidism (SHPT) (Shimada et al. 2004). SHPT promotes bone resorption by increasing the receptor activator of nuclear factor-B ligand (RANKL)/osteoprotegerin (OPG) ratio (Ma et al. 2001). The bone formed during rapid remodeling is both immature and poorly mineralized (Graciolli et al. 2017).
Indeed, the mineralization status may be dependent on the prevailing serum PTH concentrations which could explain the various mineralization states reported in ROD (Lau et al. 2013b;Miller et al. 1998). It is also possible that altered endocrine factors may directly target the expression of key phosphatases critical for skeletal mineralization. Specifically, FGF23 may inhibit matrix mineralization by suppressing TNAP expression and activity by osteoblasts resulting in the accumulation of the mineralization inhibitor, pyrophosphate (PPi) (Murali et al. 2016b). Also, PTH may induce a rapid downregulation of Phospho1 gene expression in osteogenic cells and bone marrow stromal cell lines (Chande and Bergwitz 2018;Houston et al. 2016).
Despite clear links between both TNAP and PHOSPHO1 in the control of skeletal mineralization, their roles in ROD remain unclear. PHOSPHO1 and TNAP are two of the most widely studied phosphatases involved in skeletal mineralization (Dillon et al. 2019). PHOSPHO1 is expressed at sites of mineralization and liberates Pi from phospholipid substrates for incorporation into the mineral phase (Roberts et al. 2007). Phospho1 deficient mice exhibit decreased BMD, compromised trabecular and cortical bone microarchitecture, and spontaneous greenstick fractures (Boyde et al. 2017). TNAP is an ectoenzyme and hydrolyzes PPi to allow the propagation of hydroxyapatite in the extracellular matrix (ECM), beyond the confines of the matrix vesicle membrane (Hessle et al. 2002). Mice deficient in TNAP (Alpl -/-) phenocopy infantile hypophosphatasia (HPP), an inborn error of metabolism resulting in rickets and osteomalacia (Whyte 2008). A complete absence of ECM mineralization is observed in Phospho1 -/-; Alpl -/double knockout mice and in murine metatarsals cultured in the presence of PHOSPHO1 and TNAP inhibitors demonstrating the functional co-operativity of PHOSPHO1 and TNAP for bone mineralization (Huesa et al. 2015;Yadav et al. 2011).
Despite great advances in understanding the mechanisms responsible for the altered mineralization status noted in ROD, the involvement of phosphatases is unclear.
Therefore, in this study, we examined changes in the expression of PHOSPHO1 and TNAP and bone architecture in long bones using the well-established adenine-induced model of CKD (Jia et al. 2013). We also examined the effects of PTH, FGF23, and Pi on TNAP and PHOSPHO1 expression in primary osteoblasts. Our findings support a specific role for PHOSPHO1particualarly in the altered cortical bone mineralization status in ROD.

Materials and methods
All reagents were from Sigma-Aldrich (Dorset, UK) or less otherwise stated.
2.1. Mice 6 C57BL/6 male mice (Charles River Laboratories, Currie, UK) and were used in the first in vivo study. Male Phospho1 knockout (P1KO) mice and wild-type (WT) controls, maintained on a C57BL/6 background were generated and genotyped as previously described (Yadav et al. 2011) and used in the second in vivo study. At 8-weeks of age, mice were randomly assigned a control (n = 12) or CKD (n = 12) diet ( Fig. S1A). Mice losing more than 30% of their body weight were euthanized by exposure to CO 2 and confirmed dead by cervical dislocation. All animal experiments were approved by the Roslin Institute's named veterinary surgeon and named animal care and welfare officer (NACWO), with animals maintained in accordance with the Home Office code of practice (for the housing and care of animals bred, supplied, ARRIVE guidelines or used for scientific purposes).

CKD diet and tissue collection
CKD was induced by feeding a casein-based diet containing 0.6% calcium, 0.9% phosphate, and 0.2% adenine (Envigo, Teklad Co. Ltd). Control mice received the same diet without adenine. At 13 weeks of age, all animals were sacrificed, and blood was obtained by cardiac puncture under terminal anesthesia. Femora, tibiae, and kidneys were harvested and processed and stored accordingly.

