hERG Screening 101

 
In the late 1990s a number of drugs, approved by the FDA (the U.S. Department of Health and Human Services Food and Drug Administration) and available on the market, had to be withdrawn from sale in the US when it was discovered they were implicated in deaths caused by heart malfunction. It is now known that a side effect of these drugs was the blocking of hERG channels in heart cells. This caused prolongation of "action potentials"—the electrical pulses responsible for controlling heart muscle cells. With proper control of the rate of heartbeat lost, dangerous arrhythmias could develop, which lead in some cases to death.

The hERG channel is one of a family of ion channels the first member of which was identified in the late 1980s in a mutant Drosophila melanogaster fruitfly (1). Presence of the channel was identified by leg shaking in the flies when they were anaesthetised with ether, leading to the gene that encodes the channel being dubbed the "ether-a-go-go" gene. "hERG" is an acronym for "Human Ether-a-go-go Related Gene," with the focus now on the heart rather than legs, and on humans rather than fruitflies. The hERG gene encodes the pore-forming subunits of this potassium channel, of which the biophysical properties have been described in detail (2,3).

Figure 1. Blockade of hERG can cause Long QT Syndrome. Normal heart rhythms are governed by action potential propagation in the heart. Measurement of the extracellular potential (i.e., surface electrocardiogram) from a beating heart usually provides a waveform shown in the top figure above labeled "Normal heart rhythm." The QT interval refers to the time taken during a heart beat cycle, which includes both depolarization (or the "spike") and repolarization (preparation for the next cycle). Blockade of hERG channels can prolong this QT interval and lead to potentially fatal arrhythmias, including Torsade de pointes (example in the lower figure).

hERG, An Important Cardiac Ion Channel

The heartbeat is controlled by electrical signals resembling those in nerve. Action potentials are caused by a sequence of events in which ion channels play the primary role. In heart cells in their "resting state," cell interiors have a fixed voltage relative to the outside of the cell. When an action potential is triggered by a voltage pulse from a neighbouring cell, sodium ion channels open, allowing a rapid influx of these ions into the cell. The voltage in the cell rises very quickly (Figure 2). The sodium channels remain open for only a very short time, however. The higher voltage is maintained by a second type of ion channel opening, allowing calcium ions into the cell. When the pulse ends, the cell must return to its resting state. A number of different ion channel types and the Na/Ca exchanger are responsible for this, all opening in sequence to allow potassium ions out of the cell. These ions are moving into a more electrically positive environment outside the cell, i.e., against the electrical gradient. They are pushed by the very high concentration of potassium ions in the cell relative to that in the solution outside the cell.

Figure 2. IKr blockade can prolong the cardiac ventricular action potential. Activation of Na and Ca currents is responsible for the depolarizing upswing (or spike) of the action potential. Activation of K channels and inactivation of these Na and Ca channels is responsible for the repolarizing decay of the action potential. Prolonging the cardiac action potential due to block of hERG channels increases the QT interval and can lead to harmful arrhythmias (Figure 1).

The hERG channel is an important potassium (K) channel in this sequence. K channels contribute to the final phase of the action potential that returns the cell to its resting state. If a drug suppresses the activity of the channel in any way it can lead to prolongation of the action potential. With the time span of an action potential about 300 milliseconds, it only needs blocked or partially blocked hERG channels to extend this by five or ten milliseconds to create a potentially dangerous situation. This effect was first detected, early in the 20th century, by ECG (electrocardiogram) measurements taken from the body surface near the heart. The first and last stages of ventricular action potentials detected by ECG are commonly labeled the Q and T waves, thus the 'QT interval' is the critical measurement.

Figure 3. Patch clamp is the gold standard assay for ion channels. In the whole-cell configuration of patch clamp, current flowing through all ion channels in the cell are recorded as one macroscopic current (I) in response to voltage steps. The units are typically picoamperes (pA) or nanoamperes (nA). Potassium currents (like hERG) are outward currents because K ions flow across their electro-chemical gradient from the inside of the cell membrane to the outer surface. The concentration of Compound X that blocks 50% of the channels is called the IC50 concentration (IC = inhibitory concentration).

