Abstract

Ion channels play a key role in regulating resting membrane potential
and cell excitability and are attractive targets for therapeutic intervention. Whole
cell patch-clamp electrophysiology remains the gold standard for studying ion
channel function, but these methods are complex, technically challenging and
low throughput. Automated patch-clamp technologies offer a higher throughput
alternative and are central to current ion channel drug discovery efforts, but
these methods remain relatively expensive and inaccessible. An alternative
approach to hit identification for ion channels is the use of fluorescence-based
assays. One example is thallium flux, which measures the flow of thallium
through potassium channels, detected with thallium sensing dyes. These plate-based
assays are compatible with standard laboratory automation systems and are
therefore suitable for high throughput screening and iterative compound
optimisation. However, they remain
a surrogate for channel function, and it is important to have an appropriate
panel of orthogonal and translational electrophysiology assays in place to
confirm activity at the channel of interest.

Using a thallium flux approach, we
screened a library of 100K small molecule compounds and identified 173 ‘hit’ compounds
as activators of a K2P potassium channel. These were then screened in an
automated patch-clamp assay (QPatch) to confirm activity. The aim was to
investigate which factors influence the correlation in activity between the
thallium flux and electrophysiology assay platforms.  Primarily we observed compounds with the
highest activity in thallium flux (relative to baseline) were most likely to
confirm in the automated patch clamp assay. However, the correlation between
activity in both assays was not entirely uniform and therefore a number of
other factors were investigated. The importance of compound incubation time was
tested in the thallium flux assay and highlighted that subtle differences
between methods accounted for some of the differences observed. Similarly, we
assessed whether the compound handling method impacted the response in thallium
flux. Compounds were dispensed on an acoustic dispenser (Labcyte Echo) or a
tip-based instrument (Biomek FX) and clear differences could be observed for a
subset of compounds. Finally, we looked at the properties of compounds to
assess whether calculated phys-chem properties were able to account for any of
the differences which were observed.

To
conclude, no single factor can adequately predict the correlation between the
two systems. It should be noted that fluorescent compounds and assay
interferers were removed before analysis on the QPatch. It is, however,
apparent that compound incubation time and liquid handling method can differentially
influence the response in each assay. Looking at these factors in combination,
prior to selecting compounds for electrophysiology, may help us to prioritise
screening outputs.