Rat cardiac myoblast (H9c2, 2-1) cells were acquired from the American Type Culture Collection (Manassas, VA) and cultivated in Dulbecco’s Modified Eagle Medium (GIBCO, Dublin, Ireland) containing 10% fetal bovine serum (HyClone; GE Healthcare, Bucks, UK), 100 U/mL penicillin, and 100 mg/mL streptomycin (Life Technologies, Carlsbad, CA), in a humidified atmosphere with 5% CO₂ at 37°C. The cells were cultured until 70–80% confluence and were then serum deprived for 24 h before each experiment. The cells were incubated with 0, 0.25, or 1.0 mM IS or 0, 50, or 100 ng/mL recombinant mouse FGF23 (2629-FG; R&D Systems, Minneapolis, MN) diluted in normal saline (NS) and collected after 24 and 72 h for real-time reverse transcription-polymerase chain reaction and western blotting, respectively. To perform small interfering RNA (siRNA) knockdown, the cells were transfected with On-TARGETplus SMARTpool siRNA [non-targeting control, aryl hydrocarbon receptor (AhR), and FGF23; Horizon Discovery, Cambridge, UK] using Dharmafect 1 reagent (Horizon Discovery) in accordance with the manufacturer’s instructions.

All animal experiments were conducted in accordance with the animal use protocols approved by the Committee on Ethics of Animal Experimentation at Kyushu University Graduate School of Medical Sciences (Approval number: A19-274-0). We used C57BL/6Jcl mice (CLEA Japan, Inc., Tokyo, Japan). They were maintained in an air-conditioned specific pathogen-free room at 21°C and 65% humidity, with a 12:12-h light and dark cycle (lights on at 8:00 a.m., off at 8:00 p.m.) with free access to chow and water. Experiments were reported according to the ARRIVE guidelines.

To induce LVH, 8-week-old male mice were fed a high phosphorus diet, which contained modified AIN-93G, lactose 20.0%, sucrose 2.023%, β-corn starch 20.3486%, α-corn starch 7.0%, CaCO₃ 0.55%, Ca(H₂PO₄)₂ 5.05%, and phosphate 1.5g/100g (Oriental Yeast, Tokyo, Japan) 1 day after left heminephrectomy (removal of the whole left kidney). This diet was provided for the induction of FGF23 and was administered with a continuous subcutaneous dose of 100 mg/kg IS (I3875; Sigma-Aldrich, St. Louis, MO) or 28 μl/day of NS (Otsuka Pharmaceutical Factory, Tokushima, Japan) for 4 weeks using a micro-osmotic pump (2002-0000296; Alzet, Cupertino, CA). Half of the mice were treated with a continuous intraperitoneal dose of 7.5 mg/kg H3B-6527 (H3B), which is an FGFR4 inhibitor (S8675; Selleck Biotech, Tokyo, Japan), or 3.6 μl/day of NS using a micro-osmotic pump (1004-0009922; Alzet) for 4 weeks. The 2,002-pumps were replaced biweekly. To examine the effect of IS administration on cardiac hypertrophy and fibrosis, the mice were divided into the five following groups: sham; control; IS + NS; NS + H3B; and IS + H3B. Sham mice were treated with anesthesia, skin incision, and laparotomy, but did not have pumps inserted and were fed a normal phosphorus diet (CRF-1LID10; Oriental Yeast, Tokyo, Japan). The mice were operated on day 0, observed for 4 weeks, and euthanized on day 28. The mice were anesthetized intraperitoneally with medetomidine hydrochloride (0.3 mg/kg body weight; Wako, Osaka, Japan), midazolam (4 mg/kg body weight, Sandoz, Tokyo, Japan), and butorphanol tartrate (5 mg/kg body weight, Wako). The body temperature of the mice was maintained at 37°C during the whole procedure.

The mice were euthanized on day 28. Blood samples were collected from the inferior vena cava. Serum was separated by centrifugation at 3,000 × g for 10 min, aliquoted for later analysis, and stored at −80°C. Immediately after blood collection, 50 mL ice-cold phosphate-buffered saline (14249-24; Nacalai Tesque Inc., Kyoto, Japan) (pH 7.4) was slowly perfused to harvest the heart and kidney. We separated the right kidney in half cross-sectionally for obtaining coronal sections. One half of the kidney and most of the heart were immersed in 10% formaldehyde neutral buffer solution (37152-51; Nacalai Tesque Inc.) for histological analysis. The other half of the kidney and the apex of the heart (one quarter of the heart) were snap-frozen in liquid nitrogen and stored at −80°C for protein and RNA analysis.

The hearts of mice were immersed in 10% formaldehyde neutral buffer solution for at least 48 h before being sectioned. The hearts were cut horizontally into four equal sections. The thickest wall of the left ventricle from the four sections was evaluated. Sections from each sample were subjected to hematoxylin–eosin staining using standard methods. Histological images were captured by light microscopy (Eclipse E800 microscope; Nikon, Tokyo, Japan).

