Fibroblast growth factor 23 and fibroblast growth factor receptor 4 promote cardiac metabolic remodeling in chronic kidney disease

Abstract

Chronic kidney disease (CKD) is a global health epidemic that greatly increases mortality due to cardiovascular disease. Left ventricular hypertrophy (LVH) is an important mechanism of cardiac injury in CKD. High serum levels of fibroblast growth factor (FGF) 23 in patients with CKD may contribute mechanistically to the pathogenesis of LVH by activating FGF receptor (FGFR) 4 signaling in cardiac myocytes. Mitochondrial dysfunction and cardiac metabolic remodeling are early features of cardiac injury that predate development of hypertrophy, but these mechanisms have been insufficiently studied in models of CKD. We found in wild-type mice with CKD induced by adenine diet, that morphological changes occurred in mitochondrial structure and cardiac mitochondrial and that metabolic dysfunction preceded the development of LVH. In bioengineered cardio-bundles and neonatal rat ventricular myocytes grown in vitro, FGF23-mediated activation of FGFR4 caused mitochondrial pathology, characterized by increased bioenergetic stress and increased glycolysis that preceded the development of cellular hypertrophy. The cardiac metabolic changes and associated mitochondrial alterations in mice with CKD were prevented by global and cardiac-specific deletion of FGFR4. Our findings indicate that metabolic remodeling and mitochondrial dysfunction are early cardiac complications of CKD that precede structural remodeling of the heart. Mechanistically, FGF23-mediated activation of FGFR4 causes mitochondrial dysfunction, suggesting that early pharmacologic inhibition of FGFR4 might serve as novel therapeutic intervention to prevent development of LVH and heart failure in patients with CKD.

Keywords: CKD; FGF23; FGFR4; heart failure; metabolic remodeling; mitochondrial dysfunction.

Figure 1 |. Cardiac function, remodeling, and changes to cardiac mitochondria in chronic kidney disease (CKD).

Renal and cardiac function of mice was evaluated 12 weeks after starting adenine-containing diet to induce CKD or control diet. (a) Measurement of glomerular filtration rate (GFR) and blood urea nitrogen (BUN) indicated significant kidney damage in mice fed adenine diet when compared with mice fed control diet. (b) Cardiac functional parameters of mice after 12 weeks of CKD did not indicate manifestation of impaired function or significant structural remodeling, as demonstrated by left ventricular end-systolic diameter (LVD;s), left ventricular (LV) mass, and wall thickness. (c) Expression levels of remodeling parameters Nppa, Timp1, Foxo1, and Pgc-1α indicate that hypertrophic and fibrotic remodeling had been initiated at 12 weeks of CKD. (d) Semiquantitative evaluation of mitochondria by electron microscopy revealed significant changes in cardiac mitochondrial morphology of mice with CKD. Heart tissue of mice with CKD showed a significantly increased number of damaged mitochondria per field of view and increased average mitochondrial size, further indicating mitochondrial dysfunction. (e) Representative images of electron microscopy demonstrate swelling and misalignment of mitochondria in CKD hearts. Damaged mitochondria are indicated by red arrows. Bar = 1 μm for lower-magnification images and 600 nm for higher-magnification images. (f) Mitochondrial (Mito) respiration after 12 weeks of adenine diet showed that respiration through complex I and II was significantly increased before structural cardiac remodeling was detectable. (a–c) Bar graphs represent mean ± SEM with individual values included in the graph, n ≥ 4 male mice for all experiments. *P < 0.05, **P < 0.005, ***P < 0.0005, ****P < 0.0001. Quantification of transmission electron microscopy images was performed on tissue of male and female mice. Ant A, antimycin A; JO₂, oxygen consumption rate; NADH, reduced nicotinamide adenine dinucleotide; NS, not significant; Rot, rotenone; Succ, succinate. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 2 |. Renal function, cardiac characteristics, mitochondrial changes, and metabolome of collagen type IV alpha 3 (Col4a3)^−/− mice.

