DNA-chip technologies, Part 2: State-of-the-art and competing technologies

DNA microarray product development is proceeding with one underlying principle: fast is not fast enough.

By: Cliff Henke

This is the second of a three-part series on the emerging technology. If you haven't already done so, you might like to begin with Part 1 [5] of this series.

Also in this article:

For those who marvel at how quickly DNA chips have burst upon the scene, prepare to be astonished further. Most companies report rapid progress, and suggest that as more is learned and applied, the pace of development in all fields of genetic-related diagnostics and therapeutics—including DNA microarrays—will continue to accelerate.

Labchip by Caliper Technologies (Palo Alto, CA), showing photolithographically etched fluid channels (in red). Channels are typically 70 µm wide and 10 µm deep. Photo Courtesy Caliper Technologies

There is no better example of this than the Human Genome Project (HGP). When the project was launched in 1990, government officials viewed it as perhaps a 20-year project. Proponents compared the vision of the project to that of America's goal of landing a man on the moon in the 1960s—but conducted with international cooperation and over an even greater period. Almost as soon as the HGP was launched, the date for completion was revised to 2005, then to 2003, thanks to developments in computerization, telecommunications, and molecular biology. Today, because of DNA chips and related developments, some officials now believe that the HGP can be completed by the end of this decade—three years ahead of even the revised schedules. In turn, as more is learned in this massive effort, it will accelerate DNA-chip development even more, observers say.

This is the second installment of a three-part series on DNA-chip technologies. The first [5] reviewed the theoretical underpinnings of the field and examined the market forces driving product development. This installment will look at the state of the art and the various competing technologies in this embryonic field. The final article [6] will cover the challenges to commercialization facing companies engaged in this new marketplace as well as prospective near- and long-term applications for these technologies.

Fast-Moving State of the Art

"The outlook is changing fast," says Deepak Thakkar, manager of microarray products at Genometrix (Woodlands, TX). "As an industry, we have largely been in a hibernation period, but in the past 18 months we have introduced a lot of highly targeted products based on very specific needs."

As an indicator of how rapidly the pace of development has proceeded, consider that the first DNA chip, by Affymetrix (Santa Clara, CA), was introduced only two years ago. Today two dozen or more firms are actively engaged in developing microarray technologies, and many others are developing related technologies for sample preparation and analysis.

"Until recently, most of this development work has been centered on what I would broadly call detection," says Peter Wilding, PhD, director of clinical chemistry and professor of pathology and laboratory medicine at the University of Pennsylvania (Philadelphia). "The vast majority of the products that have been introduced, however, are sold for highly specialized applications, usually in research-laboratory or drug-development applications."

Such applications usually involve some genomic-related problem. For example, the gene p450 and its eight different expressions, or mutations, have been linked to various cancers. Researchers are examining exactly which forms have the closest link to these cancers. In turn, drug developers want rapid genomic results so they can quickly determine the effectiveness of candidate therapies.

The reason that research-laboratory and drug-development applications are being developed is simple: that's where the money is. Governments, drug companies, and universities have poured billions of dollars into developing DNA chips for applications related to their agendas.

Take governments. The Defense Department's Advanced Research Projects Agency (DARPA) and the Commerce Department's Advanced Technology Program (ATP) at the National Institute of Standards and Technology (NIST) have invested millions of dollars as seed money for developing both various kinds of chips and their basic manufacturing technologies. "All the major players in this field have received research money from ATP," says Uwe Müller, director of advanced technology at Vysis, Inc. (Downers Grove, IL). "Stan Abramowitz and NIST should get a lot of credit. The ATP is a government program that is actually working, even with minute funding levels, and its support of this technology is the envy of the world."

DARPA has contributed millions more. In its case, DARPA funds projects for "dual use," meaning they must have both civilian and military applications. The prospective military use of DNA-chip technologies is for advanced detection of battlefield biological and chemical weapons.

Private-sector sources of funding have followed with billions more, initially from venture capitalists, and later from strategic partnerships, institutional investors, and public stock offerings. Yet these sources of funding have made pursuit of short-term financial objectives even more urgent, several observers point out.

