PKB/SGK-Resistant GSK3 Enhances Phosphaturia and Calciuria

BASIC RESEARCH www.jasn.org

Michael Fo¨ller,* Daniela S. Kempe,* Krishna M. Boini,† Ganesh Pathare,* Balasaheb Siraskar,* Paola Capuano,‡ Ioana Alesutan,* Mentor Sopjani,* Gerti Stange,‡ Nilufar Mohebbi,‡ Madhuri Bhandaru,* Teresa F. Ackermann,* Martin S. Judenhofer,§ Bernd J. Pichler,§ Ju¨rg Biber,‡ Carsten A. Wagner,‡ and Florian Lang‡

*Department of Physiology and §Laboratory for Preclinical Imaging and Imaging Technology of the Werner Siemens-Foundation, University of Tu¨bingen, Tu¨bingen, Germany; †Department of Pharmacology and Toxicology, Medical College of Virginia Campus, Virginia Commonwealth University, Richmond, Virginia; and ‡Institute of Physiology and Zurich Center for Integrative Human Physiology, University of Zurich, Zurich, Switzerland

ABSTRACT Insulin and IGF1-dependent signaling activates protein kinase B and serum and glucocorticoid inducible kinase (PKB/SGK), which together phosphorylate and inactivate glycogen synthase kinase GSK3. Be- cause insulin and IGF1 increase renal tubular calcium and phosphorus reabsorption, we examined GSK3 regulation of phosphate transporter activity and determined whether PKB/SGK inactivates GSK3 to enhance renal phosphate and calcium transport. Overexpression of GSK3 and the phosphate transporter NaPi-IIa in Xenopus oocytes decreased electrogenic phosphate transport compared with NaPi-IIa– expressing oocytes. PKB/SGK serine phosphorylation sites in GSK3 were mutated to alanine to create gsk3KI mice resistant to PKB/SGK inactivation. Compared with wildtype animals, gsk3KI animals exhib- ited greater urinary phosphate and calcium clearances with higher excretion rates and lower plasma concentrations. Isolated brush border membranes from gsk3KI mice showed less sodium-dependent phosphate transport and Na-phosphate co-transporter expression. Parathyroid hormone, 1,25-OH vitamin D levels, and bone mineral density were decreased in gsk3KI mice, suggesting a global dysregulation of bone mineral metabolism. Taken together, PKB/SGK phosphorylation of GSK3 increases phosphate transporter activity and reduces renal calcium and phosphate loss.

J Am Soc Nephrol 22: 873–880, 2011. doi: 10.1681/ASN.2010070757

The kidneys play a central role in the regulation of tential channel V5 (TRPV5) calcium channel ex- mineral homeostasis by mediating excretion or reab- pressed in the luminal membrane of the distal convo- sorption, respectively, of phosphate, calcium, and luted tubule and connecting tubule.11–13 The magnesium. Phosphate reabsorption occurs in the expression and activity of TRPV5 is regulated by sev- proximal tubule and is mediated by at least three dis- eral factors similarly regulating renal phosphate trans- tinct sodium-dependent phosphate cotransporters, port, such as dietary calcium intake, PTH, klotho, ac- namely NaPi-IIa (SLC34A1), NaPi-IIc (SLC34A3), and Pit-2 (SLC20A2), located in the apical brush bor- Received July 22, 2010. Accepted December 27, 2010. der membrane.1–3 Renal phosphate reabsorption is regulated by various factors including dietary phos- Published online ahead of print. Publication date available at www.jasn.org. phate intake, acid-base status, and various hormones, Correspondence: Dr. Florian Lang, Department of Physiology, such as parathyroid hormone (PTH), 1,25-(OH)2 vi- University of Tu¨bingen, Gmelinstrasse 5, D-72076 Tu¨bingen, Ger- tamin D3, fibroblast growth factor 23, insulin, and in- many. Phone: 49-7071-297-2194; Fax: 49-7071-295-618; E-mail: sulin-like growth factor 1 (IGF1).4–10 Active calcium [email protected] reabsorption is mediated by the transient receptor po- Copyright © 2011 by the American Society of Nephrology

J Am Soc Nephrol 22: 873–880, 2011 ISSN : 1046-6673/2205-873 873 BASIC RESEARCH www.jasn.org