Col4a3^−/− mice were evaluated for their renal, cardiac, and mitochondrial characteristics at 16 weeks. (a) Serum levels of intact fibroblast growth factor (FGF) 23 and blood urea nitrogen (BUN) indicate significant kidney damage at this time point, whereas (b) cardiac parameters determined by echocardiography did not yet show significant signs of cardiac hypertrophy or functional impairment. (c) Analysis of cardiac mitochondria revealed an increased number of damaged mitochondria and increased size of mitochondria in Col4a3^−/− mice compared with control animals without kidney damage. (d) Micrographs demonstrate swelling and misalignment of mitochondria in Col4a3^−/− hearts, similar to the morphologic changes observed in mice with adenine-induced chronic kidney disease; damaged mitochondria are indicated by red arrows. Bar = 1 μm for lower-magnification images and 600 nm for higher-magnification images. (e) Metabolomic analysis of serum and cardiac tissue of Col4a3^−/− mice revealed significant changes in serum acylcarnitines and amino acids, with a strong downward trend in keto acids. Changes to cardiac acylcarnitines in Col4a3^−/− mice trend toward increases predominantly in longer-chain acylcarnitines. Cardiac alanine and histidine were significantly downregulated and citrulline was upregulated, and other cardiac amino acids showed similar downward trends. Citrate showed a strong trend for upregulation in Col4a3^−/− mice, but did not reach significance. (a–c) Bar graphs represent mean ± SEM and individual values included in the graph, n ≥ 3 male mice for analysis of renal and cardiac parameters. For metabolomic analysis, 4 control animals and 6 Col4a3^−/− mice were evaluated. Quantification of transmission electron microscopy images was performed on tissue of male and female mice. *P < 0.05, **P < 0.005. KIC, ketoisocaproate; KIV, α-ketoisovalerate; KMV, ketomethylvalerate; LVMI, left ventricular mass index; NS, not significant. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 3 |. Changes to mitochondrial proteome and cardiac metabolome in chronic kidney disease (CKD).

(a,b) CKD leads to significant changes in the mitochondrial proteome of wild-type (WT) mice after 12 weeks of adenine diet, before structural remodeling becomes detectable. (a) A total of 118 mitochondrial genes were significantly regulated in the mitoproteome of mice with CKD compared with controls (Ctrls). (b) Kyoto Encyclopedia of Genes and Genomes pathway analysis showed that downregulated proteins represented fatty acid metabolism and amino acid degradation (top). The upregulated genes were enriched in ribosomal processes and translation (bottom). (c) Analysis of cardiac and serum metabolome showed significant changes in serum acylcarnitines, amino acids, and keto acids (left). Among upregulated serum amino acids were phenylalanine and citrulline. These were also upregulated in cardiac tissue (right). Here, more amino acids like alanine and histidine were significantly downregulated, and leucine and isoleucine showed a downward trend (P = 0.07). In contrast to serum, more cardiac acylcarnitines, predominantly longer-chain acylcarnitines, were significantly upregulated. For organic acids from cardiac tissue of CKD mice only, pyruvate reached significance but lactate and citrate showed trends (P = 0.07). For proteomic analysis, n = 5 for each group; for serum and cardiac metabolomics, n ≥ 4 for each group. *P < 0.05. KIC, ketoisocaproate; KIV, α-ketoisovalerate; KMV, ketomethylvalerate; NAD, nicotinamide adenine dinucleotide; NADH, reduced nicotinamide adenine dinucleotide.

Figure 4 |. Fibroblast growth factor receptor (FGFR) 4 regulates metabolic transcription and hypertrophy in bioengineered cardiobundles.

