 |
|
Eric Shusta
Assistant professor, Department of Chemical and Biological Engineering University of Wisconsin Madison |
At A Glance
Name: Eric Shusta
Position: Assistant professor, Department of Chemical and Biological Engineering, University of Wisconsin Madison, since 2001.
Background: Postdoc, William Pardridge's lab, Department of Medicine, University of California Los Angeles, 1999-2001.
PhD in chemical engineering, University of Illinois Urbana, 1999.
The blood-brain barrier prevents the uptake of drugs from the bloodstream into the brain, and has long been a vexing problem for researchers as well as physicians treating a wide array of diseases. Eric Shusta was recently awarded a four-year, $1.2 million grant from the National Institutes of Health for his work on blood-brain barrier membrane proteomics. ProteoMonitor spoke with Shusta to find out about the work.
How did you get into studying the blood-brain barrier?
I started this all off as somebody who worked with antibody production and antibody engineering. That was my PhD work. And it turned out that I got interested in trying to use antibodies to deliver drugs. That took me to the brain, which was my postdoc work with Bill Pardridge at the University of California, Los Angeles. That's where I got into this idea of trying to use antibodies to profile the membrane proteome of the blood-brain barrier.
Could you describe a bit about the blood-brain barrier and how it presents challenges to delivering drugs to the brain?
The blood-brain barrier, otherwise known as your brain blood vessels, are very impermeable. This is due to the fact that the endothelial cells that form the vessels are unique when compared to blood vessels elsewhere in the body. How are they unique? Well, it turns out they have epithelial-like type junctions, which are essentially proteins linking adjacent endothelial cells. They also lack fenestrae, which are pores found elsewhere in the body in the blood vessels. Not to mention efflux transport systems that take anything that diffuses through the membrane and pump it back out into the bloodstream. So what you have is a physical interface but also an active interface in terms of preventing the uptake of drugs into the brain. This could be small-molecule drugs, proteins, genes, you name it. That is the problem in terms of trying to deliver drugs into the brain.
Is that a problem that has been overcome over the years?
No. In fact, that's the big reason why we're working on this. The majority of medicines designed even the traditional small-molecule pharmaceuticals, say between 500 and 1,000 molecular weight are excluded from the brain. That's one of the big reasons we haven't seen that many medicines for neurological disorders.
How do the existing medicines for neurological disorders take effect?
Some of them happen to be membrane permeable, so they can go through the membranes of the cell and actually get into the brain without being targeted by efflux transport systems. And others, like L-dopa for Parkinson's disease, make use of endogenous transport systems. In the case of L-dopa, it's the large, neutral amino acid transport system. So instead of phenylalanine, you're sending in L-dopa, which differs by a couple of hydroxyl groups. So it's a molecule that looks like a natural substrate.
How many years have you been studying the blood-brain barrier, and what techniques do you use?
Seven years. Basically what I started to develop at UCLA and what the NIH grant is based on is the use of a combination of blood-brain barrier cDNA libraries and blood-brain barrier-specific antiserum to do this cloning of membrane proteins.
So it's a little different in terms of proteomics in that the idea is that we're trying to identify blood-brain barrier-specific proteins, particularly membrane proteins, using antibody-based techniques. And in a sort of off-shoot of that, we're also trying to identify the antibodies that target those proteins. The idea is that the antibodies that target blood-brain barrier-specific membrane proteins may have the advantage of targeting the brain only, which is a big problem with what's out there today, and also may allow non-invasive transport into the brain. The way it would happen is that these antibodies would target molecular transport systems that are embedded in the membranes at the blood-brain barriers.
So the technique uses a blood-brain barrier cDNA library expressed in a mammalian surrogate, such as a cell line. And then we have a blood-brain barrier membrane antiserum. It's a polyclonal that recognizes whatever the rabbit saw as immunogenic. And what we do is we take the antiserum and subtract it against other organs. So we can now make that antiserum brain specific. And then what we do is we clone the proteins that the membrane protein antiserum recognizes.
How far along are you in this project?
