Extracellular sodium regulates fibroblast growth factor 23 (FGF23) formation

Abstract

The bone-derived hormone fibroblast growth factor-23 (FGF23) has recently received much attention due to its association with chronic kidney disease and cardiovascular disease progression. Extracellular sodium concentration ([Na⁺]) plays a significant role in bone metabolism. Hyponatremia (lower serum [Na⁺]) has recently been shown to be independently associated with FGF23 levels in patients with chronic systolic heart failure. However, nothing is known about the direct impact of [Na⁺] on FGF23 production. Here, we show that an elevated [Na⁺] (+20 mM) suppressed FGF23 formation, whereas low [Na⁺] (-20 mM) increased FGF23 synthesis in the osteoblast-like cell lines UMR-106 and MC3T3-E1. Similar bidirectional changes in FGF23 abundance were observed when osmolality was altered by mannitol but not by urea, suggesting a role of tonicity in FGF23 formation. Moreover, these changes in FGF23 were inversely proportional to the expression of NFAT5 (nuclear factor of activated T cells-5), a transcription factor responsible for tonicity-mediated cellular adaptations. Furthermore, arginine vasopressin, which is often responsible for hyponatremia, did not affect FGF23 production. Next, we performed a comprehensive and unbiased RNA-seq analysis of UMR-106 cells exposed to low versus high [Na⁺], which revealed several novel genes involved in cellular adaptation to altered tonicity. Additional analysis of cells with Crisp-Cas9-mediated NFAT5 deletion indicated that NFAT5 controls numerous genes associated with FGF23 synthesis, thereby confirming its role in [Na⁺]-mediated FGF23 regulation. In line with these in vitro observations, we found that hyponatremia patients have higher FGF23 levels. Our results suggest that [Na⁺] is a critical regulator of FGF23 synthesis.

Keywords: FGF23; NFAT5; bone and kidney; extracellular sodium; hyponatremia.

Figure 1

High NaCl suppresses FGF23 formation in UMR-106 cells.A, representative immunoblots of NFAT5 and β-actin upon addition of 20 mM NaCl, 40 mM mannitol, and 40 mM urea treatment for 16 h. B, quantification of NFAT5 protein normalized with β-actin (n = 3). C, fold change Nfat5 mRNA levels; D, cFGF23 levels in cell media; E, fold change Fgf23 mRNA levels, after treating 20 mM NaCl, 40 mM mannitol, and 40 mM urea for 24 h (n = 3–4). F, lactate dehydrogenase (LDH) release in the cells following 0, 40, 80, 120, and 140 mM +NaCl treatment for 24 h (n = 3). G, fold change Fgf23 and Nfat5 mRNA levels following 0, 2, 5, 10, 20, and 30 mM +NaCl treatment for 24 h (n = 4). One-way ANOVA summary for Nfat5: F = 13.39; p < 0.0001; R sq. = 0.79, and for Fgf23: F = 56.19; p < 0.0001; R sq. = 0.94. H, fold change Fgf23 and Nfat5 mRNA levels after 20 mM +NaCl treatment over 0 to 48 h (n = 3). p value (using an unpaired t test) indicates statistically significant difference between Nfat5 and Fgf23 mRNA, when compared to 0 h (normalized to one, for each time-point). All the values are expressed in arithmetic means ± SEM. cFGF23, C-terminal fragments FGF23; FGF23, fibroblast growth factor-23; iFGF23, intact FGF23.

Figure 2

Hypotonicity elevates FGF23 formation in UMR-106 cells.A, fold change Fgf23 mRNA levels upon treating NaCl-deficient culture media by −5, −10, −15, and −20 mM NaCl for 24 h (n = 3). B, fold change Fgf23 mRNA; C, cFGF23 levels in media; D, Nfat5 mRNA, after treating cells with −20 mM NaCl-deficient culture media. The osmolality was corrected by adding mannitol (40 mM) or urea (40 mM) for 24 h (n = 3). All the values are expressed in arithmetic means ± SEM. ns (not significant) p > 0.05. cFGF23, C-terminal fragments FGF23; FGF23, fibroblast growth factor-23.