As an example of this, Müller recalls: "Before Vysis was spun off from Amoco Technology, we were working on a food diagnostic application for E. coli testing; but we put aside that project. We realized that the food processing industry is very competitive, and the low profit margins for the tests used in that industry would not bear the heavy costs of research and development for DNA chips."

Today, such economic inducements are readily apparent. Almost every major manufacturer of DNA chips has a strategic relationship with a major drug company, taking advantage of the eagerness of pharmaceutical manufacturers to find ways to screen potential drugs faster and more accurately. These relationships usually consist of "early access partnerships," in the words of Lewis Gruber, president of Hyseq (Sunnyvale, CA). "This means that partners have access to a company's technology, often for specific applications only, before others do."

"If these technologies succeed in giving pharmaceutical companies the ability to screen drugs quickly and accurately, the payback will be tremendous," agrees Thakkar. "Only one drug in 10,000 succeeds in the marketplace and it takes 12 to 15 years to develop each new one. If these technologies can reduce the length of development time to eight years, that translates into significant savings for these companies."

An example of such corporate relationships is the set of strategic agreements announced earlier this year by Affymetrix and Eos Biotechnology, Inc. (South San Francisco, CA). The agreements allow Eos broad access to Affymetrix's custom and standard GeneChip expression chips for Eos's molecular genomics research efforts in specific fields of cancer, inflammation, and cardiovascular disease.

However, some analysts, including Peter Wilding, believe that such drug and research partnerships are stifling innovation that could lead to larger markets. "If the use of chips remains an activity that is confined to research laboratories—because you need their equipment to prepare samples and do the analysis—then the markets for DNA chips will be limited," he says.

Such thoughts are leading companies into the next phase of microarray development, which is already well under way. A number of companies are developing technologies that use immunoassay reagents or amplification techniques such as polymerase chain reaction (PCR) to prepare samples right on the DNA chip. Companies are also exploring technologies for conducting sample analysis on the same chip and outputting results to a single small instrument. Taking small sample amounts and moving them through the various stages of sample preparation, detection, and analysis of test results is being made possible by advances in capillary electrophoresis. Using electrical charges and micro- and nanoscale structures, researchers have successfully demonstrated that it is possible to perform all of these stages on a single DNA-chip system.

"We think we can have such a product in the marketplace by the third quarter of 1999," says Cris McReynolds, director of business development for Cepheid (Sunnyvale, CA), a maker of sample-preparation and analysis equipment designed to be integrated with DNA-chip technologies. Others echo his comments regarding their firms' product development timetables.

Types of DNA Chips

There are three basic types of DNA chips. The first and oldest is the sequencing chip. This is also the type most commonly discussed in popular articles about this technology. With sequencing chips, such as those initially produced by Affymetrix or Hyseq, segments of DNA (usually 20 bases long) are placed in a microarray. Target samples are then introduced to the chip and the segment that the sample "sticks to" (or hybridizes with) determines the result. This design is called sequencing by hybridization (SBH), and is both an industry term and an intellectual property of Hyseq, says Gruber. Many other companies are now producing sequencing chips, most using the SBH approach. But whatever their technique, such products are intended to determine the DNA sequence of the sample.

The second variety of DNA chips is known as the expression chip. These are designed to determine the degree of expression of a certain genetic sequence by measuring the rate or amount of messenger ribonucleic acid being produced by the target gene. This is done by creating chips with a specific set of base pairs (as opposed to sequencing chips, wherein every possible base-pair combination is arrayed). Results are then compared to a reference or control, and the degree of change is noted. These chips are useful in diagnosing and treating diseases linked to particular genetic expressions, such as some forms of cancer. Vysis and Synteni (Fremont, CA) are two companies engaged in marketing expression-chip-based products and services.

