“Gene chips”, or DNA microarrays, are powerful weapons in the battle to make use of human genomic information.
The mission is clear. A powerful adversary has been created from what was meant to be an ally, and it must now be overcome. The enemy, which has reached monstrous proportions, is brimming with coded wisdom about the human body. The struggle is to tame this beast—namely, human genetic sequence data—and exploit it for all it’s worth. The success of this effort will depend on the weapons that are available and the effectiveness with which they are put to use.
One of the best additions to the arsenal is the high-density DNA microarray, otherwise known as the DNA chip or gene chip. This tool allies microbiology and nanotechnology to provide a means of highly parallel and efficient genetic analysis, replacing the “one gene at a time” approach.
Within a short time, DNA arrays have played a significant role in generating the sequence data that have been attained. Now, they offer the promise of elucidating the data’s functional implications, thus giving rise to boundless potential for drug development and the possibility of more personalized medicine. These high prospects have led to the deployment of many scientists and businesspeople in the public and private sector on the task to develop the technology and bring it to the front lines. In fact, marching orders have come from the U.S. Capitol, where, in his 1998 State of the Union address, President Clinton pronounced that “gene chips will offer a road map for prevention of illness throughout a lifetime.”
Despite this warranted excitement, the technique is still in its infancy. Much remains to be done before it can become a routine weapon in battles against the genomic “monster”.
Choose your weapon
The concept of the DNA microarray experiment is based on the traditional ideas of base pair hybridization. Perfectly matched strands of DNA will bind more tightly than those that have mismatches. Therefore, an expressed gene that has been labeled (commonly with a fluorophore or radioactive tag) can be identified and quantified in a sample by its binding properties to a particular “target” sequence, under conditions that favor perfect matches. High-density chips, in which hundreds or thousands of sequence “targets” are immobilized on a micrometer-sized substrate, make this type of analysis achievable on a very large scale.
Most current DNA microarrays fall into one of two categories, based on the makeup of their targets. In one category, the target molecules are cloned strands of complementary DNA (cDNA), and in the other, they are synthesized chains of oligonucleotides. One method of construction uses robotic arms to “touch and spot” nanoscale droplets of cDNA or presynthesized oligonucleotide solutions onto a substrate coated with a covalently binding substance. A more recently developed technique uses ink jet technology to eject samples without surface contact. In the case of cDNA microarrays, the immobilized spots must be denatured on the substrate to form single-stranded genetic targets.
An alternative to oligonucleotide arrays is concurrent synthesis of the chains directly on the chip. Affymetrix, Inc. (Santa Clara, CA), accomplishes this with a proprietary method that uses photolithographic techniques, in which locations on a glass substrate are selectively activated for chemical coupling by illumination. This process, which is used to construct the company’s GeneChip products, is the origin of the “chip” terminology used for microarrays, because of its similarities to computer chip fabrication (this article uses the terms interchangeably). It has recently been shown that ink jet methods developed by Rosetta Inpharmatics, Inc. (Kirkland, WA), can also be used to construct oligonucleotides on a substrate in situ.
Which approach, then, will predominate as the scientific community’s weapon of choice? In general, the cDNA format naturally contains full-length genetic sequences at each spot, which leads to strong hybridization signals. The oligonucleotide version presents the option of immobilizing long-strand targets or dividing the sequences into smaller sections for sharp, subtranscript-level resolution. Moreover, this type of array allows greater specificity in target sequence design. For example, variations of each sequence can be aligned next to each other. In this way, nonspecific hybridization is explicitly accounted for, and natural differences in the expression of each gene, such as alternative splice variants, can be detected. Specific mutations can also be probed, as discussed in the next section.
In general, comparisons tend to favor the oligonucleotide option for the future of microarray applications, although both methods have thus far performed valiantly in the formidable tasks of understanding disease and developing drugs.
Defining the lines of combat
The applications of DNA microarrays are diverse. Successful battles have already been fought. In drug discovery, two major fronts of action for arrays are gene expression analysis and mutational genotyping, particularly for single nucleotide polymorphisms (SNPs).
The expression level of genes in a sample can be measured by the prevalence of their corresponding RNAs. In a sample used for expression analysis, RNA is isolated, amplified, and, subsequently, reverse-transcribed to cDNA. The final products are then labeled with reporter molecules, such as fluorophores, and reacted with a microarray. Thus, the extent to which each “targeted” gene is activated in a sample can be determined by signal intensity. In this way, overall genetic mechanisms can be observed for cells that
- are found in a particular disease, such as a malignant tumor,
- are involved in an interesting biological process, or
- have been treated with a certain drug.
| Battle-worthy expressions |
| Microarray gene expression experiments are taking a leading role in elucidating the elaborate mechanisms of the human genome.
