Subscribe
Share
Share
Embed
Sounds of Science is a monthly podcast about beginnings: how a molecule becomes a drug, how a rodent elucidates a disease pathway, how a horseshoe crab morphs into an infection fighter. The podcast is produced by Eureka, the scientific blog of Charles River, a contract research organization for drug discovery and development. Tune in and begin the journey.
Mary Parker:
I am Mary Parker and welcome to this episode of Eureka's Sounds of Science. I'm joined today by a top researcher in the field of organs-on-a-chip, Dr. Kacey Ronaldson-Bouchard from Columbia University. Organ-on-a-chip technology has been around for about a decade, only getting more sophisticated in that time. However, one of the biggest hurdles to replacing traditional models with something like organ-on-a-chip is how siloed they are. Until recently, it was not possible to have more than one organ chip linked together to form a rudimentary version of a body system, heart and lungs, for example. Dr. Ronaldson-Bouchard is one researcher who is looking into combining these models to better mimic the function of the human body. Welcome Kacey.
Dr. Kacey Ronaldson-Bouchard:
Thank you.
Mary Parker:
I'm really glad to have you here and I'm really excited to learn more about your research. So to start, can you tell me a bit about your background? How did you get into this type of research?
Dr. Kacey Ronaldson-Bouchard:
Sure. So I did not start off as a biomedical engineer, which are the typical group that's doing research into organs on a chip, but I had more of a typical pre-med route where I wanted to go to med school and really have an impact on patients through the typical route, being a doctor. And then as I started in that training and I was an EMT in undergrad, I kind of realized that there were a lot of limitations, mostly being that you can only treat patients with the tools that are available or the medicines that are available. And so then somebody kind of brought up, a mentor of mine mentioned tissue engineering, and I was just immediately sold upon looking into it. I applied to my mentor lab, Gordana Vunjak-Novakovic. She was one of the leaders. So I started this journey into biomedical engineering and tissue engineering and specifically focused on how do we create medicine of the future, especially living medicine.
So cells and biologics. And our initial goal as a field for tissue engineering was, you are going to take these human cells, use them as building blocks and create these larger tissues for transplantation. And so I kind of started my PhD in this area where I was building these mini-hearts. And around that time, we're building tissues in the lab and trying to translate them clinically, but we weren't able to build very big tissues, particularly because getting nutrients and perfusion through a tissue without a really complex vasculature, is very hard and we didn't have the sophistication. But we were really good at building really, really small tissues. And so the whole field started to pivot and say, "What can we use this for?" And around the same time, drug developers and the FDA and the government were also saying, "How can we have these human models for drug discovery?"
And so everything kind of merged to apply tissue engineering to making these small models of human cells and organs in a Petri dish for things like drug discovery and personalized medicine. And also around the same time, fabrication techniques related to computer chips and the semiconductor field were being used by biomedical engineers. And so we were using them to make the homes for cells. And so then organs-on-a-chip kind of was born from that, using chips from the semiconductor computer chips name. And it just took this life of its own and everything merged around the same time, where you had tissue engineers, drug developers and biologists really wanting to build this future together. I think it was the first time ... When the grants came out for tissue chips from the NIH, I was in grad school and we worked on one of the first grants and I think it was the first time that multiple agencies had put all their money together to fund one initiative. And so that was really exciting to be a part on.
Mary Parker:
Yeah, I mean obviously the interdisciplinary approach is super important for a field like this because it isn't just biology, like you said, it is also engineering and all these other things wrapped up together. So it's really cool when those sort of things come together. So getting into it, what is an organ-on-a-chip? How is it made?
