Volume 88, Issue 6, Pages 1304-1313 (December 2015) Key role of the kidney in the regulation of fibroblast growth factor 23  Maria L. Mace, Eva Gravesen, Jacob Hofman-Bang, Klaus Olgaard, Ewa Lewin  Kidney International  Volume 88, Issue 6, Pages 1304-1313 (December 2015) DOI: 10.1038/ki.2015.231 Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 1 Impact of acute bilateral (BNX) and unilateral (UNX) nephrectomy on plasma intact fibroblast growth factor 23 (iFGF23) and the ratio of intact-to-C-terminal FGF23. (a): Plasma levels of iFGF23 after acute BNX or UNX nephrectomy and in the control group. A rapid significant increase in p-iFGF23 took place within 15min after BNX (***P<0.01). The following FGF23 levels remained stable. UNX generated a p-iFGF23 rise reaching a level in between the BNX and control group. BNX (n=11), UNX (n=9), and Control (n=6). (b) To examine whether the short-lasting acute kidney injury during the BNX procedure was responsible for the stimulated FGF23, a comparison between sham (see text) and control group was performed; however, similar p-iFGF23 levels were found (n=6). (c) The ratio of intact-to-C-terminal FGF23 was measured in the plasma of control, UNX, and BNX rats at 55min after nephrectomy. A significant increase took place in UNX and further in BNX rats (**P<0.05) (n=6). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 2 Impact of absence of parathyroid hormone (PTH) on the early intact fibroblast growth factor 23 (iFGF23) regulation after bilateral nephrectomy (BNX). Parathyroidectomy (PTX) was performed 60min prior to BNX; as a result, the circulation was depleted of PTH when BNX was performed. Still, an identical rapid p-iFGF23 increase was found in the PTX-BNX group after BNX (***P<0.01) and actually same iFGF23 levels in the BNX and PTX-BNX groups throughout the observation period. Noticeably, PTX alone did not alter p-iFGF23 levels in this early regulation; these were similar to the stable iFGF23 levels in the sham group. PTX-BNX (n=8), BNX (n=11), PTX (n=6), sham (n=6). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 3 Effect of increasing doses of fibroblast growth factor receptor (FGFR) inhibitor, PD173074, on plasma iFGF23 and FGF23 gene activity in the bone. (a) The endogenous secretion of FGF23 was inhibited by blockage of FGFR using PD173074. Maximal inhibition of FGF23 secretion was obtained, as all doses of PD173074 significantly suppressed p-FGF23 to similar levels 5h after administration (P<0.05); 10mg (n=3), 20mg (n=2), 40mg (n=6), vehicle (n=5). (b) The FGFR inhibitor, PD173074, 40mg/rat significantly reduced FGF23 gene activity in the bone (*P<0.05). FGF23 mRNA is normalized to β-actin (n=5 in each group). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 4 Effect of fibroblast growth factor receptor (FGFR) inhibition and bilateral nephrectomy (BNX) on plasma levels of iFGF23. (a) 40mg PD173074 significantly suppressed p-FGF23 at 5h (P<0.05), and then BNX was performed. No increase (NS) in p-FGF23 was demonstrated when BNX was performed after prior inhibition of FGFR compared with vehicle. BNX (n=6), vehicle (n=3). (b) To study whether the acute increase of p-iFGF23 after BNX was due to an early increase in FGF23 gene expression, the level of FGF23 mRNA in bone tissue was analyzed in controls (n=6) and 15min after BNX (n=5). Similar FGF23 mRNA levels were found. Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 5 No impact of fibroblast growth factor receptor (FGFR) inhibition on metabolism of exogenous recombinant human FGF23 (rhFGF23). rhFGF23 was administered to rats with or without prior inhibition of endogenous FGF23 by 40mg PD173074. Same disappearance curve was found in both groups, thus there was no confounding impact of PD173074 on rhFGF23 clearance. PD173074 (n=5), control (n=4). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 6 The importance of the kidney in fibroblast growth factor 23 (FGF23) metabolism. (a, b) Disappearance of exogenous recombinant human FGF23 (rhFGF23) in sham and bilateral nephrectomy (BNX) rats; rhFGF23 was administered after prior inhibition of endogenous FGF23 by PD173074. A remarkable short half-life of 4.4min was found in normal rats and a noticeable slower metabolism in BNX rats, in which half-life was prolonged to 11.8min (P<0.01). BNX (n=6), sham (n=5). (c) To investigate, whether the impact of the kidney on FGF23 metabolism could be explained by renal clearance, renal artery and vein sampling was performed demonstrating a significant renal extraction (**P<0.05). (n=6). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 7 Effect of acute severe hyperphosphatemia on plasma intact fibroblast growth factor 23 (iFGF23) levels in normal and bilateral nephrectomized (BNX) rats. Plasma concentrations of phosphate (P) after intravenous bolus of (a) 80μmol P, (c) 180μmol P, and (e) 240μmol P, respectively. An initial decline in p-P was found, yet around 30min P levels stabilized at a higher plateau in all groups. (b, d, e) The corresponding iFGF23 levels in the P-infused sham and BNX rats are shown. BNX resulted repeatedly in increased iFGF23 levels (P<0.01). No impact of the high plasma level of P on iFGF23 levels was seen in BNX and normal rats. 80 P sham (n=6), 80 P BNX (n=6); 180 P sham (n=6), 180 P BNX (n=5); 240 P sham (n=5), 240 P BNX (n=4). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 8 Effect of acute severe hypercalcemia on plasma intact fibroblast growth factor 23 (iFGF23) levels in normal and bilateral nephrectomized (BNX) rats. (a) Plasma Ca2+ after induction of acute severe hypercalcemia by intravenous administration of 140μmol Ca in normal and BNX rats. Even though very high plasma Ca2+ levels were reached, these recovered rapidly to baseline in both groups. (b) The corresponding plasma iFGF23 levels after induction of severe hypercalcemia remained stable, thus no immediate effect of high Ca2+ on iFGF23 secretion in normal and BNX rats was found. Once more, acute BNX resulted in high iFGF23 levels (n=8 in both groups). Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions

Figure 9 Acute induction of hyperphosphatemia was associated with a dose-related induction of hypocalcemia, while acute induction of hypercalcemia left plasma phosphate levels stable. (a) Plasma Ca2+ levels in P-infused rats. Acute severe hyperphosphatemia resulted not only in a significant drop in p-Ca2+ levels in both bilateral nephrectomized (BNX) and sham rats (P<0.05) but also in a dose-related effect on the severity of the hypocalcemia. In addition, the recovery from hypocalcemia was prolonged in BNX rats. (b) Plasma P levels in Ca-infused rats. In contrast, P levels were not affected by severe hypercalcemia, n as in Figures 7 and 8. Kidney International 2015 88, 1304-1313DOI: (10.1038/ki.2015.231) Copyright © 2015 International Society of Nephrology Terms and Conditions