Salidroside Stimulates Mitochondrial Biogenesis and Protects against H2O2-Induced Endothelial Dysfunction

Salidroside (SAL) is an active component of Rhodiola rosea with documented antioxidative properties. The purpose of this study is to explore the mechanism of the protective effect of SAL on hydrogen peroxide- (H2O2-) induced endothelial dysfunction. Pretreatment of the human umbilical vein endothelial cells (HUVECs) with SAL significantly reduced the cytotoxicity brought by H2O2. Functional studies on the rat aortas found that SAL rescued the endothelium-dependent relaxation and reduced superoxide anion (O2∙−) production induced by H2O2. Meanwhile, SAL pretreatment inhibited H2O2-induced nitric oxide (NO) production. The underlying mechanisms involve the inhibition of H2O2-induced activation of endothelial nitric oxide synthase (eNOS), adenosine monophosphate-activated protein kinase (AMPK), and Akt, as well as the redox sensitive transcription factor, NF-kappa B (NF-κB). SAL also increased mitochondrial mass and upregulated the mitochondrial biogenesis factors, peroxisome proliferator-activated receptor gamma-coactivator-1alpha (PGC-1α), and mitochondrial transcription factor A (TFAM) in the endothelial cells. H2O2-induced mitochondrial dysfunction, as demonstrated by reduced mitochondrial membrane potential (Δψm) and ATP production, was rescued by SAL pretreatment. Taken together, these findings implicate that SAL could protect endothelium against H2O2-induced injury via promoting mitochondrial biogenesis and function, thus preventing the overactivation of oxidative stress-related downstream signaling pathways.


Introduction
The role of oxidative stress in the development of the endothelial dysfunction has been studied extensively [1][2][3][4]. Excessive reactive oxygen species (ROS) not only reduces bioavailable nitric oxide (NO) through direct reaction to form peroxynitrite, but also leads to eNOS uncoupling and further induces more ROS production [5].
As a major cellular source of ROS, the contributions of mitochondria to the detrimental effects of cardiovascular risk factors have recently received increased attention [6][7][8]. Excessive mitochondrial ROS (mtROS) act to inspire pathologic cell-signaling cascades under conditions of comprehensive excessive oxidative stress [9][10][11]. NF-kappa B (NF-B) activation occurs secondary to excessive mitochondrial ROS production in the endothelium, participating in a range of proinflammatory and prothrombotic alterations in the endothelial cells [12,13]. Endothelial mitochondria have been found to have crucial roles in vascular path-physiology [14][15][16], and increasing evidences have indicated the importance of mitochondrial dysfunction in various vascular diseases, such as atherosclerosis, heart failure, and cardiac ischemia/reperfusion injury [17][18][19]. Previous studies have shown that dysregulation of mitochondrial biogenesis represents an early manifestation of endothelial dysfunction, shifting cell metabolism toward metabolic hypoxia in animals with impaired NO bioavailability [14]. Impairment of mitochondrial biogenesis is frequently observed in atherosclerosis and is thus likely to contribute to cellular energetic imbalance, oxidative stress, and endothelial dysfunction in these pathological conditions [20]. Since increased mitochondrial production of ROS due to impaired mitochondrial biogenesis also appears to be a key event in the development of aging-related vascular pathologies [13,21,22], identification of mechanisms that promote mitochondrial biogenesis in the endothelial cells may provide new clues on the pathogenesis of vascular disease.
The health of mitochondria is in part regulated by their biogenesis and peroxisome proliferator-activated receptor gamma-coactivator-alpha (PGC-1 ) is regarded as the key regulator [23]. In endothelial cells, PGC-1 also orchestrates cellular defenses against oxidative stress [24]. Mitochondrial transcription factor A (TFAM) is responsible for the transcriptional control of mtDNA and its translocation to the mitochondria is important to initiate mtDNA transcription and replication [25].
As described above, mitochondria are highly dynamic organelles, and their biogenesis is likely to be involved in the regulation of endothelial cell metabolism, redox regulation, and signal transduction [6,16,26]. Pathways that regulate mitochondrial biogenesis are potential therapeutic targets for the amelioration of endothelial dysfunction and vascular diseases [27].
Salidroside (SAL) is an active ingredient of the root of Rhodiola rosea, a well-known herb used to relieve high altitude sickness [28]. SAL has also been used to enhance both the physical and mental performance. SAL upregulates the levels of antioxidative enzymes glutathione peroxidase-1 and thioredoxin-1 to counteract oxidative stress [29]. Previous studies have shown that SAL promotes DNA repair enzyme Parp-1 to counteract oxidative stress [30]. Meanwhile, a recent study reported that SAL attenuated homocysteineinduced endothelial dysfunction by reducing oxidative stress [31]. As noted above, there is tightly relationship between reduction of mitochondrial biogenesis and endothelial dysfunction; the present study was conducted to determine whether SAL recovers the endothelial dysfunction induced by H 2 O 2 through stimulating mitochondrial biogenesis and counteracting the oxidative stress-related eNOS and NF-B signaling pathways.  [32]. Primary cultured human umbilical vein endothelial cells (HUVECs) were prepared as described in [33]. In brief, the umbilical cord was washed with cold PBS and then infused with 0.25% trypsin. After digestion stopped, the cells were collected by centrifuging for 10 min at 1 000 rpm. The cells were resuspended and then cultured in endothelial cell medium (ECM, Sciencell, Carlsbad, CA) at 37 ∘ C in an incubator with a humidified atmosphere of 5% CO 2 . In all experiments, cells were used at passages 2-7.

