Биоинженерия / Биомеханика_микрофлюидные_устройства / статьи / 20_BIOMGB-000005-032006_1

032006-2 Estrada et al. Biomicrofluidics 5, 032006 (2011)

(ECs) and smooth muscle cells (SMCs)), which dictates initiation and progression of wall thickening. Plaques can also suddenly rupture, leading to thrombus formation and vessel occlusion increasing the risk of myocardial infarction and stroke, which together are responsible for50% of mortalities in developed nations.2

Atherosclerotic lesions preferentially occur at vascular niches proximal to vessel branches and bends including several locations in the aortic arch, ascending and descending aorta and coronary and carotid arteries.3 At these locations, the local flow behavior is characterized as “disturbed” and is associated with low shear stress recirculation, oscillation, or lateral flow (average: <4 dynes=cm2) in comparison, straight regions of arteries with “normal” high shear stress laminar pulsatile flow (mean: 10–15 dynes=cm2, peak: <100 dynes=cm2). Several in vitro and in vivo studies evaluating the effect on shear stress associated with disturbed flow have shown generation of ECs with polygonal morphology,4,5 random alignment of actin filaments,4,5 high EC turnover,4,5 slow EC migration,4,5 increased permeability,4,5 increased gene expression,4,5 and expression of pro-inflammatory markers such as platelet derived growth factor (PDGF),6 monocyte chemotactic protein-1 (MCP-1),6 platelet endothelial cell adhesion molecule (PECAM)-1,7 vascular cell adhesion molecule (VCAM)-1,7,8 intra cellular adhesion molecule (ICAM)-1,7,8 and growth factors such as vascular endothelial cell growth factor – receptor 2 (VEGF-R2).9

In addition to shear stress, ECs in vivo are also exposed to pulsatile pressure (normal: 120=80 mm Hg) and cyclic stretch (normal: 6–20%) at around 80 bpm. However, evaluation of the effects of mechanical stretch in vitro has focused mostly on SMCs whereas ECs have received a lot less attention. Several recent studies have evaluated the response of ECs to both physiological and pathological levels of cyclic stretch.10–12 Evaluation of cyclic stretch on ECs accomplished using flexible substrates reveals alignment of ECs perpendicular to the direction of stretch13 (in the direction of flow in blood vessels) and stimulation of stretch-activated ion channels for Ca2þ ion transport.14 In vitro studies demonstrate that short term stretch results in modulation of vessel tone through synthesis of superoxide15 known to play a role in vasoconstriction whereas prolonged exposure to stretch resulted in increased generation of vasoactive mediators NO and ET-1,16 expression of angiotensin receptor II (ANG II-R),17 and pro-inflam- matory markers such as IL-8,10 MCP-1,10 ICAM-1,18 and VCAM-1.18 Physiological levels of cyclic stretch have also been shown to significantly improve endothelial barrier function whereas increased levels of stretch seen in conditions such as hypertension decreases barrier function.19 Relatively fewer in vitro studies have focused on the direct effects of pressure on EC structure and function.20 This can be attributed to the assumption that pulsatile pressure from blood flow causes the blood vessel to stretch, and therefore, the overall effect of pressure manifests itself primarily in the form of stretch. However, evaluation of pressure on ECs in vitro shows that pressure alone in the absence of stretch results in increased EC proliferation, cytoskeletal reorganization, and synthesis of ECM proteins.21–23

Conventional platforms such as parallel plate systems24 and cone in plate viscometers25 create disturbed flow conditions in the absence of pressure and stretch. Cyclic stretch is an important determinant in regulation of cytoskeletal alignment, endothelial barrier function, modulation of vessel tone, and inflammation and may play an important role particularly in areas of low shear stress.26 The effects of in vitro stimulation with stretch and shear stress have shown to have common and opposing effects on EC phenotype and function. Both shear stress and cyclic stretch cause increase in eNOS, intracellular Ca2þ, inositol triphosphate, and triacylglycerol.19 Shear stress and cyclic stretch also differentially regulate ET-1, ICAM-1, plasminogen activation inhibitor, MCP-1, and oxidative stress response.19 In general, increasing shear stress in the absence of additional stimulation results in an anti-inflammatory phenotype whereas stimulation with increasing cyclic strain alone results in a pro-inflammatory phenotype. However, the response of ECs to individual stimuli is very different from the simultaneous and coordinated stimulation sustained in vivo.27

Considering the complex interplay between different mechanical stresses, in vitro stimulation with shear stress alone may not fully replicate the in vivo signaling environment in its entirety. Therefore, concomitant simulation with pulsatile pressure and strain may be critical to generate a more physiologically relevant in vitro model of atherosclerosis. To address this issue,

camera attached to the microscope. Three images (>30 cells=image) were taken for each device at randomly selected locations, and the shape of cells was determined using the METAMORPH software (tracing boundaries of single cells) and classified as either ellipsoidal or polygonal. The ratio of ellipsoidal to polygonal cells was averaged for each condition. To determine alignment, the angle of orientation of cells with respect to the direction of fluid flow was determined and cells within 620 of the direction of fluid flow were considered aligned. For examining differences between the shape and alignment of cells in the two conditions, unpaired Student’s t- tests were performed with two-tailed significance set at p 0.05 (n ¼ 3).

