Alignment of vascular endothelial cells (ECs) in the direction of the flow is considered a key factor in maintaining endothelial integrity in an active hemodynamic environment. Our recent studies showed that exposure to oxidized LDL (oxLDL), one of the major proatherogenic lipoproteins, significantly increases the stiffness of human aortic ECs, suggesting that oxLDL may have a significant impact on the sensitivity of ECs to mechanical stimuli. In this study, we show that oxLDL strongly enhances the ability of ECs to realign in the direction of the flow and facilitates the formation of F-actin stress fibers under static and flow conditions. The impact of oxLDL on the flow-induced realignment is observed on whole cell and single-fiber levels. We also show that, similar to the effect of oxLDL on endothelial stiffness, the impact of oxLDL on flow-induced realignment can be simulated by methyl-β-cyclodextrin-induced cholesterol depletion, supporting the hypothesis that oxLDL acts as cholesterol acceptor, rather than cholesterol donor, for ECs. Finally, we propose that oxLDL/cholesterol depletion-induced sensitization of ECs to flow may be a result of an increase in cellular stiffness and a respective increase in membrane-cytoskeleton tension.
- lipid rafts
hemodynamic forces generated by blood flow are known to play prominent roles in the acute control of vascular tone, regulation of arterial structure, and localization of atherosclerotic lesions (12, 19, 31). The primary tissue that is affected by the hemodynamic environment is vascular endothelium, a single-cell layer that constitutes the inner lining of the blood vessels and exists on the blood-vascular wall interface. Multiple studies have shown that endothelial cells (ECs) respond to flow by a variety of mechanisms, including cytoskeleton rearrangement, activation of multiple signaling pathways, and changes in gene expression (for review see Refs. 9–11, 24, 25, 34). Although the exact mechanisms responsible for conversion of mechanical stimuli to the biological responses are not fully understood, it is generally accepted that the cytoskeleton provides the structural framework for transmission of mechanical signals throughout the cell. Flow-induced changes of endothelial morphology, such as realignment in the direction of flow and cell elongation, are suggested to be important for maintaining endothelial integrity in active flow environments (9, 20, 25).
Our recent studies showed that endothelial mechanical properties are strongly affected by exposure of cells to the oxidized form of low-density lipoproteins (oxLDL) (6). More specifically, exposure to oxLDL, but not the unoxidized form of LDL, resulted in a significant decrease in the deformability/increase in stiffness of aortic ECs, as measured by micropipette aspiration (microaspiration), as well as an increase in the ability of ECs to generate force on the cell-substrate interface (6). Depletion of cellular cholesterol had the same effect (6) and also increased membrane-cytoskeleton adhesion (37). Multiple studies have shown that deformability of the cellular envelope, a two-component system consisting of a membrane lipid bilayer and the underlying submembrane cytoskeleton, depends more on the properties of the cytoskeleton than on the physical properties of the lipid bilayer (29, 32, 35). Consistent with these studies, we previously showed that an increase in endothelial stiffness critically depends on the integrity of F-actin (4). In the present study, we show that an oxLDL-induced increase in endothelial stiffness is accompanied by an increased ability of ECs to realign in the direction of the flow. Similar to our earlier studies, methyl-β-cyclodextrin (MβCD)-induced depletion of cellular cholesterol has the same effects as exposure to oxLDL. We propose that an oxLDL-induced increase in endothelial stiffness may sensitize the cells to a mechanical stimulus generated by shear stress.
Cells and reagents.
Human aortic ECs (HAECs) between passages 3 and 6 were grown in EC growth medium (EGM-2, Clonetics, San Diego, CA) supplemented with 2% FBS. Bovine aortic ECs (BAECs) between passages 5 and 15 were grown in DMEM (Cell Grow, Washington, DC) supplemented with 10% FBS. Cell cultures were maintained in a humidified incubator at 37°C with 5% CO2. Cells were fed and split every 3–4 days. oxLDL (Biomedical Technologies, Stoughton, MA) was dissolved in EGM-2 supplemented with 0.2% FBS. MβCD (Sigma Chemical, St. Louis, MO) was also dissolved in 0.2% FBS-EGM-2.
