We purchased all the reagents from Sigma-Aldrich Co (St. Louis, MO, USA) unless indicated otherwise.

Mice were kept in plastic cages with standard beddings in 12-hour light—12 h dark cycles and unlimited access to food and water. We performed the experiments with the approval of the Institutional Ethics Committee of University of Debrecen under a registration number of 10/2021/DEMÁB, and all procedures conformed to the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. Animal studies were reported in compliance with the ARRIVE guidelines.

Ten male C57BL/6 mice (8–10 weeks old, n = 5/group) were randomly divided into 2 groups: control (Ctrl) and CKD. CKD was induced by an adenine-containing diet as described previously (31, 32). In the first 6 weeks the mice received a diet containing adenine (0.2%) and elevated phosphate (0.7%) followed by adenine (0.2%) and high phosphate (1.8%) diet (S8106-S075 and S8893-S006 respectively, Ssniff, Soest, Germany) for 4 weeks.

In a separate experiment we tested the effect of the hypoxia mimetic drug Daprodustat DPD (HY-17608, MedChemExpress, NJ, USA) on calcification. To this end, 15 male C57BL/6 mice (8–10 weeks old) were divided into 3 groups (Ctrl, CKD, CKD + DPD, n = 5/group). DPD was suspended in 1% methylcellulose and was administered orally at a dose of 15 mg/kg/day between weeks 7 and 10 as described previously (31). The dose of DPD is the minimal dose that corrects anemia in C57BL/6 mice which was chosen based on our previous study (31). We euthanized the mice by CO₂ inhalation at the end of the experiments, and collected blood by cardiac puncture for analysis.

Plasma phosphate, urea and creatinine levels were assessed spectrophotometrically and by a kinetic assay respectively, on a Cobas^R 6,000 device (Roche Diagnostics, Mannheim, Germany). Hematology parameters were determined from citrate-anticoagulated whole blood by a Siemens Advia-2120i analyzer (Siemens, Tarrytown, NY, USA) with the use of 800 Mouse C57BL program of Multi Species software.

OsteoSense™ dye (OsteoSense 680 EX and NEV10020EX; PerkinElmer, MA, USA) was reconstituted in DPBS in a concentration of 20 nmol/ml. We anesthetized the mice with isoflurane inhalation and injected the dye in a dose of 2 nmol/20 g body weight through the retro-orbital venous sinus. Imaging was performed 24 h post-injection. We euthanized the mice with CO₂ inhalation, perfused with 5 ml of PBS, and analyzed the isolated hearts ex vivo by an IVIS Spectrum In Vivo Imaging System (PerkinElmer, MA, USA).

After the OsteoSense™ imaging, the isolated hearts were fixed in 10% neutral buffered formalin and were embedded in paraffin blocks and cut into 4–5 µm-thick cross-sections. Sections were deparaffinized and rehydrated followed by von Kossa and Alizarin Red stainings with standard procedures. All the sections were counterstained with hematoxylin eosin. Von Kossa staining was quantified by Image J software.

Human VICs (P10462, Innoprot, Bizkaia, Spain) were maintained in Fibroblast Medium (P60108, Innoprot) supplemented with 10% FBS (10270-106, Gibco, Grand Island, NY, USA), sodium pyruvate, L-glutamine and antibiotic antimycotic solution, according to the manufacturer's protocol. Cells were cultured at 37°C in a humidified atmosphere with 5% CO₂ content. We performed the experiments on VICs derived from 3 different donors between passages 4 and 8.

To induce calcification we exposed VICs to an osteogenic medium (OM) which was obtained by supplementing the growth medium with inorganic phosphate (Pi in the form of NaH₂PO₄ and Na₂HPO₄, pH 7.4, 2.5 mmol/L, or as indicated) and Ca (CaCl₂, 0.3 mmol/L). Both growth medium and OM were changed in every other day throughout the experiments.