11–13 id-base status, and 1,25-(OH)2 vitamin D3. The role of insulin and IGF1 has remained controversial.14,15 The intracellular signaling cascades mediating the effects of these hormones on renal phosphate transporters are still incom- pletely understood. Signaling of the phosphaturic hormone PTH involves the protein kinases A and C and extracellular signal-regulated kinase,16 leading to the internalization and degradation of the NaPi-IIa cotransporter in the mouse and rat kidney.17 Signal- ing mediating the stimulating effect of insulin and IGF1 on renal phosphate reabsorption4–6 has remained ill-defined. Signaling of insulin includes stimulation of the PI3 kinase pathway with subsequent activation of protein kinase B (PKB/Akt) and the serum- and glucocorticoid-inducible kinase (SGK) isoforms.18,19 Both, Figure 1. Coexpression of GSK3 inhibits electrogenic phosphate PKB20,21 and SGK22,23 isoforms are known to phosphorylate and transport in NaPi-IIa–expressing Xenopus oocytes. (A) Arithmetic Ϯ ϭ thus to inhibit the glycogen synthase kinase GSK3. However, means SEM (n 13 to 21) of phosphate (2 mM)-induced inward nothing is known about the regulation of epithelial phosphate and currents (IPi)inXenopus oocytes injected with water (left bar), NaPi-IIa cRNA (middle bar), or NaPi-IIa and GSK3␤ cRNA (right Ca2ϩ transport by GSK3. bar). ***Statistically significant difference from absence of NaPi-IIa This study aimed to define the role of PKB/SGK-dependent cRNA (P Ͻ 0.001). #Difference from absence of GSK3␤ cRNA (P Ͻ regulation of GSK3 in the control of renal tubular calcium and 0.05). (B) Arithmetic means Ϯ SEM (n ϭ 42 to 60) of the normal- phosphate transport. To this end, renal mineral excretion was ized chemiluminescence intensity of NaPi-IIa expression in Xeno- analyzed in gene-targeted mice in which the serine residues within pus oocytes injected with water (left bar), NaPi-IIa cRNA (middle the respective PKB/SGK phosphorylation sites of GSK␣ and bar), or NaPi-IIa and GSK3␤ cRNA (right bar). ***Statistically sig- GSK3␤ had been replaced by alanin residues (GSK3␣21A/21A, nificant difference from absence of NaPi-IIa cRNA (P Ͻ 0.001). GSK3ß9A/9A). In those mice (gsk3KI), GSK␣ and GSK3␤ are resis- ###Difference from absence of GSK3␤ cRNA (P Ͻ 0.001). tant against inactivation by PKB/SGK.24 As shown before, gsk3KI (gsk3WT). Total sodium-dependent uptake of phosphate into mice are resistant to the effect of insulin on muscle glycogen syn- isolated BBMVs was significantly reduced in gsk3KI mice thase.24 (Figure 2). After inhibition of SLC34 type transporters (i.e., NaPi-IIa and NaPi-IIc) with phosphonoformic acid (PFA, 6 mM), phosphate transport was significantly reduced in both RESULTS genotypes (Figure 2). In the presence of PFA, the residual sodium-dependent phosphate transport activity was still lower in A first series of experiments analyzed the in vitro influence of gsk3KI mice than in gsk3WT mice (Figure 2). The protein ex- GSK3␤ on NaPi-IIa, a major renal tubular phosphate trans- pression of NaPi-IIa, NaPi-IIc, and Pit-2 in the brush border porter. Exposure of water-injected Xenopus oocytes to phos- membrane fraction of WT and KI murine kidneys is shown in phate (2 mM) in the bath solution did not induce any signif- Figure 3. The abundance of NaPi-IIa and of NaPi-IIc protein icant current, indicating that these oocytes do not express was significantly reduced in gsk3KI mice compared with gsk3WT significant endogenous electrogenic phosphate transport mice (Figure 3). Semiquantitative real-time PCR analysis (Figure 1A). In oocytes injected with cRNA encoding NaPi-IIa, showed significantly increased mRNA abundance of NaPi-IIa however, the addition of phosphate induced an inward current KI Ϯ ϭ and Pit-2 in gsk3 mice, suggesting that GSK3 acts partially (intraperitoneallyi)of58 10 nA (n 21 oocytes). Coexpres- ␤ upstream of the transcription of these transporters (all data sion of GSK3 significantly decreased intraperitoneallyi in NaPi-IIa–expressing oocytes (27 Ϯ 5 nA; n ϭ 17 oocytes). A chemiluminescence-based assay was used to study whether coexpression of GSK3␤ altered the membrane abundance of NaPi-IIa. As shown in Figure 1B, the surface expression of NaPi-IIa was indeed significantly reduced by coexpression of GSK3␤. The decrease of NaPi-IIa activity is not simply caused by the expression of an additional protein. The phosphate-induced current in NaPi-IIa–expressing oocytes is, for instance, similar with and without coexpression of mTOR, if Figure 2. Sodium-dependent phosphate transport activity in 25 the kinase is inhibited by rapamycin. BBMVs from gsk3KI and gsk3WT mice. Arithmetic means Ϯ SEM As a next step, activity and expression of renal sodium- (n ϭ 5 to 6 each group) of the sodium-dependent transport rates dependent phosphate transporters was assessed in brush bor- into isolated BBMVs after 1 minute in the absence (left bars) and der membrane vesicles (BBMVs) from GSK3␣␤ knockin (KI) presence (right bars) of 6 mM PFA to block phosphate transport mice (gsk3KI) and corresponding wild-type (WT) mice mediated by SLC34 family members. *P Ͻ 0.05.

874 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 873–880, 2011 www.jasn.org BASIC RESEARCH

Figure 3. Protein abundance of renal sodium-dependent phosphate cotransporters in the brush border membrane in kidneys from gsk3KI and gsk3WT mice. Western blots for NaPi-IIa, NaPi-IIc, and Pit-2. All membranes were stripped and reprobed for ␤-actin to control for loading. Densitometry was performed and the ratio of the protein of interest over ␤-actin calculated. Bar graphs present data as arithmetic means Ϯ SEM (n ϭ 5 to 6 each group).