Microaspiration of substrate-attached ECs was performed as described in our earlier study (4). Briefly, the membranes were visualized with the fluorescent membrane dye carbocyanide DiIC18 (Molecular Probes) and then aspirated using micropipettes (3–5 μm OD) pulled from borosilicate glass capillaries (SG10 glass, Richland Glass, Richland, NJ). Negative pressure was applied to a pipette by a pneumatic transducer tester (BioTek Instruments, Winooski, VT).
HAECs or BAECs were seeded at a density of 1.7 × 105/cm2 on gelatin- or collagen IV-coated slides and placed in a parallel-plate chamber connected to a recirculating flow circuit composed of a variable-speed peristaltic pump, a fluid capacitor that damped pulsation, and a reservoir with culture medium. The flow chamber consisted of a Teflon upper plate and a stainless steel bottom plate held together by eight screws. A medical-grade silicon gasket was used to seal the chamber and prevent fluid leakage. A precisely machined recess (1 × 30 × 120 mm) on the top plate defined the flow path in the chamber. The top plate also provided inlet and outlet ports and a quartz window for light transmission and sample visualization. Alternatively, in some experiments, cells were mounted onto disposable parallel-plate flow chambers with a 0.4 × 3.8 × 17 mm flow chamber (μ-slides VI, Ibidi, Integrated Biodiagnostic, Munich, Germany). No difference was observed between the experiments performed using the two types of flow chambers. In both cases, the temperature of the fluid in the reservoir is maintained constant at 37°C by immersion of the reservoir into a water bath. The level of CO2 is maintained by bubbling the gas into the fluid. The wall shear stress (τ) is calculated as follows: τ = 6μQ̇/wh2 (2), where μ is fluid viscosity (0.007 g·cm−1·s−1 for PBS), Q̇ is flow rate (ml/s), w is width of the chamber (cm), and h is height of the chamber (cm).
Imaging and analysis of cell alignment.
Images were acquired using an inverted microscope (Eclipse TE2000-U, Nikon) with a CoolSnap camera (Photometrix). Cell alignment in the direction of flow was determined by measurement of the angle between the long axis of the cell determined visually and the direction of flow using MetaVue 6.2R6 software (Molecular Devices). Cells that are aligned perfectly in the direction of the flow have an angle of 0°, and cells that are perpendicular to the direction of the flow have an angle of 90°. The full range of angles varies between −90° and 90°. When cells are not aligned, the distribution is expected to be random, whereas flow-induced alignment is expected to narrow the distribution, generating a peak coinciding with the direction of flow. Cell elongation was estimated as the ratio of cell length to cell width. Since most subconfluent cells have a clear elongated shape, the long axis was determined visually. Statistical analysis for narrowing orientation distributions was performed using a standard F-test to compare the variances of the two data sets and a standard two-sample Student's t-test, with the assumption that variances of the two data sets for cell elongation ratios are unequal. Statistical significance was set at 5% (P < 0.05).
F-actin staining was performed using a standard procedure. Briefly, cells were fixed in 4% paraformaldehyde for 20 min, permeablized with 0.1% Triton X-100 for 5 min, and incubated with 0.1 μM rhodamine-phalloidin (Sigma Chemical) for 40 min. The images were acquired using a three-dimensional (3-D) deconvolution microscope (Axiovert 100TV, Zeiss) with a ×63 plan-apochromat lens (NA 1.4), a precisely controlled xyz stage (Applied Precision, Issaquah, WA), and a scientific-grade cooled charge-coupled device camera (MicroMax, Princeton Instruments, Trenton, NJ). For creation of 3-D reconstructions of the cells, 20 optical sections, 300 nm apart along the z-axis, were obtained from each cell. Constrained iterative deconvolution and 3-D rendering were performed using Deltavision software (Softworx, Applied Precision) on a reduced instruction set computer (RISC) workstation (model 02 R10000, Silicon Graphics, Mountain View, CA).
oxLDL enhances flow-induced realignment of HAECs.
