Artificial liver tissue predicts drug safety

The development of a new drug can be a lengthy and costly process. With the investment of millions of dollars and years of research, scientists go to great lengths to try to ensure a potential drug’s safety before it reaches patients. In the search for safe and effective drugs, researchers look closely at the actions of one organ in particular: the liver, which is where drugs are broken down into useful or toxic compounds. But the human liver and mouse liver handle drugs differently, meaning that certain drugs can show no harmful effects when tested in mice, but can have devastating consequences in humans.

To create a better human liver model in which to test potential drugs, Sangeeta Bhatia’s laboratory at MIT developed artificial human liver tissue – tissue engineered from preserved human cells in the laboratory that behaves like human liver but can be implanted in mice. The MIT scientists teamed up with researchers at the Broad to test the artificial liver tissue and determined that it mimics much of the molecular activity of human tissue. Implanted in mice, the artificial tissue produced some of the same interactions and injuries that are usually only seen in humans. Their results appear online in Proceedings of the National Academy of Sciences July 11.

“Before, the way people were analyzing new liver tissue in the field of drug metabolism or toxicity was to look at about six drug metabolism genes and a few metabolites,” said David Thomas, a Broad associated researcher, instructor of medicine at Harvard Medical School, and an author of the paper. He and his colleagues at the Broad developed a set of molecular tools known as probes to survey the expression and function of a range of enzymes involved in the human liver’s activity. “What we have been able to do is look comprehensively at all of the known liver drug metabolism genes.”

Alice A. Chen, a graduate student in the Bhatia lab and first author of the paper, remembers first recognizing the opportunity for collaboration with researchers at the Broad. “We thought, ‘This is fantastic,’” said Chen. “Since we are interested in building human livers, we realized we could compare tissue that we built to the gold standard of human liver tissue culture. From there, we were able to convince ourselves that we had functioning, phenotypically similar human livers and it was worthwhile to implant them in mice and test their function.”

The researchers looked at 83 drug metabolism genes and transcription factors – master regulators that can turn on and off other genes – that are found exclusively in human liver tissue. They found that most of the enzymes are expressed in the engineered tissue, including the handful of enzymes that account for 90 percent of drug metabolism.

Once the researchers evaluated the artificial tissue and determined that the relevant human enzymes were expressed, they implanted the tissue into mice and looked for major metabolites – molecular components – that are produced in the human liver but not in the mouse liver when a drug is administered. The researchers were also able to reproduce human drug-drug interactions, which can alter the effects of two or more drugs administered at the same time. Some metabolites – known as disproportionate metabolites – do not usually appear in mouse models. Their appearance during human drug trials can set the drug development process back years.

“The upshot of all of this is that if we can identify problematic compounds and bridge the gap that exists in human and animal models, we can potentially help make drug development safer and more efficient,” said Chen.

Other research groups have used genetic methods to manipulate mice and inject human liver cells, but this process is labor intensive and difficult to predict. The approach reported here avoids many of the difficulties of these methods since it does not require genetic manipulation.

The connection between the research in the Bhatia lab and the enzyme probe set developed at the Broad was serendipitous. Thomas and his colleagues originally designed this resource to make predictions about drug toxicity using cells in a dish rather than engineered tissue in a mouse model, but by connecting with experts in other areas, they found many new applications.

“Every aspect of the projects that I work on is intimately collaborative on both the biology side and technology side,” said Thomas. He is also working with collaborators interested in analyzing whole slices of human liver tissue and a group interested in characterizing induced pluripotent stem cells. “That’s what all great projects allow you to do: they take you to unexpected places, and help you make unexpected connections and findings.”

Other researchers who contributed to this work include Luvena Ong, Robert Schwartz, Todd Golub, and Sangeeta Bhatia.