Dr. Kacey Ronaldson-Bouchard:
So there's kind of a raw definition for organs-on-a-chip, but most of it is using PDMS, so the silicon that's molded into a certain shape using the semi-computer chip fabrication techniques. So it's essentially just this small scale bioengineered system and it's meant to replicate the cellular architecture. You'll have hydrogels making the micro environment, and then any mechanical forces that you would find in human tissues. And so you make the environment and then you culture living cells in it, mostly providing perfusion, so microfluidics. You can make these microfluidics, which are really small channels, by having these molds made from the computer chip industry. And so the whole setup allows you to simulate complex biological processes in this way that you're not getting from typical cell culture methods, which are mostly just flat plastic Petri dishes, things like that, until we can control the environment using these moldable plastics.
Mary Parker:
I was reading a paper that I believe you were one of the authors on and it mentioned the importance of starting your study by figuring out how complex the organ-on-a-chip needs to be. So for whatever it is that you want to get out of it, you can make the organ more or less complex, depending on cost and other limitations. Can you talk a little bit about how you would make that determination at the beginning of a study?
Dr. Kacey Ronaldson-Bouchard:
Yeah, we mostly try to define what's the question you're going to answer. Because every time you're introducing more cells or any kind of complexity into a system, biology in itself is complex and almost chaotic, and it's hard to really focus on what you're trying to get out of it if you add too much in. And so really defining the goal of your research question: I want to model liver toxicity of this drug. And so if it's just modeling liver toxicity, you don't necessarily need all those other organs, you just want that liver. And so we try to think of it as building modular systems when we design our organs-on-a-chip, so that you can plug and play the different pieces you need to answer the question or as the question grows. So I always like to think of if we're looking at a drug discovery process, I might only start with one organ at first and I'm looking, how does the drug act in my single organ of interest?
Is it having the desired effect? What's the mechanism of action? And if yes, it's working how I think it's [inaudible 00:06:38] or it's intended to work, then I move it to the next stage and I'll combine it with something like the liver. How does my drug now work in the target organ when it's metabolized by the liver? And then I can add ... And so if it works fine at that checkpoint, I can add the heart for cardiac toxicity, the kidney for filtration and absorption. And so you can keep adding pieces so you are efficient. So for drug development, you're failing early, you're failing fast, so you can keep getting more drugs through the pipeline, and then for research you can focus in on what you're trying to ask and get rid of all the complex things that might interfere.
Mary Parker:
That makes sense. But what started researchers on this organ-on-a-chip path? Were animal models not cutting it in some cases?
Dr. Kacey Ronaldson-Bouchard:
Yeah. Specifically, so it wasn't necessarily that animal models weren't cutting it. So you saw, I think it was in the nineties, more strict regulations get introduced by the FDA to stop drugs getting to market that were unsafe. And so it most likely is that drugs weren't going to market that are safe and effective because they got flagged for being potentially toxic earlier on. And then you started to see drug development just get more bloated and it would take about 10 years and $1 million to get a drug to market. And every out of nine drugs that made it to clinical trials, only one went through.
And so it just got more and more expensive and longer. And so the key component is how do we get more drugs to market for more diseases, and make it cheaper. And looking for ... And so the animal models you have today or you had then, weren't cutting it, we needed new ways to answer questions. And so organs-on-a-chip just provided that tool and everybody was pretty eager to see what these tools could do and what questions they could answer, and try to develop them with the use cases in mind, but also very skeptical because you like the models you currently use because you know how they act. So a big initiative was to define a set of standards that organs-on-a-chip researchers and developers would use and to also really try to understand what they can do and what they can't do so you can effectively integrate them into your pipelines.
Mary Parker:
Do you think that they'll start getting more defined in terms of what they can and can't do after ... I'm just thinking about they pass the FDA Modernization Act, which means that you no longer are required to use two animal species for drug approvals. Do you think that might lead to better guidelines for what organs-on-a-chip can be used for so that they could potentially replace more traditional models?
