Klotho and phosphate are modulators of pathologic uremic cardiac remodeling

Abstract

Cardiac dysfunction in CKD is characterized by aberrant cardiac remodeling with hypertrophy and fibrosis. CKD is a state of severe systemic Klotho deficiency, and restoration of Klotho attenuates vascular calcification associated with CKD. We examined the role of Klotho in cardiac remodeling in models of Klotho deficiency-genetic Klotho hypomorphism, high dietary phosphate intake, aging, and CKD. Klotho-deficient mice exhibited cardiac dysfunction and hypertrophy before 12 weeks of age followed by fibrosis. In wild-type mice, the induction of CKD led to severe cardiovascular changes not observed in control mice. Notably, non-CKD mice fed a high-phosphate diet had lower Klotho levels and greatly accelerated cardiac remodeling associated with normal aging compared with those on a normal diet. Chronic elevation of circulating Klotho because of global overexpression alleviated the cardiac remodeling induced by either high-phosphate diet or CKD. Regardless of the cause of Klotho deficiency, the extent of cardiac hypertrophy and fibrosis correlated tightly with plasma phosphate concentration and inversely with plasma Klotho concentration, even when adjusted for all other covariables. High-fibroblast growth factor-23 concentration positively correlated with cardiac remodeling in a Klotho-deficient state but not a Klotho-replete state. In vitro, Klotho inhibited TGF-β1-, angiotensin II-, or high phosphate-induced fibrosis and abolished TGF-β1- or angiotensin II-induced hypertrophy of cardiomyocytes. In conclusion, Klotho deficiency is a novel intermediate mediator of pathologic cardiac remodeling, and fibroblast growth factor-23 may contribute to cardiac remodeling in concert with Klotho deficiency in CKD, phosphotoxicity, and aging.

Keywords: CKD; heart disease; hypertrophy; ischemia-reperfusion; transgenic mouse; uremia.

Copyright © 2015 by the American Society of Nephrology.

Figures

Figure 1.

Klotho deficiency impairs cardiac function. kl/+, WT, and Tg-Kl mice at 12 weeks of age fed with normal diet were subjected to cardiac MRI under anesthesia. (A) Left ventricular ejection fraction, (B) left ventricular stroke volume, (C) cardiac output, (D) left ventricular wall thickness at systole, and (E) left ventricular wall thickness at diastole. Data expressed as means±SDs of eight mice from each group, and statistical significance was assessed by one-way ANOVA followed by Newman–Keuls test. Significant differences were accepted when *P<0.05 or **P<0.01 between groups.

Figure 2.

Klotho-deficient mice with severe cardiac remodeling. (A, left panel) Representative macrograph of sagittal sections of the hearts (Trichrome) of kl/kl, kl/+, and WT mice at 6 and 12 weeks of age. (A, right panel) Shows semiquantification of the Trichrome-positive area over the whole-heart section (Image J). (B) Representative micrographs showing cardiac fibrosis by Trichrome staining and immunohistochemistry for α/β-MHC in the left ventricle of kl/kl and WT mice at 12 weeks of age. (C) Representative immunohistochemistry for P-Erk (green) and collagen I (blue). (D) α-SMA, P-Smad2/3, and Alexa-Fluor–WGA (red) in left ventricular sections of kl/kl and WT mice at 12 weeks old. (E) Representative immunoblots for α-actinin and β-actin. (F) P/T-Smad2/3 and P/T-Erk in left ventricular lysates from 6- and 12-week-old mice. (G) Summary of immunoblots in arbitrary units. Means±SDs (n=4 from each group). Statistical significance was assessed by one-way ANOVA followed by Newman–Keuls test. Significant differences were accepted when *P<0.05 or **P<0.01 between groups. α-SMA, α-smooth muscle actin; T, total; WGA, wheat germ agglutinin.

Figure 3.

Cardiac hypertrophy and fibrosis in CKD mice. WT mice at 12 weeks of age underwent CKD surgery (UNX-IRI) or laparotomy (sham). One month after surgery, mice were fed with high-phosphate diet for 12 weeks. (A, upper panel) Representative macrographs of sagittal sections (H&E). Scale bar, 3 mm. (A, lower panel) Summary of the ratio of HW to body weight of sham and CKD mice. (B, upper panel) Representative micrographs of left ventricular sections (Trichrome). Scale bar, 100 μm. (B, lower panel) Summary of semiquantification of the Trichrome-positive area over the whole-heart section by Image J. (C, left panel) Representative immunoblots for P/T-Smad2/3, P/T-Erk, and α-SMA. Collagen I and β-actin in left ventricular lysates from sham and CKD were induced by UNX-IRI. Summaries are shown in C, right panel. Data are means±SDs (n=4 from each group). Statistical significance was assessed by t test. Significance was accepted when **P<0.01 between groups. α-SMA, α-smooth muscle actin; H&E, hematoxylin and eosin; T, total.

Figure 4.