HAECs were preexposed to 10 μg/ml oxLDL (oxidation state 10–15 nmol/mg protein TBAR) for 1 h, as described in previous studies (3, 6). The levels of oxLDL were chosen to simulate the levels of circulating oxLDL reported in human plasma (22) and plasma of hypercholesterolemic animals (21). Control cells were exposed to the low (0.2%) serum alone for the same duration. After exposure to oxLDL, cells were mounted onto a flow apparatus and exposed to shear stress of 10 dyn/cm2 for 12 h, similar to the physiological levels of shear stress in human arteries (13). During the flow, cells were exposed to normal (2%) serum. All experiments were performed with subconfluent endothelial cultures to allow a direct comparison between sensitivity of the cells to flow and cell stiffness estimated by microaspiration, a technique that requires single cells. As shown in our earlier study (6), exposure to oxLDL significantly decreases the deformability of the cells, estimated by pulling the membrane into a micropipette (Fig. 1, inset). As expected, before exposure to flow (0 h), neither control nor oxLDL-treated cells have a preferential orientation relative to the direction of the flow (Fig. 1). Control HAECs appear to be relatively resistant to flow-induced realignment, with no significant decrease in standard deviations of the angles after 12 h of flow (Fig. 1). Exposure to oxLDL, however, dramatically enhanced flow-induced realignment of HAECs (Fig. 1). More specifically, oxLDL-treated cells showed pronounced realignment in the direction of the flow after 3 h of flow exposure, as evidenced by a decrease in standard deviations of cell orientation angles. Further realignment was observed after 6 and 12 h. A lack of flow realignment of control HAECs underscores the impact of oxLDL on the ability of these cells to sense and/or respond to flow by reorientating themselves in the direction of the flow. oxLDL also enhanced flow-induced cell elongation (Fig. 2).
OxLDL enhances alignment of F-actin stress fibers.
Earlier studies showed that oxLDL induces polymerization of actin and formation of F-actin stress fibers in several cell types (14, 26–28), including HAECs (14). Consistent with these studies, Fig. 3 shows that oxLDL also induces actin polymerization in HAECs. Under control no-flow conditions, HAECs have typical endothelial morphology, with a network of thin F-actin filaments (Fig. 3A). Exposure to oxLDL does not change the morphology of the cells but induces a marked increase in F-actin polymerization: F-actin-specific fluorescence increased from 37.7 ± 1.4 to 46.1 ± 1.8 (P < 0.05). Cells also have prominent stress fibers that crisscross the cytosolic regions (Fig. 3B). As expected (25, 34), exposure of the cells to flow induces formation of the stress fibers in control and oxLDL-treated cells (Fig. 3, C and D): F-actin specific fluorescence was 63.3 ± 6.5 and 92.2 ± 10.4 (P < 0.05) in control and oxLDL-treated cells, respectively. To analyze further the organization of F-actin in control and oxLDL-treated cells, we measured the angles of individual stress fibers in cells selected randomly from the two cell populations after the cells were exposed to flow for 6 h (30–40 individual fibers, which could be clearly identified and generally represented the brightest ones, were analyzed in each cell; curved fibers were excluded from the analysis). The angles were determined relative to the direction of the flow. This analysis shows that, consistent with the changes in cell orientations, exposure to oxLDL also significantly decreased the spread of the orientation angles and facilitated the realignment of the fibers in the direction of the flow (Fig. 3, E and F).
Cholesterol depletion enhances flow-induced realignment of HAECs.
Earlier studies showed that exposure to oxLDL results in cholesterol depletion of caveolae in vascular ECs and in homogenates of murine blood vessels (3, 23). Consistent with the notion that oxLDL induces cholesterol depletion from endothelial caveolae, we previously showed that oxLDL-induced endothelial stiffening can be simulated by MβCD-induced cholesterol depletion (6). Here we show that MβCD-induced endothelial stiffening (Fig. 4, inset) is accompanied by the facilitation of the ability of HAECs to realign in the direction of the flow (Fig. 4, A and B). The kinetics of the effect was slightly different than that of oxLDL in a sense that no significant effect was observed after 3 h of flow, but at 6 and 12 h the effects of MβCD and oxLDL were similar. As described in our earlier study using BAECs (4), MβCD-induced cholesterol depletion has no apparent effect on the degree of actin polymerization under static conditions (Fig. 5, A and B), but after the flow exposure, prominent stress fibers formed in control and MβCD-treated cells (Fig. 5, C and D). Similar to oxLDL, cholesterol depletion also significantly decreased the spread of individual actin fiber orientation (Fig. 5E).
oxLDL-induced enhancement of endothelial realignment in not unique to HAECs.