Dr. Kacey Ronaldson-Bouchard:
Yes, absolutely. So a case where I've spent a lot of time working and where you might clearly see this in the future is looking at cardiac safety. And so one of the main issues with animal models, especially rodents, their heart rates are so much higher than humans, and it's typically when you have either a low heart rate or a fast one where you get drug arrhythmias and things like that. And so this difference in heart rate can really skew the data when you're looking at things like how a drug would affect heart function. And there's this specific assay called the long QT assay, which is meant to catch arrhythmias, especially these dangerous arrhythmias. And it has helped a lot of drugs not get to market. You haven't had a drug get to market that was unsafe for this case since this guideline was introduced.
And so that's great, but it's probably preventing a lot of drugs from going through and it's really costly. Your drug can get just caught up in repeating this long QT assay for a couple of years, and spending more and more money because looking for an arrhythmia is looking for a needle in a haystack. And so while this guideline is great for the public and the population, it's keeping us safe, it's also where you might see these human models really, really come in. And so you have these cardiac-on-a-chip platforms, they're already being used in pharma and being well-defined for these specific assays. So there are working groups, one of them is the SIPA, so it's looking at these proarrhythmias and they're defining these use cases. They're sending researchers drugs that are blinded and having them send back the results, and seeing if they can predict, these drugs are high risk, these drugs are moderate, these drugs are low. So people really are coming together having defined use cases and understanding what the models can do.
Mary Parker:
That's awesome. That's going to save so much money in the future. I mean, having a drug be able to be in the early stages, identified as high risk, when in the past it might have gone through is going to be able to save the money and put more resources into different drugs that might work better for people.
Dr. Kacey Ronaldson-Bouchard:
Yeah, there could be many drugs that are just essentially sitting on the shelf that are safe but just never made it through because you had a red flag pop up for potential safety or even efficacy. It didn't work in an animal model. But we're not animals. And so that drug could be great. Aspirin probably wouldn't have made it to market and it's one of the most widely used drugs, but if it went through the same drug screening system you have today, it wouldn't have made it. And so it's looking at it from both sides. Are there drugs that we're not developing because we are worried about them. And then also how do we make sure no unsafe drugs get to market.
Mary Parker:
So getting into the multi-chip system, how is it possible to link organ chips together and what kinds of organ chips work best in a series?
Dr. Kacey Ronaldson-Bouchard:
Sure. So there's a huge need for these models of systemic human disease, so affecting multiple organs, especially because these diseases are very hard to replicate in animals. And so if you're going to make a systemic model, which is one of the benefits of animals, they're inherently systemic. They're a whole multi-organ system. We need to replicate that with our human organ-on-a-chip approach. And so there's, if you want to link models together to have this complex system, there are two approaches, mostly sharing media so you're going to mimic the blood flow. So the first approach is having organs share media. So media is a fluid that cells and these tissues are sitting in and it has all the nutrients they need and the tissues and organs are secreting things into this as well. So you can have your perfusion and microfluidics go through each organ and just share media, so you're sharing the fluid that you're communicating with.
Or the other approach, which is the approach we focus on, is having the different organ systems talk to each other through the vasculature. So we mostly take a biomimetic approach. So meaning everything we build, we are using biology's playbook as the roadmap. So we're trying to mimic what's going on in biology. And if you look in your body, your organs are located separately, but they're talking to each other or connected through vasculature. So we wanted to replicate that. And so we made this huge vasculature channel. And then at the top of each channel, you can plug in an organ of choice. So you have this well or this niche which is essentially just the home the tissue lives in, and you can tailor it to the individual organ requirements, but it has its own dedicated media that's ideal for that organ. And then right next to it you have another well or tissue niche where you can plug in a different tissue and you can keep plugging in tissues.
