Salidroside-Mediated Autophagic Targeting of Active Src and Caveolin-1 Suppresses Low-Density Lipoprotein Transcytosis across Endothelial Cells

Subendothelial retention of apolipoprotein B100-containing lipoprotein, such as low-density lipoprotein (LDL), is the initial step of atherogenesis. Activation of autophagy exhibits beneficial effects for the treatment of atherosclerosis. In our previous study, we demonstrated that hyperglycemia suppressed autophagic degradation of caveolin-1, which in turn resulted in acceleration of caveolae-mediated LDL transcytosis across endothelial cells and lipid retention. Therefore, targeting the crossed pathway in autophagy activation and LDL transcytosis interruption may be a promising antiatherosclerotic strategy. In metabolic diseases, including atherosclerosis, salidroside, a phenylpropanoid glycoside compound (3,5-dimethoxyphenyl) methyl-β-glucopyranoside), is the most important compound responsible for the therapeutic activities of Rhodiola. However, whether salidroside suppresses LDL transcytosis to alleviate atherosclerosis has not yet been elucidated. In the present study, we demonstrated that salidroside significantly decreased LDL transcytosis across endothelial cells. Salidroside-induced effects were dramatically blocked by AMPK (adenosine monophosphate-activated protein kinase) inhibitor (compound c, AMPKα siRNA) and by overexpression of exogenous tyrosine-phosphorylated caveolin-1 using transfected cells with phosphomimicking caveolin-1 on tyrosine 14 mutant plasmids (Y14D). Furthermore, we observed that salidroside promoted autophagosome formation via activating AMPK. Meanwhile, the interaction between caveolin-1 and LC3B-II, as well as the interaction between active Src (indicated by the phosphorylation of Src on tyrosine 416) and LC3B-II, was significantly increased, upon stimulation with salidroside. In addition, both bafilomycin A1 (a lysosome inhibitor) and an AMPK inhibitor (compound c) markedly prevented salidroside-induced autophagic degradation of p-Src and caveolin-1. Moreover, the phosphorylation of caveolin-1 on tyrosine 14 was disrupted due to the downregulation of p-Src and caveolin-1, thereby directly decreasing LDL transcytosis by attenuating the number of caveolae on the cell membrane and by preventing caveolae-mediated LDL endocytosis released from the cell membrane. In ApoE−/− mice, salidroside significantly delayed the formation of atherosclerotic lesions. Meanwhile, a significant increase in LC3B, accompanied by attenuated accumulation of the autophagy substrate SQSTM1, was observed in aortic endothelium of ApoE−/− mice. Taken together, our findings demonstrated that salidroside protected against atherosclerosis by inhibiting LDL transcytosis through enhancing the autophagic degradation of active Src and caveolin-1.


