TEDxUCLA 2011: Minding, Mining, Mending, Mapping

Vaults: a therapeutic delivery system


About Leonard 

Leonard H. Rome, Ph.D., is Distinguished Professor of Biological Chemistry and Associate Director of the California NanoSystems Institute. His laboratory research at UCLA centers on a novel cellular organelle called a “vault” which he and a former postdoc, Dr. Nancy Kedersha, discovered in 1986.


Thank you. I’m going to talk about turning a common household item into a therapeutic delivery system.

We start with cancer drugs. Most cancer drugs are poisons. They work by a very simple process. The idea is to kill rapidly dividing cancer cells faster than you kill normal cells.

But these drugs aren’t very smart. If Homer here has a tumor in his lung, he has to take a chemotherapeutic drug that circulates through his entire body before it can find the tumor. That causes a lot of side effects: anemia, nausea, hair loss. Homer can’t afford any more hair loss.

What Homer needs is a smart drug bus, a delivery system with a GPS device in it so that it can deliver the drugs right to the tumor. Delivering the drugs to the tumor means you can get, use higher amount of drug and have lower amounts of side effects.

But we don’t have such a smart delivery vehicle. We don’t have a drug bus with a GPS system in it. But if we wanted to develop such a system, what would it look like?

And the authors of this movie, Fantastic Voyage, back in the 1960s, thought about this because they had to design a drug delivery vehicle to try to save the life of a diplomat. And they settled on shrinking a submarine down to microscopic size, this submarine called the Proteus. And as a young boy, I was really fascinated by this movie. It was probably because of Raquel Welch’s delivery system, though.

So the Proteus didn’t work very well. As soon as it got injected into the bloodstream, it got recognized as foreign by the immune system, so it was a problem. So what kind of a drug delivery system do we want to design?

Well, first of all, it has to be the right size. It has to be less than 100 nanometers so that it can maneuver its way through the bloodstream and through the organs that are designed to filter out particles. The drug delivery system has, even though it’s small, has to have a very large payload, you want to be able to put a lot of drugs inside of it. It has to have a protective environment, protect the contents that are inside, has to be safe in humans, non-immunogenic. You’d like it to have a GPS system, some way you could target it. And finally, you’d like to be able to control the release of materials that are inside the drug delivery system.

So people have been studying all kinds of different objects — chemicals, polymers, natural things — to try to use as drug delivery systems, and probably the system that’s gotten the most study is virus particles. They have almost all the right attributes to design as drug delivery systems, but they have a big problem, and that is that the human immune system has been evolving for millions of years just to try to eliminate viruses. So they have problems.

So what you need is a material that’s naturally occurring that isn’t going to be recognized by the immune system, but has properties that make it a good drug delivery system. And back in 1986, Nancy Kedersha and I stumbled upon this particle by accident, and it seems to have some of the right properties for a drug delivery system. When we finally figured out that we had found something new, we had to name it. And so we had a little “Name That Particle” contest in the lab, and the name that got the most votes was Rome-o-some. And I mean, I might have a big ego, but it’s not that big.

So we decided that we would name the particle based on its morphology. And because the outside surface of the particle is made up of multiple arches, it reminded us of the arches, of the vaulted ceilings of cathedrals. So we called the particle a vault.

And it’s turned out that vaults are a pretty good name for the particle because they mean, it can also mean container. But a lot of my colleagues haven’t been very happy with vaults. They think maybe Rome-o-somes or raspberries or hot dogs. I don’t know, I don’t think they look like hot dogs. Well, maybe a little bit. Maybe a little bit.

So if you look at the size of a vault particle versus the size of a cell — this is a human white cell blown up two and a half million times on this screen. You can see that a vault would be this little yellow dot here. And we went, we made an antibody against the particle when we first found it and then went looking for it in human cells. And this is a human skin cell stained with an antibody that’s fluorescent-labeled against the vault particle. And we found out that there’s thousands to tens of thousands of vault particles in every cell of the human body. That means that every person has, oh, a hundred or a thousand trillion vaults in them. So this is why I call it a common household item.

So with Phoebe Stewart, we were able to look at the structure of the vault with a higher resolution using a method called cryo-EM microscopy. And there we saw that the vault was an oblong capsule. And with cryo you can look at the inside of the particle as well as the outside of the particle. And the particle appeared to be hollow with a very thin shell surrounding a very large internal volume.

But cryo is a little bit misleading. It only shows you the structure of materials there in the same exact place in every particle. If you look at a density slice through the particle, you can see that there’s actually material inside the vault. And after years of study, we figured out all the different components of the particle.

