**3.1 Experimental methods**

### **3.1.1 Asymmetric stenosis model design**

Parallel-plate flow chambers are not ubiquitous when characterizing the role of WSS in endothelial dysfunction and leukocyte adhesion. Cone-plate viscometers (Shankaran & Neelamegham, 2001), animal models (Walpola et al., 1993, 1995) and three-dimensional (3D) tissue culture models (Hinds et al., 2001) have an increasing presence in the field.

A 3D model of an idealized coronary artery with an eccentric stenosis has been developed by our research group to reveal the effect of spatial WSS gradients on both endothelial inflammation and leukocyte adhesion (Rouleau et al., 2010a, 2010b). The eccentric stenosis geometry with a 50% occlusion (i.e. 50% area reduction, orthogonal to flow) has been chosen to represent a clinically relevant atherosclerotic lesion (Brunette et al., 2008; Wexler et al., 1996). The model measures 10 cm in length with an internal diameter of 3.175 mm, Figure 3. Sylgard™ 184 silicone elastomer is cast and cured in PVC moulds to create semicompliant structures that maintain their geometric integrity through the stages of sterilization and cell culture preparation.

Fig. 3. Three-dimensional asymmetric stenosis model schematic with regional classifications defined using computational fluid dynamics and photochromic molecular flow visualization (Section 3.1.3).

### **3.1.2 Perfusion design**

The perfusion flow loop consists of a media reservoir with tubing, flow dampeners, and an 8-roller peristaltic pump head with a programmable drive to produce steady or pulsatile laminar flow at the entrance of the models, Figure 4. ECs are cultured in the internal lumen of the model until they form a continuous, confluent monolayer. Inlet WSS values of 4.5, 9 and 18 dynes/cm2 were chosen to represent moderate physiological shear in coronary arteries whereas inlet WSS values of 1.25 and 6.25 dynes/cm2 were chosen to represent moderate to high shear for *in vitro* neutrophil adhesion. The WSS field was experimentally and numerically determined (Rouleau et al., 2010b).

Fig. 4. Schematic of a steady-flow perfusion experiment with an asymmetric stenosis model.

geometry with a 50% occlusion (i.e. 50% area reduction, orthogonal to flow) has been chosen to represent a clinically relevant atherosclerotic lesion (Brunette et al., 2008; Wexler et al., 1996). The model measures 10 cm in length with an internal diameter of 3.175 mm, Figure 3. Sylgard™ 184 silicone elastomer is cast and cured in PVC moulds to create semicompliant structures that maintain their geometric integrity through the stages of

Fig. 3. Three-dimensional asymmetric stenosis model schematic with regional classifications

The perfusion flow loop consists of a media reservoir with tubing, flow dampeners, and an 8-roller peristaltic pump head with a programmable drive to produce steady or pulsatile laminar flow at the entrance of the models, Figure 4. ECs are cultured in the internal lumen of the model until they form a continuous, confluent monolayer. Inlet WSS values of 4.5, 9 and 18 dynes/cm2 were chosen to represent moderate physiological shear in coronary arteries whereas inlet WSS values of 1.25 and 6.25 dynes/cm2 were chosen to represent moderate to high shear for *in vitro* neutrophil adhesion. The WSS field was experimentally

Fig. 4. Schematic of a steady-flow perfusion experiment with an asymmetric stenosis model.

defined using computational fluid dynamics and photochromic molecular flow

sterilization and cell culture preparation.

visualization (Section 3.1.3).

and numerically determined (Rouleau et al., 2010b).

**3.1.2 Perfusion design** 

### **3.1.3 Computational fluid dynamics and photochromic molecular flow visualization**

Computational fluid dynamics (CFD) is a theoretical branch of research that relies on the power of modern computers to estimate fluid behaviour. The popularity of CFD lies in its ability to simulate physical experiments, thereby providing direction for further work or allowing for quick testing of key variables. With respect to cardiovascular flow studies, CFD is a numerical solution to a continuum of the Navier-Stokes (NS) equation. Although CFD is a powerful tool, it is only as accurate as the input data (e.g. geometry and mechanical properties). Defining the geometry and mechanics of healthy and diseased tissue is a constant endeavour in biomedical engineering (Choudhury et al., 2009; Tremblay et al., 2010). If the proper information is available, then CFD simulations are feasible, however experimental validation is still crucial. CFD has been performed for our stenosis model, yielding flow profiles at 6.25 dynes/cm2 and 1.25 dynes/cm2, respectively (Figures 5 & 6).