(a) Treatment of neonatal rat ventricular myocyte cardiobundles with fibroblast growth factor (FGF) 23 for 20 minutes significantly increased contractile force, whereas 7 days of chronic treatment led to a significant reduction in contractile force that could be rescued by coapplication of BLU9931, a selective FGFR4 inhibitor. (b) Electrophysiological function was evaluated by pacing of cardiobundles and application of Di-4-ANEPPS (6-[2-(N,N-Dibutylamino)naphthyl]ethenyl-4′-pyridinium propanesulfonate) as voltage-sensitive dye. Chronic exposure of cardiobundles to FGF23 lead to significantly longer action potential durations. (c) FGF23-treated bundles exhibited significantly lower conduction velocity that was normalized after coapplication of BLU9931. Besides functional changes, chronic FGF23 treatment also led to cardiobundle hypertrophy, indicated by the (d,g) significant increase in cross-section and (e) increased expression of hypertrophic mRNA markers Rcan1 and Trpc6. Increased expression of Rcan1 and Trpc6 was blocked by parallel treatment with BLU9931. (f) Metabolic transcription factors that were increased in chronic kidney disease mice also increased in cardiobundles after FGF23 treatment. (g) Representative images of cardiobundles indicate cellular hypertrophy after FGF23 treatment by increased myocyte cross-sections. Bars = 10 μm. (h) Gene set enrichment analysis of control and FGF23-treated cardiobundles showed an enrichment of metabolic pathways, particularly fatty acid metabolism, adipogenesis, and cholesterol homeostasis. (i) Additional enrichment was detected in pathways related to mitochondrial function, such as oxidative phosphorylation, respiratory chain, organelle fission, and organelle inner membrane. Downregulated pathways after FGF23 treatment include angiogenesis, vascular development, tumor necrosis factor (TNF)-α signaling, and P53. Bar graphs represent mean ± SEM with individual values included in the graph. n ≥ 3 for all experiments. *P < 0.05, **P < 0.005, ****P < 0.0001. APD, action potential duration; DAPI, 4′,6-diamidino-2-phenylindole; ES, enrichment score; FDR, false discovery rate; NES, normalized enrichment score. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 5 |. Fibroblast growth factor receptor (FGFR) 4 mediates metabolic remodeling in cultured cardiomyocytes.

(a,b) Cultured neonatal rat ventricular myocytes (NRVMs) responded to 48 hours of fibroblast growth factor (FGF) 23 treatment with significant hypertrophy, indicated by increased cross-sectional area and expression of prohypertrophic markers. Prohypertrophic mRNA expression and cellular hypertrophy could be mitigated by parallel treatment with the FGFR4-specific inhibitor BLU9931. (a) Bar = 30 μm. (c) NRVMs treated with FGF23 for 1 hour, before observable hypertrophy takes place, were analyzed in a Seahorse XF analyzer for extracellular acidification rate (ECAR), elevated total proton efflux rates (PERs), and glycolysis-specific PER (GlycoPER). ECAR was significantly higher in FGF23-treated cells, which could be reduced to control levels by BLU9931. PER showed elevated basal and compensatory glycolysis on FGF23 treatment; glycolysis-specific proton efflux was also increased. These FGF23-mediated effects were blocked by BLU9931 application. (c) Graphs represent 3 independent experiments. (d) Seahorse mitochondrial stress test assay showed increased basal and maximal mitochondrial respiration after FGF23 treatment of NRVMs. Adenosine triphosphate (ATP) production-linked, spare respiratory capacity and nonmitochondrial oxygen consumption rate increased in parallel after FGF23 treatment. The significant decrease in coupling efficiency and the increased proton leak indicate uncoupling of substrate oxidation and ATP synthesis after 1 hour of FGF23 treatment. Application of BLU9931 or the calcineurin inhibitor, cyclosporin A, prevented the changes to mitochondrial function caused by FGF23. Bar graphs represent mean ± SEM and individual values included in the graph. n ≥ 9 for all experiments. *P < 0.05, **P < 0.005, ****P < 0.0001. DMSO, dimethylsulfoxide; Max, maximum; NS, not significant; PBS, phosphate-buffered saline. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 6 |. Cardiac function, remodeling, and metabolomics of fibroblast growth factor receptor (FGFR) 4–Arg385 mice in the absence of chronic kidney disease (CKD).

