At A Glance
Name: Christopher Hughes
Position: Associate professor, molecular biology and biochemistry, University of California, Irvine
Background: Assistant professor, UC-Irvine — 1996-2001; Associate research scientist, Yale University School of Medicine — 1991-1995; PhD, Neuropathology, University of London — 1988
Angiogenesis remains a hot research area for potential cancer therapeutics, but methods of studying this process are still limited. Chris Hughes and colleagues at the University of California, Irvine, have developed a cell-based angiogenesis assay that may eventually prove a useful tool in high-throughput drug screening. Last week, Hughes discussed the assay and its commercial potential with Inside Bioassays.
Tell me a little bit about your background. How did you become interested in studying angiogenesis?
I did my graduate work in London at the Institute of Psychiatry in the neuropathology department, and there I was working on the blood-brain barrier. The endothelial cells that line blood vessels in the brain are novel in that they express very good tight junctions. They form an effective barrier between the blood and the brain, so there’s very little in the way of free diffusion — it’s a protective barrier for the brain. So I got interested in vascular biology through that. Then, a guy named David Male, who was an immunologist, joined the department, and together we started to look at the immunology of the blood-brain barrier, using mostly in vitro techniques. When I graduated from London University, I went to Harvard Medical School to work with Jordan Pober, who was also doing vascular biology and immunology. So I did a post doc at Harvard and one at Yale, and I got an assistant professor position here at UCI in 1996. And I’d been interested in angiogenesis for some time, and particularly, I was fascinated by the idea of cells making themselves into tubular structures. They’re seemingly simple things, but the more you think about it, the more interesting it becomes. How does the cell initiate tube formation? How does it know when it’s formed a tube? What are the mechanisms for forming tubes? When you think about it, the body is full of tubes, but we actually know very little about the initiation and the mechanisms underlying tube formation. When I came here, I decided that would be an area of research that I wanted to follow. So we set up some in vitro angiogenesis assays — not true angiogenesis, but the assays model some aspects of angiogenesis. We did a gene screen using representational difference analysis. Nobody uses it anymore, but it’s a PCR-based method of looking at differential expression of genes. We got some interesting genes out of that, including one that we continue to work on, which is called HESR1. There were about six labs that published on the same gene within a year, including us. At the time we were the only ones with any functional data on it — most people got it through homology searches. We identified it as being an important regulator of at least some stages of angiogenesis. It’s subsequently been shown to be expressed throughout the developing vasculature, and it’s a member of a family of three genes, and if you knock out two of them, you get nice vascular defects and heart defects. So it’s a transcriptional factor that regulates blood vessel formation.
This family of genes plays prominently in the assay you’ve developed?
The model that we originally set up was fairly simple and it just really modeled rearrangement of cells into cords, but there was no really good lumen formation. So we then spent quite a bit of time perfecting this in vitro assay whereby we get very good sprouting of endothelial cells in culture and form very nice tubes — they branch just like you would see in an animal. That’s the model we’re working on now, and that’s the one we’re considering for drug screening. The transcription factors are downstream of the Notch pathway, which is a well-characterized signaling pathway during development — it controls cell fate decisions, among other things. Downstream of that is the gene we work on — HESR1 — so it’s a target of Notch signaling. That, in a nutshell, is what we’re doing — we’re looking at in vitro angiogenesis assays, and in particular, we’re interested in the Notch signaling pathway and understanding what the precise role of that is. Several of the Notch genes have been knocked out in mice, and the mice die early from vascular defects. But the problem with a knockout mouse is that it’s an endpoint assay — you get a dead mouse, basically. And then you have to look and see that the blood vessels don’t look so good, but you have no idea about what they looked like before, and we don’t really understand why they stopped at this point. So what we’re trying to do with an in vitro assay, among other things, is to watch the growth of the blood vessels in the absence of Notch signaling. We can look at gene expression, we can manipulate gene expression, and we want to get at the mechanism of why Notch is needed for blood vessel development. Until now, we know that it is, but we don’t really know why.
What kind of assay techniques do you use in these assays?
