Design of Side-View Particle Image Velocimetry System for Cellular Adhesion Analysis

Technique of Side-View PIV

Building on the developments of previous work in the study of microflows and cellular adhesion, a side-view flow chamber has been integrated with particle image velocimetry (PIV) to produce a novel quantitative method of studying cellular adhesion experimentally.

Design of Side-View PIV System

The side-view PIV system consists of a square glass microchannel of dimensions 700x550 µm², two 45° mirrors coated with a double layer of Cr (50 nm) and Au (200 nm), a supportive aluminum chassis, a syringe pump, an Olympus IX 71 microscope, and a TSI PIV system.  It is possible to culture EI cell monolayers on the bottom of the microchannel to form an adhesive substrate for leukocyte binding.  For PIV data capturing, 1 µm tracer particles were seeded within the flow at a concentration of 0.0175% by volume and illuminated with a Nd-YAG 532 nm laser light source.  INSIGHT 6.0 software was used to capture sets of double frame pairs, which were subsequently subjected to aggressive processing to eliminate noise and increase the apparent particle density.  Since steady-state flow was considered, ensemble averaging was used to further increase the signal to noise ratio according to Meinhart and colleagues [46].  These methods have made PIV a manageable technique even from the side-view and in a large channel with many out-of-focus tracer particles.

Efficacy of Side-View PIV and Consistency of Experimental Results with Predictions

The side-view imaging capabilities of the side-view PIV system have been demonstrated with the visualization of 16 μm-diameter fluorescent beads on the bottom wall of the channel.  The manipulation of the focal distance enables the focus to move in front or in back of the beads, providing strong support for the ability of the system to generate side-view images.

The accuracy of side-view PIV to describe hydrodynamic activity within the microchannel were established by using the side-view PIV system to calculate velocity measurements near the bottom wall of the channel The theoretical data matches very closely to the experimental PIV-acquired data, providing strong support for the efficacy of the side-view PIV system.

The usefulness of the side-view PIV system to characterize disturbances in the flow at the cellular length scale has been demonstrated by calculating the flow profile over a 16 μm-diameter fluorescent bead and a Jurkat cell.  Furthermore, these velocity vector profiles have enabled the calculation of stream-line plots.

Shear Rate Data

Analysis of cell shear rates provides an excellent demonstration of the quantitative investigation possible with the Side-View PIV System. The shear stress over the surface of the Jurkat cell in was calculated to be 103.4 s^-1 for an upstream shear rate of 51.0 s^-1 (0.765 dyne/cm²) and 417.9 s^-1 for an upstream shear rate of 257.7 s^-1 (3.864 dyne/cm²).  The ratio of the cell shear to the upstream wall shear was 2.03 for the low shear case and 1.62 for the high shear shear case.  Thus, the normalized shear stress over the surface of the cell is decreased by 20.2% when the cell deforms.  This lends experimental support for the computational studies [33-35] that have suggested that the effective drag forces acting on adhering cells are reduced following deformation.

                       Limitations of Side-View PIV in the Study of Cell Adhesion

The primary limitations associated with the use of the side-view PIV system include the low tracer-particle seeding density, limited resolving power of the technology, and large time commitment required for data acquisition and analysis.  The low tracer-particle density is necessary to reduce noise from tracer-particles outside the focal plane.  This volume illumination element, which is a recurrent problem in mircofluidic studies, is usually circumvented by analyzing flow in channels in which the path length of the incident light through the flow region containing tracer particles is very short.  However, any microchannel for use in cellular adhesion experiments must have a width to height ratio that yields a substantial region of the channel useful for predicting the shear based on the incoming volumetric flow rate.  In addition, the 700x550 μm² microchannel is constructed of easily obtained materials and allows for easy culture of endothelial monolayers within.  The solution to the problem of low tracer-particle density was the use of an image stacking algorithm.  However, this eliminates the ability to analyze velocity profiles at discrete time points, hence doing away with the possibility for evaluation of non-steady-state flow.

The limited resolution of the side-view PIV system is a result of its optical properties.  The long working distance 20X objective is necessary for the large focal distance is conducive to use with the 45° mirrors.  Unfortunately, this objective has a rather low numerical aperture of 0.4, thus decreasing the maximum resolving power.  In addition, the mirrors are not perfectly reflective and also only redirect a finite portion of the light cone that is incident on the objective, further reducing the effective numerical aperture and the light capturing ability of the system.  Larger leukocytes were therefore chosen because the large disruptions in the flow that they produce have enabled the calculation of velocity profiles with dense vector spacing.  The side-view PIV system, in its present design, may be unable to calculate the velocity profiles over leukocytes that are 10 μm in diameter or smaller. Although impossible for use in PIV image collection, the visualization of cells using the side-view flow chamber can be greatly enhanced by the use of phase-contrast microscopy, enabling very good visualization up to magnifications as high as 100X.

Extensive time is necessary for acquiring and analyzing data with the side-view PIV system because of the complexity of the methods involved and the fact that all image processing techniques are currently user-controlled.  Thus, the ability for statistical comparisons involving the results from many cells is limited.  Future developments should include software capable of independently performing all processing techniques from the onset of image acquisition.