FGFR4-Arg385 mice did not have impaired kidney function, as indicated by blood urea nitrogen (BUN) values, but beginning left ventricular hypertrophy (LVH) is detectable by increased wall thickness at 6 months of age. (a) By 24 months of age, renal function remained unchanged, and significant LVH/heart failure with preserved ejection fraction was detected in FGFR4-Arg385 mice, indicated by robust structural remodeling and increased fractional shortening. (b) mRNA expression levels of remodeling and profibrotic and prohypertrophic markers support initiation of cardiac remodeling at 6 months of age. (c) Transmission electron microscopy showed similar changes in the mitochondria of 6-month-old FGFR4-Arg385 mice, as observed in mice with adenine-induced CKD. (c) Bar = 500 nm. (d) Metabolomic analysis of FGFR4-Arg385 mice (knock-in [KI]) at 6 months of age showed significant increase in some serum acylcarnitines and reduction in cardiac medium- and long-chain acylcarnitines compared with wild-type (WT) animals. (e) Similar to WT CKD animals, cardiac citrulline was upregulated, whereas several other amino acids, including leucine and isoleucine, were downregulated. Reduction of serum keto acids was also in line with results obtained from the adenine CKD model. (f) Organic acids also showed similar changes with a significant upregulation of pyruvate and a downregulation of lactate. Bar graphs represent mean ± SEM and individual values included in the graph. n ≥ 6 for all experiments. *P < 0.05, **P < 0.005, ***P < 0.0005. KIC, ketoisocaproate; KIV, α-ketoisovalerate; KMV, ketomethylvalerate; LV, left ventricular. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 7 |. Global deletion of fibroblast growth factor receptor (FGFR) 4 prevents left ventricular hypertrophy (LVH) and changes to cardiac mitoproteome in chronic kidney disease (CKD).

(a) Mice with global deletion of FGFR4 develop CKD to the same degree as control (Ctrl) mice after 16 weeks of adenine diet, as indicated by the increase in fibroblast growth factor (FGF) 23. (b) Wild-type (WT) animals developed LVH at 16 weeks with increased left ventricular (LV) mass, wall thickness, and ratio of heart weight/tibia length. These changes were absent in FGFR4^−/− mice. Cardiac mitoproteome ofWT and FGFR4^−/− mice was evaluated after 12 weeks of adenine feeding, before overt remodeling is observed. A total of 22 proteins were significantly regulated in WT CKD mice, but were not changed in FGFR4^−/− CKD mice, with 9 proteins downregulated and 13 upregulated. (c) Analysis showed enrichment in pathways connected to mitochondrial respiration and function. (d) Additionally, 163 proteins were identified that were only regulated in FGFR4^−/− CKD mice, with 83 downregulated and 57 upregulated proteins. (e) Enrichment analysis showed a partial normalization of mitochondrial proteins in FGFR4^−/− mice. Bar graphs represent mean ± SEM and individual values included in the graph. n ≥ 3 for all experiments. *P < 0.05, ***P < 0.0005, ****P < 0.0001. ADP, adenosine diphosphate; ATP, adenosine triphosphate; KO, knockout; NADH, reduced nicotinamide adenine dinucleotide; NAD(P)H, reduced nicotinamide adenine dinucleotide phosphate; NS, not significant.

Figure 8 |. Cardiomyocyte expression of fibroblast growth factor receptor (FGFR) 4 mediates metabolic remodeling in adenine-induced chronic kidney disease.

After 16 weeks of adenine diet, control animals and alpha-myosin heavy chain^MerCreMer–FGFR4^flox (FGFR4 conditional knockout [cKO]) mice developed kidney damage to a similar degree with elevated fibroblast growth factor (FGF) 23. (a) Left ventricular (LV) mass, wall thickness, and ratio of heart weight/tibia length were significantly lower in FGFR4 cKO mice than control animals. (b) Cardiac-specific deletion of FGFR4 also significantly reduced the cardiac expression of prohypertrophic and profibrotic markers. (c) Echocardiography showed no structural remodeling in the hearts of FGFR4 cKO mice, whereas hearts of control animals showed significant wall thickening and remodeling. (d) Analysis of the cardiac metabolome showed elevation of a greater number of medium- and long-chain acylcarnitines when compared with control animals. Expression levels of organic acids indicate a normalization of glucose use in FGFR4 cKO mice and reduction in cardiac pyruvate and citrate concentrations. Amino acids were unchanged between groups. Bar graphs represent mean ± SEM and individual values included in the graph. n ≥ 15 for all experiments. *P < 0.05, **P < 0.005. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.