Synergism of high-phosphate diet, Klotho deficiency, and aging on cardiac remodeling. kl/+, WT, and Tg-Kl mice at 6 or 12 months of age were fed normal or high-phosphate diet for 12 weeks and terminated at age 9 or 15 months, respectively. Plasma, urine, and kidney Klotho at (A) 9 and (B) 15 months old were determined by immunoprecipitation-immunoblotting or immunoblotting. Right panels summarize plasma, urine, and kidney Klotho protein levels. (C, upper panel) Representative macrographs of heart sections (Trichrome); (C, lower panel) Summarizes semiquantification of the Trichrome-positive area over the whole-heart section (Image J). (D) Representative immunohistochemistry for P-Erk, collagen I, and Alexa-Fluor–WGA in left ventricular sections of kl/+, WT, and Tg-Kl mice at 9 and 15 months old. (E) α-SMA, P-Smad2/3, and Alexa-Fluor–WGA in left ventricular sections of kl/+, WT, and Tg-Kl mice at 9 and 15 months. Data expressed as means±SDs (n=4 from each group). Statistical significance was assessed by one-way ANOVA followed by Newman–Keuls test. Significant differences were accepted when *P<0.05 or **P<0.01 between groups. α-SMA, α-smooth muscle actin; HC, heavy chain; WGA, wheat germ agglutinin.

Figure 5.

In vitro effect of soluble recombinant Klotho protein on neonatal cardiac fibroblasts and cardiomyocytes. (A and B) Cardiac fibroblasts were prepared from neonatal rat hearts and cultured in six-well plates. (A) After full confluence, TGF-β1, Ang II, and high Pi were added with Klotho or vehicle for 24 hours, and total cell lysate was immunoblotted for (left panel) TGF-β1, (center panel) Ang II, and (right panel) high Pi effect on CTGF and collagen I protein expression. (B) TGF-β1, Ang II, and high Pi were added with Klotho or vehicle for 30 minutes, and total cell lysate was prepared for immunoblot to examine (left panel) TGF-β1, (center panel) Ang II, and (right panel) high Pi on phosphorylation of Erk and Smad2/3 in cardiac fibroblasts. (Top panel) Representative immunoblots. (Middle and bottom panels) Summary of all experiments. (C and D) Cardiomyocytes were prepared from neonatal rat hearts and cultured in six-well plates. (C) On confluence, TGF-β1, Ang II, and high Pi were added with Klotho or vehicle for 24 hours, and total cell lysate was immunoblotted for (left panel) TGF-β1, (center panel) Ang II, and (right panel) high Pi effect on CTGF, collagen I, and α-actinin. (D) TGF-β1, Ang II, and high Pi were added with Klotho or vehicle for 30 minutes, and total cell protein lysate was immunoblotted for (left panel) TGF-β1, (center panel) Ang II, and (right panel) high Pi effect on phosphorylation of Erk and Smad2/3 in cardiomyocytes. (Top panel) Representative immunoblots. (Middle and bottom panels) Summary of all experiments. Data are expressed as means±SDs of three independent experiments for each group, and statistical significance was assessed by one-way ANOVA followed by Newman–Keuls test. Significant differences when *P<0.05 or **P<0.01 between groups. CTGF, connective tissue growth factor; T, total.

Figure 6.

Summary of Klotho levels and cardiac phenotype: effect of age, high-phosphate diet, and CKD on plasma Klotho levels and pathologic cardiac remodeling. (A) Plasma Klotho levels normalized to that of 3-month-old WT mice with normal renal function on a normal Pi diet (set to 100%; z axis). Groups of animals: age (3, 9, and 15 months) and dietary phosphate (normal phosphate [N Pi], 0.9%; high phosphate [H Pi], 2.0%; high-phosphate diet plus CKD [H Pi+CKD]). (B) Double log10 plot of HW/body weight (BW) and (C) double log10 plot of percentage of cardiac fibrosis versus plasma Klotho concentration in all groups of animals. The relationships between log10 plasma Klotho concentration and log10 HW/BW or log10 cardiac fibrosis were determined by linear correlation analysis, and significant association was accepted when P<0.05. (D) Double log10 plot of HW/BW and (E) double log10 plot of percentage of cardiac fibrosis versus plasma phosphate concentration in all groups of animals. The relationships between log10 plasma phosphate concentration and log10 HW/BW or log10 cardiac fibrosis were determined by linear correlation analysis, and significant association was accepted when P<0.05. (F) Unadjusted and adjusted association of cardiac hypertrophy and fibrosis with plasma parameters. The association between cardiac variables (HW/BW and percentage of cardiac fibrosis) with each variable (Klotho, FGF23, PTH, 1,25 Vit D, Pi, and creatinine) was assessed with Pearson correlation coefficients with the SAS program (v9.3; SAS Institute, Cary, NC). For each plasma parameter, the partial correlation coefficients were computed to adjust for the potential confounding effects of the other five biomarkers in evaluating the association with cardiomyopathy. P≤0.05 was considered statistically significant. r, Pearson correlation coefficient. *Adjusted for the other five plasma parameters.

Figure 7.

Association of plasma FGF23 levels with cardiac remodeling in three different plasma Klotho levels. (A and B) Plasma Klotho (x axis) and FGF23 (y axis) levels were divided into tertiles, and the effects on (A) cardiac hypertrophy and (B) cardiac fibrosis are depicted. (C and D) Associations of plasma FGF23 with (C) cardiac hypertrophy and (D) fibrosis were studied with three different plasma Klotho levels. The relationship between log10 plasma FGF23 concentration and (C) HW/body weight (BW) or (D) cardiac fibrosis was determined by Pearson correlation analysis, and significant association is accepted when P<0.05.