Placenta-on-a-chip

The ideal way to study the placenta would be to work with pregnant women themselves. But because doing that isn’t necessarily safe or ethical, science has had to make do with the next best thing. Often, this means using an animal model. They’re the closest we can get to studying the biology of a living being without risking harm to people, but because animal and human physiology is different, the findings often don’t translate. Mouse placentas are not the same as ours, for example. Another alternative is explant culture, where cells are cultured from small pieces of excised tissue. But samples generally can’t be maintained for long, lasting two or three weeks at most.

But what if we could build a model of the placenta’s physiology in the lab to study instead? This is the premise behind a new wave of ‘placentaon- a-chip’ devices: models that replicate biological processes within the placenta, built into a (typically polymer) chip. To mimic the placental environment, these chips use microfluidics – micro-scale systems for tiny amounts of fluid to pass through. At this scale, fluids behave in ways similar to how they would inside the body, and these conditions can prompt cells and other components within the devices to act as they would in vivo, too. It could take years, or even decades, for these models to progress from the lab to having clinical application. But if we get them right, they could be a game changer for maternal health.

Mimicking the placenta

Like many other pieces of lab work, creating a placenta-on-a-chip starts with a research question: what am I trying to find out? “If you are interested in transport across the placenta, you need at least two channels and you need a barrier,” says Hagar Labouta, assistant professor at the University of Manitoba’s College of Pharmacy – who is currently developing a placenta-on-a-chip device. “But if you don’t need to recreate the barrier, for example if you wanted to investigate how a therapy affects the outer layer of the placenta, a simpler model might do,” she adds.

To mimic the placenta’s main function – facilitating the transfer of nutrients and oxygen from the parent’s blood to the foetus – at the very least you’ll need a system with three parts: a compartment containing cultured cells representing the maternal side, another for the foetal side, and a membrane barrier in between them. Then, you need to pass fluid through the system to simulate blood flow and substance transfer.

Many models take a monolayer approach here, says bioengineer and biophysicist Kristina Haase, who is group leader at the European Molecular Biology Laboratory (EMBL) in Barcelona. This allows you to “sandwich” a single layer of foetal and maternal cells on either side of a perforated membrane. This can be helpful, but doesn’t replicate the complexity of the placental barrier. Namely, says Haase, “most models tend to focus on the foetal cells that come into contact with maternal blood (the trophoblast layer) rather than the blood vessels that actually supply the foetus (foetal vasculature). By modelling both, we can get a more complete picture of substance transfer between parent and foetus”.

Haase’s lab is the first to replicate foetal vasculature within a chip and is currently working on adding a trophoblast layer. Using three channels – two outer channels containing cell culture media plus one central port where cells are grown – their chip is 3D printed, with reservoirs containing different levels of fluid. This structure makes it possible to create dynamic conditions just as you’d see inside the body by tweaking fluid flow and pressure. When the team landed on the right conditions, they found that blood vessels began to form. “We can actually establish that the cells self-assemble and organise into a vasculature on their own,” says Haase.

Studying disease development

Once we’ve got a reliable model of the placenta to work with, we can simulate disease states and study what happens. “This is something that we are doing intensively for our projects,” says Labouta. “You want to model those pathological environments, not just the physiological ones.” Preeclampsia is a current research interest within both Labouta’s and Haase’s labs, with both teams working on ways to model the condition within a chip.

Haase’s group plans to put cell samples from both healthy and preeclamptic women into their chip to see how it affects the environment. “For example, [we could] see if the foetal vasculature develops quite differently. And to see whether or not the barrier function develops differently,” she says. “This will provide insight into how the disease can progress and develop.” Tweaking other variables within a device can also help mimic physiological states related to disease. For example, increasing fluid pressure can simulate hypertension, which is strongly associated with preeclampsia.

“To better reflect what happens inside the body, models for studying disease could also create membrane barriers that better limit the size of molecules able to pass through,” says Haase. “What we’ve managed to do is create a barrier that is pretty good, but not perfect. Usually, you would try to make a barrier that would limit anything less than one kilodalton from crossing to the foetal side.” Roughly, this is the size of one large molecule. “This could help us study how infections like malaria affect the placenta,” she adds.

Other labs are already looking into malaria. Last year, researchers from Florida Atlantic University and Schmidt College of Medicine created a placentaon- a-chip device modelling nutrient exchange between parent and embryo, in a malaria-infected environment. Cells were cultured on either side of a barrier – one foetal and one maternal – and researchers observed how infected blood flowed through it. They found that the blood added resistance to the barrier, making it harder for nutrients to pass through.

Testing new therapies

As well as modelling disease states, microfluidic channels can be used to test how various therapies impact the placenta. One way to do this is by deploying nanotherapies – nanoparticles that trigger the immune system into making antibodies – and seeing how they affect target cells on a chip model. Labouta is working on a device that models how nanotherapies used to treat breast cancer in pregnant women affect the placenta. Here, the challenge is to create an environment that resembles angiogenic circulation, she says. This is the process of new blood vessels forming, and it can be brought on by cancerous tumours that need the blood supply to grow. Cancer cells are formed in a 3D structure in one part of the device and then exposed to another channel that mimics angiogenic circulation. Once the nanoparticles are inserted, researchers can observe how they fare once they encounter these conditions. “Because variables inside the body, including flow conditions, can change the properties of the nanoparticles, it’s a question of both whether and in what state they reach the cancer,” Labouta explains. “We would be interested in how those nanoparticles behave, or, what are the identities of those nanoparticles [when] they reach a breast cancer cell?”

Microfluidics can also be used to synthesise the nanoparticles themselves, due to the precise level of control over the conditions within a system they allow. “For example, you could control the shape and stability of a nanoparticle much more finely. Could insights from research like this eventually inform drug development? Potentially, but there’s still a way to go,” says Labouta. “I think we’re still at the very beginning. I think we’re not even close to translation in this case.”

For Haase, creating more models using cultures from different human donors could help us better understand the variability in how disease affects pregnant women. “That will provide further insight into, for example, how antibodies might be used.”

A new route to market

Placenta-on-a-chip devices promise a more robust and reproducible way of studying human biology than what we’re doing now. “That’s probably why so many pharmaceutical companies are interested in them,” notes Haase. They can maintain cell culture for longer than explant methods and could be a more effective replacement for animal models – something Labouta says many in the field are working towards. “A few years ago, we were thinking that it would be a dream for microfluidic systems to substitute animal models of testing. It’s still a big challenge to achieve,” Labouta says.

The FDA appears to be of the same mind. In a landmark move, the agency recently removed the requirement for new drugs to pass animal trials before being approved. While it may take years to fully transition from animal tests to using chips, Labouta is optimistic that we’re heading in the right direction. “We would need less animal experiments, and also I could be submitting grant proposals without proposing animal experiments,” she says. “I hope that in five years, I will be convincing [funders] that placentation in animals is very different than in humans.”