2.3.Serum and urine biochemistry
Serum blood urea nitrogen (BUN), creatinine (Cr), Ca, Pi, and alkaline phosphatase (ALP) activity were quantified using a biochemistry analyzer (Beckman Coulter AU480, Olympus). Intact PTH (Pathway Diagnostics, Dorking, UK), FGF23 (Kainos Laboratories, Inc. Japan), N-terminal propeptide of human procollagen type I (P1NP) and carboxy-terminal telopeptide of type I collagen (αCTX) (Wuhan Fine Biotech, Wuhan, China), levels were determined by ELISA according to manufacturers' instructions. Hydrophobic bedding, LabSand (Coastline Global, CA, USA) was used to collect urine samples from which. the concentration of Cr and albumin were determined by semi-quantitative test strips (Microalbustix, Siemens) and the specific gravity (SG) was determined by a manual refractometer.

Histopathological analysis of kidney and bone tissues
The right tibiae and kidneys were fixed in 4% paraformaldehyde (PFA, for 24 hrs) and stored in 70% ethanol. Kidneys were processed to paraffin wax using standard procedures. Hematoxylin and eosin (H&E), Masson's trichrome, and von Kossa staining were performed according to standard methods. Histopathological scoring of renal interstitial inflammation, tubular atrophy, protein casts, and renal fibrosis was defined as 0 = normal; 1 = mild, involvement of < 25% of the cortex; 2 = moderate, involvement of 25 to 50% of the cortex; 3 = severe, involvement of 50 to 75% of the cortex; 4 = extensive, involvement of > 75% of the cortex. Bones were decalcified in 10% ethylenediaminetetraacetic acid (EDTA; pH 7.4) for 14 days at 4°C and processed to paraffin wax. Sections were stained by Goldner's Trichrome and reacted for tartrate-resistant acid phosphatase (TRAP). Bone histomorphometry was quantified using the BioQuant Osteo software (BIOQUANT Image Analysis Corporation, Texas, USA) using the approved ASBMR histomorphometry nomenclature (three sections/bone: six randomly selected bones from each group).
In the proximal tibial metaphysis the volume of interest (VOI) extended distally 5% from the bottom of the growth plate excluding the cortical shell. A total of 250 slices beneath this 5% were selected to exclude the primary spongiosa. In the first in vivo study, whole bone cortical analysis was performed on datasets derived from whole μ CT scans using BoneJ (version 1.13.14) a plugin for ImageJ. Following segmentation, alignment, and removal of fibula from the dataset, a minimum bone threshold was selected for each bone to separate higher density bone from soft tissues and air. The most proximal and the most distal 10% portions of tibial length were excluded from analysis, as these regions include trabecular bone. In the 2 nd in vivo study cortical analysis was performed on datasets derived from μ CT scan images at 50% of the total tibial length from the top of the tibia. BMD phantoms of known calcium hydroxyapatite mineral densities of 0.25 and 0.75 g/cm 3 were scanned and reconstructed using the same parameters as used for bone samples.

Cell viability and cytotoxicity assay
To assess the effects of Pi on osteoblast viability, the AlamarBlue assay (Thermo Fisher Scientific, Loughborough, UK), and lactate dehydrogenase (LDH) CytoTox 96 cytotoxicity assay (Promega, Southampton, UK) were performed according to manufacturer's instructions.

RNA extraction and quantitative polymerase chain reaction
The distal and proximal epiphyses of the left femorae were excised, and the diaphyseal bone marrow was removed by centrifugation at 13,000 x g for 10 mins at 4°C. The resultant cortical shafts were homogenized using a Rotor-Stator Homogenizer (Ultra-Turrax T10). RNA extraction from the homogenized bone and cultured osteoblasts was completed using the RNeasy kit (Qiagen). The RNA concentration was determined using a NanoDrop spectrophotometer (Fisher Scientific, Loughborough, UK) at a wavelength of 260 nm, and RNA purity was evaluated by the 260/280 nm ratio. RNA was reverse transcribed to complementary DNA (cDNA) using Superscript II (Invitrogen). All genes were analyzed with the SYBR green detection method (PCR Biosystems, UK) using the Stratagene Mx3000P real time QPCR system (Agilent Technologies, Santa Clara, USA). Gene expression data were normalized against housekeeping genes (Gapdh in primary osteoblasts and Atp5b in bone tissue) using MxPro software (Cheshire, UK). The relative expression of the analyzed genes was calculated and expressed as a fold change compared to control values. Primer sequences are listed in Supplementary Table S1.

