E81: Biomedical Engineering: The Future of Research

As researchers continue to look for new ways to test therapeutics in vitro, the field of biomedical engineering is offering new microphysiological systems that can reduce the need for animal models. Hear from Jeffrey Bornstein, Director of the Biomedical Engineering Centre at Draper, as he explains how it works, ways to expedite its access, and the benefits they will ultimately bring to patients.

Show Notes

• The Coming of Age of Organ-on-a-Chip Technology
• E76: Multi-Organ Chip Systems
• Charles River Joins Consortium to Advance Organ-On-A-Chip Technology
• Bold New Approaches: The Future of Organoid Intelligence
• Organoids Evolve from Academic Marvel to Industrial Tool
• Alternative Methods Advancement Project (AMAP)

• Podcast Transcript

  Jeffrey Borenstein (00:04):
  When we think about the current drug development process, we think about evaluations that are done in vitro in the laboratory and then in animal models. We think about the fact that those systems do not necessarily translate very well to human responses, and that's one of the reasons that we're looking at organoid chips and that there's so much interest in this technology.

  Mary Parker (00:37):
  I'm Mary Parker, and welcome to this episode of Eureka's Sounds of Science. I'm joined today by Dr. Jeffrey Borenstein from Draper. He's here to discuss the broad applications of biomedical engineering and bringing his years of experience to translating research into real treatments. Welcome, Jeffrey.

  Jeffrey Borenstein (00:54):
  Thank you. Great to speak with you.

  Mary Parker (00:57):
  So let's start with one of the coolest modern research innovations, in my opinion, at least microphysiological systems like for example, organs on a chip. So how do you define microphysiological systems?

  Jeffrey Borenstein (01:09):
  Yes, that's a great question. I think that people define microphysiological systems as systems that comprise engineered microenvironments with cells that have effects that recapitulate key aspects of the vivo physiology in humans. So basically you're using the cues that are enabled by microsystems, by microfabrication technologies to recapitulate effects in cells that render them similar to the tissues and organs that are found in the body.

  Mary Parker (01:44):
  So when we say something like organ on a chip like a kidney or a liver, we're not actually saying that the entire kidney or liver is on a little chip, but that parts of it can mimic what might actually happen in the body and we can test drugs on those instead of in vivo.

  Jeffrey Borenstein (02:02):
  That's right. When you think about a liver or a kidney or a lung, there are certain building blocks that are represented by small numbers of cells in relation to one another in structural and functional relationships to one another. And the idea with organ on chip or microphysiological systems is to construct a cellular microenvironment that recapitulates or that simulates those key aspects of the organs. So you don't need the entire organ or even all of the features of the organ to replicate the key aspects of what the liver, the lung, or the kidney might be doing.

  Mary Parker (02:41):
  Is it true to say that when you have these small systems, they might actually be most useful for giving you a hard no on a treatment? It seems like you're not going to get the full picture. There might be some things that come up later down the line in safety, but a system like this might tell you whether or not away a drug might be efficacious or might be too toxic.

  Jeffrey Borenstein (03:03):
  Yes, that's one of the most important applications of the technology. I think that microphysiological systems, they can perform lots of key functions in terms of understanding the body's reaction to injury or insult diseases and to understand the response the body has with regard to certain treatments. But I think that when you look at the way that microphysiological systems work, they can give you predictions of the efficacy or toxicity of a compound potentially very quickly, that if a compound is toxic to human cells, let's say, then the microphysiological system can indicate that very quickly and very easily in some cases. In other cases, it requires a more complex type of framework or interactions between different types of cells or tissues in order to do that.

  Mary Parker (04:01):
  But the worst case scenario is finding out that it's going to just definitely be toxic once you're already far into the research process. That's a big waste of time and money for the researcher.

  Jeffrey Borenstein (04:12):
  That's exactly right. And I think that the earliest applications of these microphysiological systems or organ on chip systems were in the safety and toxicology area. They were really understanding whether compounds and they could have been chemicals or they could be drugs, whether they were toxic to human cells. And those were kind of the earliest studies. And now I think studies of the efficacy of these compounds within the context of a disease model that's really taking root as an even larger opportunity for mysiological systems.

  Mary Parker (04:49):
  Speaking of which, in what ways can culturing these organ specific tissues from human derived cells change the way diseases are studied and treated?