Plainly, hERG-blocking properties can end the prospects for a potential drug. Frustratingly, there is now no way to predict from a drug's structure whether it will block hERG. Therefore, testing on these channels thus occurs early in the drug-screening procedure. A first screen run on a compound library of 100,000 compounds might be expected to return 100-odd compounds worthy of further testing. In many companies these will all be tested for hERG blocking before any further investigation is carried out—there is no point in going on with a compound that can never get to market.

The PatchXpress™ 7000A screening platform from Axon Instruments, Inc. is ideally suited to conduct such testing. With high-quality electrical current recording while each compound is added it can capture the very rapid peak of activity indicative of hERG activation, and its absence if blocking occurs. It can be easily set up to test each compound at various concentrations, adding successively higher concentrations. In cases where hERG blocking does occur, the concentration dependence of this block can be measured and compared with the concentrations that produce the candidate drug's beneficial effects on other targets. The anatomy of a hERG current waveform is shown in Figure 4.

Figure 4. Anatomy of a current waveform in a single CHO-hERG cell. An example of whole-cell hERG current is shown here. From a holding potential of -80 mV, the voltage is first stepped to -50 mV for 500 ms. This step to a voltage in which hERG channels are not opened is important for leak subtraction. From -50 mV the voltage is stepped to +20 mV for 2 seconds. At this voltage, hERG channels open and steady-state current is observed. From +20 mV, the voltage is stepped back down to -50 mV. An immediate increase in hERG current amplitude is observed for the following reasons. The inactivation rate constant is faster than the deactivation rate constant. This means that inactivation is quickly removed, but there are many channels that have not proceeded to the closed state from the opened state. This results in the observed "rebound" or tail current. Typically, this tail current amplitude is measured and the leak current measured at -50 mV is subtracted out. (Recording by Iris Yang, Axon Instruments, Inc.)

The hundreds of compounds for testing at this degree of accuracy provide a formidable challenge to pharmaceutical companies, who in the past have relied on individual electrophysiologists working at manual patch-clamp rigs to work through the compounds (Figure 3). This process is slow, tedious, and boring, wasting the talents of highly trained and creative scientists. The PatchXpress can automate the entire process resulting in faster and cheaper screening of compounds, with data comparable in quality to the results of manual experiments.

Figure 5. Definitions of "cell health" parameters. Measuring membrane parameters is an important part of accurate patch clamp experiments and is featured in the PatchXpress. The seal resistance (Rseal) is the resistance between the cell membrane and the patch electrode substrate (before the whole cell configuration is obtained). The PatchXpress SealChip electrode arrays are capable of achieving giga-ohm seals ("gigaseals") which is important for long lasting recordings that are devoid of artifacts caused by fluctuations in leak current. The membrane resistance (Rm) is the resistance measured once the whole-cell configuration is achieved. This parameter includes the leak resistance, the input resistance and Rseal. Rm is governed by the lowest of these three resistance values. Typically, in PatchXpress experiments, Rm is > 500 MΩ. The access resistance (Ra), often called series resistance or Rs, refers to the resistance between amplifier and ground (or between Vactual and Vcommand as shown in the figure). Ra includes the electrode resistance (Re), which is usually around 2 MΩ, and the other resistances in the pathway through the cell to ground. Ra is an extremely important parameter that needs to be monitored throughout a patch clamp experiment. When Ra values are high, voltage errors can occur and this can lead to artefactual distortions in the whole-cell current. These distortions would be particularly harmful in a drug screening campaign (i.e., what looks like a hit might actually result from a change in voltage caused by a deviation in Ra).