Introduction
Atherosclerotic cardio-or cerebrovascular diseases are common causes of morbidity and mortality worldwide [1].
Subendothelial retention of apolipoprotein B-(APOB/a-poB100-) containing lipoproteins, such as low-density lipoprotein (LDL), is the initial step of atherogenesis [2][3][4]. The gap between vascular endothelial cells is roughly 3-6 nm in diameter, which only allows water and inorganic salts, and several small molecules, to pass through. However, the diameter of LDL is 20-50 nm; thus, the only way for LDL particles to traffic across the intact endothelial barrier is through a transporting process called transcytosis [3]. In endothelial cells, LDL transcytosis is predominantly mediated by caveolae, which are specialized lipid rafts that form 50-100 nm flask-shaped invaginations in the plasma membrane [5][6][7].
Caveolin-1 and cavin-1 are two essential and structural components of caveolae and represent small invaginations of the plasma membrane that form lipid vesicles [7]. Caveolin-1 was initially identified as a substrate for c-src tyrosine kinase, which phosphorylates caveolin-1 on tyrosine 14 [8,9]. Tyrosine-phosphorylated caveolin-1 (p-caveolin-1) can drive caveolae reconformation and subsequent internalization from the cell membrane [10]. Furthermore, tyrosinephosphorylated caveolin-1 increased the number of caveolae on the cell membrane by promoting the expression of caveolin-1 and cavin-1 via transcriptional regulation of early growth response-1 [11].
Macroautophagy/autophagy is a homeostatic process that occurs in all eukaryotic cells and involves sequestration of cytoplasmic components in double-membraned autophagosomes that subsequently fuse with lysosomes in which their cargo is delivered for degradation and recycling purposes [12].
Autophagy acts as a therapeutic target for preventing and ameliorating atherosclerosis via numerous pathways to protect cells against oxidative stress, inflammation, and apoptosis [13][14][15][16]. In our previous study, we demonstrated that high glucose suppressed autophagic targeting of caveolin-1. Therefore, more caveolin-1 was accumulated in the cytosol and utilized to increase the caveolae on the cell membrane to facilitate the transcytosis of LDL across endothelial cells [17].
Salidroside, a phenylpropanoid glycoside compound, has been shown to alleviate metabolic diseases, such as atherosclerosis and diabetes, by modulating various synergistic pathways that control oxidative stress, inflammation, mitochondria, autophagy, and cell death, as well as AMPK signaling [18]. In our previous study, we demonstrated that salidroside effectively alleviated the progression of insulin resistance and atherosclerosis by activating AMPK to suppress reactive oxygen species generation and inflammasome activation [19][20][21].
In the present study, we demonstrated that treatment with salidroside enhanced the autophagic degradation of active Src and caveolin-1 by activating AMPK. Subsequently, the expression of caveolin-1 and p-caveolin-1 was decreased, thereby resulting in inhibition of LDL transcytosis across endothelial cells to reduce lipid accumulation in the vascular wall.
2.3. LDL Labeling. LDL and FITC were mixed at a ratio of 6 mg LDL : 1 mmol FITC and incubated at 37°C for 2 h. The mixture was dialyzed against PBS for 72 h at 4°C to remove unbound FITC.

Flow
Cytometry of FITC-LDL Uptake Analyses. HUVECs were seeded in 12-well plates (Roche Diagnostics Corporation, 3513) and were incubated with serum-free Opti-MEM (Gibco, 31985-070), containing FITC-LDL (50 μg/mL) for 3 h. Cells were harvested using 0.125% trypsin (without EDTA), and LDL uptake was measured by flow cytometry (Mindry, Bricyte E6). HUVECs treated with naïve LDL were used as a negative control. To obtain the real fluorescence due to FITC-LDL uptake, the background fluorescence was subtracted from the mean FITC-LDL fluorescent intensity of each sample, FSC/FSS scatter diagrams were plotted, and 2000 events were recorded.

