2015 Fast Congress Banner
2015 Fast Congress Banner

Understanding Drug Toxicity: 3D Tissue Constructs, Organ Chips and the Hermeneutic Circle of Biology

Back to Podcast Download Page 

Ann Nguyen:

Greetings and welcome to this podcast from Cambridge Healthtech Institute for the Second Annual FAST Congress on Functional Analysis & Screening Technologies, which runs November 17th through the 19th in Boston, Massachusetts. I'm Ann Nguyen, Associate Conference Producer. With us today is our closing keynote speaker from the Engineering Functional 3D Models and Organotypic Culture Models for Toxicology conferences, Dr. John Wikswo, Founding Director of the Vanderbilt Institute for Integrative Biosystems Research and Education and Gordon A. Cain University Professor at Vanderbilt University. John, very glad to have you here.

John Wikswo:

It's a pleasure to chat with you today.

Ann Nguyen:

Your research background includes superconducting magnetometry plus cardiac, neural and gastric electric and magnetic field modeling and measurement. Your current work includes organs-on-chips, microfabricated devices and software for studying how living cells respond to drugs and toxins. How did you get from there to here?

John Wikswo:

That's an interesting story. I became interested in the physics of biological systems as I was entering graduate school at Stanford in 1970, and so the interface between biology and physics has been a central theme for my entire career and that has led to a great appreciation of the complexity and elegance of physiology. I got into superconductivity by starting out as a low-temperature physics technician as an undergraduate doing plumbing, carpentry and wiring for a low-temperature physics lab. When I went to Stanford, I went to measure the magnetic field of the brain, but the instruments we needed didn't exist yet. So I instead worked on the magnetic field of the heart and measured both cardiac volume changes and electrical activity. I came to Vanderbilt to measure magnetic fields of isolated nerves, which we did, and the fact that we had the superconducting magnetometers led us into gastro-magnetism and we currently have our clinical facility for studying the magnetic fields of gastric electrical activity.

The same instruments allowed me to do a tremendous amount of work on non-destructive testing, and just before I became a cell biologist, we were doing non-destructive testing of aging aircraft. So the circle that I have traveled is even broader than you might have imagined. But a dozen a years ago, a friend of mine and I sat down and decided where should we be in a dozen years, and came to the conclusion that biological physics needed microfluidics in order to control systems. I started a major effort at Vanderbilt to instrument and control cells with microfluidic devices and have set up a number of very powerful collaborations, primarily with analytical chemists, pathologists and cell biologists. So we were studying in depth the activation of immune cells by various chemical and environmental means.

A number of government agencies came out with requests for proposals for organs-on-a-chip to study drug discovery. The first one was from the Defense Advanced Research Projects Agency, DARPA, and there have been others to see whether microfluidic technologies and 3-dimensional tissue culture could be applied to studying the problems of drug interactions and organ metabolism of drugs. The goal of these agencies was to devise a means to understand the mechanism of action, toxicity and efficacy of drugs long before the 10-year typical period when Phase III clinical trials start failing after investment of a tremendous amount of money. I looked at these calls and realized that in fact the work we had been doing on cellular instrumentation and control were in fact ideally suited for instruments and controlling 3-dimensional tissue constructs in the middle of an organ bioreactor.

The fundamental problem that I see in much of our understanding of biology, and in particular our understanding of drug effects from in vitro studies, is that the growth of a monolayer of cells on either plastic, glass or a layer of protein on plastic or glass is that these monoculture monolayer cell constructs look nothing like human tissue and the cells behave in many ways quite differently than they would in tissue. In a monolayer cell culture, there's a thousand times more media in terms of volume than the cells, and hence any signals that the cells secrete are diluted by a factor of a thousand and cell-cell communication, called paracrine signaling, and communication of the cell with itself, termed autocrine signaling, those signals can be lost just by dilution. In contrast, tissue is approximately 80% to 90% intracellular space and 10% to 20% extracellular space so that when a cell secretes something the neighboring cell really does see it. It's not diluted.

The challenge that I think is facing organs-on-a-chip, the most significant challenge, is to figure out how to first of all grow cells in 3-dimensional tissue constructs, and there are a whole lot of people working on that for quite a few years making very good progress. I think much of the early work in this field has been done by people trying to understand the cancer microenvironment, where in fact you have proteases secreted by cancer cells that affect the extracellular matrix. So 3D tissue culture is a growing field. The difficulty is that if you try to make constructs that are larger than one or two hundred microns in diameter, the cells in the center will die. Just as the interior of a tumor becomes necrotic.