Figure 3

No evidence for FGF23 regulation by AVP.A, agarose gel electrophoresis for detection of rat V_1a and V₂ receptor mRNA in rat liver, kidney, and UMR-106 cells. The cDNA (2 μg) obtained by RT was amplified for 34 cycles by PCR. B, V_1a, V_2, and V_1b mRNA expressions quantified by qRT-PCR in UMR-106 cells (n = 3). ND: not detected. C, fold change Fgf23 mRNA abundance relative to Tbp in UMR-106 cells treated without (Ctrl) or with the indicated concentrations of AVP for 24 h. (n = 3). D, fold change Fgf23 levels in UMR-106 cells treated without (Ctrl) or with 100 nM AVP for indicated durations (n = 3). E, fold change Fgf23 levels in UMR-106 cells treated without (Ctrl) or with the indicated concentrations of V_1a agonist, ([Phe²]OVT, [Phe²,Orn⁸]vasotocin) for 24 h. (n = 3). F, fold change Fgf23 levels in UMR-106 cells treated without (Ctrl) or with the indicated concentrations of V₂ agonist, ddAVP for 24 h (n = 3). G, fold change Fgf23 mRNA in UMR-106 cells treated with −20 mM NaCl with or without AVP and V_1a agonist (each 10 nM) for 24 h (n = 3). H, serum cFGF23 and iFGF23 levels in wildtype and V_1a KO mice (n = 4–5, each group). AVP, arginine vasopressin; cFGF23, C-terminal fragments FGF23; FGF23, fibroblast growth factor-23; iFGF23, intact FGF23.

Figure 4

The [Na⁺]-mediated regulation of FGF23 in UMR-106 cells requires NFAT5.A, original immunoblot of NFAT5 and β-actin after -NaCl (−20 mM) and +NaCl (+20 mM) treatment for 24 h in control and NFAT5^KO cells. B, Nfat5 mRNA levels in control and NFAT5^KO cells after -NaCl and +NaCl treatment for 24 h (n = 3). Nfat5 mRNA levels in NFAT5^KO cells were not detected. C, LDH release in cell supernatants treated with different NaCl concentrations (n = 3). p value (using an unpaired t test) indicates statistically significant difference between LDH levels in control and NFAT5^KO cell supernatant at that particular +NaCl concentration. The two-way ANOVA reveals a p < 0.0001 for the treatment effect and p = 0.0028 for the interaction. D, Fgf23 mRNA levels measured by qRT-PCR in control and NFAT5^KO UMR-106 cells after -NaCl versus +NaCl treatment for 24 h (n = 3, each group). E, venn diagrams showing a number of upregulated/downregulated genes in -NaCl versus +NaCl treatment for 24 h. F, heatmap of gene expression levels in control and NFAT5^KO cells upon -NaCl versus +NaCl treatment. G, volcano plot of upregulated/downregulated genes in control cells upon -NaCl versus +NaCl treatment. These genes were differentially expressed in control, but not in NFAT5^KO cells upon -NaCl versus +NaCl treatment. H, GO analysis of the upregulated genes in control cells that remain unchanged in NFAT5^KO cells upon -NaCl versus +NaCl treatment. I, FGF23-regulating genes that are statistically significantly (p < 0.05) upregulated/downregulated in control cells but remain unchanged in NFAT5^KO cells upon -NaCl versus +NaCl treatment. J, Phex mRNA levels measured by qRT-PCR upon -NaCl versus +NaCl for 24 h in UMR-106 cells. Osmolality was corrected to +NaCl by adding mannitol or urea (n = 4). FGF23, fibroblast growth factor-23; LDH, lactate dehydrogenase; NFAT5, nuclear factor of activated T cells 5.

Figure 5

Hyponatremia patients have higher FGF23 levels.A–D, serum [Na⁺] (A), cFGF23 (B), iFGF23 (C), and phosphate (D) in healthy matching controls and hyponatremic patients (n = 6, each group). E, correlation analysis between serum [Na⁺] and serum cFGF23/iFGF23. r= −0.6138, p value (two-tailed) = 0.0337 for [Na⁺] versus cFGF23; r = −0.4830, p value (two-tailed)= 0.1117 for [Na⁺] versus iFGF23. cFGF23, C-terminal fragments FGF23; FGF23, fibroblast growth factor-23; iFGF23, intact FGF23.