Futuretech Podcast Interview Transcript
Interviewer: Richard Jacobs
James J. Hickman, PhD
Chief Scientist
Interview Transcript
JACOBS
Hello. This is Richard Jacobs from the Futuretech Health Podcasts and I have James J. Hickman. He’s the founder of a company called Hesperos. Hesperos website is Hesperos INC.com. Tell me about the company and how the idea for this company came together as well.
HICKMAN
Hesperos started in 2000 as a holding company for my IP, but I started collaborating with Michael Shuler about 8/9 years ago. We basically took my technology I had been developing which are functional in vitro systems and serum free medium; added his multi organ metabolic systems and came with something that really was better than the two parts separately. So we started the company as a real company in the beginning of 2015 and so we’ve been around now since then. We’re focusing on a human on a chip technology that comes from what Michael Shuler had created the term body on a chip is about 20 years old. They came up with the first patents on taking different cell types and linking them with fluidic flow but you then combine that with technologies I’ve been developing which is a function system. So do you know what a functional system means? If you go into a doctor’s office he doesn’t immediately take your blood and look at your biomarkers okay what he does is he says well you know how you doing? Look at your eyes, are you walking, are you limping? Ect. Those are functional readouts. He’s looking at how your nerves are controlling your muscles, looking at how your brain is functioning in terms of cognition. And then he listens to your heart, he looks and says you know it is your heart beating, he’s looking at your heart rate and then he might non-invasively take your blood pressure these are all also functional readouts. It's what we try and do. We recreate those function like muscle contraction, nerves innervating muscle in causing that contraction, cardiac electrical activity, cardiac force neuronal communication neurons basically sending electrical impulses to each other, and we combine that in a multi organ system where the fluid is recirculating and so that the organs are all talking to each other. So when we put liver in there then we can actually look at not only a drug but the general metabolizes a drug into different compounds. Some of which can be toxic in some cases. They’re deliberately the drug they want to see whether or not you know that when their liver metabolizes that’s the more active form of the drug and so we can then look at what they call the parent compound and the metabolites in the same system. That’s pretty much what a human on a chip system does. We can add in barrier tissues like GI tract, your intestines, to control as compounds or how that gets into the system. We can also add another barrier tissue called the blood-brain barrier which is the interface between your blood supply and your brain. Your brain really doesn’t like all the components in the blood. It likes to carefully filter those out. People don’t realize that there’s in your bloodstream it's about 8% protein and in your cerebral spinal fluid which bathes your neurons it’s actually .001% because your brain likes to really control the interactions because there’s so much communication not only with electrode impulses but also with molecules going on. So we can build a blood-brain barrier into the system as well. We’re working with kidneys and other types of organs, as well.
JACOBS
I don’t know how you’re supposed to make a model of a multi-organ system on a chip. I mean each organ in itself is super complicated and the fluid flowing between them you know it has all kinds of things with nutrition going in it. How far have you been able to model and is it correlating to real-world phenomena? How good is the system you’ve made?
HICKMAN
We’ve actually been able to show in a four organ system the proper response to drugs we’ve been able to and that’s been published. We just got a paper accepted in Science Translational Medicine where we looked at for the first time the idea of looking at efficacy which does a drug work and toxicity in the same system. So they call this off target toxicity. With a stunning collaboration with Roche Pharmaceuticals, we asked well can you take a non multidrug-resistant cancer put it in the same research rating fluid that you have cardiac tissue, because most your chemotherapeutics have some sort of cardiac toxicity, and the presence of a liver, because also chemotherapeutics are toxic to liver, but they have a lot of chemotherapeutics that are converted into toxic species by the liver. So we then took in that four compartment system two different types of cancer: a cardiac force measurement, a cardiac electrical measurement and the liver. So it's basically a five chambered system recirculating and looking at a drug called oxofin. Which is a standard chemotherapeutic and showed that the tamofixin itself wasn’t very effective in reducing perforation of either cancer but it’s metabolite hydrocytamofixin was for the non multidrug-resistant cancer. One of the reasons that cancer cells become multidrug-resistant is there’s these pumps inside of them that pump out the chemotherapeutic as fast as it comes in but you can add a second drug in that blocks that pump in this case we use something called rappamill. So which is a drug drug combination and we were then shown that we could stop the proliferation and even the multidrug-resistant cancer. At the same time we were monitoring the cardiac function and the liver function, and we saw the electrical activity and the force both were affected and they were decreased. In the presence of the tamoxifen and hydroxytomocasin but recovered. So they recovered after a couple of days. Which is again if you're trying to then understand how a chemotherapeutic might affect an individual patient. That ‘s really good information for a physician to have but when we added in the second drug to block the multidrug-resistant cancer pushing the drug out of the cells and it decreased proliferation but it also increased toxicity. So the drug drug combination increased the toxicity and it didn’t recover as well. So again if a physician was looking at this data, looking at your cancer he’d be able to say okay I got to try a different PGP or inhibitor of this pump then the verapamil. So this way he could actually do this testing which right now they do on people. So now we could actually take your cancer and look and see how different drugs, different combinations of drugs and be able to test us to be able to inform a physician. It wouldn’t be a diagnostic because that involves the FDA and everything else like that but a physician tries to take as much information as he can from as many sources as they can and this now enables that to happen in a system and something that could be done within a couple weeks.
