[music playing]

Oby: From the campus of Harvard Medical School, this is Think Research. A podcast devoted to the stories behind clinical research. I'm Oby.

Brendan: And I'm Brendan. And we are your hosts. Think Research is brought to you by Harvard Catalyst, Harvard University's Clinical and Translational Science Center.

Oby: And by NCATS, the National Center for Advancing Translational Sciences.

Brendan: Stem cells are invaluable raw materials of the body that allow for specialized approaches to medicine for specific tissues or organs. If under the right conditions, stem cells can mimic a kidney or lung cell then researchers are able to mimic treatment for human biology outside of the body. Dr. Samirah Musah and her lab at Duke University have created an organ on a chip model that replicates kidney function outside of the body.

On today's episode, Dr. Musah joins us to talk about the applications for this technology. Dr. Musah, thank you very much for joining us. And welcome to Think Research.

Samirah Musah: Thank you so much for having me here. It's great to be with you. Thank you.

Brendan: So you're an assistant professor in the Department of Biomedical Engineering at Duke and you have your own lab, can you tell us a little bit about the focus of your lab?

Oby: Yeah. First of all, thank you very much again. So my lab really is at the interface of engineering and medicine. And as you mentioned, I hold a joint appointment with both in biomedical engineering and also in the Department of Medicine.

So a lot of our work is really focused on concepts from fundamental cell biology to stem cell biology, engineering microenvironment to control stem cell behavior. Basically telling stem cells what to become. We know that these cells have this remarkable capacity to self renew but also differentiate into almost any tissue in the body. But it's challenging to actually get them to become a desired cell type if you don't give them the proper signals. And sometimes the challenge also arises from not knowing what signals are required to get a cell to become a heart cell or a lung cell or a kidney cell.

So part of our interest in the lab is trying to uncover some of the mechanisms or signals or cues that the cells need in order to get them to become a desired cell type. And a lot of interest in my lab is really in deriving kidney cells. But also we do some work with neuronal cells. But generally, trying to find what environment we can provide these cells to become a desired cell type.

Now part of the reasons why we want to do this is to be able to develop models that would allow us to mimic human biology outside of the human body. So if you can imagine, if we can tell a stem cell to become a kidney cell in a dish, then we can use that to study how human tissues and organs develop. And we can potentially develop models that would allow us to better understand human diseases. And naturally, if we're able to develop disease models, then we can use them as a platform for discovery therapeutics in a way that would be directly relevant to humans. So these are some of the motivation and in the directions that we're also taking in the lab.

Brendan: Yeah, so you mentioned the kidney and that's one of the things I wanted to ask you about. And maybe you could just tell us a little bit about that work and why you wanted to look at the kidney. And I know that you've had grants from DARPA and other organizations to do some of this work. So tell us a little bit about why this-- why you focus on the kidney and maybe a little bit about this organ on a chip idea that some people might not be familiar with.

Samirah Musah: Yeah, absolutely. And that's a really great question. And I should mention that actually this work started as a challenge from DARPA. So that's really how I started it.

When I was a postdoc we were challenged by DARPA to develop models of about 10 different human organs, in vitro models that can be used to understand human biological responses and potentially connect the organs to study human pharmacokinetics, basically how the body responds to drugs and processes it right. So one of the organs we could not engineer-- at least we could not develop a functional model of the human kidney was because we didn't have access to human kidney cells. As you can imagine, to get human kidney cells will require a highly invasive procedure.

And it's also one of the organs that we know it does not regenerate. At least the functional unit-- one of the key functional units in the kidneys, which is the glomerulus, we don't have evidence that that functional unit can regenerate itself. So if a patient has damage to this part of the kidney, they are more likely to progress to have organ failure. And because there is just no way to repair that.

So that means that we didn't have a commercial source of these cells. And so without those cells we wouldn't be able to develop these in vitro models. So for me with a stem cell background and knowing that these cells have capacity to grow indefinitely and if given appropriate cues to differentiate into a desired cell type, I thought, well, can we develop a method to tell stem cells to become the population of kidney cells that we don't have access to. And of course it didn't take me too long to discover that there were no methods to do that.

So we had to come up with a method until-- but it was exciting and quite a motivator just knowing that there's so much about the kidneys that we don't yet know. Both from a developmental standpoint and disease mechanisms, I knew that if we could do this, that is, we could develop a method to derive the kidney cells that we wanted from stem cells, it could open up a whole new way for studying human kidney biology and developing models that would help us better understand disease and potentially even discover therapeutics. So I started off by trying-- basically started off by trying to find out if we could come up with a method to tell stem cells to differentiate into the cells that we needed.

