January 1, 2001; pages 14 A-15 A
Copyright 2001 American Chemical Society
RESEARCH PROFILES
Chip-based mosaic immunoassays
For immunoassays, the perpetual goal is to increase throughput by reducing the amount of reagent binding. In the current issue of Analytical Chemistry (pp 8-12), and Bruno Michel at IBMÅfs Zurich Research Laboratory (Switzerland) describe a new method for doing just that. They call their approach microfabricated mosaic ("micromosaic") immunoassays and report that the assays consume only nanoliter quantities of reagents and perform incubations within a few seconds or minutes.
The works draws upon an ongoing effort, led by Michel, which is centered around microcontact processing to develop microfabricated devices. The researchers abandon microwells, which are traditionally used for immunoassays, in favor of a flat poly (dimethylsiloxane) substrate, onto which they deposit narrow stripes of antigen using a network of microfluidic channels fabricated in silicon. Then, a second network of microchannels, perpendicular to the first, delivers the solutions to be analyzed. (The researchers use simple capillary forces to induce flow in the channels.) Binding the analytes results in a mosaic pattern of tiny squares--similar to the grids of dots seen with DNA microarrays--which can be analyzed with a fluorescence microscope.
Researchers have long recognized the value of such a microarray-type format for immunoassays. However, despite some recent success in this area, protein-based assays have generally been more difficult than DNA-based assays to miniaturize and integrate into smaller, highly sensitive, practical formats, Delamarche says. For example, gridlike patterns of DNA oligonucleotides can be "built" on a surface using photolithography, but the same is not possible for proteins, because they can be composed of hundreds of amino acids. (Antibodies generally have about1400 amino acids.) In addition, proteins can easily lose their three-dimensional structure when handled under "denaturing" conditions.
The key to making the microfabricated device work was the ability to deposit the proteins with a high resolution, says Bernard. To do this, researchers have used various approaches, including inkjet printing, drop-on-demand techniques, microcontact printing, and soft lithography, which is what the IBM team chose. "What is really pleasing to see," Delamarche says, "is that soft lithography not only has the potential to help microfabrication, but it can handle fragile proteins and pattern them on a surface."
In preliminary experiments, the sensitivity and reliability of the new approach compared well with those of traditional immuno assays, the researchers say. In the simple case of two-step assays (antigen-antibody pairs), binding and detection occurred over a concentration range of <1-1000 nM. More sophisticated sandwich-type assays (such as enzyme-linked immunosorbent assays) could also be performed, and in these cases, the signal intensity could be scaled with the amount of analyte present.
Another encouraging finding was that dilute solutions of proteins could be used. Proteins from the solution are deposited onto the surface of the substrate as the solution flows inside the microchannel, Delamarche explains.
The challenge is that solutions of proteins are typically dilute, and the volume is typically small--possibly less than a nanoliter, depending on the geometry of the channel. To ensure that enough proteins flow through the microchannel to entirely coat the exposed region of the substrate, the researchers place a spongelike flow-promoting pad at the far end of every microchannel. As the dry pad soaks up solution, it helps draw fluid through the channel.
The short coating and binding times seen for this device are a result of the small channel dimensions, which prevent mass transport limitations. "This point is really important, and most people donÅft realize it," Delamarche says. People doing conventional assays in microwells typically wait 30 min while the proteins slowly diffuse in all directions, including toward the surface, but inside the deviceÅfs 10-nm-high microchannels, the diffusion path of the protein from the solution to the surface is considerably smaller. Most of the proteins will diffuse to the surface of the substrate within seconds and adsorb. "By performing both the patterned delivery of antigens and antibodies to a surface using microchannels, we make an important economy of solutions, of proteins, and time," Delamarche adds.
The researchers note that no blinded experiments were performed (although, in some cases, unexpected cross-reactivityÅ\in which antibodies from one species recognized and bound to antibodies from another species--was observed). The reason, Delamarche explains, is that "we were not comfortable developing this work on unknown ground." However, now that the preliminary work is finished, the team will start using real-life samples to better evaluate the diagnostic potential of the technique.
-- Sandra Katzman