Real-time results

Table 1. Comparison of DNA microarrays and real-time qPCR.
	DNA microarray	Real-time qPCR
Sample preparation time	4–8 h	1.5 h
Minimum sample volume	4 x 106 cells 50–100 µg of mRNA	1 to 1 x 104 cells 0.01–100 ng of mRNA
Turnaround/data generation time	2 days/sample	1.5–3.0 h/plate
Number of samples per run	1	~40 per 96-well plate ~165 per 384-well plate
Maximum number of targets per sample	500–40,000	4
Cost per sample	$2000–$8000+	$2–$5

Although LSETs facilitate the elucidation of disease-specific gene expression phenotypes and the discovery of targets for drug action, they are low-throughput and extraordinarily expensive (Table 1). A single sample can require days of analysis and cost thousands of dollars, making LSETspractical only for initial discovery applications. For this reason, quantitative PCR (qPCR) and quantitative reverse transcription (qRT-) PCR are much better suited to subsequent studies of drug action and target evaluation.

Quantitative PCR and qRT-PCR rely on the concept that the amount of PCR product present after each amplification cycle is directly related to the number of target copies originally present in the sample. Protocols such as competitive and end-point PCR have been developed to perform qPCR. Although generally successful, these methods are time-consuming and difficult to optimize. Problems arise because PCR reactions become unstable near the plateau phase, and errors can be introduced in the subsequent analysis. The creation of quantitative real-time detection techniques has greatly simplified qPCR and qRT-PCR. The time between sample preparation and the final resultis measured in hours, and individual assays typically cost a few dollars. Errors introduced by other methods are absent in real-time qPCR.The technique uses a fluorescent reporter to directly measure the amount of product in the reaction tube during each amplification cycle. And quantification iscarried out at the beginning of the log phase of amplification, which is the most stable and efficient segment of a PCR reaction.

Detection
Two reporter schemes are used in real-time qPCR, one of which employs intercalating dyes that fluoresce more brightly when bound to double-stranded (ds) DNA. Early work in this area used ethidium bromide—which is approximately 40 times brighter when bound to dsDNA than when free in solution—in the PCR mix, and fluorescence was read after each cycle. More recently, the greater dynamic range of SYBR-green (200:1 fluorescent differential) has made it the dye of choice. Dyes are inexpensive and require only that a primer set specific for the gene of interest be optimized for performance. But there are a few caveats to their use. Dyes report all dsDNA made in the PCR reaction, which makes them unsuitable for reactions that generate side products. And their nonspecific nature means that only one target may be quantified per reaction.

Fluorescent resonance energy transfer (FRET) probes overcome the problems of intercalating dyes. There are several different probe technologies, but all involve oligonucleotides that are designed to hybridize to the DNA sequence amplified by the PCR primers and are dually labeled with a fluorochrome and quencher in close proximity. When free in solution, energy released by the fluorochrome is absorbed by the quencher. But when the probe binds its target, the fluorochrome and quencher are physically separated, which allows light to be emitted by the fluorochrome.

FRET probes have two main advantages. First, because they hybridize only with and report the amount of the specific PCR product, they are particularly suited to PCR reactions that generate side products. Second, by labeling probes with different fluorochromes, several different products can be quantified in a single PCR tube. This is called multiplex PCR and, at present, up to four products can be detected. The caveats to the use of FRET probes are that they can be difficult to design with the proper binding characteristics, they are more expensive than SYBR-green, and optimizing multiplex PCR can be problematic.

Sensitivity
One final consideration is the extreme sensitivity of PCR. Microarrays and other LSETs are inefficient at detecting poorly expressed RNA, revealing small changes in gene expression, and looking at small sample quantities. Quantitative RT-PCR, on the other hand, has been used to detect genes expressed as single mRNA copies, can theoretically detect changes in gene expression as small as 5%, and has been used to examine gene expression in single cells. This sensitivity is important since the sequencing of various genomes will reveal many genes that LSETs cannot detect because of low expression levels or expression in rare cell types.

When working with any of the mRNA-based gene expression techniques, it should be remembered that RNA level does not always reflect the level of protein produced by the cell. Although this correlation generally holds true, some situations have been observed in which the gene is highly expressed but translational, rather than transcriptional, mechanisms control the amount of protein produced by the cell.

Applications
Once severalpotential targets have been identified for a diseased tissue, such as genes expressed preferentially by or mutated within a particular cancer, real-time qPCR is a very efficient method for screening patients to ensure target validity and for assessing target expression over a large patient set. Screening within the patient population is necessary because even the most ubiquitous mutations seen in cancers are present in only a fraction of patients. This is evident from the observed rates of p53 and ras mutations.

A further extension of screening individual patients is the diagnosis of disease subtypes. Some reviewers have speculated about performing patient diagnosis by using microarrays that contain gene markers for hundreds to thousands of specific diseases—a somewhat fanciful notion considering the exorbitant cost of even a small gene array.A patient exhibiting a specific type of cancer does not need to be screened for every infectious and genetic disease known. Real-time PCR is much better suited to these applications.

One can envision small diagnostic kits designed to measure gene expression patterns within a particular disease that would allow determination of subtypes of the condition. This would then allow drugs to be selected that are tailored for that subtype. For example, colon cancer progresses through a number of cell stages that appear to have unique gene expression profiles and varying susceptibility to chemotherapeutics. Rapid determination of disease stage with real-time qRT-PCR will make treatment with stage-specific drugs possible. This may allow more accurate dosage schemes of drugs specific to an individual patient’s disease or reduce the need for multidrug cocktails designed to treat a family of related neoplasms.

The most attractive targets for drug action are those genes and proteins specifically expressed within a diseased tissue. Therefore, real-time qRT-PCR can be used to look for expression of the target within other tissues of the body, as a way to anticipate possible drug side effects. Once targets have been identified by LSETs, PCR can be used to rapidly compare panels of different tissues and cell types. In the search for anti-osteoporosis targets, immunohistochemistry and in situ hybridization experiments showed that cathepsin K was only expressed by osteoclasts. This result could have been obtained more expediently by real-time qRT-PCR.

In the final analysis, real-time qPCR is a valuable tool for use in gene-based drug development protocols. Although it does not replace the use of LSETs in the initial discovery steps of drug development, it complements and augments the utility of DNA microarrays and SAGE by offering a rapid, low-cost method of pursuing interesting findings.

Suggested reading

• Drake, F. H.; et al. Cathepsin K, but not cathepsins B, L, or S, is abundantly expressed in human osteoclasts. J. Biol. Chem. 1996, 271, 12511–12516.
• Hedenfalk, I.; et al. Gene-expression profiles in hereditary breast cancer. New Engl. J. Med. 2001, 344, 539–548.
• Higuchi, R.; et al. Kinetic PCR analysis: Real-time monitoring of DNA amplification reactions. Biotechnology 1993, 11, 1026–1030.
• Lodish, H. F. Translational control of protein synthesis. Ann. Rev. Biochem. 1976, 45, 39–72.
• Velculescu, V. E.; et al. Serial analysis of gene expression. Science 1995, 270, 484–487.