All the cells in an organism share the same complement of genes in the DNA of their genomes. Therefore the difference in their form and function rely on selective use of these genes, and any given cell is thought to express approximately 20,000 genes from the complete library of an estimated 100,000 in the human genome. Gene expression changes underlie not only development and cellular differentiation, which produces the panoply of cellular variants, but also disease states such as cancer and aging-related diseases. Therefore analyses of gene expression have the potential not only to increase understanding of fundamental biology at the molecular level but are also important to the development of therapies that target actions at that level and so to the future of medicine.

Looking below the surface

The classification of disease states to date has relied predominantly on direct observation of tissue materials by pathologists. For example, the classification of cancers, which is the bedrock of diagnosis and so prognosis and treatment choice, remains dominated by observation of gross cellular characteristics. While the skill and intuition of the pathologist is not to be underestimated, this approach is inherently subjective and superficial. The variation of how individual pathologists will categorize a given sample combines with the fact that cancers with similar histological appearances may have different behaviors. Together, these introduce uncertainty. The analysis of gene expression patterns is inherently quantitative and so not subjective, and because it directly examines the molecular changes which drive the disease state, it looks below the surface.

mRNA - the best link in the chain of gene expression to measure

Gene expression is the process whereby a sequence of DNA, the blueprint, directs the production of a protein, the machine that the blueprint describes. All the work in the body is done by proteins: they are the enzymes that digest your food, the cytoskeleton that gives your cells and therefore you, shape, the receptors that make your brain function and the antibodies that protect you from disease. As the central dogma of molecular biology states: DNA makes RNA makes protein. So what is this RNA intermediate? It is a nucleic acid, akin to DNA, but is a transient, condensed copy of the gene made only when it is time to express that protein. To use an elaborate analogy, one could think of the DNA genome as being kept in a room (the nucleus) for protection. When a given gene is to be expressed, rather than bringing that DNA piece outside the room, copies of it are made and sent outside the room. These copies are called messenger RNA or mRNA, and they direct the synthesis of the protein using the ribosome. Because thousands of these mRNA copies are made and then destroyed when they are no longer needed, they represent an amplified and accurate target to measure of when the gene that produced them is active. Furthermore, because they are nucleic acids they show the same exquisitely specific base-pairing interactions that DNA does. This makes it cheaper, faster and easier to detect them than proteins.

How to measure mRNA levels

Traditionally mRNA levels are measured by Northern analysis, wherein RNA is separated according to size in a gel and transferred to a membrane to make it available for binding by a probe. The probe is a labeled segment of known nucleic acids, derived from the gene under investigation, which is detected on film, and revealing a band of measurable size and intensity. Probes are typically used one at a time, in a protocol that takes 2 days plus up to another week to obtain the desired film record. Although Northerns can be reused up to 10 times or so, this method does not lend itself to high throughput analysis. Furthermore, although it is quantitative it is not highly sensitive.

More recently, polymerase chain reaction based methods have been applied to RNA analysis. The most sophisticated approach currently in use, real-time RT-PCR, does allow highly sensitive and quantitative analysis of mRNA levels. However, this method is also limited in its throughput by the requirement of setting up separate reactions for each probe/template combination.

High-density probe arrayed Biochips offer an ideal solution to the demands of high throughput expression analysis by essentially performing many hybridization based analyses at once. This relies on several significant modifications of the traditional Northern:

• The "probe" is immobilized and not the template. As a result many probes, targeting many genes, can be analyzed in one experiment. In typical arrays, several dozen to hundreds of short DNA probes are bound to a surface and hybridized with a labeled mixture of mRNA molecules. For reasons of chemistry, mRNA itself is rarely used. Rather it is first converted enzymatically into cDNA (for copy DNA) which is more stable, and can be labeled in the process.
• Miniaturization of probe spots allows the number of spots per surface area to be increased to where the throughput rate begins to offset the material and time investment of creating and using the biochip. Miniaturization also provides the benefits of using less of the biological material that is being studied, as well as making the reaction proceed faster.
• Fluorescence detection with its benefits of safety, sensitivity and high range of linear response can now be used, as surfaces the size of a biochip can be interrogated with modified fluorescence microscopes.

High throughput gene expression analysis: the promise

The ability to analyze expression levels of several dozen genes in a given biological sample accurately, rapidly, and cheaply allows a whole new set of questions to be asked, which will bring answers that will lead to new diagnostic and therapeutic approaches to human disease. A good example is the classification of tumors discussed above: in the future diagnoses based on the histologic appearance of a tumor will be strengthened and refined by gene expression profiles. In a recent paper (1) in Science, Dr. Eric Lander and colleagues demonstrated that blinded, de novo classification of blood malignancies on the basis of high throughput gene expression faithfully recreated the categories long established by conventional means. Intriguingly, their analysis also pointed to further sub-classifications which may have prognostics significance. It is highly likely that the same will hold true for other cancers. Furthermore, the wide open extremely high-density analyses that are within easy reach of some laboratories and companies, will identify the restricted set of genes that are diagnostic for a given disease state. This will open the door for smaller, more targeted, niche-chips that can be more easily be applied in a clinical setting as a new standard in molecular diagnosis. These advances in understanding will also significantly accelerate the pace of molecular medicine: treatments based on an understanding of the molecular defects in the patient being considered.

Clemson C3B
©1998-2008 Clemson University. All Rights Reserved.
This page is maintained by the Web Coordinator
Rev 11 Nov 2007