Rapid, label-free, contactless measurement of membrane potential in excitable H9c2 cardiomyoblasts using ζ-potential

The measurement of cell membrane potential (V m) is important for understanding ion channel function. V m plays a role in several routine cellular functions and diseases, particularly in excitable cells such as muscle and nerve. However, measuring V m is difficult, relying either on labour-intensive direct measurement of single cells (intracellular electrodes, patch clamp) or indirect measurement of fluorescence intensity, using V m-sensitive labels. Here we demonstrate a direct measurement technique based on determination of the cell’s ζ-potential, the electrical potential at the hydrodynamic shear plane, approximately 1 nm beyond the cell surface. We demonstrate this principle using excitable H9c2 cardiomyoblasts, measured in both polarised and depolarised states, before and after extracellular intervention to alter cell ion concentration. Given widespread availability of ζ-potential measurement apparatus (most typically in chemistry and materials science settings), this offers a new method of measuring V m without the need for fluorescence measurements or calibration curves.

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
The cell membrane potential (V m ) is a key biophysical component of cell functions such as cellular excitability, ion transport, cell signaling, secretion, motility, proliferation, energy storage, among others [1][2][3][4][5][6][7].It arises from the partitioning of ion species (principally Na + , K + and Cl − ) on either side of the cell membrane through the use of ion transporters [7,8].This normally maintains a non-equilibrium steady state, known as the resting membrane potential (RMP).However, specialised cells (commonly referred to as excitable cells) are able to use millisecond-scale transients in V m called action potentials (APs)-involving rapid depolarisation and repolarisation of V m -in order to perform vital cell functions.There are two cell types which exploit excitability to function; nerve cells (neurons) use APs to communicate with other neurons, whilst muscle cells use APs to control contraction, as well as to transmit synchronisation signals in the heart.A typical cardiac cell AP illustrated in figure 1.The AP comprises three distinct phases evoked by changes in the permeability of the membrane to distinct ion transports.The resting, depolarized, and repolarized membrane potentials form the basis of the overall AP waveform.The initial depolarization takes the form of a rapid upstroke driven by an ingress of positive ions inside the cell (often sodium); the repolarization phase is caused by a release of positive ions from the cytosol, mainly potassium, to reach the RMP [9,10].
Whilst measurement of V m is vital to understanding the function of excitable cells, it is difficult to measure using the current conventional methods, such as optical (fluorescent probes) or non-optical techniques (electrode based methods), as reviewed in [12].Initial experiments in the study of APs (electrophysiology) used giant squid axons, which are of sufficiently large diameter that wire electrodes could be inserted into the cytoplasm for the recording of V m [8].Subsequent experiments used smaller intracellular microelectrodes, which were later superseded by the Patch Clamp that allowed direct intracellular measurement through a nanoscopic transmembrane pore [13].However, although such methods produce precise data of single cells with high temporal resolution, they are slow (typically allowing analysis of a few cells per day), require a high degree of operator skill, and destroy the cells in the processes.
There exists a need for a simple method of membrane potential analysis on the benchtop.With rare exceptions (such as the proton ionophore method used to measure V m in red blood cells and mitochondria [14], or the use of radioactive traces such as rubidium [15]), such methods are almost exclusively based on the use of a V m -sensitive fluorescent dye [16].This usually takes the form of the production of a calibration curve using cells in media with different ion concentrations, and inferring the membrane potential from this, or using flow cytometry to produce relative measures of membrane potential rather than absolute measurements.Again, neither technique offers rapid measurement, and fluorescent labelling methods can be confounded when, for example, the drug is a substrate for a transporter molecule such as those found in drug-resistant A schematic of a cardiomyocyte action potential.From a resting potential of ca.−80 mV, the cell rapidly depolarises, exhibits a short spike of ca.+20 mV, then depolarises again to ca. 0 mV until the end of the action potential.Adapted from [11].
cancer cells [17].Consequently, there remains an unmet need for rapid and direct measurement of V m on the benchtop, labelfree and without a requirement for operator input.
An alternative approach to the measurement of V m is to examine the electrical potential in more easily accessible places.For example, cells (as with any object immersed in water) generate a small extracellular electrical potential which can be measured by observing the electrical mobility of the cell in an external electric field.This potential (commonly referred to as the zeta potential or ζ-potential) arises due to the cell's surface charge, and is manifested at the boundary of the cell's hydrodynamic plane of shear (or slip plane) [18].According to the standard model of ζ-potential, the voltage derives entirely from fixed surface charge, and it is commonly used in materials science to assess the stability of suspensoid solutions.However, it has been known since the 1970s that ζ-potential is modulated by V m , with evidence of a linear relationship being observed in cancer cells [19], bacteria [20], red blood cells [21], platelets [22] and slime molds [23].The mechanism by which this effect occurs has not been fully elucidated, but is understood to be due principally to capacitive coupling of V m across the electrical double layer outside the cell, as well as the membrane itself.
Whilst the phenomenon was described in the colloid science community, it was not widely understood or exploited in the wider biological community.However, the significance of the effect is profound; it suggests that cells can modulate their extracellular environment, including modulating interaction with ions [21], antibodies [22] and potentially, other cells.However, it also suggests the possibility that it might be exploited to directly observe V m .
Dukhin [24], determined that the relationship between V m and change in ζ-potential (∆ζ) are described by the equation: C m and C dl are the capacitances of the cell membrane and electrical double layer, respectively.Whilst the latter can be calculated for a known medium, the former can be difficult to measure independently of other parameters.Hughes et al [21] observed similar behaviour in red blood cells, and defined the parameter of proportionality Ξ = ∆ζ/V m .This was found to be around 0.32 for normal red blood cells, but lowered to 0.19 for cells treated with neuraminidase and below 0.1 for those treated with DMSO.
Given that the membrane capacitance of most cells is defined by the lipid composition of the membrane, which is again similar in most cell types, we propose that using this as a rule-of-thumb allows direct measurement of V m through the measurement of ζ-potential using an off-the-shelf ζ-potential instrument.Intriguingly, whilst prior research [19][20][21][22][23] demonstrated a clear relationship between ζ and V m , none considered using measurement of ζ as a proxy for V m ; this reflected the research priorities of the researchers who made the observations, who (being largely drawn from the colloid science community) were more interested in the phenomenon as a complex electrochemical system, and less interested in the technique as a simple, practical method of V m measurement.
The importance of measuring V m in live excitable cells stems from their ability to develop irregular AP patterns when expressing defective ion transporters.This is particularly true for cardiomyocytes that exhibit, even at normal state, various AP tracings that are tightly linked to the heterogeneity of their function across the myocardium [25].Nonetheless, mutations in the gene coding for specific ion transporters that are implicated in the onset and propagation of APs across cardiomyocytes are translated into abnormal V m coupled to cardiac dysfunction and disease.For example, abnormal electrical activity of the heart has been linked to mutations in the sodium, calcium, and potassium channels [26][27][28] depicting characterized diseases of the heart.
In this paper we present a ζ-potential-based technique of V m measurement, and use this to study a cardiomyocyte cell line in both polarised and depolarised states, with the results comparing favourably with those taken with considerably more complex methods.The widespread availability of instruments for ζ-potential measurement mean that this technique offers entirely new and rapid methods of V m measurement in cell biology.Using excitable H9c2 cardiomyoblasts, we evaluate the dynamics of the ζ-potential both at rest and in KCldepolarized cells.Our data establishes, for the first time, a link between the ζ-potential and V m in live cells which allows the direct, dynamic measurement of V m .This offers a new method for cell electrophysiology, and could be of medical significance to address diseases of excitable cells.

