Testing the toxicity of pharmaceutical candidates in lab rats before the compounds are judged safe enough for human clinical trials is notoriously unreliable. Often compounds that appear safe in the rodents prove to be toxic in humans.

To come up with a better way of predicting toxicity, Gabriela Cezar, assistant professor of animal science at the University of Wisconsin-Madison, is turning to human embryonic stem cells. In the current issue of Stem Cells and Development, Cezar and her colleagues reveal a novel way to test drug toxicity: by monitoring the behavior of embryonic stem cells exposed to a drug-candidate compound. Studying how potential drugs affect embryonic stem cells could provide a far more accurate prediction of a drug’s potential toxicity than conventional animal models can.

During normal development, embryonic stem cells produce molecules that direct cellular metabolism and differentiation. Cezar hypothesized that exposure to a toxic drug may skew concentrations of these molecules, disrupting cell-to-cell interactions and causing a biological cascade resulting in potential developmental disorders.

As a proof of concept, Cezar’s group looked at human embryonic stem cells’ response to valproate, an anti-epileptic drug that has been linked with cases of autism and spina bifida in the offspring of mothers treated with the drug. “Developmental disorders and birth defects start in utero during pregnancy, and we have no way to measure or look at mechanisms that could be participating in the onset of these diseases,” says Cezar. “With human embryonic stem cells, we can recapitulate development of the human brain and measure concrete changes of chemicals to drugs like valproate.”

In the experiment, Cezar introduced various dosages of valproate, from very low to high, into three sets of embryonic stem-cell cultures, altering the dosages over different lengths of time. Control groups contained stem cells not exposed to the drug. Cezar then ran each sample through a mass spectrometer, which measured concentrations of the molecules present in culture.

Compared with the control group, samples with valproate exhibited significant changes in the concentrations of two key molecules: glutamate and kynurenin. Both molecules are heavily involved in early brain development, and Cezar found that exposure to valproate caused spikes in each molecule’s concentrations, indicating that such molecules may serve as biomarkers for a drug’s potential toxicity.

“We’re predicting toxicity to humans in human cells,” says Cezar. “Discovering these measurable molecules of toxicity, we hope to present other serious adverse reactions that are caused by testing drugs in animals, with the hope of bringing safer drugs to patients.”

However, using embryonic stem cells as testing grounds for drug safety is still a relatively new concept, and according to some scientists, much more research is needed before it can be determined that the method is viable. Steven Tannenbaum, professor of chemistry and toxicology at MIT, says that drug metabolism in the body is a complex process. In particular, drugs taken into the body are processed first in the liver, taking on different forms before traveling through the rest of the body, and into the womb. “More than 90 percent of drugs are metabolized in the liver to other forms of the drug, some of which might be toxic,” says Tannenbaum. “This group has taken valproic acid, which is normally extensively metabolized in the body, and exposed it under unrealistic conditions.”

Cezar says that a possible solution may be to direct embryonic stem cells to grow into liver cells before exposing them to drugs–a project that she may take up in the future. “As long as we can make the cells from human embryonic stem cells, then once we have the mature cells in a dish, we could discover biomarkers in liver toxicity,” says Cezar.

“It’s a very versatile platform.”