To provide hypoxic environment we placed the cells into a modular incubator chamber (Billups-Rothenberg Inc, Del Mar, CA, USA). We filled the chamber with a gas mixture of 1% O₂, 5% CO₂, and 94% of N₂ (Linde, Dublin, Ireland) and applied a continuous slow flow (0.1 L/min) of the gas throughout the experiment. For normoxia, we used a gas mixture of 21% O₂, 5% CO₂, and 74% of N₂. In other experiments, we used hypoxia mimetic drugs such as desferrioxamine (DFO, 40 μmol/L), CoCl₂ (200 μmol/L) and DPD (20 µmol/L) or the HIF-1 inhibitor chetomin (Tocris, Bristol, United Kingdom, 12 nmol/L).

At the end of the experiment we washed the cells with PBS, and fixed with 4% paraformaldehyde for 20 min. After rinsing with PBS we stained the cells with Alizarin Red S solution (2%, pH 4.2) for 10 min at room temperature. Following this we applied several washes with deionized water to remove unbound dye. After taking pictures of the staining, we dissolved the dye in 100 µl of 100 mmol/L hexadecylpyridinium-chloride and determined optical density at 560 nm. Experiments were repeated at least three times minimum in triplicates.

VICs cultured in 96-well plates were washed with PBS and decalcified with HCl for 30 min at room temperature. We measured Ca content from HCl-containing supernatants with QuantiChrom Calcium Assay kit (Gentaur, Kampenhout, Belgium). To obtain protein concentration, we washed the cells with PBS and lysed in a lysis buffer containing NaOH (0.1 mol/L) and sodium dodecyl sulphate (0.1%). We determined protein concentration with BCA protein assay kit (ThermoFisher, Waltham, MA, USA) and nomalized Ca content of the cells to protein content. Experiments were repeated at least three times in triplicates.

VICs were cultured in 6-well plates. After removing the medium, we added 100 μl of EDTA (0.5 mol/L, pH 6.9) to the wells. We quantified OCN content of the EDTA-solubilized samples by an enzyme-linked immunosorbent assay (Bio-Techne R&D Systems, Minneapolis, MN, USA). OCN content was normalized to protein content and expressed as ng OCN/mg protein. Experiments were repeated at least three times in duplicates.

RNA was isolated from the hearts of the mice with Tri reagent (Molecular Research Center, Cincinnati, OH, USA) according to the manufacturer's protocol. To prepare cDNA we used High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, USA). The qPCR was carried out on a BioRad CFX96 Real-time System (Bio-Rad, Hercules, CA, USA) with the use of iTaq™ Universal SYBR® Green Supermix (Bio-Rad) and predesigned primers to detect mRNA levels of HIF-1α, HIF-2α, Runx2 and Sox9 (Table 1). We used the comparative Ct method to calculate the expression level of the transcripts, and mouse HPRT was used for normalization as internal control. Experiments were repeated at least three times in triplicates.

We lysed VICs in Laemmli lysis buffer and the cell lysate was resolved by SDS-PAGE (7.5%–10%). Proteins were blotted onto nitrocellulose membranes (Amersham, GE Healthcare, Chicago, IL, USA). Western blotting was performed with the use of primary antibodies listed in Table 2. Secondary antibodies—horseradish peroxidase linked rabbit (NA-934) and mouse IgG (NA-931) (Amersham)—were applied at a concentration of 0.5 µg/ml. Blots were developed with enhanced chemiluminescence system Clarity Western ECL (BioRad, Hercules, CA, USA). Chemiluminescent signals were either detected on an x-ray film or with a C-Digit Blot Scanner (LI-COR Biosciences, Lincoln, NE, USA). Following the development, all membranes were stripped and re-probed for β-actin using anti-β-actin antibody at a concentration of 0.5 µg/ml (sc-47778, Santa Cruz Biotechnology Inc., Dallas, TX, USA). We used the inbuilt software of the C-Digit Blot Scanner for quantification. Experiments were repeated three times.

We used Lipofectamine RNAiMAX reagent (Invitrogen, Carlsbad, CA, USA) to transfect VICs with siRNA. We followed the protocol that was provided by the manufacturer. The siRNA for HIF-1α (AM16708, ID: 106498) and HIF-2α (AM16708, ID: 106446) and silencer negative control #1 (4390843) were purchased from Invitrogen. To confirm the efficiency of silencing we performed Western blot analysis. Experiments were repeated at least three times.

