Array of Hope for Gene Technology.

Using the same technology that drives inkjet printers, a Kirkland, Washington-based bioinformatics company is developing sophisticated DNA microarrays that may help researchers measure and analyze gene expression faster, more economically, and with greater precision than ever before possible. These small glass slides may revolutionize the field of toxicogenomics, helping scientists target new drugs, discover gene functions, determine biologic pathways, and better understand illnesses such as cancer, cystic fibrosis, and cardiovascular disease at the molecular level.

The FlexJetTM system, as the microarray product is known, was pioneered by the group of scientists who founded Rosetta Inpharmatics--Stephen Friend, Leland Hartwell, Leroy Hood, and Jasper Rine--along with Alan-Philippe Blanchard, who heads Rosetta's FlexJet technology development team. The system combines modern printing technology with DNA synthesis techniques to print tiny arrays of thousands of different gene sequences onto a single glass slide. An "inkjet synthesizer" propels molecular strands of DNA onto the surface of a slide, "printing" arrays of DNA molecules in a process not unlike the manner in which an inkjet printer deposits ink onto paper, forming distinct patterns of characters and images.

Microarrays Demystified

DNA microarrays contain literally thousands of unique DNA sequences, or probes, which are deposited in an orderly arrangement onto a solid substrate such as a glass slide. Each probe corresponds to a DNA sequence within one or more genes. Microarrays give researchers a glimpse at genetic activity within a cell by indicating which genes are being expressed and to what extent. Researchers then analyze the microarrays with sophisticated software to discover, for example, which genes control biologic processes, study toxic responses within cells, or chart molecular pathways to disease. All of this information can point to possible new drugs to treat disease.

Such chips can be made in one of several ways. One method employs photolithography, a method that is similar to that used to manufacture electronic microchips, in which a series of patterned masks and chemical processing are used to build an array of predetermined DNA sequences on the chip, base by base. Other methods of DNA deposition include spray flux, reagent spotting, and the inkjet technique.

To analyze gene expression using such microarrays, mRNA (or messenger RNA) is first extracted from a cell or tissue culture, then converted to short lengths of cDNA (or complementary DNA). The cDNA is then amplified--copied in an enzymatic reaction using polymerase chain reaction, or PCR--and tagged with a fluorescent label. The pool of amplified sample cDNA mixes and binds to complementary probe DNA sequences on the microarray, a process known as hybridization. A laser then scans the array, causing the sample DNAs with fluorescent tags to generate signals of intensity in proportion to their original abundance in the mRNA pool. Later, these hybridization data are analyzed using a sophisticated hardware/software system called Rosetta ResolverTM to reveal gene expression patterns.

FlexJet's History

The FlexJet microarrays had their genesis in work begun in the early 1990s by Blanchard. Working at the California Institute of Technology alongside Hood--whose laboratory was well known for DNA sequencing--Blanchard investigated the possibility that nucleotides, or precursors used to bind DNA molecules, could replace the dyes in an inkjet and thus be propelled through a nozzle onto a glass slide. For decades inkjets had been used to deposit precise amounts of dye onto specific locations on a piece of paper. Why not co-opt that technology, he and his colleagues wondered, and adapt it to the synthesis of genetic material on a glass slide?

In time, the team took their research to the University of Washington in Seattle, where Hood was the founder and chairman of the Department of Molecular Biotechnology. While there, Hood and Blanchard began to have conversations with colleagues at the Fred Hutchinson Cancer Research Center including center director Leland Hartwell and Stephen Friend, who led the center's program in molecular pharmacology. Hartwell and Friend were interested in the question of how arrays could be used to follow the effects of compounds on cells. Friend explains the collaboration, saying, "We combined our interest in pattern recognition and their flexible inkjet technology."

Blanchard soon developed a prototype model to prove that his original idea for inkjet synthesis of DNA was possible. The prototype incorporated an Epson printer head and "was part robust machine and part science project," as Friend puts it. The machine could synthesize long oligonucleotides--or stretches of DNA--on a glass slide with the same accuracy that an inkjet printer could spit ink onto paper.