The challenge becomes how to keep these constructs alive and you don't do it with passive diffusion. You do it with perfusion, either through artificial capillary beds, or there are a couple of groups working on perfused networks, vascular networks. Once you have built 3D tissue constructs that represent tissue of specific organs, whether it's liver or heart, kidney, brain, you then have to figure out how to keep them alive and how they can talk to each other. That requires in my mind very small, very low-volume pumps, valves and interconnects.

Our group has been working specifically to take the technologies we've developed over the past decade and apply them to the connection of different organs to build what is effectively a homunculus. A homunculus is the classic representation of a model of the human typically drawn across the skull to show how the sensory regions are distributed. But homunculus is a more general concept. It's basically a miniature representation of a human. And so what we are trying to do is develop microfabricated 3D tissue constructs of different organs that are connected together to form functional units with volumes that are small enough that when the liver metabolizes a drug and produces a cardiotoxic metabolite, the concentration and time course of that concentration in the model, recapitulates the action in the human.

We're not trying to build a complete human-on-a-chip. That would be a folly because if we built a complete microhuman or a complete millihuman, a microhuman or a millihuman would be too complicated to understand. So what we're trying to do is build simplified models. A physicist would call them a toy model. Simple enough to allow you to build a good intuition but too simple to be exact, and the various organ-on-a-chip programs are trying to combine multiple organs in different programs, and different projects have different organs and different collections of organs, but the idea is that you take multiple organs.

For example brain, liver, kidney and gut. With the gut and the liver being the main routes by which nutrients enter into the body and are metabolized. The kidney on how they are secreted or not secreted and the brain as a very important target organ. The goal is can you make a system that is simple enough that you can put a drug into it, have the liver metabolize the drug into something that's either neurotoxic and cardiotoxic, and detect it before that drug goes into human clinical trials.

Ann Nguyen:

How has research on tissue- and organ-scale toxicology evolved over the last decade or so with regard to in vitro organotypic culture models?

John Wikswo:

Much of what has been learned about toxicology over the past decades has been gained through studies on animals and also humans. The goal of using animals was that you could do experiments on animals that you would not conceive of doing on people, particularly until a drug had been developed or the mechanism of action of a toxin had been identified. The difficulty is that the biochemistry and physiology of a rat or a mouse or a rabbit or a guinea pig or a dog or a sheep or a pig does not match that of a human. And so drugs that are toxic to humans may not be toxic to the animal of choice and vice versa.

There was a goal then to see whether you could cut costs and increase the throughput of an analysis by going to cell culture, and you then saw a great deal of effort of trying to grow cells in tissue culture to study the toxicology of the drugs on the cells. While a great deal has been learned by doing this, particularly about intracellular mechanisms of action and signaling and metabolism, I believe that many people in the community are disappointed by the difficulty in extrapolating from planar cell culture studies all the way to humans. You can get some interesting possibly self-consistent results that may not extrapolate to humans.

The next step, then, we say okay is the problem that we're using either animal cells, canine kidney cells for example, or immortalized cells, HeLa cells, or more modern, very specific cell lines. The immortalized cells are effectively immortal because either they've had their telomeres adjusted or they have been otherwise affected, typically by cancer, to become immortal. It's not clear that the biochemistry and physiology of these immortalized cell lines looks anything like normal human biochemistry and physiology.

The evolution of toxicology is basically going from initially people, because that's where you saw the toxins taking effect, to animals, to planar tissue constructs. There's a growing interest in organoids, which are essentially 3D constructs mammospheres from reconstructed breast epithelial cells or comparable spheroids from liver or gut. But the evolution is that everybody is trying to get more and more realistic cellular microenvironments. I view in vitro organotypic culture as the next very important step and I think it has the potential of revolutionizing our understanding of emergent behavior in biology. Interactions between two organs that you didn't anticipate is an emergent phenomenon.

The trick is the engineering challenges have to be solved and I'm convinced that they will be solved. There are a large number of people working on it and once you recognize the challenges you can immediately move to solve them. I've taken a rather active role in pointing out the problems as I see it and I'm working as fast as I can with my group to solve as many of these problems as we can.

In a short answer to your question, tissue- and organ-scale toxicology is evolving extremely nicely over the past several years. The national and international efforts on organs-on-a-chip are moving in exactly the correct direction. There's a lot of work to be done and the problems are not easy to solve, but I think the net result is we'll come up with a much better understanding of biology and in particular how to control it or at least prevent damage to biology.