Fig. 5. Velocity profile, with an appreciable recirculation zone, in the asymmetric tissue culture model at 6.25 dynes/cm2.

Fig. 6. Flow analysis of the asymmetric tissue culture model. (A) CFD normalized WSS contour plot. (B) Velocity profile at 1.25 dynes/cm2. (C,D) PMFV velocity profiles at the inlet and peak, respectively.

We have used the photochromic molecular flow visualization (PMFV) technique to validate our CFD flow in the stenosis model (Ethier C.R. et al., 2000; Mahinpey et al., 2004). Using the flow profile at a given position, we are able to estimate the shear stress acting on the wall of a channel. Photochromic species reversibly change conformation when excited by a light source, such as a laser, resulting in an observable colour change. A PMFV setup includes a solution of photochromic dye, a laser, and a high-resolution camera. In practice, the photochromic dye solution is pumped through a micro-channel, the laser is triggered and the resulting pulse passes through the solution orthogonal to flow, activating any dye it contacts (Couch et al., 1996; Park et al., 1999). At this moment, a narrow column of visible dye will appear within the solution. The solution is in motion, however, so a fraction of a second later, the excited dye will have displaced with flow. This displacement is recorded by a camera, providing a snapshot of the flow profile and subsequently, the shear stress on the opposing walls. Photochromic visualization results are presented in Figure 6 (c) & (d) to validate our CFD simulation.

### **3.1.4 Cellular analysis**

EC morphology is often a good predictor of EC phenotype. Healthy ECs elongate in the direction of flow, whereas dysfunctional ECs may become randomly oriented and cobblestone in appearance (Dartsch & Betz, 1989). The shape index (SI) is a metric of EC morphology defined as (Nerem et al., 1981):

$$SI = \frac{4\pi \cdot Area}{Perimeter^2} \tag{7}$$

Generally, elongated cells have a lower SI than rounded cells. In concert with the SI, the angle of orientation evaluates the proportion of EC elongation relative to the direction of flow. ECs in regions of observed elongation will be narrowly distributed near 0° (i.e. in the axis of flow) while regions that appear cobblestoned will have a much wider distribution.

Protein and mRNA regulation of inflammatory markers and transcription factors defines the endothelial phenotype and relates to leukocyte adhesion and atherogenesis. Gene regulation is quantified for large cultures using Q-PCR, while the resultant protein expression is observed using Western Blotting. Regional inflammation around the stenosis can be observed using immunostaining and confocal microscopy for adhesion molecules and the translocation of inflammatory transcription factors to EC nuclei. Our analysis includes, but is not limited to, specific inflammatory markers and adhesion molecules, including: ICAM-1, VCAM-1, E-selectin and NF-κB.

### **3.2 Results and discussion**

### **3.2.1 Endothelial cell morphology**

The stenosis model was first used to investigate the morphological effects of shear gradients caused by the stenosis. A morphological response is one of the last measurable changes which occurs in the cascade of events following introduction of flow. As ECs experience a steady WSS, they tend to become elongated and aligned in the direction of flow, representing a healthy endothelium. When exposed to low shear magnitude or WSS spatial gradients, ECs tend to take on a more cobblestone and random morphology which is indicative of an unhealthy endothelium (Helmlinger et al., 1991; Levesque et al., 1986; Levesque & Nerem, 1985; Nerem et al., 1981; Nerem, 1993).

Perfusion experiments to evaluate EC response were run at wall shear stress values of 4.5, 9 and 18 dyne/cm2 which corresponded to Reynold's numbers of 50, 100 and 200, respectively. It was found that at all times and inlet WSS values the shape indices in the inlet and outlet of the stenosis model were statistically similar to that of a straight model. These results demonstrate that these values can be a good reference point for morphological changes in the regions surrounding the stenosis. Furthermore, the longer the perfusion time, the more elongation was observed in the direction of flow.