So both approaches allow you to of connect as many tissues as you would want in a row by either sharing the media or having vasculature go under the different tissue niches or through them. And so you can really optimize for your question. And so we typically take the approach of the vasculature, because we want separated media for our individual organs. And that's because we only work with stem cells. And when you're taking a stem cell and differentiating it or making it become a committed cell like a cardiac cardiomyocyte or a liver hepatocyte, it is very reliant on the media that it's cultured in. cardiomyocyteAnd so if you were to have a common media or shared media, you have to have that media suit all the organs. And we didn't want to compromise the maturity of our differentiated organs because we really work a lot at getting these stem cell models to be more mature, since one of their main drawbacks for drug discovery is that when you use stem cell models, they're not as physiologically relevant as we want them to be, because you're starting from a stem cell, so very early on in development, and we are mostly using drugs in adults that have had 18 plus years of development, and so we don't have full maturation.
And so when you link tissues, you want to make sure that you're keeping the original model as biologically relevant as you can. And then in terms of what organs work best to link together, that just goes back to the question being answered. So you always want to keep them as simple as you can, focus on the main organs for your question or purpose. But typically for organs-on-a-chip, you see the liver, heart, lungs, and kidneys as the most commonly used ones because the liver is critical for drug metabolism, heart for cardiac function, and every drug has to be safe for the heart. Respiratory processes, so lung is the most seen and was the first developed by Dan Huh and Don Ingber. It was that lung-on-a-chip which was the seminal paper for organs-on-a-chip, that got everything started, but you also had earlier works from organ-on-a-chip as well. It just wasn't really coined that. But still these microfluidic systems. And so which ones you link is just really the ones available and what your question is.
Mary Parker:
So for example, if you wanted to see how well something is filtered in the liver before it goes to the heart, you could do the liver going to the heart sort of model and just link those two. Okay, cool.
Dr. Kacey Ronaldson-Bouchard:
And your route of delivery. So you can get drugs in different ways. And so we mainly think of them in three ways. So through the vasculature intravenously; through the skin, so drugs are absorbed through the skin or you're inhaling things through the lungs. Gut systems are available. They're just not widely linked with other systems at the moment because gut has bacteria. This is a burgeoning area, but that's the main route of typical drugs as it's you're ingesting it, but most organ chips skip that and go straight to getting it into the vasculature. And so you also pick your organs based on your route of delivery.
Mary Parker:
Yeah, that makes sense. How many chips can you link together so far?
Dr. Kacey Ronaldson-Bouchard:
We typically work with either two or four, because that's where we find our complexity maxes out. And we haven't really had a situation where we've wanted to model more than that, probably because the different diseases or use cases we're specifically looking into, but other teams have connected as many as 10, all in a row. But typically you're going to see two because you have your organ of interest and then your auxiliary organ like the liver or the heart, and then four if you want to have multiple auxiliary organs, so your liver, kidney and heart plus your organ of interest.
Mary Parker:
And how difficult would it be to add nerve models to the mix? Could you potentially predict pain symptoms from a drug if you could have nerves-on-a-chip?
Dr. Kacey Ronaldson-Bouchard:
So integrating nerves-on-a-chip and innervating the tissues is challenging, but certainly feasible. And lots of people are doing it. Nerve cells and neurons, they're quite sensitive and they require very specific conditions to survive and function, so even more so than just stem cells. However, we're getting quite good at microfabricating environments, and you can take a couple of approaches to integrating or innervating your tissues.
So you can think of it as installing the electrical wiring in your house where your nerves are the wiring and your organ's the house. So we don't really have the tools to just easily go in and do this with high resolution. So you either grow the nerves from the beginning by mixing it in with your tissue and letting it self form, or you'll grow the nerves separately and then add them to the tissue. And this is kind of the approach that works the best, is having two tissue homes right next to each other and one is the nerve home and one is the tissue home, so let's say a muscle. And so you have these two wells right next to each other, but then you'll create these small paths between them so the nerves can grow into and reach the tissue.
And so this is really useful for studying things like neuromuscular junction. You can have optimal media conditions for both, but it provides this physical support system. And you can also send a signal on the nerve side and watch it go through the nerves, through the neuromuscular junction and actually make that muscle contract.
