COI Funded Project: Genomic Approaches to Assessing the Global Impact of Marine Pollution
Grant Funded: 2001
The specific objective of the research proposed here is to address
the question of whether cDNA microarrays ("gene chips") constructed
with genes of one species can be used to evaluate changes in gene
expression in other species, and to demonstrate the use of these
chips in assessing the effects of contaminants in the ocean. This
would represent a major step toward our long-term objective of defining
the significance of chemical exposure in different types of animals,
in marine environments around the world.
Genomic efforts have opened the door to development of tools such as microarrays to profile the expression of thousands of genes simultaneously. However, a major obstacle to employing this technology in the oceans is that currently, the development and use of gene chips is largely restricted to a select group of laboratory species for which genome sequence data has been obtained.
In the past month, zebrafish gene chips (5000+ genes) have been generated (in a collaborative venture), making this proposal possible. Use of these chips to detect broad scale gene changes in response to TCDD in zebrafish is underway. We will determine whether these chips constructed with genes from one species (zebrafish) can be validly used to assess gene expression profiles in other fish species, particularly marine fish (Fundulus heteroclitus). Subsequently, we will determine the utility of these chips in assessing the spectrum of gene changes associated with chemical exposure in the environment, targeting species and locations that are significant in coastal contaminant issues (Fundulus) and to open ocean contaminant questions (midwater fishes).
If "cross-species" use is proven, it will open the door to deploying these chips for broad scale use in assessing the status of ocean health. The results of this study could establish, finally, the basis for unprecedented assessment of chemical effects in marine systems.
Funding from the Ocean Life and Coastal Institutes would allow us to prove the concept and thus enable us to seek large scale funding from governmental and international sources.
The specific objectives of the research were 1) to determine whether cDNA microarrays ("gene chips") constructed with genes of one species can be used to evaluate changes in gene expression in other species, and 2) to demonstrate the use of these chips in assessing the effects of contaminants in the ocean. This would represent a major step toward our long-term objective of defining the significance of chemical exposure in different types of animals, in marine environments around the world. The results of the work accomplished indicate that, with further optimization, cross-species hyridization in microarrays should be an extremely informative tool.
Anthropogenic chemicals that are of concern to animal and human health occur worldwide, in all environments and in virtually all organisms. It is well known that the oceans are the ultimate sink for these chemicals. Concerns about these chemicals derive from results of laboratory experiments and from overt diseases including reproductive and developmental abnormalities in animals where the chemicals occur at high levels, primarily in coastal environments. For more than 25 years there have been vigorous efforts to determine the health effects of such contaminants occurring at lower levels in the world's oceans, in coastal as well as in open ocean waters. These efforts generally have been inconclusive, and have been hindered because the scope of the effort has been completely inadequate relative to the scale required to answer the questions.
The sequencing of all or large parts of the genomes of a growing number of species has opened the door to development of tools for assessing the expression of many, even all genes in a species. One such tool is the "microarray", or gene chip, which enables investigators to profile the expression of thousands of genes simultaneously. This has become a major tool for investigating the relationships between changes in health or environmental condition and changes in gene expression. Although still very young, this field of "toxicogenomics" is already making great strides in the areas of elucidating molecular mechanisms of toxicity and defining chemical-specific expression profiles. Ultimately, the goal of much of this work is the development of diagnostic and predictive biomarkers for pre-clinical, clinical, and environmental applications. DNA microarrays are the primary tool being used in such work.
A major obstacle to employing this technology in the oceans has been that there are still few species for which many gene sequences are known, and fewer still for which the entire genome is known. Up to now, the development and use of gene chips has been restricted largely to a select group of laboratory species for which genome sequence data has been obtained. As part of the research supported here, we determined whether gene chips constructed with genes from one species (zebrafish) can be used to assess gene expression in other fish species, particularly marine fish (the marsh killifish Fundulus heteroclitus).
Research Plan Undertaken
In developing the proof of concept, our approach was intended to be in three phases: Phase 1, the production of fish gene chips and demonstrating that they can reveal broad scale gene changes in response to chemicals.
Phase II, determining whether these chips constructed with genes from one species can be validly used to assess gene expression profiles in other fish species.
Phase III, determining the utility of these chips in assessing the spectrum of gene changes associated with chemical exposure in the environment, targeting species and locations where there are significant questions concerning contaminant effects, i.e., in the coastal environment (Fundulus) and the open ocean (mid-water fishes).
