Thread as a Matrix for Biomedical Assays
Original Entry: Peter Foster, AP 225, Fall 2011
Authors: Meital Reches, Katherine A. Mirica, Rohit Dasgupta, Michael D. Dickey, Manish J. Butte, and George M. Whitesides
Publication: Reches et al. Thread as a Matrix for Biomedical Assays. Acs Appl Mater Inter (2010) vol. 2 (6) pp. 1722-1728
This paper details a fairly ingenious application of the capillarity of cotton string. The general idea is that one can think of string like a one dimensional microfluidic device. When one end of the string makes contact with a (water based) liquid, capillary effects will pull the fluid up through the string. Strings can be chemically treated so that they will change their color in the presence of a given chemical and thus can serve as a biomedical diagnostic tool.
Mercerized cotton (diameter = 0.3 mm) was chosen as the thread because of its ready availability and the reasonable rate at which water flows up the string ( ~0.23 cm/s). Three different layouts for the string are shown in Figure 1. In Figure 1 (A), the strings are sandwiched in parallel between two pieces of tape, with one end of each string exposed. They term this the "woven array". Figure 1 (B) shows the "branching array". The design is similar to the woven array, but instead of having multiple strings laid out in parallel, there is a single input that travels to a branching point, from which liquid can travel down other strings to different detection zones. The advantage of the branching array is that because the strings aren't laid out in parallel, one can put the same number of tests on one device with a smaller footprint compared with the woven array. Figure 1 (C) shows the "sewn array". Here, the string is simply sewn through a piece of plastic. Clear nail polish was used in order to seals the holes left from sewing to ensure that liquid can only travel through the string. This advantage of this assay is that the device could be sewn into almost anything (bandages, diapers, etc.).
Figure 2 shows results of using the 3 assays to detect ketone, nitrate, and protein. These were chosen because the presence of any of these three components in urine can indicate problems (high ketone concentration could mean diabetic ketoacidosis, high nitrate could mean a urinary tract infection, high protein could mean kidney dysfunction). To detect protein the detection region was soaked in 250 mM citric acid (pH 1.8) and 3.3 mM tetrabromophenol blue (TBPB) in 95% ethanol, which will turn blue in the presence of protein. To detect nitrate, the detection region was soaked in 2 mg/mL sulfanilamide, 1.7 mg/mL 3-hydroxy-1,2,3,4-tetrahydroben- zo(h)quinoline, and 25 mg/mL tartaric acid in methanol, which turns red if nitrate is present. To detect ketone, the detection region was soaked in 20 mg/mL sodium phosphate, 20 mg/mL sodium borate, 10 mg/mL glycine, a 0.5 μL solution of 20 mg/mL nitroprusside, 30 mg/mL polyethylenglycol (PEG, Mw ) 2000), and 2 mg/ mL poly(acrylic acid), which turns purple in the presence of keytone. A solution containing 1 μM Bovine Serum Albumin (BSA), 1 mg/mL lithium acetate, and 0.2 mM sodium nitrite was applied to each device and Figure 2 shows that for these concentrations, all three assays work well. Color changes could be seen ~10 seconds after the test began, and the maximum intensity was seen ~6 minutes later. Tests with different concentrations of protein, nitrate, and ketone were examined in order to determine if one could discern the initial concentrations of protein, nitrate, or ketone based on the final color intensity in the detection zone. However, it was found that this was not the case.
The final experiments in this paper had to do with carrying out enzymatic assays in beads of polyacrylamide gel suspended on the strings. The idea behind using the polyacrylamide beads has to do with storage of these assays. If the enzymes were stored simply on the thread, then during storage the enzymes would eventually become dehydrated and would have lowered activities. Keeping the enzymes in the hydrated environment of the beads should somewhat prevent this. Figure 3 shows before and after pictures of two different tests, A and B corresponding to glucose detection, and C and D corresponding to alkaline phosphatase. For the glucose test, bead 1 was soaked in 0.6 M aqueous solution of potassium iodide and bead 2 was soaked in a 1:5 horseradish peroxidase/glucose oxidase aqueous solution. A solution of 50 mM glucose was applied and one could see the full color change to red/brown ~15 minutes after application. For the detection of alkaline phosphatase, the beads were soaked in nitrotetrazolium blue chloride (1.5 mg/mL) and 5-bromo-4-chloro-3-indolyl phosphate (1 mg/mL) dissolved in 100 mM Tris buffer pH 9.5. After application of 10 ul of 140 units/liter alkaline phosphatase, the color change to blue could be seen after ~10 minutes.
The devices described in this paper are (in my opinion) extraordinarily clever. In my mind, the strength of these assays lies in the fact that they are made of inexpensive components, being composed chiefly of thread and tape. The low price could translate into high distribution, making devices based on these principles more readily available in underdeveloped countries. The other main strength is that they only need a relatively low sample size, with most samples used in this paper being of order 10 uL. If the assay is used to test blood, the required volume is small enough that a simple finger prick would be sufficient instead of having to draw blood. The main weakness of these tests is their binary nature. They seem to be able to determine if a given substance is present or not present but give no quantitative information on the substance's concentration. This is alright for some tests (e.g. protein in urine), but less helpful in other situations (e.g. glucose in blood). Even with this limitation, this is an exciting idea, and I'm looking forward to seeing the future impact of thread based biomedical assays.
 Reches et al. Thread as a Matrix for Biomedical Assays. Acs Appl Mater Inter (2010) vol. 2 (6) pp. 1722-1728