Yale Environment Review

Yale Environment Review (YER) is a student-run review that provides weekly updates on environmental research findings.

Designing safer chemicals

By Antoine2K via Shutterstock

Shen, Longzhu Q. et al “Probabilistic diagram for designing chemicals with reduced potency to incur cytotoxicity” Green Chemistry (2016) 18: 4461-4467. DOI: 10.1039/c6gc01058j

Tuesday, June 6, 2017
More than 83 percent of chemicals have no safety information. Most businesses don’t design them for safety, and the government doesn’t test most of them for safety. Yet thousands of chemicals are in our water and soil, potentially causing human harm and costing billions to cleanup. How can we tell if new chemicals will cause damage to humans before they are made?

Chemists build molecules like Lego blocks. The chemicals they make are in the products we use every day — all the materials we use in our work, home, and personal life. However, 83 percent of these chemicals lack safety data. These products end up in the environment after we throw them away. There are now thousands of these chemicals in our environment that cause harm to humans and other animals. Is there any way a chemist can predict how toxic a chemical will be before the product is even produced, marketed, and sold?
This is not easy, but it is possible. Chemists normally tell if a chemical is toxic by giving some to an animal and observing if there are any negative effects on the animal’s health. However, animal research is expensive and raises ethical concerns. In a recent paper published in the journal Green Chemistry, a team of researchers led by Longzhu Shen at Yale University developed a way to predict whether a chemical will be safe based on its chemical properties. These properties are similar to choosing the color and shape of the Lego pieces when building a structure.
The structure of our cells is what makes us human. We have trillions of cells in our body with many types in our different organs. Instead of using live animals, the Yale researchers examined an EPA database of the effects of chemicals on human cells in test tubes. The researched selected 1,006 different chemicals with 37 toxic effects on human cells in the laboratory. These toxic effects included cell death, changes in cell growth, and cell shape. These effects occurred in many different organs (liver, breast, skin, blood, lungs, intestines, and cervix) where disease and cancer is observed from chemical exposure.
The researchers used the data to predict whether or not a chemical was safe using a statistical model. Here safe means that the chemical had no toxic effects on the organ cells in the database. The researchers knew that chemists construct chemicals that provide certain effects (the color and shape of the Lego block). For example: does the chemical dissolve in water? Does the chemical make its surroundings more acidic? How much electric charge does it have? Humans represent these molecular properties on paper by drawing lines and letters. However, on the scale of a cell, the molecule is an amorphous blob with different surface textures, vibrating into other molecules. How the created chemicals collide with human cells determines if a chemical will cause us harm.
Instead of simply assuming that more of a chemical would cause more damage, they took what is called a probabilistic approach. That is, given a certain chemical property, what is the probability that it may cause harm to cells in one or more places in the human body? Their model was tested against chemicals in the database and correctly predicted toxic effects with 77-percent precision. This model gives chemists safety guidelines to consult while they are designing new molecules. The model’s high predictive power showed that toxic effects on human can be largely explained by physical molecular properties that are not specific to an individual chemical, but rather certain combinations of Lego colors and shapes.
The power of this research is that it shows how businesses can design chemicals to achieve similar function (e.g. cleaning your clothes or washing your hair) with reduced toxic effects to humans. By comparing a proposed chemical’s properties with the statistical model, chemists can estimate potential toxicity to humans before a chemical is mass-produced. Applying this technique for every new chemical would potentially avoid the societal pain of unintended damage to human health and billions of dollars in future cleanup costs.