Wellcome Trust Sanger Institute, Atlas Group
Name: Michelle Teng
Position: Postdoc, Wellcome Trust Sanger Institute, Atlas Group (PI: John McCafferty)
Background: PhD, immunology, Gonville and Caius College, University of Cambridge, 1999-2002; undergraduate degree, biochemistry, Imperial College of London and National University of Singapore, 1994-1999
Scientists at the Wellcome Trust Sanger Institute in the UK and Erasmus Medical Center in The Netherlands have published research demonstrating how C. elegans expressing mammalian G-protein coupled receptors can be used to screen for small-molecule modulators of GPCRs and potential drug candidates.
The work, which appears in the July 20 issue of the open-access journal BMC Biology, takes advantage of a simple phenotypic readout: the physical avoidance behavior of GPCR-expressing nematodes to specific ligands. Michelle Teng, lead author on the paper, took a few moments this week to discuss with CBA News the implications of this research for drug screening.
Have researchers expressed mammalian GPCRs in other small animals or organisms in the past?
Not us, although there has been work in the past by others. We were not interested in mammalian models because we wanted something that was quick. John McCafferty’s group runs the Atlas [of Protein Expression] Project, which is an open resource that generates proteins and antibodies using phage display, with the readout in the form of flow cytometry or immunohistochemistry. They establish global protein expression patterns using tissue arrays.
There was one particular group of proteins that was difficult to tackle, and that we wanted to express, and that was GPCRs – that’s where I came in. The whole approach was to look at a creative, novel way of expressing GPCRs that would enable us to look at GPCR-ligand interactions without involving cell culture. We wanted to use the cell culture method; in fact, we were exploring baculoviral systems that had been previously reported, but we didn’t have a lot of success with it.
There was a published paper that has been out for a few years now that reported olfactory receptors being expressed in the olfactory neurons of C. elegans [Milani et al, Neuroreport, 2002 Dec 20; 13(18): 2515-20]. It was actually a very small report, and they didn’t do a lot of experiments with it, but they did demonstrate that cloning a rat olfactory receptor confers these animals a new type of behavior – it gives them an attractive response to volatile ligands. If you clone it into avoidance neurons, it generates the avoidance behavior. That gave us the idea that you could probably express other GPCRs of medical importance.
We wanted to explore the possibility of expressing it in gustatory neurons that are exposed to the environment, so that would give you access to soluble ligands that the worms can detect. The olfactory neurons are actually embedded in a sheath underneath the nose, so it would not be accessible to ligands. That’s why we went for the set of taste neurons, so to speak. These neurons are quite unique – they are polymodels, meaning that they can respond to a lot of other types of stimuli, like touch. One neuron, for example, the ASH, detects avoidance to SDS, for example, or other detergents. That was one of the neurons we looked at in the study, and the other one was ADL, which is not so well studied. But because the promoter that we used was to drive expression in one or more neurons, we thought it would generate a higher chance of success of getting a response from the animals.
The first time we did it, it was completely blind – just a matter of picking the right reporter and promoter to clone it upstream from our GPCR. We were really happy when it worked, and that’s how we developed the protocol. The original protocol was actually developed by a scientist who looked at mutants that respond to a particular repellent. That wasn’t a high-throughput method – it was on a large agar plate with a barrier of test compound drawn right across it, so you actually need quite a lot of the compound. We decided to adapt this to a higher throughput form, in a four-well rectangular plate, where each well is actually a long rectangle. You put the worms in one end of the rectangle, put a barrier of compound down, and put a volatile attractant on the other end, to give it some sort of motivation to cross the line. Within a half hour you almost always get your answer – beyond that, there is no significant difference in response. So this is amenable to high throughput, in a sense.
You write in the paper that this method would be useful for testing activation of GCPRs …
Right, and the avoidance behavior is the readout. You measure the avoidance behavior. Obviously, you have to compare it to the wild type that doesn’t express the GPCR. And obviously there are caveats, such as: What if the worms are innately repelled by a test compound? This is highly unlikely, unless it expresses a similar receptor in its own neurons.
Do you think the throughput of this will be amenable to drug discovery?