And to make long story short, we successively accomplished that by deriving-- by coming up with a method that allowed us to develop one of the most important cell types in the kidney, which are kidney glomerulus protosites. In fact, these cells work with vascular endothelial cells to form that filtration barrier in your kidneys where your blood gets filtered. That's really at the interface where toxins get removed from the body. So you can imagine if that gets damaged a patient cannot remove toxins from their body and that could lead to a whole lot of other complications.

So for us, being able to derive the cells, was a huge step forward. But we also happened to be in a-- or I happened to be in an environment where we didn't do just the biology but we really have the capability to also engineer tissues or use concepts from micro engineering approaches and even computer design approaches at the bits at the time to develop these in vitro models that would allow us to mimic a lot of the biophysics of an organ like how fluid flows to an organ or tissue, what kind of mechanical simulation or factors affect the tissues function. So we decided to engineer a model that would allow us to basically repopulate this engineered system with the cells that we derived from stem cells and essentially create an in vitro system that would recapitulate some of the structure and functional characteristics of the kidneys function.

So that's really how we started and then we're able to build this model which we call the organ on a chip system. And are now using it for various applications including understanding how some of the mechanisms involved in tissue development. And also developing these platforms that we're now getting better understanding of how the kidney function and how disease progresses.

Brendan: Yeah, I wonder if you could talk a little bit about what does this look like? This in vitro model of organ on a chip. Like when you use it how do you-- I guess, you talked about fluids flowing through and seeing how blood flows through organs, what does this actually look like?

Samirah Musah: So you can think of-- so the material itself, you can think of gummy bears. So the texture of the material itself is gummy bears to feel like. I think gummy bears when I think of, well, if I could turn this into food what could we really make of it.

So it's a polymer. And so it has the structure that is quite flexible. And you can imagine if you had holes inside of a gummy bear but there are multiple holes that separates different channels. So how you create a channel but you can flush water through it and it will come out a specific path line that you create.

We designed it so that we know that in the body you have multiple cell types. And the cells come together to form tissues. And you have different types of tissues in each organ that work together.

So we created these fluidic paths in this chip that allowed the cells to interface with each other. So you can imagine having a dedicated path for blood to flow but there's another separate path where when you filter molecules or you remove toxins from your kidneys, you pee it out as urine, right? That's part of-- one of the ways we remove toxins from our body. So we have another channel where we can have that filtrate or the urine or pee go and then come out.

So doing that allow us to basically populate the separate channels with the specific cell types in the way-- in a way that resembles how they would be in the human body. And so we're able to connect these chips. Almost like you go to the hospital and you're hooked up to machines you can imagine these chips being hooked up to machines that allow them to have blood flowing or their urine being removed.

And so a setup like that allows us to study not just whether cells-- whether cells actually like the environment but whether we can flow blood, whether when you flow blood that has some kind of toxin in it, will that be removed so that urinary compartment that we created and under what conditions is that damage so that a patient might not be able to remove that toxin. And so that-- it gives us a lot of flexibility and really unprecedented ways to model some of these very complex biological responses completely outside of the human body but in a way that's directly relevant to the patient.

One other feature of this that I think is truly remarkable is just being able to couple this technology with iPS, induced pluripotent stem cell research. So you can-- as you may already know, we can take any patient cell, reprogram them into-- it could be blood cell or skin cells-- reprogram those cells into stem cells. And now that we have a method of telling stem cells to become kidney cells that means that we can generate the kidney cells from any patient without going directly into their kidneys.

We can take kidney cells from their urine. Or-- so in a very non-invasive way get cells from the patient, reprogram them to stem cells and we would have unlimited supply of these stem cells. So that we can differentiate them into the kidney cells or lung cells even if you want to see how their lungs work like.

So that means that we're able to create models that are specific to a patient. So we could have you on a chip, we can have Andrea on a chip, right? And it's a really-- it's a powerful way to do personalized medicine because we know that the readouts that we see would be specific to that individual.

And if you have a number of these chips, you can test a number of different patients or individuals and see whether it's common or different between how they respond to a drug or how a mechanism of disease unfold in different populations. So that drugs can potentially be discovered that works for everybody or perhaps works best in some population and another set of therapeutic works best in another population. So I like to think of it almost like a clinical trial in vitro or a clinical trial on a chip. When you're able to test so many different patients biology outside of that patient and potentially discover therapeutics that would still be directly relevant for that individual.

