Differential Gene Expression: Technology Provides Quick Access to DNA Data

HOW IT WORKS: This diagram shows how Genzyme's SAGE method flows from one stage to another. In 1997, Genzyme Corp. of Cambridge, Mass., was shopping for new technology. "We were creating a cancer company, to put together a group of technologies that would give us a gene-based approach to cancer drug discovery," recalls Susan Primrose, director of business development at the now-formed Genzyme Molecular Oncology (GMO) in Framingham, Mass. They found a key to their project in SAGE (Serial Analysis of Gene Expression), then a proprietary technology of PharmaGenics Inc. of Allendale, N.J. GMO snapped up both the technology and the company. SAGE, which rapidly and quantitatively measures the differences in gene expression between two samples, quickly became a fixture in GMO's research program. "We believe it will be one of our mainstay technologies," Primrose says. SAGE is just one approach to a broader technique, called differential gene expression, which is attracting increasing attention in the biotechnology industry. Biogen Inc. of Cambridge, Mass., in partnership with CuraGen Corp. of New Haven, Conn., also has invested heavily in the method, while companies such as CuraGen and Synteni Inc., of Fremont, Calif., have developed proprietary technologies to offer as services to other companies or to academic labs. In its varied forms, differential gene expression seeks to answer the timeless question asked by microbiologists staring at cell lines through microscopes: What's really going on in there? The technique gives a quantitative measure of the genes that are expressed in a control cell line, compared to those in a cell system that has been altered in some way. If the messenger RNA (mRNA) isolated from a transformed cell line shows that a particular fragment is suddenly being expressed at two or three times the level in the control line, this difference could be playing an important role in the process being studied. A scientist running an experiment with SAGE, for example, would isolate mRNA from two cell lines-one normal and one cancerous-and then use GMO's proprietary technology to transform the mRNA into fragments of complementary DNA (cDNA), each with a tag of 10 to 14 base pairs in length that uniquely identifies the fragment. After gel separation, the results are scanned into a computer and fed into a database that gives a graphic account of which gene fragments have been up- or down-regulated, and to what degree. This whole-genome approach has the advantage of unearthing novel genes when they suddenly pop up--or are suddenly dampened--in an altered cell. But the information it generates can become difficult to manage. "Given that we don't yet know [the location and identity of all the human genes], many of the genes that show up in these experiments will be new. That's both exciting and frustrating," says Joe Davy, vice president of research at Biogen, which is contracting with CuraGen to handle its differential gene expression technology. INVALUABLE INSIGHT: Temple Smith, director of the Biomolecular Engineering Research Center at Boston University, believes that differential gene expression could give insights into disease by showing what variations of specific genes are expressed by different members of the population. Biogen sends its samples to CuraGen, which has "set up a factory" to handle bulk orders from its clients, says Gregory T. Wendt, CuraGen's executive vice president. The company uses pairs of complementary tags that are five to eight base pairs in length. In the human genome, each tag will occur 500 to 1,000 times. The company's proprietary technology allows it to identify the fragment based on the identity of the tag and the distance between the pair after separation. With 96 unique pairs of tags--each finding 500 to 1,000 places in the genome--CuraGen figures it should have most of the estimated 100,000 human genes covered. Once the fragments from the sample are tagged and separated by gel electrophoresis, the data goes into CuraGen's proprietary database, accessible by Biogen over a secure Internet connection. The technique is a big step up from older gene-discovery technologies, because it gives a better hint at which directions to follow. Classically, researchers in quest of novel genes might have used subtractive hybridization. The researcher would first mix single stranded mRNAs from two samples. Those single strands that are produced in both cell lines find their complementary strands and hybridize to form double strands, which are easily separated from the rest. The single strands left over could only have been expressed by one or the other of the cell lines, not both. That technique has some major drawbacks. "It requires you to make assumptions about when is the right time to look for new genes, and [the hybridization and cloning] is so labor-intensive," Davy contends. "You're making a commitment to large amounts of work with relatively little beginning information, and lots of assumptions." He offers the example of a protein that causes a progenitor cell line to differentiate