Senior principal scientist
Schering-Plough Research Institute
At A Glance
Name: Steve Sorota
Position: Senior principal scientist, Schering-Plough Research Institute, Kenilworth, NJ
Background: Principal scientist, Schering-Plough Research Institute, 1999-2002; Associate research scientist/assistant professor, pharmacology, Columbia University, 1988-1999; Postdoc, pharmacology, Columbia University, 1986-1988; PhD, pharmacology, University of Connecticut Health Center, 1986.
Steve Sorota has been investigating ion channel regulation in cardiac cells and tissue for nearly two decades, during which time he has witnessed the emergence of highly parallel patch clamp technology — such as Molecular Devices' IonWorks HT — as a go-to tool for high-throughput drug screening against ion channels.
Recently, he and colleagues at the Schering-Plough Research Institute conducted a study comparing IonWorks with more traditional ion-channel assays for screening modulators of the cardiac cell ion channel, hERG. The results, published in the Feb. 2005 issue of Assay and Drug Development Technologies, revealed some interesting insights about the performance of IonWorks in such an assay.
Last week, Sorota took a few minutes to discuss the results with Cell-Based Assay News.
What type of research do you conduct at the Schering Plough Research Institute?
I've done work related to safety as well as discovery, and my focus has been on ion channel research. The research has encompassed a number of different methodologies and approaches, including doing cell-based screening, lead optimization and lead selection, tissue electrophysiology, single-cell electrophysiology, and pre-clinical proof-of-concept studies.
In the Assay and Drug Development Technologies paper, your group compared Molecular Devices' IonWorks to two other types of assays: conventional hERG patch clamp, and functional screens based on rubidium flux. What was the basis of choosing these for comparison?
The conventional patch clamp we selected because it really is the gold standard — it's the most reliable method for evaluating ion channel function, particularly when it's used by a well-trained investigator. And anytime you're trying to validate a higher throughput method, one is obligated to make the comparison to conventional patch clamp, because we know that is the highest fidelity procedure we have for looking at ion channel function.
We selected the IonWorks at a time when it was new to the market and it offered a level of throughput that we thought would match what we needed to look at a hERG liability screen. In terms of rubidium, that was an existing screen that was used pretty widely and still is a commonly employed procedure — a higher-throughput procedure for looking at potassium channel function, in general, and in this case, more specifically, hERG channel function. We settled on rubidium versus, say, a fluorescence-based screen because in our hands, the hERG rubidium efflux screen performed better than a fluorescence-based screen using voltage-sensitive dyes. That was primarily because the rubidium efflux screen had less severe potency shifts than we observed in the fluorescence-based screen.
One of the major conclusions of the paper was that IonWorks did not outperform the rubidium efflux screen. Is that a result that your group expected?
We were surprised. The IonWorks permits one to actually voltage-clamp single cells in parallel, and in theory, there was reason to expect that this would enable one to have a better performing screen than, for example, a cell-based screen where one does not have complete control over transmembrane potentials, such as the rubidium-efflux screen. After we convinced ourselves that the IonWorks hERG screen first of all was working fine and had a smaller average potency shift than the rubidium-efflux screen, we in fact thought that the IonWorks would outperform the rubidium-efflux screen in a single-concentration mode. But what we neglected to consider at that point was that both screens exhibited a compound-specific potency shift. Just as an example, a couple of compounds — E-4031 and quinidine — did not show any potency shift on the IonWorks; whereas the IC50 for another compound — pimozide — was 12-fold higher on the IonWorks than it was in comparison to conventional patch clamp. So this compound-specific and variable potency shift made it difficult to select an appropriate screening concentration for unknown compounds, and we think that's the reason why we essentially had comparable performance on the rubidium efflux-based screen compared to the IonWorks hERG screen.
So this is not so much non-functionality of IonWorks as it is certain conditions that researchers have to keep in mind when conducting this type of assay?
One has to remember that when we became involved, the technology was fairly new, and I think that the potency shift issue probably surprised us and others. There certainly was not any malfunction — the device did parallel patch clamp, and it did permit us to develop a perfectly valid assay for the ability of compounds to block hERG. Most screens will have some degree of predictive value, and the predictive value of most screens is not ideal, and that was true in this case, also. In terms of the device, I think it worked fine. In terms of the intended function of the instrument, we did not get to a point where we could develop a screen that outperformed a screen based on rubidium efflux.
What are the major advantages of a parallel patch-clamp device such as this?
Before the IonWorks, if you wanted to do cell-based ion-channel drug screening, one would typically use flux assays or a fluorescent dye where the dyes would either be sensitive to voltage or to intracellular calcium levels. The limitation to both of these other alternatives that were in place before IonWorks was the inability to fully control the transmembrane potential. If one is working with a voltage-gated ion channel, it's a distinct advantage to be able to clamp the transmembrane potential to a level where you know what state the channel is going to be in. Typically one would want to go from closed to open, and keep the channels in some sort of active form if you could, either by a single pulse or repetitive pulsing. In some cases the inability to directly control the transmembrane potential was an issue that could be pretty readily surmounted; an example of where one could do that would be voltage-gated calcium channels, where KCl depolarization in combination with an intracellular calcium dye gave you pretty reliable screens. In other cases, the inability to directly control transmembrane potential created a much more serious obstacle, and the IonWorks gets around that by and large by permitting one to directly clamp transmembrane potentials, and by doing so it can facilitate assay development for certain voltage-gated channels. The other advantage of it is that in terms of other voltage-gated patch clamp devices that are out there, the IonWorks has the highest potential throughput of any of the devices that are still on the market.
What are your opinions of high-throughput patch-clamp devices in general in terms of efficiency and data quality, and improvements that you'd like to see made?
Part of this answer is dependent on how I've interpreted data that's been presented by others as opposed to direct experience with some of the devices. I think in terms of efficiency and data quality, there has been significant progress made by a number of vendors over the years, but there is probably still room for improvement. So we've talked a little about potency shifts already, but other factors that probably can be improved upon include overall success rates; the stability of recordings, by which I mean how long one can record from a cell without the recording configuration falling apart in some way; and cell-to-cell reproducibility can probably be improved on most of the available automated patch-clamp devices.
You mentioned that IonWorks had the highest throughput potential of anything you had seen at the time. Are they still leading the way in this area?
I think Molecular Devices is an important player in this field — they have the IonWorks as well as the PatchXpress. There certainly are other vendors that are actively developing instruments.
Can you discuss any technologies on the horizon for these types of screens that might be an alternative to highly parallel patch clamping?
If you're talking about cell-based assays, specifically, I'm not so sure I would say there is anything out there. And with my bias, I think that if one wants to look at ion channels, I think cell-based assays are usually the way to go, as opposed to doing binding assays in membranes. So no, I think that what's on the horizon is further improvement of the currently available parallel patch-clamp devices.
Is there anything else at a level above cell-based, such as tissue?
That would not really lead to the type of throughput one would need. There are procedures out there where people are doing things like surface recordings from cells, and trying to make inferences based on surface recordings, but my bias there is that direct recording of ion currents is probably going to turn out to be somewhat more reliable than trying to make inferences about ion channel function based on surface electrical activity.