  Jeffrey Borenstein (04:59):
  So I think that's a great question and that's really hits at one of the tremendous opportunities of organ on chip technology. When we think about the current drug development process, we think about evaluations that are done in vitro in the laboratory and then in animal models. We think about the fact that those systems do not necessarily translate very well to human responses. And that's one of the reasons that we're looking at Organon chips and that there's so much interest in this technology. So one of the most fundamental things that Organon chips can do is to give you particular responses that humans would be expected to have to a compound or to a disease. But I think when you take that to the next level, and that really gets at the future of this technology in precision medicine, you can look at the responses of particular demographic populations or populations that may suffer from certain diseases or have certain conditions called comorbidities and how those populations will respond to a disease or to a drug. And I think an example of that is if you look at particular populations such as patients who have diabetes or heart disease, they might respond differently to an infection than a healthy patient. And you can look at that in the context of these organs on chips by using cells that come from those particular patients.

  Mary Parker (06:29):
  That's a really good point. You don't think so much about these sort of advanced technologies being, you mostly think of them as being used to study new diseases and come up with brand new treatments for cancer or other rare diseases. You don't think of them as much being used to figure out the best kind of antibiotic to give someone who has certain conditions. When you have a million drugs to treat one thing, it is a really good thing to be able to know which one is going to help each individual type of patient.

  Jeffrey Borenstein (06:57):
  Yes, exactly. And I think that that's really one of the key opportunities here is that we know that animal studies have been very powerful in drug development throughout the course of the history of the field. And we know also that there have been a great utility made with immortalized cell lines that in spite of the limitations, these technologies have taken us a long way with regard to drug development. But one of the things that they fundamentally cannot do is to reflect the nature of a disease or a drug treatment against the particular features of a population of humans. And that population of humans may be by gender, it may be by age, it may be by certain ethnic backgrounds or people who suffer from particular types of diseases. And so there really is no opportunity with animals or with immortalized cell lines to get into the deeper aspects of how the disease and the treatment are affecting populations of patients that have specific characteristics.

  Mary Parker (08:10):
  That makes perfect sense. And kind of segueing into our next topic, this obviously would create a huge amount of data for human scientists to have to go through. So can you explain how machine learning can be leveraged or be useful for physiological systems?

  Jeffrey Borenstein (08:27):
  This topic of machine learning, thinking about the whole field of artificial intelligence as we speak, leaders in the field of AI are all gathering and our nation's capital to talk about the future of this technology, the challenges and opportunities that it presents. And I think this is a tremendous area for drug development. And I think what we're looking at is how do we combine ai, machine learning, deep learning, neural networks, kind of the whole hierarchy of these technologies with microphysiological systems and the machine learning kind of capabilities can operate at various stages of the drug development process. So before you get to the microphysiological system, they may be playing a role of identifying candidate compounds by doing modeling, protein folding and various complex models that look at how the candidate drugs may target specific receptors in the body. And so there's that front end where the AI is creating kind of the candidate libraries of compounds and then further downstream, these microphysiological systems are generating huge amounts of data. And so it's really going to take kind of an AI framework to be able to handle that data to be able to draw from those huge data sets conclusions regarding particular treatments, how well they're working and mechanisms of disease and treatment.

  Mary Parker (10:02):
  Yeah, I think it's always good to point out whenever we bring up machine learning or AI that when it's used for a scientific purpose like this, it is not the same thing as AI generated art or chat GPT or the things that people write funny headlines about. This is basically just doing what a computer was meant to do, which is large scale data processing. It's not being asked to create anything that's still being left to the humans. It's just analyzing in a way that a human brain can't fit all those data points in at once.

  Jeffrey Borenstein (10:34):
  Absolutely. When you think about the way that AI is working in drug development, a lot of these powerful new tools that have emerged from laboratories across the country and the world, what they've done is they've analyzed the sort of biophysics involved in disease processes and in response to treatments and using a lot of complex biology and chemistry and physics within these models, which are tremendously computationally intensive, they come up with a framework where you can make predictions. And those predictions are extremely useful because what they do is they help us narrow down the scope of the experiments that we have to do so that we can kind of focus in on certain types of treatments and certain mechanisms of action and rule out others instead of having to do all those untold numbers of experiments without having the guidance of the AI tools.

  Mary Parker (11:38):
  So what disease areas can these approaches impact?