JACOBS
What is the recirculating fluid made of? If it's proprietary, just say so. How are the organs modeled?
HICKMAN
I developed the first serum-free medium about 25 years ago. Most people doing this research are still stuck using calf serum. The problem with calf serum is its calf serum. It's also from a baby. When they actually drain the fluid out of the calf is either dead or terrified. And so there‘s all these hormones or things like that in there that aren’t really good for cells. But there’s also a lot of things that basically say you need to be a calf. By using serum a lot of cells will D differentiate and this is a problem that most labs have. By using a serum free medium, which is not proprietary, we published it and variation of it, we’ve been able to figure out how to keep four and five cell types alive for up to four weeks in this circulating serum-free medium and that paper was just published in Advanced Functional Materials. Where we took a prop for 28 days. 28 days being a magical period because if you’re going to measure toxicity in an animal for systemic talks generally they will do that experiment for 28 days. We use a different pumping mechanism as well. Instead of using a peristaltic pump or something that’s actually a compressive type pump, we just simply use gravity and we believe this is actually better for the cells. It causes less stress on the cells and the other key thing is we’re not trying to reproduce the anatomy. People try and say well 3D is obviously better than 2D, not always. If I’m trying to make an organ that I’m going to implant in somebody’s body obviously you need to have it be 3D. If I’m trying to create a system to test the function, all we got to do is recreate the function. So what we do is try to get the cells in a healthy mature state and integrate them with these silicon based devices called micro electromechanical systems, or bio MEMS as it’s been called. This allows us to create hybrid 3D systems which allows us to have the function of the organs without having to preserve the anatomy or reproduce the anatomy.
JACOBS
The cells themselves, how do you get them? Let’s say the liver. Do you biopsy someone’s liver and culture those cells?
HICKMAN
People donate livers and they know they’re going to be used in research. We can basically buy primary liver cells and because of our serum free formulation we can get them to not be differentiated in a fairly simple culture system. Whereas most people have to go to great lengths to be able to try and keep them differentiated in their systems but with the serum free mediums we don’t have to do that. With the cardiac cells we create from induced pluripotent stem cells. The neurons are from induced pluripotent stem cells. The muscles are from stomach cells which are expanded from a biopsy. We can also now get to the point and be able to do muscle from induced pluripotent stem cells. Some of these differentiate ourselves. Some can just be bought. There's a company called CDI. You can get cells from Sigma, from Lonza, there’s lots of different places that sell human cells.
JACOBS
I’ve heard of people making organoids to model let’s say cardiac function or other functions. How would you compare what you’re making to organoids? What is an organoid compared to your modeling?
HICKMAN
People will try and reproduce the anatomy and that’s what an organoid will do for you. It will give you that anatomy of a subsystem from the body. That’s really good for things like developmental tox if you’re trying to look and see what a compound does is starting to develop. The problem with organoids is really hard to figure out the function from them. Because it is really difficult to assay what’s going on inside an organoid. Generally, what you have to do you actually have to take it, you know fix it, slice it open and then do an immune estate site of chemistry to look and see what is happening with the cells because they haven’t yet figured out how to integrate. A way of measuring function in the systems is there's crude way so kind of look at some beading and everything but the problem with an organoid is that you get an oxygen gradient in them because they haven't figured out how to vascularized them yet and so you get different levels of cells dying within gradients within the organoids. And that changes their function so it's really good for looking at anatomy, but it's not really good for looking at function.