Cell Viability Assays.
To study the effect of SAL on H 2 O 2induced cytotoxicity, HUVECs were inoculated at a density of 2 × 10 4 per well in 96-well plates and cultured overnight; the cell viability was evaluated using the Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto) [34]. Briefly, the HUVECs treated with H 2 O 2 (Sigma, 4 h) or SAL (National Institute for Food and Drug Control, purity > 98%, 24 h) at the indicated concentration in OPTI-MEM (Gibco) before cell viability was measured. Moreover, to study the effect of SAL on H 2 O 2 -induced inhibition of cell viability, the cells were cultured as described above; after treatment with SAL for 20 h, the H 2 O 2 (100 M) was added to the medium for another 4 hours. At the end of the time period, the culture medium was removed. The cells were washed twice with PBS and incubated with CCK-8 solution at 37 ∘ C for 30 min. The absorbance was measured using a microplate reader with a test wavelength of 450 nm.

Free Radical Measurement in Cell Free
System. The effects of SAL on scavenging hydroxyl radical (OH • ), superoxide radical (O 2 − ), and H 2 O 2 were measured with commercially available kits (Nanjing Jiancheng Bioengineering Institute, China) according to the manufacturer's instructions. In brief, the OH • was generated by the Fenton reaction and then treated with a chromogenic substrate nitrotetrazolium blue chloride (NBT) to yield a stable colored substance, which was measured using a microplate reader with a test wavelength of 450 nm. O 2 − was generated by the xanthine/xanthine oxidase system, O 2 − was detected by nitrite method, and the absorbance at 550 nm was measured. The reaction product of H 2 O 2 and molybdic acid can be detected at 405 nm. Deionized water and ascorbic acid (Vc) were used as the blank and positive controls, respectively. The inhibition rate = (optical density of blank control groups-optical density of treatment groups)/optical density of blank control groups.