The third type of chip is devoted to comparative genomic hybridization. It is designed to help clinicians determine the relative amount of a given genetic sequence in a particular patient. "A certain amount of unusual genetic expression is normal, but it becomes a cancer out of control only when the level of expression reaches a dangerous level. As an extreme example, many breast cancer tumors—particularly at the end stages of the disease—are so violently aberrated genomically that they don't even have 23 chromosomes anymore," Müller points out. "This type of chip is designed to look at the level of aberration." This is usually done by using a healthy tissue sample as a reference and comparing it with a sample from the diseased tumor.

Design Goals

In the first installment of this series we reviewed the types of technologies companies are using to manufacture their products. While those basic technologies have not changed, rapid progress is being made in all of them to increase density (the number of arrays per chip) for the next generation of products in development. Whether via robotic deposition or microlithography, development is proceeding in pursuit of two basic design goals: increased sensitivity and reliability, and systems integration.

In the case of the former, the "holy grail is to get better discrimination," says Lance Fors, president of Third Wave Technologies (Madison, WI). "The key to better discrimination is improved signal-to-noise ratio." This is especially important since so many DNA chips depend on PCR-based amplification of the sample, which can undermine discrimination. "A challenge facing us is to find a tumor cell amidst a thousand healthy cells in a biopsied sample; it's like finding a needle in a haystack," he says.

Third Wave and other companies are working to reduce needed sample size by decreasing the amount of reference sample, thus reducing the amount of PCR amplification needed. The reduction of the amounts needed is being accomplished through advances in microfluidic technology. For example, capillary electrophoresis breakthroughs and nanoscale fabrication development have enabled companies to reduce needed reference and target sample sizes to microliter and picoliter scales.

In addition, says Wilding, piezoelectric charges applied to these tiny capillary tubes can move samples from preparation to reaction on a single chip. This achievement has laid the groundwork for the second goal, systems integration. "To do this you must have a microfluidics platform," he says.

"Systems integration is no longer a pipe dream," Wilding continues. "Five years ago an industry survey identified a dozen or so companies developing chips, nearly all of which were involved in detection applications. Three were involved in reactions, and none were doing sample preparation. Today nearly everyone is involved in all of these applications, and all have recognized the importance of an integrated approach."

An example of this work is the collaboration among the U.S. Department of Energy's Argonne National Laboratory, Engelhard Institute in Russia, Motorola, and the Packard Instrument Co. This consortium is working on a microarray-based system that would perform sample preparation (using PCR amplification) and reaction on a single chip and within a single analytical device.

"I think you'll see integrated systems that will address some key markets in clinical chemistry sometime in 1999," predicts Thakkar. He agrees with Wilding that these will be integrated systems designed for ease of use.

Ultimately, within the next 10 years, Wilding and others predict, DNA chips and integrated processing will penetrate the emerging point-of-care market. To achieve this, however, myriad commercialization hurdles must be overcome.

Conclusion

DNA-chip development is proceeding at a pace so rapid that it surprises even the most optimistic members of this fast-emerging industry. "The question 'Will DNA chips succeed?' is a dumb one," notes Wilding. "The only real questions are how they will develop and how quickly."

As companies look to market products outside the research-laboratory and drug-development environments, they will face enormous regulatory and market challenges. Chief among the regulatory challenges is developing a set of industry standards for quality assurance. Today each company has its own. "This will involve some of the big companies like Affymetrix taking the lead, and it will require a real partnership between FDA and all the key players in the industry," says Thakkar.

The biggest marketplace hurdle will be to introduce products at prices that are competitive with existing technologies, says Wilding. "In today's reimbursement climate for health care, products must be better as well as cheaper," he points out. These are issues that will be discussed further in the third and final installment of this series, which will discuss future developments.

Cliff Henke is a freelance writer based in Southern California.

Cepheid bridges the sample-volume gap

Kurt Petersen

Regardless of the details of any particular technique, diagnostic applications of DNA-chip technologies must contend with one overriding and frequently overlooked requirement: they must be sensitive to low-concentration target analytes. Modern DNA diagnostic assays face mounting demands to detect organisms or DNA mutations at very low concentrations, often less than 100 copies per ml, in raw biological samples such as blood or urine.