CHARLES BERGQUIST: This is Science Friday. I’m Charles Bergquist.

FLORA LICHTMAN: And I’m Flora Lichtman. Last week on the show, we talked about how pharma companies are using artificial intelligence to speed up the drug research and development pipeline, and this week, we’re zeroing in on another part of that process, animal testing. Before a new drug can begin clinical trials in humans, it gets tested on animals, but things are changing.

Late last year, Congress passed the FDA Modernization Act 2.0, which cleared the way for new drugs to skip animal testing, but can we expect to phase out animal testing altogether? And what technologies might make that possible?

To help us answer those questions and more is my guest, Dr. Thomas Hartung, Director of the Johns Hopkins Center for Alternatives to Animal Testing based in Baltimore, Maryland. Dr. Hartung, welcome to Science Friday.

THOMAS HARTUNG: Thanks for having me.

FLORA LICHTMAN: OK, so what is driving this push to phase out animal testing for drugs? Is this about ethical concerns, or for scientific, or for safety reasons?

THOMAS HARTUNG: It is for all of, these and also for economical reasons. You have to imagine, we are not 70 kilogram rats. A lot of wrong decisions have been based on this wrong assumption. Pharma companies fail in 95% of the cases they get something into humans.

FLORA LICHTMAN: They fail in 95% of the cases.

THOMAS HARTUNG: Yeah, that’s the failure rate of clinical trials.

FLORA LICHTMAN: Wow.

THOMAS HARTUNG: And if we have anything which is more relevant to humans, which makes them spend their money on better drugs, they want to go for it, even though we have learned the hard way for drugs like Alzheimer that very often, our animal models have been so misleading. If you could cure tons of animals, but we cannot cure patients.

FLORA LICHTMAN: So we need something that approximates humans better than animals. Is that the idea?

THOMAS HARTUNG: That’s exactly the point, and this was theory until most recently, when stem cell technologies really allowed us to produce human relevant models.

FLORA LICHTMAN: So let’s talk about some of these alternative approaches to animals. I know you work on organoids. What is an organoid?

THOMAS HARTUNG: An organoid is a structure which has organ-like properties, and it is self developing out of stem cells.

FLORA LICHTMAN: A human stem cell?

THOMAS HARTUNG: A human stem cell, yes. It is only since 2006, so for bio scientific measures, it is really very recent that we do have ethically not problematic stem cells from humans because Yamanaka developed a technology to take some skin cells, and reprogram them, and make them like an embryonic stem cell.

FLORA LICHTMAN: And so you can use these to make tiny organs?

THOMAS HARTUNG: Exactly. So in our case, for example, we are producing brain organoids, so we’re able to produce thousands of tiny, tiny little balls. They’re just as big as the pin of a needle, and then we can use these and run our experiments as if it was a piece of human brain.

FLORA LICHTMAN: Do they– do they look like tiny brains? Like if I were to look at it under the microscope, would it be like a microscopic version of a brain?

THOMAS HARTUNG: No, not at all. They look like a basketball, I would say. They’re just perfectly round, bit of a rough surface.

FLORA LICHTMAN: Do they match the architecture of the brain in any way?

THOMAS HARTUNG: To some extent. We find, first of all, all of the cells of the brain, which is already a big one because cultures in the past had typically just the neurons, but 50% of the brain is actually help our cells, which is very important. They form circuits, so the cells are talking to each other like the brain cells do.

But also, these helper cells are starting to wrap themselves around the long tubes, which are coming out of these neurons, which are connecting them. We call this myelin, and this is a very, very important feature of the brain. Makes it 100 times faster.

FLORA LICHTMAN: Wow. That is amazing. I’ve also heard about organs on a chip. Is that similar, or are they different?

THOMAS HARTUNG: Yeah, and the chip is not a chip to snack on.

[LAUGHING]

This refers to systems where you can perfuse, which means to get liquid, like blood, flowing through the organ. We can pump liquid through these systems, and by doing so, we can get them larger.

They are much better maintained. They get their oxygen, they get their sugar, and whatever they need. So these systems are highly engineered, very small, like a computer chip, and this allows us to really assemble even combinations of different organs. We are talking now for human organ systems or multi-organ systems.

FLORA LICHTMAN: Oh, wow, so you can link the chips together, like you could link the brain to the lungs to the heart kind of thing?

THOMAS HARTUNG: Exactly. There’s some people who have developed plug and play type of systems, where you can really say, oh, for this experiment, I need a liver and a heart. Yeah, perhaps a brain would not [INAUDIBLE]. So it is fantastic how this disruptive technology has been developing over the last decade.

FLORA LICHTMAN: What part of the drug testing process do you think that these organoids, or organs on a chip, where do you think they could be most effective in the drug testing process? Where do you think they would outperform animals?

THOMAS HARTUNG: I think they are outperforming them instantly because they’re human. You have to understand that half of the drugs which come to the market nowadays are actually human proteins or antibodies against human proteins.

Testing them in animals is mostly useless. Sometimes they work in some monkeys. This holds both for safety. This is one of the important things. We want to know that the drug is safe before we go into humans, but also for efficacy. Is it really helping? Is it curing disease?

FLORA LICHTMAN: Are they cheaper than animals to make?