2
Oxidative Medicine and Cellular Longevity 2.6. LDL Transcytosis. The amount of LDL transcytosis was measured according to previous reports [23]. In brief, HUVECs were seeded on polyester transwell membranes (Costar, 3470). The integrity of the cell monolayer was evaluated as previously described [24]. Two inserts containing cell monolayers with equal integrity were assigned to the same group, and two different inserts were considered: the noncompetitive insert and the competitive insert. The competitive insert was treated with FITC-LDL (50 μg·mL −1 ) and sixfold excess of unlabeled LDL to determine paracellular transport, whereas the noncompetitive insert was treated with FITC-LDL (50 μg·mL −1 ) alone to determine the total transport of transendothelial LDL. Samples were collected from the lower chambers and further dialyzed against PBS (Thermo Fisher Scientific,10010031) to remove free FITC. Samples were embedded in EMbed-812 (EMS, 14120), and after polymerization of the resin at 60°C for 36 h, serial sections were cut using an ultramicrotome (Leica, Germany) and mounted on formvar-coated slot grids (EMS, GA300-Cu). Sections were stained with 4% uranyl acetate and lead citrate and examined under a Tecnai G2 F20 S-TWIN transmission electron microscope (FEI, American).
2.11. Confocal Imaging Analyses of LDL Uptake. Cells were incubated with DiI-LDL (50 μg/mL) for 24 h and then treated as indicated. Images were obtained with a confocal laser scanning microscope (FV3000; Olympus) using a 40x objective. Fluorescence images were analyzed using ImageJ software (Santa Clara, CA, USA). Individual microscopic fields were randomly selected to include at least 15 cells, and the numbers of cells were counted. The fluorescence intensities were normalized to the number of cells.
2.12. Experiments in ApoE -/-Mice. Animals were treated in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health and approved by the local animal care committee. All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals [25]. Every effort was made to minimize animal suffering and reduce the number of animals used. 12-week-old male ApoE −/− mice were purchased from HFK (Beijing, HFK Bioscience Co., Ltd., Beijing). All animals were maintained in a controlled environment with a light/ dark cycle of 12 h, a temperature of 20 ± 2°C, and a humidity of 50 ± 2%. After giving them a week to adapt to this environment, mice were fed a standard diet for 2 weeks and then followed by a "Western diet" (21% fat, 0.15% cholesterol) for 8 weeks. ApoE -/mice (n = 7) were randomly assigned into the vehicle group and the salidroside group. The vehicle group received p.o. with 0.9% saline. The salidroside group received p.o. with salidroside (purity > 98%, National Institute for Food and Drug Control, Beijing, China; 50 mg·kg −1 ·day −1 ; n = 7) for 8 weeks and then killed under anesthesia. The ascending aorta was collected for Oil Red O staining. Serial cross-sections (8 μm) of the heart throughout the entire aortic valve area were cut in a cryostat (Leica CM1900), and the atherosclerotic lesions were stained with Oil Red O. Plaque size was quantified using the ImageJ pro plus software as described previously [26,27]. The sections stained with LC3B (1 : 50; Cell Signaling Technology, 3868) or SQSTM1 (1 : 50; Cell Signaling Technology, 39749) were examined under light microscopy at a magnification of ×400 with a semiquantitative scoring system (0 to 4) by a method described previously [28].
2.13. Statistical Analyses. Data are expressed as the mean ± SEM from at least three independent experiments. Individual group statistical comparisons were analyzed by an unpaired Student t-test, whereas multiple-group comparisons were evaluated by one-way ANOVA with post hoc testing. p < 0:05 was considered statistically significant.

Decreased LDL Uptake in Salidroside-Treated HUVECs.
LDL uptake by endothelial cells is an intermediate step of LDL transcytosis [29]. Therefore, the intracellular concentration of LDL in HUVECs reflects the activity of LDL transcytosis. In the present study, confocal imaging was used to determine the uptake of DiI-labeled LDL. Meanwhile, flow cytometry analyses were performed to evaluate FITClabeled LDL uptake in HUVECs in which the fluorescent intensity of individual cells reflected the extent of LDL uptake. Figure h)). However, Y14F plasmid transfection did not significantly influence salidrosideinhibited LDL uptake (Figures 2(e)-2(h)). Taken together, treatment with salidroside attenuated LDL uptake by activating AMPK and by suppressing caveolin-1 phosphorylation.

Salidroside Activates AMPK and Increases Autophagic
Degradation of p-Src and Caveolin-1 in a Dose-Dependent Manner. The AMPK signaling pathway is a pathway to induce formation of the phagophore, a crescent-shaped double membrane that expands and fuses to form a doublemembrane vesicle known as the autophagosome [30][31][32]. Therefore, we next evaluated the treatment effect of salidroside on AMPK activation. As depicted in Figures 3(a) and 3(b), AMPK phosphorylation was significantly upregulated by salidroside treatment in a concentration-dependent manner (0.1 μM, 1 μM, and 10 μM),which was consistent with the findings presented in our previous study in which we showed that salidroside activated AMPK by increasing the ratio of AMP : ATP [19][20][21]33]. In general, the covalent conjugation of a soluble form of LC3B (LC3B-I) with phosphatidylethanolamine to form a nonsoluble form (LC3B-II) is a major hallmark of autophagy [34]. We monitored changes in autophagy by analyzing the abundance of LC3B-II and the classical autophagic substrate SQSTM1 (sequestosome 1) as well as c Casitas B-cell lymphoma (c-Cbl) [35]. As shown in Figures 3(a) and 3(b), salidroside treatment resulted in a significant increase in LC3B-II and attenuated accumulation of the autophagy substrates SQSTM1 and c-Cbl. Moreover, active Src (as represented by p-Src) has been shown to be a target for the autophagosome, which was subsequently delivered to the lysosome for degradation through interaction with c-Cbl [36]. As illustrated in Figures 3(a) and 3(b), accumulation of p-Src was decreased by salidroside treatment, which was consistent with a downregulated expression of c-Cbl. In our previous study, we demonstrated that caveolin-1 was recruited to autophagosome for autolysosome degradation by direct interaction with LC3B-II [17]. Moreover, downregulation of caveolin-1 may    Oxidative Medicine and Cellular Longevity lead to a loss in cavin-1 stability [37][38][39]. As depicted in Figures 3(a) and 3(b), upon salidroside treatment, protein levels of the caveolae structure proteins (caveolin-1 and cavin-1) were inhibited, which was accompanied by a disrupted phosphorylation of caveolin-1. Additionally, salidroside treatment resulted in a significant increase in LC3B-II and attenuated accumulation of the autophagy substrates (SQSTM1, c-Cbl, p-Src, and caveolin-1) in a time-dependent manner, which was consistent with a downregulated expression of cavin-1 (Figures 3(c) and 3(d)). Taken together, treatment with salidroside may activate autophagy and thereby induce autophagic degradation of p-Src and caveolin-1.