Protein extraction and isolation from brush border membrane vesicles (BBMV) of kidney
Kidneys were

Western blot analysis
Protein from cultured osteoblasts and right femorae diaphyseal cortical bone (with marrow removed) was extracted in RIPA buffer containing protease inhibitor cocktail after homogenization. Protein concentrations were determined using the BCA protein assay kit (Life Technologies). Proteins were separated using a 10% Bis-Tris protein gel (Thermo Fisher Scientific). After blocking in 5% skimmed milk/Tris-buffered saline with Tween 20 (TBST) or LI-COR buffer at room temperature (RT) for 1 hour, the membranes were incubated sequentially with primary and secondary antibodies (Tables S2 and S3). Western blot analysis of proteins from primary osteoblasts was performed using the Odyssey infrared detection system (LI-COR). Western blot analysis of proteins from bone tissues was undertaken using the ultra-sensitive enhanced chemiluminescence detection system (Thermo Fisher Scientific). The blots were imaged by the GeneGnome XRQ chemiluminescence imaging system (Syngene, UK). Densitometry of the protein bands was analyzed with Image J software (NIH) for quantification.

Quantification of ECM mineralization
Cultured osteoblasts were fixed in 4% PFA for 10 mins at RT and stained with aqueous 2% (w/v) Alizarin red solution for 10 mins at RT. The bound stain was solubilized in 10% cetylpyridinium chloride and the optical density was measured by spectrophotometry at 570 nm.

Statistical analysis
Quantitative data are expressed as the mean ± standard error of the mean (SEM) of at least three biological replicates per experiment. The precise number (n) is indicated in the relevant table and figure legends. Statistical analysis was performed using a two-tailed Student's t-test or one-way analysis of variance (ANOVA) followed by Tukey's range test, as appropriate. Statistical analysis was implemented by the GraphPad Prism software. A p < 0.05 was considered to be significant and noted as "*". p values of < 0.01, < 0.001, and < 0.0001 were noted as "**", "***" and "****" respectively.

Verification of the CKD-MBD mouse model
Before investigating TNAP and PHOSPHO1 expression in experimental ROD we first confirmed that our mouse model presents with the characteristic serum biochemistries and kidney pathology of CKD-MBD. The CKD mice lost bodyweight and presented with the expected changes to serum and urine analyte levels at the end of the study ( Fig. S1B and Table 1). The kidneys of CKD mice presented with various pathologies including tubular atrophy, protein casts, interstitial inflammation, and renal fibrosis ( Fig S2). Furthermore, transcripts encoding kidney injury biomarkers

Cortical BMD is increased in CKD mice and influenced by PHOSPHO1 status
Trabecular BMD, bone volume/tissue volume (BV/TV), thickness (Th), number (N), structural model index (SMI) and connectivity density (Conn Dn.) of the tibiae were all decreased in CKD mice when compared to controls (Fig. 2). Cortical bone parameters were also altered in CKD mice; cortical BMD was generally increased whereas cross-sectional area (CSA), cortical thickness, resistance to torsion, and Imin and Imax were all generally lower over the entire tibial length of CKD mice (Figs. 3a,c,d,f,i,j). Consistent with the thinner cortex the medullary area, the endosteal perimeter was increased and the periosteal perimeter decreased in the CKD mice (Figs.  (Table 1).
The increased cortical BMD in CKD-MBD mice (Fig. 3a) aligns with the higher PHOSPHO1 expression in the cortical bone shafts, despite being an unexpected finding in the CKD-MBD mice (Figs. 1A-C). To explore this further, we next examined bone from PHOSPHO1 deficient (P1KO) mice maintained on the 0.2% adenine supplemented diet for 5 weeks. Cortical analysis was performed on datasets derived from μ CT scan images at 50% of the total tibial length as this region of bone from CKD mice had a higher BMD than control counterparts (Fig 3a). As previously noted, (Fig. 3a), the cortical BMD of WT CKD mice was increased compared to WT control mice but, in contrast, no such increase was apparent in P1KO CKD mice, which had a BMD similar to their respective P1KO controls but as expected lower than the BMD of WT control mice (Fig. 4). Structural cortical bone changes were also influenced by the absence of PHOSPHO1 in the P1KO mice; the CKD-induced increases in porosity and decreases in BV/TV, CSA, and Th noted in WT CKD mice were all blunted in P1KO CKD mice compared to P1KO control mice (Fig. 4).