Microfilament arrangement and adherns junction formation using immunofluorescence microscopy

HAECs were fixed with 4% paraformaldehyde in 1 PBS for 20 min, washed two times with wash buffer (1 PBS containing 0.05% Tween-20 (Fisher Scientific, Fair Lawn, NJ)) and permeabilized with 0.5% Triton X-100 (Fisher Scientific, Fair Lawn, NJ) for 2 min at room temperature. Then, cells were washed two times with wash buffer, blocked with 1% BSA in 1 PBS freshly prepared for 30 min, and incubated with primary antibody anti-human mouse b-Catenin (1:50; Santa Cruz Biotechnology, Santa Cruz, CA) at room temperature for 1 h. Cells were washed three times with wash buffer for 5 min each time and incubated at room temperature with the second antibody, fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse (1:100; Millipore, Billerica, MA). After 1 h, cells were washed three times with wash buffer. For negative controls, the same procedure was performed without adding the primary antibody. For the detection of F-actin, cells were washed two times with wash buffer, blocked with 1% BSA in 1 PBS for 30 min, and incubated at room temperature for 1 h with TRITC-conjugated phalloidin (1:100; Millipore, Billerica, MA). Light Diagnostics mounting fluid (Millipore, Billerica, MA) was added to the cells, and cells were examined using a Nikon Eclipse A1 Confocal Microscopy System (Nikon Instruments, Melville, NY).

Evaluation of inflammatory markers using flow cytometry

The surface area of the ECCM allows culture of >15 000 HAECs. By obtaining primary antibodies conjugated to four different fluorophores (FITC, PE, PerCP, and APC), simultaneous evaluation of four different markers can be accomplished. HAECs cultured within the ECCM under normal and disturbed flow conditions for 24 h were trypsinized and resuspended in 250 lL of flow cytometry buffer (1 PBS containing 2% paraformaldehyde and 1% bovine serum albumin). 10 lL of each antibody (anti-human PECAM-1-FITC, VEGF-R2-PE, VCAM-1- PerCP, and ICAM-1-APC) were added to the cell sample and incubated at room temperature in the dark for 45 min. Samples were washed with an additional 250 lL of flow cytometry buffer, centrifuged at 350g for 5 min, and finally resuspended in 200 lL of flow cytometry buffer processed. Results were normalized to isotype controls and analyzed using flow cytometry data analysis software (WINMDI and BD CELL QUEST PRO). Paired Student’s t-tests with two-tailed significance set at p 0.05 was used to quantify differences in expression of inflammatory markers (n ¼ 3).

RESULTS

Generation of pressure, flow, strain, and shear stress waveforms

The ECCM was used to generate pressure, flow, strain, and shear stress waveforms associated with normal (pressure 118=83 mm Hg systolic=diastolic, 13% constant strain, 6% cyclic strain, average flow of 28 ml=min, and average shear stress of 11 dynes=cm2) and disturbed flow (pressure 118=83 mm Hg systolic=diastolic, 13% constant strain, 6% cyclic strain, average flow of 3 ml=min, and average shear stress of 1.3 dyne=cm2) conditions (Fig. 2). A setup similar to previously described was used to generate normal flow conditions whereas the setup was slightly modified by removal of the one-way valve and reduction of flow rate to generate reciprocating disturbed flow conditions.

FIG. 2. Pressure, flow, strain and shear stress waveforms generated using the ECCM to culture ECs under (A) normal flow (pressure 118=83 mm Hg systolic=diastolic, 13% constant strain, 6% cyclic strain, average flow of 28 ml=min, and average shear stress of 11 dynes=cm2) and (B) disturbed flow conditions (pressure 118=83 mm Hg systolic=diastolic, 13% constant strain, 6% cyclic strain, average flow of 3 ml=min, and average shear stress of 1.3 dyne=cm2). These waveforms closely match the clinical waveforms reported in literature.

Replication of shear waveform in the infrarenal segment of the abdominal aorta proximal to the inferior mesenteric artery

Once disturbed flow was enabled within the ECCM, the pulmonary compliance and resistance as well as the aortic compliance and resistances were modulated to replicate the flow and shear stress profiles seen in the infrarenal segment proximal to the inferior mesenteric artery, an atherosclerosis susceptible region in the abdominal aorta. For the waveforms generated, the mean shear stress was 1.37 dynes=cm2, the maximum shear stress was 7.5 dynes=cm2 and the minimum shear stress was 4.4 dynes=cm2, which are similar to in vivo measured values.29 The measured OSI and NEG values are 0.27 and 0.38, respectively, and again these values closely match in vivo values to within 65% error.29 Figure 3 shows the shear waveform over 1 beat and the calculated OSI and NEG values. This condition was used for all disturbed flow experiments.

