To use the dual flow modality, cells seeded within and/or on one or both sides of a permeable membrane, or cells seeded in a scaffold disc, or cells in situ (in a slice of tissue), are interposed between the two gaskets and flow is applied. For example, cells are seeded on a porous membrane (0.2 micron Anapore membrane, Nunc International, Denmark) that is interposed between two gaskets. The cell monolayer is interposed between two identical fluid layers, each with an independent inlet and outlet (Figure 1). The dual configuration provides relatively even flow across both surfaces of the monolayer, where the thin membrane (thickness is chosen by the user) between the regimes suspends the specimen in a physiologic manner. However, as mentioned previously, the membrane can be interchanged with a variety of other substrates, depending on the specific needs of the study, e.g. a layer in which cells are seeded or embedded can be placed between the gaskets or a slice of tissue with cells in situ can also be interposed between the gaskets. Furthermore, single-flow can also be achieved with a solid coverslip in place of the membrane.

Following the general flow design, the chamber itself consists of an upper and lower reservoir that is dictated by compressing rubber gaskets between two polycarbonate cases (Figure 2). The solid cases are fabricated such that the inlet and outlet channels are encased within the polycarbonate, in which the flow descends from the inlet channel, into the upper reservoir (formed by the geometry of the rubber gasket), and then exits via the polycarbonate outlet channel. Aside from the inlet/outlet channels, the flow regime, from the perspective of the cellular monolayer, is completely governed by the rubber gasket geometry as well as a glass coverslip above the monolayer which enables microscope visualization into the chamber. It should be noted that single-flow is achieved by the replacing the membrane with a glass coverslip, or by using a glass plug which completely "fills" the lower reservoir, decreasing the overall thickness of the chamber for ease of use when single-flow is desired.

Five different gasket geometries are created (Figure 3) to provide a variety of controlled stress fields to the cell monolayer, depending on the desired application,. Each gasket is compatible with the general parallel plate configuration; geometric differences control the specific spatial variation of shear stress imparted to the cell. In gasket I the geometry is designed to provide a constant shear stress to the majority of the cellular monolayer. In this case, the geometry contains zones of widening and tapering geometry that surround the parallel walls near the chamber midplane. In gasket II and III, the gasket geometry contains symmetric and asymmetric wide flow zones, whose walls taper immediately and linearly as they approach the outlet. Thus, the geometry provides specific gradients of stress to a monolayer. Finally, the last two gasket designs are designed to provide planar jet like flow, where zones with geometries that widen and taper abruptly define two different flow geometry lengths. Again, the geometry is created such that specific gradients of stress would be imparted to the cells, as opposed to the constant stress environment of gasket I.

Computational models are created for each design to predict the fluid environment and imparted shear stress in each gasket. The inlet and outlet geometries do not change within the polycarbonate sections of the chamber. Hence, it is assumed that the entering and exiting flow are identical for each case and only the geometry within the rubber gaskets is modelled. Furthermore, the chamber design can accommodate a variety of gasket thicknesses. For the computational fluid dynamics (CFD) study, all gaskets are created with a thickness of 250 μm for comparison with previous predictions for commercial chambers 20. In addition, gasket I is modelled for both 250 μm and 500 μm thicknesses to demonstrate flow variation attributable to gasket thickness. Gasket I is also modelled for single- and dual-flow scenarios, where the membrane is given a porosity of 50% and permeability of 10^-16 m². All other gaskets are modelled in single-flow mode. Inter-geometric variations in dual flow mode can be inferred from gasket I simulation results. In all cases, the cells are not included in the gasket fluid geometry; the effect of flow on cells has been reported previously 20. For each flow model (gaskets I-V), the number of volumes were 89760, 117130, 117130, 179380, and 189180, respectively. The mesh density was increased until grid independence was achieved for each gasket (average finite volume = 10^-13 m³).