Bafilomycin A 1 Blocked Salidroside-Induced Autophagic
Degradation of p-Src and Caveolin-1. Our data using trans-mission electron microscopy showed an increase in autophagic vacuoles in the cytoplasm of HUVECs (Figure 4(a)). The increase of autophagosomes in cells is an intermediate process within the autophagic flux and reflects a balance between the rate of formation and degradation [40]. Thus, the increase in autophagosomes in salidroside-treated cells may be explained as follows: (1) salidroside increased the formation of autophagosomes, (2) salidroside prevented the fusion of lysosomes and autophagosomes, or (3) salidroside inhibited lysosome activity. To test these possibilities, bafilomycin A 1 , a lysosome inhibitor that can prevent fusion of lysosomes and autophagosomes and the degradation activity of lysosomes, was used. LC3B that was tagged at its N terminus using a fluorescent protein, such as GFP (GFP-LC3B), has previously been used to monitor autophagy through direct fluorescence microscopy that was measured as an increase      Oxidative Medicine and Cellular Longevity in punctate LC3B [40]. Therefore, overexpression of HUVECs with GFP-LC3B was performed by transfecting HUVECs with GFP-LC3B constructs. As shown in Figures 4(b) and 4(c), both salidroside and bafilomycin A 1 significantly enhanced the appearance of LC3B punctate. Moreover, simultaneous treatment with salidroside and bafilomycin A 1 further increased the number of LC3B punctate when compared with bafilomycin A 1 treatment alone (Figures 4(b) and 4(c)). Subsequently, a tandem fluorescent-tagged GFP-RFP-LC3B construct was used to assess the roles of salidroside in autophagic flux. The GFP of this tandem autophagosome reporter is sensitive to pH and quenched in the acidic environment of the lysosome, whereas the RFP is resistant. Therefore, the fusion of autophagosomes with lysosomes results in the loss of yellow puncta and the appearance of red-only puncta. As shown in Figures 4(d) and 4(e), salidroside increased the numbers of both yellow puncta and red-only puncta. In addition, coincubation of salidroside with bafilomycin A 1 significantly enhanced the abundance of LC3B-II; however, no significant effect was observed on the amount of SQSTM1 when compared to bafilomycin A 1 treatment alone (Figures 4(f) and  4(g)). Interestingly, coincubation of salidroside and bafilomycin A 1 restored accumulation of the autophagic substrate (c-Cbl, p-Src, SQSTM1, and caveolin-1), when compared to salidroside treatment alone (Figures 4(f) and 4(g)). Taken together, these findings demonstrated that salidroside disrupted the autophagic accumulation of p-Src and caveolin-1 through facilitating the formation of autophagosomes in endothelial cells.