And there's been a lot of work in funding nerve models and developing them because there's a huge need for models that can both predict neurotoxicity or the painful side effects from drugs, and also pain. And so things like [inaudible 00:20:28], and study how do we develop drugs that effectively relieve pain but are safe, in particular when it comes to avoiding addiction and these addictive loops that some drugs tend to activate and most pain drugs currently do. So we want to have drugs that manage pain effectively, but don't activate those negative pathways that are really unsafe.
Mary Parker:
How can these linked chips help with studying, say, metastasis and patient immune response to cancer drugs? I understand you guys do a lot of work with oncology.
Dr. Kacey Ronaldson-Bouchard:
So going back to how many organs should you link together? we started just talking to people developers, drug developers, researchers, and seeing when the use case for more than one organ would come up. And time and time again it came up for metastasis, metastasis, metastasis. This is a clear case where you want more than one organ. And so we had already been doing work in the cancer field and modeling just cancer-in-a-chip. And so we really shifted to let's use our multi-organ system here and really study it. And so you can mimic the interactions between the cells and also understand where they're going to go. So if you're looking at metastasis, there are two main theories that explain how cancer cells spread. And these might be completely wrong theories, but these are the ones people go with.
So first you have the seed and soil theory, and so it suggests that the metastatic cells are the seeds and they have a preference for certain environments which are the soil. And you do see this in a lot of cancers where breast cancer cells tend to metastasize here, and prostate cancer cells tend to metastasize elsewhere. And so if we look at prostate cancer cells, they love to go to the bone. They really like that niche. If you look at prostate cancer animal models, they do not go to the bone. They want to go anywhere else. And so that was a clear case where there's this discrepancy between animal models and humans.
And so you have the seed and soil, we can plug in the metastatic cells and have all these organ chips and see why do they pick the bone over the liver? Why do cells never really metastasize to the heart? Is it the micro environment? Are there certain growth factors or signals that the cells are giving out? Is it the extracellular matrix? What helps cancer cells thrive in a new environment and where do they not go and how can we harness that? And so you can kind of figure out they like the bone because they like to break down the bone and absorb the good things from the bone, but they might not like another tissue because it's not good soil.
And then on the other hand, you have the other theory is more of like a mechanical or anatomical. And so it says that the spread of metastatic cells just follows the patterns of blood flow and lymphatic drainage. And so it's going to occur in the organs that receive most of the blood flow, especially when you look at where the primary tumor is and then what's downstream. And this also rings true because the liver and lungs are really common types of metastasis and those are your key filtration points for blood returning from everywhere in the body. And so it's not that these theories are mutually exclusive, but they compliment each other and there's not really been a great way to see, I have this cancer cell, it's leaving the primary tumor, it's going into the circulation and then it's going here.
We can't really zoom in on a mouse and distract these cells without opening up the mouse, and we're definitely not going to do that with humans. But you have these multi-organomic chip systems that let you answer all these questions, look at these theories, but they give you this really comprehensive understanding of how metastasis might occur. There's so many unanswered questions on how cancer forms, how it evolves, how it spreads. We've used animal models for many years, having the organ-on-a-chip just as this really great tool and everybody we've worked with who does animal models, they're never really opposed to switching. They're usually super grateful to have this other tool.
Mary Parker:
So how might these systems be useful when studying drug interactions?
Dr. Kacey Ronaldson-Bouchard:
So they're great for studying drug interactions and especially the circulation of drugs and where they're going. And so you can link things like the liver and kidney and identify these drug-drug interactions, any toxic effects, and then optimize ... One of the main things you can do is optimize the drug dosing and combinations. And so you're getting the safer, more effective treatments. And particularly when you're going into a phase one clinical trial, just to look at safety, you're not really able to tinker with dosing and drug-drug interactions. And so here you have this parallel human platform that lets you optimize that, because the long-term goal is you want to get every patient the right drug at the right time and at the right dose and the right drug-drug combination. And so there's a ton of different combinations that you want to optimize, and that is quite hard to do in the models that exist being either human or animal models.