According to Friend, it took about a year to validate the commercial viability of such a machine, and the FlexJet system was on its way. That done, Rosetta chose not to manufacture the arrays themselves. They instead forged an alliance with Hewlett-Packard's tests and measurements subsidiary, Agilent Technologies, which would assume the role of manufacturing and marketing the array system. (The Palo Alto, California-based company became fully independent of Hewlett-Packard in June 2000.) Since Rosetta's FlexJet arrays complement Agilent's own DNA array technology program, the two companies formed a partnership. "We've been able to shift the know-how, intelligence, and intellectual property to Agilent," Friend explains. "This allows Rosetta to focus on how to interpret the patterns that are projected into the microarray."

According to Wilson Woo, marketing manager for Agilent's bioscience products , his company has significantly developed and enhanced the technology through its manufacturing process.

Flexible, Fast, and Specific

The major advantages of FlexJet, according to Friend, are its flexibility and its specificity. "It allows very specific regions of a gene to capture what's going on in the cell," he says.

Woo agrees. "The advantage is, you can custom-design what DNA goes onto the glass slide," he says. The process of probe selection is software-driven; researchers can program changes into the next phase of the experiment without having to redesign the entire array. This makes inkjet arrays ideal for hypothesis-driven experiments: "You form a hypothesis, do the experiment, find out something different, then you can change the design," Woo says.

Flexibility and reproducibility are two of the reasons the FlexJet arrays hold so much promise. Traditional DNA arrays--in which cDNAs, or gene fragments, are placed onto glass slides by hand--are expensive to make and analyze in terms of the cost of both time and materials. Such arrays require a researcher to presynthesize tens of thousands of fragments and then put them on the slides. And any researcher who wants to tweak the experiment a bit or change one of the sequences must start over. "It's expensive to put down a different set [of gene fragments]," Friend adds.

Photolithography offers an improvement over that method, but it too has limitations. For each nucleotide, four chromium masks must be made. "You have to make a mask for each new design process," Friend says, and that can be time-consuming.

By contrast, the inkjet arrays offer "fast turnaround with custom DNA," says Woo. Fabricators can go from design to array within a few days. Depending on the format, size, and labor required, Woo says, the arrays can cost anywhere from a few hundred to a little over a thousand dollars each.

Another advantage this technique offers is noncontact printing: the printer head applies droplets of genetic material to the substrate surface without actually touching it. This provides a more uniform shape than pin-spotting, a technique in which a tiny pin touches the glass to apply the cDNAs. This benefit, contends Woo, is critical for achieving reliable data analysis. Pin-spotting is also slower and consumes more material, whereas inkjet arrays require only a very small volume of genetic material. "We can print five times more microarrays using the same amount of genetic material," Woo says.

Drug and Disease Discoveries

Drug discovery promises to be one of the most useful applications of the FlexJet system. With fewer samples and less testing, it is now possible to build up a database of known areas within genes that are responsible for certain functions, an advantage that promises to be very useful in making new drugs. Inkjet arrays have been used to analyze the effects on DNA of compounds from cadmium to phenobarbitol, pesticides to heavy metals.

Working as a molecular geneticist in the early 1990s, Friend became aware of lots of discoveries being made about genes that are important to disease, but, he says, "they were so slow to reach the patient." That didn't sit well with Friend, who is also a pediatric oncologist. Friend says the force that propels Rosetta's work is the quest to use this technology to benefit patients.

The same motivation drives Jerry Radich and his fellow researchers at the Fred Hutchinson Cancer Research Center--a center known for pioneering cancer treatments, including the first bone marrow transplant. Radich's team, which works in the Clinical Research Division of the center's Program in Genetics/Genomics, is using FlexJet microarrays in looking for genes involved in the progression of chronic myeloid leukemia (CML) to pinpoint better treatments for leukemia patients.