Brendan: Yeah, talk a little bit more about the-- so you talked about the tissue-- tissue regeneration issue and the kidney does not regenerate very well if you have damage to it. So how does this technology help with that? Could you grow a kidney and then transplant into a body to replace a damaged kidney?

Samirah Musah: So we hope to be able to do that. But I have to say that is one of-- it's really one of the most exciting projects in the lab because it's also somewhat counterintuitive, right? Because it's just we know that-- we know that we don't-- there's no evidence that the kidneys can regenerate. But the fact that we can model the kidney function in using a device that's really small, like a computer memory stick, that's how small it is, but you can imagine making that bigger depending on how much functionality you want to have or how many functional units you want to have. But I imagine that if we can model the kidneys function on these-- using these devices, then someday we could potentially use them to replace lost organs or tissues in patients. So that's one way I imagine this being useful for potentially organ transplantation purposes or organ replacement-- possibilities for organ replacement therapies.

Another strategy that were taken that is quite unique approach is to trying to say, well, we know this organ does not naturally regenerate or it's not very good at repairing itself when this particular functional unit is damaged, but we have an idea about what other organs can repair themselves. Like we know that the liver when cut out, if a portion of the liver cut off, it can potentially regrow. And we have some idea about other tissues in the body that have that kind of self regrow capacity.

So what we're doing in my lab is saying, well, even though we know these cells aren't very good at doing these things, what can we learn from the tissues that can do that to potentially engineer new pathways or modulate signaling pathways to allow ourselves to respond differently to injury. So instead of maybe dying when they are injured, we can activate pathways that would allow them to self renew, for example, and then we specialize. Since we know how to tell stem cells to become the cells that we want, we believe that if we can tell the cell not to die or respond in a very adverse way, we might be able to change its fate.

Brendan: Have you identified any of these pathways? Like you mentioned the liver and how that organ is able to regenerate, are there similar pathways between the kidney and the liver that you see like, OK, you know, when liver's damaged, this pathway is active, when kidney's damage, the same pathways not active.

Samirah Musah: Yeah, so because this unpublished I can't say too much.

Brendan: OK.

Samirah Musah: But I'm happy to share some really-- so we did find some-- there are some pathways we thought. So we thought, well, if these cells don't regenerate this pathway has to be down regulated. And it actually was the opposite of what we observed. So we then immediately thought, well, this must-- this pathway must work differently in kidney cells than in other cells like liver cells.

So we did find that, we know the signaling pathway is a network, right, so it's not a linear process. But we did find that some of the interacting partners of some of the transcription factors that are regulated seem to be differentially modified or regulated in kidney cells than other tissues. And so we believe that these differences in what classes of molecules or what classes of pathways are active or repressed around the same time could be a way to have almost like a set of signals that would work in concert to regulate cell fate decision.

So that has been really exciting. When we started off with a hypothesis and it seemed to be something else but it's leading us into unexpected directions that's quite exciting. So we're still working on that but hoping to have a story on that out soon.

Brendan: OK, perfect. And one of the other areas we are looking at is drug toxicity. And you mentioned that, obviously, the kidney filters blood and removes toxins from the body and this is really important in drug trials. Tell us about why that's important in drug trials and how you're using this technology.

Samirah Musah: Yeah, absolutely. So we know that during drug discovery pipeline-- drug discovery pipeline, we have about 75% of drugs or more actually fail during clinical trials. Mainly at large due to toxicity to the kidneys. So one of the reasons why this is usually not observed before drugs get into human clinical trials is because, of course, we know that animals are often used to do some of these studies and then there were no good in vitro models of human kidney.

So we envision that now that we have this platform that's able to mimic some of the key functional characteristics of the human kidneys, we could also use it as a platform for screening drugs and finding out how drug candidates work on human tissue and potentially use that process to eliminate drugs that would be toxic. Because we know that this is also one of the functional unit in the kidney that if a drug damages it, it means that the patient could end up on dialysis or need a organ transplant. So if we can actually rule out some of these potential toxins in vitro before it goes into humans, it really can be an important way to bridge those animal studies and the human clinical trials.

So that's really the key motivation for that part given that we know so many drugs are actually nephrotoxic. Even drugs that are-- a lot of over the counter drugs are quite nephrotoxic. And although sometimes that kind of effect is just as gradual but it builds up over time and patients end up with chronic kidney disease and things like that. So we hope that we can use this platform for screening some of these drugs.