Pi, PTH, and FGF23 perturb ECM mineralization and the expression of key mineralization markers in primary osteoblasts
To determine the causes of the mineralization defects noted in the CKD mice, we At concentrations of 2 mM and above, Pi significantly down-regulated Phospho1, Alpl, and Bglap mRNA expression (p < 0.01, Fig. 5A). In contrast, Enpp1, Spp1 and Slc20a1 expression was increased at the higher Pi concentrations (p < 0.05, Fig. 5A).
Cell viability as assessed by Alamar blue and LDH release was unaffected at all Pi concentrations tested (Fig. S6). PHOSPHO1 and TNAP protein expression were also inhibited by increasing Pi concentrations whereas the addition of Pi, 3 mM and above, increased the formation of mineralized bone nodules in a dose-dependent manner (p < 0.001, Figs. 5B, C).
Administration of PTH at > 5 nM downregulated the expression of Phospho1, Alpl and Bglap (p < 0.01, Fig. 6A). Enpp1, Slc20a2, and Runx2 expression were also decreased but only at higher PTH concentrations (p < 0.05, Fig. 6A). Reduction of PHOSPHO1 and TNAP protein expression by increasing PTH concentrations mirrored the changes in gene expression (Fig. 6B). The addition of PTH inhibited ECM mineralization and this was noted at concentrations as low as 0.5 nM.
Exposure to FGF23 had little effect on the expression of the genes studied although both Phospho1 and Alpl expression were decreased but only at the highest FGF23 concentrations (p < 0.05, Fig. 7A). The addition of Klotho to the FGF23 supplemented cultures had no further effects on gene expression when compared with FGF23 alone (data not shown). A similar trend was also noted at the protein level where PHOSPHO1 and TNAP expression decreased in a FGF23 concentration-dependent manner but this change did not reach statistical significance from control-treated cultures (Fig. 7B). A similar response was observed in the presence of FGF23 and Klotho (data not shown). FGF23 with or without klotho had no effects on ECM mineralization of primary osteoblasts at the concentrations tested ( Fig. 7C and data not shown).

Discussion
This study has shown that PHOSPHO1 and TNAP, two phosphatases required for bone mineralization, have altered expression in ROD. Specifically, the ROD phenotype was characterized by increased cortical BMD and this response may be mediated by increased PHOSPHO1 expression. However, the altered PHOSPHO1 expression is unlikely to be a direct result of the altered endocrine milieu as PTH, mice. The loss of bone is likely to be multi-factorial but PTH enhanced bone resorption via altered RANKL and OPG expression is likely to predominate (Ma et al. 2001). In the early stages of CKD, the low bone turnover disease results from bone cell inactivity due to PTH resistance, as well as reduced calcitriol levels, and accumulation of uremic toxins (Couttenye et al. 1999). When renal function further deteriorates, the chronically increased PTH levels overcome peripheral PTH resistance and activate the indolent bone cells, leading to high turnover bone disease (Drüeke and Massy 2016). Bone resorption predominates in both high and low bone turnover disease and the resultant elevated serum Ca and Pi levels promote bone extra-skeletal (vascular) calcification (Zheng et al. 2016). In agreement with the results of this present study, others have also reported increased cortical porosity and compromised bone architecture in CKD animal models although inconsistent effects on the cortical and trabecular compartments have been reported (Jia et al. 2013;Metzger et al. 2021;Miller et al. 1998;Ogirima et al. 2006). Although patients with CKD have been reported to have lower cortical BMD inconsistencies in trabecular and cortical BMD in CKD animal models also exist (Lau et al. 2013b;Nickolas et al. 2013). Specifically, in a mouse nephrectomy model in which serum Pi levels were unchanged, trabecular and cortical BMD were increased and decreased, respectively which was the opposite to that found in this present study (Lau et al. 2013b). The increased trabecular BMD was not influenced by dietary phosphate content whereas the decreased cortical BMD was only noted in mice fed a high phosphate (0.9%) and not a normal phosphate (0.5%) containing diet (Lau et al. 2013b). In this present study, mice were fed 0.9% phosphate-containing diet and analysis revealed that at no location along the entire cortical bone shaft was BMD lower in the CKD mice. The spectrum of bone phenotypes reported in CKD-MBD models may reflect the differing serum PTH levels at the point of the study, as progressive SHPT is linked with different effects on bone quality and structure (Miller et al. 1998). Furthermore, whether differential expression of PHOSPHO1 and TNAP within the trabecular and cortical bone compartments contributes to the divergent BMD response is unclear and requires further investigation.
The high bone turnover status in SHPT will contribute to bone that is less mineralized, a hallmark of stage 4 and 5 CKD, and lead to reduced mechanical strength and increased risk of fractures (Drüeke and Massy 2016). Similarly, in this present study, PTH-induced skeletal remodeling is likely to, at least in part, explain the poorly mineralized trabecular bone noted in this study although PTH exposure can also inhibit osteoblast differentiation and thus indirectly delay osteoid production and matrix mineralization (Qin et al. 2004). Additionally, the CKD-driven increase in osteocyte secretion of Wnt/β-catenin signaling inhibitors, such as FGF23, dickkopf 1, and sclerostin may negatively affect osteoblast function and contribute to the mineralization defect in ROD (Evenepoel et al. 2015;Murali et al. 2016a). The results of this present study offer changes to PHOSPHO1 and TNAP osteoblast expression as an additional/alternative explanation for the altered bone mineralization status associated with ROD.
Monitoring serum ALP has been regarded as a useful serum marker of bone turnover in ROD however its expression in bone, functioning as a phosphatase capable of mineralizing osteoid has not to our knowledge been explored in the pathogenesis of ROD (Bervoets et al. 2003). The decreased Alpl expression in CKD cortical bone was not however consistent with the observed increased cortical BMD and we hypothesize that the latter is possibly driven by increased PHOSPHO1 expression (Huesa et al. 2015). To examine this further we determined cortical BMD and other structural parameters in P1KO CKD mice. In the absence of PHOSPHO1, cortical BMD in control mice was decreased as previously reported and the increased BMD in cortical bone of CKD wild-type mice was not observed in the P1KO CKD mice (Yadav et al. 2011). Furthermore, other structural parameters such as cortical porosity, thickness and CSA were also improved in P1KO CKD mice. Whilst supportive of our Several studies have reported that mineralizing cells including odontoblasts, chondrocytes, and osteoblasts are sensitive to P i and respond by altering the expression of mineralization-associated genes and transcription factors (Foster et al. 2006;Khoshniat et al. 2011). The regulation of biomineralization by Pi may be related to its ability to stimulate MV release and/or the accumulation of type III NaP(i) transporter (PiT1) in osteoblasts promoting Pi uptake and ECM mineralization (Foster et al. 2006;Yoshiko et al. 2007). In relation to this present study, the availability of exogenous Pi to promote osteoblast matrix mineralization by-passes the requirement for Pi production from phosphocholine and phosphoethanolamine by PHOSPHO1 and PPi by TNAP (Ciancaglini et al. 2010;Houston et al. 2004;Roberts et al. 2007) and may explain the concentration-dependent decrease in PHOSPHO1 and TNAP expression by exogenous Pi as part of a negative feedback mechanism (Foster et al. 2006). In this regard, human PHOSPHO1 shares approximately 30% homology at the amino acid level with a tomato phosphate starvation-induced gene product, LePS2, which possesses phosphatase activity that can convert organic phosphorus into available Pi. Intriguingly, LePS2 expression is tightly and negatively regulated by Pi availability and is thus induced in the absence, but repressed in the presence of Pi (Stenzel et al. 2003). It is unknown if such a Pi negative feedback mechanism controls PHOSPHO1 expression but the increased osteoclast resorption observed in ROD will bring about the release of Pi which will contribute to the observed hyperphosphatemia and impede the skeleton from exerting its normal reservoir function when serum Pi concentrations increase (Hruska et al. 2008). In such a scenario, the resulting Pi stress conditions experienced by the skeleton may drive higher PHOSPHO1 expression in a similar way to the LePS2 protein and other phosphatases such as OsACP1 a PHOSPHO1-like acid phosphatase in rice (Deng et al. 2022).
In summary, this study has identified PHOSPHO1 as a possible mediator in the development of the cortical bone phenotype in ROD, thus providing a foundation for future research to explore potential therapies to improve bone health in CKD-MBD.