expressed as mRNA expression normalized to HPRT mRNA abundance; see Concise Methods; NaPi-IIa: gsk3WT: 29.0 Ϯ Figure 4. Plasma phosphate concentration, urinary phosphate 3.2, n ϭ 6; gsk3KI: 59.97 Ϯ 11.12, n ϭ 5; P Ͻ 0.05; NaPi-IIc: excretion, and fractional phosphate excretion in gsk3KI and WT gsk3WT: 0.22 Ϯ 0.01, n ϭ 6; gsk3KI: 0.22 Ϯ 0.03, n ϭ 5; not gsk3 mice. Arithmetic means Ϯ SEM (n ϭ 6 to 10) of plasma WT Ϯ ϭ KI Ϯ phosphate concentration (top left panel) and urinary phosphate significant; Pit-2: gsk3 : 0.16 0.01, n 6; gsk3 : 0.34 KI 0.04, n ϭ 5; P Ͻ 0.01). excretion (top right panel) in GSK3 knockin mice (gsk3 , closed bars) and corresponding wild-type mice (gsk3WT, open bars). The Next, in vivo experiments were performed in gsk3KI and bottom panel depicts the arithmetic means Ϯ SEM (n ϭ 5to6)of WT 26 gsk3 mice. As reported earlier, the body weight was similar the fractional phosphate excretion of GSK3 knockin mice and wild KI Ϯ ϭ WT Ϯ in gsk3 mice (29.1 0.3 g, n 10) and gsk3 mice (28.9 type mice maintained under a normal diet (left bars) or main- 1.4 g, n ϭ 10). The urinary flow rate was significantly higher in tained for 3 days under a phosphorus-deficient diet (right bars). KI gsk3 mice (95.2 Ϯ 10.1 ␮l/24 h per gram body weight, n ϭ 6) *P Ͻ 0.05 versus respective value of gsk3WT mice. than in gsk3WT mice (24.1 Ϯ 4.8 ␮l/24 h per gram body weight, n ϭ 6). Similarly, the creatinine clearance was significantly The phosphaturia could further have been caused by inhi- higher in gsk3KI mice (0.39 Ϯ 0.06 ml/24 h, n ϭ 6) than in bition of renal tubular phosphate reabsorption by PTH. How- gsk3WT mice (0.19 Ϯ 0.02 ml/24 h, n ϭ 6). ever, as shown in Figure 5A (left panel), the plasma concentra- As shown in Figure 4 (top right panel), the urinary phos- tion of PTH was significantly lower in gsk3KI than in gsk3WT phate excretion was significantly larger in gsk3KI than in gsk3WT mice. A decreased plasma PTH concentration is expected to

mice. In a further series of experiments, the mice were fed a reduce the formation of 1,25-(OH)2 vitamin D3. As shown in control diet followed by a phosphorus-deficient diet for 3 days. Figure 5A (right panel), the plasma concentration of 1,25- KI Before the phosphorus-deficient diet, the fractional phosphate (OH)2 D3 was indeed significantly lower in gsk3 than in excretion was 31.1 Ϯ 6.9% (n ϭ 5) in gsk3WT and 63.9 Ϯ 13.6% gsk3WT mice. The transcript levels of CYP27B1 (25-hydroxyvi- ϭ KI ␣ (n 6) in gsk3 mice (Figure 4, bottom panel). On the third tamin D3 1 -hydroxylase), which catalyzes the last step in the day of the phosphorus-deficient diet, the fractional phosphate synthesis of the active 1,25-(OH)2D3, tended to be higher in excretion amounted to 1.4 Ϯ 0.1% (n ϭ 5) in gsk3WT mice and gsk3KI than in gsk3WT mice; this difference, however, did not to 31.8 Ϯ 11.1% (n ϭ 6) in gsk3KI mice, values significantly reach statistical significance (Figure 5B). Ͻ (P 0.05) different (Figure 4, bottom panel). A decrease of the plasma PTH and 1,25-(OH)2 vitamin D3 In theory, the enhanced urinary phosphate excretion could concentration is further expected to blunt the renal tubular have been secondary to enhanced plasma phosphate concen- Ca2ϩ reabsorption. As shown in Figure 6 (right panel), the tration. However, the plasma phosphate concentration was renal Ca2ϩ excretion was indeed larger in gsk3KI than in gsk3WT significantly lower in gsk3KI than in gsk3WT mice (Figure 4, top mice. The plasma Ca2ϩ concentration tended to be lower in left panel). gsk3KI than in gsk3WT mice; this difference, however, did not

J Am Soc Nephrol 22: 873–880, 2011 GSK-Sensitive Mineral Metabolism 875 BASIC RESEARCH www.jasn.org

KI Figure 5. Plasma PTH and 1,25-(OH)2D3 concentration in gsk3 and gsk3WT mice. (A) Arithmetic means Ϯ SEM of plasma PTH (left ϭ ϭ panel, n 6) and 1,25-(OH)2D3 (right panel, n 17 to 19) concentration in GSK3 knockin mice (gsk3KI, closed bars) and corresponding wild-type mice (gsk3WT, open bars). *P Ͻ 0.05 versus respective value of gsk3WT mice. (B) mRNA expression of

CYP27B1 (25-hydroxyvitamin D3 1a-hydroxylase). Arithmetic means Ϯ SEM (n ϭ 5to6). reach statistical significance (Figure 6, left panel). Further experiments addressed the fractional calcium excretion before Figure 7. Bone density of gsk3KI and gsk3WT mice. (A) Arithmetic and after a phosphorus-deficient diet for 3 days. Before the low means Ϯ SEM (n ϭ 4 each group) of bone density in GSK3 phosphate diet, the fractional calcium excretion was 1.6 Ϯ knockin mice (gsk3KI, closed bar) and corresponding wild-type 0.2% (n ϭ 4) in gsk3WT mice and 2.4 Ϯ 0.2% (n ϭ 5) in gsk3KI mice (gsk3WT, open bar). *P Ͻ 0.05 versus respective value of mice, values significantly (P Ͻ 0.05) different. On the third day gsk3WT mice. (B) Original microCT image of the leg of a GSK3 KI of the phosphorus-deficient diet, the fractional calcium excre- knockin mice (gsk3 , right image) and of a corresponding wild- WT tion amounted to 0.8 Ϯ 0.1% (n ϭ 4) in gsk3WT mice and to type mouse (gsk3 , left image). (C) Cuts from microCT images in 11.7 Ϯ 5.4% (n ϭ 5) in gsk3KI mice. tranaxial, coronal, and sagittal directions through the femur and ϩ knee. These high-resolution microCT images display the degra- Renal loss of both Ca2 and phosphate together with low dation of trabecular bone and compacta in GSK3 knockin mice (gsk3KI, right image) and in wild-type mice (gsk3WT, left image).