JACOBS
mmm it is amazing that you're able to model the function of these organs and in such simple ways. Kind of funny, just everyone had the assumption that they had to make organoids and make them more and more sophisticated and maybe even 3D printed organs and model things properly but you're able to do that in a much simpler way.
HICKMAN
Well I've been working on it for 25 years. So that's a benefit of Hesperos that you know Mike Shuler basically invented the field 25 years ago. I've been working on these systems for 25 years. Mike's got over 100 publications or 70, 75 to 100 publications in this space. I've got over 50 publications in the space. You know we have large research groups that are also feeding information into the company. We’ve basically licensed the whole suite of patents from UCF where I'm a professor and also from Cornell where Mike's a professor into the company. So that's a real benefit that Hesperos has over lots of companies that are kind of coming up. Where they really don't know the history of these things where when somebody says yeah 3D is better than 2D without knowing the history they just sort of assume oh that's true. But it's not true. It is true if you're trying to create in vivo tissue engineering but it's not necessary for in-vitro tissue engineering for many different systems. It is for some and we do go to 3D when we need to. But in most cases it's not necessary if you're just trying to reproduce function.
JACOBS
So if you got into the point where you've tested drugs in your system. You've gotten certain results when you've modified where you've tested that drug than in people. Where you've modified its use and tested people. Have you seen the correlation?
HICKMAN
We have not advanced far enough yet. Although we're getting there to be able to have a drug that was tested in our systems get approved for going into a clinical trial. We're heading in that direction because to be able to really convince the FDA to go into a human they like to have a lot of computer modeling on what the different possibilities are. Generally that is called a PKPD model so you know how long the drug is sticking around and actually how much of it is going to each organ. So you get some idea of the residence time and everything for a compound. So what we can do is we can create these PKPD models or pharma kinetic pharma dynamic models of our in vitro systems and we're using those and building those to help predict what we think we might happen in vivo. Okay and that's how we're going to transition our data. To be able to better inform what's going on in a clinical trial and hopefully convince the FDA that somebody's drug will be safe to go into a human without doing an animal model. And this becomes really really important if you think about rare diseases. Where in many cases the compounds they don't want to have the risk of going into somebody if they don't have an animal model and so it's really hard to get you know drugs approved for some rare diseases because it's just no good way to assay it beforehand and the last thing I want to do is put a drug into somebody who has a rare disease and have to be a catastrophic problem. You think about all the gene therapies you know where they started off just trying to go into humans and then people died as a result of it. They didn't have a really good model system for that. So we can actually be a really good model system because we can take the cells from the rare diseases and build what they call a phenotypic model to be able to reproduce aspects of the disease. Basically show that we can reverse it or ameliorate the effects of that generally hereditary or some other type of rare disease. At the same time monitoring the tox. This work could be a huge benefit for looking at rare diseases, as well as looking at more common diseases and conditions for the general population.
JACOBS
Why not start with a drug that's in use already. So you can get clinical data and look at its impact on other organ systems later and find a second use for it or to see if it has toxicity that maybe wasn't observed when the first came out and then it can be modified to have less side effects. Why not look there?
HICKMAN
Absolutely that's something that we're definitely working with people on right now. Which is a drug to say has been used for something else or has safety data and may be able to show that for another indication that we have efficacy. Exactly one of the things that we're trying to do right now. Hesperos is leading in the area of trying to do that with drug companies.
JACOBS
Are you seeing any overall themes? I mean if you look at enough drugs perhaps you'll see an overall theme in your modeling that would maybe steer people in a faster direction towards getting the right drug for people?
HICKMAN
Well we're not. We don't have enough replicates at this point in enough disease models to be able to look at that. We are finding evidence that some drugs that were toxic the reason they are toxic is because they over accumulate in certain organs and that those organs actually do metablize on them and it's not the liver. Which is why it was really hard to see. We are seeing some of that and we're working with drug companies on that. We are doing what was suggested earlier which is looking at drugs that have either been failed in clinical trials, have actually been approved and then pulled, and testing against our system. We're really pretty good at being able to predict results in why they failed. We're also working with certain pharmaceutical companies to examine some of their failed drugs and to see whether or not we can predict the results that they saw you know not until late in the process and we're having pretty good success rate there, as well.
JACOBS
I would think this should be mandatory for a lot of clinical trials at certain stages so that before you do the trial and get a bad outcome. You can at least model them and that you know change your initial starting condition so it goes better.
