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

Proceedings for TestSmart -- Endocrine Disruptors

Breakout Group VI - Application of Genomics and Proteomics to EDs

Chair

Bernard Robaire
McGill University

Panel Members

Kevin Gaido
CIIT

Barbara Hales
McGill University

Overview, Bernard Robaire

  1. What are genomics and proteomics?
    Genomics and proteomics refer to methods for analyzing gene expression profiles. Genomics generally refers to the study of DNA or RNA expression in cells; proteomics refers to the study of protein expression in cells. The logic underlying these approaches derives from the following biological paradigm:

    Figure 1

    Sequencing of genomes (human, mouse, rat, Drosophila) gives us a map, but doesn't tell us where to go.

    Another critical issue is understanding how genes are turned on or off. Methylation and acetylation of DNA are important mechanisms for regulating transcription of specific genes, and methods for studying methylation are improving rapidly. Histones are also critical in determining which stretches of DNA are available for transcription.

  2. What techniques are used to analyze gene expression?

    Genomics currently focuses primarily on analyzing RNA expression. Data acquisition is relatively straightforward and consists of expression profiling, cDNA arrays or ESTs. The medium used can vary. It is possible to examine up to 8000 sequences at once with a plastic array. Tissue specific arrays are being developed, which would be particularly useful in ED screening and testing (e.g., could examine gene expression in testes vs. prostate, etc.)

    The most difficult aspect of genomics is data analysis and interpretation. It is easy to get data from arrays - but then what do you do with it? Cluster analysis? Some statisticians feel that normalized stats are not appropriate for analyzing array data. Every 2-3 months there seems to be a new approach. One need that must be addressed before this methodology can be applied to ED screening and testing is a unified approach to data analysis and interpretation.

    Other technical questions include storage and retrieval of data. Some advocate standard structure, others recommend putting the data on web sites. Who owns the data? How can you effectively share the data without losing ownership?

    The methodology used in proteomics consists of resolving a protein mixture on 2-D gels and then identifying specific proteins in the resultant spots. This methodology is rapidly evolving. Problems associated with the methodology include post-translational modification of proteins that can alter their migration patterns in gels (and thus different spots are not necessarily different proteins) and protein-protein interactions. Questions exist concerning whether or not this tool is too sensitive to use for standardized screening and testing. At the moment, it is difficult to classify the data.

    Another approach to proteomics involves antibody arrays. These are commercially available; however, this methodology has many of the same problems observed with genomics, e.g., data analysis, interpretation, storage and retrieval. A relatively novel approach to proteomics involves analyzing expression pattern of transcription factors, which is done by examining binding of proteins to DNA, e.g., analyzing protein-DNA interactions. This method is also evolving rapidly.

    Summary of techniques: if you know what your targets are, use non-array techniques; if you have a discovery question, use arrays.

  3. Major questions to be addressed when considering application of genomics and proteomics to ED screening and testing include:
    1. What are the appropriate species and/or strains?
    2. What are the appropriate windows of exposure?
    3. Which tissues should be analyzed?
    4. Which cell lines are true reflections of in vivo biological function?

Presentation by Barbara Hale: Preliminary data from a TSRI group project illustrating application of genomics to study of ED

  1. Tributyltin is an anti-fouling agent used in marine paint. Is it an ED? It is known to cause imposex in fish
  2. Hypothesis tested: in utero exposure to tributyltin at environmental levels will have adverse reproductive consequences.
  3. Study design consists of exposing rodents to tributyltin in utero, then following effects on reproductive organs at varying life stages including fetal, postnatal, and adult animals. The endpoints being considered are gene expression as assessed by cDNA arrays, and histology.
  4. Observations include differential effects on gene expression. However, not sure what these differences mean since obvious histological changes are not noted.

Presentation by Kevin Gaido: Data from CIIT research regarding the use of genomics to discover the mechanism by which phthalate causes anti-androgenic effects.

  1. Phthalates act as anti-androgen, but do not interact with the receptor. Use genomics as a discovery tool to determine how these chemicals cause anti-androgenic effects in the absence of interactions with the receptor.
  2. Rats were exposed in utero to phthalates, then looked at mRNA expression in fetal testes using cDNA microarray. Data from array studies were confirmed using PCR.
  3. Phthalates blocked expression of genes that control cellular uptake of cholesterol; also blocked expression of other key genes in steroidogenesis. This model of phthalate endocrine disruption derived directly from discoveries made using genomics.

Summary of General Discussion: Four Key Questions

  1. Which endpoints of relevance to endocrine disruption are particularly amenable to genomics and proteomics technologies? Where do these methodologies fit in the ED screening and testing program?

    Endpoints particularly amenable to genomics and proteomics:

    • Gene transcription events downstream of estrogen, androgen or thyroid hormone binding to nuclear receptors.
    • However, changes at the genomic level are an exposure tool, cannot be used for regulation, but rather for signaling potential for endocrine disruption

    Suggested Applications in ED screening and testing:

    • For prioritizing chemicals for screening and testing (b/c of potential for high throughput)
    • Use to refine animal testing, e.g., uses data from genomics to assist in selection of appropriate Tier I and Tier II tests
    • As a screening tool - develop a profile of gene expression changes associated with endocrine disruption, or identify a key gene that is modified by EDs. At present, the biological relevance of most chemical-induced genetic changes are not really known, so in the short-term genomics and proteomics should be used in conjunction with in vivo tests; as data base increases, it may be possible determine which genetic changes are biologically relevant and thus replace in vivo tests with genomics.
    • Use genomics to develop biomarkers of exposure
    • Use genomics as a discovery tool for identifying potential mechanisms of endocrine disruption
  2. What is the best approach to determine the sensitivity and specificity of these methodologies, and to validate and standardize these technologies?

    Approaches to determining sensitivity and specificity:

    • In tissues, there is lots of "biological noise" requiring replicates of samples to identify what is a real change, and what is just noise. The number of replicates required differs from system to system.
    • Need to confirm results from arrays to follow-up experiments using RT-PCR and/or in situ hybridization or immunocytochemistry
    • Need to link array data to biological effects
    • Need to develop a standard approach to understanding and communicating variability in array data

    Issues pertaining to validation and standardization:

    • Need to first simulate research to identify optimal platform or test for ED screening or testing, thus there is the need to increase funding opportunities for grants focused on examining the potential for genomics to replace toxicity testing
    • Yet eventually need to optimize and standardize sample collection and storage, sample preparation, data analysis and interpretation, so that results can be compared across labs and across samples
  3. What is the best approach to establish relevance of genomics data to endpoints or effects of biological significance?
    • This is the most profound question - the technical issues will be relatively straightforward to resolve, but without answering this question first, genomics data in and of itself will have little meaning.
    • Need to determine where in the continuum of biological response the gene expression changes occur, e.g, early events (indicator) vs. response at the tissue, organ or organism level.
    • Incorporate into existing uterotrophic and Hershberger assays, which have well-defined biological endpoints, to determine if unique gene expression profile is linked to effects on these reproductive organs
    • Once key genes have been identified, you can use an imaging approach to link genomics and proteomics to events going on in the intact animal in real time
  4. What is the feasibility of collecting tissues from ongoing endocrine screening/testing programs for subsequent or parallel analysis by genomics and/or proteomics?
    • Techniques exist for preserving integrity of RNA samples (collect into RNA later and store at -80°C)
    • Need to determine the appropriate time in ongoing programs to collect samples
    • Need to develop procedures for documenting the integrity of the RNA samples.
    • Questions remain regarding who has access to stored tissues