PTH and 1,25-(OH)2 vitamin D3 levels is expected to compro- mise the mineralization of bone. Thus, bone density was determined. As shown in Figure 7, femur bone density was indeed significantly lower in gsk3KI than in gsk3WT mice.

DISCUSSION

This study confirmed the previous observation26 that creati- Figure 6. Plasma calcium concentration and urinary calcium excretion in gsk3KI and gsk3WT mice. Arithmetic means Ϯ SEM (n ϭ nine clearance and urinary flow rate are higher in mice express- 6 to 19) of plasma calcium concentration (left panel) and urinary ing protein kinase B(PKB)/serum and glucocorticoid-induc- calcium excretion (right panel) in GSK3 knockin mice (gsk3KI, ible kinase (SGK)-resistant glycogen synthase kinase GSK3 KI closed bars) and corresponding wild-type mice (gsk3WT, open (gsk3 ). PKB/SGK resistance was accomplished by introduc- bars). *P Ͻ 0.05 versus respective value of gsk3WT mice. ing a mutation of GSK3ß in which the serine of the PKB phos-

876 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 873–880, 2011 www.jasn.org BASIC RESEARCH

9A/9A 2ϩ phorylation site was replaced by an alanine (GSK3ß ) and PTH and 1,25-(OH)2D3 stimulate renal tubular Ca reab- at the same time a mutation in GSK3␣ in which the serine of sorption via upregulation of the TRPV5 channel.12,13,33 The

the PKB phosphorylation site was replaced by an alanine decreased plasma PTH and 1,25-(OH)2D3 concentration in (GSK3ß12A/12A). Because inhibition of GSK3 mediates the ef- gsk3KI mice could thus contribute to the calciuria of gsk3KI fect of insulin on glycogen synthase,24,27 the effect of insulin on mice by affecting translation, trafficking, and/or activity of muscle glycogen synthase is abrogated in gsk3KI mice.24 TRPV5 channels. More importantly, these observations showed a completely Despite the decreased PTH plasma concentration, the bone novel function of GSK3, i.e., a powerful role in the regulation mass was significantly decreased in gsk3KI mice. Thus, the renal of renal mineral excretion. Compared with wild-type mice Ca2ϩ and phosphate loss of gsk3KI mice may impair mineral- (gsk3WT), gsk3KI mice are both hypercalciuric and hyperphos- ization of bone. Phosphate is known to inhibit the generation phaturic. Renal phosphate excretion is mostly determined by of new osteoclasts and to stimulate apoptosis of mature oste- ϩ the expression and activity of Na -dependent phosphate oclasts.34 Moreover, Akt/PKB-dependent regulation of GSK3 transporters in the brush border membrane of the proximal may more directly participate in the regulation of bone cells. tubule. Their expression and activity is regulated by various Akt/PKB regulates the survival of both osteoblasts35 and oste- hormones and factors, but only very little is known about in- oclasts.36,37 tracellular signaling events mediating the control of these In conclusion, Akt/PKB-dependent regulation of GSK3 transporters.1,2,8,10 Our observations showed that GSK3 plays participates in the control of renal tubular phosphate trans- an important role as evident from the stimulation of NaPi-IIa port. Loss of Akt/PKB-dependent inhibition of GSK3␣␤ leads activity in Xenopus oocytes by GSK3 and the decreased phos- to renal phosphate wasting, which presumably contributes to or phate fluxes in isolated BBMVs from gsk3KI mice. Reduced even accounts for the decrease of PTH release with resulting cal-

activity of SLC34 transporters (NaPi-IIa and NaPi-IIc) and ciuria, decreased formation of 1,25-(OH)2 D3, and decreased other PFA-resistant phosphate transporters was detected. The mineralization of bone. These observations thus show a novel, expression of Pit2, resistant to PFA, was, however, not altered, powerful element in the regulation of mineral metabolism. suggesting that either only its activity is affected by GSK3 or that other unknown PFA-resistant phosphate transporters contribute to the residual activity in the presence of PFA. CONCISE METHODS The phosphaturia of gsk3KI mice cannot be explained by an increased plasma phosphate concentration, which was actually In Vitro Expression significantly decreased. Moreover, the phosphaturia of gsk3KI For generation of cRNA, constructs were used encoding wild-type mice contrasts the decreased plasma level of PTH, a major regu- human NaPi-IIa38 and human GSK3␤. The cRNA was generated as lator of renal phosphate transport.8 PTH stimulates the internal- described previously.39 GSK3␤ cDNA was from the “Deutsches Res- ization and subsequent degradation of NaPi-IIa.8,17 The hormone sourcenzentrum fu¨r Genomforschung,” Berlin, Germany. For elec- thus enhances renal phosphate excretion and decreases plasma trophysiology, Xenopus oocytes were prepared as described previ- phosphate concentration. A decreased plasma PTH concentra- ously.40,41 Ten nanograms of NaPi-IIa cRNA was injected on the first tion is thus expected to enhance renal phosphate reabsorption day and 7.5 ng of GSK3␤ cRNA was injected on the second day after and to increase plasma phosphate concentration, the opposite of preparation of Xenopus oocytes. All experiments were performed at what was observed. A decreased plasma phosphate concentration room temperature 3 days after the second injection. Two electrode inhibits PTH release.28 However, the decrease of the plasma phos- voltage-clamp recordings were performed at a holding potential of phate level is very discrete and hardly accounts for the low plasma Ϫ50 mV. The data were filtered at 10 Hz and recorded with a Digidata

