Neil Anderson and Dr. Adam Woolley, Department of Chemistry and Biochemistry
Biomarkers have a tremendous capacity for use in screening for various diseases, including cancer, diabetes, and bacterial infections. Existing immunoassay methods can detect a broad spectrum of biomarkers. However, they are costly to perform for small numbers of samples and require bulky, expensive instrumentation, making these methods unfit for point-of-care use in the developing world. The focus of my research in the past 10 months has been to help develop a small microfluidic system capable of detecting and quantifying biomarker concentrations with simple, easy-to-read results and that requires little or no technical training before use. Our group has developed and worked to optimize a microfluidic assay platform with promising results. By altering device parameters such as solution viscosity, top layer thickness, and channel depth, we have been able to detect streptavidin concentrations of 1 ng/mL with these devices and are optimistic that continued optimization will yield even lower levels of detection. My primary responsibility during this process has been to fabricate devices for testing.
The overarching goal of the project is to develop a simple, label-free point-of-care microfluidic immunoassay platform with selectivity and sensitivity comparable to ELISA but which does not require any advanced instrumentation or a technically-trained physician.
The design concept used for this platform was a microchannel system made of polydimethylsiloxane (PDMS) in which a fluid sample itself can be used to read the results of the test by measuring with a ruler the total distance the fluid flows through a microchannel. Devices like these could potentially be produced to directly analyze blood, urine or other bodily fluids. My primary research responsibility was to fabricate the various forms and variations of these PDMS immunoassays. The focus of my experimentation was to optimize these PDMS flow-valve devices for improved detection and quantitation. I was also responsible for some device testing and immunoassay mold fabrication. Finally, I was responsible for determining the viability of a single-well device design, which could potentially make sample introduction easier (eliminating the need for micropipets) if this type of test were to be used for blood or urine samples. The other group members and I have demonstrated the effectiveness of these PDMS immunoassays using biotin-streptavidin interactions to simulate the binding of potential biomarkers and antibodies.
We made PDMS devices with 0.45, 0.50, 0.55, and 0.60 mm thick top layers to determine the optimal top layer thickness. The first devices I made were straight-channel devices. Channels were 35 mm long, 58 μm wide, and 6.1 μm deep. I tested these devices by placing 1 μL water in one of the wells at the ends of the channels.
The limit of detection of these straight-channel devices was limited by the short maximum flow distances. Hence, devices with deeper (13-17 μm), serpentine channels were devised that provided a much larger possible flow lengths (120–360 mm) with the same 58 μm channel width. These devices gave better, more quantitative results with lower limits of detection. The majority of the devices I made were serpentine-channel devices.
I also fabricated and tested single-well devices by plasma bonding regular serpentine devices in such a way that the channel ends hung off the edge of the glass slides. The top layers were cut flush with the edge of the slide, leaving a channel open to the side of the device. The resulting devices were essentially the same as the original straight-channel devices, but with deeper channels and only one well. My testing revealed that these devices were largely nonfunctional. Sample introduction was very difficult, and debris from the cutting step frequently blocked the entrance to the channel. In addition, the top layer frequently was peeled away from the bottom (thin) PDMS layer at the device edge during sample introduction, rendering the devices unusable.
After demonstrating proof of concept with straight-channel devices, we began studying and optimizing serpentine channel devices to find lower levels of detection. My contribution was to fabricate hundreds of devices with different top layer and channel parameters to optimize the sensitivity and quantitative ability of the devices. The primary parameters studied were PDMS top layer thickness, and channel depth. Analysis of PDMS top layer thickness shows that thinner top layers (4.5 mm) yield shorter flow distances and improved limits of detection.
Thinner top layers than this result in curling of the top layer and are difficult to work with. Optimal channel depth is somewhat unclear. 13 μm channels exhibit shorter absolute flow distances than 17 μm channels, but with greater scattering of data (R2=0.951 vs. 0.978). By producing devices with 0.45 mm top layers and 17 μm channels, the lowest streptavidin concentration detected was 1 ng/mL.
This novel flow-valve approach is a promising platform for biomarker detection. Many important biomarkers exist in ng/mL concentrations in human body fluids, with virtually all known biomarkers existing at concentrations of 1 pg/mL or greater. Continued device optimization could very well result in devices sensitive enough to detect such low biomarker concentrations.
My testing of straight-channel devices helps demonstrate the viability of the flow-valve assay approach, the lab has obtained promising results, detecting strepatavidin concentrations as low as 1 ng/mL. Future work will further optimize streptavidin-biotin devices by experimenting with different channel depth:width ratios as well as developing devices with significantly deeper channels (16–30 μm). Thinner top layers will also be investigated by using silicone sheets (0.1–0.4 mm) instead of poured and cured PDMS. After determining the lowest quantifiable streptavidin concentration in these flow-valve devices, an antigenantibody flow valve system will be developed, and real biomarker quantitation of hepatitis B surface antigen will be explored.
The results for our research group were recently published: Anal. Chem., 2012, 84 (16), pp 7057–7063.