12
Oxidative Medicine and Cellular Longevity

Salidroside Increases Src-LC3B and Caveolin-1-LC3B
Complexes. During autophagy, autophagic cargo is recruited to the autophagosome by interacting with the autophagosomal marker protein LC3B [35]. Interaction with the Src-LC3B complex is mediated by c-Cbl. Therefore, another pathway to demonstrate that salidroside promoted the autophagic degradation of p-Src and caveolin-1 was to investigate the interaction between p-Src, caveolin-1, and LC3B by coimmunoprecipitation experiments. As shown in Figures 5(a) and 5(b), the association between p-Src and c-Cbl and LC3B-II in HUVECs was significantly increased by salidroside treatment. Furthermore, salidroside treatment significantly increased the binding of caveolin-1 and LC3B-II (Figures 5(c) and 5(d)). Thus, these findings further demonstrated that salidroside accelerated the autophagic degradation of p-Src and caveolin-1.
3.6. Compound C, an AMPK Inhibitor, Prevented Salidroside-Induced Autophagic Degradation of p-Src and Caveolin-1. In this study, we also investigated whether salidroside induced the autophagic degradation of p-Src and caveolin-1 by activating AMPK. As presented in Figures 6(a) and 6(b), coincubation of salidroside and compound c attenuated the level of p-AMPK, when compared to salidroside treatment alone. In addition, the abundance of LC3B-II as induced by salidroside treatment was also decreased by administration of compound c. Conversely, simultaneous treatment of salidroside and compound c restored salidroside-induced autophagic degradation of SQSTM1, c-Cbl, p-Src, and caveolin-1 (Figures 6(a) and 6(b)). As a result, the restored expression of caveolin-1 enhanced the stability of cavin-1 (Figures 6(a) and 6(b)). Collectively, salidroside treatment promoted the autophagic degradation of p-Src and caveolin-1 by activating AMPK.

Salidroside Alleviates Atherosclerotic Lesion Formation in
ApoE -/-Mice. Compared with the vehicle group, the relative surface area of the atherosclerotic lesion was markedly smaller in the salidroside (50 mg/kg) group (Figures 7(a)  and 7(b). Furthermore, salidroside significantly delayed the

Discussion
The retention of LDL particles in the subendothelial space following LDL transcytosis across endothelial cells initiates the development of atherosclerosis [5,23,41]. In previous studies, it has been shown that autophagy is an emerging therapeutic target for preventing the development and progression of atherosclerosis [13,14,16,42]. Thus, targeting the crossed pathway to activate autophagy and interrupt LDL transcytosis may be a promising antiatherosclerotic strategy. In the present study, we demonstrated that salidroside treatment decreased LDL transcytosis across endothelial cells by inducing the autophagic degradation of active Src and caveolin-1. Transcytosis of LDL particles across endothelial cells is predominantly mediated by caveolae, which are presented at a high frequency in endothelial cells (~10,000 per cells) [43]. In the process of LDL transcytosis across endothelial cells, LDL particles are first to be endocytosed by caveolae at the lumen side of the endothelial cell membrane, subsequently trafficked to the basolateral side, then exocytosed to the subendothelial space [44]. Overexpression of cavin-1 and caveolin-1 increased the number of caveolae in the cell

