|Courtesy of Gabrielle LeBlanc, NINDS|
Purkinje cells in the mouse cerebellum expressing the calbindin gene
The GenSAT project, named after NASA's LANDSAT, which shows population densities or other features as captured by satellites, will show regions of density of highly activated gene expression as seen through fluoroscopic microscopes. "We expect this to be a tremendous resource for the community, because now they won't have to do this piecemeal on a lab-by-lab basis, but everyone will have a common pool of knowledge that they can look at and analyze how genes are expressed," said LeBlanc, who demonstrated a prototype of the database at the neuroscience meeting. She also described how the database could interact with microarray and mutational analysis for the better understanding of gene function.
LeBlanc noted that the GenSAT initiative is needed because the rush of information coming from the Human Genome Project is just "too daunting to be undertaken by a single lab. In the past, labs interested in a particular gene did a limited sampling to determine where that gene was expressed in the nervous system, but those samples are typically done on too small a scale and in a [non]standardized fashion." With the voluminous amount of information coming from the Human Genome Project, she added, "there is a need to know where these genes are expressed."
Under a $4.6 million contract for the first year, Nathaniel Heintz of Rockefeller University and Gregor Eichele of Baylor University are beginning to collect data on the first 250 genes this year, with the expectation that the information will be ready for viewing by neuroscientists later this year. In subsequent years, the number of genes imaged is targeted at about 1,000 each year. LeBlanc added that if all goes well, this would involve about a $25 million contract for the institutions over a five-year period.
Normal mouse genes will be imaged first due to the high-quality tissue samples that are easily obtained from this species and because the mouse shares many genes with humans. The plan is to eventually include genes from knockout mice that are models for human disease conditions, as well as animals subjected to experimentation (models for drug and alcohol addiction). Samples will be taken in a standardized fashion, "and we'll choose those dissections that are of general interest to the neuroscience community, such as the hippocampus, which is involved in learning and memory, and the substantia nigra, which is involved in Parkinson's disease. Knowing how these genes are expressed and then the disease in which they are expressed will give us a clue to their function," LeBlanc explained.
Imaging Gene Expression
To obtain the images, tissue sections are obtained through standardized planes and examined for the expression of a particular gene using a fluorescent probe. When these genes are illuminated under a fluorescence microscope, the regions in the nervous system where the gene is most highly activated within the nervous system are clearly shown. These fluorescent images are then collected as computer files and placed in the public database for study by interested scientists.
LeBlanc says that the nervous system "offers an unparalleled opportunity" to study gene expression, because it contains an abundance of genes that carry out a wide range of functions. Although the human body may contain 120,000 genes--the exact number is not known--only certain subsets of genes are expressed at any one time. Getting a complete picture of gene expression for the nervous system requires the development of probes for each of the 60,000-plus genes that could potentially be expressed and then using the probes to test for the presence or absence of gene expression, a task LeBlanc characterizes as "too formidable" to be undertaken by individual laboratories.
|Courtesy of Gabrielle LeBlanc, NINDS|
The GenSAT project will map the expression patterns of thousands of different genes in the nervous system
Other databases that house information on specific areas of study also exist. One such database is NeuronDB (senselab.med.yale.edu/senselab/NeuronDB/default.asp), a federally supported neuroscience database that contains information on olfactory genes and proteins. To date, the Neuron DB has more than 1,000 entries. Its purpose is "to aid in the sequencing of this enormous gene family consisting of up to 1,000 members--the largest family in the genome," according to Gordon Shepherd of Yale University. The NeuronDB is now large enough that it has been adopted by the National Institute on Deafness and Communicative Disorders, which also just supplied four years of funding for the database.
Shepherd commented that NeuronDB database users can easily identify other laboratories involved, and the tissues and species in which genes were expressed. An innovation, he added, is a "special database where users can deposit unpublished sequences and run searches to see whether there are other similar sequences still unpublished." In this case, the two laboratories can get together to determine how best to finish the job. According to Shepherd, another database that is in the middle stages of development at Yale will house data on what properties are expressed by a particular cell in specific brain regions, "and then we look at dendrites, to see how they interact with a particular cell."
Other efforts at Yale include the construction of the Cell Properties Database (CellPropDB, senselab.med.yale.edu/senselab/CellPropDB/default.asp) to hold data on membrane properties of different types of neurons. This data identifies synaptic receptors, ion channels, and the neurotransmitters that are associated with each type of neuron, in an initial subset of about 20 neuron types. Shepherd said that this type of database presents some challenges in that the properties being observed are involved in the rapid process of signaling.
Other hurdles, he added, involve how data submission should be automated, either as sequence data directly from authors or as data coming from PubMed, as well as determining ways of providing automated links to citations. Shepherd said that he expects CellPropDB to be particularly useful in the future when maps of the brain containing information on how genes are expressed in a particular region become available.
Also at Yale is ModelDB (senselab.med.yale.edu/senselab/ModelDB/default.asp), a widely used neuronal modeling program created by Michael Hines. Hines and colleagues at the Yale Center for Medical Informatics are working closely with Shepherd's group in constructing the databases and in carrying out fundamental research in this area.
The ModelDB database, Shepherd said, contains the models "that give insight into how the properties are integrated in normal function." This database, the CellPropDB, which assembles data for a single neuron, and the NeuronDB, which assembles data for a certain compartment, can facilitate drug discovery by allowing researchers to experiment with a variety of receptors and channels to discover drugs with more specificity and fewer adverse effects.
Still other databases that deal with the brain are being developed or used, said LeBlanc of NINDS, including "a gene expression database that has data already in the literature," and another that is under development by the European Union. The unique aspect of GenSAT, how-ever, is that it will provide not just data, but actual images. "Most gene databases give sort of a quantitative information, and they're doing large pieces, whereas we're concentrating on more of a cell-by-cell basis," says LeBlanc. S
Jean McCann (firstname.lastname@example.org) is a science writer in Cleveland Heights, Ohio.