The Center for Alternatives to Animal Testing is an academic center affiliated with the Division of Toxicological Sciences in the Department of Environmental Health Sciences of the Johns Hopkins University Bloomberg School of Public Health.

 

Johns Hopkins School of Public Health

Abstract for TestSmart--A Humane and Efficient Approach to Screening Information Data Sets (SIDS) Data

Thoughts on High Throughput Analysis

Gihan Tennekoon
University of Pennsylvania

The advent of molecular biology has and will continue to revolutionarize the field of toxicology. It will no doubt take some time to fully understand and use all of the information available. However, utilization of the newer methods and information will go a long way in reducing the use of animals through the refinement of the methods and perhaps replacement of some of the tests currently conducted in animals. Nevertheless, validation of the in vitro data will require some animal use. Although my comments will be restricted to high throughput analyses for acute toxicity, similar strategies are equally applicable in ecotoxicity.

Until relatively recently, the promise of molecular biology was tempered by the number of cloned genes and thus many of the studies were restricted to the expression of a few genes. In fact it was felt that generating permanent cell lines with a representative promoter regulating a reporter gene such as luciferase would be ideal as a rapid screening tool. This approach had its appeal but was soon supplanted by newer methods where hundreds to thousands genes could be assessed. In part this is due to the progress in the human genome project which will be completed by year 2003. At present there are 1 million ESTs with about 53,000 unique human genes. The sequence of C elegans just having been completed. In addition to sequencing the human genome, in the next 15 years it is likely that the 3 dimensional structure of almost all of the proteins will be known. Because of this rapid advancement there has been considerable amount of thought given as to how best to handle the information overload, but again progress has been made in this area as well.

An approach to assess tissue toxicity is to develop patterns of gene expression when classes of chemicals or toxins are exposed to the respective tissue including the use of embryonic stem cells to study teratogenicity. These in vitro tests will rely of using organotypic cultures, purified cells from the respective tissue or immortalized cells. Clearly each have their own set of advantages and disadvantages. Since this has been discussed at length over several years I will not enter into this discussion at the present. The question is whether in response to classes of high production volume chemicals or other toxins the tissue in question will increase transcription of sets of genes and the corresponding proteins that will have a predictive value.

The methods available to investigate genes expressed in tissue is done using one two broad categories. One is by direct sequencing of EST sequences or by serial analysis of gene expression (SAGE) and the second broad category uses hybridization methods that include Northern blots, ribonuclease protection assays, differential displays and DNA arrays. It is this latter method of DNA arrays that provides us with the best opportunity to obtain profiles of gene expression in response to different chemicals, and it is this method that requires validation. The DNA arrays utilizes two different strategies, one of which is to spot DNA clones onto glass slides or nylon membranes. At present glass has many advantages that include the ability to covalently attach the DNA, the fact that glass is more durable, non porous, has low baseline fluorescence and the ability to perform simultaneous incubations. The second strategy is to use oligonucleotide arrays and knowing the gene sequence oligonucleotide arrays can be generated. The advantages of the DNA arrays is that it can be routinely performed and repeated many times with reproducible results and aside from high throughput analyses they can be used for mapping. linkage studies, detection of mutations and gene expression. The clonesets can be obtained from Genomic Research or Research Genetics, macroarrays on nylon filters obtained from Genome Systems, Research Genetics and Clonetech and the high density microarrays from Incyte Systems and Affymatrix.

The scenario for testing would be generate total RNA from tissue or cells that have and have not been exposed to a particular chemical. From the two different sets cDNA will be made with one being labeled with cye3 (not exposed to the chemical) or cye5 (exposed to the chemical) The DNA on the arrays is denatured and simultaneously hybridized with the cDNA labeled with either cye3 or cye5 at 42°C in the presence of formamide. Once the DNA has hybridized the slides are scanned using a laser scanner that can detect 10-18 dye molecules in a 100 mM spot. Thus genes induced by the chemical will be detected with cye5 (red) while genes equally expressed under the two conditions will appear yellow (cye3-green and cye5-red) and those genes that are down-regulated will appear green. The goal will be to obtain a recognizable "footprint" of genes expressed by families of chemicals which will help in the screening process. Since many academic and private industries are likely to use this methodology it is imperative that the information be stored in a central place and it would worth the EPA discussing this with the National Center for Bioinformatics that is currently invloved in this endevour. The next step in the refinement of high throughput analysis is to develop a means of taking the information from the DNA arrays and determining from a mixture of cells in tissues which cell makes the relevant transcript. This will require modifying the present methods to develop high throughput in situ hybridization but done will help in validating and refining the present methods of testing. Finally since not all transcripts result in protein changes it very feasible that could be addressed by making a "protein" chip similar to the DNA chips.

An example of the power of this method appeared when DNA arrays were used to assess the response of human fibroblasts to serum. On the chips these investigators used they had spotted 10000 genes and when they scanned the slides there were many immediate early genes that were upregulated and when all of the data was collected and expressed genes clustered not surprisingly it became apparent that the profile of genes expressed were identical to that seen with wound healing. For instance cell division genes were upregulated as were genes involved in coagulation, clot dissolution, chemotaxis, increase angiogenesis and those involved in migration and proliferation of fibroblasts, keratinocytes and melanocytes. We and others are using "RNA profiling" to study cancer biology where a normal cells undergoes transformation to a more malignant cell such a Schwann cell to a schwannoma or a melanocyte to a melanoma.

In conclusion the methods aimed at obtaining patterns of gene expression using the DNA arrays and other complementary methods have immediate applicability as screeing tools and should as such be validated. The information obtained by academic and industrial concerns should be placed and made available to all investigators. The database should be kept at a federal funded center such as the EPA with links to the National Bioinformatics Center. This will most certainly help in the refinement and reduction of animal use.