So it turns out the shell around the outside is a single protein in multiple copies and we call that protein the major vault protein, or MVP for short. And inside the particle there are also three other components: one, a small RNA, which we call vault RNA; an RNA binding protein called TEP1; and an enzyme called VPARP, all inside the particle.

The cryo-EM also lets you look at, tell you the size of the particle precisely. And the particle’s about 70 nanometers by 42 nanometers. So what’s a nanometer? We’ve been talking about nanometers today. So a penny’s about 20 millimeters across. A human hair is a tenth of a millimeter, which is 100,000 nanometers in diameter. So with relation to a vault, you’d have to line up 2,400 vaults side by side in order to extend across the diameter of a human hair. Hair grows about six inches a year, which turns out to be 10 nanometers a second. So since I started my talk, your hair has grown 3,000 nanometers.

So if we’re gonna use vaults as a recombinant, as a, as a delivery vehicle, we’re gonna learn to make make the vault in the lab. And the way we did that, Andy Steven in my lab decided that he would express the major vault protein in a cell that doesn’t have vaults. And so he used an insect virus which infects insect cells and he put into the insect virus a copy of the DNA that encodes the major vault protein. And we infected that virus into insect cells in culture, the insect cells became little protein production factories, and they made hundreds of thousands of copies of the major vault protein.

And to our surprise, all of the information that was necessary to assemble the vault particle was contained within the primary sequence of the major vault protein, so vaults self-assembled in these cells from the major vault protein. And we could purify those vaults, and structurally they looked identical to vaults that we could isolate from natural human tissue.

So the only difference, though, was that the recombinant vaults that we made in the laboratory were empty. They were truly containers with a big empty volume in the inside surrounded by this protein, the major vault protein. So the major vault protein’s quite interesting, and by using this expression system in the baculovirus system, which is what we used in insect cells, we were able to make vaults and alter the structure of the major vault protein in order to try to figure out how was it arranged in the vault particle.

Proteins have a beginning called the N terminus, and they have an end called the C terminus. And proteins are just a string of amino acids, like beads on a string. If we magnify the chain, you could see the amino acids making up the protein. So using molecular biology techniques, we added extra amino acids to the N terminus of the major vault protein, we produced vaults in the insect cells, and we purified those vaults and went looking for those extra amino acids. And the only place we could find extra amino acids was at the waist of the particle on the inside, where you see this pink density.

When we put extra amino acids at the other end of the vault particle, the C terminus, and then we produced vaults, we found that there was an extra blob of density at the top and bottom of the particle. So that type of study allowed us to propose a model for how the major vault protein might be oriented within the particle, and that model we published in 2004 and it’s shown here. And there, we speculated that the major vault protein started at the waist of the particle where its N terminus was facing the inside and it traveled up to the cap of the particle where its C terminus was facing the end of the vault. And then about 40 copies of the major vault protein came together to form the half-vault, and then two half-vaults came together to form the full vault, and that’s why all the N termini were in the middle, and that’s why the C termini were at the top and the bottom.

And working for three years, Dan Anderson and David Eisenberg’s laboratory here at UCLA got a crystal that defracted at eight angstroms, and he was able to determine a low-resolution crystal structure of the vault particle, which is shown here, the half-vault. And what you see is every other chain of the MVP being colored red, blue, red, blue, and it’s meandering from its N terminus at the waist to its C terminus at the cap.

A couple years later, a group that we’d never even known was working on vaults by Hideki Tanaka published a 3.5 angstrom resolution structure of the vault particle, which is shown here. They worked on a vault crystal for five years before they got this structure. So it’s an incredible structure made up of about 80 copies of a single protein that self-assemble and enclose a huge volume, and they’re naturally occurring so they might be terrific for using as a drug delivery vehicle.

So if we’re going to make them into a delivery vehicle, deliver therapeutics, we have to learn how to package things inside. And Val Kickhoefer in my lab showed by studying one of the natural vault proteins that’s inside the vault, this protein called VPARP, she gave us a clue as to how we could package any protein inside the vault.

What Valerie found was that there was a sequence of amino acids on the VPARP protein that acted like a ZIP code that directed this protein into the inside of the vault. And she figured that she could just attach this sequence of amino acids to any protein, and maybe any protein would be packaged into the vault.

So she picked a protein called luciferase, which is an enzyme that gives the firefly its glow. We use it in the lab a lot because you can follow it around, because it glows. So using molecular biology techniques, she attached the ZIP code to the luciferase protein, she expressed this protein in the insect cell system with the major vault protein, and the vaults that got made look normal. The inside had the luciferase protein packaged in the inside in two rings of density, pseudo-colored here in red. So we had succeeded in making little nano light bulbs, if you will.