Vascular Function Measurement.
The thoracic segments of the rat aorta were dissected and the surrounding connective tissues were cleaned off. Each aorta was cut into ring segments of 2∼3 mm in length. The aortic rings were mounted in organ chambers filled with Krebs-Henseleit (KH) solution at 37 ∘ C with constant bubbling of 95% oxygen/5% carbon dioxide. KH solution contained (mM) 133 NaCl, 4.75 KCl, 1.5 CaCl 2 , 1.25 MgCl 2 , 25 NaHCO 3 , and 11 D-glucose. Isometric tension was recorded with a force transducer (RM6240C, Chengdu Instrument Factory). A basal tension of 20 mN was applied to each vascular ring. After being placed in organ baths for 90 min, 0.5 M phenylephrine (PE, Sigma) was first administered to the rings to test their contractility and then 1 M acetylcholine (ACh, Sigma) was administered to assess the integrity of the endothelial layer.
Rings with less than 80% relaxation response to ACh were discarded. The aortic rings were pretreated with or without 10 M SAL for 30 min before H 2 O 2 100 M was added to the bath. After precontracting with PE (0.1 M), ACh (1 × 10 −8 ∼1 × 10 −4 M) was added cumulatively to the bath to evoke the endothelium-dependent relaxation. The relaxation rate is the ratio between the tension relaxed by ACh and the tension contracted by PE.

NF-B Activity
Assay. An ELISA-based assay was used to measure the NF-B activity, as described previously [37][38][39]. In brief, after stimulation, cells were rinsed twice with cold PBS, and then RIPA containing a protease inhibitor cocktail (Roche, Basel, Switzerland) was added. After incubation on ice for 30 min, the lysate was centrifuged for 15 min at 14 000 rpm and the supernatant was collected. After being quantified with BCA reagent (Pierce, Rockford, IL), the cell extracts were incubated in a 96-well plate coated with the oligonucleotide containing the NF-B consensus-binding site (5 -GGGACTTTCC-3 ). Activated transcription factors from extracts specifically bound to the respective immobilized oligonucleotide. NF-B activity was then detected with the primary antibody to NF-B p65 (1 : 1000, Proteintech, China) and secondary antibody conjugated to horseradish peroxidase (1 : 10 1000, Abbkine, CA). Tetramethylbenzidine (100 L, Sigma) was added in each microwell at 37 ∘ C before adding 100 L of stopping solution (2M H 2 SO 4 ). NF-B activity was finally determined as absorbance values measured with a microplate reader at a wavelength of 450 nm.

Measurement of Mitochondrial Mass.
Mitochondria mass was determined by using MitoTracker green, a mitochondrial-selective membrane potential-independent dye [40]. The cells grown on cover slips coated with 2% gelatin were incubated with SAL (1, 10 M) for 24 h. At the end of the incubation, suspensions were removed and the cells were incubated with 200 nM MitoTracker green (Beyotime Institute of Biotechnology, China) in 37 ∘ C for 30 min. The images were captured with a fluorescence microscope (Olympus FV500) using 40× magnification objective [41]. The integrated fluorescence intensities were measured using the Image-Pro Plus software and normalized to the number of cells.

Measurement of Intracellular ATP Levels.
After treatment, the HUVECs from each well of a 6-well dish (5 × 10 5 /well) were washed twice with cold PBS and lysed with 0.5 M perchloric acid and briefly sonicated (5 to 10 times of a 1 sec burst) until cells were clearly disrupted. Samples were then neutralized with 2 M KOH and then centrifuged to remove the precipitate. ATP content was analyzed by HPLC (Agilent, Palo Alto, CA) with an LC-18T reverse-phase column (Agilent, Palo Alto, CA) at a flow rate of 0.3 mL/min, and the absorbance at 254 nm was recorded. The elution peak was compared with ATP standards (National Institute for Food and Drug Control, Beijing) to confirm its identity.