### *Effect of wall shear stress gradients on endothelial cell morphology*

The stenosis model allows the observation of the morphological response of the ECs to WSS gradients. Figure 7 shows the WSS patterns within the model as a function of position. A

1996; Park et al., 1999). At this moment, a narrow column of visible dye will appear within the solution. The solution is in motion, however, so a fraction of a second later, the excited dye will have displaced with flow. This displacement is recorded by a camera, providing a snapshot of the flow profile and subsequently, the shear stress on the opposing walls. Photochromic

EC morphology is often a good predictor of EC phenotype. Healthy ECs elongate in the direction of flow, whereas dysfunctional ECs may become randomly oriented and cobblestone in appearance (Dartsch & Betz, 1989). The shape index (SI) is a metric of EC

ܵܫ ൌ ସగή

Generally, elongated cells have a lower SI than rounded cells. In concert with the SI, the angle of orientation evaluates the proportion of EC elongation relative to the direction of flow. ECs in regions of observed elongation will be narrowly distributed near 0° (i.e. in the axis of flow) while regions that appear cobblestoned will have a much wider distribution. Protein and mRNA regulation of inflammatory markers and transcription factors defines the endothelial phenotype and relates to leukocyte adhesion and atherogenesis. Gene regulation is quantified for large cultures using Q-PCR, while the resultant protein expression is observed using Western Blotting. Regional inflammation around the stenosis can be observed using immunostaining and confocal microscopy for adhesion molecules and the translocation of inflammatory transcription factors to EC nuclei. Our analysis includes, but is not limited to, specific inflammatory markers and adhesion molecules, including: ICAM-

The stenosis model was first used to investigate the morphological effects of shear gradients caused by the stenosis. A morphological response is one of the last measurable changes which occurs in the cascade of events following introduction of flow. As ECs experience a steady WSS, they tend to become elongated and aligned in the direction of flow, representing a healthy endothelium. When exposed to low shear magnitude or WSS spatial gradients, ECs tend to take on a more cobblestone and random morphology which is indicative of an unhealthy endothelium (Helmlinger et al., 1991; Levesque et al., 1986;

Perfusion experiments to evaluate EC response were run at wall shear stress values of 4.5, 9 and 18 dyne/cm2 which corresponded to Reynold's numbers of 50, 100 and 200, respectively. It was found that at all times and inlet WSS values the shape indices in the inlet and outlet of the stenosis model were statistically similar to that of a straight model. These results demonstrate that these values can be a good reference point for morphological changes in the regions surrounding the stenosis. Furthermore, the longer the perfusion time,

The stenosis model allows the observation of the morphological response of the ECs to WSS gradients. Figure 7 shows the WSS patterns within the model as a function of position. A

௧<sup>మ</sup> (7)

visualization results are presented in Figure 6 (c) & (d) to validate our CFD simulation.

**3.1.4 Cellular analysis** 

morphology defined as (Nerem et al., 1981):

1, VCAM-1, E-selectin and NF-κB.

Levesque & Nerem, 1985; Nerem et al., 1981; Nerem, 1993).

the more elongation was observed in the direction of flow. *Effect of wall shear stress gradients on endothelial cell morphology* 

**3.2 Results and discussion 3.2.1 Endothelial cell morphology**  positive shear gradient is found in the proximal region of the stenosis, reaching a peak WSS just upstream of the apex. A negative WSS gradient is observed in the recirculation zone of the stenosis, however, downstream of the flow reattachment point, laminar Hagen-Poiseuille flow resumes.

Fig. 7. CFD simulation of the WSS and WSS gradient Profile along the bottom central axis of the stenosis model (**–** Normalized WSS Magnitude; **–** Normalized WSS gradient).

It was found that after sufficient time had passed, morphological trends formed throughout the model. The inlet and outlet of the model showed similar shape indices as the straight tube controls, making them acceptable internal controls. These uniform internal controls can be compared to the other regions of the models. The deceleration, or recirculation zone (depending on the flow patterns, governed by the Re), showed the highest shape index, perhaps indicative of the most inflammatory response. The acceleration, or proximal zone, showed a slightly elevated shape index compared to the control regions, though this value was still statistically lower than that found in the deceleration zone. It was expected that the increased shear in this zone would result in more elongated ECs, however the results suggest that WSS gradients can have a more drastic effect on endothelium health than WSS magnitude alone.

It can be concluded that the deceleration zone could potentially present an inflamed endothelium and therefore one would predict to see an increase in regional expression of proteins linked to inflammation in that region. In turn, this should also lead to the largest amount of neutrophil adhesion in the recirculation zone.