16
Oxidative Medicine and Cellular Longevity membrane, which upregulated LDL transcytosis [41]. The phosphorylation of caveolin-1 on tyrosine 14 triggers caveolae internalization from the plasma membrane and thereby accelerates caveolae-mediated LDL endocytosis [10]. Furthermore, phosphorylation of caveolin-1 on tyrosine 14 induced caveolae formation by enhancing caveolin-1 and cavin-1 transcription [11]. Activating AMPK induced formation of the phagophore or isolation membrane, a crescent-shaped double membrane that expands and fuses to form a double-membrane vesicle, the autophagosome [14]. In our previous study, we demonstrated that salidroside treatment moderately depolarized Δ ψm and increased the cellular AMP : ATP ratio, which ultimately activated AMPK [19,20].
In an established in vitro model of LDL transcytosis across the endothelial cell monolayer, we first documented that salidroside interfered with LDL transcytosis that was restored by the AMPK inhibitor, compound c and AMPKα siRNA. Overexpression of exogenous p-caveolin-1 in salidroside-treated HUVECs by the transfection of the cells with a Y14D construct (phosphomimicking caveolin-1 on tyrosine 14) significantly reversed salidroside-suppressed LDL transcytosis.
An intermediate event of LDL transcytosis across endothelial cells is that LDL particles are endocytosed into cells but still need to be excreted out the other side of the cells. Therefore, the concentration of LDL within the cells reflects the amount of LDL transcytosis. Indeed, observations in these two sets of experiments were consistent with the data obtained in the in vitro transcytosis model as mentioned above. The transfection of AMPKα siRNA, ATG5 siRNA, and Y14D constructs prevented the downregulation of LDL uptake upon salidroside treatment. Thus, these data provided compelling evidence, showing that salidroside inhibited LDL transcytosis.
Next, the treatment effect of salidroside on autophagy was elucidated. In our study, we determined that treatment with salidroside enhanced the phosphorylation of AMPK on Thr172 and increased the abundance of LC3B-II in a dose-dependent manner as well as in a time-dependent manner. Conversely, accumulation of autophagic substrates, including SQSTM1, c-Cbl, p-Src, and caveolin-1, was significantly disrupted by salidroside treatment. As a result, pcaveolin-1 expression was attenuated. The lost stability of caveolin-1 may lead to increased degradation of cavin-1. This phenomenon was also observed in our study.
Furthermore, transmission electron microscopy was performed to further ascertain that salidroside treatment activated autophagy. We found that treatment with salidroside significantly increased autophagic vacuoles in the cytoplasm. Moreover, bafilomycin A 1 , a type of drug that is used to prevent the fusion of lysosomes and autophagosomes and the degradation activity of lysosomes [34], further increased the number of LC3B punctate in salidrosidetreated cells. Furthermore, the autophagic degradation of caveolin-1, p-Src, c-Cbl, and SQSTM1 was blocked by bafilomycin A 1 . Taken together, these results strongly suggested that salidroside activated autophagy.
In our previous study, we demonstrated that caveolin-1 was recruited to the autophagosome for autophagic degradation by directly interacting with LC3B-II [17]. Interestingly, active Src (indicated by autophosphorylation of Src on Tyr 416), which induced the phosphorylation of caveolin-1 on tyrosine 14, was also found to be a target of autophagosome for degradation by their association with autophagy cargo, c-Cbl [36]. Therefore, we next examined whether treatment with salidroside enhanced the interaction between c-Cbl, p-Src, and LC3B. We observed that salidroside strongly enhanced the association of c-Cbl, p-Src, and LC3B. In addition, binding of caveolin-1 and LC3B was also significantly increased. Thus, these data indicated that treatment with salidroside stimulated the autophagic degradation of p-Src and caveolin-1.
Activation of AMPK directly initiates the formation of autophagosomes. Moreover, the expression of p-AMPK, indicating the activation of AMPK, is increased upon salidroside stimulation. To further elucidate the role of AMPK in salidroside-induced autophagic degradation of p-Src and caveolin-1, in this study, compound c, which suppresses AMPK activation, was used. Our data showed that administration of compound c markedly attenuated the activation of salidroside-stimulated AMPK. Moreover, the increased amount of LC3B-IIas facilitated by salidroside, an autophagic marker, was also decreased by treatment with compound c. In addition, compound c restored the accumulation of SQSTM1, c-Cbl, p-Src, and caveolin-1 in salidroside-treated cells, which was consistent with the results of bafilomycin Combined, these results strongly demonstrated that salidroside promoted autophagic degradation of p-Src and caveolin-1. Consistent with our previous study [19], salidroside treatment significantly alleviated the development of atherosclerosis in ApoE -/mice. Meanwhile, consistent with the above in vitro findings, the expression of LC3B was increased, while the accumulation of SQSTM1 was attenuated in endothelium of atherosclerotic lesions. These observations further supported the in vivo significance of salidroside as strategies for the prevention or treatment of atherosclerosis.
In conclusion (Figure 8), upon salidroside stimulation, AMPK was activated, and subsequent formation of the autophagosome was promoted. As a result, increased amounts of p-Src and caveolin-1 are recruited to the autophagosome through interaction with LC3B-II, which in turn enhanced the autophagic degradation of p-Src and caveolin-1. Reduced accumulation of p-Src and caveolin-1 disrupted caveolin-1 phosphorylation on tyrosine 14, which ultimately inhibited caveolae-mediated LDL transcytosis across endothelial cells by decreasing the number of caveolae and by suppressing caveolae-mediated LDL endocytosis.