Assessment of Mitochondrial Membrane Potential (Δ m).
JC-1 is a positively charged fluorescent compound which is taken up by mitochondria proportionally to the inner mitochondrial membrane potential [42]. The ratio of red (J-aggregate)/green (monomeric JC-1) emission is directly proportional to the Δ m. HUVECs were grown on 96well plates treated with SAL (10 M) for 20 h, and then H 2 O 2 (100 M) was added for another 4 h. Cells were rinsed with PBS and incubated in 100 L JC-1 staining solution at 37 ∘ C for 20 min. Cells were then rinsed twice with JC-1 washing solution and analyzed with a fluorescence spectrophotometer. J-aggregates were recorded with an excitation wavelength of 535 nm and an emission wavelength of 610 nm, respectively, and monomeric JC-1 was recorded with an excitation wavelength of 485 nm and an emission wavelength of 535 nm, respectively.
2.11. Western Blot Analysis. Cells were homogenized in icecold RIPA lysis buffer containing protease inhibitor cocktail and phosSTOP (Roche, Basel, Switzerland). Equal amounts of protein (60 g) were mixed with the loading buffer (Beyotime Institute of Biotechnology), boiled for 10 min, and separated by SDS-PAGE. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes (PVDF, Millipore, Temecula, CA). The membranes were blocked for 1 h in 5% milk. The membranes were then incubated overnight at 4 ∘ C with one of the following specific primary antibodies: rabbit anti-eNOS ser1177, anti-eNOS, anti-AMPK thr172, anti-AMPK , anti--actin (1 : 1 000, Cell Signaling Technology, Beverly, MA), anti-Akt ser473 (1 : 1 000, EPITOMICS, CA), anti-Akt (1 : 600, Proteintech), anti-TFAM (1 : 300, Proteintech), and anti-PGC-1 (1 : 200, Santa Cruz, CA). After washing, the membranes were incubated for 2 h at room temperature with secondary antibodies (Goat antirabbit IgG, goat anti-mouse IgG, 1 : 10 000, Abbkine, CA) and then washed. Finally, the blots were developed with enhanced chemiluminescence detection reagents (Thermo Scientific, Waltham, MA). Membranes were scanned using the Micro-Chemi bioimage analyzer (NDR, Israel) and quantified using Image J program and normalized against -actin. 2.12. Statistical Analysis. All data in this study are expressed as the mean ± SEM from at least three separate experiments. SPSS 13.0 was used for statistical analysis. Individual group statistical comparisons were analysed by unpaired Student's -test with Bonferroni correction, and multiple groups comparisons were evaluated by one-way ANOVA with post hoc testing. A probability value of < 0.05 was considered statistically significant.

SAL Alleviates the Cytotoxicity Induced by H 2 O 2 in
HUVECs. In this study, H 2 O 2 was used to induce oxidative stress in HUVECs. After exposure to H 2 O 2 (100-1000 M) for 4 h, HUVECs viability was reduced (Figure 1(a)). SAL at the concentration below 10 M had no obvious effect on cell viability compared to control (Figure 1(b)). Pretreatment with SAL could reduce cell death induced by H 2 O 2 in a concentration-dependent manner (Figure 1(c)).

SAL Inhibited the Activation of Transcription Factor NF-B Induced by H 2 O 2 in HUVECs.
Activation of the transcription factor NF-B has been associated with endothelial cells dysfunction and vascular inflammation in atherogenesis [43]. In line with previous studies, exposure to H 2 O 2 could result in transient activation of NF-B (Figure 5

SAL Induced Mitochondrial Biogenesis in Endothelial
Cells. Our results showed that SAL (10 M) increased the fluorescent intensity of MitoTracker green, suggesting that SAL increased the mitochondrial mass in HUVECs (Figure 6(a)). Consistent with this, the expression of PCG-1 and TFAM, the key regulators of mitochondrial biogenesis were significantly increased in HUVECs incubated with SAL ( Figures  6(b) and 6(c)).