PTH levels. The plasma 1,25-(OH)2 D3 concentration were de- A/D-D/A converter and Chart V.4.2 software for data acquisition and creased in gsk3KI mice. Low plasma PTH levels are expected to analysis (Axon Instruments). The control solution (superfusate/

decrease the plasma 1,25-(OH)2 D3 concentration, because PTH ND96) contained 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2,1mM ␣ 29 stimulates the renal 1 -hydroxylase, the rate-limiting enzyme of MgCl2, and 5 mM HEPES, pH 7.4. Phosphate (2 mM) was added to 29,30 1,25-(OH)2 D3 formation. On the other hand, formation of induce NaPi-IIa–dependent currents. The flow rate of the superfu- 1,25-(OH)2 D3 is stimulated in a PTH-independent manner by sion was 20 ml/min, and a complete exchange of the bath solution was cellular phosphate depletion.31 The transcript levels of the 1␣- reached within about 10 seconds. For the determination of NaPi-IIa hydroxylase tended to be enhanced in gsk3KI mice. The possibility surface expression, defolliculated oocytes were incubated with pri- must be considered that GSK3 influences 1␣-hydroxylase and mary rabbit anti-NaPi-IIa antibody (diluted 1:500; Lifespan Biosci-

1,25-(OH)2 D3 formation by further mechanisms in addition to its ences, Seattle, WA) and secondary, peroxidase-conjugated goat anti- influence on phosphate balance and PTH plasma concentration. rabbit antibody (diluted 1:1000; Cell Signaling, Danvers, MA). ␮ Because 1,25-(OH)2 D3 stimulates intestinal phosphate ab- Individual oocytes were placed in 96-well plates with 10 l of Super 32 sorption, decreased 1,25-(OH)2 D3 formation is expected to Signal ELISA Femto Maximum Sensitivity Substrate (Pierce, Rock- decrease intestinal phosphate absorption. Impaired intestinal ford, IL). The chemiluminescence of the oocytes was quantified in a phosphate uptake could contribute to the development of hy- luminometer (WalterWallac2 plate reader; Perkin Elmer, Ju¨gesheim, pophosphatemia. Germany) by integrating the signal over a period of 1 second. Results

J Am Soc Nephrol 22: 873–880, 2011 GSK-Sensitive Mineral Metabolism 877 BASIC RESEARCH www.jasn.org display normalized arbitrary light units that are proportional to the RNA Extraction and Real-Time RT-PCR detector voltage. Snap-frozen kidneys (five kidneys for each condition) were homoge- nized in RLT-Buffer (Qiagen, Basel, Switzerland) and supplemented In Vivo Experiments with ß-mercaptoethanol, resulting in a final concentration of 1%. All animal experiments were conducted according to the guidelines of the Total RNA was extracted and reverse transcribed as described previ- American Physiologic Society and approved by the respective authorities. ously.9 Quantitative real-time qRT-PCR was performed on the ABI Mice were generated in which the codon encoding Ser9 of the GSK3ß gene PRISM 7700 Sequence Detection System (Applied Biosystems). was changed to encode nonphosphorylatable alanine (GSK3ß9A/9A), and si- Primers for all genes of interest were as described.9 For analysis of the multaneously, the codon encoding Ser21 of GSK3␣ was changed to en- data, the threshold was set to 0.06 because this value had been deter- code the nonphosphorylatable GSK3␣21A/21A, thus yielding the GSK3␣/ mined to be in the linear range of the amplification curves for all mRNAs ␤21A/21A/9A/9A double knockin mouse (gsk3KI) as described previously.24 in all experimental runs. The expression of the gene of interest was calcu- The mice were compared with corresponding wild-type mice (gsk3WT). lated in relation to hypoxanthine guanine phosphoribosyl transferase. [Ct(HPRT) Ϫ Ct(test gene)] The mice were fed a control diet (Altromin, Lage, Germany) con- Relative expression ratios were calculated as Rϭ2 , taining 7523 mg/kg phosphorus or a phosphorus-deficient diet con- where Ct represents the cycle number at the threshold 0.06. taining 131 mg/kg phosphorus, as indicated. They had free access to tap drinking water. BBMV Preparation and Phosphate Transport Assays To determine creatinine clearance, urinary flow rate, and urinary BBMVs were prepared from rat kidney cortex and outer medulla us- ϩ excretion of Ca2ϩ and phosphate, the mice were placed individually in ing the Mg2 precipitation technique as described previously.43,44 metabolic cages (Techniplast, Hohenpeissenberg, Germany) for 24- The phosphate transport rate into BBMVs was measured in freshly hour urine collection as described previously.42 They were allowed a prepared BBMVs at 25°C in the presence of inward gradients of 100