So how are we going to target vaults to cells, such as cancer cells? How are we going to give vaults their GPS system? And turns out that antibodies are a great specific way to target things because the antibodies have a very high specificity for the target that they’re directed against.

So using the fact that we can express peptides on the outside of the vault by putting them at the C terminus of the major vault protein, we made a vault that had an antibody binding site, an antibody binding site at the top and bottom of the vault. And we mixed that vault with antibodies that would have been, that were made against cancer cells so that those antibodies would be immobilized very tightly on the outside of the vault, and with our hope that the antibody combining sites being still available, would allow those vaults to target the cancer cells.

And when we add these vaults to cancer cells, we have to be able to follow them. So we put a green fluorescent protein inside the vault, so the vaults would fluoresce green, and we could look at them under a fluorescent microscope. And so when we added these vaults to cancer cells in culture, we see that the green fluorescent vaults bind to the cancer cells and actually show us where the cells are in a culture when we look by fluorescence. If we leave the antibody off the vaults, we don’t see any binding of the vaults to the cancer cells.

So we have a lot of different applications in our laboratory now that are under study for using vaults as therapeutic delivery vehicles. We’re just like an ordinary academic lab, we’re spread all over everywhere, we’re doing everything. But I only have time to talk to you about two different applications today that are probably furthest along in our search to turn vaults into therapeutic delivery vehicles. And that is an example about cancer therapeutics, and one about vaccines.

So with Steve Dubinett and Shervan Sharma here at UCLA, we’ve been trying to engineer vaults as cancer therapeutics, and Steve and Shervan have been working on a small molecule called a chemokine. This chemokine, which is called CCL21, attracts immune cells. And often tumors put out signals that allow them to hide from the immune system. And with Steve and Shervan have found, that if they inject this chemokine into a tumor, it wakes up the immune system and starts to attract immune cells to the tumor.

This is a really interesting therapy. But the problem with CCL21 is it’s a very small molecule and it rapidly dissipates out of the tumor. So with Steve and Shervan, we decided we would engineer a vault particle with CCL21 packaged inside, bound to the inside in such a way that it would still come out of the vault very slowly. And we, again, we use that targeting ZIP code that I told you about before, made vaults with CCL21 on the inside.

And before we could see whether or not they would be active in cancer, we had to show that the CCL21 was still active. And you do that by what’s called a chemoattraction assay. You put the chemoattractant on one side of a membrane, immune cells on the other side, and you watch the attraction of the cells across the membrane.

And when we did that with the CCL21 vault, to our surprise, we found out that the CCL21 vault was about 100 times better at chemoattraction than the chemo atttractant and CCL21 alone. And we think that’s because these cells sense a chemical gradient and that the very slow release of the CCL21 out of the vault sets up a very steep gradient of the chemical.

If we inject empty vaults into a mouse tumor, the tumor grows very quickly. If, on the other hand, we inject the CCL21 vault into that tumor, the tumor is slowed dramatically, its growth. And if we look at the tumor itself, we see that there is a massive infiltration of immune cells, marked as these small green cells in this slide, into the tumor.

So we really think that the CCL21 vault is working to attract immune cells, wake up the immune system to the tumor. And we’re hoping that this will be the first application for vaults that we can move into a human clinical trial. And the next stage is to make a vault that’s acceptable to the FDA so we can move this to a clinical trial.

Next example I want to talk about is vaccine delivery. Worked with Kathy Kelly at UCLA, has asked the question, “Could we use vaults as vaccines?” Kathy works on chlamydia. Chlamydia is the most common sexually transmitted disease in the United States, three million cases a year, and there’s no good antibody therapy for chlamydia.

So with Kathy, we engineered a vault particle with a major outer membrane protein of chlamydia engineered in the inside, packaged using the same strategy I told you about. And we immunized the mice with a nasal spray of these vaults. This protein MOMP, the MOMP vault, the protein itself is not, does not give the mouse immunity against chlamydia. If we look at an empty vault and we do a chlamydia challenge in the mouse, the mouse is infected with chlamydia. But if we immunize the mouse with a nasal spray of the MOMP vault, the mouse is completely protected from a chlamydia infection. So the good news is, if you’re a female mouse, you’ll be protected from from chlamydia, okay? Okay.

So again, we’re hoping that we can turn this therapy into a human therapy. We’re working on a human vault with chlamydia antigens in it that we could take to a clinical trial. We’re not quite ready to use vaults therapeutically, but I hope that you’ll see that this particle has great potential as a therapeutic delivery vehicle.

And one day, maybe we’ll have turned a common household item into a therapeutic delivery vehicle. Thank you very much.