Transparency document
The Transparency document associated with this article can be found, in online version.

Declaration of competing interest
The authors declare that they have no competing interest.

Ethical Approval
All experimental protocols were approved by Roslin Institute's Animal Users Committee and the animals were maintained in accordance with UK Home Office guidelines for the care and use of laboratory animals, and with the ARRIVE guidelines.

Acknowledgments
We thank Elaine Seawright and Yao- Tang   BUN, blood urea nitrogen; Cr, creatinine; Ca, calcium; Pi, phosphorus; ALP, alkaline phosphatase; PTH, parathyroid hormone; FGF23, fibroblast growth factor 23; SG, specific gravity; UACR, urine albumin/urine creatinine ratio; UCr/SCr, urine creatinine/serum creatinine. Four mice on the CKD diet lost > 30% bodyweight between 4 and 5 weeks and were removed from the study. The data are represented as the means ± SEM.   Mice in the CKD-MBD group were fed a casein-based diet containing 0.2% adenine diet for 5 weeks. Mice in the CTL group were fed a casein-based diet without adenine.

Figure legends
(B) The bodyweight of the CKD-MBD mice progressively decreased during the 5-weeks on the adenine supplemented diet. The data are represented as the means ± SEM. **** p < 0.0001 as compared to CTL mice of the same age. Note, 4 mice in the CKD-MBD group lost more than 30% of their body weight and were removed from the study.  The data are represented as the mean ± SEM (n ≥ 3); NS, not significance from the 1 mM control group.