The level of ROS was measured with CM-H2DCFDA assay (Life Technologies, Carlsbad, CA, USA). The cells were loaded with the dye (10 μmol/L, 30 min), then washed thoroughly with HBSS. After a 4-hour treatment the cells were washed with HBSS and the fluorescence intensity was evaluated with the use of 488 nm excitation and 533 nm emission wavelengths. In some experiments, we applied the ROS inhibitor N-acetyl cysteine (NAC, 1 mmol/L) during the treatment. Experiments were repeated at least three times in quadruplicates.

We performed an MTT assay to measure cell viability. A solution of 3-[4, 5-Dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide (0.5 mg/mL in HBSS) was incubated with the cells for 4 h. Following this, we removed the MTT solution and dissolved the formazan crystals in 100 μl of DMSO. Using DMSO as a blank, we measured optical density of the samples at 570 nm. Experiments were repeated at least three times in quadruplicates.

We show all the results as mean ± SD. We used GraphPad Prism software (version 8.01, San Diego, CA, USA) to perform statistical analyses. Normality of distribution was assessed by Shapiro-Wilk test. All data passed normality and equal variance tests, therefore we used parametric tests to determine p values. Two-tailed Student's t-test (in case of two groups) and one-way ANOVA followed by Tukey's post hoc test (in case of more than two groups) were used to determine statistically significant differences between the groups. A value of p < 0.05 was considered significant.

Cardiac VC is the main cause of cardiovascular disease and mortality in CKD patients. We induced CKD in C57BL/6 mice with a two-phase diet containing adenine (0.2%) and moderately elevated phosphate (0.7%) in the first 6 weeks and adenine (0.2%) and high phosphate (1.8%) in the following 4 weeks. Control mice (Ctrl) received a standard mice diet with 0.3% phosphate content (Figure 1A). The development of CKD was associated with significant decrease in body weight (Figure 1B), and increased urea, creatinine and phosphate levels in plasma (Figures 1C–E). To address whether CKD induces osteogenic and hypoxia pathways, we determined mRNA levels of osteogenic transcription factors Runx2 and Sox9 and hypoxia markers HIF-1α and HIF-2α in the heart of Ctrl and CKD mice. Both osteogenic and hypoxia markers were elevated in the heart tissue of CKD mice in comparison to Ctrl (Figures 1F,G). Furthermore, to evaluate osteogenic activity in mouse hearts we performed OsteoSense™ staining in Ctrl and CKD mice. Fluorescent intensity of the heart tissue was higher in CKD mice compared to Ctrl mice (4.21 × 10⁸ vs. 6.99 × 10⁸p/s, p < 0.001, Figure 1H).

Osteogenic trans-differentiation and extracellular matrix (ECM) mineralization of VICs play a major role in the development of cardiac VC. To set up an in vitro model of VC we treated VICs with osteogenic medium (OM: growth medium supplemented with 2.5 mmol/L Pi and 0.3 mmol/L Ca). In response to OM we observed time-dependent upregulation of Runx2 and Sox9, the master transcription factors regulating osteogenesis and chondrogenesis respectively, as well as alkaline phosphatase (ALP) (Figures 2A,B). OM triggered calcification of VICs which was assessed by Alizarin Red staining and Ca measurement from HCl-solubilized ECM (Figures 2C,D). Furthermore, OM induced deposition of the Ca-binding protein osteocalcin (OCN) in the ECM (Figure 2E). Along with these responses, OM also triggered a hypoxia response in VICs, characterized by elevated protein expression of HIF-1α, HIF-2α and Glut-1 (Figures 2F,G).

Recent works highlighted that hypoxia signaling is activated in calcifying aorta and showed that hypoxia inducible factors (HIFs) play a critical role in osteogenic differentiation of VSMCs (26–28).