One prominent study, published in Nature (2000, 403, 503–511), analyzed samples from patients with diffuse large B-cell lymphoma (DLBCL). The study used a microarray called the Lymphochip, which has more than 17,000 cDNA clone targets that are expressed in normal and malignant lymphoid cells. Using an algorithmic approach to cluster the sample by similarities in gene expression, the researchers determined two distinct, differently expressed forms of DLBCL, which subsequently were shown to accurately correlate with the phenotypic treatment responses of the patients.
In another example (Nature 2000, 406, 532–535), researchers screened highly metastatic cells against a collection of custom-made oligo arrays and identified 32 genes that have potential involvement in metastasis.
The effectiveness of using microarrays for studying treatment response was demonstrated by scientists who profiled isoniazid (INH)-treated strains of Mycobacterium tuberculosis (Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 12833–12838). INH is a common antituberculosis drug that has encountered increasing amounts of resistance from mutant strains. Samples from treated M. tuberculosis were hybridized to arrays that contained virtually the entire pathogen genome. Several new potential drug targets were observed, and a "signature" expression profile was generated for the drug's activity, which may have use in the design of new antituberculosis compounds. |
By hybridizing a collection of samples to identical microarrays, highly parallel comparisons can be made. This presents the possibility of generating genetic “signatures” that will allow specific and accurate distinctions between medical conditions (beyond our current standard of diagnosis) and explicit predictions of drug response to be made. In addition, genes can efficiently be identified as potential drug targets. Progress is being made in these areas with cDNA chips and oligonucleotide arrays (see box at right: “Battle-worthy expressions”).
Investigating mutational differences in biological samples represents a different type of approach to the drug discovery process. It involves looking at the subtle variations in particular genes that are crucial determinants of individual disease susceptibility or drug response.
Of particular interest to the medical community is identifying single base pair variations, or SNPs, and mapping them into genotypic classes that predict causal relationships to phenotypes. Millions of SNPs have been entered into public databases. For the genotyping effort to keep up, thousands of samples of interest will have to be analyzed, each for many SNPs. High-density oligonucleotide arrays offer a good alternative for making the process more efficient. By representing each gene on a chip by a set of its alleles, the mutations that are expressed in a sample can be identified.
Another approach that has been proposed for SNP genotyping is to combine microarray technology with other genotyping techniques. For example, single base extension (SBE) genotyping entails reacting cloned samples with marker-specific SBE primers that will bind to specific polymorphisms. Before this reaction, the primers can each be labeled with unique oligonucleotide sequence tags. The complex solution of genotyped products can be hybridized to an array that contains reverse complements to these unique tags, therefore separating the SNPs on the chip. In this manner, a generic array can be used for any set of SNPs that needs to be analyzed.
Arming the troops
Clearly, DNA arrays offer impressive firepower to those who wish to “slay” some aspect of the genome. However, the struggle is far from complete. There are countless potential expression systems to be profiled and SNPs to be genotyped. To meet these challenges and fully reap the benefits of the technology, scientists in a wide range of facilities, from universities to multinational corporations, must gain access to DNA arrays. Several alternatives have arisen in the genomic market for arming oneself with microarray capabilities.
| Table 1 |
| A selection of companies that market microarrays (M), microarray services (S), and array-making instrumentation (I) for expression (E) and genotyping (G) applications. |
|
M
|
S
|
|
| Company |
E |
G |
E |
G |
I |
|
|
|
|
|
|
Affymetrix, Inc.
www.affymetrix.com |
X |
X |
|
|
X |
Amersham Pharmacia Biotech AB
www.apbiotech.com |
|
|
|
|
X |
BioRobotics
www.biorobotics.co.uk |
|
|
|
|
X |
Cartesian Technologies, Inc.
www.cartesiantech.com |
|
|
|
|
X |
CLONTECH Laboratories Inc.
www.clontech.com |
X |
|
|
|
|
DNAmicroarray.com, Inc.
www.dnamicroarray.com |
X |
X |
X |
X |
|
GeneMachines
www.genemachines.com |
|
|
|
|
X |
GenPak Inc.
www.genpakdna.com |
|
|
|
|
X |
GeneScan Europe AG
www.biochip.com |
X |
X |
|
|
|
Genometrix, Inc.
www.genometrix.com |
|
|
X |
X |
|
IncyteGenomics, Inc.
