CESE Plus: a computational approach for accelerating electrophysiology research

---

Sandy Vascotto PhD, Sergey Missan PhD, Chicuong La PhD, LLB

Since the seminal work by Hodgkin and Huxley demonstrating that excitability of cells is controlled by ions crossing the cell membrane, the field of electrophysiology has expanded to become a mainstay in identifying new targets and securing the safety of lead compounds in the pharmaceutical industry.

Electrophysiology investigates properties of membrane proteins called ion channels and pumps, and their contribution to electrical activities of the cells. The physiological outcomes of cell excitability include the coordinated heart beat, nerve impulse generation and propagation, activation of sensory receptors, muscle contraction, hormone release, and cell ion homeostasis.

Cell excitability can be evaluated at different levels; that of the individual ion channels and pumps, a change in the conductance or potential measured at isolated cell membrane, propagation of the excitation wave (the action potential) from cell to cell, or electrical activity of the whole heart or brain (Figure 1).

Regardless of the level of interest, the measured signal depends on the activity of a population of ion channels and pumps that determine the flow of particular ions across the cell membrane. As such, the goal of many modern electrophysiology experiments is to determine how a test compound affects the biophysical properties of selected ion channels, how ion channels are affected by regulatory cell signalling pathways, or how a gene modification affects the gating properties of a channel.

Challenges in Modern Electrophysiology Research
While the past 30 years have brought unprecedented technical advancements to the field of electrophysiology, many of the underlying experiments remain tedious and time consuming. Measurements of minuscule currents (pA-nA range) or voltages (mV range) across a cell membrane using microscopic glass electrodes or pipettes pose great technical challenges.

For example, sensitive “patch clamp” recordings require one to learn fresh cell isolation from heart or brain tissue, microforging of borosilicate patch pipettes, positioning of recording pipette in contact with cell membrane, and formation of stable gigaohm seal between the pipette and cell surface.

Additionally, one must understand the influence of electronic compensatory circuitry on the recording characteristics, recognize and correct for multiple factors that reduce the signal to noise ratio of the measurement, and maintain the optimal conditions of cell preparation during the recordings. An expert patch clamper might only evaluate seven to nine cell recordings in a day, whereas 15-20 data points are needed for a determination of a simple dose-response relationship. Such practical and technical demands can significantly hamper workflow particularly when compounded by scientific uncertainty. To this effect, the interpretation of the results obtained in a typical electrophysiological study is rather difficult.

Due to the intrinsic variability in the individual experiments and a large number of contributing factors (cell condition, intracellular ion concentrations, activity of cell regulatory pathways, recording stability) reproducibility of results can be challenging. Consequently, the number of experiments required to achieve certainty must be increased. Finally, there is no bridge between the various levels of analysis in electrophysiology. Different experimental setups and analysis routines are used to record currents through ion channels, action potentials, or the ECGs (i.e., voltage clamp, field tissue recordings, animal/clinical models respectively) and correlations are indirectly made between them. Thus, technical and practical workflow limitations complicate electrophysiology experiments and their interpretation.

Computer Simulations in Electrophysiology - A Cardiac Example
Since Hodgkin and Huxley’s mathematical description of the ionic mechanism of neuron excitation, electrophysiologists have pursued better computational models to describe the specific system within which they retain interests. The cell model in electrophysiology is comprised of a system of non-linear differential equations which describe the biophysical properties of individual membrane conductances. The sum of all currents at a particular time may be used to calculate a new membrane potential, and ultimately the action potential properties.

These principles have been extended to describe electrical properties of various excitable cells, including those of cardiac myocytes. To date, over 50 models of the electrical activity of cardiac myocytes have been published. They describe electrical properties of heart cells in different species (guinea pig, rat, mouse, rabbit, canine, human) and myocytes from different regions of the heart (ventricle, atria, SA node). The latest generation of the models is very complex, describing not only individual currents through ion channels and pumps, but also changes in the intracellular ionic concentrations, including Ca2+ buffering and release from sarcoplasmic reticulum. In addition, specialized models have been created to assess the influence of phosphorylation pathways, mechanical stress, and mutations of individual ion channels on the action potential amplitude and shape.

Traditionally such models have been created using different programming languages and approaches and non-standard notation to define variable names and units, making it extremely difficult to use and compare their outputs.

CESE Plus Platform and Simucore Models
Cell Electrophysiology Simulation Environment Plus (CESE Plus) is a unified simulation platform for running complex electrophysiological cell models (Figure 2a). It was created as a scientific effort to unify the many approaches used by model developers and to enable mainstream electrophysiologists to use simulations in their research. CESE Plus provides all supporting infrastructure for performing simulations “out of the box” - differential equation solvers, access to model parameters, data visualizations, analysis, and import/export. Using the Simucore framework, the models have consistent names for common variables.

This reduces learning time and allows for the easy transfer of simulation protocols from one system to another. The models are implemented using one programming language and a common approach based on sound industrial design (Figure 2b).

1) Simucore Models Emulate Individual Ion Channels and Action Potentials
Currently, a number of popular and well-tested models of cell electrical activity are available for CESE Plus (e.g., guinea pig, rabbit, canine and human ventricular and atrial myocyte models). The baseline action potentials generated from such models are comparable to the actual experimental traces (Figure 3a). CESE Plus provides a number of powerful data visualization possibilities designed for quick comparison of outputs between different simulation runs in the same model, between different models, or between a model and imported experimental data. The simulated traces can be analyzed, and properties such as peak amplitudes or mean values within regions of interest measured. The current-voltage relationships can be determined, plotted, printed and exported for further analysis.

