DENVER, CO — Toxicology is getting a facelift with an infusion of genomics and proteomics–and powerful computing–that will help researchers predict adverse effects to chemicals in the environment, as well as the effectiveness of drugs used to treat breast cancer patients.
Knowing how a person's body would allow a toxic effect to occur at the gene and protein level, may someday make it easier for doctors to decide what treatment is more appropriate. Such information also would lead to the development of better biomarkers to detect toxicity earlier during medical treatment, researchers said today at the 2003 American Association for the Advancement of Science (AAAS) Annual Meeting.
Like a wad of gum stuck in a child's hair, a small breast cancer cell from a tissue mass requires delicate unraveling and extraction. Traditionally, researchers have studied the genetics of the whole tumor mass because of the difficulty faced when trying to excise the cell of interest. Now, Emanuel Petricoin of the Food and Drug Administration has been working with a special video-game like microscope that pulls selected cells out, like a tractor beam. When combined with new protein array technologies he and his colleague Lance Liotta (National Cancer Institute) have also been developing, the new microscope technology has allowed him to study toxicity and the cancer-associated "miswiring" of the cellular circuitry in breast cancer patients receiving different kinds of drugs or combination of drugs.
After the cell is isolated from the tumor mass, it is ready to have its proteome (the protein complement encoded by a person's genome) analyzed. To do this, Petricoin has also developed protein array tests to detect subtle changes in protein phosphorylation along many key "nodes" within the cellular circuitry. Phosphorylation occurs when a phosphate group is removed from ATP, a component in many cellular circuits. While it is possible to know of the existence of a gene product in a cell, Petricoin takes it a step further to find out what it is doing–he is concerned not just with the receptor of a molecular drug, but whether or not there is a continuous energy cascade or a plug, for a given cellular activity. As a result, he is able to look at a cell-signaling pathway globally.
"This protein array is able to look at signal pathway profiling, cellular circuitry, and protein-protein interactions on a chip, to look at toxic and cancer related signaling events," Petricoin said.
"We're also using this to monitor toxicity simultaneously. In clinical trials, we're going to use this to target the patient's response and change drugs to fit the profile. The promise of proteomics is that while DNA is the information archive, proteins actually do the work. They are the targets, the biomarkers, or the molecular machines that cause the cell to die, grow, or differentiate," Petricoin said.
By profiling these tumors and assigning them with a computer-system tool to their proteomic mass spectroscopy, to generate a complex protein "barcode" serum, the computer can datamine between different known toxicities for patients, and have a better basis upon which to assign a drug treatment.
Work by Richard Paules at the National Institute of Environmental Health Sciences, another AAAS speaker, may provide the scientific community with more insight on how exactly genes get regulated to allow the manifestation of a state of toxicity. This would be helpful for drug development, and for wider chemical development and safety evaluation of these compounds to determine toxicity before synthesis.
The database project that Paules leads contains data on gene expression effects, based on the exposure to multiple substances. By using rodents as the model, the point is to generate information that is useful to people. "One of our key efforts right now is prediction of adverse effects, to profile adverse effects," Paules said. "We're doing studies trying to measure classical parameters, and linking them with gene expression changes we can see."
The chemical substances that the researchers are studying are well-characterized and well-studied agents that have relevance to human exposure, Paules said. "The goal is to use these compounds in different combinations that share similar effects among members, and also members that do not share that. Some of these compounds are drugs, so they will have pharmaceutic effects. All of these compounds have nonspecific effects due to the initial stress of a xenobiotic. Some of that is reversible. Some of them are carcinogens, known for humans or mice," Paules explained. The researchers are also using nontoxic, inert isomers as controls.
"One of our long-range goals is to use genomics to evaluate an exposure rapidly, with accuracy, with fewer numbers of animals in a reproducible, predictive manner," Paules said.
In addition to knowing whether a gene is being up or down-regulated, researchers are also learning which genes are most informative and discriminatory, carrying the most information.
"That's useful in helping evaluate impact of exposure," Paules said.