Sitting at a computer connected to a large microscope, Salman Khetani calls up a kaleidoscopic image: green islands of human liver cells in a hexagonal pattern, surrounded by a red sea of support cells. Sangeeta Bhatia, Khetani’s advisor, says that the cells have been carefully patterned to hit the liver “sweet spot.”Arranged just so – in 37 colonies about 1,200 micrometers from each other – the cells behave as though they were in the human body.
When grown in the lab using existing methods, liver cells can survive for a day or two, but over the course of a week, they lose the ability to perform their liver-specific functions and then die. Bhatia and Khetani’s cells, on the other hand, function for about a month. They secrete the blood protein albumin, synthesize urea, and make the enzymes necessary to break down drugs and toxins. Bhatia believes that the cells act enough like human tissue that they could be used to screen new drugs for liver toxicity or to study metabolism and, possibly, hepatitis C, a virus that grows only in human tissue. Indeed, the researchers have already developed a drug toxicity test that uses liver cells arranged in their signature hexagonal pattern.
[For images of this lab, its researchers, and their processes for growing liver tissues, click here.]
In addition to being a major health problem, liver toxicity is the primary reason pharmaceutical companies recall existing drugs or abandon new ones that are under development. Bhatia says that’s because “when you’re developing new [drugs], there aren’t really good models of human liver.”Instead, drug companies rely on cancer cells, dying liver cells, or rat tissue – poor substitutes for fully functioning human liver tissue. Bhatia and Khetani believe they can supply a better model.
Bhatia, an associate professor in the Department of Health Sciences and Technology and the Department of Electrical Engineering and Computer Science at MIT, developed her patterning technique using rat cells, when she was in graduate school in the mid-1990s. At the time, she was interested in using micropatterning, an emerging technique for physically arranging cells in culture, to build a dialysis-like device to support patients with liver disease. For her PhD, Bhatia worked on using the technique to bolster cell function and was particularly interested in finicky cells like liver cells (also called hepatocytes).
Inspired by the work of others in her lab who were growing multiple cell types in the same cultures and combining fibroblasts – supportive cells that normally live in connective tissue – with skin cells, she tried micropatterning fibroblasts alongside her hepatocytes. Micropatterning more than one cell type at a time and regulating the interaction that hepatocytes had with each other, and with the secondary cells, was an innovation. The fibroblasts Bhatia borrowed for her experiments turned out to be particularly good at bolstering liver functions. She describes her breakthrough as “a happy, lucky thing that I just stumbled upon.”
Even though there are no fibroblasts in the human liver, their presence in Bhatia’s cultures coddles the hepatocytes and keeps them functioning. Part of the reason that cells behave like liver, lung, or muscle cells is their environment: signals from neighboring cells, physical forces, and the matrix of supportive proteins stabilizing them. As successful as the method has proven to be, Bhatia is still investigating what exactly causes each patterned hepatocyte island to behave like liver tissue.
For his PhD, Khetani, now a postdoc at MIT, was able to apply Bhatia’s technique to human hepatocytes, making possible the development of the toxicity test. Bhatia says Khetani’s work “was a logical extension of mine, but again surprisingly, human hepatocytes turned out to be even more sensitive to clustering than rat hepatocytes.”
Others have attempted to grow functioning liver tissues on scaffolds. But, says Khetani, this approach lets the cells do their own organizing, so the architecture of the resulting models is different every time. Bhatia and Khetani, by contrast, precisely specify the organization of the cells in their model, giving them tighter control over functionality.
To verify that their micropatterned liver cells actually behave like hepatocytes in the human body, Bhatia and Khetani put them through a series of rigorous tests. They analyzed the cells’ gene-expression profile and measured the amount of drug-metabolizing enzymes they produced. They exposed the cells to a battery of substances known to be either benign or toxic to the human liver, from caffeine to cadmium. To test the toxicity of a drug, Khetani creates a solution of the desired concentration and pipettes it into a set of wells, where it’s incubated with the liver tissues. Then he looks for changes in hepatocyte function or cell death.
Drug companies could, says Bhatia, use this assay to compare several chemically similar compounds and eliminate toxic ones early in the drug development process. “If I were at a drug company,”she asks, “and my medicinal chemists gave me four compounds, could I have picked which one would have been the most toxic using my assay?”The answer seems to be yes. She and Khetani have compared chemically similar drugs known to be benign or toxic to the liver and confirmed that the new assay can measure differences in toxicity.
Strengths and Limitations
Bhatia’s assay is good at detecting drugs toxic to the general population, but it may not uncover drugs with adverse effects on only a small number of people. It might not, that is, have detected the trouble with Rezulin, a diabetes drug that caused liver damage in many patients and which the U.S. Food and Drug Administration ordered off the market in 2000. The liver cells in the assays do not represent a wide enough sample of the population to predict such effects, though it’s theoretically possible to test a drug on cells from thousands of different livers.
The assay is unique in being able to test drugs for chronic toxicity, which is caused by low-level repeat exposure, “which is actually the way we take our drugs clinically,” says Bhatia – one pill a day. Bhatia’s model could be used to test the effects of a drug over four to six weeks. Existing models simply cannot detect chronic effects because liver cells die so quickly in culture. Bhatia says that pharmaceutical companies know that potential drugs that become toxic only over time are slipping through the cracks, but the FDA does not require chronic toxicity tests. “We’re in this kind of funny position where we’ve developed a really powerful tool and have to convince people to use it.”
The miniature tissues can also be used to detect acute toxicity, which has much more immediate effects. Acute toxicity can be studied using an existing method, with simple cultures of hepatocytes that die within a week. But Bhatia believes that her assay will be more efficient: because it uses wells, it requires a lower volume of drugs, and the micropatterning means fewer hepatocytes are required.
Bhatia is developing her test for commercialization, and several pharmaceutical companies are interested. She and Khetani are also looking into other uses for the assay – for example, studying interactions between drugs and how liver cells transport drugs. “My hope is that the assay would make drugs safer, better labeled, and would help ensure that toxic drugs never reach patients,”says Bhatia.
The dark secret behind those cute AI-generated animal images
Google Brain has revealed its own image-making AI, called Imagen. But don't expect to see anything that isn't wholesome.
The hype around DeepMind’s new AI model misses what’s actually cool about it
Some worry that the chatter about these tools is doing the whole field a disservice.
The walls are closing in on Clearview AI
The controversial face recognition company was just fined $10 million for scraping UK faces from the web. That might not be the end of it.
This horse-riding astronaut is a milestone in AI’s journey to make sense of the world
OpenAI’s latest picture-making AI is amazing—but raises questions about what we mean by intelligence.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.