As pointed out above, HAECs are relatively resistant to flow-induced realignment. A question arises, therefore: is the effect of oxLDL on flow-induced realignment unique to HAECs precisely because of their inability to realign efficiently under normal-cholesterol conditions? To address this question, we tested whether oxLDL has a similar effect on BAECs, which are known to realign very well and have been used as a model for endothelial realignment in multiple earlier studies. Figure 6 shows that, similar to HAECs, exposure to oxLDL (10 μg/ml, 1 h) also enhances flow-induced realignment of BAECs. It is also interesting to note that we showed previously that BAECs are significantly stiffer than HAECs, as measured by microaspiration (5), an observation consistent with our hypothesis that an increase in cell stiffness is associated with an increased sensitivity of ECs to flow-induced realignment.
oxLDL-induced effect on endothelial realignment is partially reversed by acetylated LDL.
Earlier studies showed that oxLDL and acetylated LDL (acLDL) have the opposite effects on the level of endothelial cholesterol: oxLDL depleted cholesterol from endothelial caveolae (3), and acLDL enriched ECs with cholesterol (15). Indeed, acLDL is a known cholesterol donor and is widely used in lipoprotein research to load cells with cholesterol and simulate the physiological effects of LDL (16, 30). Furthermore, we showed earlier that exposure of aortic ECs to 50 μg/ml acLDL for 24 h increases endothelial cholesterol approximately twofold, similar to the increase observed in vivo under hypercholesterolemic conditions (15). Therefore, we tested whether the effect of oxLDL may be reversed by acLDL. In this series of experiments, cells were exposed to oxLDL (10 μg/ml), acLDL (50 μg/ml), or oxLDL + acLDL for 24 h. As shown in Fig. 7, the oxLDL-induced effect on flow-induced realignment was reversed by the simultaneous exposure to acLDL, supporting our hypothesis that changes in membrane cholesterol play a key role in the observed effects.
A hallmark of the endothelial response to flow is EC realignment in the direction of the flow. The realignment is believed to be important in distributing the hemodynamic forces applied to the luminal surface of the cells to the attachment sites on the basement membrane and, thus, preventing endothelial detachment, which would jeopardize the integrity of the endothelial layer (17, 41). It was also suggested that alignment of ECs in the direction of the flow reduces the shear gradients on the luminal cell surface, which also is expected to have a protective effect under high-shear conditions (1). Our study shows that EC realignment in the direction of the flow is enhanced when cells are preexposed to oxLDL.
Multiple studies have shown that flow-induced EC realignment is accompanied by the formation of thick bundles of F-actin filaments containing myosin and α-actinin, also known as F-actin stress fibers (25, 34). Furthermore, although F-actin stress fibers are found in a variety of cell types when cultured in vitro, this structure appears to be unique to vascular ECs in vivo (41). The most pronounced fibers oriented in the direction of the flow are observed in regions exposed to high-velocity flow, such as the left ventricle and aorta (41). Typically, flow-induced generation of the stress fibers occurs before EC realignment in the direction of the flow, and the fibers can be randomly distributed or oriented along the long axis of the cells (17, 25, 34). It is also known that the integrity of F-actin fibers is essential for flow-induced EC realignment (25, 34, 40). In the present study, we investigated how acute exposure to oxLDL alters the organization of F-actin in ECs exposed to steady unidirectional laminar flow. First, we show that oxLDL enhances F-actin polymerization in cells maintained under static conditions and in cells exposed to flow. These observations are in agreement with earlier studies showing that acute exposure to oxLDL induces formation of F-actin stress fibers, increases endothelial contractility (14), and induces activation of RhoA (14, 18, 36), which is well known to induce stress fiber formation. A longer (24 h) exposure to oxLDL resulted in disappearance of individual stress fibers and clustering of F-actin in the peripheral regions of the cells (43). Formation of actin stress fibers and formation of peripheral an actin band were suggested to contribute to an increase in endothelial permeability (14, 43). oxLDL was also reported to induce actin polymerization in macrophages (26, 27). In the present study, we extended these observations further to show that oxLDL also facilitates the reorientation of actin fibers in the direction of the flow, demonstrating that enhancement of flow-induced realignment of aortic ECs occurs on whole cell and individual fiber levels. Further studies are needed to determine the impact of oxLDL on the sensitivity of ECs to disturbed flow characteristic to atheroprone regions.