Johnson & Johnson Pharmaceutical Research & Development, L.L.C.
(J&JPRD) Experiments

Dr. Adrienne Dubin and her colleagues (J&JPRD, San Diego) conducted experiments with known hERG channel blockers using the PatchXpress 7000A and compared the results to results obtained with a conventional patch clamp rig. She first reported these results at the Inaugural PatchXpress User's meeting (4). Dr. Dubin reported a success rate of 50% for experiments lasting at least 15 minutes (with Rm > 200 MΩ and Ra < 15 MΩ, throughout; see Figure 5 for details about these membrane parameters). Dose-response experiments were conducted with eight known hERG blockers (Figure 6). Seven of these eight compounds showed similar potency whether tested on the PatchXpress or with a conventional patch rig. One compound, flunarizine, was relatively right-shifted when tested on the PatchXpress. Further studies done at Axon showed that similar flunarizine potency could be achieved with the PatchXpress if the compound solution was made up fresh and used in glass coated compound plates just before the experiment. An example from a single SealChip16 is shown in Figure 7.

Figure 6. Determination of the rank order potency of eight compounds and the fold difference in IC50 compared to previous electrophysiological studies. The rank order potency of the compounds to block hERG in the PatchXpress assay was similar to that determined using conventional with one exception: flunarizine, a notoriously insoluble compound, was 6-fold less potent. Further studies performed at Axon demonstrated similar potencies between conventional patch and the PatchXpress with flunarizine when the compound was made fresh and glass-coated compound plates were used. Optimal potency determinations required very healthy cells and since single cells in a particular batch of cells show variability, each cell underwent a quality assurance protocol. pIC50 (-logIC50) values for 8 compounds were determined from conventional whole cell patch clamp technique (EP T162, EDMS-PSDB-2207503; gold bars) and the PatchXpress (green bars).

Figure 7. Example of dose-response experiments from a single SealChip.

Flunarizine, like terfenadine, is a very hydrophobic compound that is difficult to get into solution. As shown in Figure 8, Quintiles reports similar IC50 values for terfenadine using the PatchXpress (28 nM) and conventional patch (27 nM).

Figure 8. Terfenadine dose-response experiment:            conventional patch vs. PatchXpress.

In her presentation at the User meeting, Dr. Dubin reported that no false positives or negatives were observed with the PatchXpress. To test for false positives, 33 compounds known not to block hERG channels (i.e., less than 50% block at 10 µM) were applied to cells patch clamped with the PatchXpress. None of the 33 compounds tested were detected as blockers in this screen. To test for false negatives, 29 known hERG blockers were applied and all 29 compounds acted as blockers on the PatchXpress. The results reported by Dr. Dubin were conducted over 13 days with 36 SealChips.

	Mean	SEM
Re (MΩ)	2.02	0.01
Rseal (GΩ)	2.53	0.07
Rm (GΩ)	1.44	0.04
Ra (MΩ)	5.87	0.09
Example data from 532 cells (HEK-hERG cells)

Table 1. Cell health statistics.

Screening Services

Aptuit

Aviva Biosciences

BioFocus DPI

ChanTest, Inc.

Evotec AG

MDS Pharma Services

Millipore

Acknowledgments

Many thanks to James Rountree (Axon Instruments, Inc.) for contributing to the writing of this page. Dr. Kirsty MacFarlane (Quintiles) kindly provided the terfenadine data and cell health statistics table.

References

1. Jan, L.Y. and Jan, Y.N. (1990). A Superfamily of Ion Channels. Nature, 345(6277):672.
2. Sanguinetti, M.C., Jiang, C., Curran, M.E., and Keating, M.T. (1995). A Mechanistic Link Between an Inherited and an Acquired Cardiac Arrhythmia: HERG encodes the Ikr potassium channel. Cell, 81:299-307.
3. Trudeau, M.C., Warmke, J.W., Ganetzky, B., and Robertson, G.A. (1995). HERG, a Human Inward Rectifier in the Voltage-Gated Potassium Channel Family. Science, 269:92-95.
4. Dubin, A. (2004). HERG Potassium Channel Activity Assayed with the PatchXpress Planar Patch Clamp. Inaugural PatchXpress User's Meeting, February 12, 2004 (Baltimore, MD).