Cell preparation
H9c2 cardiomyoblasts were procured from ATCC and propagated according to the manufacturer's protocol.In brief, cells were cultured and maintained in 6-well plates in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin at 37 • C in a 5% CO 2 incubator.On the day of the experiments, cells were washed quickly with phosphate buffered saline and trypsined with 0.1% trypsin for 2 min at 37 • C. Detached cells were collected by centrifugation (700 × g for 5 min at room temperature) and were resuspended in 1 ml of DMEM, or isotonic solution, as described in the next section.All reagents were supplied by Gibco.

ζ-potential measurement
Cell solutions were measured using a Malvern Zetasizer Lab Series Blue and DTS1070 capillary cells (Malvern Panalytical Ltd, Malvern UK).Three technical repeats were performed per sample, and three biological repeats were carried out.Experiments were performed in three solutions; culture medium, and a mixture of culture medium and an isotonic solution comprising 8.5% w/v sucrose and 0.5% w/v dextrose, in 1:9 and 1:99 (culture medium:sugar solution) ratios.These are referred to here as 'high conductivity', 'medium conductivity' and 'low conductivity' solutions, respectively.The conductivity of the suspending solution was measured using the zetasizer at time of analysis.Experiments were performed at 25 • C.
Cells were measured in one of three states.In the first (referred to here as 'polarised'), suspended cells were measured without treatment.In the second state (herewith, 'depolarised') cells were stimulated into APs using a KCl solution by raising the net KCl concentration in the solution to 25 mM, causing the cells to depolarise V m above the threshold for AP initiation [29][30][31].In the third state, cells were permeabilised using Tween-20 at a final concentration of 2 µg ml −1 for 5 min.Permeabilisation was confirmed using Trypan Blue, as shown in figure 2.
Analysis provided results in two different formats.For medium and low conductivities, data were acquired as a histogram of cells at a range of measured potentials approximately 3 mV apart.The data set for each experiment was normalised against its maximum value, then combined and analysed using GraphPad Prism.For each curve, a Gaussian curve was fitted to determine the mean value for each condition.For the high conductivity condition, the Zetasizer is unable to produce a histogram in order to avoid solution heating, and instead presented only the modal value of ζ-potential for the sample.