In terms of the mechanism, we propose that the effect of oxLDL is mediated by changes in the cholesterol content of the membrane. Several lines of evidence support this hypothesis. Earlier studies showed that oxLDL induces depletion of cholesterol from the caveolae followed by the internalization of caveolin (3). In addition, we recently showed that oxLDL also induces internalization of another major lipid raft marker, GM1, and that this effect is simulated by MβCD-induced cholesterol depletion (6). Consistent with these observations, we show here that oxLDL-induced facilitation of endothelial realignment is also simulated by cholesterol depletion. It is noteworthy, however, that oxLDL-induced enhancement of endothelial realignment is observed earlier than the effect of MβCD-induced cholesterol depletion, suggesting that cholesterol depletion may not fully account for the effect of oxLDL. Alternatively, it is also possible that oxLDL is more specific and efficient than MβCD in removing cholesterol from a specific membrane compartment, such as caveolae, whereas MβCD may not only decrease cholesterol in all membrane domains but may also induce redistribution of cholesterol from one membrane domain to another (44). Finally, we show here that oxLDL-induced enhancement of endothelial realignment is rescued by exposure of the cells to acLDL, whereas exposure to acLDL alone has no significant effect. The latter observation is also consistent with our previous study showing that cholesterol enrichment has no impact of endothelial stiffness (4).
What is the possible link between oxLDL/cholesterol depletion and the sensitization of cells to flow? Our recent studies showed that oxLDL and MβCD result in an increase in stiffness of vascular ECs (4–6), and we propose that changes in endothelial biomechanics play a critical role in the sensitization of the cells of flow. Indeed, formation of stress fibers is expected to increase cell stiffness, consistent with the earlier studies showing that shear stress alone also induces an increase in cell stiffness (33, 34). An increase in stress fiber content and an associated increase in cell stiffness are expected to result in an increase in cortical tension that develops in the submembrane cortical region of the cytoskeleton. Our general hypothesis is that oxLDL-induced endothelial stiffening increases cortical cytoskeleton tension and results in sensitization of the cells to a mechanical signal generated by shear stress (Fig. 8). The exact nature of the flow sensor, however, is not known, despite the numerous studies addressing this question. The putative candidates for flow sensors include integrins, ion channels, G proteins (12, 25), and, more recently the endothelial glycocalyx (42). The latter appears to be in apparent contradiction with our studies, because diet-induced dyslipidemia, in general, and oxLDL, in particular, were shown to result in partial degradation of the endothelial glycocalyx (7, 38, 39). oxLDL-induced degradation of the glycocalyx, however, appears to be transient, with full recovery after ∼1 h (7, 39) and, therefore, may not have a major impact on flow-induced endothelial realignment. Further studies are needed to elucidate specific molecular pathways responsible for oxLDL-induced endothelial stiffening and flow sensitization.
Finally, several lines of evidence suggest that oxLDL-induced endothelial stiffening may occur in vivo. Cholesterol depletion of endothelial caveolae is observed not only in response to oxLDL in vitro, but also under dyslipidemic conditions in vivo. Specifically, cholesterol content is lower in caveolae isolated from the vessels of cholesterol-fed apolipoprotein E-deficient mice, a mouse model for atherosclerosis, than in caveolae isolated from control animals (23). Furthermore, ECs covering fatty streaks in the vessels of cholesterol-fed rabbits were shown to have fewer caveolae, an effect that was also simulated by cholesterol depletion in cultured ECs (8). Consistent with these studies, we previously showed that dyslipidemia also results in an increase in endothelial stiffness in a diet-induced porcine model of atherosclerosis (6). We propose, therefore, that oxLDL-induced changes in endothelial biomechanics play an important role in endothelial physiology in vivo.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-073965 and HL-083298.
We thank Dr. Peter Davies for sharing with us the flow apparatus and for multiple discussions throughout this project. We also thank Mete Civelek for help in the initial stages of the experiments and Dr. Min Joo for help with the statistical analysis of the variances.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society