Ann Nguyen:

Sounds very promising. And finally, can you give us a brief preview of your keynote presentation -- including a little bit about the hermeneutic circle -- slated for November 18th.

John Wikswo:

Thank you for asking about the hermeneutic circle. It's something that I find quite fascinating. In the study of language -- and this is most evident when you study a foreign language -- is that you have difficulty understanding the language until you understand the meaning of the words, but you can't understand the meaning of the words unless you understand the context in which they are spoken. You can imagine the statement, "the man was scattered." Another: "the men were scattered" and "his ashes were scattered." Those three uses of the word "scatter" are quite different and context-dependent. So hermeneutics is the study of meaning, and a generalization of the field is that you cannot understand the whole until you understand the parts, and you cannot understand the parts until you understand the whole.

So for The past decade when I've been teaching classes on systems biology, I viewed it as the end game in a problem of reductionist biology. Actually, it's the post-reductionist end game of biology and this is what I'm going to talk about in November. We started out studying people and animals thousands of years ago. The Greeks began to understand physiology and the challenge that I see is how can you understand the animal or the human until you understand all of the molecules, and how can you understand all of the molecules until you understand the human. So there's a hermeneutic circle of biology.

I used to think it was simply reductionist. You start from the animal, work your way down. We've reached the genome. We have knowledge of everything and hence we should be able to work our way promptly back, and initially I thought that systems biology was going to be the approach back. In fact what I thought was computational systems biology would be the way back. You have the theory of everything. You have all the information on the genome and its regulation and you simply start reconstructing in mathematical models the entire human.

The difficulty that I ran into when I started thinking about this problem was that you start counting equations required to do what I just described and you quickly come up with estimates that suggest that you might have Avogadro's number of coupled ordinary differential equations or even worse, coupled partial differential equations that keep track of spatiotemporal effects in cell-cell communication. It's beyond comprehension to solve Avogadro's number of equations, so I gave it a name. Albert Einstein got the unit the Einstein is a mole of photons and so I've decided that Leibniz, the inventor of differential calculus, needs a unit. And a Leibniz is a mole of differential equations.

My frustration was that having watched the final years of reductionist biology where we gained tremendous insights into the genome and its regulation and you say, well, how do we go back? I've come to the conclusion that you go back with the parallel of the way you got there. We started with animals. The field started doing isolated organ preparations. Intense effort in the 60's on isolated organ preparations, some of which are still used today in drug development and physiology.

In the 60's and 70's, cellular biology took off with the development of cultured cells. Proteomics has allowed us to understand proteins. We've gotten to the genome and its regulation so you work your way back. You can now today engineer proteins that have either alterations at specific locations or, for example, a green fluorescent protein tag at a location or reporters to report binding. Today people are able to use the combination of mathematical modeling and structural biology to create essentially instrumented proteins. You can similarly create instrumented cells through genetic engineering. Optogenetics is a stunning field where you can both sense and control cells.

I've decided that the hermeneutic circle of biology says that we're going to go from the knowledge of the parts back to the knowledge of the whole through a hybrid approach of mathematical modeling and experimental devices and I think we're today at the point of building carefully instrumented 3-dimensional tissue constructs. People are figuring out how to put reporter constructs inside of cells to report what the cells are doing, building 3D tissue constructs. The next step is to put the constructs together both mathematically with models and more importantly with homunculus, the in vitro model of a subset of human organs.

What I see the grand challenge in front of us is to close the hermeneutic circle of biology, where we start with knowledge of the animal and the human, go to a thorough understanding of the genome and its regulation, that's the parts, and then work our way back through again the combination of mathematical modeling, challenging engineering and instrumented 3D constructs that talk to each other, and we close the circle with a highly instrumented homunculus that lets us understand what drugs and toxins are doing to both the homunculus and us.

Ann Nguyen:

Got it. Well, John, these are some very thought-provoking and very broad insights cutting across a lot of different areas of study. Thank you so much for sharing your time and your insights about them today.

John Wikswo:

It's been a pleasure chatting with you. I look forward to seeing you in November.

Ann Nguyen:

That was Dr. John Wikswo of Vanderbilt University. He'll be giving his closing keynote presentation at the Engineering Functional 3D Models and Organotypic Culture Models for Toxicology conferences at the FAST Congress, which runs November 17th through the 19th in Boston. If you'd like to hear him in person, visit www.fastcongress.com for registration information and enter the keycode “Podcast”. I'm Ann Nguyen. As always, thanks for listening.