Current studies of the central nervous system (CNS) are assigning an increasing number of activities to astrocytes. To better understand astrocyte function, we are studying transcriptional regulation of the human gene encoding glial fibrillary acidic protein (GFAP), the major component of astrocyte intermediate filaments. As the activity of this gene is regulated developmentally and in response to injury, these studies should lead to the identification of signaling pathways operating in these processes. In addition, by identifying the region of the GFAP gene responsible for its specific expression in astrocytes, we can direct expression of other genes to these cells in transgenic animals. Studies in the past year were performed in the areas summarized below. 1. Fine structure analysis of the GFAP promoter. Using studies with cultured cells, we previously found that a 120 bp sequence upstream of the GFAP coding region was sufficient to direct transcription to astrocytes. To further pinpoint the regions required for this specificity, small blocks of bases within this sequence have been mutated. Mutation of any of three different blocks was found to significantly increase expression in a non-astrocytic cell line, while having little effect on expression in an astrocytic cell line, suggesting that specificity is under negative control. These findings are now being tested in mice by using transgenes composed of the wild type or mutated GFAP promoter linked to a lacZ reporter gene. 2. Modification of GFAP transgenes to enhance their utility. One goal is to produce more active GFAP promoters. For this purpose, we have altered our standard GFAP promoter by inserting into it multiple copies of enhancer subregions that we have previously identified. In cell culture experiments, these """"""""super"""""""" GFAP promoters displayed over 100 times the activity of our standard promoter, while retaining cell specificity. They are now being tested in transgenic mice. Another goal is to develop astrocyte-specific transgenes that can be turned on and off by the investigator. Mice have been made which carry the coding region of a tetracycline- or an ecdysone-regulated transcriptional transactivator under control of the GFAP promoter, and RT-PCR has shown that we have several lines that express each transgene. Lines of indicator mice, which carry the transactivator binding site upstream on a lacZ coding region, are now being made to be crossed with the GFAP-transactivator transgenics to test their efficacy. In the course of this work we developed a new method for detecting transgene expression in neurons and astrocytes of mice without having to sacrifice the animals. When the intent is to establish lines of transgene-expressing animals, this method, which involves RT-PCR of RNA isolated from an eye, saves the time and expense of having to first breed the founder mice to determine if the transgene is active. 3. Use of GFAP-driven transgenes for gene therapy. We participated in collaborative studies demonstrating the feasibility of using the standard GFAP promoter to express a tyrosine hydroxylase cDNA for gene therapy for Parkinson?s Disease, and to express the herpes simplex virus thymidine kinase coding region for gene therapy for gliomas. These collaborations are now beginning to test the suitability of using the """"""""super"""""""" GFAP promoters. 4. Role of the GFAP gene in Alexander?s Disease. We have previously shown that expression of a human GFAP transgene in mice produces a pathological state indistinguishable from Alexander?s Disease, a rare but fatal neurological disorder of humans. This result implicates the GFAP gene in the etiology of this disease. Since the sequences of the human and mouse GFAP proteins are not identical, either over-expression of the protein or a mutated sequence could be a candidate cause. We have obtained DNA from several Alexander?s Disease patients and succeeded in using PCR to amplify each of the 9 GFAP exons as well as 1800 bp of the 5?-flanking promoter. Preliminary sequence analyses of these fragments suggest that a majority of the patients were heterozygous for mutations in the GFAP coding region.