The fluid environment is simulated for each model using a computational solver (CFD-ACE software package, Huntsville, AL) that calculates the velocity, pressure, and shear stress distribution. Flow is induced via a pressure gradient for each model to achieve a target wall shear stress at the exposed surface of the cell (at the interface with the flow channel's lower surface) is 1 dyn/cm² 2324257. Also, no-slip conditions are enforced on the walls of each gasket. For all models, the continuity equation (1) and Navier-Stokes equations (2) are solved using a 2^nd order upwind-discretization scheme in three dimensions, and the wall shear stress is calculated from the wall strain rate (3).

∇·u = 0

p(u·∇u) = -∇p + μ∇²u

τ w a l l = μ γ ˙ MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xI8qiVKYPFjYdHaVhbbf9v8qqaqFr0xc9vqFj0dXdbba91qpepeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaacciGae8hXdq3aaSbaaSqaaiabdEha3jabdggaHjabdYgaSjabdYgaSbqabaGccqGH9aqpcqWF8oqBcuWFZoWzgaGaaaaa@3809@

where u is the velocity vector, ρ is the fluid density, p is the pressure, μ is the fluid viscosity, τ is shear stress, and γ˙ MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaacciGaf83SdCMbaiaaaaa@2D8A@ is the strain rate. In all cases, the fluid is assumed to be similar to water and given a viscosity = 0.001 kg/ms and density = 1000 kg/m³. Flow is simulated for steady conditions with a local convergence criterion of 0.0001, where conservation of mass is used to validate each solution. It is important to note that, while simulations are calculated to achieve a target shear stress of 1 dyn/cm², the results of these simulations can be scaled up for flow regimes in the laminar range (appropriate assumptions for flow regimes up to and beyond 20 dyn/cm², i.e. regimes commonly applied for mechanotransduction research).

In order to evaluate the chamber design in a realistic laboratory setting, bench-top testing is performed to determine ease of use and performance. During testing, both strengths and weaknesses are noted in comparison to the previously evaluated commercial chambers 20. Ease of use testing consists of assembling the chamber for fluid/cell experiments, with emphasis placed on the choice of the materials and geometry for each design and how the chambers interface with standard laboratory equipment used in flow/cell studies (i.e. tubing, microscope, syringe pump, etc.). However, performance testing consists of actual flow through the assembled chamber, where problems associated with leaking, bubbling, visualization, etc. are the main areas of interest. In both modes, the novel design is compared with existing designs, providing a basis to evaluate improvement over the current state-of-the-art.

For the first gasket geometry, the maximum velocity and wall shear stress are calculated for three cases: (a) 250 μm gasket thickness, (b) 500 μm thickness, and (c) dual-flow 250 μm thickness (each layer). The maximum velocity in each case is highest near the inlet/outlet, with a nearly constant magnitude throughout the rest of the geometry (displayed within the plane) (Figure 4). As the thickness or gasket height is doubled from 250 to 500 μm, the maximum velocity (at the center height) increases proportionally. However, when the 250 μm gasket is used in dual-flow mode, the maximum velocity increases by 20% as compared to the single-flow setup (Figure 5). In all cases the calculated velocity is designed to impart a target wall shear stress of 1 dyn/cm² on the chamber bottom (at the cell monolayer location). Here, the shear stress distributions are similar for the three cases (Figure 6). As seen in midplane and centerline plots, the shear stress achieves the target magnitude near the middle-third of the geometry. The imparted stress is nearly constant over the midplane, where over 95% (9.5 mm) achieves the target stress within 5% (Figure 7). In addition, nearly 45% (8.2 mm) of the centerline (inlet to outlet distance) attains the target stress within 5%, thus 47% of the total gasket area experiences a shear stress within 5% of the target value (Figure 8). If an 8 mm region of interest is considered, centered on the chamber floor, in order to compare to commercial chambers, it is found that 60% of the region is within 1% of the target stress, over 92% is within 2.5% of the target, and the entire region is within 10% of the target stress. As the allowable range of shear stress (within the target) increases, the area of the gasket that feels this imparted stress increases as well.