We've published the basic technique where we pulled out several proteins that were selectively expressed in the membranes of the blood-brain barrier. Now we're in the state where we're doing this all by flow cytometry, so we can clone many more. We're sitting on an enriched pool right now, but we haven't gotten to the point of sequencing yet.
The other technique that we're using is with antibody libraries. In the polyclonal antiserum, we never know what antibody is doing the binding. So we can't really use it as a delivery antibody. What we're trying to do now is expand this technology by using antibody libraries these are the libraries that you would have in vitro, maybe on phase for phage display. We use yeast, so it's on the surface of the yeast particle.
What's really neat about an antibody library is you can trace the antibody to its encoding DNA. So if you can imagine doing the same sort of subtractive profiling, but now I know what antibody is doing the binding what antibody is binding to what receptor. So in that way you've identified both a blood-brain barrier-specific membrane receptor, like a transporter, and you have the antibody that binds to it. It's a simultaneous identification of the membrane protein and the targeting antibody. So then that allows us to maybe do some drug delivery with those novel antibodies.
I think both techniques are somewhat compatible. As you know, if you raise an antiserum in an animal, as with the first strategy, there are some proteins that are likely not immunogenic, so you wouldn't get complete coverage. And in the second set, there's a different set of principles governing whether or not your antibody binds to a membrane protein. So they're really somewhat compatible, and they're both fairly high throughput. They're not completely comprehensive like, say, mass spec, but really our niche is to identify specificity and have an antibody that binds to that particular protein. That's the niche that this technique falls into quite well.
So the ultimate goal is to have these antibodies that target the blood-brain barrier membrane proteins?
Yes, particularly transporters.
Are there some blood-brain barrier membrane proteins that have been identified that are being used to deliver drugs?
I can give you examples of what's out there that we're trying to get better than. I unfortunately can't tell you what's most exciting in my lab now, because it's not published.
The two that have been most widely tested are antibodies against the transferrin receptor and the insulin receptor, both of which are present at the blood-brain barrier. This has allowed the delivery of small molecules, proteins, liposomes, you name it. It does so because these two receptors employ a receptor-mediated transcytosis mechanism.
The reason that we're still looking for other things is that they're not very efficient, and they lack targeting, which means the drug goes everywhere in the body. So that's where the subtractive techniques come in to try to look for organ specificity, in this case the brain. And that's where the antibody library stuff comes in, because we want to know what those antibodies are so we can use them for delivery. So we're essentially looking for the next generation replacements for the anti-transferrin and anti-insulin receptor systems. Not to mention we're also trying to find out what proteins are selectively expressed on the blood-brain barrier endothelium from more of a scientific perspective, and less of an application perspective.
So you could put your drug cargo, which could be small molecules, you can connect it directly to proteins, or you can deliver particles like liposomes that are loaded with drugs. You could decorate the liposomes with antibodies.
What diseases in particular do you think this would benefit?
Stroke would be a target. Some of the proof-of-concept work done with this general approach was done with stroke models, using neurotrophins as a therapeutic.
Another example would be the recent clinical trials in injecting another type of neurotrophin for Parkinson's disease. This was done in human patients, but they had to essentially perform a craniotomy and inject the drug directly into the target tissue.
There are many others I'm probably not aware of as well, in terms of what pharmaceutical companies are working on but probably can't deliver [into the body].
What we're trying to do is something no more invasive than an intravenous injection. Where we're at right now is we've identified some novel antibodies and receptors that show some promise.
Have you filed for patents, and do you plan on commercializing the antibodies?
Yes, we've filed for patents, and if they keep passing all the tests in terms of does it actually get into brain endothelial cells, does it get across, does it work in vivo then maybe we'll think about [commercialization]. We're in those sorts of stages now where we're doing more directed testing towards whether or not these can be used as drug transporters.
Do you have any plans to partner with a pharmaceutical company?
This would all be licensed through the University of Wisconsin Madison. Were it to prove robust enough, we might target a [pharma] company to license the technology.
Do you do any of the more traditional mass-spec based proteomic work at all?
No. We are looking at ways to integrate that into our system, but as of yet, that really isn't ready.
