Boehringer Ingelheim Pharma
At A Glance
Name: Martin Valler
Position: Group leader, high-throughput screening, Boehringer Ingelheim Pharma, Biberach, Germany
Background: Group leader, GlaxoWellcome R&D, 1995-2001; Senior research scientist/research scientist, Glaxo Group Research, 1986-1995; PhD, University College Cardiff.
Martin Valler and colleagues at Boehringer Ingelheim in Germany recently published research in the Journal of Biomolecular Screening demonstrating the use of Applied Biosystems' InteraX protein-protein assay system in high-throughput drug discovery. Though Valler's group works with all forms of screening, it has recently significantly increased its use of high-throughput cell-based assays such as InteraX, as well as high-content image-based screening. Valler took a few moments last week to discuss these trends with CBA News.
What is your division responsible for at Boehringer Ingelheim?
I'm responsible for a group called high-throughput screening, and that is part of a department which has responsibilities for structures for handling compounds, structural research, and computational chemistry; so computational methods for lead discovery. And the entire department is in lead discovery. Our specific responsibility is for screening development; screening; some assay development in collaboration with a therapeutic area, for customers really; and secondary screening. We're taking on an increasing proportion of secondary screening the profiling of compounds. But we don't, at this point, get involved in what you would call pure lead optimization. And we have some responsibilities for hits to leads, which is sort of the intermediate stage. So the key themes would be assay adaptation, very flexible assay biology, high-throughput screening, and a lot of adaptation to our automated systems.
In terms of cell-based assays not necessarily high-content, but also high-throughput how involved is your group?
Increasingly. There is certainly an increasing demand for cell-based assays because our therapeutic area customers value the extra information, and that's not just in this so-called high-content area. That's talking about cell-based assays in general. As a proportion, a good 50 percent, or perhaps more than that now, of our total screening workload is in cell-based formats. And that's in primary and secondary screening. That's increased substantially over the past four years.
You mentioned that it's not necessarily high-content screening or imaging … can you give some examples of some of the other types of technologies you're referring to?
What our main cell-based technology has been is fluorescence imaging with FLIPR, either for calcium, for looking at agonists, or indeed looking at antagonists, which we've increasingly used FLIPR for directly; or using it with membrane-potential dyes to look at channels, binding channels, transporters, etc. So that is certainly the major input, and this is what we've built up in the past three years. This is high-throughput, fully automated we have two fully automated systems with FLIPR readers, and they can deliver a throughput of up to 60,000 compounds a day, up to about 100 384-well plates.
How did you come to experiment with the Applied Biosystems technology?
We had seen the technology published although not that many publications and we cite some of the first papers in our most recent paper. It's a very intriguing concept, and we've found it to be a very promising way to produce a very generic readout that spans the gap between in vitro binding assays, and functional assays in the sense of a pure unmodified cell-based assay. And it had that potential for providing a sort of bridge between those two worlds. That was the main attraction. It's based on this idea of the fusion proteins, which are halfway natural, but they're also a cellular reporter for this binding function. And also, the whole concept is very generic. We try to develop technologies that we don't have to develop multiple times one after the other. If we can produce a platform which we can then use for multiple targets, that is very much a benefit for us, because we've spent a very large amount of time here either in technology development or target development. We're really trying to minimize that, and this technology has the promise to do that.
How does the InteraX assay work in the context of high-throughput screening? It is based on a beta-galactosidase reporter assay, right?
There aren't really any alternatives that are directly comparable. The beta-galactosidase function in the InteraX system is used purely as a reporter, and theoretically it could be other enzymes as well, but Applied Biosystems has particularly optimized it for beta-galactosidase. But there are other enzymes that also exhibit this complementation functionality you have these two fragments which on their own are inactive, but when they are forced together by association of their partners, then you get this activity. So it's not unique to beta-galactosidase, but the ability to see the interaction of two partners in their membrane, in their native environment, through a reporter system that should have relatively little interference on the native functionality of the membrane proteins, is new and interesting to us. As an alternative the paper that we did was with EGFR but I think the alternative would be for someone to look for cellular proliferation in the presence of EGF and competing substances, so the readout would be functional through the actual native activity of EGF. That would be much more difficult or time consuming.