SAL Restores H 2 O 2 -Induced Mitochondrial Dysfunction.
We used independent parameters to evaluate mitochondrial function ATP production and ΔΨm. As shown in Figure 7,

Discussion
There is increasing attention on the relationship between oxidative stress and endothelial cell injury [44][45][46]. The present study employed H 2 O 2 -induced oxidative stress in HUVECs as a cellular model to study the protective effect of SAL. In our experiments, pretreatment of SAL significantly prevents the impaired viability of HUVECs caused by H 2 O 2 exposure. These results are in accordance with previous studies [47] and further confirmed the protective effects of SAL against injury induced by H 2 O 2 . Moreover, the antioxidative mechanism of SAL is not due to the direct reaction between SAL and H 2 O 2 ( Figure 2). Overproduced ROS are known to harm the normal vascular function by limiting the beneficial effects of endothelium derived NO [48]. The enhanced production and release of ROS and/or the diminished bioavailability of NO within vascular wall lead to endothelial dysfunction that is widely Oxidative Medicine and Cellular Longevity believed to be the early key event in the pathogenesis of various vascular complications [4,49]. Although H 2 O 2 elicits relaxation in rat, mouse, and rabbit aortas [50][51][52][53] and stimulates eNOS, resulting in higher NO levels [54], the effect of prolonged elevation of H 2 O 2 is to impair endotheliumdependent relaxation [55]. Previous study showed that SAL prevented homocysteine-induced endothelial dysfunction through curtailing oxidative stress [31]. In our study, pretreatment with H 2 O 2 induced significant impairment of endothelial dependent relaxation, while SAL had the capacity to rescue this impairment. As expected, incubation of HUVECs with H 2 O 2 strikingly increased intracellular O 2 •− , and this can be suppressed by pretreatment with SAL. These results strongly suggest that SAL inhibits H 2 O 2 -induced ROS production, contributing to the restoration of the endotheliumdependent vasorelaxation.
NO, derived from the action of eNOS in endothelial cells, is one of the most important mediators in the regulation of endothelial functions [56]. Phosphorylation of eNOS at ser1177 activates eNOS, while increased oxidative flux directly scavenges NO to lower NO bioavailability [57] and subsequently impairs endothelium-dependent vasodilatation [58]. Interestingly, as shown in many previous studies, H 2 O 2 directly upregulated the levels of NO in endothelial cells, suggesting that overproduction of NO by endothelial cells in response to H 2 O 2 stimulation was intended to protect the cells, rather than damage cells [59,60]. An intriguing question that arises from this study is why SAL tends to inhibit H 2 O 2 -induced eNOS activation and NO production, while it per se appears to upregulate eNOS expression and activation (Figures 4 and 3(c)). We speculated here that the upregulation of eNOS activity and NO production by long-term SAL treatment may be initiated by a transient mild mitochondrial depolarization and reduced oxygen demand, which in turn increase the tolerance of affected cells to subsequent oxidative insult that is greater in severity. But this needs to be further proved.
There are evidences that NO can repress the activation of NF-B through degradation of I B [61] or inhibition of NF-B DNA binding [62]. We also analyzed the effect of SAL on the activation of NF-B induced by H 2 O 2 . Our date showed that SAL decreased the basal activity of NF-B, meanwhile blocking the activation of NF-B induced by H 2 O 2 ( Figure 5). H 2 O 2 at the concentration of 100 M slightly reduced rather than increased the NF-B activity; meanwhile, pretreatment with SAL further reduced the activity. The possible explanation is that H 2 O 2 per se induces the formation of active dimmers of P65 and P50 subunits and leads to their translocation to the nucleus. However, within the nucleus, the Cys 62 residue on the P50 subunit is oxidized to sulfenic acid and is further followed by Sglutathionylation, which inhibits the binding of NF-B to the DNA [63][64][65]. To strengthen this hypothesis, we detected the direct effect of H 2 O 2 on DNA binding activity. We used TNF-, a classical NF-B inducer to treat HUVECs. After cell lysis and total protein extraction, we incubated the proteins with increasing concentrations of H 2 O 2 (0.1, 1, 10, and 100 M) for 30 min, and then the activities of NF-B were detected as described in Section 2. The results indicated that H 2 O 2 dose dependently inhibited the DNA binding activity of activated NF-B. Since NF-B is a redox sensitive transcription factor, its activation participates in inflammation and mitochondrial biogenesis impairment [66]. We speculated that the protective effect of SAL on HUVECs might be related to an interference with the NF-B signaling pathway.