2-day habituation period, during which food and water intake, uri- mM NaCl or 100 mM KCl and 0.1 mM K-phosphate. The substrate Pi 32 ␮ nary flow rate, and excretion of electrolytes were recorded every day to was prepared with 0.125 mM K2HPO4 and P(1 Ci/ml) to yield a ascertain that the mice were adapted to the new environment. Subse- final concentration of 0.1 mM, close to the expected apparent KmPi for ϩ quently, 24-hour collection of urine was performed for 3 consecutive Na -dependent transport in renal BBMVs. The stop solution con- days to obtain the urinary parameters. To assure quantitative urine tained 100 mM mannitol, 5 mM Tris-HCl, pH 7.4, 150 mM NaCl, and ϩ collection, metabolic cages were siliconized, and urine was collected 5mMPi. The Na dependence was established by incubating BBMVs under water-saturated oil. in solutions in which KCl replaced NaCl equimolarly. Phosphate up- The phosphate concentration was determined colorimetrically take was determined after 60 seconds, representing initial linear con- using commercial diagnostic kits (Roche Diagnostics, Mannheim, ditions, and after 120 minutes, to determine the equilibrium values. 2ϩ ϩ Germany). The urinary concentration of Ca was measured by To distinguish between Na -dependent Pi uptake mediated by SLC34 ϩ flame photometry (ELEX 6361; Eppendorf) or by a photometric family members (e.g., NaPi-IIa and NaPi-IIc) and other Na -depen- method according to the manufacturer’s instructions (dri-chem dent phosphate transporters such as SLC20 family members (e.g., clinical chemistry analyzer FUJI FDC 3500i; Sysmex, Norsted, Ger- Pit-1 and Pit-2), we used trisodium PFA (final concentration, 6 mM) many). The creatinine concentration in urine was determined us- added to the same solution with 107 mM NaCl. PFA has previously ing the Jaffe reaction (Sigma, St. Louis, MO); creatinine and cal- been shown to have a higher selectivity for SLC34 than for SLC20 cium concentration in plasma were measured using a photometric phosphate transporters at this concentration.3 The total protein con- method. The plasma intact parathormone concentration was mea- centration was measured using the Bio-Rad Protein Assay kit (Bio- sured using an ELISA kit (Immunotopics, San Clemante, CA). A Rad, Hercules, CA). BBMVs were stored at Ϫ80°C until further use. radioimmunassay kit was used to determine the concentration of

1,25-(OH)2D3 (IDS, Boldon, UK) in plasma. Western Blotting For the analysis of bone density, animals were killed, and legs were After measurement of the protein concentration (Bio-Rad), 10 ␮gof amputated and fixated in formalin. The samples were scanned with a renal brush border membrane proteins was solubilized in loading high-resolution microCAT-II (Siemens Preclinical Solutions) small buffer containing dithiothreitol and separated on 8% polyacrylamide animal computed tomography (CT) scanner using a field of view of gels. For immunoblotting, the proteins were transferred electropho- 3.1 ϫ 3.1 ϫ 4.8 cm3. The x-ray tube parameters were set at 80 kVp and retically to polyvinylidene fluoride membranes (Immobilon-P; Milli- 400 ␮A. The images were acquired with 720 angular projections (ex- pore, Bedford, MA). After blocking with 5% milk powder in Tris- posure time, 1200 ms per projection) over 360° and binned with a buffered saline/0.1% Tween-20 for 60 minutes, the blots were incubated factor of two, yielding a spatial resolution of approximately 38 ␮m. with the primary antibodies: rabbit polyclonal anti-NaPi-IIa (1:6000),45 The total scan time was 24 minutes. Reconstructed CT images were rabbit polyclonal anti-NaPi-IIc (1:10,000),9 rabbit polyclonal anti- analyzed with the Inveon Research Workplace software (Siemens Pre- Pit-2 (1:3000; kindly provided by Dr. V. Sorribas, University of Zara- clinical Solutions) by drawing a standard-sized container around the goza, Spain),3 and mouse monoclonal anti-␤-actin antibody (42 kD; femur and applying a region growth routine to segment the trabecular Sigma; 1:5000) either for 2 hours at room temperature or overnight at bone structure. For all samples, the same upper and lower density 4°C. Membranes were incubated for 1 hour at room temperature with threshold was applied, and the relative numbers of trabecular bone secondary goat anti-rabbit or donkey anti-mouse antibodies 1:5000 density were compared. linked to alkaline phosphatase (Promega) or to horseradish peroxi-