The readout for this assay is simple luminescence detection?
Well, yes, in a way. What you see is an activity, an enzyme activity, but how it differs from, say, a fragmented GFP, is that you produce here an enzyme, and you can let this run with a substrate. In this case it's a luminescent substrate, but it could be other substrates. But the advantage here is that you get an additional amplification, so instead of looking at the fluorescence of one molecule of GFP, with this system you're producing a luciferase luminescent signal that is amplified many times by the activity of the enzyme. So it has that potential for being significantly more sensitive, I would say.
Do you use any particular instrumentation?
We use the substrate that comes as part of this, which is actually a luminescent substrate, and the instrument we use is the TopCount, from PerkinElmer, which you can say is actually overkill for the sensitivity you actually need. That is really more of a scintillation counter, actually, but Applied Biosystems claims that this could be used with most luminometers. We were measuring for one second per well, which is very short. I think with a normal luminometer you could measure for a few seconds and get away with it it's quite a strong signal.
Cell-based assays in general seem to have been more difficult to implement in high-throughput screening than secondary screening, et cetera. Do you see that, and if so, why?
The major burden that we've seen with cell-based assays is first of all the establishment of a functioning cell line that actually gives a robust signal; and then actually translating that signal into a high-throughput environment. That is absolutely the major burden. The actual implementation, the running of it once it's set up, is not so much of a problem. We have automated systems that can handle plate transport, pipetting; we have capacity for holding plates, and we use 384-well plates so the throughput is pretty reasonable. Where we see the major problems are in the early phases: Getting a clone that works properly, things like DMSO sensitivity we have always had problems here, because we try to have the highest concentration of compound that we can because we want to achieve a relatively sensitive assay. But then as soon as we get anywhere near 0.5 DMSO or thereabouts, the cells will typically not tolerate that. And also the stability of a signal through multiple passages is an issue. The actual cell culture keeping them in culture, plating them out, producing nice stable reproducible plates and well it takes a while for the whole process. The good thing is that if we get a stable, running system, then the screen will run very smoothly. The initial phases can take several months, though. It's amazing how long it can sometimes take.
Is a lot of this work done in house, or is much of it contracted out?
We have experience with companies that have very well-optimized cell lines that are ready to use. Companies like Applied Biosystems in this case we obtained cell lines from them. But also, for instance, with our high-content screening, we obtain our cell lines from GE Healthcare, the old Amersham group, and also from Euroscreen, for instance. All of these companies produce cell lines with well-established properties. But we still have many projects where we would do that work internally. We're lucky within Boehringer Ingelheim in that we have a biopharmaceutical branch that helps us considerably in producing cell lines with exactly these properties.
Do you see your group moving more into image-based cellular screening?
The area of high-content screening is really critical for us right now, and we've put a lot of time and money and effort into building up those systems. Initially, where they stand right now, we use them most for secondary screening. And this is for profiling compounds that are coming out of our more conventional screens. So we use the IN Cell 3000 system from GE, and the nice thing about that system and its automation, is that it does have the throughput to enable primary screening. Simply where we are right now is building out our capability, and we have not ruled out that we would do primary screening with it. The capacity and throughput is probably on the order of 40 384-well plates a day, which is maybe 25,000 compounds a day, which is not ultra-high-throughput screening by our standards. But at the same time, it is a throughput that is primary-screening compatible. We will certainly aim in that direction, but right now we look at primarily multiple projects with secondary screening, and they've been very interested in the results they're getting. This is a great development area for us. What we'll need in the future, as always with industry, is that you have to build up and show demand, and then you get the capacity you need. We're using the capacity we've got, but we do have demand, and questions about whether we can do primary screening. The answer is 'Yes, we can,' but we'll have to extend our capacities in order to do that.