Moderate increases in NO stimulate mitochondrial biogenesis, mainly through cGMP-dependent gene expression and activation of regulatory factors including PGC-1 and TFAM [67,68]. In cardiomyocytes, coimmunoprecipitation experiments demonstrated that the p65 subunit of NF-B is constitutively bound to PGC-1 coactivator and blocks its activation of gene transcription and that NF-B activation increases this binding [69]. Moreover, PGC-1 overexpression inhibits NF-B activation in human aortic smooth muscular and endothelial cells [70]. So it is reasonable that the agents stimulating PGC-1 expression and mitochondrial biogenesis in the endothelial cells are beneficial to prevent the development of cardiovascular disease. Here we reported for the first time that SAL increased mitochondrial mass in HUVECs (Figure 6(a)). Inducible mitochondrial biogenesis is very important in vascular health [27,71]. Moreover, mitochondrial proliferation reduces the flow of electron per unit mitochondria; SAL-induced mitochondrial biogenesis may contribute to the reduction of mitochondrial ROS production in HUVECs. To determine whether the effect of SAL on mitochondrial biogenesis is a consequence of the activation of mitochondrial biogenesis regulatory factors, we examined the expression of PGC-1 and TFAM. We found that SAL increased the expression of PGC-1 and TFAM (Figures 6(b) and 6(c)).
In addition to stimulating mitochondrial biogenesis, PGC-1 also contributes to the induction of ROS detoxifying enzymes, including catalase, superoxide dismutase, and heme oxygenase [72][73][74][75]. These findings imply that PGC-1mediated mitochondrial biogenesis in oxidative injured cells seems to offer a good source of "healthy mitochondria" which detoxify mtROS by a large antioxidant defense system containing numerous redox enzymes of the electron-transport chain, ultimately decreasing net ROS production. These effects of SAL on mitochondrial biogenesis may partially explain its antioxidant properties.
Many evidences have revealed that H 2 O 2 caused endothelial cell injury by inducing mitochondrial dysfunction [76,77]. Due to localization to the inner mitochondrial membrane, lack of histone-like coverage and a less efficient DNA repair system compared with nuclear DNA, mtDNA is prone to oxidative stress [78]. Furthermore, the mutation is more likely to affect gene integrity because of the absence of intron in the mitochondrial genome [79]. Damaged mitochondria produce less ATP but more greater amounts of ROS, potentiating the signal and entering a vicious circle, which aggravate cardiovascular diseases [80]. As shown in Figure 7, SAL pretreatment abrogated the H 2 O 2 -induced collapse of ΔΨm and rescued the mitochondrial function, proved by increased ATP production. These effects of SAL decrease the potential for mtDNA damaged by H 2 O 2 ; moreover, SAL enhanced mitochondrial biogenesis, which provides healthy mitochondria to replace the mitochondrial components damaged by ROS and maintain normal mitochondrial function. The protective effect of SAL on mitochondria is of paramount importance to maintain the endothelial homeostasis.
In summary, the present study demonstrated a novel mechanism of SAL to protect the endothelial cells from oxidative damage. By stimulating mitochondrial biogenesis ( Figure 8) and counteracting the reactive oxygen species burst within mitochondria, SAL administration prevented the overactivation of several signaling pathways evoked by damaging oxidative stimuli and preserved the viability of endothelial cells, as well as the endothelial dependent vessel functions. Novel compounds with mitochondria biogenesis stimulating activities may become potential drug candidates for the prevention or treatment of such disorders associated with oxidative stress as metabolic syndrome or cancer and so forth.