Advanced glycation end products accelerate calcification in VSMCs through HIF-1α/PDK4 activation and suppress glucose metabolism

Arterial media calcification is associated with diabetes mellitus. Previous studies have shown that advanced glycation end products (AGEs) are responsible for vascular smooth muscle cell (VSMC) calcification, but the underlying mechanisms remain unclear. Hypoxia-inducible factor-1α (HIF-1α), one of the major factors during hypoxia, and pyruvate dehydrogenase kinase 4 (PDK4), an important mitochondrial matrix enzyme in cellular metabolism shift, have been reported in VSMC calcification. The potential link among HIF-1α, PDK4, and AGEs-induced vascular calcification was investigated in this study. We observed that AGEs elevated HIF-1α and PDK4 expression levels in a dose-dependent manner and that maximal stimulation was attained at 24 h. Two important HIF-1α-regulated genes, vascular endothelial growth factor A (VEGFA) and glucose transporter 1 (GLUT-1), were significantly increased after AGEs exposure. Stabilization or nuclear translocation of HIF-1α increased PDK4 expression. PDK4 inhibition attenuated AGEs-induced VSMC calcification, which was evaluated by measuring the calcium content, alkaline phosphatase (ALP) activity and runt-related transcription factor 2 (RUNX2) expression levels and by Alizarin red S staining. In addition, the glucose consumption, lactate production, key enzymes of glucose metabolism and oxygen consumption rate (OCR) were decreased during AGEs-induced VSMC calcification. In conclusion, this study suggests that AGEs accelerate vascular calcification partly through the HIF-1α/PDK4 pathway and suppress glucose metabolism.

AGEs enhanced HIF-1α and PDK4 expression during VSMC calcification. To investigate the effects of AGE-BSA treatment on HIF-1α and PDK4 expression, VSMCs were treated with AGE-BSA (0, 50, 100, 200, and 400 μg/ml) containing 10 mM β-GP for 24 h. HIF-1α and PDK4 protein and mRNA expression levels were determined by western blotting and qRT-PCR. We found that the protein and mRNA expression levels of HIF-1α and PDK4 were significantly increased in a dose-dependent manner ( Fig. 2A,C). Then, we incubated VSMCs with AGE-BSA (200 μg/ml) containing 10 mM β-GP for 0, 6,12,24,48, and 72 h. The protein and mRNA expression levels of HIF-1α and PDK4 were analyzed by western blotting and qRT-PCR. We observed that HIF-1α and PDK4 protein and mRNA expression levels were increased in AGE-BSA-treated groups compared with the normal control groups, and this increase was maximal after 24 h of stimulation (Fig. 2B,D). Taken together, these results indicate that HIF-1α and PDK4 transcription and translation are increased during AGEs-induced VSMC calcification.
AGEs promoted HIF-1α nuclear translocation and HIF-1α-regulated gene expression. As mentioned above, HIF-1α expression levels were elevated in VSMCs treated with AGE-BSA. To further demonstrate that increased transcriptional expression of HIF-1α is associated with AGE-BSA treatment, VSMCs were incubated with AGE-BSA or BSA (200 μg/ml) containing 10 mM β-GP for 24 h. HIF-1α nuclear translocation was visualized by immunofluorescence with a confocal microscope. We observed that AGE-BSA significantly promoted HIF-1α translocation into the nucleus compared with the BSA group (Fig. 3A). Western blotting also showed that  Figure 1A), further suggesting that AGE-BSA can promote HIF-1α nuclear translocation.

PDK4 knockdown alleviated VSMC calcification.