Well, we can imagine it could also be useful for discovering new drugs, right? So druring a drug discovery and design phase, if you know that a drug-- you know immediately that a drug is toxic to kidney tissues, for example, it might help influence or guide the design strategy for optimizing the drug. A chemist might decide, well, maybe I want to change some functional groups on this drug to make sure that it's not so toxic to these cells or these tissues and organs in the body. So we're hoping that these technologies could be useful in all these different disciplines.

Brendan: And so you would-- so that would come into the drug discovery process pretty early, right? Like even before phase 1 possibly?

Samirah Musah: That's definitely my hope, absolutely. Is to be able to use these models before it gets into human patients so that we can identify what could potentially be toxic. We've had some-- now that we're trying to validate the model for some of these applications, some of the things we also doing is to look at drugs that we know failed during a clinical trial because of toxicity to the kidneys but we're not predicted by the animal models that were used and then using these platforms to show that it can indeed predict some of these toxicities. Or if the companies had used a technology like this that they could have actually identified some of these toxicities.

I think that's just-- I think that just perhaps reassuring for a technology like that but the true use of it, I think, the power really comes from being able to do the test before you even potentially cause a damage in a patient.

Brendan: You talked about being able to intervene or get that information about toxicity before it go-- before a drug goes into a human. How are you trying to get this platform-- how are you trying to foster adoption of the platform while you're validating it? I imagine it must be complicated to say-- go to a company like, I don't know, a big pharma company and say like, I'm working on it, we're pretty sure, or have you had those conversations with companies and what's the response been?

Samirah Musah: Yes, I think I probably can't talk too much on this. But where we found-- I think I would say, yes, we are talking to some companies and I think I'll probably just have to leave it there.

Brendan: Sure.

Samirah Musah: Yeah, but we definitely are looking at-- so for some of the companies that are interested in technology like this, they want-- they are interested in seeing whether a-- predict something they missed before, right? And so they can to really see that OK. This could-- because we know that when drugs fail during clinical trial it's not just about-- just about the company and their stocks and their wealth is affected but the patients that are involved, right, and the image also can impact the company severely. And it could, I mean, people-- companies sometimes shut down because of something like that and it almost--

So there there's a lot of-- there are a lot of other things that could-- adverse effects that could come out of something failing. And if you look at the amount of money that goes into these clinical trials, it's really just a fraction of it to test it on a platform outside of the human body that could potentially get information that you don't have. Because until you actually take that drug into patient-- into human patients you wouldn't fully know how they would respond especially when the kidneys are involved.

We know that a lot of lab animals, especially mice are used in these studies, but they don't develop kidney disease the way humans do. I mean, a lot of the drugs and chemicals that would destroy you and I's kidneys, one would be fine-- mice would be fine eating like several times more of that drug. So there are simply not a good model for knowing everything about a drug.

So it's important-- and we know already that they're both developmental differences in how human kidney develops and mice kidneys develop as well as function. So it's perhaps not to-- I think it's conceivable that you will see differences in terms of how drugs are processed. So my hope is that this could be-- this would be useful for really minimizing some of the challenges that occur when trying to extrapolate animal studies to human clinical trials.

A lot of our current collaborations also include clinicians coming to actually get-- a lot of clinicians coming directly to us, especially nephrologist because this is a field that there's still a lot of research going on, there-- the tools that they have available, it's really tough to find therapeutics that would work when it gets to humans. So we're getting a lot of collaborations with clinicians who either observe a certain pattern in patients in terms of how they're responding to therapies or how some families-- you might have family members that are responding differently to some of their therapeutics and they want to be able to model them-- model that in vitro to better figure out how a patient is going to respond or why they're responding differently to a drug so that they can customize the therapy for them. So we definitely do have collaborations focused in that area as well.

Brendan: The clinicians must be really excited about this.

Samirah Musah: Absolutely. Yeah.

Brendan: Yeah. Great. Well, Dr. Musah, thank you very much for your time and we really enjoyed having this conversation with you.

Brendan: Thank you so much. It's a pleasure. I really appreciate you having me on here. Thank you.

Brendan: Thank you for listening. If you've enjoyed this podcast, please rate us on iTunes and help us spread the word about the amazing research taking place across the Harvard community.

Oby: To learn more about the guests on this episode, visit our website, catalyst.harvard.edu/thinkresearch.

[music playing]