ζ-potential depolarises when cells depolarise
We first investigated the effect of KCl-induced depolarisation on the ζ-potential of H9c2 cells.We found that under normal resting conditions, the cells exhibited a Gaussian distribution in ζ-potential whose mean varied with medium conductivity, as shown in figure 3, but which was typically in the range of -2-30 mV.The addition of KCl in all cases caused nearinstant depolarisation to typically −2 mV.Depolarisation of the sample was so rapid that it could only be observed by setting equilibration time to one second and pre-conditioning the sample prior to the addition of KCl.This still demonstrated the rapid shift shown in figure 4. We then investigated cells following permeabilisation using Tween-20, which induces nanoscopic pores in the membrane and allows equilibration of intracellular and extracellular spaces, which abolishes V m (also shown in figure 3); we hypothesised that this reduces the cell membrane potential to 0 mV, such that the observed ζpotential is dependent only on the surface charge of the cell and allowing the determination of ζ', the potential arising only from fixed surface charge.When this was performed, the permeabilised cells exhibited a mean ζ-potential that was depolarised with respect to the resting state cells, but was more polarised than those exposed to KCl.This suggests that in light of V m -mediated alteration in ζ-potential, the resting state is negatively polarised and that the KCl-depolarised state reflects a mild positive polarisation in V m , in line with expectation.

Effect of medium ion concentration
Measurements in high conductivity solutions were only able to use reported modal values of ζ-potential; for the given resolution this suggests a quantization error of ±1.6 mV.Averaged across three biological repeats we observed a ζ-potential of −18.5 mV in polarized cells, which reduced to −7.28 mV after depolarization.When cells were permeabilised, this increased the ζ-potential to −9.3 mV.
Similar behaviour was observed in medium and low conductivity solutions.In the medium conductivity solutions, cells in the initial polarised state yielded mean ζ-potential after Gaussian fitting of −17.5 mV, reducing to −2.76 mV after  KCl treatment, and −4.94 mV after permeabilisation.There was a statistically significant difference between polarised, depolarised and permeabilised cells, with p < 0.0001.In low conductivity solutions, values were similar to those observed in medium conductivity solutions; mean ζ-potential for polarised cells was −21.8 mV, depolarised −1.77 mV, and permeabilised −1.91 mV (Data presented in table 1).
As would be anticipated, addition of KCl altered the medium conductivity; in high conductivity solution, the conductivity rose from a mean of 1570 ± 90 mSm −1 to 1880 ± 120 mSm −1 after addition of KCl, and further to 1910 ± 100 mSm −1 after permeabilisation (which results in the cytoplasm equilibrating with the suspending solution).The medium conductivity solution was measured at 144 ± 33 mSm −1 , rising to 228 ± 41 mSm −1 after KCl stimulation and 417 ± 85 mSm −1 after permeabilisation; the lower initial conductivity means that the effect of adding KCl has a more profound impact on solution conductivity, which is exacerbated by cytoplasm equilibration after permeabilisation.Finally, the low conductivity solution was measured at 14.4 ± 0.3 mSm −1 in normal state, 30.6 ±1 mSm −1 after KCl, and 46 ± 2.1 mSm −1 after permeabilisation.The lower increment in the low conductivity case after KCl treatment (16.2 mSm −1 ) when compared to the increment for medium conductivity solutions (189 mSm −1 ) suggests that the rise is due to cells effluxing ions as part of the depolarisation process, rather than due to the KCl itself, since the same quantity of KCl was added in both cases, but the cell concentration in the low conductivity solution is substantially lower.