To explore whether AGEs accelerate VSMC calcification through a PDK4-dependent pathway, small interfering RNA (siRNA) was used for PDK4 knockdown. The knockdown efficiency was nearly 75% after VSMCs transfected with siRNA against PDK4 for 24 h compared with that in the scrambled siRNA group (Fig. 5A). PDK4 siRNA also presented effective transfection efficiency at day 7, although the efficiency was weaker than that observed on the previous 3 days (Supplementary Figure 3). AGE-BSA treatment alone without calcium medium did not obviously cause VSMC calcification and played only an accelerating role (Supplementary Figure 4), and thus, AGE-BSA was used in the presence of β-GP. VSMCs transfected with PDK4 siRNA or scrambled siRNA were cultured in calcium medium with or without AGE-BSA (200 μg/ml). Since RUNX2 is an important factor during VSMC osteoblastic differentiation, we postulated that PDK4 is an upstream molecule of RUNX2. As shown in Fig. 5B, AGE-BSA treatment significantly increased the RUNX2 expression level, as previously reported 12 , and PDK4 knockdown via siRNA down-regulated RUNX2 protein levels, suggesting that the AGEs/PDK4/RUNX2 signaling pathway is present in VSMC calcification. In addition, PDK4 inhibition decreased AGEs-induced ALP activity and calcium deposition content (Fig. 5C). Furthermore, PDK4 inhibition via DCA markedly decreased AGEs-induced calcified nodule formation as shown by Alizarin red S staining (Fig. 5D). All these results reveal that PDK4 plays an important role in AGEs-induced VSMC calcification. Effect of AGEs on glucose metabolism changes during VSMC calcification. Since HIF-1α has been reported as a glycolysis inducer 29 and PDKs are regulators of the cellular energy metabolism shift, we attempted to investigate whether AGEs could enhance glycolysis during VSMC calcification through HIF-1α and PDK4. We first cultured VSMCs with AGE-BSA in calcium medium to detect lactate production and glucose consumption. Unexpectedly, the level of lactate, an important product of glycolysis, was decreased in a dose-dependent manner at different time points, and glucose consumption was also decreased in a dose-dependent manner at different time points (Fig. 6A). We next analyzed key enzymes of glucose metabolism; VSMCs were treated with AGE-BSA in the presence of calcium medium for 24 h, and qRT-PCR showed that hexokinase (HK), lactate dehydrogenase (LDH), isocitrate dehydrogenase (IDH), and glucose-6-phosphate dehydrogenase (G6PD) expression levels decreased as the concentration of AGE-BSA increased, but glucose 6-phosphatase (G6pase) levels did not change (Fig. 6B). To determine whether AGEs could influence the overall metabolism, the oxygen consumption rate (OCR) was also measured, and AGEs significantly inhibited OCR (Supplementary Figure 5). These results indicate that AGEs suppress glycolysis, aerobic oxidation, the pentose phosphate pathway, and mitochondrial respiratory capacity in VSMCs, leading to decreased glucose consumption and increased glycogen synthesis. Despite the suppressive effect of AGEs on glucose metabolism, we still speculated that PDK4 knockdown could induce glycolysis activation; VSMCs transfected with PDK4 siRNA were used to detect lactate production. As shown in Fig. 6C, PDK4 knockdown accelerated AGE-BSA-down-regulated lactate production, which shows that AGEs and PDK4 have contradictory roles in the regulation of glucose metabolism.

Discussion
This study demonstrated that AGEs increase the transcriptional and translational expression of HIF-1α and PDK4. Moreover, HIF-1α translocation and target gene expression were promoted by AGEs, and HIF-1α stabilization and nuclear translocation could regulate PDK4 expression. PDK4 knockdown by siRNA suppressed AGEs-induced VSMC calcification, as shown by RUNX2 protein levels, ALP activity, calcium deposition, and calcium nodule staining. In addition, we observed that AGEs suppressed glucose metabolism and that PDK4 siRNA silencing enhanced AGEs-down-regulated glycolysis. Taken together, we found that AGEs accelerate VSMC calcification through the HIF-1α/PDK4 signaling pathway and hinder glucose metabolism.
Vascular calcification and vascular events are strongly correlated 30 . Coronary artery calcification has been reported to be a predictor of adverse cardiac events in asymptomatic patients 31 . Oxidative stress, inflammation, apoptosis, and metabolic shifts contribute to vascular calcification through several downstream signaling cascades 12,32,33 . Recent clinical trials demonstrate that high fibroblast growth factor-23 (FGF-23) levels and high dose plus long-term statin therapy are also related to vascular calcification 34,35 . Based on previous reports, AGEs may be associated with the abovementioned vascular risks 12,32,33 ; however, the role of AGEs in the pathogenesis of diabetic vascular calcification remains unclear.
Since HIF-1α has been demonstrated to be a glycolysis promoter 26,29 and an important factor in phosphate-induced VSMC calcification 18 , we explored whether HIF-1α is involved in AGE-induced vascular calcification. We were the first to report that AGEs enhance HIF-1α transcriptional and translational expression in calcified VSMCs. In addition, AGEs induced HIF-1α translocation, further indicating the role of AGEs in transcriptional regulation. Several papers have shown that two important HIF-1α target genes, GLUT-1, a specific glucose transporter that hydrogen bonds with glucose as it moves through the membrane channel 36 , and VEGFA, which increases vascular permeability and angiogenesis 37 , induce osteoblastic differentiation and vascular calcification in different types of cells [38][39][40] . Our results demonstrated that GLUT-1 and VEGFA expression levels in VSMCs treated with AGEs were markedly increased compared with those in control VSMCs, further suggesting that HIF-1α and downstream genes participate in AGEs-induced VSMC calcification.