Mary Parker:
That leads to the next part of my question, which is can you imagine this technology being used for personalized medicine like taking drugs that are already approved and testing them on a patient's cells to see which one they might best respond to?
Dr. Kacey Ronaldson-Bouchard:
Absolutely. I think it's a really promising application of it. In cancer, you're seeing that done because one of the limitations in actually implementing this for personalized medicine is the cells. So getting the personalized cells. So for cancer, the cells exist. You can take them from the body and put them right into the platform. For a lot of other tissues or organs, you're not really just taking them directly from the patient, you're taking like a blood cell and then transforming that into an induced pluripotent stem cell and then taking that stem cell and growing all your other organs. And so that just takes time and money right now. And so as that process gets faster, then it'll be more reasonable to make everybody's personalized avatar in a Petri dish for precision medicine. But for right now, it really applies well to things like cancer and the immune system where you have this direct access to the cells and this lack of animal models.
Mary Parker:
What is the potential for these systems to replace traditional animal and cell models? Where might they fall short and where might they be a big improvement over what we already use?
Dr. Kacey Ronaldson-Bouchard:
So I think they have the potential to really reduce our reliance on animal models. Right now, it's pretty much the only way you can answer questions for drug development or precision medicine, is having animal model data. There's this kind of this joke that Reviewer Three is always asking for more animal model data, but with these human systems you have sometimes a quicker tool that's more of a better fit for your question, readily available. So it's not just, oh, I need an animal model. Because if there's an animal model that doesn't exist, you're making it from scratch and this can take quite some time, but it doesn't take as long for you to build the organ-on-a-chip model. So not only do you have more models available, but they are human. They're inherently human, and so you're getting human responses, but they're not fully complex like the animals.
And so you have this give and take, and I think you'll use them in this complementary approach, and you're not going to completely replace animal models, but you're going to use organs on a chip as a way to answer questions you can't with animal models and then reduce their usage. So I guess I'd say they fall short for full systemic interactions. Long-term studies, you can keep organs-on-a-chip alive for a month, three months, but people haven't really been keeping them alive for much more than that, a year max, people have done. And then reproductive processes like developmental toxicology where you need this full reproductive process to occur, you're not going to have an embryo grow in a Petri dish or an organ-on-a-chip model, so you would need an animal model for that.
And then I think the areas for improvement are mostly centered around you now have this ability to have more information so you can make more informed decisions earlier on in drug development or even research in the lab. So you can take a lean approach to moving forward where you're failing early, failing fast, or answering questions really early on, rather than doing all this work, spending all this time and money and then having it not translate. And so you have more information available. They're human, you can have these genetic models. You can also genetically make a cell without the mutation, so an isogenic cell. So you can look into just what the mutation is doing specifically, and how that compares to what that same cell with the same genetic background would do without that mutation. And so these are easy to do in just a monolayer or an organ-on-a-chip system.
And then looking at disease progression. So in an organ-on-a-chip model, you can make this model of an organ and its disease and you can essentially see into it at all points and time in the progression of the disease. You are not really opening up animals and humans and being able to see with the resolution that we can see in an organ-on-a-chip system. And so you get a lot more information about how these cells interact and fibrosis grows over time or systemic inflammation, any disease, how it actually progresses, is nicely modeled in organ-on-a-chip or in a Petri dish, because you have direct insight into what's going on. And then you can also play with things like, if I remove this factor or add this factor, what's the effect on the progression?
Mary Parker:
Very cool. This is really, really cool stuff. I'm so grateful that you were able to take time to talk with us about it. I appreciate you coming on the podcast.
Dr. Kacey Ronaldson-Bouchard:
Yeah, thanks so much for having me. I love talking about this and I'm really excited to see where the field goes and what comes next even.
Mary Parker:
Yeah, definitely. Me too. Thank you so much.