Evaluation of cell shape and alignment

Phase contrast microscopy was used to characterize HAEC shape and alignment. Image analysis of cells cultured under both normal and disturbed flow conditions was used to determine percentage of cells per frame exhibiting ellipsoidal shape and percentage of cells per captured frame that were aligned (<20 from the direction of flow). Microscope images and image processed data clearly indicates that cells attain an ellipsoidal shape and show good alignment in the direction of flow with normal high shear stress laminar flow, whereas cells cultured under disturbed flow exhibit rounded or cuboidal shape and random orientation (Fig. 4). The percentage of ellipsoidal and aligned cells in the normal flow condition was statistically significant in comparison (p ¼0.05) for a sample size of n ¼ 3.

FIG. 4. Phase contrast images of cells cultured under conditions of (A) normal and (B) disturbed flow. Graphs quantifying % of cells cultured under both conditions that assume (C) ellipsoidal shape and (D) alignment in the direction of flow following culture for 24 h within the ECCM. Differences in shape and alignment are statistically significant (p < 0.05).

pathophysiologic conditions in the body can be accurately replicated. This system has been used to generate realistic mechanical stresses associated with conditions like heart failure, hyper=hypo-tension and brady=tachy-cardia where there is no sustained retrograde flow. This manuscript describes a new application of this technology for culture of ECs under conditions of disturbed flow seen in atherosclerosis susceptible regions resulting in generation of low and oscillatory shear stress while maintaining pressure and strain waveforms at normal levels as seen in vivo. Results clearly show that the flow and shear waveforms generated, accurately mimic conditions seen in the infrarenal segment of the abdominal aorta at the wall distal to the inferior mesenteric artery where the average shear stress value is 1.3 dynes=cm2, the OSI value is 0.26, and the NEG index is 0.38. This model therefore represents the first demonstration of an in vitro model of atherosclerosis with realistic concomitant stimulation with shear stress, pressure, and strain.

To demonstrate proof of concept of the ability to accurately replicate normal and disturbed flow conditions within the ECCM we cultured HAECs within this system and evaluated morphological, structural, and functional differences between cells cultured under both conditions. Evaluation of cell shape and alignment using image analysis of phase contrast imaging clearly demonstrates that >70% of cells cultured under normal conditions assume an ellipsoidal shape and >85% of cells align in the direction of fluid flow (within 20 ). However, analysis of cells cultured under disturbed flow conditions shows that <10% of cells assume an ellipsoidal shape and <30 % of cells align in the direction of the flow.

Actin microfilaments (F-actin) in the cytoskeleton of cells play important roles in the establishment and maintenance of structural stability and are involved in variety of signaling

FIG. 7. Flow cytometric analysis of expression of HAEC activation and growth factors (A) PECAM-1, (B) VEGF-R2, (C) VCAM-1, and (D) ICAM-1 on HAECs cultured under normal and disturbed flow conditions. Plots show small but statistically insignificant increase in expression of these markers in HAECs cultured under disturbed flow in comparison to HAECs cultured under normal flow. (E) Graph summarizing flow cytomety data.

white blood cells (WBCs) between blood and tissues, especially at sites of infection and inflammation. In ECs subject to normal flow, cell-cell junctions are usually depicted with a linear morphology along the boundaries between adjacent cells and in contact with cortical F-actin. However, in conditions of disturbed flow, the cell-cell contacts are discontinuous with an irregular morphology. Adherens junctions are based on the transmembrane adhesive receptor that belongs to the Cadherin family, specifically vascular endothelial (VE)-Cadherin. VE-Cadherin is constitutively linked through its cytoplasmic tail directly to b- or c-Catenin and indirectly to a-Catenin via b- or c-Catenin. Staining for b-Catenin within HAECs cultured under both conditions again confirms in vivo observations. Cells exposed to normal flow exhibit continuous and higher levels of expression of b-Catenin whereas cells cultured under disturbed flow exhibit low and intermittent expression.

Several studies30–34 evaluating ECs under conditions of disturbed flow report increased expression of EC activation markers involved in the binding of activated leukocyte populations like monocytes, macrophages and possibly granulocytes on the endothelium like VCAM-1, ICAM-1, PECAM, and growth factor VEGF-R2, which is involved in EC repair and regeneration. All of these markers are known to be involved in the initiation and progression of atherosclerosis lesions. However, our data indicates flow conditions alone result in small and statistically insignificant changes (p ¼0.05) in expression levels of these four markers in normal and disturbed flow. This is consistent with events leading to lesion formation and progression in regions of disturbed flow in vivo where activation is triggered only in the presence of proinflammatory stimulation via various factors such as hypercholesterolemia, hyperglycemia, hypertension, smoking, and elevated levels of serum C-reactive protein. Our data therefore indicates that in the absence of pro-inflammatory stimulation, markers known to be involved in initiation and progression of atherosclerotic lesions should be comparable to levels in cells cultured under normal conditions.

These results establish proof of concept of the ability of the ECCM to generate a physiologically relevant in vitro cell culture model of atherosclerosis where the specificity of cellular level systems is maintained and the in vivo mechanical loading environment is accurately replicated. This model can be used to understand the molecular basis of events involved in the transduction of altered mechanical stresses in disturbed flow regions that represent a significant risk factor in atherosclerotic lesions formation. This system also provides an ideal platform for drug discovery and development of treatment options for atherosclerosis.