To address whether hypoxia signaling is implicated in osteogenic differentiation of VICs, we used siRNA to downregulate protein expressions of HIF-1α and HIF-2α, the regulatory subunits of the HIF complexes. Western blots revealed that the gene silencing approaches were successful (Supplementary Figures S1A,B). Knockdown of either HIF-1α or HIF-2α was associated with decreased calcification of VICs as assessed by Alizarin Red staining (Figures 3A,B) suggesting that HIF pathways are not only activated upon osteogenic stimulation, but they are actively participated in the calcification process.

After defining the crucial involvement of hypoxia signaling in phosphate-induced calcification of VICs we asked whether hypoxia influences OM-induced osteogenic differentiation and calcification of VICs. First, we exposed VICs to normoxia (21% O₂) or hypoxia (1% O₂) for 24 h and evaluated protein expressions of HIF-1α, HIF-2α and Glut-1. As expected, hypoxia triggered a hypoxia response in VICs characterized by elevated protein expression of HIF-1α, HIF-2α and Glut-1 (Figure 4A). Then we treated VICs with OM (2.5 mmol/L Pi, 0.3 mmol/L Ca) under normoxic (21% O₂) and hypoxic (1% O₂) conditions for 24 and 48 h. Compared to control, OM slightly increased Runx2 and Sox9 expressions under normoxic condition after 48 h of exposure (Figure 4B). On the other hand, hypoxia strongly upregulated Runx2 expression even in the absence of OM stimulation (Figure 4B). Osteogenic stimuli could not further increase Runx2 expression under hypoxia (Figure 4B). Compared to normoxia, Sox9 expression was elevated under hypoxia at each condition (Figure 4B). These results suggest that hypoxia may exaggerate osteogenic reprogramming of VICs.

Next, we addressed the effect of hypoxia on ECM calcification in VICs. We induced VICs calcification with OM containing calcium (0.3 mmol/L excess) and different amounts of excess Pi (1.5; 2.0; 2.5 mmol/L) under normoxic and hypoxic conditions. As revealed by Alizarin Red staining and calcium measurement, hypoxia potentiated the pro-calcification effect of Pi at each tested concentrations (Figures 4C,D). Then we investigated time-dependency of VICs calcification under normoxic and hypoxic conditions. Alizarin Red staining showed positivity after 2 days of OM exposure under hypoxic condition, whereas calcification became detectable only on day 6 under normoxia (Figure 4E). Calcium measurement from HCl-solubilized ECM also supported the finding that hypoxia potentiates and accelerates Pi-induced calcification of VICs (Figure 4F).

To see whether HIF signaling was involved in hypoxia-induced acceleration of VICs calcification, first we applied the HIF inhibitor chetomin and investigated OM-induced calcification under hypoxic condition. As shown by Alizarin Red staining and calcium measurement, chetomin inhibited calcification of VICs (Figures 5A,B). Then we knocked-down HIF-1α, HIF-2α or both with the use of target-specific siRNAs under hypoxia. Western blots revealed that the gene silencing approaches were successful (Supplementary Figures S1C,D). Silencing of either HIF-1α or HIF-2α resulted in attenuation, whereas silencing of both HIF-α subunits caused complete inhibition of hypoxia-induced calcification (Figures 5C–E), supporting the involvement of HIF signaling in hypoxia-induced VICs calcification.

Recent evidence suggested a causative role for excess ROS-mediated oxidative stress in the osteogenic differentiation of VICs (13, 16). To explore whether unfettered production of ROS is implicated in VICs calcification under hypoxia we measured ROS production in control and OM-stimulated VICs under normoxic and hypoxic conditions. Osteogenic stimulation increased ROS production under normoxia (Figure 6A). Compared to normoxia, hypoxia increased ROS production in VICs in both control and OM conditions (Figure 6A). The glutathione precursor, N-acetyl-cysteine (NAC) attenuated excessive ROS production in all conditions (Figure 6A).