  Jeffrey Borenstein (11:42):
  Very important question. So I think that there are an almost unlimited number of examples of how these disease areas, various disease areas could be impacted by this technology. So I'll focus on a couple right here. I think one of the first ones that comes to mind is the area of infectious diseases research. And I think as we all know from the COVID-19 pandemic experience, this is a critically important area. And so one of the questions that I often talk about when I'm giving presentations on this topic is when we think about the tremendous technical advances that were made in vaccines, we note that in the therapeutics and the treatment for patients who are sick with something like COVID-19, that process proceeded more slowly for the field of drug development. And it wasn't until a couple of years into the pandemic that we really had identified treatments that were available orally dispensable treatments that worked against COVID-19.

  (12:52):
  So what we've been doing at our lab at Draper is modeling the SARS cov to two infection, the pathogen that causes COVID-19, and modeling that in essentially a lung microphysiological system. And kind of just a couple of months into the pandemic, we were modeling this infection and you have to do it in what's called a high containment laboratory because sars cov to two is a very dangerous pathogen. You can't work with it kind of out on an open bench. So you have to work in these high containment laboratories and you essentially infect the lung on a chip or the lung microphysiological system with the pathogen with the virus, and then the virus replicates and you observe how the virus will replicate depending on whether it's a particular variant of COVID-19 versus another variant. And then you introduce specific treatments. And so I think you can sort of see a paradigm shift from the days where you would only rely on animal models.

  (14:00):
  For instance, you would infect certain species with COVID-19 and then treat them with a drug. And that's limited because of the predictive power of animal versus human responses, the throughput and kind of the ability to study, probe these mechanisms of disease. And now we can sort of envision doing this on a well plate based system, an instrumented well plate, which is essentially a microphysiological mimic of a lung. And then you can study dozens of treatments in a single experiment and identify which ones actually have an antiviral action against the infection and which ones do not. And so I think that's a powerful example of how you can go from the previous paradigm of animal model based systems to a human based system that has high throughput in is laboratory based. That will give you much more accurate predictions of what's going to happen in the clinic.

  Mary Parker (15:07):
  Theoretically in the future, we could replace animals entirely. I think for the present, we still like to use them for looking at the whole systemic issue of a drug being introduced. But before you get to that point, these would be an excellent way to work out all the kinks before you even get to the regulatory level where you start needing to use those models.

  Jeffrey Borenstein (15:29):
  That's right. And I think a lot of the people that we talk to and the stakeholders in the field are really interested in developing these microphysiological systems to create analogs, laboratory analogs for animal studies. So the vision isn't necessarily replacing the animal models, but it's creating a laboratory analog so that you can look at the correlation between an animal response and the response of a microphysiological system that is cultured with animal cells. And that gives you kind of a very powerful tool to be able to understand how these platforms, these engineered platforms, can assist you in predictive power for safety and efficacy. Because we're still a long way from the ability of a microphysiological system to predict all of the complex systemic effects that you can see in an organism such as an animal model.

  Mary Parker (16:28):
  How can organ on a chip technology and other physiological systems be expedited so that organizations like Charles River can make it an accessible model for our clients?

  Jeffrey Borenstein (16:37):
  Yes, that's a great question. So I think one of the key aspects here is how do we scale these technologies? Because I think a lot of people who've seen sort of reports of these physiological systems in the news or in the literature, they kind of envision them as small scale research devices. And so they're very useful. They may recapitulate some very interesting effect like the breathing mechanism, the breathing action in the lung or the beating of the heart at a cellular scale. But I think that the key here is going to be to scale these up because drug development is not done in small research scale devices. Drug development is really done in higher throughput systems, multi-well plates that may be 96, 3 84 or 1536 well plates. And so we really need to take the microphysiological systems to that next level of throughput and also validation in order for companies like Charles River Laboratories to be able to make use of those in the pipeline. And I think that's been one of the key challenges for physiological systems is moving into that sort of higher throughput and more robust operation.

  Mary Parker (18:01):
  In what ways can CROs like us support the evolution of next generation drug development technology? From your perspective as a researcher, what could we do better?

