The result was an enormous amount of chemical data — descriptions of tens of thousands of chemicals and their properties. But those chemicals are just the tip of the iceberg: For every chemical that has been tested, there are tens of thousands that haven’t. Thomas Hartung, a public health toxicologist at Johns Hopkins who directs the Center for Alternatives to Animal Testing, estimates that there are 100,000 chemicals in everyday products alone, 90 percent of which have never been tested.
But therein lies the beauty of chemistry. When it comes to toxicology, it turns out that chemically similar substances usually have similar effects on humans. Consider the case of bisphenol-S. The chemical came into wide use when its relative, bisphenol-A, was found to mimic estrogen and linked to a host of diseases and disorders. Bisphenol-S, which is chemically similar to bisphenol-A, was touted as an alternative for baby bottles and plastic cups — but earlier this year, scientists announced that it appears to affect animal embryos in similar ways.
Since similar chemicals have similar effects, reasoned Hartung, why not map their relationships? His team embarked on a dizzying data project, sorting out REACH data for 10,000 chemicals and adding info from over 800,000 research studies conducted on those chemicals alone. They created a computer model that shows which chemicals are similar to other chemicals with the click of a button. The results were published last month in the journal ALTEX.
“This is an eye-opener,” Hartung said. He sees the model as a kind of shortcut that could save companies billions of dollars in testing, set priorities for public health inquiries and even change the way chemists produce new substances.
The database could have a surprising secondary benefit: It could eventually help us cut down on animal testing. By laying bare chemicals’ similarities, it could reduce the need to test chemical toxicity in animals when testing has already been performed on similar substances. That could save billions of dollars — along with the lives and well-being of countless lab animals.
The team’s chemical map also highlights which substances have been tested again and again on animals by different organizations. “We found two chemicals which were tested 90 times in rabbit eyes,” Hartung said. “Sixty-nine chemicals were tested 45 times in rabbit eyes.” Since rabbit eyes are used to approximate reactions in human eyes, the developers of substances similar to those that were already cleared could have saved themselves — and the bunnies — a lot of trouble.
Hartung hopes to make the database publicly available at some point but faces legislative, legal and logistical hurdles. Since chemical companies technically own the data, they are hesitant to open it up to potential competitors. He’s involved in ongoing talks with the European Union and manufacturers and is also talking with the FDA and EPA about creating a similar database in the United States.
Ultimately, he said, all parties “have the same agenda as we have” and want to understand the world’s chemical landscape, save money, and reduce animal testing. But Hartung will have to reach a delicate consensus with industry and government for the data or the tool ever to be made public. “We’re trying to find a way here,” he said. For now, his team is busily adding new data to the tool — and a world of chemical connections is becoming ever more clear.