We are actually looking into that. Right now, we’ve only just carried out these experiments about six months ago. High throughput is definitely going to incur quite a bit of cost and more manpower, because so far, I’ve done most of the experiments myself. We have explored platforms such as 96-well plates that would make it a possible drug screening tool. The plates that we’ve used have a membrane, where you could embed your test compound, and if the worms are actually repelled by it, they won’t cross the membrane to the other side, and you actually count the worms after X amount of time. It’s doable. We haven’t done it in a high-throughput manner, but we’ve tested it on a small scale.
Is there any issue with interpreting the worms’ responses? Does a person just look at the worms and make a decision, or have you explored automated analysis methods?
There is someone in our group that is looking into developing software that would count worms at the end of 30 minutes, for example. You would put the plate on an automated stage, and the software would run itself, and the stage would move, and you would just take pictures of each well.
Is there any difficulty in handling the worms and preparing them for screening purposes?
The whole beauty of the C. elegans system is that it is so easy to handle. You don’t need extra-sterile techniques because they grow on bugs. You grow a lot of bacteria, and the worms feed on the bacteria. The fact that they are so easy to go and you don’t need sterile technique means that you can grow a lot of worms in a short amount of time for testing.
Has your group applied for patents on this?
Unfortunately, the group that I previously mentioned, which first expressed the olfactory receptors, has actually filed a patent that covers an enormous amount of ground. We are actually exploring other aspects of transgenic C. elegans for screening purposes. They did talk about expression of mammalian GPCRs in gustatory neurons for screening. I don’t think they covered every single aspect of it. We have to look at the patent again.
So there is an interest in commercializing this for drug discovery?
Can you use any other phenotypic response besides avoidance to assess GPCR activation, such as fluorescent or luminescent reporters?
These animals are actually transgenically expressing a GFP construct, so their gut is fluorescent. If you look at the supplementary materials for the paper, we describe how the animals can be sorted using conventional flow cytometry, and they actually glow green. It’s really easy to sort them according to size and fluorescence. You can actually use the fluorescence as a readout, but technically speaking it’s not that straightforward, given that these animals are mosaic. The levels of brightness vary between different animals.
There is another way of doing it, but it’s not that practical – it’s mainly for academic purposes. But you could do calcium imaging, if you want to look at whether an animal responds to a chemical. In calcium imaging, you have a construct that you clone into your neuron in question, along with your GPCR of interest, and you look at changes in fluorescence intensity to measure it quantitatively.
This type of approach is commonly used in cell-based assays, but it would be more difficult with nematodes?
Yes, because you’re talking about high magnification, immobilizing the worm on an agar pad, and then bathing the worm in your test solution. All of that takes an enormous amount of time. If you want to quantify your response, however, that would be a way to do it.
Why would someone want to use the nematode approach for GPCR screening as opposed to cell-based assays? Is it more physiologically relevant?
That is one main reason – it’s an in vivo assay. It’s not straight in vivo, but you are expressing this in a live animal, so you are almost emulating what’s happening in real life. But the real beauty of this, which I think most people have not realized, [is that] these animals actually feed on bacteria. The main drawback of a cell-based approach is that you need incredibly pure compounds. It has to be endotoxin-free, and very pure. But with the worms, we can test very crude extracts. If they can feed on fungi and bacteria, then that means you can throw very crude extracts at the animal to do screens. You get a very quick answer, as well. This could be the first part in a test, to see if a drug generates a response. If it does, then you can go to more detailed studies in cell-based assays. Obviously it will cut down on your costs, and you will get much quicker answers – as opposed to, for example, testing hundreds of 96-well plates with cell membranes, which would be very costly.
So what are your next steps for developing this assay?
We’re going to explore the use of crude extracts more to screen potential ligands, or even using non-ligands to further validate the method. It would be very exciting if we could use crude extracts, because it means people wouldn’t have to go into a cell-based assay anymore. They could use this to validate their results without going on to more expensive methods.
One last thing – an interesting aside of this is that human receptors can talk to worm proteins. I think biologically speaking that is very interesting. Essentially there is not that much difference between man and worm.