PDK4, an important mitochondrial matrix enzyme that controls metabolism shift, has been shown to be sensitive to intracellular ROS levels in our laboratory 12 . HIF-1α stabilization is also sensitive to cellular oxygen and mtROS 25 . Marycz K et al. reported that the oxidative stress/HIF-1α/PDK4 axis plays a key role in adipose stem cell (ASC) osteogenic differentiation 41 . AGEs were also shown to increase oxidative stress in VSMC calcification, as mentioned previously 11 . Therefore, whether AGEs-induced HIF-1α activation increases PDK4 expression levels in VSMC calcification is of interest. We found that HIF-1α stabilization or translocation up-regulated PDK4 expression, indicating that the AGEs/HIF-1α/PDK4 axis exists in VSMC calcification. In addition, PDK4 stimulation leads to mitochondrial dysfunction and excessive mtROS 14,42 ; thus, a reciprocal loop among mtROS, HIF-1α, and PDK4 may be involved in AGEs-induced calcified VSMCs.
The relationship between PDK4 and vascular calcification has been previously elucidated 12,14 . This study further confirmed the role of PDK4 in AGEs-induced VSMC calcification. We discovered that PDK4 knockdown by siRNA reduced RUNX2 expression, which is a key factor for osteogenic gene expression 43 . ALP activity, calcium deposition, and calcified nodule formation were also attenuated after PDK4 inhibition. Our study shows for the first time that AGEs accelerate VSMC calcification through a PDK4-dependent pathway. However, PDK4 is associated with metabolic dysfunction, which leads to excessive ROS formation 13 . PDK4 inhibition may also down-regulate upstream signaling through a reciprocal loop, since HIF-1α and target genes are critical for vascular calcification 18,38-40 as well as oxidative stress 11 . PDK4 interference may only partially alleviate VSMC calcification. Thus, it may be more accurate to conclude that AGEs accelerate VSMC calcification partly through a PDK4-dependent pathway.
Glucose metabolism plays an important role in vascular reactivity 44 , especially in VSMCs, which exhibit high glucose consumption and lactate production levels even under normal and well-oxygenated conditions 45 . Approximately 30% of the adenosine triphosphate (ATP) supply in VSMCs is derived from aerobic glycolysis, and 90% of the glycolysis flux contributes to lactate production 46 . During injury and atherogenesis, VSMCs will transdifferentiate from a contractile to a synthetic phenotype 47 . Lactate, the end product of glycolysis, has been reported to have an important role in promoting the synthetic phenotype in VSMCs 48 . Lactate can also induce osteoblast differentiation via HIF-1α 49 , and these findings demonstrate that enhanced glycolysis is important for VSMC phenotype differentiation and vascular function changes.
In VSMCs, PDK4 and HIF-1α are both promoters of glycolysis 12,26 ; therefore, we speculated that AGEs may stimulate glycolysis. Unexpectedly, we found that AGEs significantly suppress lactate production and glucose utilization during VSMC calcification. In addition, expression levels of HK, LDH, IDH, and G6PD, which are related to glucose metabolism but not G6pase, and mitochondrial respiratory capacity as measured by OCR was decreased after treatment with AGEs. Although reduced G6PD expression may be due to the inhibition of glucose-6-phosphate, and the down-regulation of IDH expression may be attributed to mitochondrial dysfunction caused by PDK4 activation 13 , all these findings suggest that AGEs inhibit glycolysis and impair normal mitochondrial function during VSMC calcification. A recent report has suggested that during atherogenesis, VSMCs have increased mitochondrial dysfunction and use of glycolysis; enhanced glycolysis could be a compensatory response to energetic failure 50 . In our study, AGEs induced mitochondrial dysfunction during VSMC calcification as measured by the expression of key enzymes of Kreb's cycle and by OCR. Mitochondrial dysfunction may be associated with AGEs-induced oxidative stress, leading to mitochondrial membrane potential decline and an impaired mitochondrial respiratory chain 51,52 . However, why AGEs inhibit the glycolysis response are still unclear, although in human umbilical vein endothelial cells (HUVECs), glycolysis also declined after AGEs exposure 51 . PDK4 activation elevated the lactate concentration in VSMC supernatant in our experiments, a possible explanation might be that AGEs suppress glycolysis as a whole, counteracting the effects from HIF-1α and PDK4. Further investigation of the basic mechanisms of the AGEs-induced glucose metabolism shift will be critical for treatment of diabetic complications.