HICKMAN
We believe that's where the future is going to be. I mean you know right now animal models are predictive. When you go from an animal model into a drug discovery drug approval process about only 11% of the drugs survive clinical trials due to tox or not being efficacious. Animal models are just really poor predictors of what's going to happen in the human. Now people will say you know some things have worked that's absolutely correct because you spent trillions of dollars something's got to work. In some cases animal models have been pretty predictive and it worked well but overall the percentage is very low. I mean we've cured over 200 diseases in mice that have not translated to humans. This is not anecdotal. This has actually been quantified and looked at in the industry. There's been publications. There is a very nice publication by first author Cooke from Astra Zeneca from 2004 I believe. Where they looked at all their drugs and to why they failed and the reasons the animal models don’t work. Again going back to this whole idea you know in rare diseases you don't have animal model for most of them and you're not going to because it's too expensive.
JACOBS
So do you have interest from drug companies that they want to use your modeling as a part of the clinical trial process so that they go in with the best foot forward?
HICKMAN
We're getting there. Most of them are very skeptical still to some degree. You know some of the smaller drug companies are a little bit more accepting. We're now you know we've been around for about three three and a half years. We’ve had some success. We got this one paper coming out. I think will be a game-changer in Science Translational Medicine. We have another paper with another pharmaceutical company where we've been able to use our PKPD models of our in vitro system to predict animal data that they had got. So we're getting to that point now where we think they're starting to accept this as you know something they should start thinking about incorporating into their approval process. Again what we'd like to do is be able to start convincing the regulatory authorities that our systems are as good as if not better than animals and actually start reducing substantially the amount of animal data that they have to use for clinical trials if not eventually one day trying to eliminate it.
JACOBS
How fast can you model a given drug?
HICKMAN
It really depends upon the drug and the system and kind of what they want. You know if you want something like a neuromuscular junction. Something like you want to test a drug against ALS and you want to have a compound simple test that does it restore function in a mutant model of the neuromuscular junction. We could probably do that test in a couple months. We basically think that we are also shortening the time dramatically in terms of drug discovery as well. Because our system really can work directly with a medicinal chemist. People don't realize one of the biggest problems in drug discovery is they screen hundreds of thousands of compounds and come up with a couple of candidates for a disease or their target or somehow a phenotypic model they've created. They then have to take variations of this using their computer models predicting its solubility, its efficacy and things like that but then the medicinal chemist has to choose, because they're going to go from milligram quantities to kilograms before they can go into an animal trial. They have to pick one because it's really expensive to scale that up. With our systems they could test four or five six different variations of their primary compound because we only require milligrams; we're not going into an animal model. So this is a way of actually getting the best candidate not just guessing at what hopefully might be the best candidate. I think that's really where some of the new work that we're doing is pointing to and again we really shorten in the process we believe.
JACOBS
Why not challenge you know a couple of drug companies that are about to go through a trial or that have gone through a trial and say hey I'm gonna figure out some proprietary data without you telling me if you let me test your drug on our model and I'm gonna shock you with what I can tell you it's gonna correlate with what you've seen.
HICKMAN
If somebody had the money to do that we'd be happy to do it. We're a small startup company and we don't have investors. You know Mike Shuler and I set this company up basically using our own funds. So we don't have a huge big bankroll out there. We're using NIH grants, we're using the customers we have to slowly grow the company. We're trying to keep control of the destiny of the company and not have profit be the only motive for forming the company. We also think that we can do a lot of social good with the company. Again rare disease spaces, areas that drug companies aren't particularly interested in that still need attention.
JACOBS
How much would it cost for you to assay a certain drug you said it may take several months at least but what does the ballpark cost look like?
HICKMAN
Really depends, but certainly less than million dollars.
JACOBS
Even for a startup that’s a lot. That makes sense.
JACOBS
oh yeah yeah if somebody was gonna come up with the money I mean we'd be happy to do it but for us to come up with a couple hundred thousand dollars or something else like that we just can't do that we just we don't have the funds.
JACOBS
So what targets do you see are the most likely for you? What areas of the whole drug discovery efficacy testing model do you find multiple ones or seeing which one test?