With the capacity to create powerful simulation protocols that re-create voltage clamp and current clamp experiments, researchers have the opportunity to compare properties of simulated currents with “real life” records.

Supporting improved workflow, simulation at a single cell level enables the seamless translation of channel activity to an action potential - a task that would normally require the use of an entirely separate experimental setup. For example, if the amount of channel block by a test compound was determined in a patch clamp experiment, the result can be instantaneously modeled in a range of conditions (different heart pacing rates, ionic concentrations), and effects of the block on contributors of the action potential evaluated (Figure 3b).

Typically, measurement of the impact of a test compound on multiple ion channel targets would require months of experimental work. Simulations provide the scientist with capacity to prospectively hypothesize on interaction between an array of factors and subsequently plan and test to better resolve the mechanism of action. Finally, these simulations may be extended to such systems where experiments are costly or complicated due to regulatory issues (e.g., use of human cardiac myocytes or tissue preparations).

2) CESE Plus Expands Experimental Options, and Bridges Ion Channel data to Potential Clinical Risk Indicators
While in an experimental setting one has very limited options in terms of experimental design and capacity for interpretation, within the CESE Plus environment the scientist can run a series of experiments and controls for confounding variables in parallel. The CESE Plus variable control system - VirtuClamp - allows one to fully investigate the impact of a theoretically complex series of factors in combination on ion channel activity. In essence, this allows one to perform the simulations that could require months of wet-lab experimentation - under conditions of uncertainty and inconsistency between material preparations - and theoretically test them in a single day under validated postulates that are guaranteed to be consistent. The results of such experimentation can be extended to the relevant action potentials and scrutinized for indicators of cardiac safety risk.

This is particularly relevant for in vitro assessments during pharmacological discovery and cardiotoxicity testing.

An example of such an application is presented in the simulated impact of a theoretical compound on the HERG (IKr) in Figure 4. The results extended to the action potential show that with a block of IKr by 70% we see a 26% increase in the action potential duration (APD90) – or the time from the overshoot to 90% repolarization (Figure 4b). Many studies have shown that a 5-10% increase in the APD90 increases the risk of life-threatening arrhythmias, such as Torsades de pointes (TDP). Furthermore, a block of IKr by 70% together with an increase in heart rate from 120 to 180 bpm, using CESE Plus, reveal two classic biomarkers of arrhythmic risk: premature beats and action potential triangulation (Figure 4c).

Another great benefit of computer simulations is the ability to test the physiological effects of a multi-channel block by a potential drug compound. As demonstrated in Figure 4d, a compound that blocked the IKr (potassium current) by 70% in addition to blocking the ICa,L (calcium current) by 50% generates an action potential that has a 31% reduction in amplitude during the plateau phase but no significant change in APD90 due to mutually-canceling effects of IKr and ICa,L blocks. Therefore, this compound has low risk of arrhythmogenicity at these conditions. While in general many indicators of risk are comprised of difficult to discern subtle changes in electrophysiological parameters, simulations generate clear results under well-defined experimental conditions.

Conclusions
Electrophysiology employs powerful methods to discover the basic mechanisms underlying cell electrical activity and their importance as safety indicators in pharmaceutical applications. The disjoint between experimental systems and technical limitations, however, pose some difficulties for a widespread adoption of practical electrophysiology studies. In contrast, simulations allow one to assess multiple “what if” scenarios in a fraction of the time required for a single electrophysiological recording. The precise control of almost any cellular physiological parameter is possible either before, or during simulation. A complex pharmacological profile of a test compound which affects multiple ion channel targets can be re-created in a selected model, and its potential effects on action potential properties evaluated.

CESE Plus is an advanced tool that can bridge the gap between experimental data derived from patch clamp experiments to determining action potential properties, and cardiac safety predictions. In summary, the integration of simulations in CESE Plus into the electrophysiology workflow has a potential to save time, money, and uncertainty, while expanding the ones capacity to hypothesize and interpret results.

CESE Plus is available through Simulogic Inc. Visit www.simulogic.com for further information on simulation tools for electrophysiology.

Figure 1

Different levels of computer models of cardiac electrical activity and their simulated output. Results obtained at each level can be used to determine the cardiac safety of lead compounds in basic research and pharmaceutical applications. Simulations allow one to transcend levels from the single channel to whole heart propagation within the same model – a feature not available in classical electrophysiology.

Figure 2
CESE Plus electrophysiology simulation platform

a) Examples of the CESE Plus user interface. b) The main components of Simucore cell models.

Figure 3
CESE Plus compatible models emulate
the properties of membrane currents and
action potentials

a) Action potentials derived with Simucore models reproduce experimental traces. Inset: example action potential recorded from guinea pig ventricular myocyte. b) Simulated 70% block of an actual ion channel (HERG – IKr) – dotted square – propagates to parameters of the model and affects properties of the action potential. Superimposed traces for control and IKr block conditions are shown.

Figure 4
CESE Plus simulates complex combinatorial
drug effects

a) Simulated ventricular action potential under control conditions at 120 beats per minute (bpm). b) A 70% block of the IKr increases the APD90 indicative of arrhythmia risk. c) An increase of heart rate to 180 bpm results in a premature heart beat. d) Simulated administration of a compound that blocks 70% of the IKr and 50% of the ICa,L does not significantly affect APD90.