into insulin-producing (beta) cells over the course of five days. Theresearcher wants to know what's going on genetically, what genes are expressed, and most importantly, when? "Say I want to find an (intermediary) factor that's responsible for (the transformation to beta cells). When, in that five-day period, is the hypothetical factor being expressed? I have to make a decision and propose that it's expressed early on, or in the middle, or at the end--and it's an arbitrary [choice]," he says. In contrast, differential gene expression offers a much more elegant approach. The researcher can sample the cell line every hour, for example, and generate a series of snapshots detailing the up-regulation and down-regulation of genes. "You have the gene expression history of a process laid out in front of you, and you can then decide which genes to focus on before you do sequencing and lots of cloning," explains Davy. The whole-genome approach taken by CuraGen and Biogen is complemented by chip technologies from companies such as Synteni and Affymetrix Inc. of Santa Clara, Calif. Affymetrix produces chips with gene fragments attached. Synteni calls its product a micro-array, but the device works similarly. Messenger RNA or cDNA from samples is labeled, applied to the chip, and allowed to hybridize. The chip may then be scanned by the researcher or sent back to the manufacturer for analysis. Of course, pre-made chips only show data for those genetic fragments they were manufactured to contain, and therefore aren't useful in discovering completely novel genes. But what they lack in flexibility, the chips may make up for in ease of use, consistency, and reproducibility. "Our customers feel more at ease with the chip-based approach," comments Affymetrix vice president Ed Hurwitz, "because they are synthetic and have to pass quality- control standards." Although chips aren't useful for uncovering completely novel genes, they can be used to find novel members of a known class of genes. Hurwitz notes that many members of a gene family may share common domains. The Affymetrix chips use probes of 20 to 25 base pairs in length to fish for the genes of interest. "By selecting probes very carefully, you can identify members of a gene family," he says. "That's not possible with other technologies." PICTURE TOOL: Synteni's GEM tool produces a micro-array for analysis. Discovery of novel genes may have kick-started differential gene expression, but it isn't likely to be the technology's primary use in the next century. For example, companies are already using the technique to probe drug candidates. By measuring gene expression in a cell line before and after treatment with the drug, they hope to get a better measure of its specificity and effectiveness. There are also initiatives under way to test known toxins against cell lines and build a database to which unknown toxins can be compared, according to Dari Shalon, president of Synteni. Temple Smith, director of the Biomolecular Engineering Research Center at Boston University, and one of the initial developers of the bioinformatics division at Incyte Pharmaceuticals Inc., Palo Alto, Calif., believes that differential gene expression could give insights into disease by showing what variations of specific genes are expressed by different members of the population. "We'll have the whole human genome by 2005," notes Smith. "Then one could build a chip to measure variations of the p53 (tumor-suppressor gene). You could then attach probes aimed at the known mutations, then look at an individual's p53 and see which alleles are expressed." That information is already being correlated with known mutations for various kinds of cancer, and when the human genome is completely sequenced, chip technology will allow researchers to rapidly identify complex genetic relationships and novel mutations that may be leading to disease, says Smith. Myriad Genetics Inc. of Salt Lake City, for example, offers its CardiaRisk® analysis of the angiotensinogen gene, which determines whether an individual has the salt-sensitive form of hypertension. According to the company, such patients have more than triple the risk of a heart attack compared to patients with a normal copy of the gene. The company says the test may also be useful for determining the best antihypertensive agent to prescribe. There is much more to come. Within a year or two, Shalon expects differential gene expression to play an important role in pharmacogenetics. Since individuals respond to drugs differently, and many rare side effects may be caused by a genetic predisposition, Shalon anticipates that differential gene expression will be used to track a patient's response to a drug. "A patient may be undergoing an adverse reaction to a drug before he or she is even aware of it," Shalon observes. One day soon, such telltale genetic changes may be caught before they can escalate. James Kling, a science writer based in Bellingham, Wash., can be reached online at http://nasw.org/users/jkling/.

Differential Gene Expression: Technology Provides Quick Access to DNA Data

Interested in reading more?

Become a Member of