According to Radich, the disease progresses from the chronic phase into an accelerated phase, and then into the blast crisis phase, the terminal phase of the disease during which it has the greatest resistance to treatment. It is critical to administer chemotherapy well before this final, acute phase. However, the first phase--the chronic phase--can last anywhere from 6 months to 15 years, making it difficult to predict when a patient is about to progress to the next, more severe phase of the disease. So Radich and his colleagues want to compare the expression of genes in the chronic phase with those in the blast crisis phase.

"We don't understand what drives the molecular clock in CML; no one understands the genetic progression during the chronic phase," Radich says. He believes that DNA microarrays may give insight into other malignancies, as well as guide future medical diagnostics. He hopes one day to have a genetic test that can be administered during the early phase of CML to show whether a leukemia patient will be likely to experience a shorter or longer chronic phase. This information would help physicians time chemotherapy treatments for maximum effectiveness.

So far, Radich's preliminary data have shown clear changes in the gene expression profiles that occur relatively early in the progression of the disease, results he calls "extremely promising." Using the FlexJet microarrays, he has compared samples from approximately 20 chronic-phase patients with samples from 10 blast crisis-phase patients. He plans to analyze and compare samples from 100 more patients to ensure that the results gained are statistically robust.

Transforming Science

The dizzying possibilities offered by such sophisticated microarrays call into question some long-accepted conventions in the world of science. In the past, researchers have tended to conduct experiments, derive conclusions, then go on to the next project, sometimes without saving data that might prove quite valuable to the next researcher down the line. Friend believes scientists haven't yet learned how to connect with each other to create core databases accessible to others who are interested in similar questions. "There needs to be a coordinated effort. Scientists have not found rewards sufficient to make them eager to link together," Friend laments. The unfortunate result, he says, is that "a lot of the power of technology gets reduced simply to being a separate finding."

Still, Friend says, using arrays in general is transforming the way scientists approach their research. "Biology has been done by asking very specific questions about `my favorite gene,'" Friend says. Asking such specific questions yields information about one small area of interest. Instead of performing experiment after experiment--turning up one genetic clue at a time--scientists can capitalize on array technology to yield a whole set of potentially meaningful genetic clues at one time.

"Arrays have the ability to sense the whole genome, and get [feedback] on all the information in a cell," Friend explains. "This allows you to see unanticipated or unwanted [cellular] effects. This is a significant advantage over the old methods of analyzing compounds and patient samples."

While it is certainly impressive to have the capability to learn so much about one cell in one fell swoop, the questions arise whether at some point so much information is too much, and how much of it is truly useful. Friend admits that making sense of the overwhelming quantity of gene expression data obtained from a single microarray does present a dilemma. "There are 50,000 dimensions of information for one cell," Friend concedes. "The challenge will be in developing tools that allow one to extract that information."

Radich echoes this concern: "How do you make sense of the data? How do you extract the information from the noise?" Still, he is enthusiastic about the promise the FlexJet microarrays show in his research on leukemia treatments. "It's changed the way we look at things," he reflects. With the microarray technology, scientists not only find answers to their initial queries, but they also find patterns in genes that they weren't looking for at all. Such discoveries could lead down new roads toward altogether new destinations and yield important answers to questions scientists haven't yet thought of posing. "The potential is fantastic," Radich says. "We'll probably understand what the patterns of ... genes mean before we actually find out how they work together."

Suggested Reading

Blanchard AP, Friend SH. Cheap DNA arrays: it's not all smoke and mirrors. Nat Biotechnol 17(10):953 (1999).

Friend SH. Genomic approaches to drug discovery. Adv Oncol 16:2-11 (2000).

Friend SH. How DNA microarrays and expression profiling will affect clinical practice. BMJ 319(7220): 1306-1307 (1999).

Hughes TR, et al. Functional discovery via a compendium of expression profiles. Cell 102(1):109-126 (2000).