  Jeffrey Borenstein (18:11):
  So A CRO like yours can play a very powerful role in this entire cycle of technology development. And I think we've even seen examples of that already. I think that one of the things that we can do is to generate data sets that allow us to validate these microphysiological systems. And so if we think about the microphysiological systems that might predict cancer drug efficacy, for instance, we think about the fact that the current cancer development pipeline relies very heavily on syngenetic mouse models, for instance. And so one of the key aspects is to be able to look at how particular drugs, and they may be small molecules, chemotherapies, they may be immune checkpoint inhibitors, a very powerful new emerging approach in cancer therapy. And we can look at how these treatments fare in animal studies and then compare them side by side with microphysiological systems in kind of a benchmarking kind of a way. Because I think that both the CROs like yourselves and the regulatory agencies, the drug developers, the drug companies are all eager to see how well validated these microphysiological systems are so that they can start relying on them in their processes and in their operations. So I think that benchmarking and that validation is something that A CRO like Charles River Laboratories is ideally suited to provide.

  Mary Parker (19:55):
  Have you done any research yourself on regeneration and have you used physiological systems for that?

  Jeffrey Borenstein (20:03):
  That's a very interesting question. So I think that what that question kind of touches on is the origins of the field of organs on chips. So I could even say personally for myself, when I started in the field of tissue engineering and regenerative medicine over 25 years ago, the goal was really laboratory grown organs and tissues for end stage organ failure and for injury and disease treatment. And so the vision at that time was to take these technologies, these engineered tissue constructs and scale them up to human organ scale for of a range of diseases, let's say lung, liver, kidney failure, those kinds of treatments, heart failure. And so the field of microphysiological systems in many ways evolved from that field of regenerative medicine because I think that people realized our cells and our collaborators among them that well before the time that you could grow an entire replacement liver in the laboratory, you could create mini livers, if you will, that could be used for testing drugs that might treat liver diseases such as liver fibrosis or something like that.

  (21:29):
  So it was about 20 years ago that we started working in the field of organ on chip where we took those tissue engineering technologies and started applying them in collaboration with drug companies. And I think that your question is an interesting one because then you can imagine that coming back full circle where you can take these microphysiological systems and study how they might be useful in the whole field of tissue regeneration and how one might treat, for instance, tissue or organ loss due to burn injury or other disease states, and then create kind of a regeneration scaffold or platform for those patients.

  Mary Parker (22:13):
  I remember many years ago people getting very excited about 3D printing and talking about 3D printing organs and things like that. That seems to have not really materialized, but that doesn't mean that following that line of research didn't lead to anything useful. I mean, you just zigged and zagged to the point of getting to being able to use these organs on a chip for the purposes you just described.

  Jeffrey Borenstein (22:37):
  Yes. I think 3D printing is a very interesting element to consider here because sort of in its most basic form, I think 3D printing can be used to create the scaffold itself. And so when one thinks about a microphysiological system, how do you create those scaffolds? And the scaffold is supposed to recapitulate in some fashion the way that cells and tissues are structured and organized in the body. So many people use lithography based techniques to try to achieve those microenvironments. But I think 3D printing represents another possibility in another opportunity to create this 3D architecture in terms of a scaffold that will then guide tissue development and replicate let's say a key structure in the liver, the lung, the kidney, the heart. But then when we think about 3D bioprinting, we're taking that to the next level where we're not just printing the scaffold, but we're depositing the cells in a precisely arranged manner within that scaffold.

  (23:43):
  And so I think you're right, that is a technology that is still being developed. It's a very exciting technology. I think it's not really the main focus right now of the microphysiological systems field, which is much more looking at these kind of plate based systems where you can organize the cells in a roughly planer configuration and introduce flow and integrate sensors and these kinds of things. But I think it's all part of the future where we're bringing more and more of these powerful technologies to the table to ever more closely achieve a recapitulation of the physiology of human organs.

  Mary Parker (24:29):
  And we kind of touched on this earlier, but it seems like a more realistic near goal would be personalized medicine, like using these on, say, someone who has liver failure. What would be the best drug treatment for them?

  Jeffrey Borenstein (24:43):
  Yes. One of the most exciting opportunities I think in the entire field is in this whole precision medicine arena and maybe no more compelling example than in cancer therapy. And so we started a project many years ago and have been continuing on this path and working towards the goal of creating systems that can predict the personalized cancer treatment for particular patients. And so the vision here is to when the cancer is detected and a biopsy is taken to take tissue from that biopsy and prepare it in such a way that you can integrate it into a microphysiological system. And the goal there of the microphysiological system is to replicate the microenvironment that the tumor is in the body so that you're not taking the tissue out and having it degrade or essentially stop functioning in some way so that the tumor microenvironment creates that in vivo microenvironment, and then you test particular drugs to see which is the most efficacious.

