Elevated FGF23 and disordered renal mineral handling with reduced bone mineralization in chronically erythropoietin over-expressing transgenic mice

Abstract

Fibroblast Growth Factor 23 (FGF23) is a phosphaturic factor causing increased renal phosphate excretion as well as suppression of 1,25 (OH)₂-vitamin D_3. Highly elevated FGF23 can promote development of rickets and osteomalacia. We and others previously reported that acute application of erythropoietin (EPO) stimulates FGF23 production. Considering that EPO is clinically used as chronic treatment against anemia, we used here the Tg6 mouse model that constitutively overexpresses human EPO in an oxygen-independent manner, to examine the consequences of long-term EPO therapy on mineral and bone metabolism. Six to eight weeks old female Tg6 mice showed elevated intact and C-terminal fragment of FGF23 but normal plasma levels of PTH, calcitriol, calcium and phosphate. Renal function showed moderate alterations with higher urea and creatinine clearance and mild albuminuria. Renal phosphate excretion was normal whereas mild hypercalciuria was found. Renal expression of the key proteins TRPV5 and calbindin D28k involved in active calcium reabsorption was reduced in Tg6 mice. Plasma levels of the bone turnover marker osteocalcin were comparable between groups. However, urinary excretion of deoxypyridinoline (DPD) was lower in Tg6 mice. MicroCT analysis showed reduced total, cortical, and trabecular bone mineral density in femora from Tg6 mice. Our data reveal that chronic elevation of EPO is associated with high FGF23 levels and disturbed mineral homeostasis resulting in reduced bone mineral density. These observations imply the need to study the impact of therapeutically applied EPO on bone mineralization in patients, especially those suffering from chronic kidney disease.

Figure 1

Over-expression of EPO induced high hematocrit, low iron, low liver hepcidin expression, and splenomegaly in Tg6 mice. Biological markers of EPO activity were assessed in 6–8 weeks old female Tg6 mice and their littermates: (A) hematocrit, (B) total plasma iron, (C) liver hepcidin mRNA expression, and (D) spleen weight normalized to body weight. Data are means ± s.e.m.; n = 5–7 mice for each genotype. Significance was determined by unpaired t test and indicated as *p < 0.05, **p < 0.01.

Figure 2

Markers of kidney function in Tg6 mice. Markers of kidney function were examined in 6–8 weeks old female Tg6 mice and their wildtype littermates: (A) urine volume per 24 hours, (B) plasma urea, (C) plasma creatinine, (D) urine urea, (E) urine creatinine, (F) urea clearance (ml/min), (G) creatinine clearance (ml/min), and (H) urine albumin excretion on a Coomassie blue stained gel with bovine serum albumin (BSA) as a positive control. Data are means ± s.e.m.; n = 5–12 for each group of mice. Significance was determined by unpaired t test and indicated as *p < 0.05, **p < 0.01.

Figure 3

Higher urinary excretion of Ca²⁺ and elevated levels of both forms of FGF23 in Tg6 mice. (A,B) Plasma and urine concentrations of phosphate (Pi) and total calcium (Ca²⁺) in 6–8 weeks old female Tg6 mice and their littermates. (C) Plasma intact FGF23, (D) plasma C-terminal fragment of FGF23, (E), plasma 1,25 (OH)₂ vitamin D_3, and (F) plasma parathyroid hormone (PTH). (G) Renal α-klotho protein expression was detected by immunoblotting in kidney extracts of Tg6 mice and their controls. Membranes were reblotted for β-actin. The β-actin to α-klotho ratio is shown as a graph. Data are means ± s.e.m.; n = 5–10 for each group of mice. Significance was determined by unpaired t test and indicated as *p < 0.05, **p < 0.01.

Figure 4

Chronic EPO over-expression induced FGF23 expression in spleen and liver of Tg6 mice. FGF23 mRNA levels were examined in (A) femur, (B) liver, and (C) spleen of 6–8 weeks old female Tg6 mice versus control. Expression levels of FGF23 were normalized to ribosomal 18 s. Data are means ± s.e.m.; n = 5–7 for each group of mice. Significance was determined by unpaired t test and indicated ***p < 0.001.

Figure 5

Normal 1,25 (OH)₂ vitamin D₃ metabolism in Tg6 mice. Expression of the enzymes metabolizing 1,25 (OH)₂ vitamin D₃ and the vitamin D receptor (VDR) was examined by immunoblotting in kidney and ileum of 6–8 weeks old female Tg6 mice and their controls. (A) Renal expression of Cyp24a1 and the vitamin D receptor (VDR) and (B) ileal expression of Cyp24a1 and the VDR. For immunoblots, all membranes were stripped and re-probed for β-actin. Graphs show the ratio of the protein of interest over β-actin. Data are means ± s.e.m.; n = 5 for each group of mice.

Figure 6

EPO over-expression reduced expression of renal proteins involved in Ca²⁺ and phosphate transport in Tg6 mice. The protein expression of the (A) sodium phosphate co-transporter NaPi-IIa, (B) the epithelial Ca²⁺ channel TRPV5, and (C) the cytoplasmic Ca²⁺ -binding protein Calbindin-D28k protein was tested by immunoblotting of kidney extracts from 6–8 weeks old female Tg6 mice and their littermate controls. All membranes were stripped and reprobed for β-actin. Graphs show the ratio of the protein of interest over β-actin. Data are means ± s.e.m.; n = 5 for each group of mice. Significance was determined by unpaired t test and indicated as *p < 0.05 and **p < 0.01.

Figure 7

Bone phenotype of 6–8 weeks old female Tg6 mice. The bone phenotype of 6–8 weeks old female Tg6 mice and their wildtype littermates was examined by assessment of bone turnover markers in plasma and urine. (A) Plasma osteocalcin and (B) urinary DPD excretion/24 hrs. (C) Total, cortical and trabecular bone mineral density (BMD).

Figure 8

Bone phenotype of 6–8 weeks old female Tg6 mice. The bone phenotype of 6–8 weeks old female Tg6 mice and their wildtype littermates was examined by μCT. Representative cross-sections microCT images of distal femora of WT versus Tg6 are shown. Dark grey area indicates mineralized bone. Data are means ± s.e.m.; n = 5–9 for each group of mice. Significance was determined by unpaired t test and indicated as *p < 0.05 and ***p < 0.001.