Generation of fish gene arrays was completed with zebrafish (Danio rerio), a model vertebrate species for which substantial genome sequence information is available. The support from the Ocean Institutes was key in completion of these chips that were constructed as part of collaboration with a large zebrafish research group at Mass General Hospital. The chips were made with DNA pieces obtained from a zebrafish heart preparation, as the original objective for use of these chips was to assess gene changes associated with the developmental defects caused by halogenated aromatic hydrocarbons including selected polychlorinated biphenyls (PCBs) and the highly toxic chemical TCDD or dioxin. The chips were used in several experiments to assess changes in gene expression in embryonic fish exposed to TCDD. A first example of a result is shown in Figure 1, (which appeared also in the proposal). The results with multiple exposures and doses have provided a strong picture of the changes caused by TCDD. (These results are summarized in the Joint Program Ph.D. thesis of Dr. Heather Handley, which was supported in part by the Institute funding.)
The large body of DNA sequence data needed to support microarray design severely limits the number of species for which microarrays are available. Fundulus heteroclitus, the salt-marsh killifish, is a small marine fish that has been used extensively in both developmental biology and ecotoxicology. At the outset of the studies, the F. heteroclitus genome was poorly characterized. While additional DNA sequence data are now available for F. heteroclitus, for nearly all other environmentally and economically important fish species extensive sequence still are lacking. There is evidence that microarrays constructed with material specific to one species can be used to assay gene expression in closely related species. Thus, it was of interest to determine whether zebrafish microarrays might be used to study gene expression in other fish species.
To this end, we prepared labeled cDNA from both zebrafish and F. heteroclitus heart RNA, and compared the strength and patterns of hybridization to zebrafish cDNA microarrays. mRNA from Fundulus heteroclitus adult heart tissue (provided by Dr. Sibel Karchner) and total RNA from zebrafish adult heart tissue was used to generate appropriately labeled (amino-allyl post-labeled) cDNA, according to protocols developed in phase I. Single-color hybridizations to AH-001 arrays (that is, zebrafish Adult Heart first generation arrays) were performed at 55?C, with all other conditions as previously defined in Phase I.
Side-by-side visual inspection of same-species and cross-species hybridizations revealed obvious similarities in patterns of relative signal strength among features (Figure 2).
To quantify this relationship, feature (i.e. spot) intensities from 3787 features on Cy3 hybridizations were compared directly (Figure 3). In the vast majority of cases, same-species hybridization produced higher fluorescence intensity. Several hundred features with cross-species fluorescence intensities less than or equal to 2-fold higher than same-species intensities were separated from the main body of data. Each group showed moderate levels of correlation between same-species and cross-species fluorescence intensities (R² = 0.59 and 0.75).
These results suggest that, while less efficient than same-species hybridization, cross-species hybridization to zebrafish microarrays may be used to detect gene expression in fish species for which DNA arrays are not available. A general correlation between same-species and inter-species hybridization results was readily apparent upon inspection of either hybridization images or resulting numerical data.
The phase III of the study was not completed within the time frame of the grant. However, samples of Fundulus heteroclitus were collected from coastal systems of know contaminant levels (low to high). Samples of mid-water fishes also were collected from the Northwest Atlantic. The coastal and mid-water samples are archived and will be analyzed as time and funding allows, to complete the originally proposed studies.
Further work on cross-species hybridization. Neither the results of this work nor published investigations of inter-species hybridization have adequately addressed the potential for non-specific hybridization. In the current case, significant "outliers"and only moderate support for a regression trend-line fitted to the main body of data both suggest an unexplained source of variation affecting some subset of Fundulus genes. In other studies, a high level of variation in cross-species results for 6% of arrayed human genes also suggested gene-specific artifacts. Increasing the stringency of cross-species hybridizations could decrease the variation. Further work is needed to determine an ideal hybridization temperature for use of Fundulus heteroclitus samples with zebrafish arrays. However, it is also possible that a large portion of variability is due to actual biological differences (i.e., differences in basal levels of expression of certain genes). Thus, it would be interesting to assay non-specific hybridization using individual Fundulus gene transcripts. Such work has yielded important information about specificity of arrayed cDNA probes, and might contribute to the development of general guidelines for conditions of cross-species hybridization.
In summary, the results establish the proof of concept, showing that microarrays with genes sequences from one species can be used with other species. The next step in these studies will be to employ the gene chips in hand to assess the spectrum of gene changes associated with chemical exposure in the environment.
Support from the two Ocean Institutes has opened the door to further defining the conditions for cross-species hybridization and eventually deploying these or similar chips for broad scale use in assessing the status of ocean health. That support is most gratefully acknowledged.
We also acknowledge Dr. Mark Fishman, formerly at the Mass General Hospital and now with Novartis, for collaboration in providing laboratory space and materials that made it possible for Heather Handley to produce the zebrafish arrays. That collaboration represents a very large "in-kind" contribution to this research. We acknowledge Dr. Larry Madin and particularly Mr. Erich Horgan for their collaborative assistance in obtaining samples of mid-water fishes, also a large "in-kind" contribution.
Originally published: January 1, 2001