878 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 873–880, 2011 www.jasn.org BASIC RESEARCH dase (Amersham). The protein signal was detected with the appropri- 11. Boros S, Bindels RJ, Hoenderop JG: Active Ca(2ϩ) reabsorption in the ate substrates (Millipore) using the DIANA III-chemiluminescence connecting tubule. Pflugers Arch 458: 99–109, 2009 detection system (Raytest, Straubenhardt, Germany). All images were 12. Hoenderop JG, Nilius B, Bindels RJ: Calcium absorption across epi- thelia. Physiol Rev 85: 373–422, 2005 analyzed using the software Advanced Image Data Analyzer AIDA, 13. Woudenberg-Vrenken TE, Bindels RJ, Hoenderop JG: The role of Raytest to calculate the protein of interest/␤-actin ratio. transient receptor potential channels in kidney disease. Nat Rev Neph- rol 5: 441–449, 2009 14. DeFronzo RA, Cooke CR, Andres R, Faloona GR, Davis PJ: The effect Statistical Analysis Ϯ of insulin on renal handling of sodium, potassium, calcium, and phos- Data are provided as means SEM; n represents the number of in- phate in man. J Clin Invest 55: 845–855, 1975 dependent experiments. All data were tested for significance using 15. Hoskins B, Scott JM: Evidence for a direct action of insulin to increase paired or unpaired t tests or ANOVA. GraphPad InStat version 3.00 renal reabsorption of calcium and for an irreversible defect in renal for Windows 95 (GraphPad Software, San Diego, CA) was used. Only ability to conserve calcium due to prolonged absence of insulin. results with P Ͻ 0.05 were considered statistically significant. Diabetes 33: 991–994, 1984 16. Bacic D, Schulz N, Biber J, Kaissling B, Murer H, Wagner CA: Involve- ment of the MAPK-kinase pathway in the PTH-mediated regulation of the proximal tubule type IIa Naϩ/Pi cotransporter in mouse kidney. ACKNOWLEDGMENTS Pflugers Arch 446: 52–60, 2003 17. Bacic D, Lehir M, Biber J, Kaissling B, Murer H, Wagner CA: The renal Naϩ/phosphate cotransporter NaPi-IIa is internalized via the receptor- The authors acknowledge the technical assistance of E. Faber, Daniel mediated endocytic route in response to parathyroid hormone. Kid- Bukala, and Michael Stu¨bs and the meticulous preparation of the ney Int 69: 495–503, 2006 manuscript by T. Loch and L. Subasic. This study was supported by 18. Hawkins PT, Anderson KE, Davidson K, Stephens LR: Signalling the Deutsche Forschungsgemeinschaft (GK 1302) and by a grant from through Class I PI3Ks in mammalian cells. Biochem Soc Trans 34: 647–662, 2006 the Swiss National Science Foundation to C.A.W. (3100A0-122217). 19. Lang F, Bohmer C, Palmada M, Seebohm G, Strutz-Seebohm N, Vallon V: (Patho)physiological significance of the serum- and glucocorticoid-inducible kinase isoforms. Physiol Rev 86: 1151–1178, 2006 DISCLOSURES 20. Cross HR, Radda GK, Clarke K: The role of Naϩ/Kϩ ATPase activity None. during low flow ischemia in preventing myocardial injury: A 31P, 23Na and 87Rb NMR spectroscopic study. Magn Reson Med 34: 673–685, 1995 21. Shaw M, Cohen P, Alessi DR: Further evidence that the inhibition of REFERENCES glycogen synthase kinase-3beta by IGF-1 is mediated by PDK1/PKB- induced phosphorylation of Ser-9 and not by dephosphorylation of 1. Biber J, Hernando N, Forster I, Murer H: Regulation of phosphate Tyr-216. FEBS Lett 416: 307–311, 1997 transport in proximal tubules. Pflugers Arch 458: 39–52, 2009 22. Sakoda H, Gotoh Y, Katagiri H, Kurokawa M, Ono H, Onishi Y, Anai 2. Murer H, Forster I, Biber J: The sodium phosphate cotransporter M, Ogihara T, Fujishiro M, Fukushima Y, Abe M, Shojima N, Kikuchi family SLC34. Pflugers Arch 447: 763–767, 2004 M, Oka Y, Hirai H, Asano T: Differing roles of Akt and serum- and 3. Villa-Bellosta R, Ravera S, Sorribas V, Stange G, Levi M, Murer H, Biber glucocorticoid-regulated kinase in glucose metabolism, DNA syn- J, Forster IC: The Naϩ-Pi cotransporter PiT-2 (SLC20A2) is expressed thesis, and oncogenic activity. J Biol Chem 278: 25802–25807, in the apical membrane of rat renal proximal tubules and regulated by 2003 dietary Pi. Am J Physiol Renal Physiol 296: F691–F699, 2009 23. Wyatt AW, Hussain A, Amann K, Klingel K, Kandolf R, Artunc F, 4. Allon M: Effects of insulin and glucose on renal phosphate reabsorp- Grahammer F, Huang DY, Vallon V, Kuhl D, Lang F: DOCA-induced tion: Interactions with dietary phosphate. J Am Soc Nephrol 2: 1593– phosphorylation of glycogen synthase kinase 3beta. Cell Physiol 1600, 1992 Biochem 17: 137–144, 2006 5. DeFronzo RA, Goldberg M, Agus ZS: The effects of glucose and 24. McManus EJ, Sakamoto K, Armit LJ, Ronaldson L, Shpiro N, Mar- insulin on renal electrolyte transport. J Clin Invest 58: 83–90, 1976 quez R, Alessi DR: Role that phosphorylation of GSK3 plays in 6. Feld S, Hirschberg R: Insulinlike growth factor I and the kidney. Trends insulin and Wnt signalling defined by knockin analysis. EMBO J 24: Endocrinol Metab 7: 85–93, 1996 1571–1583, 2005 7. Jehle AW, Forgo J, Biber J, Lederer E, Krapf R, Murer H: IGF-I and 25. Kempe DS, Dermaku-Sopjani M, Frohlich H, Sopjani M, Umbach A, vanadate stimulate Na/Pi-cotransport in OK cells by increasing type II Puchchakayala G, Capasso A, Weiss F, Stubs M, Foller M, Lang F: Na/Pi-cotransporter protein stability. Pflugers Arch 437: 149–154, Rapamycin-induced phosphaturia. Nephrol Dial Transplant 25: 2938– 1998 2944, 2010 8. Murer H, Hernando N, Forster I, Biber J: Proximal tubular phosphate 26. Boini KM, Bhandaru M, Mack A, Lang F: Steroid hormone release as reabsorption: Molecular mechanisms. Physiol Rev 80: 1373–1409, well as renal water and electrolyte excretion of mice expressing 2000 PKB/SGK-resistant GSK3. Pflugers Arch 456: 1207–1216, 2008 9. Nowik M, Picard N, Stange G, Capuano P, Tenenhouse HS, Biber J, 27. Cohen P, Goedert M: GSK3 inhibitors: Development and therapeutic Murer H, Wagner CA: Renal phosphaturia during metabolic acidosis potential. Nat Rev Drug Discov 3: 479–487, 2004 revisited: molecular mechanisms for decreased renal phosphate reab- 28. Martin DR, Ritter CS, Slatopolsky E, Brown AJ: Acute regulation of sorption. Pflugers Arch 457: 539–549, 2008 parathyroid hormone by dietary phosphate. Am J Physiol Endocrinol 10. Picard N, Capuano P, Stange G, Mihailova M, Kaissling B, Murer H, Metab 289: E729–E734, 2005 Biber J, Wagner CA: Acute parathyroid hormone differentially regu- 29. Portale AA, Miller WL: Human 25-hydroxyvitamin D-1alpha-hydroxy- lates renal brush border membrane phosphate cotransporters. lase: Cloning, mutations, and gene expression. Pediatr Nephrol 14: Pflugers Arch 460: 677–687, 2010 620–625, 2000

