Segregating the effects of ferric citrate-mediated iron utilization and FGF23 in a mouse model of CKD

Abstract

Ferric citrate (FC) is an approved therapy for chronic kidney disease (CKD) patients as a phosphate (Pi) binder for dialysis-dependent CKD, and for iron deficiency anemia (IDA) in non-dialysis CKD. Elevated Pi and IDA both lead to increased FGF23, however, the roles of iron and FGF23 during CKD remain unclear. To this end, iron and Pi metabolism were tested in a mouse model of CKD (0.2% adenine) ± 0.5% FC for 6 weeks, with and without osteocyte deletion of Fgf23 (flox-Fgf23/Dmp1-Cre). Intact FGF23 (iFGF23) increased in all CKD mice but was lower in Cre⁺ mice with or without FC, thus the Dmp1-Cre effectively reduced FGF23. Cre⁺ mice fed AD-only had higher serum Pi than Cre^- pre- and post-diet, and the Cre⁺ mice had higher BUN regardless of FC treatment. Total serum iron was higher in all mice receiving FC, and liver Tfrc, Bmp6, and hepcidin mRNAs were increased regardless of genotype; liver IL-6 showed decreased mRNA in FC-fed mice. The renal 1,25-dihydroxyvitamin D (1,25D) anabolic enzyme Cyp27b1 had higher mRNA and the catabolic Cyp24a1 showed lower mRNA in FC-fed mice. Finally, mice with loss of FGF23 had higher bone cortical porosity, whereas Raman spectroscopy showed no changes in matrix mineral parameters. Thus, FC- and FGF23-dependent and -independent actions were identified in CKD; loss of FGF23 was associated with higher serum Pi and BUN, demonstrating that FGF23 was protective of mineral metabolism. In contrast, FC maintained serum iron and corrected inflammation mediators, potentially providing ancillary benefit.

Keywords: CKD; GF23; ferric citrate; iron; kidney; klotho.

FIGURE 1

FGF23 with conditional deletion and Ferric citrate (FC) treatment. (a) Post‐treatment intact FGF23 demonstrated significantly lower levels in Cre⁺ mice versus Cre⁻ counterparts; no differences were observed in iFGF23 levels in the FC treated mice versus AD only groups. Three statistical outliers removed. (b) Both Cre⁺ mouse groups showed lower bone Fgf23 mRNA expression compared to Cre⁻ groups and the FC‐treated Cre⁻ group had lower Fgf23 mRNA versus AD‐only controls (n = 4–8 mice per group; *p < 0.05; **p < 0.01 versus genotype, same treatment; ^# p < 0.05 versus treatment, same genotype; ^$ p < 0.05 versus different treatment and genotype via two‐way ANOVA with a Tukey post‐hoc test). Gray points represent data outliers; the dotted line represents relative levels of gene expression in the casein‐diet controls

FIGURE 2

Vitamin D metabolic genes in chronic kidney disease. (a) Kidney Cyp27b1 mRNA levels were higher in the mice fed Ferric citrate (FC) regardless of genotype. (b) Kidney Cyp24a1 mRNA levels were lower in the mice on FC regardless of genotype. (n = 4–8 mice per group; ^# p < 0.05, ^## p < 0.01 versus treatment, same genotype via two‐way ANOVA with a Tukey post‐hoc test). Dotted line represents relative levels of gene expression in casein‐diet controls

FIGURE 3

Markers of inflammation and fibrosis. (a) Liver IL‐6 mRNA levels were lower in the mice fed Ferric citrate than those receiving AD only. (b) Liver Col1a1 mRNA levels, (c) kidney Col1a1 mRNA levels, and (d) kidney Egr1 mRNA expression, were not different across groups. (n = 4–8 mice per group; ^# p < 0.05, ^## p < 0.01 versus treatment, same genotype via two‐way ANOVA with a Tukey post‐hoc test). Dotted line represents relative levels of gene expression in the casein‐diet controls

FIGURE 4

Genes controlling iron utilization. (a) Liver Tfrc mRNA expression was higher in the mice receiving Ferric citrate (FC) regardless of genotype, and Tfrc mRNA in Cre⁺ mice fed FC was higher than the Cre⁻ control. (b) Liver Bmp6 mRNA levels were higher in the mice treated with FC regardless of genotype. (c) Liver Hamp mRNA was elevated in the mice fed FC regardless of genotype. (n = 4–8 mice per group; *p < 0.05 versus genotype, same treatment; ^# p < 0.05; ^## p < 0.01 versus treatment, same genotype via two‐way ANOVA with a Tukey post‐hoc test). Dotted line represents relative levels of gene expression in the casein‐diet controls

FIGURE 5

Cortical bone structure. (a) Cortical thickness, (b) cortical bone area, and (c) cortical tissue area were not different between groups. (d) Cortical porosity was higher in the Cre⁺ Ferric citrate (FC)‐treated mice. (e) The median µCT image for each group is shown. (n = 4−8 mice per group; **p < 0.01 versus genotype, same treatment; ^# p < 0.05, ^## p < 0.01 versus treatment, same genotype via two‐way ANOVA with a Tukey post‐hoc test)