  (25:57):
  And by doing that, you can eliminate a very lengthy and dangerous process currently where the oncologists are testing different drugs on patients because they don't know which one is going to work best for their cancer. So it's essentially kind of the next level of a targeted therapy where you can take that particular patient's cancer itself and put it into these microphysiological system and test dozens of drugs in a manner of a few days to weeks at most, and then determine that in fact, this is the treatment that's going to work best for this patient and move quickly into that treatment state. That's really one of the most exciting examples of this technology that I can think of in terms of a clinical application.

  Mary Parker (26:47):
  That's where the use of high throughput makes perfect sense, especially with something like cancer. Time is of the essence.

  Jeffrey Borenstein (26:54):
  That's right. And I think there's so many possible combinations of treatments. In many cases there are combinations of immune checkpoint inhibitors and small molecule chemotherapeutic compounds and other treatments, cell therapies. If you can test these many combinations in the laboratory and then augment that perhaps as you mentioned AI earlier with even larger data sets, testing in terms of the machine learning type of environment or platform, then you can really quickly identify treatments that are efficacious for a particular patient's cancer.

  Mary Parker (27:33):
  Is there any research that you've been involved with recently that we haven't talked about already that you think is just really super cool?

  Jeffrey Borenstein (27:42):
  So I think that the research that we've been involved in recently takes several different directions, and I think we've talked a little bit about the cancer treatment. We've talked about some of the infectious disease research. And I think that when you think about the next generation, a technology that is very cool is the idea of putting these organ systems together in what we call a multi-organ system or MOS. And this is where we start to approach what, let's say an animal model can do for us to look at systemic effects. And so it's a vision that's been around for many years. I think one of the early examples that we demonstrated in collaboration with Northwestern University Medical School, we called it EBIT avatar, essentially kind of a play on the term avatar, was a replica of key aspects of the female reproductive tract. Because one of the hardest things to do in the field is to understand reproductive toxicities.

  (28:47):
  It's one of the most difficult applications to study for obvious reasons. So I think what happens in a lot of cases is that women who suffer from particular diseases are simply restricted from those treatments because of the potential risk of impact to their reproductive system, to a growing fetus. And so the idea was to put together these systems where you combine a series of organs in a circuit and then create something that starts to approach what we might call a body on a chip. So I think we've been involved in this for many years now, and I think more recently there's a great interest in many of the potentially emerging pathogens or other types of threats that are out there that one may need to create these multi-organ systems in order to understand the effects that may exist when a threat agent is sort of absorbed into the gut or the airway.

  (29:52):
  It may be breathed in through the respiratory tract and then interacts with various tissues and organs in the body. There may be some metabolism in the liver. There may be some processing by the kidney. And so if you only have a kidney on a chip or a liver on a chip or a lung on a chip, you're only going to see those interactions. But with these multi-organ systems, you can really start to replicate the complexity of the human physiology in understanding how these various diseases or other agents may interact with the body. And then the same thing goes of course for the treatments where the treatment may require or may function through interactions between various organs in the body. So I think really one of the coolest developments in the field is this multi-organ system technology.

  Mary Parker (30:46):
  If anybody who's listening to this podcast is interested in learning more about multi-organ ship systems, they can go back and listen to episode 76 where I discuss that topic with Dr. Casey Ronaldson Bouchard from Columbia University. So I'm glad to hear there's crossover in that. So many scientists think this is a pretty cool technology.

  Jeffrey Borenstein (31:05):
  Absolutely. It's one of the things that I think the field is most intensely interested in now, because I think what it does is there's tremendous opportunity in the multi-organ systems, but there are also tremendous challenges, as I'm sure you heard in that episode, the way that these organs are scaled with one another, and then how you flow a blood, like substitute a blood substitute through the organs so that it essentially is in contact with all the different types of cells in the system. So there's many challenges, but many opportunities there.

  Mary Parker (31:41):
  Well, thank you Jeffrey, for joining us and sharing your expertise with us. It's been really interesting to hear about these systems.

  Jeffrey Borenstein (31:48):
  It's been my pleasure to speak with you. Thank you so much for having me.

  Mary Parker (31:52):
  Jeffrey Borenstein is the Laboratory Fellow in the bioengineering division at Draper. Stay tuned for the next episode of Sounds of Science. Until then, you can subscribe to Sounds of Science on Apple Podcasts, Spotify, Stitcher, or wherever you get your podcasts. Thanks for listening.