Our study also has some limitations. A calcification medium composed of 0.25 mmol/L L-ascorbic acid and 10 −8 M dexamethasone in addition to β-GP is a conventional calcification medium 53 , and our results suggested that VSMCs treated with conventional calcification medium calcify more easily than those treated with β-GP alone (Supplementary Figure 6). Therefore, the conventional calcification medium is a better choice for further research. In addition, although HIF-1α has been reported to bind the PDK4 promoter in results of a luciferase reporter assay 22 , chromatin immunoprecipitation could better prove the direct interaction.
In summary, this study demonstrates that AGEs enhance vascular calcification through the HIF-1α/PDK4 pathway. In addition, glucose metabolism is suppressed during AGE-induced VSMC calcification. Therefore, inhibition of the AGEs/HIF-1α/PDK4 pathway might be an effective approach for the prevention of diabetic vascular calcification. However, the basic mechanisms of the AGEs-mediated glucose metabolism shift remain to be investigated in depth. Alizarin red S staining. VSMCs were fixed in 4% paraformaldehyde for 30 min at room temperature, washed twice with PBS, and then stained with 1% Alizarin red S (pH 8.4) for 30 min at 37 °C. Then, excess Alizarin red S reagent was removed by washing twice with PBS. The calcium nodules were observed under a microscope.

Materials and Methods
Measurement of calcium content. VSMCs were decalcified with 0.6 M HCl for 24 h at 37 °C, and then, cells were washed three times with PBS and solubilized with 0.1 M NaOH containing 0.1% SDS. The calcium content in VSMCs was measured using the calcium assay kit and normalized to the total protein content with the BCA protein assay kit. ALP activity assay. VSMCs were solubilized with RIPA lysis buffer. After centrifugation, the supernatants were examined by the ALP activity kit and normalized to total protein content with the BCA protein assay kit.
Measurement of lactate production and glucose consumption. The supernatants collected from cultured VSMCs were examined by the lactate assay kit and glucose assay kit according to the manufacturer's instructions.
Immunofluorescence staining. VSMCs were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 for 20 min, and then blocked with 5% BSA for 0.5 h at room temperature. Primary antibody (anti-HIF-1α, 1:200) was incubated with cells overnight at 4 °C, and then, the cells were incubated with appropriate second antibodies for 0.5 h in the dark. Nuclei were stained with DAPI for 15 min. The images were visualized using a confocal microscope (FV10i, Olympus, Japan). Real-time qRT-PCR. Total RNA was isolated using TRIzol according to the manufacturer's instructions.
RNA purity was evaluated based on the A260/A280 ratio using a Merinton SMA4000. Reverse transcription (RT) was performed with Prime Script TM Master Mix (Takara). Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) was performed on a StepOne Plus system (ABI) using SYBR Green Mix. PCR primers are shown in Table 1. Results were normalized to the expression level of β-actin.
Western blot analysis. VSMCs were lysed with RIPA lysis buffer containing protease inhibitors for 30 min at 4 °C. After centrifugation, the supernatants were harvested, and the protein concentration was measured using the BCA protein assay kit. Subsequently, 60 μg of total protein was loaded onto an SDS-PAGE gel and then transferred onto nitrocellulose membranes. The membranes were blocked with 5% non-fat milk for one hour. Subsequently, the membranes were incubated with different primary antibodies (HIF-1α: 1:1000, PDK4: 1:2000, RUNX2: 1:1000, β-actin: 1:2000) overnight at 4 °C and then visualized using anti-rabbit IgG (1:5000) conjugated with horseradish peroxidase for 1 h at room temperature. The blots were detected using ECL, and the results were quantified by Image-Pro Plus 6.0 software and then normalized to β-actin.

Statistical analysis.
All experiments were independently repeated at least three times. All data are presented as the mean ± standard deviation (SD). Statistical analyses were performed using Statistical Package for Social Science (SPSS) 22.0 software (SPSS, Chicago, IL, USA). Data were plotted using GraphPad Prism software (GraphPad Prism 7.0; GraphPad Software Inc., La Jolla, CA). Student's t-test was used to compare two variables, and one-way analysis of variance (ANOVA) was used to compare more than two groups. All statistical tests were two-tailed, and all data followed a normal distribution. Values of P < 0.05 were considered to be statistically significant.