HICKMAN
Well so far we've been pretty successful in being predictive for whatever model that people have asked us to build. Because that's we have some standard assay that we do but most of what everything we do is somebody comes to us and says we want you to build this. Yeah you published this part over here but we want you to take out the muscle and put in the kidney and because our system is very easily reconfigurable we can do that. So most everything we build is custom. They work pretty well in predicting what the compound was supposed to do and what they were expecting. So we just want to keep building upon that and it's publicly especially you know trying to publish that research and again just get it out there. That you know these things are not just us talking that they work. They've actually been validated peer reviewed and published in very high-impact journals so they can actually start thinking scientifically that these things actually work.
JACOBS
Ya know that's fantastic it's amazing. I'm just amazed you're able to model this stuff in such a simple way, but to such a simple way. It seems like it would need incredible sophistication to do this but yet it's working.
HICKMAN
Well one of the things is we don't sell them. We don't sell the systems. We actually do it as a service based because it is very uncomplicated. We want to have control over that because it is a complicated system that we're quite good at, but it would take amazing amounts of training to get other people up to speed you know if we just transferred our systems to them. So our’s is a service based company. Unlike some of the other companies out there like tissues or emulate which are trying to sell their systems. Either they have to keep their systems really simple or they're if they're finding difficulty getting people to seamlessly add them into their research flow. Because they're complex these are pretty complex systems. So far there's only five companies out there that can do multi organ systems at this point.
JACOBS
Well very good so with them what do you see is happening in the next three to five years? What do you hope will be the result of what you are working on?
HICKMAN
I expect in three to five years that this will become a mainstream technology in the drug development process. I think that we will be able to get drugs using only human-on-a-chip data into clinical trials. I expect in some cases in rare diseases we might be able to go directly from our data into patients if there's clinical safety data. You know these are the kind of things that we see and are counting on to push the technology and really fundamentally change how the drug discovery process occurs.
JACOBS
What is one of your ideas like the most sophisticated system that you'd ever want to make? How many organs would it include and what would it look like?
HICKMAN
The most sophisticated system would be something where we could do systemic tox. So right now you have to go into an animal to do systemic tox if you just put the drug in there after it's gone through all the other testing and say okay what is it we're going you know are we going to see something unknown that we didn't expect in the animals. Then they go for 28 days. They sacrifice the animals and they look and see if there's anything unexpected there. The day we can actually build a multi-organ system that is accepted as a substitute for systemic tox you won't need animals anymore.
JACOBS
Hopefully that will happen.
HICKMAN
We hope so - we're really pushing really hard for that. So exactly how many organs we need for that, not sure. Some people that published ten organs already. Michael Shuler you know in his academic lab has published a 13 chamber with different cell types in it. So we know it's possible to put many many types of organs in these systems. It's just whether or not it's necessary because the problem is as you put in more and more organs the cost goes up and the complexity goes up. So we try to keep it as simple as possible to answer questions. If we can predict systemic tox that will be the most complex system that you would need in the drug discovery process.
JACOBS
You're going to try to do that or you need to stay in here so confidence right now that would take you out?
HICKMAN
We actually got pretty far down the road in my academic lab. We were collaborating with L'Oreal for a number of years until their head of exploratory research retired. We got pretty far down the road of creating a systemic tox model. Because you know place like L'Oreal, you know the EU has banned use of animals in cosmetics. They still have problems you know they still have to be able to predict systemic tox. It's not like the drug discovery you know industry where they can still use animals it's not a problem, cosmetics industry can't use them so we got pretty far down the road that was the one paper in Advanced Functional Materials showing that we can get four organs interconnected for 28 days and be able to show functional activity for the entire time. That was one of the results from that project with L'Oreal.
JACOBS
What's the best way for folks to find out more and you get in contact?
HICKMAN
Visit our website and then send us an email or give us a call and we're quite happy to talk to people. What we do webinars all the time with companies who are interested in working with us. We kind of show them the data. We then talk with them about what they want to try and be able to do and we're very honest with people if we can do what we tell them we can do. it if we're not really sure we'll say okay this is what the research we have to do or you know just somebody else a competitor who can do a better then we send them off to the competitor.
JACOBS
You have a any legacy webinars that you could link to you know maybe in the show notes we can look for you?
HICKMAN
There is some information on the website. We are putting together a webinar for it to put on the website but it's a brand new website. Just created a couple months ago. We're still adding to it but we should have something like a webinar for the basic systems. But most of the technologies described on the website.
JACOBS
Okay well very good I appreciate you coming on the call and it's amazing stuff you're working on so.
HICKMAN
I appreciate talking to you as well and quite happy to help and I again I appreciate the opportunity to to talk to you about our research.