Let me back up for a second and say that many companies that screen for angiogenesis molecules look at endothelial cell proliferation, and so they’re looking at drugs that affect this. Why are they doing that? Is it because they think that endothelial cell proliferation is the critical step, or is it because proliferation assays are easy to scale up? And to a large extent, they’re doing those assays because they can. But, is targeting endothelial cell proliferation the best, or the only way of blocking angiogenesis? We don’t believe it is. It’s quite possible that by blocking endothelial cell proliferation you’ll also block proliferation of other cells, but that could be a lack of specificity. So we were interested in asking: What are the unique components of angiogenesis? What would be a better, more specific target? And one of the things that may not be unique in the strict sense of the word is tube formation — branching and anastomosis. If you target proliferation, there’s a chance you target all proliferating cells; whereas if you target branching, you’re really getting down to some specific functions of endothelial cells. Those kinds of things are harder to assay for. You can’t just do it blindly in terms of throwing in tritiated thymidine and seeing how much is incorporated. You really have to look at the cultures and decide what the morphology is. What we do is fairly labor-intensive right now, and we’d like to find ways to alleviate that. But we look at the cultures, and we’ve gotten pretty good at it — we can look and say: This doesn’t look right, or this looks fine. And then one can look closer and say: Are there tubes forming, and are these vessels actually forming lumens, or are they just remaining as cords of cells? We can count the number of tips, so if you imagine the way this assay works: These endothelial cells are coated on beads, which are about 100 microns in diameter, and these beads are embedded in a gel. And what happens over time is that cells sprout from the beads, and you get sort of a starburst pattern. Each sprout has a leading tip, and so one can count the number of tips, and that will give you not only idea of the number of sprouts, but also the degree of branching. For example, if you count the number of sprouts at the base, and then you count the number of tips, then you have zero branching. We can also measure the length of the sprout. Many of these things can be automated, and we’re looking at newer software that may be able to handle these kinds of things, at least semi-automatically. We can also look at the diameter of the sprouts, which can be important. We can do BrdU staining, and then use fluorescence to look at the number of proliferating cells. That is something that could very clearly be automated, if you’re just looking at fluorescent dots. With that, you’re getting back to just endothelial proliferation, but here it’s in a very specific location — a developing vessel, and not just a monolayer of endothelials growing in a dish, which, I’m not sure what that model is, but it sure isn’t angiogenesis. Our assay gets as close to a developing blood vessel as we can while still being in vitro.
Is this bead technique commercially available?
No, it’s not. There were some papers from a German group several years ago that went quite a long way towards getting this working, but they never got good tube formation, and it only worked with certain types of cells. We’ve made a lot of modifications to that such that we now get reliable tube formation. That’s a big issue in angiogenesis assays — whether you’re truly getting lumens, and we absolutely do. That makes our assay unique. People are doing similar things, but ours is by far the best in vitro angiogenesis assay around right now.
Right now you use confocal microscopy for this?
Exactly. We use confocal and we use two-photon, which is nice, because you can go much deeper. We can stain these cultures in a kind of whole-mount stain — we can stain a whole gel and visualize the whole network, the whole 3D structure. And we can stain for extracellular matrix proteins, stain for membrane and cell junction proteins — if you’ve got a good antibody, we can stain this.
Do you have an interest in developing this assay commercially?
We do, actually. I’ve been talking to some people in the last few weeks about ways that we might be able to commercialize this, and the way I see this is that we could set up some kind of company that could do angiogenesis screening using this model. Advantages of this are the potential for high-throughput, and we can monitor over time, which you can’t do very easily in vivo without sacking the animal, and again, you’re looking at endpoint. We can do time-lapse if necessary. We can follow the development of the vessel. We can get back way more information about a particular drug’s effect on angiogenesis. So you could imagine a company would screen all their drugs against endothelial cells growing in a 2D dish, looking at proliferation, and maybe they get a few. But they may have drugs there that block branching, tube formation, and anastomosis, and they’re never going to know it. But, if they put them into our assay, they will see that. Basically, it opens up the field much wider for identifying drugs that hit important aspects of angiogenesis that most people just aren’t assaying for. There is a “tube-formation” assay out there, which is the Matrigel assay. This is an extracellular matrix preparation that contains a lot of different proteins. It’s not particularly well-analyzed and quantified, but what happens is if you put endothelial and several other cell types on this is that they will change from a monolayer and will form networks of cords. People talk about this as being an angiogenesis assay, and I strongly disagree with that. You very rarely get proper lumen formation in this assay, and you occasionally get these slit-like intracellular lumens. So they may be trying to form lumens, but they don’t really go through the process. So you get this kind of network that looks vaguely like a capillary network, but I’m not convinced by it. Angiogenesis requires protein synthesis and RNA synthesis. There are reports out there that you can get these networks forming even when you block RNA and protein synthesis, so you have to ask: What exactly is this modeling? It’s really rearrangement and migration of endothelial cells, but it’s not really angiogenesis.
Have you been in talks yet with anyone?
We haven’t. I really started thinking about this when I was scheduled to [give a talk] on it, and then when I got your call I thought about it a bit more. And even yesterday, one of the students that’s been working on this gave his advancement, and one of the people in his advancement committee asked if we’d ever thought about commercializing this as a drug assay, and he said: Funny you should mention that! But that’s as far as it’s gone, but I do want to get into talks with people. I’m hoping this might be a conduit if there are people out there that would like to talk to us about this, and we would be very interested.