Roll-Your-Own Microarrays

CDNA microarrays hold great promise for characterizing disease and performing genetic studies, but they're not exactly an out-of-the-box technology just yet. Often the scientists must prepare their own chips. Yet, this process is limited by the amount of space on the array itself—forcing researchers to make choices about which genes to include in their sample. Clinical microarray applications are further limited by the availability of sufficient cell numbers for testing purposes. Illuminating Pathology: Researchers are using the power of microarrays to advance clinical medicine. In 2000, two groups of researchers took a step towards establishing the clinical efficacy of gene expression profiling when they correlated that data with a clinical outcome in diffuse large B-cell lymphoma (DLBCL) and a phenotype in cutaneous malignant melanoma. For their analysis of melanoma samples, Jeffrey Trent, director of intramural research at the National Institutes of Health's National Human Genome Research Institute, and colleagues used a microarray containing 8,150 human cDNAs—representing 6,971 unique genes—from a Cooperative Research and Development Agreement with Huntsville, Ala.-based ResGen, a subsidiary of Invitrogen Corp. of Carlsbad, Calif.1 In those days, there were a limited number of known genes to select from. "The first thing you want is genes that have a name and activity associated with them, because if one of those is indicated as important, you have something immediate to follow up on," says Michael Bittner, who is a coauthor on the paper. There were few enough known genes available that the chips had plenty of room for more, so the team "threw in expressed sequence tags [ESTs] to fill up the spots on the chip," he relates. "People want to include ESTs in hopes of finding novel genes for long-term study." To decide which ESTs to use, the team used several criteria. They wanted ESTs from chromosome regions that had already been implicated in melanoma or cancer, and they also favored ESTs that appeared more than once in the sequence databases, since such repeat performances suggest that the EST is not just an experimental artifact. Finally, the investigators sought larger ESTs, because larger ESTs tend to give better signals than do smaller ones. The team led by Louis M. Staudt, senior investigator, at the Center for Cancer Research, National Cancer Institute, also constructed a specialized microarray for their research, which they called a "lymphochip." For the purposes of this study—an attempt to use gene expression to classify DLBCLs—Staudt and his team focused on genes that are expressed primarily in lymphoid cells and genes that are known to be important to immunology or cancer.2 The researchers, suspecting that cells in the germinal center may be source cells for the lymphomas, pulled 12,069 of the 17,856 selected cDNA clones from a germinal center library. The germinal center is a structure enriched for B- and T-cells wherein B-cells undergo hypermutation of antibody genes to improve the antibody response. To round out the chip, the team selected cDNA clones from libraries derived from DLBCL, follicular lymphoma, and chronic lymphocytic leukemia, as well as genes that are induced or repressed during B- or T-cell activation, and a set of genes that play a role in lymphocyte or cancer biology. About a quarter of the genes on the microarray were represented by two or more separate cDNA clones; these provided internal controls to ensure the reproducibility of the quantitative results. Strange as it seems, a microarray's size actually limits the number of clinical samples that researchers can use. "They call them microarrays, but they're far from micro," observes Bittner. Even small microarrays are usually about 20 x 20 mm in size, and obtaining enough sample to cover the entire area—usually a minimum of about 10 million cells—can be a problem. In the case of melanoma, it's often difficult to get enough cells from an early-stage tumor, because surgical excision and testing pretty much destroy all of it. "People are working on amplification methods, but if you survey the literature, there are a few publications of such a thing working, but nobody uses the same technique twice. It's a lab-to-lab thing." That problem forced Trent's group to use advanced tumors, but they would like to eventually perform microarray analysis on early-stage tumors as well, because the early developmental stages can have a profound influence on the resulting tumor, Bittner believes. Nevertheless, research marches on: "We're still working on it," he says. Jim Kling (jkling@nasw.org) is a freelance writer in Washington, D.C. References 1.Bittner et al., "Molecular classification of cutaneous malignant melanoma by gene expression profiling," Nature, 406:536-40, 2000. 2.A.A. Alizadeh et al., "Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling," Nature, 403:503-11, 2000.

Roll-Your-Own Microarrays

Interested in reading more?

Become a Member of