For the second gasket, the instantaneous widening and subsequent gradual tapering of the walls yield a unique velocity and stress distribution (Figure 9). As flow enters the wide zone, the velocity dissipates immediately, similar to planar jet flow. However, due to the subsequent tapering walls of the gasket, the maximum velocity is almost sustained along the centerline (within the expansion zone), approaching zero in the outer regions of the gasket. This distribution corresponds to defined gradients of shear stress imparted to the chamber floor, again for a target of 1 dyn/cm². The shear stress reaches a maximum near the entrance to the widening and tapering zones. Between the zones of fluid "expansion" and "contraction" (corresponding to widening and tapering of the flow channel walls), varying gradients of stress are found within the gasket. Focusing on three different locations within the gasket, initially a sharp gradient in stress is found between the gasket center and walls (Figure 10). However, moving away from the entrance to the "fluid expansion" zone, the shear stress decreases at the center and increases near the walls, essentially smoothing the stress distribution. Thus, the shear stress and local variation in stress are dependent on their location along the centerline.

Gasket III is the asymmetric version of the previous gasket design, where only one half of the fluid expansion zone (where the channel area widens) is created. Here, the velocity again appears similar to planar jet flow due to the widening geometry. However due to the asymmetry and the no-slip condition, the velocity decreases sharply as it approaches the "centerline" wall (Figure 11). Again, this distribution causes defined gradients in shear stress on the chamber floor, where the maximum stress is found 0.75 mm from the "centerline" wall (Figure 12). The largest gradient is stress is found near the entrance to the widening channel, where the imparted stress is smoothed with increasing distance from the entrance. However in all cases, a sharp gradient of shear stress is found near the "centerline" wall, providing unique and defined stress distributions within the gasket.

The fourth gasket imposes the most abrupt widening and tapering geometry, where the resulting fluid regime parallels planar jet flow (Figure 13). The maximum velocity is found in the smaller sections of the gasket, near the inlet and outlet, where the magnitude dissipates within the widened flow area. Here, flow near the walls of the widened area has a velocity of nearly zero along the entire length of gasket. This profile causes an increased shear stress near the gasket centerline, with values decreasing with increasing distance from the center. At a location near the entrance to the widened area, the shear stress is nearly constant near the centerline, where a dramatic gradient is found roughly 1 mm from the center (Figure 14). As flow moves away from the entrance, the gradient in wall shear stress (from the center to walls) is decreased and a nearly constant stress is found at the midplane. This gasket geometry induces the largest change in stress between the widening entrance and the midplane, again providing unique and defined stress distributions.

Finally, Gasket V is a variation of the planar jet geometry of the previous gasket, where the length of the widened area is decreased. As a result, the required velocity decreases from that of Gasket IV in the inlet/outlet channels and near the transition zones to the widened channel area (Figure 15). The subsequent tapering of the area causes the change in shear stress between the entrance to the widened geometry and the midplane to be less dramatic than the previous geometry (Figure 16). The maximum shear stress at line 1 is again found at the centerline, however for a 33% reduction in the length of the expanded area (from Gasket IV), the maximum shear stress at this location is reduced by 29%. In addition, the centerline shear stress at lines 2 and 3 are similar to Gasket IV, however the decrease in stress near the gasket walls is greater for this geometry. Thus, the gradient of shear stress at the midplane is greater from the centerline to the wall due to tapering of the flow area.

Bench-top testing of the novel chamber prototype shows improvements, in the ease of use as well as in performance, compared to other commercial chambers (Table 1). The chamber is relatively easy to assemble, using a gasket compression design with four thumb screws, where silicon grease is used to ensure proper sealing and eliminate leakage found in other chambers. However, in conditions where sterility is crucial, assembly is more difficult, and the novel prototype requires comparable assembly times to the commercial chambers. In performance testing, the novel prototype reduces leakage and bubbling, up to flow rate of 60 mL/min which is a rate 10 times higher than that at which the commercial chambers show leakage and bubbling. Overall, the strength of the novel prototype chamber lies in its compression design, which eliminates flow deviations (leakage, bubbles) and allows actual flow profiles to better conform with those predicted computationally.