www.incyte.com |
X |
|
X |
|
|
Research Genetics
www.resgen.com |
X |
|
X |
|
X |
Mergen Ltd.
www.mergen-ltd.com |
X |
|
X |
|
|
NEN Life Science Products, Inc.
www.nen.com |
X |
|
|
|
|
Packard Biochip Technologies, LLC
www.packardinst.com |
|
|
|
|
X |
Radius Biosciences, Inc.
www.ultranet.com/~radius |
X |
X |
|
|
X |
Sigma-Genosys Ltd.
www.genosys.com |
X |
|
|
|
|
Stratagene
www.stratagene.com |
X |
|
|
|
|
To satisfy diverse research needs, numerous companies have put multitudes of chips on the market (see Table 1 at right). Interest in broad determinations of gene function and gene relationships is addressed with arrays that attempt to generally cover all aspects of the human genome. And many disease-, function-, and tissue-specific chips are offered for more focused experiments, such as those discussed in the box. Arrays that are designed for SNP genotyping are also available, notably from Affymetrix, which offers platforms for both the specific and generic approaches mentioned in the previous section.
Finally, researchers often find that no chip exists for the information they seek. To meet such needs, several companies sell custom-made chips.
Microarray companies often include peripheral materials with their chips, such as scanners, software for data analysis, and kits for fluorescent or radioactive sample labeling. Others save clients the labor of setting up array experiments by offering services in which samples can be sent in for genotyping or expression profiling, on pre-made or custom-designed chips.
Despite these benefits, a limiting factor for purchasing chips or services is cost. Most arrays can be used for only a limited number of experiments, and some, such as Affymetrix’s GeneChip products, are “one-time use only”. The cost of each array, together with the expense of peripheral supplies and service fees, is substantial. The high costs require a significant investment from individual researchers or small companies that plan to do complex and continuous experiments, especially because numerous repetitions often must be performed.
This situation has improved somewhat over the years because of increased competition and programs such as Affymetrix’s AcademicAccess, which offers discount-pricing arrangements to academic organizations for its GeneChip products. Furthermore, as new, less expensive methods of making arrays are developed, prices will go down. One example is the ink jet method, which is inexpensive and increases the sequence density of microarrays while improving repeatability of spot size and concentration (1). Agilent Technologies, Inc. (Palo Alto, CA), which has exclusive rights to Rosetta Inpharmatics’ ink jet technology, may be one of the first out of the starting block to market chips made in this fashion, although only organizations that have a specific technology-sharing agreement with Agilent currently have access.
Still, purchasing DNA chips is not a comfortable or feasible expenditure for many, particularly academics. An alternative is for researchers to make their own. Although the level of functionality achieved by companies that invest big money in chips may not be easily attainable by “in-lab” construction, the option provides a flexible, economical alternative.
In fact, this approach has been used in many notable published studies, including two discussed in the box (the Lymphochip and the Mycobacterium tuberculosis array). Several companies have taken the route of marketing the instrumentation for constructing arrays (see Table 1). Those who are up to the task could follow the free instructions posted on the Web site of Patrick Brown, a Stanford University professor, for building an arrayer similar to that used in his laboratory (http://cmgm.stanford.edu/pbrown/mguide/index. html). Many instruments on the market, as well as those used by Brown, work by the “touch and spot” method for immobilizing cDNA or oligonucleotides on the substrate, although several companies do offer noncontact ink jet dispensing instruments.
Onward soldiers
More people and organizations are being recruited into the science and business of microarrays than ever before, which translates into significant progress. At the same time, there is a lot of room for improvement in the technique itself, particularly in the areas of sensitivity and selectivity. Microarrays are being widely investigated in an effort to improve such areas as target construction and hybridization detection.
It is not clear which aspects of today’s DNA microarrays will remain and which will change as the technology advances. It is evident, though, that development will help in the fight to conquer the complex “beast” that is the human genome.
Reference
- MacNeil, J. S. Modern Drug Discovery 2000, 3 (6), 71–74.
Further reading
- DNA Microarray (Genome Chip)—Monitoring the Genome on a Chip. www.gene-chips.com; maintained by Leming Shi.
- Hirschhorn, J. N., et al. Proc. Natl. Acad. Sci. U.S.A. 2000, 97 (22), 12164–12169.
- Pease, A. C., et al. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 5022–5026.
- Schena, M., et al. Science 1995, 270, 467–470.
David Filmore is a staff editor of Modern Drug Discovery. Send your comments or questions regarding this article to mdd@acs.org.