Discussion
It is evident from figure 4 that the ζ-potential of cardiomyoblasts responds near-instantly to depolarisation, and that the measured value of ζ contains contributions of both V m and the surface potential.In both high and medium conductivity solutions, the ζ-potential of permeabilised cells lay between the polarised and depolarised ζ-potential measurements; if we consider the permeabilised cells as having equilibrated with the medium, we can regard V m as 0 mV and that the ζ-potential is a reflection of the cell surface alone.If we subsequently subtract this value from the polarised and depolarised ζ-potential values, we can calculate the contribution from V m , and the two contribution to ζ of the resting potential, when polarised (ζ pol ) or depolarised (ζ depol ).We can then consider how this relates to the corresponding resting potentials V pol and V depol .
Considering cells in high conductivity solution first, we found mean values of ζ pol = −9.2mV and ζ depol = + 2.12 mV.The normal resting potential of cardiomyocytes is approximately −80 mV, suggesting a value of Ξ (the proportion of V m that alters ζ) of 11.5 %.This in turn suggests that V depol = 18.4 mV, which is equal to the peak value of V m observed at the start of the AP in figure 1.
Moving to the medium conductivity solution, we find that ζ pol = − 12.6 mV and ζ depol = −2.2mV.If we apply the same value of Ξ this yields V pol of −109.6 mV and V depol of +19.1 mV.This suggests that the depolarisation spike remains at the same potential but that the resting state has become hyperpolarised compared to that in 'normal' media.This is in line with observations of cardiomyocytes in low-K+ extracellular medium, a condition called hypokalaemia [32].In this condition, the RMP hyperpolarises to ca. −105 mV [33], which is in line with our observations.The addition of KCl to induce hyperpolarisation means that the cells in depolarised state are not hypokalaemic, and hence the values of V depol described above are comparable.
In the final, lowest conductivity solution we observed greater differences from expected behaviour; the cells showed a ζ pol of −19.9 mV, equating to a V pol of −173 mV.Conversely, the cells exhibited a ζ depol of 0.4 mV, equating to a V depol of only 1.2 mV.The value of V m suggests that the AP has missed the initial spike altogether, and instead has depolarised to the second section of the AP where V m is at or near zero.This is again in line with expectations; the initial phase of the AP is driven by sodium influx, and in the low conductivity case, there is insufficient Na + in the medium to generate the requisite spike.The value of V pol is even more hyperpolarised than that observed in the medium conductivity case; this may be indicative of significant cell dysregulation, but may also indicate a change in the membrane capacitances of equation ( 1) resulting in a different, higher value of Ξ.Indeed, one would expect a dependence between medium ion concentration and Ξ, since the value of C dl in equation ( 1) should theoretically vary with the double layer thickness, which has an inverse square root relationship to ion concentration.The high and medium conductivity solutions here vary in solution ion concentration by an order of magnitude, which should result in a change of approximately ×3 in C dl ; however, the results of analysis for V pol and V depol in both cases align near-exactly to published electrophysiology when Ξ = 0.115.This suggests one of two possibilities; either that C dl is invariant, or that both the capacitances (C dl and C m ) scale with medium concentration.However, a constant value is consistent with observations in other cell types [21,22].Given the complexity of the electrical properties of double layers and their effects in cells that have been revealed recently [21,22,34], it appears likely that Ξ does remain constant but that the underlying mechanism may be more complex than predicted by equation (1); that said, it is possible that the low conductivity solution may have caused an increase in Ξ which would explain the elevated estimate of V pol .For example, a value of Ξ = 0.18 would be realistic given known values of V m in hypokalaemia, and remains in line with observed values in red blood cells for example [21].

Conclusion
For the first time, we have demonstrated that measurement of cellular ζ-potential is a useful tool for the measurement of cellular membrane potential.Using excitable H9c2 cardiomyoblasts in three ionic solutions and in three states (polarised, depolarised and permeabilised) we have measured values of membrane potential in line with published values by conventional methods.This work suggests that ζ-potential measurement presents a rapid, low-cost alternative to single-cell measures such as Patch Clamp, that allows ensemble measurements, requires no specialist skills, and is free from operator bias.This offers significant benefits for widening access to electrophysiological measurement, offering tremendous benefits both for laboratory cell biology and for novel diagnostic measurement of excitable cells.

Figure 1 .
Figure 1.A schematic of a cardiomyocyte action potential.From a resting potential of ca.−80 mV, the cell rapidly depolarises, exhibits a short spike of ca.+20 mV, then depolarises again to ca. 0 mV until the end of the action potential.Adapted from[11].

Figure 2 .
Figure 2. H9c2 cardiomyoblasts trypsinized and stained with trypan blue, without (Left panel) and with (right panel) Tween (2 g ml −1 for 5 min).Accumulation of trypan blue in the cytosol of the cell is noticed in permeabilized cells (right panel) indicating perforated plasma membrane.Images were taken using an Olympus CKX53 microscope and ×20 objective.

Figure 3 .
Figure 3.The mean ζ-potential distribution for cells in medium conductivity.

Figure 4 .
Figure 4. Zetasizer traces taken at 2 s intervals (darkest first), beginning 25 s after the addition of KCl at the cuvette.The ζpotential peak migrates to its final position 33 s after the addition of

Table 1 .
Mean ζ-potential values H9C2 cells in the three medium conductivities, in their resting state (polarised), and following KCl treatment to stimulate action potentials (depolarised) and subsequent treatment with Tween-20 (permeabilised).