Total RNA was extracted from H9c2 cells using the MAXWELL^®16 LEV simply RNA tissue Kit (Promega, Madison, WI) and the MAXWELL^® 16 instrument (Promega) in accordance with the manufacturer’s instructions. Complementary DNA was synthesized from 1 μg of total RNA with the PrimeScript RT Reagent Kit (Takara Bio Inc., Shiga, Japan). Real-time polymerase chain reaction (PCR) was performed using SYBR Premix Ex Taq™ (Takara Bio Inc.) and the Applied Biosystems 7,500 Real Time PCR System (Applied Biosystems, Foster, CA). Rat Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was amplified as an internal control. Expression analysis was performed by the Delta-Delta Ct method using 7500 Software v2.3 (Applied Biosystems). The primers used for these experiments were as follows: rat atrial natriuretic factor (ANF), forward 5′-ATGGGCTCCTTCTCCATCAC-3′ and reverse 5′-TTCATCGGTATGCTCGCTCA-3′; rat brain natriuretic peptide (BNP), forward 5′-TGGGAAGTCCTAGCCAGTCT-3′ and reverse 5′-GATCCGGTCTATCTTCTGCC-3′; rat β-myosin heavy chain (beta MHC), forward 5′-CTAGGAGGCGGAGGAACAG-3′ and reverse 5′-CTTGGCGCCAATGTCACG-3′; rat alpha smooth muscle actin (alpha SMA), forward 5′-GACACCAGGGAGTGATGGTT-3′ and reverse 5′-GTTAGCAAGGTCCGATGCTC-3′; rat collagen I, forward 5′-TGCCGTGACCTCAAGATGTG-3′ and reverse 5′-CACAAGCGTGCTGTAGGTGA-3′; rat FGF23, forward 5′-GCAACATTTTTGGATCGTATCA-3′ and reverse 5′-GATGCTT CGGTGACAGGTAGA-3′; rat N-acetylgalactosaminyltransferase 3 (GALNT3), forward 5′-GTTGCTAGGAGCAACAGTCGCA-3′ and reverse 5′-AGTTCACCGTGGTAGTATTGTAGT-3′; and rat GAPDH, forward 5′-GGCACAGTCAAGGCTGAGAATG-3′ and reverse 5′-ATGGTGGTGAAGACGCCAGTA-3′.

To perform western blot analysis, H9c2 cells were harvested with cell lysis buffer (Mammalian Protein Extraction Reagent, 78501; Pierce Thermo Scientific, Tokyo, Japan) containing protease inhibitors (#04080-11; Nacalai Tesque Inc.) and phosphatase inhibitors (#07575-51; Nacalai Tesque Inc.) on ice for 15 min. The supernatants of protein lysates were collected after 10 min of centrifugation at 10,000 × g. The protein concentrations of cell lysates were determined using a bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific). The samples (5 μg) were separated on 5–20% sodium dodecyl sulfate-polyacrylamide gels (#2331830; Atto, Tokyo, Japan) and transferred onto polyvinylidene difluoride membranes (BioRad, Hercules, CA) using Trans-Blot Turbo (BioRad). After being blocked with Blocking One (#03953-95; Nacalai Tesque Inc.) for 30 min at room temperature, the membranes were washed in Tris-buffered saline containing 0.1% Tween 20 (polyoxyethylene sorbitan monolaurate, 35624-15; Nacalai Tesque Inc.) three times for 10 min and incubated with primary antibodies at 4°C overnight. The following antibodies were used as primary antibodies: monoclonal anti-rat FGF23 antibody (1:500, MAB2629; R&D Systems); polyclonal anti-goat FGF23 antibody (1:1000, ab123502; Abcam, Cambridge, UK); polyclonal anti-rabbit FGFR4 antibody (1:1000, ab119378; Abcam); polyclonal anti-rabbit FGFR4 (phospho Y642; pFGFR4) antibody (1:1000, ab192589; Abcam); polyclonal anti-rabbit furin antibody (1:1000, PA1-062; Thermo Fisher Scientific); polyclonal anti-rabbit hypoxia-inducible factor 1 alpha (HIF1α) antibody (1:1000, NB100-134; Novus Biologicals, Centennial, CO); polyclonal anti-rabbit polypeptide GALNT3 antibody (1:1000, SAB2106736; Sigma-Aldrich); and anti-rabbit α/β tubulin (1:1000, CST#2148; Cell Signaling Technology, Danvers, MA). After being washed in Tris-buffered saline containing 0.1% Tween 20 three times, the membranes were incubated with the following horseradish peroxidase-conjugated secondary antibodies: donkey anti-rabbit IgG antibody (1:5000, NA934; GE Healthcare, Bucks, UK) and goat anti-rat IgG antibody (1:10,000, NA935; GE Healthcare) for 1 h. The bands were detected by the enhanced chemiluminescent method (ECL prime; GE Healthcare or Chemi-Lumi One Ultra; Nacalai Tesque Inc.), captured using a chemiluminescence imaging system (AE-9300 Ez-capture MG; Atto), and analyzed with ImageJ Software (National Institutes of Health, Bethesda, MD).