J Am Soc Nephrol 22: 873–880, 2011 GSK-Sensitive Mineral Metabolism 879 BASIC RESEARCH www.jasn.org

30. Kato S: Genetic mutation in the human 25-hydroxyvitamin D3 1alpha- 38. Busch AE, Wagner CA, Schuster A, Waldegger S, Biber J, Murer H, hydroxylase gene causes vitamin D-dependent rickets type I. Mol Cell Lang F: Properties of electrogenic Pi transport by a human renal Endocrinol 156: 7–12, 1999 brush border Naϩ/Pi transporter. J Am Soc Nephrol 6: 1547–1551, 31. Perwad F, Azam N, Zhang MY, Yamashita T, Tenenhouse HS, Portale 1995 AA: Dietary and serum phosphorus regulate fibroblast growth factor 39. Strutz-Seebohm N, Shojaiefard M, Christie D, Tavare J, Seebohm 23 expression and 1,25-dihydroxyvitamin D metabolism in mice. En- G, Lang F: PIKfyve in the SGK1 mediated regulation of the creatine docrinology 146: 5358–5364, 2005 transporter SLC6A8. Cell Physiol Biochem 20: 729–734, 2007 32. Brown AJ, Finch J, Slatopolsky E: Differential effects of 19-nor-1,25- 40. Boehmer C, Palmada M, Klaus F, Jeyaraj S, Lindner R, Laufer J, Daniel dihydroxyvitamin D (2) and 1,25-dihydroxyvitamin D (3) on intestinal H, Lang F: The peptide transporter PEPT2 is targeted by the protein calcium and phosphate transport. J Lab Clin Med 139: 279–284, 2002 kinase SGK1 and the scaffold protein NHERF2. Cell Physiol Biochem 33. Friedman PA: Mechanisms of renal calcium transport. Exp Nephrol 8: 22: 705–714, 2008 343–350, 2000 41. Boehmer C, Laufer J, Jeyaraj S, Klaus F, Lindner R, Lang F, Palmada M: 34. Kanatani M, Sugimoto T, Kano J, Kanzawa M, Chihara K: Effect of high Modulation of the voltage-gated potassium channel Kv1.5 by the phosphate concentration on osteoclast differentiation as well as bone- SGK1 protein kinase involves inhibition of channel ubiquitination. Cell resorbing activity. J Cell Physiol 196: 180–189, 2003 Physiol Biochem 22: 591–600, 2008 35. Chaudhary LR, Hruska KA: The cell survival signal Akt is differentially 42. Vallon V: In vivo studies of the genetically modified mouse kidney. activated by PDGF-BB, EGF, and FGF-2 in osteoblastic cells. J Cell Nephron Physiol 94: 1–5, 2003 Biochem 81: 304–311, 2001 43. Biber J, Stieger B, Haase W, Murer H: A high yield preparation for rat 36. Kwak HB, Sun HM, Ha H, Lee JH, Kim HN, Lee ZH: AG490, a Jak2- kidney brush border membranes. Different behaviour of lysosomal specific inhibitor, induces osteoclast survival by activating the Akt and markers. Biochim Biophys Acta 647: 169–176, 1981 ERK signaling pathway. Mol Cells 26: 436–442, 2008 44. Biber J, Stieger B, Stange G, Murer H: Isolation of renal proximal 37. Lee SE, Chung WJ, Kwak HB, Chung CH, Kwack KB, Lee ZH, Kim HH: tubular brush-border membranes. Nat Protoc 2: 1356–1359, 2007 Tumor necrosis factor-alpha supports the survival of osteoclasts 45. Custer M, Lotscher M, Biber J, Murer H, Kaissling B: Expression of through the activation of Akt and ERK. J Biol Chem 276: 49343– Na-P(i) cotransport in rat kidney: Localization by RT-PCR and immu- 49349, 2001 nohistochemistry. Am J Physiol 266: F767–F774, 1994

880 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 873–880, 2011