THOMAS HARTUNG: On the long term, absolutely. So let’s say, oh, brain organoids, for example, they are less than $1 a piece, compared to a rat which is $30 plus, but this is not really relevant. If you develop a drug this is $2.4 billion, before you get to the market, and this does not matter whether you have an additional animal experiment or not. It is important that you put your money on the right thing.

FLORA LICHTMAN: So you’re saying it’s peanuts? That the difference between the cost, it’s whether you can avoid failing?

THOMAS HARTUNG: That’s exactly it. It is all an economical thing and timing. Animal tests can take enormously long, so for example, if you would like to know whether something is producing cancer, it takes you five years to have the result because you’re treating the animals for two years, and then it takes you about two years to cut them into thin slices to find any possible tumor, and with organizing and reporting, that’s easily four to five years.

FLORA LICHTMAN: Is it faster with these organs on a chip or organoids?

THOMAS HARTUNG: In general, yes. These systems are typically done in the month range. You don’t need to wait until it has grown to something you can find with the microscope. Our modern technologies allow us to use the gene expression, for example. What has changed in the cells in order to say this is going to be cancer?

FLORA LICHTMAN: On the show, we’ve talked about how researchers are using– how researchers are using artificial intelligence to find new drugs. Are people using AI for drug safety and efficacy testing?

THOMAS HARTUNG: Yeah, the AI for safety is actually exactly my own background. For efficacy, it is really about generating ideas, finding things which we might not have thought about, and that’s where artificial intelligence is helpful at the moment. About 40% of the drugs which are under development at the moment have been designed with some AI.

For safety, it is fascinating that AI can help us to mine all of the existing information, which means very often, it can find nothing on the substance itself that’s a new chemical often, but it finds something on things which look pretty similar. By looking at all of the chemicals which are similar and all of the properties we know, they are astonishingly good in predicting also the safety of this substance.

FLORA LICHTMAN: So basically, if you have a drug that looks like another drug, you can– and you know that other drug has safety issues, you can then deduce that this drug might have safety issues, and AI can do that analysis faster than anything else.

THOMAS HARTUNG: Exactly. We have a database now which has information on 900,000 chemicals, and it includes about 100 million structures. So you mention all of this information AI can mine and learn from, humans can’t. And whenever you have big data, then AI is shining.

FLORA LICHTMAN: Yeah, I mean, we know that there are problems with big data, too. I mean a big part of the conversation around AI is that it replicates human bias, and we drug testing has had this issue in particular. Like I know there’s been studies showing that because drug dosage trials are tested on men, women have been overmedicated. How do we develop a tool that is more useful than what we already have if we’re pulling on data that has bias in it?

THOMAS HARTUNG: I mean, you have always to look is this plausible? Can the answer be in the data? And we certainly must not hand over to AI, make it autonomous. It is, at the moment, a fantastic tool to get information on a silver platter to take better decisions. It is definitively not yet, and probably will never be, a tool which just you press a button and it says go to the market.

FLORA LICHTMAN: So animal testing started as a safety check, test a new drug in animals, make sure it’s not toxic before it goes to humans. If we skip the animal testing, do you feel like we could be missing a key safety measure?

THOMAS HARTUNG: I’m actually more optimistic that we are actually doing something good here. I mean, the first thing science is incredibly fast in changing. 1, 2 years, our knowledge in the life sciences is doubling.

And completely in contrast to this, these animal tests have been introduced between the 20s of the last century and the 70s. I was not yet born or in kindergarten when these animal tests were introduced, so there must be something in the box which is reflecting us humans better, and this is why I believe it is time for change, and it is a matter of change management more.

The alternative methods, whether it’s these organoids, organ on chip systems, whether it’s AI, they are actually outperforming most of these animal tests.

FLORA LICHTMAN: How far along are we with these alternatives to animal testing? I mean, are we a proof of concept phase? Are pharmaceutical companies already using these techniques?

THOMAS HARTUNG: Yes, pharmaceutical companies use everything, which gives them relevant information and making them move faster. I’m very much in safety. Toxicology is the field. The pharma industry is already spending four times more on cell cultures and computational methods for safety than they spend on animals. It is only that for the final step of registering their drugs, they also have to deliver package with animals.

FLORA LICHTMAN: What is your long term vision? Do you imagine that we’re going to eliminate animal testing altogether?

THOMAS HARTUNG: I would say we will need some animal testing. For example, we have to develop veterinary drugs. So human drugs have to be tested in human trials. You need animal trials.

I can also not really imagine how we would measure effects on behavior in animals in a non-sentient system. So it’s probably some animal testing will stay, but this black box testing, put something in the animal, and wait whether something is happening, and then fighting for the next two years that this is human relevant, that’s not a healthy process.

FLORA LICHTMAN: That’s all the time we have. I’d like to thank my guest, Dr. Thomas Hartung, Director of the Johns Hopkins Center for Alternatives to Animal Testing, based in Baltimore, Maryland.

THOMAS HARTUNG: Thanks for having me.

Copyright © 2023 Science Friday Initiative. All rights reserved. Science Friday transcripts are produced on a tight deadline by 3Play Media. Fidelity to the original aired/published audio or video file might vary, and text might be updated or amended in the future. For the authoritative record of Science Friday’s programming, please visit the original aired/published recording. For terms of use and more information, visit our policies pages at http://www.sciencefriday.com/about/policies/.