Data were analyzed using the JMP 14.0 software program (SAS Institute, Tokyo, Japan). Continuous variables are expressed as the mean ± standard error. Differences between two groups were compared by Student’s t-test. Differences among groups were compared by one-way analysis of variance followed by Tukey’s honest significant difference tests. P < 0.05 was considered statistically significant for all tests.

To determine if FGF23 is associated with hypertrophic signaling in cultured rat cardiac myoblast cells (H9c2), we incubated H9c2 cells with 0, 50, or 100 ng/mL recombinant mouse FGF23. These concentrations were chosen on the basis of previous reports (17–19). The cells were collected after 24 h for real-time PCR. The mRNA levels of FGF23, the LVH markers ANF, BNP, and beta MHC, and the fibrotic markers alpha SMA and collagen I were upregulated (Figure 1).

We determined the underlying mechanisms by which IS induces cardiomyocyte hypertrophy by treating H9c2 cells with IS. The H9c2 cells were cultured with 1 mM IS for 24 h, which has been reported as a clinically relevant concentration of IS in severe CKD (20–23). The mRNA levels of beta MHC, BNP, alpha SMA, collagen I, and FGF23 were upregulated (Figure 2). We also examined whether IS affects the cell size in cultured cardiomyocytes. We found that IS induced hypertrophy in H9c2 cells (Supplementary Figure 1). IS is known as a protein-bound toxin (20). Therefore, we examined whether IS induces hypertrophic and pro-fibrotic signaling in H9c2 cells with 4% albumin-containing medium as previously reported (23). IS induced hypertrophic and pro-fibrotic signaling in H9c2 cells in the 4% albumin-containing medium (Supplementary Figure 2).

In H9c2 cells cultured with 1 mM IS for 72 h, FGF23 protein expression and FGFR4 phosphorylation were elevated in cell lysates (Figures 3A–D). This FGFR4 phosphorylation by IS also occurred even in the 4% albumin-containing medium (Supplementary Figure 3). Furin, which is a protein that cleaves intact FGF23, was not downregulated by IS (Figures 3E, F).

These data suggest that IS increases FGF23 protein expression and FGFR4 phosphorylation in cardiomyocytes, and induces hypertrophic and pro-fibrotic signaling in cardiomyocytes.

In our mild CKD mouse model, we observed a heavier heart weight and greater left ventricular wall thickness in IS + NS mice than in the other mice (Figures 4A–C). Body weight, blood pressure, creatinine concentrations, calcium concentrations, and phosphate concentrations were not significantly different among the groups (Table 1). Serum IS concentrations tended to be increased in the IS-treated groups (Table 1). Inhibition of FGFR4 reduced heart weight and left ventricular wall thickness in the IS-treated groups (Figures 4A–C). Although serum FGF23 concentrations were not significantly different among the experimental groups (Table 1), FGF23 protein expression in the heart was markedly increased in IS-injected mice (Figures 5A, B). To examine the relationship between IS and FGFR4, we treated H9c2 cells with IS with or without an FGFR4 inhibitor for 24 h (Figure 6). IS upregulated mRNA expression levels of the hypertrophic markers ANF, BNP, and beta MHC, and FGFR4 inhibition reduced their expression.

To examine the relationship between IS and FGF23, we suppressed AhR, which is the receptor of IS, by siRNA in H9c2 cells and then treated cells with 1 mM IS for 6 h (for mRNA) or 48 h (for protein expression). The increase in FGF23 expression in IS-treated myocytes was suppressed by AhR siRNA (Figures 7A–C). HIF1α, which is a hypoxic stress marker that regulates FGF23 production (24), was upregulated by IS. This effect was suppressed by AhR siRNA (Figures 7D, E). GALNT3, which regulates FGF23 O-glycosylation (25), was also upregulated by FGF23 and IS (Figures 7F, G). The increased expression of GALNT3 in IS-treated myocytes was suppressed by AhR siRNA (Figures 7H–J). These data suggest that IS affects FGF23 processing by upregulating GALNT3 and by an inflammatory effect.

To examine the relationship between IS and FGF23, we suppressed FGF23 by siRNA in H9c2 cells and then treated cells with 1 mM IS for 24 h (for mRNA) or 48 h (for protein expression) (Figures 8F, 9A). The increase in mRNA levels of ANF, beta MHC, BNP, alpha SMA, collagen I, and FGF23 was suppressed by FGF23 siRNA (Figures 8A–F). The increase in FGF23 and FGFR4 phosphorylation in IS-treated myocytes was suppressed by FGF23 siRNA (Figures 9A–D).