|Arrays Arrive in the Brain|
|Profiling the molecular response to seizure in the mouse brain|
Edward R. Winstead
September 29, 2000
Seizures almost always result in the death of brain cells, but the extent of the damage varies greatly among individuals. Certain genes switch on in response to injury, and this activity can help brain cells survive. Experiments in the mouse show that responsiveness is inherited. Genetically diverse mice, for instance, can experience the same drug-induced convulsions for the same length of time in a controlled environment and have very different outcomessome may lose relatively few brain cells compared to others.
A handful of human genes that offer modest protection following seizures have been identified, but molecular responses to trauma are likely to be varied and complex, involving different genes in different people. As a prelude to identifying human genes, scientists recently added a genomic twist to the mouse seizure experiment. Researchers in California compared the behavior of genes in the hippocampus region of the brain between two strains of mice prior to and following a seizure. Most seizures in humans occur in the hippocampus.
The seizure experiments were part of a broader study comparing the activity of some 7,000 genes in six regions of the brain between two genetically diverse strains of mice. An analysis revealed clear region-specific differences in the activity of genes, but the differences involved fewer than 100 genes.
"We are very happy that so few genes were different," says Carrolee Barlow, of the Salk Institute for Biological Studies. "We didn't know what to expect, but many people thought we'd find thousands of differences between the regions." She says this result validates an approach that was unproven in tissue as complex as the mammalian brain.
Gathering and interpreting meaningful data on thousands of genes requires expertise in mouse genetics, anatomy, and surgeryas well as in microarrays. Gene microarrays are slides arrayed with DNA that record the activity of thousands of genes simultaneously. Researchers recently used the technology to make finer distinctions between forms of cancer based on tumor gene expression. (See The Evolving Art of Arrays)
Barlow has expertise in brain disorders but not in microarrays. She teamed up with David Lockhart and colleagues at the Genomics Institute of the Novartis Research Foundation. Lockhart has consulted for Affymetrix Inc., which builds gene microarrays, and he wanted to test their tools in the brain.
"Microarrays give you a global perspective, but they also give you specific genes that might be involved," says Lockhart. "This allows you to form hypotheses about how the biology works, which is what you really want to know." The study profiled gene activity in the cortex, hippocampus, amygdala, entorhinal cortex, midbrain, and cerebellum.
For the test drive of the technology, the researchers focused on a well-studied phenomenon in the brain. Considerable research has been done on genes and seizures, and microarray results were compared to these data. Virtually all of the known genes that were expressed differentially turned out to have previously defined roles in the central nervous system, according to a paper describing the research in the current issue of the Proceedings of the National Academy of Sciences.
"This study is state of the art," says Daniel H. Geschwind, of the University of California, Los Angeles, School of Medicine. "Neurobiologists probably will do the same kind of study in the years ahead, and these researchers have provided a road map of sorts." Geschwind, who wrote a commentary on the research for Proceedings, notes, however, that the results reflect the limitations of a relatively new technology.
Size and sensitivity are two limitations of current mouse microarrays. Because only a portion of the mouse genome was represented on the array, the true number of genes that were differentially expressed in at least one region of the brain between the two strains is greater than 100. "But if we extrapolate based on size estimates of the mouse genome, the number is still less than 1,000," says Barlow.
Current microarrays are also limited in their ability to detect genes that are expressed at very low levels. "Clearly some genes were slipping under the radar," says Geschwind.
Some genes were meant to slip under the radar, says David Lockhart. For the purposes of the study, the researchers used a conservative definition of changes in gene activity. The decision was a strategic one. "At this point we really didn't want to get it wrong," he explains. "This is the first study of this type, and we are laying the foundation for more comprehensive studies with additional strains, additional points in time, and more genes."
Based on the rigorous standards and confirmation of results through traditional technologies, Lockhart and Barlow are confident that the genes they say changeda total of 73really did change. Geschwind reviewed the results, and he shares their confidence. "Seventy-three is very likely to be the right number," he says. "The data are strong and very carefully analyzed."
The strains of mice, C57BL/6 and 129SvEv, have well-characterized behavioral differences associated with the brain. The mice differ not only in seizure response but also in anxiety levels, learning skills, and alcohol tolerance. Mouse strains are essentially populations of identical twins, and Geschwind notes in his commentary that inbred mouse strains "may offer the most powerful route to understanding the influences of environment on gene expression, because genetic variability is controlled."
Carrolee Barlow emphasizes that understanding why individuals differ in their response to seizures requires looking at sets of genes. "Collections of genes working together rather than single genes are what make the difference," she says, noting that these kinds of studies can't be done in humans. The microarray data suggest that from a researcher's perspective, the mouse brain is a manageable model system. "We're looking at the molecular structure rather than the three-dimensional structure. And yes it's complicated, but it isn't too complicated for us to figure out."
. . .