Apoptotic cell death and the release of apoptotic bodies is an important calcification mechanism. Excess ROS production can trigger cell death, therefore next we investigated cell viability in control and OM-treated VICs under normoxia and hypoxia after 4 days of exposure in the presence or absence of NAC. Osteogenic stimulation triggered a decline in cell viability in normoxia and even more cell death was observed in hypoxia (Figure 6B). NAC prevented OM-induced cell death under both normoxia and hypoxia (Figure 6B). Attenuation of unfettered ROS production and cell death by NAC was associated with complete inhibition of OM-induced VICs calcification as revealed by Alizarin red staining and calcium measurements (Figures 6C,D).

Hypoxia mimetic drugs mimic the effect of real hypoxia through the stabilization of HIFα subunits. We investigated three different hypoxia mimetic drugs, cobalt-chloride (CoCl₂), desferrioxamine (DFO) and Daprodustat (DPD), to see whether they influence Pi-induced VICs calcification under normoxic condition. We treated VICs with CoCl₂ (200 µmol/L), DFO (40 µmol/L) or DPD (20 µmol/L) for 24 h and first we evaluated protein expressions of HIF-1α and HIF-2α from whole cell lysate (Figure 7A). Hypoxia mimetics increased both HIF-1α and HIF-2α levels markedly in VICs.

Next, we investigated the effects of hypoxia mimetic drugs on OM-induced calcification of VICs. We treated VICs with OM (0.3 mmol/L excess Ca, 2.5 mmol/L excess Pi) in the presence or absence of CoCl₂ (200 µmol/L), DFO (40 µmol/L) or DPD (20 µmol/L). Alizarin Red staining and calcium measurement were performed on day 5. We observed that all the three tested hypoxia mimetic drugs enhanced OM-induced calcification in VICs (Figures 7B,C). These results suggest that not only real hypoxia but also chemical activation of the HIF pathways enhances calcification of VICs.

Silencing of either HIF-1α or HIF-2α resulted in partial inhibition of OM + DPD-induced calcification as assessed by Alizarin Red staining (Figures 7D,E), pointing out the contribution of HIF signaling to the promotion of VIC calcification by DPD.

DPD is a hypoxia mimetic drug that is used to treat anemia in CKD patients in Japan. After seeing that DPD enhances VICs calcification in vitro we addressed its effect on VC in the adenine-induced CKD model in male mice. Fifteen C57BL/6 mice (8–10 weeks old, male) were randomly assigned to 3 groups, Ctrl, CKD, and CKD + DPD (Figure 8A). CKD was induced with a diet containing adenine and elevated phosphate (Figure 1A). After 6 weeks, these mice showed signs of deteriorating kidney function characterized by elevated levels of plasma urea, creatinine and phosphate levels (Supplementary Figure S2). Then we increased phosphate content of the diet, and started to administer DPD (15 mg/body weight kg/day orally) in the next 4 weeks of the experiment (Figure 8A). At 10 weeks we terminated the experiment. At this time point, anemia was developed in CKD mice, characterized by reduced Hb concentration, decreased red blood cell count and low hematocrit levels (Table 3). DPD efficiently corrected CKD-associated anemia resulting in normalized Hb concentration, red blood cell count and hematocrit levels, similar to the controls with normal renal function (Table 3). Plasma urea, creatinine and phosphate levels were similarly high in DPD- and vehicle-treated CKD mice (Figures 8B,D). To address the effect of DPD on heart calcification we performed OsteoSense^TM staining and detected higher amount of hydroxyapatite deposition in the hearts derived from DPD-treated CKD mice compared to vehicle-treated CKD mice (2.35 × 10⁹ ± 0.3 × 10⁹ vs. 1.38 × 10⁹ ± 0.17 × 10⁹p/s, p < 0.05) (Figure 8E). Additionally, we performed histological analysis of hearts derived from Ctrl, CKD and CKD + DPD mice to detect VC. We found stronger von Kossa and alizarin red staining in heart valves of CKD + DPD mice compared to CKD, whereas no calcification was detectable in the heart of Ctrl mice (Figure 8F). These results suggest that DPD—at the dose that is efficient to correct CKD-associated anemia -, can accelerate VC in male mice with CKD.